SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ? NUMBER 33 Great Basin Aquatic Systems History Robert Her shier, DavidB. Madsen, and Donald R. Currey EDITORS fSS'JIH) DEC 112002 SMRHSONIAN INSTITUTION Smithsonian Institution Press Washington, D.C. 2002 ABSTRACT Hershler, Robert, David B. Madsen, and Donald R. Currey, editors. Great Basin Aquatic Sys- tems History. Smithsonian Contributions to the Earth Sciences, number 33,405 pages, 190 fig- ures, 21 tables, 2002.?The 14 papers collected herein treat diverse aspects of the aquatic history of the Great Basin of the western United States and collectively attempt to summarize and integrate portions of the vast body of new information on this subject that has been acquired since the last such compilation was published in 1948. In the first section, four papers (Lowenstein, Negrini, Reheis et al., Sack) focus on physical aspects of Great Basin paleolake histories, whereas a fifth paper (Oviatt) summarizes the contributions to the study of Bonne- ville Basin lacustrine history made by two early giants of the field, Grove Karl Gilbert and Ernst Antevs. In the second section, four papers synthesize perspectives on Great Basin aquatic history provided by diatoms and ostracods (Bradbury and Forester), fishes (Smith et al.), aquatic insects (Polhemus and Polhemus), and aquatic snails (Hershler and Sada), whereas a fifth (Sada and Vinyard) summarizes the conservation status of the diverse aquatic biota that is endemic to the region. In the final section, three papers integrate terrestrial biotic evidence per- taining to Great Basin aquatic history derived from pollen from cores (Davis), floristics (Wigand and Rhode), and the mammal record (Grayson), whereas a fourth (Madsen) examines the relationship between Great Basin lakes and human inhabitants of the region. Although diverse in scope and topic, the papers in this volume are nonetheless linked by an appreciation that integration of geological, biological, and anthropological evidence is a necessary and fun- damental key to a mature understanding of Great Basin aquatic systems history. OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recorded in the Institution's annual report, Annals of the Smithsonian Institution. SERIES COVER DESIGN: Volcanic eruption at the island of Krakatau in 1883. Library of Congress Cataloging-in-Publication Data Great Basin aquatic systems history / Robert Hershler, David B. Madsen, and Donald R. Curry, editors. p. cm. ? (Smithsonian contributions to the earth sciences ; no. 33) Includes bibliographical references. 1. Limnology?Great Basin. 2. Paleolimnology?Great Basin. I. Hershler, Robert. II. Madsen, David B. III. Currey, Donald R. IV. Series. QE1 .S227 no. 33 [QH104.5.G68] 550 s?dc21 [551.48'0979] 2002026685 ? The paper used in this publication meets the minimum requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48?1984. Contents Page Introduction, by David B. Madsen, Robert Hershler, and Donald R. Currey 1 LAKE HISTORIES Pluvial Lake Sizes in the Northwestern Great Basin throughout the Quaternary Period, by Robert M. Negrini 11 Pliocene to Middle Pleistocene Lakes in the Western Great Basin: Ages and Connec- tions, by Marith C. Reheis, Andrei M. Sarna-Wojcicki, Richard L. Reynolds, Charles A. Repenning, and Martin D. Mifflin 53 Pleistocene Lakes and Paleoclimates (0 to 200 Ka) in Death Valley, California, by Tim K. Lowenstein 109 Bonneville Basin Lacustrine History: The Contributions of G.K. Gilbert and Ernst Antevs, by Charles G. Oviatt 121 Fluvial Linkages in Lake Bonneville Subbasin Integration, by Dorothy Sack 129 AQUATIC BIOTIC PERSPECTIVES Environment and Paleolimnology of Owens Lake, California: A Record of Climate and Hydrology for the Last 50,000 Years, by J. Platt Bradbury and R.M. Forester 145 Biogeography and Timing of Evolutionary Events among Great Basin Fishes, by G.R. Smith, T.E. Dowling, K.W. Gobalet, T. Lugaski, D.K. Shiozawa, and R.P. Evans 175 Basins and Ranges: The Biogeography of Aquatic True Bugs (Insecta: Heteroptera) in the Great Basin, by Dan A. Polhemus and John T. Polhemus 235 Biogeography of Great Basin Aquatic Snails of the Genus Pyrgulopsis, by Robert Hershler and Donald W. Sada 255 Anthropogenic Changes in Biogeography of Great Basin Aquatic Biota, by Donald W. Sada and Gary L. Vinyard 277 NON-AQUATIC BIOTIC AND HYDROCLIMATIC PERSPECTIVES Late Neogene Environmental History of the Northern Bonneville Basin: A Review of Palynological Studies, by Owen K. Davis 295 Great Basin Vegetation History and Aquatic Systems: The Last 150,000 Years, by Peter E. Wigand and David Rhode 309 Great Basin Mammals and Late Quaternary Climate History, by Donald K. Gray- son 369 Great Basin Peoples and Late Quaternary Aquatic History, by David B. Madsen ... 387 m Introduction David B. Madsen, Robert Her shier, andDonaldR. Currey It has long been known that the huge, arid region of internal drainage in western North America known as the Great Basin contained a series of large, intermontane lakes during wetter, or pluvial, periods of the late Cenozoic. During the last 100 years numerous geologists and biologists, including many preemi- nent scientists beginning with Grove Karl Gilbert (1843-1918) and Israel Cook Russell (1852-1906), have studied the extent, integration, and contemporaneity of these lakes, and the rela- tion between regional hydrographic history and biogeography. The first attempt to synthesize this information was in June 1942, when the Western Division of the American Society of Ichthyologists and Herpetologists met in Salt Lake City, Utah. The highlight of the meeting was a symposium on the glacial and postglacial history of the Great Basin, which was jointly arranged with the Pacific Division of the American Association for the Advancement of Science (Croker, 1942). Papers pre- sented at the meeting included summaries of the geological de- velopment of the Great Basin, pluvial lake distribution and per- tinent ichthyological evidence, and paleoclimates, which were later collected in The Great Basin, with Emphasis on Glacial and Postglacial Times (Antevs, 1948; Blackwelder, 1948; Hubbs and Miller, 1948). This modest volume has stood as a landmark in the study of what Grayson (1993:xvi) referred to as the "natural prehistory" of the Great Basin. Immense strides have been made in the study of Great Basin aquatic history during the last 50 years. On the physical side, our understanding of geological evolution of the Great Basin has been revitalized in the wake of modern concepts of moun- tain building, tectonics, and volcanology. Geochronology has provided many new dating methods to aid interpretation of key geologic events. New methodologies also have enabled much more precise delineation and dating of ancient lake beds and better discernment as to whether such data represent records of long-lived lakes or ephemeral water bodies (playas). Climatic David B. Madsen, Environmental Sciences, Utah Geological Survey, Salt Lake City, Utah 84114, USA. Robert Hershler, Department of Sys- tematic Biology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0118, USA. DonaldR. Currey, Department of Geography, University of Utah, Salt Lake City, Utah 84112, USA. modeling has developed into an important field that further aids elucidation of pattern and sequence of aquatic-system de- velopment. On the biological side, evidence derived from fishes (both living and fossil) has been expanded and refined, and other aquatic biota now has been similarly studied. The coupling of biogeographic analysis with phylogenetics has pro- vided explicit, testable methods of inferring historical relation- ships among geographic areas (drainages). The blossoming field of phylogeography (Bermingham and Moritz, 1998) of- fers additional tools for fruitful analysis at the population level, although this has been little applied to aquatic biota of the Great Basin (but see Duvernell and Turner, 1998). The purpose of this volume is to assemble and integrate this large body of relevant data from the geological and biological schools as it pertains to the history of aquatic systems in the Great Basin. This is hardly a new concept, but instead it reiter- ates the spirit of Hubbs and Miller (1948:120): Even greater advances should ensue, when the biological and geological ap- proaches are more definitely integrated. We eagerly look forward to the time when we may join with the glacial geologist to journey, helping hand in hand, across the desert mountains and flats of the Great Basin, together attempting to unlock some of the doors that are yet closed to us. This volume is a product of the Great Basin Aquatic System History (GBASH) Symposium, which was convened in Salt Lake City at the University of Utah from 17 September to 21 September 1997 to celebrate the 50th anniversary of the publi- cation of papers from the original 1942 symposium. More than 75 individuals attended the GBASH conference, including in- vited speakers, poster presenters, representatives of land man- agement agencies, and the general public. Participants enjoyed a highly structured program designed to summarize recent ad- vances in Great Basin natural history in an integrated, interdis- ciplinary fashion. Presenters included an array of geologists, climatologists, botanists, zoologists, and anthropologists under the general direction of conference chair Donald R. Currey. The presentation format for the 25 invited papers followed the same tripartite structure we employ herein. Fourteen of those papers were revised subsequent to the symposium and are col- lected in this volume. Together these papers complement the more generalized treatments of Great Basin aquatic history that are currently available. SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 1.?The limits of the Great Basin, as defined by areas of internal drain- age (after King, 1986). Defining the Great Basin The Great Basin is a vast complex of internal drainage sys- tems lying in interior western North America (Figure 1). These systems are composed primarily of high, north-to-south trend- ing mountains and long, narrow valleys into which mountain streams and rivers drain. Depending upon how it is defined (and it can be defined in a number of ways), the Great Basin extends as far north as 44? north latitude, as far south as 33? north latitude, as far east as 110? west longitude, and as far west as 121? west longitude. Although the many definitions of the Great Basin are appropriate for specific purposes, none ex- ceeds that of John C. Fremont's (1845:175) original description in terms of simple utility: The Great Basin?a term I apply to the intermediate region between the Rocky mountains and the next range [the Sierra Nevada and the Cascade Range], con- taining many lakes, with their own system of rivers and creeks, (of which the Great Salt is the principal,) and which have no connexion with the ocean, or the great rivers which flow into it. A number of general accounts reviewing the natural history of the Great Basin are available (e.g., Houghton, 1976; McPhee, 1981; Trimble, 1989; and particularly, Grayson, 1993), and there is also a variety of extensive treatments of specific topics (Lanner, 1983; Ryser, 1985; Fiero, 1986; Moz- ingo, 1987; and Sigler and Sigler, 1987). As a result, we have not attempted to produce yet another generalized description of the Great Basin herein. Furthermore, because a variety of bio- logical, geomorphological, and hydrological topics are covered in the papers that follow, it is important to recognize that the physiographic, hydrographic, and floristic definitions of the Great Basin employed in these papers are somewhat different. Physiographically, the Great Basin is part of the Basin and Range Province (Hunt, 1967) that, although coterminous with the hydrographic Great Basin in the north, extends into south- ern Arizona and New Mexico and covers much of northern Mexico (Figure 2). The fault-block mountains that characterize the province began forming during the middle Tertiary, approx- imately 40 million years (My) before present (B.P.) (King, 1977). This system of mountains and valleys was formed as part of an ongoing extensional process related to plate tectonics (Zoback et al., 1981), and the Great Basin continues to expand as the Pacific Plate shears off the western margin of the North American plate (Stewart, 1971; Fiero, 1986). This east-west expansion has created a series of north-to-south trending fault- block mountains separated by long, narrow graben and/or half- graben valleys filled with Quaternary alluvium. The larger val- leys exceed 100 miles (161 km) in length and 30 miles (48 km) in width. The Great Basin is bounded on the west and east by the higher ranges of the Sierra Nevada and the western Rocky FIGURE 2.?The limits of the Basin and Range Province (after Stewart, 1971; Fiero, 1986). NUMBER 33 FIGURE 3.?The limits of the floristic Great Basin (after Cronquist et al., 1972). Mountains, respectively, but many of the smaller interior ranges reach almost as high because the Great Basin is dome- shaped in cross section. Many ranges reach 12,000-13,000 ft (3660-3960 m) and extend 5000-6000 ft (1524-1829 m) above valley floors. The biotic Great Basin is defined primarily by the distribu- tion of plant communities rather than by the animals that are dependent upon them (Gleason and Cronquist, 1964). The most widely employed floristic definition is that of Cronquist et al. (1972) (Figure 3). In this interpretation, the floristic Great Ba- sin extends from the Sierra Nevada on the west to the Wasatch Front ranges of Utah on the east (Cronquist et al., 1972; Os- mond et al., 1990). The eastern boundary is not as sharp as that on the west, however, and many early treatments of Great Ba- sin flora (e.g., Shreve, 1942) included much of the northern Colorado Plateau in the floristic Great Basin. Plant communi- ties are dominated by sagebrush (Artemisia tridentata), shad- scale (Atriplex confertifolia), and saltbush (Atriplex canescens) on valley floors, juniper (Juniperous osteosperma) and single- leaf pinon (Pinus monophyla) on lower mountain slopes in the central and southern Great Basin, and white fir (Abies con- color), Engelmann spruce (Picea emglemannii), blue spruce (Picea pungens), Douglas fir (Pseudotsuga menziesii), and quaking aspen (Populus tremuloides) at higher elevations. These communities extend north into the Snake River Plain, and the floristic Great Basin, as a result, does not conform well with the Great Basin as defined either physiographically or hy- drographically. There is similar terminological confusion re- garding the southern margin of the floristic Great Basin be- cause this boundary is defined mostly by the appearance of creosote bush (Larrea tridentata), yucca {Yucca brevifolia), and other plants of the Mojave Desert south of about 37? north latitude. When either a hydrographic or physiographic defini- tion of the Great Basin is employed, the Great Basin actually contains three of the four biotic North American deserts (Fig- ure 4). The Great Basin can be described most definitively in terms of a series of contiguous interior drainage systems. Although subject to some debate over whether or not a few individual drainage basins do indeed have internal drainage (e.g., Mifflin and Wheat, 1979; Smith and Street-Perrott, 1983; King, 1986), this hydrographic definition is the most detailed of those avail- able because drainage basin boundaries can be identified with precision. Several hydrographic definitions have been pro- duced, which differ because of the varying number of drainage NM FIGURE 4.?The three biotically defined North American deserts with the hydrographically defined Great Basin (after Shreve, 1942; Spaulding et al., 1983). SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 5.?Hydrographic basins within the Great Basin (after G.Q. King, unpublished data). basins included. A hydrographic definition of the Great Basin is most useful herein because our interest is in Great Basin aquatic systems, and we employ that of King (1986), who iden- tified 187 separate drainage basins within the Great Basin as a whole (Figure 5). Many of these individual drainages coalesced to form larger drainage units during cooler and wetter periods when lake lev- els rose and overflowed into adjoining basins. The largest of these, by far, are the Lahontan Basin on the west and the Bonneville Basin on the east, where the relatively more mas- sive Wasatch Mountains and Sierra Nevada, respectively, cap- ture substantial amounts of precipitation. During warm and dry intervals, the rivers of the Lahontan Basin, such as the Truckee, the Humboldt, the Carson, and the Walker Rivers, flowed into separate basins and form individual lakes, such as Pyramid and Walker Lakes. Under higher precipitation and lower evaporation regimes, however, these basins filled differ- entially, forming variable combinations of basin systems as lake waters rose (Davis, 1982). At their maximum, these wa- ters formed a single lake, covering nearly 8670 square miles (22,440 km2), during the last glacial cycle. The situation is similar in the eastern Great Basin where the Sevier and Great NUMBER 33 V * i 200 km FIGURE 6.?Large lakes within the Great Basin during the height of the last pluvial period (18-16 Ka) (after Wil- liams and Bedinger, 1984). Salt Lake subbasins at times have been either completely sepa- rate, connected by an overflow channel, or coalesced into a single lake. At its height, Lake Bonneville extended more than 500 miles (805 km) from north to south, covered more than 19,900 square miles (51,640 km2), and overflowed into the Snake River and the Columbia River drainage system. At such times the Bonneville Basin was not, hydrographically speak- ing, part of the Great Basin at all. A number of other basins within the Great Basin also over- flowed during wet periods, but they formed a series of inter- connected lakes rather than a single, large lake. The most extensive chain of lakes and basins were formed in the south- western Great Basin, where Mono, Owens, China, Searles, and Panamint Lakes (and other smaller lake basins) were connected at the peak of pluvial periods by a series of overflow channels that eventually extended into Death Valley (Smith and Street- Perrott, 1983). A number of shorter, interconnected lake sys- tems, such as the Malheur Lake-Harney Lake chain in Oregon, were found in other parts of the Great Basin. Most of the basins in the central and southern Great Basin contained only isolated lakes during moist intervals because the mountains that form their margins are limited both in extent and elevation; however, several of these isolated lakes, such as Pleistocene Lake Frank- lin in central Nevada, were relatively large because of the high elevation of both the basin floor and the surrounding mountain crests in the central Great Basin. More than 80 percent of the 187 Great Basin drainages contain playas, and most of these basins likely supported lakes at least seasonally during the peak of the last pluvial interval (e.g., King, 1986; Williams and Be- dinger, 1984) (Figure 6). SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 7.?Major subregions of the Great Basin (after Houghton, 1976). Major subdivisions within the Great Basin have been defined in a number of ways, much like the basin as a whole. These subdivisions, also like the basin as a whole, differ markedly de- pending upon the criteria that are employed in making the divi- sions. Cronquist et al. (1972) divided the Great Basin floristic province into 15 separate sections. This number can be reduced considerably because parts of seven of these sections lie well outside the Great Basin, as defined either hydrographically or physiographically, on the Snake River Plain and the Colorado Plateau. On the other hand, they include none of the plant com- munities of the Mojave or Sonoran Deserts in the southern hy- drographic Great Basin. Many of these floristic subdivisions cross drainage system boundaries. Both Hunt (1974) and Houghton (1976) divided the Great Basin into five major sub- regions based upon a combination of hydrographic and physio- graphic factors (Figure 7). These subdivisions are in relatively close agreement, differing primarily in the labels that are ap- plied to them, and they provide a far simpler and more coherent division of the Great Basin than that of Cronquist et al. (1972). These include the Bonneville Basin to the east of the central basin area, the Lahontan Basin to the west of the central area, a Northwest Lakes (or Lava and Lakes) area to the north of the Lahontan Basin, and a Southern or Death Valley system to the south of the central area. The confusion surrounding various biotic, physiographic, and hydrographic definitions of the Great Basin is symptomatic of the difficulties in understanding the history of the Great Ba- sin aquatic systems. The Great Basin is related in many differ- ent ways to the areas around it, and it is difficult to tease out how changes in those relationships through time have differen- tially affected Great Basin biomes. For example, at times the Bonneville Basin has been hydrographically part of the Great Basin, and at other times it has been part of the Columbia River drainage system. As a result, the aquatic biota of this basin has a history that is both richer and more complex than that of many other Great Basin drainages. These different definitions are also illustrative of the difficulties in making comparisons across drainages within the Great Basin. Lakes in the south, for example, are surrounded by vegetation that is so markedly dif- ferent from that surrounding lakes in the north that floristically they are not even in the Great Basin. Yet it is these differences that lie at the heart of the Great Basin and that, by extension, lie at the heart of the papers in this volume. The Great Basin is at once both obvious and enigmatic, making attempts to explore and interpret its natural history both irritating and rewarding. Chronological Aspects of Great Basin Aquatic Systems A Great Basin aquatic systems history must include an un- derstanding of the changing climatic conditions that have largely structured fluctuations in the size of Great Basin lakes and the development of biotic communities in the region. These long-term climatic changes are related most directly to alterations in the relative distance between the earth and the sun and, hence, vary with perturbations in the earth's orbit and axis. These climatic changes operate on a number of different scales, ranging from about 20,000 to 100,000 or more years, and, depending upon where the earth is in each of these Mi- lankovitch cycles, they can reinforce or counteract each other. For example, the wobble in the earth's axis alternately moder- ates and exacerbates the change in the amount of energy re- ceived from the sun as the earth's orbit changes on a longer scale. For Late Quaternary climates, the 19,000- and 23,000- year precession cycles and the 41,000-year obliquity cycles are particularly important (e.g., Grootes and Stuiver, 1997). Cli- matic change on a millennial scale of 1000-2000 years also has been recognized in proxy records from ice cores and from cores taken from the ocean and large continental lakes. As yet, however, it is not clear how short-term Dansgaard-Oeschger cyclical variations may relate to the celestial mechanics that drive the longer cycles. All of these cycles appear to produce globally linked climatic changes, although the nature of these changes varies spatially. The longer-term cycles defined in the oxygen isotope record of deep-sea cores are also readily recognized in the fluctuation histories of many Great Basin lakes (e.g., Smith and Bischoff, 1997; Benson et al., 1998). Several of the papers in this volume explore the effects of these longer-term climatic cycles on Great Basin lakes, and there is no need to discuss the issue here. Less well known, however, is the relationship between millen- nial-scale climatic changes and the fluctuation histories of NUMBER 33 Great Basin lakes. Most dramatically, it is becoming increas- ingly clear that the transitions between the steady-state condi- tions that characterize these millennial-scale cycles are rela- tively abrupt, often on the order of a decade or less, and certainly within the lifetime of a human being. Evidence from Greenland ice cores and from cores taken from the North At- lantic Ocean indicates that these cycles assumed a characteris- tic form during the glacial period. They were initiated by a rapid rise in temperature of 5?-8? C in a matter of decades, gradually returned to moderate conditions over the course of about 1000 years, and ended with a rapid return to very cold temperatures prior to the start of a warming event that initiated a new cycle. The amplitude of the cycles appears to have been much smaller during the Holocene (on the order of 1??3? C), but the pattern seems to have been relatively similar. Twenty- four cycles have been identified in the last glacial period, and nine have been recognized in these detailed ice core records during the 12,000 years of recorded human occupation in the Great Basin (e.g., Bond et al., 1997). These millennial-scale cy- cles extend back at least 100,000 years (e.g., Brooks et al., 1996) and appear to be part of the entire Pleistocene record. Although it is becoming evident that change in the water budget of many Great Basin lakes is correlated with millennial- scale cycles evident in the ice core record (e.g., Benson et al., 1997; Oviatt, 1997), all of the late glacial and Holocene cycles have yet to be recognized locally. This is probably due to a combination of factors, such as the relatively coarse nature of local environmental records and the presence of other, longer cycles with periodicities of 6100 and 11,100 years that may ob- scure local manifestation of millennial-scale fluctuations (e.g., Mayewski et al., 1997). Once detailed and sensitive records, such as those from tree rings and lake varves, are available, many, but perhaps not all, of these cycles may become evident. Regardless, it is as important to incorporate a notion of time into an understanding of Great Basin aquatic systems as it is to understand the spatial dimensions that define them. At times Great Basin valleys were filled with water, and at other times they became completely desiccated; lakes came and went; hy- drological systems connected and disconnected. These cyclical changes, together with the basin and range topography, com- prise the basic structure that underlies Great Basin aquatic sys- tems. Organization of the Volume The first section of this volume includes papers treating the physical history of Great Basin paleolakes. Negrini's compre- hensive summary of pluvial lake histories in the northwestern Great Basin (and evaluation of Antevs' model for Quaternary climate change) will prove useful not only to geologists but also to those seeking information relative to aquatic biogeogra- phy. Millennial-scale oscillations are evident in this sequence, but it is apparently slightly out of phase with those of the east- ern Great Basin lakes. Similarly, the paper by Reheis et al., which carefully documents new evidence supporting an en- larged Lake Lahontan drainage system during the late Neo- gene, also fits well the cross-disciplinary focus of this sympo- sium. They suggest that lake levels were as much as 70 m higher during the early to middle Pleistocene than during the late Pleistocene highstand?a sequence that is also quite differ- ent from that found in the eastern Great Basin. The highly dy- namic nature of Quaternary paleoclimates and associated hy- drological cycles in the Great Basin is emphasized by Lowenstein, who reviews pertinent high-resolution data from lake beds of the Death Valley hydrographic system. Lowen- stein makes it clear that rapid climate-change events evident in the Greenland ice core and in North Atlantic core records also are present in Great Basin lake histories. Oviatt summarizes the lasting contributions of Gilbert and Antevs, two of the contrib- utors to the 1948 Great Basin volume, to the study of aquatic history of the Bonneville Basin. Sack marshals geomorphic and hydrologic evidence to unravel changes in subbasin relation- ships within a component of the Bonneville Basin, again indi- cating that complexity is a hallmark of Great Basin aquatic sys- tems history. The section on aquatic biology focuses on the latest contribu- tions that biotic data have made to our understanding of Great Basin aquatic systems history, and it is intended as a supple- ment to the earlier papers collected by Harper and Reveal (1978). Regional aquatic biota includes various taxa that dis- perse readily, speciate slowly, and whose distribution and oc- currence more reflects ecological and climatic conditions than history of drainage interconnections. The fossil record of such organisms thus can provide a detailed picture of environmental history. In the paper by Bradbury and Forester, high-resolution paleontologic data for two such groups from lake bed cores (di- atoms and ostracods) are integrated with other biotic and phys- ical evidence to infer the late Pleistocene history of Owens Lake within the context of Sierra Nevada glaciation and re- gional climatic history. On the other end of the spectrum, speci- ose groups, whose members are often locally endemic, can pro- vide important clues of drainage interconnections, although this utility is strongly dependent upon the availability of robust phylogenetic hypotheses as the basis for analysis. Fishes have long paved the way in this arena, and the paper by Smith et al. represents a significant advance from the classical treatment of Hubbs and Miller (1948). In this paper phylogenies of several groups, based on mitochondrial DNA sequences, together with a comprehensive review of fossil evidence, are used to refine previous interpretations of historic drainage relationships and rates of fish evolution. Fish evidence suggests the importance of old drainage relationships and vicariant events that preceded the late Pleistocene pluvial period, as presaged by Minckley et al. (1986). Papers on aquatic snails (Hershler and Sada) and in- sects (Polhemus and Polhemus) suggest additional ancient drainage relationships that predate modern topography, al- though these hypotheses await testing within the context of phylogenetic analysis. Great Basin biogeography has changed SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES (extirpation of biota, introduction of exotic taxa, etc.) as a re- grettable consequence of anthropogenic activities relating to water development in this arid region. It is appropriate to ask whether, using the phraseology of Kornfield and Echelle (1984:251), attention to "tending of the flock" is currently suf- ficient to ensure long-term persistence of the unique aquatic ec- osystems in the Great Basin. The chapter by Sada and Vinyard summarizes the conservation status of aquatic biota endemic to the Great Basin. One cannot but conclude that conflicting de- velopment and conservation needs can be better balanced. The final section on nonaquatic biology and hydroclimatic perspectives is focused on the climatic mechanisms that drive changes in Great Basin lake levels and on the biological com- munities that surround the Great Basin lakes. Because the his- tory of these biotic systems is, in turn, mostly determined by fluctuations in the size of the lakes, these changes are necessar- ily an integral part of Great Basin aquatic systems history. The papers in this section vary between generalized accounts appli- cable to the Great Basin as a whole and specific accounts re- lated to a single lake-basin history, and between consideration of very long-term changes and of relatively short-term records. As such, these papers reflect the breadth of studies involving Great Basin ecosystems that are currently underway. Davis fo- cuses on the long-term vegetational history of the Bonneville Basin, as delineated from analysis of the pollen in deep cores from the Great Salt Lake. He suggests that there was a gradual evolution of desert flora during the last five million years that was punctuated by a number of noteworthy events. Pluvial-in- terpluvial cycles began after about 750,000 years ago, and Davis suggests that an increase in pollen concentration indi- cates the Bear River was diverted into the Bonneville Basin shortly before about 300,000 years ago. Wigand and Rhode provide a detailed compilation of Great Basin floristic changes during the last 150,000 years, focusing primarily on the last 15,000-20,000 years. Using an extensive array of data derived from pollen, midden, and tree-ring records from the northwest- ern, southern, and northeastern Great Basin, they suggest that Great Basin ecosystems have not evolved in a holistic fashion. Rather, Great Basin floristic communities are a continually changing mosaic of individual organisms, and Wigand and Rhode suggest that modern and prehistoric communities are not appropriate analogues for one another. Grayson uses a de- tailed, small-mammal record derived from the collecting activ- ity of owls to assess the nature of climatic change during the Pleistocene-Holocene transition and the early Holocene in the Great Salt Lake basin. He suggests that the early Holocene was both cooler and wetter than at present, in contrast to climate models that indicate otherwise. He too notes that faunal com- munities are composed of species that react differently to changing climatic conditions and finds support for species- richness response models that predict increasing diversity with increasing precipitation. Finally, Madsen reviews the relation- ship between Great Basin lakes and Great Basin peoples. He notes that this relationship is mostly structured by the dichot- omy between the relatively rich and productive riverine and lacustrine ecosystems and the relatively dry and less productive landscapes that surround them. As a result, changes in the size and distribution of lakes have been critical components in the development of prehistoric human society in the Great Basin. Acknowledgments We thank the authors for their timely submissions and for the patience they displayed with delays in the editorial process. We also thank the many colleagues who served as reviewers of these contributions. The Smithsonian Institution provided funds to support the publication of this volume. This institu- tion, together with the University of Utah and the Utah Geolog- ical Survey, also provided support for the GBASH conference. Linda Williams and Peter Hovingh assisted in many ways with the organization of the conference. We thank Guy King for per- mission to use his unpublished map of Great Basin drainage units. Literature Cited Antevs, E. 1948. Climatic Changes and Pre-White Man. In The Great Basin, with Emphasis on Glacial and Postglacial Times. Bulletin of the Univer- sity of Utah, 38(20):167-191. [Biological Series, 10(7).] Benson, L.V., J.W. Burdett, M. Kashgarian, S.P. Lund, F.M. Phillips, and O.R. Rye 1997. Climatic and Hydrologic Oscillations in the Owens Lake Basin and Adjacent Sierra Nevada, California. Science, 272:746-749. Benson, L.V., S.P. Lund, J.W. Burdett, M. Kashgarian, T.P. Rose, J.P. Smoot, and M. Schwartz 1998. Correlation of Late-Pleistocene Lake-Level Oscillations in Mono Lake, California, with North Atlantic Climate Events. Quaternary Research, 49:1-10. Bermingham, E., and C. Moritz, editors 1998. Special Issue?Phylogeography. Molecular Ecology, 7(4):367-545. Blackwelder, E. 1948. The Geological Background. In The Great Basin, with Emphasis on Glacial and Postglacial Times. Bulletin of the University of Utah, 38(20):3-16. [Biological Series, 10(7).] Bond, G., W. Showers, M. Cheseby, R. Lotti, P. Almasi, P. de Menocal, P. Priore, H. Cullen, I. Hajdas, and G. Bonani 1997. A Pervasive Millennial-Scale Cycle in North Atlantic Holocene and Glacial Climates. Science, 278:1257-1266. Bradbury, J.P, and R.M. Forester 2002. Environment and Paleolimnology of Owens Lake, California: A Record of Climate and Hydrology for the Last 50,000 Years. In R. NUMBER 33 Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:145-173. Brooks, E.J., T. Sowers, and J. Orchardo 1996. Rapid Variation in Atmospheric Methane Concentration during the Past 110,000 Years. Science, 273:1087-1092. Croker, R.S. 1942. Western Division Meetings. Copeia, 1942(3):200. Cronquist, A., A. Homgren, N.H. Holmgren, and J.L. Reveal 1972. Intermountain Flora: Vascular Plants of the Intermountain West, U.S.A. Volume l:iii + 270 pages. New York: Hafner. Davis, J.O. 1982. Bits and Pieces: The Last 35,000 Years in the Lahontan Basin. In D.B. Madsen and J.F. O'Connell, editors, Man and Environment in the Great Basin. Papers of the Society for American Archaeology, 2: 53-75. Davis, O.K. 2002. Late Neogene Environmental History of the Northern Bonneville Basin: A Review of Palynological Studies. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems His- tory. Smithsonian Contributions to the Earth Sciences, 33:295-307. Duvernell, D.D., and B.J. Turner 1998. Evolutionary Genetics of Death Valley Pupfish Populations: Mito- chondrial DNA Sequence Variation and Population Structure. Mo- lecular Ecology, 7:279-288. Fiero, B. 1986. Geology of the Great Basin. 198 pages. Reno: University of Nevada Press. Fremont, J.C. 1845. Report of the Exploring Expedition to the Rocky Mountains in the Year 1842, and to Oregon and California in the Years 1843-1844. 693 pages. Washington, D.C.: Goles and Seaton. Gleason, H.A., and A. Cronquist 1964. The Natural Geography of Plants, viii + 420 pages. New York: Co- lumbia University Press. Grayson, D.K. 1993. The Desert's Past: A Natural Prehistory of the Great Basin, xvii + 356 pages. Washington, D.C.: Smithsonian Institution Press. 2002. Great Basin Mammals and Late Quaternary Climate History. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:369-385. Grootes, P.M., and M. Stuiver 1997. Oxygen 18/16 Variability in Greenland Snow and Ice with 10"3- to 105-Year Time Resolution. Journal of Geophysical Research, 102(C12):26,445-26,470. Harper, K.T., and J.L. Reveal, editors 1978. Intermountain Biogeography: A Symposium. Great Basin Natural- ist Memoirs, 2: 268 pages. Hershler, R., and D.W. Sada 2002. Biogeography of Great Basin Aquatic Snails of the Genus Pyrgulop- sis. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:255-276. Houghton, S.G. 1976. A Trace of Desert Waters: The Great Basin Story. 287 pages. Glen- dale, California: Arthur H. Clark. Hubbs, C.L., and R.R. Miller 1948. The Zoological Evidence: Correlation between Fish Distribution and Hydrographic History in the Desert Basins of Western United States. In The Great Basin, with Emphasis on Glacial and Postgla- cial Times. Bulletin of the University of Utah, 38(20): 17-166, 1 map. [Biological Series, 10(7).] Hunt, C.B. 1967. Physiography of the United States. 480 pages. San Francisco: W.H. Freeman. 1974. Natural Regions of the United States and Canada, xii + 725 pages. San Francisco: W.H. Freeman. King, G.Q. 1986. Morphological Regions of the Great Basin. Forum of the Associa- tion for Arid Land Studies, 2:65-72. King, P.B. 1977. The Evolution of North America. Revised edition, xv + 197 pages. Princeton, New Jersey: Princeton University Press. Kornfield, I., and A.A. Echelle 1984. Who's Tending the Flock? In A.A. Echelle and I. Kornfield, editors, Evolution of Fish Species Flocks, pages 251-254. Orono: University of Maine at Orono Press. Lanner, R.M. 1983. Trees of the Great Basin, xvi + 215 pages. Reno: University of Ne- vada Press. Lowenstein, T.K. 2002. Pleistocene Lakes and Paleoclimates (0 to 200 Ka) in Death Valley, California. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:109-120. Madsen, D.B. 2002. Great Basin Peoples and Late Quaternary Aquatic History. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:387^105. Mayewski, P.A., L.D. Meeker, M.S. Twickler, S. Whitlow, Q. Yang, W.B. Lyons, and M. Prentice 1997. Major Features and Forcing of High-Latitude North Hemisphere At- mospheric Circulation Using a 110,000-Year-Long Glaciochemical Series. Journal of Geophysical Research, 102(C12):26,345-26,366. McPhee, J.A. 1981. Basin and Range. 216 pages. New York: Farrar, Strauus, Giroux. Mifflin, M.D., and M.M. Wheat 1979. Pluvial Lakes and Estimated Pluvial Climates of Nevada. Bulletin of the Nevada Bureau of Mines and Geology, 94: 57 pages, plate 1. Minckley, W.L., D.A. Hendrickson, and C.E. Bond 1986. Geography of Western North American Freshwater Fishes: Descrip- tion and Relationships to Intracontinental Tectonism. In C.H. Hocutt and E.O. Wiley, editors, Zoogeography of North American Freshwa- ter Fishes, pages 519-613 (+bibliography). New York: John Wiley & Sons. Mozingo, H.N. 1987. Shrubs of the Great Basin, xvii + 342 pages, 24 unnumbered plates. Reno: University of Nevada Press. Negrini, R.M. 2002. Pluvial Lake Sizes in the Northwestern Great Basin throughout the Quaternary Period. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contri- butions to the Earth Sciences, 33:11-52. Osmond, C.B., L.F. Pitelka, and G.M. Hidy, editors 1990. Plant Biology of the Basin and Range. Ecological Studies, 80: xii + 375 pages, plates 1-16. Berlin: Springer-Verlag. Oviatt, C.G. 1997. Lake Bonneville Fluctuations and Global Climate Change. Geology, 25:155-158. 2002. Bonneville Basin Lacustrine History: The Contributions of G.K. Gilbert and Ernst Antevs. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:121-128. Polhemus, D.A., and J.T. Polhemus 2002. Basins and Ranges: The Biogeography of Aquatic True Bugs (In- secta: Heteroptera) in the Great Basin. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:235-254. 10 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Reheis, M.C., A.M. Sarna-Wojcicki, R.L. Reynolds, C.A. Repenning, and M.D. Mifflin 2002. Pliocene to Middle Pleistocene Lakes in the Western Great Basin: Ages and Connections. In R. Hershler, D.B. Madsen, and D.R. Cur- rey, editors, Great Basin Aquatic Systems History. Smithsonian Con- tributions to the Earth Sciences, 33:53-108. Ryser, F.A. 1985. Birds of the Great Basin, xv + 604 pages. Reno: University of Ne- vada Press. Sack, D. 2002. Fluvial Linkages in Lake Bonneville Subbasin Integration. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:129-144. Sada, D.W., and G.L. Vinyard 2002. Anthropogenic Changes in Biogeography of Great Basin Aquatic Biota. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:277-293. Shreve, F. 1942. The Desert Vegetation of North America. Botanical Review, 8:195? 246. Sigler, W.F., and J.W. Sigler 1987. Fishes of the Great Basin, xvi + 425 pages. Reno: University of Ne- vada Press. Smith, G.I., and J.L. Bischoff, editors 1997. An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Soci- ety of America, 317: 165 pages. Boulder, Colorado: Geological So- ciety of America. Smith, G.I., and F.A. Street-Perrott 1983. Pluvial Lakes of the Western United States. In S.C. Porter, editor, Late Quaternary Environments of United States, Volume 1: The Late Pleistocene, pages 190-212. Tucson: University of Arizona Press. Smith, G.R., T.E. Dowling, K.W. Gobalet, T. Lugaski, D.K. Shiozawa, and R.P. Evans 2002. Biogeography and Timing of Evolutionary Events among Great Ba- sin Fishes. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:175-234. Spaulding, W.G., E.B. Leopold, and T.R. Van Devender 1983. Late Wisconsin Paleoecology of the American Southwest. In S.C. Porter, editor, Late Quaternary Environments of United States; Vol- ume 1: The Late Pleistocene, pages 259-293. Tucson: University of Arizona Press. Stewart, J.H. 1971. Basin and Range Structure; A System of Horsts and Grabens Pro- duced by Deep-Seated Extension. Bulletin of the Geological Society of America, 82(4): 1019-1044. Trimble, S.H. 1989. The Sagebrush Ocean: A Natural History of the Great Basin, xiv + 248 pages, 32 unnumbered plates. Reno: University of Nevada Press. Wigand, P., and D. Rhode 2002. Great Basin Vegetation History and Aquatic Systems: The Last 150,000 Years. In R. Hershler, D.B. Madsen, and D.R. Currey, edi- tors, Great Basin Aquatic Systems History. Smithsonian Contribu- tions to the Earth Sciences, 33:309-367. Williams, T.R., and M.S. Bedinger 1984. Selected Geologic and Hydrologic Characteristics of the Basin and Range Province, Western United States: Pleistocene Lakes and Marshes. Miscellaneous Investigations of the United States Geolog- ical Survey, Map 1-1522-D. Zoback, M.L., R.E. Anderson, and G.A. Thompson 1981. Cainozoic Evolution of the State of Stress and Style of Tectonism of the Basin and Range Province of the Western United States. Philo- sophical Transactions of the Royal Society of London, A300:407? 434. Pluvial Lake Sizes in the Northwestern Great Basin throughout the Quaternary Period Robert M. Negrini ABSTRACT Sediments deposited by pluvial lakes of the northwestern Great Basin are good repositories of paleoclimate data, which have been used to test models of global climate change whose scale ranges from multidecade to megayear. The principal driving models include (1) Milankovitch-scale (106-104 yr) jet stream migration in reponse to changes in the sizes of continental glaciers, (2) mil- lennial-scale (104-102 yr) climate change as a hemispheric response to large-scale climate changes in the North Atlantic Ocean (i.e., Dansgaard-Oeschger cycles, Heinrich Events, and Bond Cycles), and (3) Antevs' (1948) phenomenological model of climate change in the Holocene, which primarily predicts, on the average, a relatively warm and dry interval in the middle part of the Holocene. Changes in the Earth's orbital geometry are the ulti- mate causes for Milankovitch-scale climate changes. Changes in the solar constant might prove to be the ultimate cause for climate change with periods long enough to overlap the part of the spec- trum occupied by the Holocene changes and by the millennial- scale climate changes; however, the magnitude of change expected from solar processes would be much smaller than those associated with millennial-scale climate changes, so if there is a causal con- nection, it would have to be strongly nonlinear. Climate records from the northwestern Great Basin, particularly those from pluvial Lakes Modoc, Lahontan, and Chewaucan, sup- port models related to Milankovitch-scale climate changes by exhibiting the appropriate periodicities and/or phases of predicted climate change and by exhibiting the appropriate dynamic spectral content of climate change throughout time (e.g., the change from a dominance of periods lasting 400 thousand years (Ky) and 41 Ky to a dominance of 100 Ky starting at -900 thousand years ago Robert M. Negrini, Department of Physics and Geology, California State University, Bakersfield, California 93311, USA. Reviewers: Andrew S. Cohen, Department of Geosciences, University of Arizona, Tucson, Arizona 85761, USA; Timothy K. Lowenstein, Department of Geological Sciences, State University of New York at Binghamton, Binghamton, New York 13902, USA; James F. Luhr, Department of Mineral Sciences, National Museum of Natural His- tory, Smithsonian Institution, Washington, D.C. 20560-0129, USA; and one anonymous reviewer. (Ka)). Records of millennial-scale change from the northwestern Great Basin are beginning to appear in the literature, but at the time of this review, no published records have been dated precisely enough for comparison with the archetypical records from North Atlantic marine and ice cores. A collective record associated with pluvial lakes throughout the northwestern Great Basin indicates a weak trend toward overall lower lake levels and warmer, drier cli- mate during the middle Holocene, thereby supporting Antevs' (1948) model of Holocene climate change for the Great Basin. Introduction Approximately 50 years ago, Ernst Antevs (1948) proposed a model of Quaternary climate change for the Great Basin. The inductive reasoning that led to this model was based in part upon observations made by Allison (1945) and Russell (1895) of changes in pluvial lake levels through time in the northwest- ern Great Basin. The pluvial lakes of the internally drained Great Basin hydrological province have continued to be a valu- able source of nonmarine paleoclimate data primarily because the sizes of these lakes are especially sensitive to changes in precipitation and evaporation (e.g., Morrison and Frye, 1965; Mifflin and Wheat, 1979; Smith and Street-Perrott, 1983; Ben- son and Thompson, 1987a; Benson et al., 1990; Morrison, 1991a; Spaulding, 1991; Oviatt et al., 1992; Grayson, 1993; Benson et al., 1995). Furthermore, these size changes are re- corded by the presence of a wide range of observable features in the lake system, including tufa deposits, strandlines, and biota, and by a wide range of measurable features, including the geochemistry, physical lithology, and magnetism of lake sediments (Currey, 1994b). Finally, because most of the basins have been actively subsiding along marginal faults for much of the late Cenozoic era, their deposits potentially carry paleocli- mate records throughout the Quaternary and earlier into the lat- est Tertiary period. Records from lakes in the northwestern part of the Great Ba- sin are set apart from those of the rest of the Great Basin because 11 12 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES this region is located downwind from the volcanoes of the Cas- cade Range. Consequently, volcanic ashes are ubiquitous in lake-bottom sediments and are commonly found in context with other geomorphic features associated with these lacustrine sys- tems. For example, a 17 m section of lake-bottom sediments ex- posed near Summer Lake, Oregon, contains more than 50 tephra layers, and the overlying sand dunes contain fragments of the Mazama tephra (Allison, 1982; Davis, 1985). The geochemical, petrographic, and, to a lesser extent, the paleomagnetic signa- tures of such ashes allow excellent intrabasinal correlations among outcrops and cores as well as extrabasinal correlations with other well-dated sequences of lake sediments or volcanic deposits. The absolute ages of the ashes often can be determined directly by using a variety of radiometric and luminescence methods, thus extending the age control of these sediments be- yond the range of radiocarbon dating (Berger, 1991; Sarna- Wojcicki and Davis, 1991). This paper is a review of research on northwestern Great Ba- sin pluvial lake levels published since Antevs' 1948 paper. An- te vs' original hypotheses are summarized first, followed by a review of improvements made on these models during the last 50 years. Next is a brief review of other climate-change models that potentially impact the northwestern Great Basin. Finally, advances in the associated observations (i.e., the prehistory of northwestern Great Basin lakes) made during the last several decades are summarized, focusing on studies done after previ- ously published, related reviews (e.g., Davis, 1982; Smith and Street-Perrot, 1983; Benson and Thompson, 1987a; Morrison, 1991a, 1991b; Grayson, 1993). The spatial coverage includes pluvial Lake Lahontan and all of the Great Basin pluvial lakes to the north and west of Lake Lahontan for which published re- search is available (Figure 1). The temporal scope of this paper is the Quaternary period, during which the late Tertiary Great Basin was influenced by the advance and retreat of continental glaciers. The latest Quaternary receives the most emphasis be- cause of the prevalence of data from that time interval. Unless otherwise noted, this paper follows Russell (1895), Hubbs and Miller (1948), and Morrison and Davis (1984a) in the use of the same proper name for all successive pluvial lakes that occu- pied each major basin. Models of Great Basin Climate Change ANTEVS' JET STREAM MIGRATION MODEL FOR PLEISTOCENE CLIMATE CHANGE Near the end of the nineteenth century, Russell (1895) sur- mised that the rise and fall of pluvial Lake Lahontan was syn- chronous with the advance and retreat of continental glaciers in North America. Antevs (1925, 1938, 1948, 1955) agreed with this premise and advanced the idea into a more complete model whereby pluvial episodes in the Great Basin resulted from the southward deflection of the northern jet stream storm track (and associated weather regimes) by North American continen- tal ice sheets. Some of the details in Antevs' model have since been proven to be wrong. For example, his absolute chronolo- gies of climate change were in error principally because his dating methods were limited to long distance correlations of discontinuous varve sequences and to salt balance calculations. In addition, he was overly attached to the assumption of pre- cisely synchronous climate change in Europe and North Amer- ica (Antevs, 1955:324; 1962; Frye et al., 1965:57). Neverthe- less, Antevs' model linking ice-sheet extent to pluvial lake sizes has, to a first order, stood the test of time and, as a result, has become the prevailing model for research on high-ampli- tude Quaternary climate change in the Great Basin (e.g., Ben- son and Thompson, 1987a; Benson et al., 1990; Morrison, 1991a; Spaulding, 1991; Benson etal., 1995). IMPROVEMENTS TO ANTEVS' PLEISTOCENE MODEL Estimating the Size of the Ice Sheets through Time DIRECT EVIDENCE.?The timing of lake-level change pre- dicted by Antevs' model is, of course, dependent upon the size of the continental glaciers through time. With the development of additional absolute dating techniques and the inclusion of additional data, constraints on the extent of North American continental glaciers through time have improved considerably since Antevs' 1948 paper (e.g., Crandell, 1965; Frye et al., 1965; Goldthwait et al., 1965; Lemke et al., 1965; Muller, 1965; Pewe et al., 1965; Schafer and Hartshorn, 1965; Wayne and Zumberge, 1965; Wright and Ruhe, 1965; Hamilton and Thorson, 1983; Mickelson et al., 1983; Waitt and Thorson, 1983; Andrews, 1987; Booth, 1987). This improved knowledge base has led to more robust testing of Antevs' jet stream migra- tion model. Unfortunately, even these advances cannot make up for the obvious fact that most of the direct evidence for the extent of the North American continental glaciers during the Quaternary period has been erased by the last glacial advance. As a result, abundant direct evidence of glacial size is only available for the last maximum and subsequent retreat posi- tions. INDIRECT EVIDENCE.?One of most important advances in Quaternary climate models during the last 50 years was the dis- covery that the relative concentration of the oxygen isotopes in the shells of marine microfossils provides a good first-order proxy for global ice volume for the entire Quaternary period (Shackleton, 1967; Imbrie and Kipp, 1971; Imbrie and Imbrie, 1979). This concentration is conventionally expressed as 818O, the difference between a ratio of 18O and 16O in a given sample and that same ratio observed in a standard sample. Of particular importance to Antevs' model and to the Great Basin is the use- fulness of this proxy to represent the size of the North Ameri- can ice sheets through time. Indeed, despite low-amplitude, high-frequency discrepancies and an approximately 1000-year time lag, the decrease of 518O in marine microfossils has been shown to occur more or less synchronously with the 14-6 Ka NUMBER 33 13 Lake Fort Rock Lake Malheur Lake Modoc Lake Buck Lake Round Lake Malheur Lake Laki Chewauc Oregon_ Nevada N Lake Lahontan Walker Lake FIGURE 1.?Pluvial lakes of the northwestern Great Basin. Lake sizes during pluvial maxima are shown in gray. Modern remnant lakes are shown in black. The names of pluvial lakes discussed in this paper are shown in itali- cized print with large font size; the names of relevant modern remnants are indicated with smaller font. retreat of the North American ice sheets (Mix, 1987; Ruddiman, 1987). Considering that many of the radiocarbon ages used to date both the marine records and the terrestrial limits of the ice sheets were from entirely different materials exposed to radi- cally different geochemical environments, the timing of the two records agrees quite well. Furthermore, there is no reason to be- lieve that this result is not representative of previous glaciations. Another advantage in using marine isotope records as a proxy for ice-sheet size is that their ages can be determined with an accuracy, precision, and continuity previously un- known for middle- to late-Pleistocene sedimentary records. The dating method used assumes that a large component of cli- mate change is driven by variations in solar insolation incident on the earth's northern hemisphere, which in turn are driven by variations in the earth's orbital geometry, the so-called Milank- ovitch cycles (Hays et al., 1976; Imbrie et al., 1984; Martinson et al., 1987; Berger et al., 1992). These variations consist of the precession of the earth's axis of rotation about the ecliptic nor- mal (an ~21 Ky period), the oscillation of the axial obliquity angle (an ~41 Ky period), and the oscillation of the eccentricity of the orbit (an ~100 Ky period). Marine oxygen isotope records are dated by tuning their phase and amplitude spectra 14 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES to those of the Milankovitch cycles. This analysis includes the expected lag in the melting of the ice sheets and the associated lowering of marine 518O. Because the timing of Milankovitch cycles can be accurately and precisely extrapolated backward in time for hundreds of thousands of years from modern and historical astronomical observations, the age control for the ox- ygen isotope global ice volume proxy is arguably precise to within a few thousand years to at least as far back as 300 Ka (Martinson et al., 1987). There is strong support for the accu- racy of the marine oxygen isotope chronology well back into the early Quaternary (Johnson, 1982; Baksi et al., 1992; Izett and Obradovich, 1994; Edwards et al., 1997; Kerr, 1997). It should be noted that terrestrial temperature proxies from ice cores and vein calcites, which should respond more quickly to insolation variations than the marine oxygen isotope records, at times agree very precisely with the marine records but at other times are quite discordant (e.g., Winograd et al., 1992; Wino- grad et al., 1997). This inconsistency calls into question the constancy of the lag times used in the orbital tuning of the ma- rine oxygen isotope chronology and, thus, the accuracy of this chronology. An abrupt increase in mean 518O, a corresponding decrease in the mean percentage of CaCO3, and a coeval increase in variability of both of these indices in marine sediments point to the onset of ice sheets in the northern hemisphere at about 2.5 Ma (Ruddiman and Wright, 1987, and references therein). These early ice sheets apparently fluctuated in size at obliquity- like periods of 41 Ky from -2.5 Ma until -900 Ka (Ruddiman et al., 1986; Ruddiman and Wright, 1987; Ruddiman and Raymo, 1988). From 900 Ka to the present, the 23 Ky preces- sion-like period became as important as the 41 Ky period, and both became subordinate to the 100 Ky eccentricity-like period (Imbrie, 1985; Ruddiman and Wright, 1987). The above-de- scribed character of the marine-sediment proxies for ice-sheet extent can be used in conjunction with Antevs' model to pre- dict similar features in records of pluvial lake-level sizes through time. Furthermore, assuming that the lake-level data have the requisite amplitude resolution and age control, finer- scale features in the marine proxies, such as the oxygen isotope stage (OIS) and the substage boundaries of Hays et al. (1976), also can conceivably be employed as predictive tools. Quantifying the Jet Stream Model ATMOSPHERIC CIRCULATION MODELS.?Spurred by ad- vances in computer technology during the last 25 years, An- tevs' model has been expressed quantitatively as part of a more general effort to model global atmospheric circulation in re- sponse to a variety of factors (Kutzbach, 1987; COHMAP, 1988; Kutzbach and Ruddiman, 1993). From such models, jet- stream position versus time has been predicted in response to changing global climatic conditions for the last 18 Ky. The models are based upon the fundamental laws of physics and are constrained by a myriad of boundary conditions, including the surface area and elevation of continental ice sheets, latitude- and time-dependent Milankovitch-driven changes in solar inso- lation, atmospheric chemistry (e.g., the concentration of CO2), sea-surface temperatures, albedo of the earth's surface, etc. Such efforts have shown that, to a first order, significant changes in precipitation, temperature, and weather distribution patterns should, indeed, occur in response to the advance of the continental ice sheets as proposed in Antevs' original model for the Great Basin (Kutzbach and Wright, 1985; Manabe and Broccoli, 1985; Kutzbach and Guetter, 1986; Kutzbach, 1987; COHMAP, 1988). Although the results of the above-described models gener- ally support Antevs' jet stream migration model, the spatial and temporal grid spacings in the models are relatively large com- pared with the spatial and temporal resolution and the accuracy of the observational database. That is, the general atmospheric circulation models do not generate specific-enough predictions regarding the effect of the jet stream on the magnitude and tim- ing of pluvial lake size in a region as small as the northwestern Great Basin. For example, such models show a maximum southward mean position of a strong winter jet stream at ~25?-35? N from before 18 Ka to as late as 15-12 Ka (Kutz- bach, 1987, fig. 11; Thompson et al., 1993). In contrast, north- western Great Basin lakes are spaced only tens to hundreds of kilometers apart, and the lake-level prehistory of at least one of them (Lake Lahontan) is known throughout this time interval with a precision as small as a few hundred years (see "Latest Pleistocene," below). As knowledge of the lake-level prehisto- ries for the other lakes approaches that for Lake Lahontan, higher-resolution models will be required. EFFECT OF JET STREAM MIGRATION ON LAKES.?More-lo- calized modeling has been conducted to determine if predicted climate conditions would result in lake-size changes that agree with observations (Mifflin and Wheat, 1979; Benson, 1986; Benson and Paillet, 1989; Hostetler and Bartlein, 1990; Hostetler and Benson, 1990; Hostetler, 1991; Freidel, 1993). Input for these models includes the general climate conditions dictated by the output of the larger-scale models, independent observations of pluvial climate conditions (e.g., temperature and precipitation estimates from fossil pollen analyses), or ex- treme values from modern measurements (e.g., river discharge values from the 1983 wet year). Such efforts have demon- strated that pluvial lakes can indeed reach their maximum sizes under the conditions corresponding to a mean annual position of the jet stream storm track near or south of the latitude of the lake in question in response to decreased evaporation from lower temperatures and increased cloud cover and in response to enhanced precipitation in the drainage basins. The amount of required increase in precipitation is reduced significantly by greatly reduced evaporation if the temperature decreases enough to freeze the lakes during the winter. NUMBER 33 15 DANSGAARD-OESCHGER CYCLES, HEINRICH EVENTS, AND BOND CYCLES AS MODELS FOR HIGH-AMPLITUDE PLEISTOCENE CLIMATE CHANGE In the last two decades, high-frequency climate events and cycles not explainable by Milankovitch cycles have been ob- served in several different types of climate records, especially high-resolution ice cores and marine sediment cores (Broecker, 1994). In particular, Dansgaard-Oeschger (D-O) cycles are variations in several climate parameters and were first ob- served in Greenland ice cores (Dansgaard et al., 1982; Dans- gaard et al., 1993). They have a distinctive square-wave mor- phology and periods of a few thousand years. These cycles usually oscillate about a progressively cooler temperature mean until a relatively instantaneous Heinrich event occurs. Heinrich events are marked by distinct marine sedimentary de- posits that indicate extensive ice rafting from North America into the North Atlantic Ocean (Heinrich, 1988). These litho- logic deposits are coeval with the coldest of the cold parts of the D-0 cycles since the last Heinrich event. A rapid warming usually follows, marking the beginning of a new sequence of D-0 cycles that, in turn, end in the next Heinrich event (Bond et al., 1993). Each set of D-0 cycles and its associated terminal Heinrich event is collectively referred to as a Bond cycle (Bro- ecker, 1994). Within Bond cycles, each cold part of the higher- frequency D-0 cycles also are correlated with high percentages of lithic detritus presumably rafted on icebergs (Bond and Lotti, 1995). The balance of evidence suggests the above phenomena are global in nature. Heinrich events, in fact, appear to have oc- curred near major global climate boundaries (e.g., at the transi- tion from OIS 6 to 5, from 5 to 4, and from 2 to 1 (Broecker, 1994). The most-favored mechanisms for connecting Heinrich events (and now the cool parts of the D-0 cycles) with global climate change are (1) the impact of fresh water into the North Atlantic Ocean on the marine thermohaline circulation and re- lated atmospheric/oceanic transfer of heat and water vapor, and (2) the impact on the position of prevailing weather pat- terns caused by a rapid retreat of the Laurentide ice sheet (Bro- ecker, 1994; Bond and Lotti, 1995). Although the causes of D- O cycles, Heinrich events, and Bond cycles have not been iden- tified unequivocally, their widespread nature and distinct tim- ing have led to their use as potential phenomenological models for high-frequency climate change in the Great Basin (e.g., Oviatt, 1997; Benson et al., 1998). ANTEVS' NEOTHERMAL MODEL FOR HOLOCENE CLIMATE CHANGE The understanding of Holocene climate change is important for several reasons, including its bearing on the distribution and nature of archeological sites (Baumhoff and Heizer, 1965; Aik- ens, 1983; Mehringer, 1986; Madsen, 2002) and its importance in establishing a baseline for anthropogenic climate change (Scuderi, 1993). Unfortunately, for the most part, models of Holocene climate change are still in the inductive stage. They are either entirely descriptive or consist of time-varying global phenomena that have only speculative, generalized links to cli- mate. Part of the problem for the northwestern Great Basin lies in the unremarkable nature of the observations relative to those in the latest Pleistocene. Holocene pluvial lake-level changes in this region either are very low in magnitude or, if high-magni- tude changes do exist, are of too high a frequency (< 100-year periods) to be easily observed in the geological record. Despite these problems, the existing models of Holocene climate change are worth reviewing as points of departure. Antevs (1948, 1955) based his model for Holocene (or, in his terms, "Neothermal") climate change in the Great Basin on ob- servational evidence from North America and Sweden and on the assumption of synchronous climate change for the entire Northern Hemisphere (Grayson, 1993). Loosely interpreted, Antevs divided the Holocene into three intervals of roughly equal duration, as shown in Figure 2a. The early Holocene Anathermal age (10-7.5 calendar thousand years ago (cal Ka); 8.4-6.6 radiocarbon thousand years before present (14C Ky B.P.)) was characterized by modern-like temperatures that grad- ually increased throughout the interval and by aridity levels that were slightly moister than today. The middle Holocene Altithermal age (7.5-4.5 cal Ka; 6.6-4.0 14C Ky B.P.) was dis- tinctly warmer than present, and it was characterized by higher aridity. In the late Holocene Medithermal age (4.5 cal Ka to present), conditions returned to cooler temperatures and less aridity, although they were slightly more arid than the Anather- mal. Quantitative modeling has not yet been attempted to deter- mine the extent of lake-level rise and fall in response to Ho- locene climate change; however, assuming Antevs' model is correct, it is reasonable to extrapolate from this model that northwestern Great Basin pluvial lakes would have been at late Holocene levels or slightly higher during the early Holocene and at their lowest levels during the middle Holocene. As with Antevs' jet stream migration model, the particulars of his Neothermal model have fallen prey to additional data that have been analyzed with more modern techniques. For ex- ample, his attribution of moister conditions for the Anathermal age is based in part upon several meters of lake sediments found above the Mazama tephra in an outcrop that is 25 m above the present-day lake level of Summer Lake, Oregon (Al- lison, 1945; Antevs, 1948). After the advent of geochemical fingerprinting of tephra layers and better absolute dating tech- niques, it became clear that Allison's tephra was, instead, the considerably older (~45-50 Ky old) Mount St. Helens Cy tephra (Allison, 1966a; Davis, 1985; Berger and Davis, 1992). Generally speaking, however, when data are averaged tempo- rally over hundreds of years and spatially over the entire north- western Great Basin, Antevs' original hunch has proven to be more or less correct, although a strict adherence to this model is commonly advised against, mostly because of its limited ap- plicability to very specific localities and times (Bryan and 16 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Medithermal Altithermal Anathermal 800- jQQ. Q. CO ?JO 400- o 0J -34- early Holocene cooling event luminosity changes and higher overall solar luminosity i i i 10 11 12 cal Kyr B.P. FIGURE 2.?a, Approximate age range of Antevs' Neothermal ages, b, Frequency and severity of Holocene erup- tions as indicated by residual sulfate anomalies from the GISP2 ice core (after Zielinski et al., 1994, fig. 4). c, Holocene temperature as indicated by oxygen isotope data from the GISP2 ice core (after Stuiver et al., 1995, fig. 5). d, Solar luminosity throughout the middle and late Holocene as inferred from 14C concentration anomalies corrected for variations in magnetic field intensity (after Damon et al., 1989, fig. 4; Magny, 1993, fig. 3). NUMBER 33 17 Gruhn, 1964; Mehringer, 1986; Grayson, 1993; Wigand and Rhode, 2002). Aside from the issue of its applicability, Antevs' model is less than satisfying because it is entirely descriptive in nature. That is, there is no cause attributed to the transitions between, and the durations of, Antevs' three Neothermal ages. Toward this end, Antevs' model appears to be consistent, at least in a qualitative sense, with his jet stream migration model until the end of the middle Holocene. That is, the moist, not-quite-hot Anathermal could be considered to be related to slightly more frequent storms in the northern Great Basin if the storm track had not quite retreated to its modern position associated with ice-free conditions. The Altithermal, which reached its peak hot, dry conditions 5.5-6.0 cal Ka, corresponds quite well with the final disappearance of the North American ice sheets. The return of moderately warm, arid to semiarid conditions in the Medithermal age after 4.5 cal Ka is not explained by this model because the North American ice sheets have not returned. Results from quantitative modeling, however, indicate that the northern jet stream storm track would have reached its modern position and intensity (or lack thereof) by 9 Ka or ear- lier, which is well before the peak of the Altithermal age (Kutz- bach, 1983, 1987). Thus, subsequent reductions in the size of an already insignificant ice sheet should not have affected Great Basin climate. On the other hand, although atmospheric circulation models also take into account the more subtle shift in weather patterns caused solely by Milankovitch-driven changes in insolation in the absence of ice sheets, the low-am- plitude changes in precipitation and temperature that result di- rectly from such insolation changes are usually below the sen- sitivity level of these models (Kutzbach, 1987). On a more localized scale, Benson (1986) has shown that changes in the evaporation rate that were induced solely by insolation changes were an order of magnitude lower than those needed to fill plu- vial Lake Lahontan in conjunction with increased discharge. Thus, if insolation variations were nonetheless responsible for the subtle changes in lake level observed in the northwestern Great Basin during the Holocene, then they must have acted in the absence of the other factors that Benson found to be much more important (e.g., the difference between air and water-sur- face temperatures, the type of clouds, and the degree of cloudi- ness). In any case, the peak value of summer insolation at the latitude of the northwestern Great Basin occurred at ~11 Ka, well before the onset of the middle Holocene Alithermal age. Again, timing is a problem with regard to nondescriptive mod- eling of Holocene climate in the northwestern Great Basin. OTHER MODELS FOR HOLOCENE CLIMATE CHANGE Volcanic Eruptions Large volcanic eruptions significantly increase the concen- tration of atmospheric aerosols and, as a result, decrease global temperatures slightly (<1.5? C) for a few years until the aero- sols settle out (Stuiver et al., 1995). The 1991 eruption of Mount Pinatubo lowered the earth's mean temperature by as much as ~0.5? C for a year or two, as determined both by direct observations and by models that postulated the effects on the atmosphere caused by the observed increase in aerosol concen- tration (Hansen et al., 1992; Houghton et al., 1996, fig. 14). The effects of larger eruptions on climate (e.g., Tambora in 1815), as inferred from increased levels of acidity in Greenland ice cores, do not seem to last much longer (Hammer et al., 1980, fig. 2). Despite the short-term nature of the effects of single erup- tions on climate, some time intervals containing frequent, large eruptions may coincide with intervals characterized by lower temperatures over much of the earth. For example, Hammer et al. (1980, fig. 5) demonstrated longer-term (decade to century) correlations between periods of lowered Northern Hemisphere temperature (by 0.5?-1.0? C) and intensified and/or more fre- quent volcanism. On the millennial scale, Stuiver et al. (1995) used 818O-derived temperature estimates from the GISP2 ice core to argue that volcanism caused no more than a 0.1? C av- erage decrease in temperature during the last 1000 years. The last 1000 years, however, may not be representative of the entire Holocene. Zielinski et al. (1994) presented an in- triguing record of volcanic eruptions for the past 9000 years, based upon their analysis of sulfate concentrations in the GISP2 ice core (Figure 2b). They found that extremely large eruptions occurred more frequently during the early Holocene (-9.0-7.0 cal Ka) as opposed to the remainder of the Holocene. These authors tentatively concluded that this increased volcan- ism contributed to observed climatic cooling in the Northern Hemisphere during a time interval that otherwise would have been warmer because of higher summer insolation values in the Northern Hemisphere. Thus, Holocene climate change could conceivably be explained by the combined effects of volcanism and insolation variations. That is, the highest insolation values in the early Holocene could have been counteracted by the ef- fects of anomalous volcanism. The middle Holocene (Antevs' Altithermal age) would have been a time of higher tempera- tures driven by still-high insolation values, but the cooling ef- fect of the anomalous early Holocene volcanism would have ended. As Northern Hemisphere insolation values decreased steadily to present values, cooler temperatures would have pre- vailed in the late Holocene. This combined model is appealing, but it remains to be shown that the larger, more frequent erup- tions of the early Holocene would result in a significantly larger temperature anomaly than the mere 0.1? C effect deter- mined by Stuiver et al. (1995) for the last 1000 years. It also must be shown that the increased level of sulfate in the GISP2 ice core during the early Holocene was caused by globally sig- nificant eruptions rather than by local events in the vicinity of Greenland (Zielinski et al., 1994). 18 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Solar Variability Low-amplitude temperature change during the Holocene also has been attributed to variations in solar insolation caused, in turn, by inherent variability in the sun's luminosity on the or- der of 0.1% (Stuiver, 1965, 1980; Denton and Karlen, 1973; Reid, 1987; Stuiver et al., 1991; Stuiver et al., 1995). This causal relationship is most directly based upon correlations of the timing of various global climate estimates during the last century with both the number of sunspots (Eddy, 1977; Reid, 1987) and the period of the sunspot cycle (Friis-Christensen and Lassen, 1991). Such variations in solar luminosity can ac- count for an increase of as much as 0.25? C, which is approxi- mately 50% of the long-term temperature increase observed since the middle of the nineteenth century. These values are based upon recent studies that modeled the effect on global temperature caused by the relative contributions of heat input from changes in solar luminosity and from changes in the con- centrations of greenhouse gasses and aerosols (Kelly and Wig- ley, 1992; Lacis and Carlson, 1992; Schlesinger and Raman- kutty, 1992). Proxy indices have been generated to extend estimates of changes in solar activity back through the last several tens of thousands of years. These indices are based upon observed cor- relations of lower solar luminosity with higher concentrations of cosmogenic radioisotopes (Stuiver, 1965; Stuiver and Quay, 1980; Damon et al., 1989; Stuiver et al, 1991; Stuiver et al., 1997). One such proxy index (Figure Id) is based upon A14C, which is a ratio expressing the difference between measured ra- diocarbon activity and the expected activity of a sample of a particular age (Stuiver and Pollach, 1977; Sonett et al., 1990; Stuiver et al., 1991). The data shown in Figure Id are residual values of A14C after the removal of variations caused by well- documented changes in the intensity of the earth's magnetic field (Damon et al., 1989). Such proxies for solar luminosity have been compared with various estimates of climate change throughout the Holocene (e.g., glacier extent around the world and lake-level fluctua- tions in Europe), and correlations have been found corre- sponding to periods of duration of-90, 200, and 2200-2300 years (Damon and Jirikowic, 1992; Magny, 1993; Stuiver and Braziunas, 1993; O'Brien et al., 1995; Stuiver et al., 1995). The largest magnitude of temperature change at these peri- ods, as inferred from Holocene variations of 518O in the GISP2 ice core (0.4? C; Stuiver et al., 1995), is reasonably close to the previously mentioned modeled results for solar- induced temperature change during the last century or so (see above). Assuming that changes in the solar constant during the Ho- locene really do contribute to the A14C variations shown in Fig- ure Id, one could conceivably use this data to predict the tim- ing and duration of the relatively subtle Holocene climate changes observed in the Great Basin. For example, the overall low values of A14C from 7.0 to 3.5 cal Ka would indicate an extended period of higher solar luminosity during a time period that, for the most part, corresponds with Antevs' Altithermal age. The same time period also was characterized by higher- frequency variations in luminosity, suggesting that one might expect to find high-frequency climate changes in the middle Holocene superimposed on an overall warmer climate. Finally, relatively high values of A14C shown in Figure 2t/at time inter- vals of 6.0-5.2, 3.4-1.7, and 0.6-0.2 cal Ka predict low values of solar luminosity at those times. Magny (1993) noted that cooler, moister climate in the European Alps indeed occurred at approximately those times, as evidenced by three dominant intervals of lake-level transgressions during 6.5-4.5, 3.5-2.0, and 0.60-0 cal Ka. Furthermore, the solar-luminosity proxy and an earlier period (8.7-8.0 cal Ka) of transgressions noted by Magny (1993) correlate reasonably well with the low-am- plitude, century- to millennial-scale, global cooling events (8.8-7.8, 6.1-5.0, 3.1-2.4, and 0.6-0 cal Ka) recognized both in glaciochemical time series from the GISP2 ice core and in previously published, coeval paleoclimate indicators from around the globe (O'Brien et al., 1995). This set of observa- tions suggests that solar variability is worthy of further investi- gation as a potential cause of low-amplitude climate change in the Holocene. A Short-Term, High-Amplitude Cooling Event in the Early Holocene A spike-like, early Holocene cooling event centered at -8.4-8.0 cal Ka (e.g., Figure 2c) is recognized in several cli- mate records from around the world (Alley et al., 1997) and may be potentially observable in pluvial lake records from the northwestern Great Basin. The geographic distribution of the type of climate change observed at each locality is similar to that of the terminal Pleistocene, Younger Dryas event (Alley et al., 1997, and references therein). Furthermore, the amplitudes of the anomalies in several climate proxies for the early Ho- locene cooling event are comparable (approximately one-half) to those of the same proxies for the Younger Dryas event. For example, a temperature decrease in Greenland of about 6? C was implied for the early Holocene event by a 2%o decrease in 518O in the GISP2 ice core (Alley et al., 1997). The Younger Dryas event has often been attributed to diminished thermoha- line circulation induced by a sudden influx of fresh water into the North Atlantic Ocean (Broecker et al., 1990; Alley et al., 1997), suggesting that a similar influx caused the early Ho- locene event; however, Alley et al. (1997) noted that no such influx of fresh water is evident in the geologic record during this time interval. Alternatively, this cooling event is coincident with the time interval containing four closely spaced, very large volcanic eruptions, as inferred by the residual sulfate re- cord from the GISP2 ice core (Zielinski et al., 1994). It is tempting to draw a causal relationship between the two phe- nomena, but, again, it has yet to be shown that large eruptions disrupt climate for periods significantly greater than a couple of years (see "Volcanic Eruptions," above). NUMBER 33 19 Pluvial Lake Sizes through Time in the Northwestern Great Basin METHODS During the last 50 years, pluvial lake studies in the north- western Great Basin have benefited from an increasingly di- verse array of new or improved methods with which to deter- mine the timing and magnitude of lake-size changes. These methods include radiometric and luminescence dating (Rosholt et al., 1991; Berger, 1988), petrographic and geochemical fingerprinting of tephra layers (Sarna-Wojcicki and Davis, 1991), paleoecology of lacustral biota (Forester, 1987; Bradbury, 1991, 1992, 1997), trace element and isotope geochemistry of lake sediments (Stuiver, 1970; De Dekker and Forester, 1988; Benson and Peterman, 1995; Benson et al., 1996), particle-size analysis of sediments (McCave and Syvitski, 1992), paleomagnetic-polarity stratigraphy and secu- lar-variation dating as applied to lacustrine sediments and vol- canic tephra (Reynolds, 1977, 1979; Verosub, 1988, Rosholt et al., 1991; Negrini and Davis, 1992), environmental magnetism (Thompson and Oldfield, 1986; Verosub and Roberts, 1995), improvements in sedimentological and geomorphological models of lacustrine systems in arid environments (Currey, 1994a, 1994b), and the empirical and theoretical analyses of post-depositional deformation of basins in this region (Pezzo- pane, 1993; Bills, unpublished). The age estimates of several key northwestern Great Basin tephra layers have been improved since the last review paper was written on northwestern Great Basin lakes, and previous reviews (e.g., Davis, 1982; Smith and Street-Perrot, 1983; Benson and Thompson, 1987a; Morrison, 1991a, 1991b; Gray- son, 1993) have not covered the use of paleomagnetic secular variation as a dating tool in the northwestern Great Basin. These two topics will be addressed below. Recent Work on the Ages of Key Tephra Beds in the Northwestern Great Basin EARLY PLEISTOCENE TEPHRA.?The ages of a few early Pleistocene tephra layers have been revised recently to slightly older ages, in part as a result of improved technology of the laser-fusion 40Ar/39Ar method (Table 1). For example, the ages of the Bishop and Lava Creek B tephra layers (and, by as- sociation, the Rye Patch Dam) have recently been revised downward by an average of a few percent to ~760 ? 2 Ky (Sarna-Wojcicki and Pringle, 1992; Izett and Obradovich, 1994) and 665 ? 10 Ky (Izett et al., 1992), respectively. The same technique has been used to revise downward the age of the Brunhes/Matuyama geomagnetic polarity reversal to agree better with that predicted by marine 818O stratigraphy (e.g., Johnson, 1982; Baksi et al., 1992; Izett and Obradovich, 1994). MIDDLE PLEISTOCENE TEPHRA.?Several middle Pleis- tocene tephra layers have ages that are poorly constrained be- cause they are either too old for the easy acquisition of precise luminescence or radiocarbon ages, too young, or of the wrong chemical composition, mineralogy, and/or grain size for the easy acquisition of precise fission-track or potassium-argon ages on distal volcanic ashes (Naeser and Naeser, 1987; Rosholt et al., 1991; Sarna-Wojcicki and Davis, 1991). As a result, many ashes in this age range have age estimates with TABLE 1.?Tephra layers, from youngest to oldest, found in the pluvial lakes of the northwestern Great Basin. Tephra Tarupah Flat Mazama Tsoyawata Mount St. Helens Mp Trego Hot Springs Wono Mount St. Helens Cy Olema Pumice Castle Set Summer Lake 2 Summer Lake V Wadsworth Summer Lake GG Summer Lake KK Rockland Dibekulewe Lava Creek B Rye Patch Dam Glass Mt. D Glass Mt. G Rio Dell Bear Gulch Pluvial lakes Lahontan Lahontan, Chewaucan Lahontan, Modoc Lahontan, Chewaucan Lahontan, Chewaucan, Modoc Lahontan, Chewaucan Lahontan, Chewaucan Chewaucan, Modoc Chewaucan, Modoc, Malheur Chewaucan Chewaucan, Modoc Lahontan, Modoc Chewaucan Chewaucan, Modoc Lahontan, Modoc Lahontan, Modoc Lahontan, Modoc Lahontan, Modoc Lahontan Lahontan Modoc Modoc Age 600-1500 calyrB.P. 6.845 ?0.05 14CKyB.P. 7.015 ?0.045 14CKyB.P. 20.5 ? -0.5 14CKyB.P. 23.2?0.3 14CKyB.P. 27.3 ?0.3 14CKyB.P. 45-50 Ky 50-55 Ky 72?6Ky 67.3 ? 7.2 Ky 155-200 Ky 155-200 Ky ~190?~30Ky ~200?~30Ky 410 Ky -510 Ky 665?10Ky -675 Ky -0.9-1.0 My -1.0-1.1 My 1.45 My 1.94 My References Davis (1978) Davis (1978), Bacon (1983); Davis (1985) Davis (1978), Bacon (1983); Reick et al. (1992) Davis (1978), Davis (1985); Negrini et al. (1988); Negrini and Davis (1992) Davis (1978), Davis (1985); Reick et al. (1992); Benson et al. (1997) Davis (1978), Davis (1985); Benson et al. (1997) Davis (1978), Davis (1985); Berger and Busacca (1995) Reick et al. (1992); this paper Bacon (1983); Davis (1985); Reick et al. (1992); Botkin and Carambelas (1992a, 1992b); this paper Davis (1985); Berger (1991) Davis (1985); Negrini et al. (1994) Davis (1978), Reick et al. (1991) Davis (1985); Herrero-Bervera et al. (1994); Negrini et al. (1994) Davis (1985); Berger (1991); Herrero-Bervera et al. (1994) Sarna-Wojcicki et al. (1985); Sarna-Wojcicki et al. (1991) Davis (1978); Negrini, Verosub et al. (1987); Sarna-Wojcicki et al. (1997) Davis (1978); Reick et al. (1991); Izett et al. (1992) Davis (1978); Reick et al. (1991); this paper Sarna-Wojcicki et al. (1991); Parker (1993) Morrison and Davis (1984a); Sarna-Wojcicki et al. (1991) Izett (1981); Sarna-Wojcicki et al. (1987); Sarna-Wojcicki et al. (1991) Sarna-Wojcicki et al. (1991) 20 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 0 5- Q.(D Q 10- 15- ? I Mt St Helens Mp(C-14) Trego Hot Springs (C-14, tl_) -Wono(C-14) MtSt Helens Cy(TL) Pumice Castle (K-Ar) N(TL) Prominent Unconformity GG ("excursion" age) GG (Ar-Ar) 0 100 200 Age (Ky) FIGURE 3.?Chronology of Lake Chewaucan sediments exposed in the Ana River canyon north of Summer Lake, Oregon. Age control is based upon a variety of methods that are summarized in Negrini et al. (1994) and in this paper. Tephra names are after Davis (1985). Note the pronounced decrease in precision for dates older than 150 Ky. Shaded region includes all age estimates as allowed by standard deviations of these ages. NUMBER 33 21 uncertainties of 5%-20%. This problem is exemplified by the tephra beds from the pluvial Lake Chewaucan sediments ex- posed in the Ana River canyon near the town of Summer Lake, Oregon (Figure 1). As shown in Figure 3, age uncertainties of tephra from lower in the section are much greater than those from the upper part of the section (see Negrini et al., 1994, for a review of the age control at this section). The large scatter in tephra ages is reflected in the age-range (-180-230 Ky) as- signed to the Pringle Falls/Biwa I geomagnetic excursion. This is a geomagnetic phenomenon recorded in several sedi- mentary records, including two that contain Summer Lake tephra layer GG at the Ana River locality (Negrini et al., 1994) and two that contain the correlative tephra layer D at the Prin- gle Falls, Oregon, locality 170 km to the northwest (Herrero- Bervera et al., 1994). McWilliams (1995) has suggested that the true age of the excursion is toward the older part of this range, based upon the similarity of its transitional virtual geo- magnetic pole (VGP1) path with that of an excursion found in 223 ? 4 Ky old volcanic rocks from New Zealand. Conversely, a correlation of the Pringle Falls event with an excursion found in a marine sediment record from the Blake/Bahama Outer Ridge in the northwestern Atlantic Ocean (Henyey et al., 1995) suggests that the Pringle Falls excursion correlates with the 190 ? 10 Ky old Jamaica event found throughout the world in marine sediments (e.g., Tauxe and Shackleton, 1994; Yamazaki and Ioka, 1994; Henyey et al., 1995; Weeks et al., 1 VGPs are magnetic pole positions pointed to by each paleomagnetic vector recorded at a particular site. Because the earth's magnetic field is not a perfect dipole, coeval VGPs recorded at different sites will not necessarily coincide. 1995; Guyodo and Valet, 1996; Lehman et al., 1996). The cor- relation with the New Zealand excursion is problematical be- cause the VGP path of the Pringle Falls excursion has previ- ously been shown to have a common morphology with the VGP paths of several different polarity transitions (Trie et al., 1991). The correlation with the Jamaica event is more con- vincing because it is based upon distinct paleomagnetic direc- tional and intensity features found in both the excursion itself and in the subsequent secular variation (Negrini et al., 1994; Henyey et al., 1995). PUMICE CASTLE TEPHRA SET AND THE OLEMA AND MOUNT ST. HELENS CY TEPHRA LAYERS.?The three tephra layers comprising the Pumice Castle set were first identified by Davis (1985) in lake sediments from a sequence of 54 water- lain tephra beds in the Ana River canyon near Summer Lake, Oregon (Figure 4). Based upon their major element composi- tions, Davis correlated Summer Lake tephra layers 4, 6, and 8 with the Pumice Castle-like 2, the Pumice Castle, and the Pumice Castle-like 1 tephra layers, respectively, from Crater Lake, Oregon (Bacon, 1983; Davis, 1985). Negrini et al. (1988) added support to this geochemical correlation by com- paring the paleomagnetic direction of Summer Lake tephra 6 with an identical direction for the Pumice Castle tephra mea- sured near the Crater Lake volcano by D. Champion of the United States Geological Survey (USGS) at Menlo Park. The absolute ages of the two older tephra are constrained by brack- eting lavas that are dated by K-Ar at 117 ? 6 Ka and 72 ? 6 Ka; the youngest tephra lies above the 72 ? 6 Ky old lava and be- low a 52 ? 4 Ky old andesite flow, which also was dated with Lake Chewaucan sediments Crater Lake volcanics Tephra 2 (67.3 ? 7.5 Ky) poorly developed soil on tills and glaciated surfaces -50 cm ? lava (52 ? 8.0 Ky) ? Pumice Castle-like 1 ? lava (72 ? 6.0 Ky) ? Pumice Castle ? Pumice Castle-like 2 ? lava (117 ?6.0 Ky) FIGURE 4.?Stratigraphic relationships supporting ~70 Ky old age of the Pumice Castle tephra set (after Bacon, 1983; Davis, 1985). These three tephra layers are nearly the same age because they are found within 10 cm of each other in fine-grained lacustrine sediments at several different localities separated by several kilometers in the Summer Lake subbasin. Thus, they are approximately as old as the 72 + 6 Ky old lava flow that separates two of them at Crater Lake, their proximal location. This age is supported by the thermoluminescence date deter- mined for tephra 2 in the Lake Chewaucan sediments (Berger and Davis, 1992). 22 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the K-Ar method (Bacon, 1983; C.R. Bacon and M.A. Lan- phere of the USGS, in litt. to J.O. Davis, 1984; Davis, 1985). Reick et al. (1992) reported finding the Pumice Castle-like 2 tephra, or perhaps the Pumice Castle tephra, in a core taken from the sediments of pluvial Lake Modoc. (The core was named Tulelake for the nearby town of Tulelake, California, rather than nearby Tule Lake, a remnant of Lake Modoc.) They determined this ash was deposited -120 Ka, in part because of its position below two samples showing negative paleomag- netic inclinations that were provisionally correlated to the 100 -130 Ka Blake Reversed-Polarity Subchron (Reick et al., 1992:419, 422). This age was considered to be reasonably close to the above-mentioned 117 ? 6 Ka K-Ar date on the lava flow below these tephra. Ongoing work on the bottom sediments of pluvial Lake Chewaucan indicates that an age of ~70 Ky, near the middle of OIS 4, is a more reasonable estimate for the Pumice Castle set of tephra beds. The three ash layers in this set are observed to be separated by only a few centimeters of lacustrine silty clays in several exposures 1-2 km apart along the Ana River canyon and in a core taken several kilometers away at an Ana River outcrop (Davis, 1985; Negrini et al., 1988; Erbes, 1996). Be- cause this relationship argues that the three ashes were depos- ited within no more than a few hundred years of each other, and because a 72 ? 6 Ky old lava flow is found between the correla- tives of the upper two tephra found at Crater Lake (Figure 4), all three tephra layers in the Pumice Castle set must be approx- imately the same age as the 72 ? 6 Ky old lava flow. This con- clusion is supported strongly by a thermoluminescence date of 67.3 ? 7.2 Ky on Summer Lake tephra 2 located 35 cm below the Pumice Castle set (Berger and Davis, 1992). The ages of the Mount St. Helens Cy tephra and its correla- tives (e.g., the Marble Bluff bed) also have been the subject of controversy. The ages of these tephra layers have most often been quoted in the range of 35-30 14C Ky B.P. (e.g., Davis, 1978, 1985, and references therein) on the basis of radiocarbon ages of materials found in close proximity to the tephra. In con- trast, radiocarbon dating of some of the near-vent tephra depos- its infers an older age for these tephra layers of perhaps greater than 40 14C Ky B.P., which in turn, infers a calendar date of at least 45-40 Ka (Mullineaux, 1986; Crandell, 1987; McDonald and Busacca, 1992). Thermoluminescence dates on the Cy tephra (46.3 ?-4.8 Ky) in lacustrine sediments at Summer Lake, Oregon, and on the slightly older Cw tephra (-55 ? 5 Ky) in east- ern Washington loess sediments support the older ages (Berger, 1991; Berger and Davis, 1992; Berger and Busacca, 1995). The Olema tephra, originally found in cores from Clear Lake and Tule Lake, California (Sarna-Wojcicki et al., 1988), also was found in the Lake Chewaucan sediments of the WL core (N. Foit, in litt., 1993; Erbes, 1996). Given its position relative to the Pumice Castle set and Mount St. Helens Cy tephra layers in the WL core and given the above-described revisions of the ages of the latter tephra layers, the estimated age of the Olema bed is probably closer to 55-50 Ky than to the previous estimate of 75-55 Ky (Sarna-Wojcicki et al., 1988; Reick et al., 1992). WONO AND TREGO HOT SPRINGS TEPHRA.?New ages for the Wono tephra (27.3+0.3 14C Ky B.P.) and the Trego Hot Springs bed (23.2?0.3 14C Ky B.P.) have been published by Benson et al. (1997). These new ages are based upon 11 precise accelerator mass spectrometer (AMS) radiocarbon dates from adjacent organic samples in a core from Pyramid Lake. Paleomagnetic Secular Variations for Correlating Fine-Grained, Lacustrine Sediments Paleomagnetic secular variation (PSV) records from fine- grained sediments must satisfy strict criteria in order to estab- lish that they record the ambient geomagnetic field vector (dec- lination, inclination, and paleointensity) at or near the time of deposition (Verosub, 1988). This is especially true of the paleo- intensity component (Tauxe, 1993). If these criteria are satis- fied, PSV records can be used as a method for high-resolution (century- to millennial-scale) correlation among fine-grained sediments of Quaternary age over a spatial range of continental scale. Anchored by a solid foundation of PSV reference curves for the last 30 Ky (Lund et al., 1988; Lund, 1996, and refer- ences therein), this methodology has been used on many sedi- ments of the Great Basin (Negrini and Davis, 1992; Ewing, 1996; Liddicoat and Coe, 1997; Benson et al., 1998; Benson et al., unpublished). The potential value of this method is illus- trated in Figure 5a in which the PSV signal from a North At- lantic Ocean core is compared with that from the sediments of Lake Chewaucan. This particular interval of time immediately follows the 190 Ky Pringle Falls/Jamaica/Biwa I geomagnetic excursion (Herrero-Bervera et al., 1994; Negrini et al., 1994; Henyey et al., 1995) discussed in the previous section. Note that this example demonstrates the utility of the PSV method for direct correlations between marine and terrestrial sedi- ments. Clearly, there is an advantage to such an independent correlation of marine and nonmarine climate records without relying on either circular reasoning (i.e., the correlation of cli- mate events to prove that the events occurred simultaneously) or a comparison of radiocarbon dates from entirely different geochemical environments. Another advantage of this method is that, under ideal conditions, it allows for continuous age con- trol throughout the sedimentary section. Obviously, as with all relative age-dating methods, this tech- nique is most valuable when used in conjunction with an inde- pendent method, such as tephrochronology (Verosub, 1988; Negrini, 1993; Liddicoat and Coe, 1997). For example, the ac- curacy of the 14C chronology in the sediments of pluvial Lake Russell is confirmed using such an approach. Figure 5b is a plot of the inclination component of the PSV records from pluvial Lakes Chewaucan and Russell. The Lake Chewaucan record is from the B&B core taken from the middle of the Summer Lake play a (Erbes, 1996). The Lake Russell record is after Lund et al. (1988). The recently redated Wono tephra layer (Benson et NUMBER 33 23 Inclination Inclination 40 80 40 80 170 180 Declination RelativeDeclination Inclination 0 30 60 0 30 60 90 20T projection of Wono tephra age onto Lake Russell inclination record using Lake Chewaucan inclination record 30 Normalized Paleointensity 0 2 Normalized Paleointensity 0 2 1 i i Lake Chewaucan Lake Russell 170 180 Lake Chewaucan Northwestern Atlantic Ocean FIGURE 5.?a, Correlation of paleomagnetic vector components recorded in the sediments of Lake Chewaucan (Negrini et al., 1994) with those recorded in marine sediments from the northwestern Atlantic Ocean (S.P. Lund, unpublished data). The bottom of the correlated intervals begins at the end of the -190 Ky old Pringle Falls/ Jamaica/Biwa I geomagnetic excursion. Inclination and declination are in degrees. Declination of the marine sediments is relative rather than absolute because the core was not azimuthally oriented. The paleointensity data in both records are normalized to the mean as per Tauxe (1993:340). The dashed correlation lines of all three components were fixed by the two prominent inclination troughs shaded in black, b, Correlation of inclination records from pluvial Lakes Russell (Lund et al., 1988) and Chewaucan (Negrini et al., 1993). The correlation model is supported by the projection of the new age (27.3 + 0.3 14C Ky B.P.) for the Wono tephra layer (Benson et al., 1997) onto the previously dated PSV record from Lake Russell. al., 1997) is not found in the Lake Russell record but appears in the Lake Chewaucan section not too far above the interval of sediments that contains a record of the Mono Lake excursion (Negrini, 1994). Using the PSV records, the new age of this tephra layer (27.3 ?0.3 14C Ky B.P.) can be precisely projected onto the Lake Russell record (Figure 5b), thereby confirming the accuracy of this portion of the Lake Russell chronology de- termined previously by radiocarbon dating of carbonates (e.g., Benson et al., 1990; Benson et al., unpublished). PLUVIAL LAKE RECORDS Late Pliocene to Middle Pleistocene (3 Ma-30 Kd) LAKE LAHONTAN.?Our knowledge of the pluvial lacus- trine history of the northwestern Great Basin is bimodal in na- ture. On the one hand, no other pluvial lake in the Great Ba- sin, with the exception of Lake Bonneville, has been studied as extensively as Lake Lahontan (Figures 1, 6). On the other hand, although significant advances have been made during 24 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 118? __OREGON CALIFORNIA[" NEVADA J BLACK ROCKPLAYA BOUNDARY OF LAKE LAHONTAN M EXPLANATION PRIMARY SILLS SECONDARY SILLS PYRAMID LAKE HUMBOLDT-/'/[ ) PRESENT-DAY LAKES OR PLAYAS PRIMARY SILLS Pronto Chocolate Adrian Valley Darwin Pass Mud Lake Slough Astor Pass Emerson Pass MAJOR SUBBASINS Smoke Creek/ Black Rock Desert Carson Desert Buena Vista Walker Lake Pyramid Lake Winnemucca Dry Lake Honey Lake FIGURE 6.?Detailed map of Lake Lahontan pluvial system (from Benson et al., 1990, fig. 2). the last 50 years regarding the prehistory of the other pluvial lakes in the northwestern Great Basin, we still know little about these other lakes relative to what is known about Lake Lahontan. The emphasis on Lake Lahontan is understandable. At high- stand, this pluvial system covered more than 22,000 square ki- lometers, approximately 50% more than all of the other pluvial lakes in the northwestern Great Basin combined (Grayson, 1993). Its sheer size leads to the probability of finding key ex- posures of critical geomorphic features and lake sediments. Furthermore, the presence of large remnant lakes, particularly Pyramid Lake and Walker Lake, provide for modern analogs and, at times, for sites of continuously deposited sediments, en- abling the potential retrieval of long cores containing nearly continuous climate records. Despite the complexity of the Lake Lahontan system (six major streams feeding seven major subbasins that are separated at different elevations by seven sills), the overall lake-level his- tory has converged toward a coherent story for the latest Pleis- tocene (see "Latest Pleistocene," below). The prehistory of this pluvial system prior to 30 Ka is less well developed for the ob- vious reason that most of the evidence either has been super- posed by the sediments of the last highstand or has been eroded or degraded through time. Nevertheless, great exposures of NUMBER 33 25 sediments as old as late Pliocene in age can be found in can- yons cut by the Humboldt, Walker, Carson, and Truckee Rivers (Russell, 1895; Morrison and Davis, 1984a, 1984b; Morrison, 1991b). These exposures and their associated geomorphic fea- tures (e.g., strandlines) are the foundation for the Quaternary stratigraphy of the Lake Lahontan area, which was first devel- oped by Russell (1895) and later refined by Morrison and oth- ers (Morrison, 1964, 1991b; Morrison and Frye, 1965; Morri- son and Davis, 1984a, 1984b). The latest stratigraphic framework for this region consists of a series of alloforma- tions, as shown in Figure 7. According to Morrison and Davis (1984a: 255), an alloformation is defined as "a mappable strati- form body of sedimentary rock that is defined and identified on the basis of its bounding discontinuities." For example, lacus- tral alloformations include all deposits of a single lacustrine ep- isode, such as bars, deep-water sediments, deltas, tufas, etc. In- terlacustral episodes could even include lake sediments in the larger remnants of a great lake, e.g., Pyramid Lake (Morrison, 1991b). This usage follows the Stratigraphic Code of the North American Commission on Stratigraphic Nomenclature (1983). Restrictions on the use of such terminology, such as the need to define alloformations as mappable units, limit the ability to di- rectly correlate the contacts between allounits with climate- change boundaries (e.g., see later discussion of Mount St. Helens Cy tephra as a marker for the contact of the Wyemaha and Sehoo Alloformations). Given adequate spatial sampling of the Quaternary sediments of the Lake Lahontan system, perhaps with the help of an exten- sive coring program, it is this author's opinion that allounits corresponding to each and every marine OIS cycle will be found somewhere in the system. Only after such data is col- lected will Quaternary geologists be able to make final deci- sions on which groups of cycles best fit into collected allofor- mations or, indeed, whether a separate alloformation should be assigned to each cycle. With this end in mind, the present state of the older Lahontan stratigraphy and its paleoclimatic impli- cations are reviewed below. The oldest sediments in the Lake Lahontan area are the "deep-lake/deltaic sediments" (Figure 7) that lie below the in- terlacustral Lovelock Alloformation in at least three locations: (1) 700 m south of Rye Patch Dam on the west side of the Humboldt River (Morrison and Davis, 1984a:256); (2) possi- bly in the Truckee River valley, 6.5 km south of Nixon, Nevada (Morrison, 1991b); and (3) at various locations in the badlands near Weber Reservoir (Morrison and Davis, 1984a; Morrison, 1991b). The absolute age of these sediments is constrained to be at least 800 Ky, based upon preliminary magnetostrati- graphic results and on the identification of the ~1.0-1.1 My old Glass Mountain G tephra layer in an upper unit of these sediments near the Weber Reservoir (Morrison and Davis, 1984a; Morrison, 1991b, and references therein). A lower pre- Lovelock lacustrine unit also appears near Weber Reservoir. Its relative antiquity is inferred by the fact that it is extensively faulted and folded and that these structures are truncated by an TIME years x io3 10 - 20 - ? 1OO - 200 - CO flC > 300 Ik O COI 400 o 2 500 ? UJ 2 600 - 700 - 800 - 900 - J ALLO3TRATIQRAPHIC AND PEDOSTRATIQRAPHIC UNITS FALLON AF(?hallow-la TOYEH QEOSOL _ _(aubaerlal) . Upper member.^moderate lalceaL Middle member (deep lake Thinollte unit (shallow lake! Lower member (deep lake! CHURCHILL QEOSOL AND WYEMAHA ALLOFORMATION (subaerial/shallow-lake) EETZA ALLOFORMATION (sediments of three deep- lake cycles) PAIUTE ALLOFORMATION AND COON QEOSOL COMPLEX (alluvium and eolian sand with many paleosols) RYE PATCH ALLOFORMATION(deep lake sediments) ? LOVELOCK ALLOFORMATION AND PEDOCOMPLEX alluvium with many paleosols) DEEP-LAKE/DELTAIC 8EDIMENTS JLAKE CYCLES(aolld) ISOIL FORMINGIEPI800E8 I (pattern) Ro FIGURE 7.?Stratigraphic units of the Lake Lahontan pluvial system (from Morrison, 1991b, fig. 6). angular unconformity that separates it from the unit containing the Glass Mountain G tephra layer. Although these older lacustrine units are presumed to predate the Lovelock Formation, this formation may itself be quite old at some localities. For example, the ~0.9-1.0 My old Glass Mountain D tephra has been found in the upper part of the Lovelock Alloformation approximately 10 m below the Rye Patch Alloformation at a locality in the Humboldt River can- yon (Parker, 1993, and references therein). In fact, Parker (1993) used magnetostratigraphy to show that the 25 m of sedi- ments found below this tephra were at least ~2.0 My old. No 26 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES lacustrine sediments were observed in this section below the Rye Patch Alloformation, prompting Parker (1993) to presume that these 25 m of sediments are within the Lovelock Allofor- mation. If one interprets the stratigraphy depicted in Figure 7 in a strict sense, then one must conclude either that the pre-Love- lock lacustrine sediments were deposited in an areally re- stricted lake or that an unconformity exists in Parker's section (i.e., the pre-Lovelock lacustral units in this part of the Hum- boldt River canyon were eroded, and the lower part of Parker's section belongs to an older interlacustral alloformation). Addi- tional tephrochronological and magnetostratigraphical studies in the Humboldt River canyon and Weber Reservoir badlands are required to test the continuous nature of older alloforma- tions in the Lahontan Basin. The two allolacustral members of the Rye Patch Alloforma- tion contain two tephra layers in the Humboldt River canyon, namely the Rye Patch Dam bed and the Lava Creek B ash (Morrison, 1991b; Davis, 1978). Morrison (1991b) quoted ages of 630 Ky and 620 Ky for these two ashes, respectively. As dis- cussed above, the Lava Creek B has since been redated at 665 ? 5 Ka by Izett et al. (1992), which puts the Rye Patch Dam bed at a slightly older age of ~675 Ky. These tephra lay- ers imply a correlation of this lacustral episode with OIS 16. The Dibekulewe tephra of Davis (1978) is found above the Lava Creek B tephra in the Humboldt River canyon (Morrison and Davis, 1984a; Negrini, Verosub, et al., 1987). Morrison (1991b) quoted an age of 610 Ky for the Dibekulewe tephra and placed it at the top of the Rye Patch Alloformation. This age estimate was probably derived from the work of Reick et al. (1992), who found this tephra bed in close proximity to the Lava Creek B ash in the Tulelake core. Although this was their preferred age, these authors acknowledged that the age could be much closer to that of the overlying Rockland ash (410 Ky) because of the likelihood of disconformities in the interval be- tween the Lava Creek B and Rockland ashes. A more recent study on a long core from Owens Lake placed the age of the Dibekulewe tephra at approximately 510 Ky based upon con- stant-mass accumulation rates in this core (Sarna-Wojcicki et al., 1997). This age is more in line with an earlier estimate by Negrini, Verosub, et al. (1987:10,625) that was based upon the relative position of these three tephra in a sedimentary section near Blairsden, California. If the Dibekulewe is ~500 Ky old, then either the Rye Patch Alloformation is younger or, alterna- tively, the Dibekulewe ash is within the overlying Paiute Allo- formation. Indeed, the Rockland ash lies within the Paiute Al- loformation in the Humboldt River canyon (Parker, 1993); thus, the upper age limit of the base of this alloformation is 410 Ky. The Eetza Alloformation corresponds roughly to Russell's (1895) lower lacustral clay. This alloformation contains at least three major lacustral allomembers, which consist primarily of silts and clays separated by thick lenses of alluvium or collu- vium. All told there are possibly as many as six to eight lake cycles recorded in exposures along the Truckee River between Nixon and Wadsworth, Nevada (Morrison and Davis, 1984a; Morrison, 1991b). The Wadsworth tephra of Davis (1978) is found toward the top of this alloformation near Pyramid Lake. The age of this tephra was estimated to be in the range of 155 -200 Ky (Morrison, 1991b; Reick et al., 1992, and references therein). Because of this age and the observation that the Rock- land tephra resides below the Eetza in the Paiute Alloformation in the Humboldt River canyon, Morrison (1991b) correlated the Eetza lacustral episodes with OIS stages 10, 8, and 6. This correlation is supported by 230Th ages of 290 ? 90 Ky and 128 ? 10 Ky for samples from near the bottom and top of the Eetza, respectively (Kaufman and Broecker, 1965). The Wyemaha Alloformation represents a long interlacustral interval from ~130 Ka to ~35 Ka and is characterized by eolian and alluvial coarse-grain deposits and well-developed paleo- sols at higher elevations and by finer-grained sediments depos- ited in shallow lakes in the depocenters of the major subbasins (Morrison, 1991a). Morrison proposed that the boundary be- tween the Wyemaha Alloformation and the overlying lacustral Sehoo Alloformation should be put at the Mount St. Helens Cy tephra, now thought to have formed 45-50 Ka (see above). This is a reasonable suggestion as it would make the Wye- maha/Sehoo transition relatively easy to map. This tephra layer is thick, is widespread throughout much of the Lahontan Basin, is distinctive both petrographically and geochemically, and is commonly found at the base of lacustrine sequences associated with the Sehoo Alloformation (Davis, 1978:45). There is a trade off, however, between the mappability of this contact and its correlation with global climate transitions. The Sehoo Allo- formation is commonly considered to be roughly coeval with OIS 2 and perhaps with the early Holocene (Davis, 1978; Morrison, 1991b). If the Mount St. Helens Cy tephra layer is indeed as old as it appears to be based upon recent chronomet- ric studies (see above), then the base of the Sehoo Alloforma- tion is pushed back to early OIS 3 time. Perhaps the 27.3 14C Ky B.P. Wono tephra would be a better choice for the base of the Sehoo Alloformation. It is as widespread as the Mount St. Helens Cy tephra layer in the Lahontan Basin (Davis, 1978) and has a more appropriate age. Furthermore, it was deposited in a lake of moderate depth preceding the last significant reces- sion of Lake Lahontan, which, in turn, preceded the prominent transgressive phase correlated to the OIS 2 ice sheet advance (Benson et al., 1995, 1997). The 23.2 14C Ky B.P. Trego Hot Springs tephra layer is, in some sense, still a better choice as it was probably deposited in the shallow-water lakes of the reces- sion immediately prior to the OIS 2-related transgression. The Trego Hot Springs tephra, however, is not as widespread as the Wono bed (Davis, 1978), probably because it was deposited under shallower lake conditions. In addition to its limited application for exact correlations to global or regional climate change, the Lake Lahontan stratigra- phy places only very loose constraints on the size of this plu- vial lake versus time. That is, one can determine whether or not the lake existed during a given time interval but not the exact NUMBER 33 27 elevation of the shoreline. To most effectively use the geologic record to determine the effects of climate change on the lake, one must determine the surface area of the lake versus time (Smith and Street-Perrot, 1983; Benson and Thompson, 1987a; Benson and Paillet, 1989). To do this, one must, of course, de- termine the precise elevation of the lake surface. Although this has been done exceedingly well for the time period after 30 Ka (see next section), very little has been done to determine abso- lute lake levels for earlier time periods. Morrison (1991b) has done this in a general way by documenting the maximum ele- vation occupied by lacustral alloformations. Reheis et al. (2002) also have obtained estimates of the lake size during plu- vial maxima in the early and middle Pleistocene by mapping highly eroded but still discernable strandlines and near-shore deposits around the Lahontan Basin and nearby basins. In do- ing so, they have reconstructed an enormous lake correspond- ing to OIS 16 that would have covered Reno and Carson City, Nevada, and would have backed the Humboldt River at least to the town of Battle Mountain, Nevada. Lao and Benson (1988, and references therein) and Szabo et al. (1996) determined lake-level minima by dating tufa deposits with the uranium- series method. The results of the earlier study indicated that the surface of Lake Lahontan in the Walker subbasin (the first to fill up if it is fed by the Walker River) would have reached a level more or less equal to its late Pleistocene maximum of 1330 m at 290 ?90 Ka, -200 Ka, 128 ? 10 Ka, and 49 ?5 Ka. According to the model of Benson and Thompson (1987b), Pyramid Lake, Winnemucca Dry Lake, and Smoke Creek/ Black Rock Desert subbasins would have coalesced and risen to their common elevation of 1207 m at Emerson Pass. It is conceivable that all of these subbasins could have joined the Walker subbasin at its 1330 m elevation if they received suffi- cient spillover from the Walker subbasin through the Carson River subbasin. Szabo et al. (1996) combined dates from sev- eral studies in the Pyramid Lake subbasin near the elevation of the 1207 spillpoint into the Smoke Creek/Black Rock Desert subbasin. They found that a lake filling at least the local subba- sins would have been at or above the Emerson Pass sill for much of the middle Pleistocene, especially in the time interval from 250 Ky to 170 Ky. Also, these subbasins would have filled at least twice during the time period from 65 Ka to 30 Ka. LAKE MODOC AND ENVIRONS.?The longest, most continu- ous pluvial lake record in the northwestern Great Basin comes from the sediments of pluvial Lake Modoc (Figure 1) and nearby Buck Lake and Round Lake in south-central Oregon (Klamath County) and northernmost California (Modoc and Siskiyou Counties), respectively. Upper and Lower Klamath Lakes and Tule Lake are the remnants of pluvial Lake Modoc, which once filled the subbasins that these lakes occupy. At plu- vial maximum, the surface of Lake Modoc stood at an eleva- tion of 1292 m, approximately 63 m above the surface of mod- ern Tule Lake (Dicken, 1980). The Lake Modoc system is not considered to be part of the Great Basin hydrologic province as defined by Grayson (1993) because most of it presently drains into the Pacific Ocean via surface flow through the Klamath River and via groundwater flow through permeable volcanic rocks into the Klamath River and Pit River systems. Neverthe- less, it is often included as a source of Great Basin paleocli- mate data because it lies within the botanical Great Basin and its size has responded dramatically to climate changes in the past regardless of its connection to the Pacific Ocean (Dicken, 1980; Bradbury, 1991; Grayson, 1993). Also, as with most plu- vial systems in the hydrographic Great Basin, Lake Modoc has been a site of lacustrine sedimentation for millions of years as a consequence of its presence within the Basin and Range struc- tural province (Dicken, 1980). Sediment cores were taken from seven sites in this system by the USGS (Adam et al., 1989; Adam, 1993; Klamath Core Study Team, 1995). At two of the sites (Tulelake, 334 m; and Butte Valley, 102 m) coring was completed into sediments older than 3 My. At three of the other sites (Wocus Marsh, 52 m; Buck Lake, 41m; and Round Lake, 50 m) the oldest cored sediments were from 1.4 to ~1.0 Ma. The Grass Lake core (29 m) penetrated sediments no older than ~400 Ky, and the Cale- donia Marsh cores (15.3 m) bottomed out in sediments younger than 100 Ky. The ages of all of the cores were con- strained using tephrochronology (Adam et al., 1989; Reick et al., 1992). Several well-known tephra were identified, includ- ing the 1.94 My old Bear Gulch ash, the 1.45 My old Rio Dell ash, the 665 Ky old Lava Creek B ash, the 410 Ky old Rock- land ash (Sarna-Wojcicki et al., 1985; Meyer et al., 1991), the 50-55 Ky old Olema ash (Sarna-Wojcicki et al., 1988), a pre- cursor ash (the Tsoyawata bed) to the Mount Mazama eruption (Davis, 1978; Bacon, 1983), and several ashes that correlate with the collection of tephra found nearby at Summer Lake, Oregon, such as one of the Pumice Castle tephra set (Bacon, 1983; Davis, 1985; Reick et al., 1992). The ages of the Tule- lake, Butte Valley, and Buck Lake cores were also constrained by magnetostratigraphy (Reick et al., 1992; Roberts et al., 1996; Rosenbaum et al., 1996). Radiocarbon dating was con- ducted on samples from the top of the Grass Lake, Caledonia Marsh, and Wocus Marsh cores with limited success because of inverted ages, low concentrations of carbon, and a prepon- derance of anomalously young ages (Bradbury, 1991; Klamath Core Study Team, 1995). Despite similar difficulties encoun- tered with the Tulelake core, Bradbury (1991, fig. 6) (Figure 8) was able to assemble a chronology for the top of this core by using selected radiocarbon dates and tephra. In addition to age control, several other studies have begun on the Klamath Basin cores that have generated data from which interpretations of pluvial lake conditions can be made. These include lithologic description, palynology, diatom and ostracode analyses, environmental magnetism, and geochem- istry. An initial study on the Tulelake core found that two ma- jor shifts in the character of pollen type versus depth curves occurred at -2.5 Ma and -1.65 Ma (Adam et al., 1989). The 2.5 My shift is best expressed by a slight overall increase in 28 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES a a.o 0 75 0 percent 75 0 500 300 750 500 diatoms/mm FIGURE 8.?a, Diatom biostratigra- phy for the entire Tulelake core from Bradbury (1992, fig. 6). For detailed descriptions of all genera, groups, and species the reader is referred to Bradbury (1991, 1992). b, Fragi- laria and Aulacoseira diatom pro- files for the last 180 Ky, after Brad- bury (1997, fig. 9). Note positions of key tephra layers. Fragilaria Aulacoseira 50 age (Ky) 100 150 Tephra T2023 percent Artemisia count and by the transition toward higher-frequency variations in both Artemisia and TCT (Taxodiaceae + Cupressaceae + Taxaceae, which includes Western Juniper). This older shift was correlated with the onset of continental glaciers in North America and the beginning of the Praetiglian cold period in Europe. Some researchers believe that the Plio/ Pleistocene boundary should be moved back to this point in time, specifically to the Gauss/Matuyama polarity chronozone boundary, rather than the currently accepted time, the top of the Olduvai normal-polarity subchronozone (see Morrison, 1991a, and references therein). The latter candidate for the Plio/Pleistocene boundary is also expressed in the Tulelake core. After 1.65 Ma there are common interruptions in pollen preservation, a drop in TCT amplitude, increases in the peak percentages of Artemisia, Cheno-ams, and Poaceae (grasses), and high-amplitude oscillations in total algae, all of which were interpreted to indicate a pattern of alternating lacustrine and nonlacustrine phases (Adam et al., 1989). Two diagnostic species of diatoms, Anomoeoneis costata (Kutzing) Husted (saline water) and Cocconeis placentula Ehrenberg (shallow water), were most abundant between 90 and 50 m of depth (800 to <400 Ka), indicating an overall drying of the lake and a hydrologically closed lake that was rich in solutes. These ob- servations are consistent with well-preserved ostracodes and with high values for "factor 3," an association of geochemical entities (e.g., (carbonate), Mg, and Sr) indicative of dry lake conditions (Adam et al., 1989). Bradbury (1991, 1992, 1997) later used a more detailed and comprehensive diatom stratigraphy to reconstruct lake condi- tions in the Tule Lake subbasin throughout the entire three mil- lion years represented by that core. In these reports he made several intriguing interpretations that bear on global climate- change models. First, he noted an ~40 Ky periodicity in total diatom concentration in the two densely sampled time intervals (3.0-2.5 Ma and 2.05-1.85 Ma) prior to 1.0 My. This observa- tion is consistent with the dominance of the obliquity period in the late Pliocene and early Pleistocene observed in marine oxy- gen isotope records (Ruddiman et al., 1986; Ruddiman and Raymo, 1988). Second, based upon the gradual disappearance of Aulacoseira solida (Eulenstein) Krammer and the subse- NUMBER 33 29 quent takeover by a Fragilaria-Stephanodiscus?Cyclotella as- semblage, he noted a transition from "more equable and prob- ably warmer Pliocene climates to more variable, more seasonal, drier, and probably colder Pleistocene climates" (Bradbury, 1992:284). Third, three broad peaks in Aulacoseira solida alternating with broad peaks in Fragilaria species oscil- late with a period approximately equal to the 413 Ky eccentric- ity-driven insolation cycle (Ruddiman and Raymo, 1988). Fourth, from 900 to ~300 Ka, the diatom stratigraphy in the Tulelake core is characterized by several peaks in Fragilaria sp. concentration that are more or less in tune with glacial stages recorded by marine oxygen isotope curves (Bradbury, 1992, fig. 7). Furthermore, these peaks are terminated by in- creases in diatoms that prefer more saline or alkaline waters (e.g., Anomoeoneis costata, Cocconeis placentula). Bradbury interpreted these data to imply that, during glacial periods, Tule Lake was probably an open, freshwater marsh formed during cold and somewhat dry climates. During interglacials, effective moisture decreased but winter precipitation was prob- ably significant as the presence of Stephanodiscus and Aula- coseira species in these intervals suggests at least seasonally high lake levels and dilute waters. Fifth, Bradbury (1992) noted that the more or less uniform response to glacial cycles during the 900-300 Ka time period breaks down in the latest Pleistocene. The interval from 300 to ~180 Ka is represented by Anomoeoneis costata, a diatom that prefers "exceptionally shallow and alkaline conditions." Thus, the low effective mois- ture interval that might correspond to OIS 7 appears to begin 80 Ky too soon in the Tulelake core. The interval from ~ 170 to ~130 Ka contains abundant Fragilaria, which would be ex- pected for a glacial interval like OIS 6 based upon its occur- rence patterns in this core from 900 to 300 Ka; however, as Fragilaria abundance declines, the diatom concentration be- comes characterized by deeper-water diatoms (e.g., Aula- coseira ambigua (Grunow) Simonson, Cyclotella meneghini- ana Kutzing, Cyclotella bodanica Grunow) until A. ambigua decreases at ~50 Ka. Thus, if the 900-300 Ka diatom assem- blage is a good analog for the latest Pleistocene, then it appears as if the entire interval from 170 to 50 Ka is characterized by a glacial interval; that is, the OIS 5 interglacial is not represented in the Tulelake core. Bradbury (1991:228) noted that diatoms such as Fragilaria, which characterized pluvial intervals in the middle Pleistocene, might also do well during a warm-dry in- terpluvial period like OIS 5; however, the relatively low abun- dances of Fragilaria during previous interglacials is left open for question. Two cores from the Klamath Basin have proven to be fertile ground for the nascent field of environmental magnetism. Rob- erts et al. (1996) conducted sediment magnetic measurements on bulk samples and reflected light microscopy and x-ray de- fraction (XRD) analysis on magnetic separates from the Butte Valley core. They found that most of the magnetism in the 3.0 My of core was carried by titanomagnetite, which was most likely derived from basalt detritus in the catchment area. Greigite (Fe3S4), an authigenic magnetic mineral, was found in several horizons in the age interval from 3.0 to ~1.0 Ma. Be- cause the formation of greigite depends upon the availability of dissolved sulfate, Roberts et al. (1996) tentatively con- cluded that greater residence time of lake water may have caused the slightly anoxic conditions under which the avail- ability of dissolved sulfate would have been enhanced. Thus, lake conditions were significantly different, at least in inter- vals, for the periods before and after ~1.0 Ma. Notably, a nine- fold increase in sedimentation rate also occurred at ~ 1.0 Ma, supporting a major change in the environment of deposition at this time. Climate change during the middle Pleistocene (~400 Ka to -300-200 Ka) also was interpreted with the help of an environ- mental magnetism record from Buck Lake, 15 km west of Kla- math Basin (Rosenbaum et al., 1996). Variations in magnetic properties were found to correlate strongly with changes in cli- mate, as indicated by pollen type. In particular, zones of high magnetic susceptibility, hence, high concentrations of magne- tite, coincide with cold and dry climate zones, and zones char- acterized by low susceptibilities were identified with warm cli- mate zones. The magnetic variations preceded the changes in pollen, an observation that prompted Rosenbaum et al. (1996) to propose that Buck Lake magnetic properties responded more quickly to climate change than did vegetation. High hematite concentrations, inferred from two ratios (HIRM and the S pa- rameter) calculated from measurements of isothermal remnant magnetization (Thompson and Oldfield, 1986), were found to be in phase with high magnetite concentrations. The overall content of heavy minerals, as indicated by the concentrations of Ti and Zr (immobile elements mainly held in heavy minerals), was also found to be in phase with increases in magnetic-min- eral content; however, because the magnetic-mineral concen- trations varied by two orders of magnitude between cold and warm climates and, in contrast, the heavy-mineral concentra- tion varied only by a factor of three, Rosenbaum et al. (1996) concluded that the low amount of magnetic minerals during warm climates relative to other heavy minerals was caused by sourcing from soil and highly weathered rock in the catchment. In contrast, during pluvials, greater runoff would have brought in a greater concentration of relatively fresh detritus derived from basaltic bedrock. LAKE CHEWAUCAN.?Four subbasins (Summer Lake, Upper and Lower Chewaucan Marshes, and Lake Abert) combined during pluvial maxima to become pluvial Lake Chewaucan in south-central Oregon (Figure 1). At its maximum, the lake was 113 m deep, and its surface stood at an elevation of 1328 m. The climate prehistory of this pluvial lake (Allison, 1945) was made well known when Antevs (1948) used it as a constraint for his Neothermal climate model. Although this prehistory proved to be wrong in light of the application of more modern chronological techniques (see earlier discussion under "Other Models for Holocene Climate Change"), pluvial Lake Chewau- can has continued to be the focus of climate-change research 30 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES CO o 0 50 Mill lower ? i i i ii i higheri i 111 n i 500 , i ij more saline 0 J fresher JLi 20 I 1 1 ? coarser f~ i * 0 finer ? \i i i i i i i i ^ 200 outcrop magneticsusceptibility (SI units) 250 lower higher WL core magnetic susceptibility (SI units) i i i i i II i i II 1 outcrop salinity index (dimensionless) saZ WLcore salinity index (dimensionless) WL core silt/clay ratio (dimensionless) - y, marine 0-18 (norm III alized) III interpluvial (Stage 7) interpluvial (Stage 3) pluvial (Stage 4) interpluvial (Stage 5) pluvial (Stage 6) 0.001 0.001 500 1.0 -1.0 FIGURE 9.?Lake-size proxies for Summer Lake subbasin of pluvial Lake Chewaucan compared with a marine oxygen isotope proxy for ice-sheet extent (Martinson et al., 1987, fig. 18). Magnetic susceptibility has been shown to correlate with other lake-level proxies by Erbes (1996). The salinity index is a weighted ratio of salino- phile ostracode faunal counts over counts of ostracodes that prefer fresh water (Palacios-Fest et al., 1993). The outcrop samples are from the Ana River canyon near the northern margin of the subbasin. The WL core samples are from a locality closer to the depocenter, near the northern edge of the modern playa lake. owing to the combination of excellent exposures of pluvial lake sediments, abundant tephra layers, climate-controlled geomor- phology, and well-preserved records of fossil biota and magne- tism (e.g., Allison, 1982; Grayson, 1993). Furthermore, in the last 50 years many studies on archeology, climate modeling, geochemistry, geomorphology, fossil magnetism, hydrology, mineralogy, paleontology, sedimentology, structural geology, and tephrochronology have set the stage for continuing paleo- climate and related research in this basin (Conrad, 1953; Brown, 1957; Phillips and Van Denburgh, 1971; Van Den- burgh, 1975; Allison, 1982; Negrini et al., 1984; Davis, 1985; Negrini et al., 1988; Simpson, 1990; Banfield et al., 1991a, 1991b; Berger and Davis, 1992; Gobalet and Negrini, 1992; Negrini and Davis, 1992; Freidel, 1993; Palacios-Fest et al., 1993; Pezzopanne, 1993; Negrini et al., 1994; Oetting, 1994; Roberts et al., 1994; Wigand et al., 1995; Erbes, 1996; Wigand and Rhode, 2002). The lake-level history for the Chewaucan system during the last 200-300 Ky has been estimated from several proxy indica- tors derived from the analysis of bottom sediments exposed in outcrops and sampled from cores (Figure 9). The correlation of cores with outcrops and the age control of the resulting com- posite section (Figure 3) are based upon a variety of methods, including lithostratigraphy, tephrochronology, sediment mag- netism, paleomagnetic secular variation, radiometric methods, and thermoluminescence (Davis, 1985; Berger and Davis, 1992; Negrini et al. 1994; Erbes, 1996). Notably, this chronol- ogy was constructed completely independent from a priori as- sumptions regarding correlations to climate-change events. Variations in lithology type, sediment grain size, and the con- centration of magnetite grains, as indicated by magnetic sus- ceptibility, all indicate that the environment of deposition changed systematically through time more or less in tune with the marine OIS signal (Figure 9). Complementary variations NUMBER 33 31 through time in the paleosalinity of the lake, as estimated by the faunal distribution and geochemistry of fossil ostracodes, and in the shallowness of the lake, as estimated by the noisi- ness of the trace-element concentration of the same ostracodes, strongly suggest that the changes in the depositional environ- ment were caused by increases in lake size as the ice sheets ad- vanced and vice-versa (Figure 9; Palacios-Fest et al., 1993; Er- bes, 1996). Thus, both the nature of the variations of climate proxies in the Chewaucan sediment and their timing indicate that, to a first order, the lake-level history of pluvial Lake Chewaucan supports Antevs' jet stream migration model. Marine OIS 4 is well resolved in the Summer Lake record (Figure 9). In fact, the freshest water, hence the deepest lake, appears to have occurred in OIS 4 rather than in OIS 6 or 2. There are two explanations for this observation. First, because the signatures and durations of OIS 4 are either subdued, short- ened, or nonexistent in lake-level proxies from southern Great Basin lakes (e.g., Bradbury, 1997; Lowenstein, 2002), it may be that the southernmost position of the jet stream occupied a more northerly latitude (i.e., over Summer Lake) than it did during the OIS 2 glacial maximum (Lowenstein, pers. comm., 1997). More specifically, if the jet stream remained for a while at its southernmost point over southern Oregon during OIS 4, this lesser glacial stage would have had a very profound effect on Lake Chewaucan. In contrast, if the jet stream moved quickly through Oregon on its way south and back during the OIS 2 event, then the pluvial episode seen at Summer Lake during OIS 2 may have been relatively minor. A second and equally plausible explanation is the following. Lake Chewau- can could have been deeper and fresher during OIS 6 and 2 than it was during OIS 4, but the records of the freshest lakes were erased by deflation during the subsequent interpluvial maxima. The latter explanation is supported by the gap in sedi- mentation shown in Figure 9 from -160 to 115 Ka and also by the fact that the sediments at the surface of the modern Summer Lake playa are no younger than 23 14C Ky B.P. (Negrini et al., 1993). Note that without good independent age control one would never realize that cold/wet and hot/dry extrema corre- sponding to both sides of glacial terminations had been erased by deflation. Thus, the Summer Lake example serves as a warning to those drawing conclusions regarding climate ex- trema from poorly dated paleoclimate records associated with deflation-prone desert basins. In addition to climate variations, drainage diversions also could have affected the surface level of Summer Lake. In fact, the Chewaucan River, the main stream entering the Chewaucan system, is currently diverted from the Summer Lake subbasin by its own fan-delta complex (Allison, 1982; Davis, 1985). It is unlikely, however, that such diversions would have resulted in a lake-level history that matches the OIS signal so well (Figure 9). Thus, climate variation is probably the major influence on lake level in this subbasin. A similar study of the Lake Abert subbasin could provide a test for this hypothesis. In an early study of the paleomagnetism at this locality, Milankovitch-like periodicities were reported in spectra of the magnetic vector time series (Negrini and Verosub, 1987; Ne- grini, Darke, et al., 1987; Negrini et al., 1988). This was cer- tainly a provocative result but has since proven to be incorrect by improvements in the chronology (e.g., Negrini et al., 1994). Even though the PSV time series are no longer considered to be related to climate change, they still have potential as tools for high-resolution correlation of other climate records with the Chewaucan record (see above). Comparison of Lake Chewaucan and Lake Modoc Re- cords: The Lake Modoc climate record, as described in a pre- ceding section, contrasts greatly with the Chewaucan record. Specifically, OIS 4 is a major feature in the Chewaucan record but is barely noticed in the Tulelake record. Furthermore, the OIS 5 interglacial episode at Lake Modoc is characterized by continuous deposition in an environment similar to that of gla- cial episodes earlier in the Pleistocene. Most of the discrepancies between these records can be explained by differing interpreta- tion of chronologies rather than by radically different climates in regions that are in close proximity. Figure 10 presents a compar- ison of the two records after the Lake Modoc chronology has been reinterpreted by aligning the records using four tephra lay- ers common to both records. The age scale assumes that the chronology for the Lake Chewaucan record is accurate, but be- cause the comparison is based upon the relative position of the four tephra layers, the comparison would still hold up if the Chewaucan chronology changes with improved age control. With the reinterpreted chronology, the two records now have much in common with each other and with the marine OIS sig- nal (Figure 10). In particular, the interval of time from ~190 to -160 Ka is characterized in both records by more pluvial con- ditions, and a strong OIS 4 signal is now observed in both records. Note also that the reinterpretation forces a major OIS 5 unconformity on the Lake Modoc record. This is not entirely unreasonable considering that such an unconformity occurs in the Chewaucan record. PLUVIAL LAKE MALHEUR.?A large region (2400 km2) of very low relief (< 10 m) in eastern Oregon presently contains Malheur, Mud, Harney, and Silver Lakes (Figure 1). Presently, all except Malheur are ephemeral lakes. In 1986, after a series of abnormally wet years, Malheur overflowed through a sill called The Narrows into Mud Lake and then into Harney Lake (Hostetler and Bartlein, 1990). At its peak, this combined lake reached a depth of only 4.4 m, and its surface stood at an eleva- tion of 1250.5 m. Had the effective moisture level increased so that the lake rose another one meter or so, then the combined lake also would have overflowed into the Silver Lake basin. If the lake deepened yet another three meters to a surface eleva- tion of-1254 m, it would have overflowed into the Columbia River system. One of the tephra layers from the -70 Ky old Pumice Castle set was identified in lacustrine sediments at an elevation of 1247 m near the base of two trenches associated with archeological 32 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 50 100 15 To E T5 HI 150 200 - Lake Chewaucan (Ana River Outcrop) magnetic susceptibility 0.001 jsl units) 0.01 Lake Modoc (Tulelake Core) Marine 0-18 0 40 0 40 ? 0 Olema Tephra Tulelake T2382 Pumice Castle Set OIS 4 expressed in Tule Lake core as high percentages of Fragilaria or Aulacoseira gap in record corresponding to unconformity observed in Chewaucan sediments gap in record corresponding to unconformity predicted for Tule Lake sediments between Pumice Castle and "V" tephra layers Tulelake T1193 ??????????????????????????I Summer Lake Tephra V Tulelake T2023 Summer Lake Tephra OIS 6 expressed as high percentage of Fragilaria 50 100 150 200 500 0 0 40 0 40 ostracode salinity index % Fragilaria % Aulacoseira FIGURE 10.?Comparison of Lake Chewaucan and Lake Modoc/Tulelake core (from Figure 8) paleolake envi- ronments. Pluvial conditions increase to the right in all diagrams. Relative age-control is anchored by four com- mon ash layers, as per Reick et al. (1992). The absolute chronology of the Lake Modoc record is reinterpreted in terms of preferred ages for the tephra layers, as discussed in the text. The positions of the these tephra layers in the original Tulelake chronology are shown in Figure 86. The new chronology for the Lake Modoc record resolves the anomalous occurrence of Fragilaria sp. in the OIS 5 interpluvial, which was noted by Bradbury (1991, 1992), by placing this occurrence in OIS 6 and OIS 4. This chronology also indicates a major gap in the Tulelake record between the Pumice Castle tephra and the Summer Lake V tephra layers. investigations near the southern margins of modern Malheur Lake (Elston, 1992; Botkin and Carembelas, 1992a, 1992b). In one of the trenches, the ash is a continuous layer found near the base of at least one meter of stratified lacustrine sandy clays. Thus, for approximately 70 Ky, during the middle of OIS 4, pluvial Lake Malheur had enough water in it to deposit more than one meter of sediments on a lake bottom at 1248 m, which is two meters above the present-day lake surface. If the OIS 4 lake had a depth of at least 2 m, then pluvial Lake Malheur would have been at least as large as it was in 1986. Pluvial Lake Malheur also appears to have included all of the four major subbasins at -32 14C Ky B.P., as evidenced by a ra- diocarbon date on mollusk shells collected on the "lakeside face of a gravel beach ridge" (Gehr, 1980:167). The elevation of this ridge is 1251 m. Latest Pleistocene (30 to -10 Ka) LAKE LAHONTAN.?The late-Pleistocene pluvial history of Lake Lahontan has been the subject of several reviews published NUMBER 33 33 in the last 50 years (e.g., Morrison, 1965; Mifflin and Wheat, 1979; Davis, 1982; Smith and Street-Perrot, 1983; Benson and Thompson, 1987a, 1987b; Dansie et al., 1988; Benson et al., 1990; Morrison, 1991b; Grayson, 1993). This current review draws heavily on those earlier reviews and adds supplemental material from subsequent research. Extensive radiocarbon dating associated with tufas, organic materials from packrat middens and other terrestrial animals, tephra, wood, shells, bulk sediments, fishing cordage, etc., have allowed a detailed and coherent picture to emerge for the past 30,000 years (especially the last 25,000 years) of lake- level prehistory for the Lahontan Basin. The pioneering radio- metric studies in the Lake Lahontan area were done on carbon- ates by Broecker and Orr (1958) and Broecker and Kaufman (1965). Many of these dates have been contested by Benson (1978), particularly those indicating an unexpectedly young age (11.5-9.5 14C Ky B.P.) for the last 1330 m Lahontan high- stand, because they were found to be inconsistent with the stratigraphic framework as reviewed in Morrison and Frye (1965). Since the work of Broecker's group, more than 100 ra- diocarbon dates have been done on Lahontan Basin materials, most of them by L. Benson and colleagues (e.g., Benson, 1978, 1994; Benson et al., 1995). This group has demonstrated that dense, lithoid tufas and aragonitic gastropods yield consistent dates, whereas massive forms of tufa (e.g., dendritic tufa) yield significantly younger ages. Through careful petrographic anal- ysis, Benson (1978) demonstrated that the latter type of tufa contains abundant evidence for the introduction of secondary carbon (e.g., shell fragments in pores and precipitated acicular carbonate) well after the initial formation of the tufa. Thus, massive tufa is usually not an appropriate material for radiocar- bon dating. After careful selection of dateable materials from his petrographic analysis and from previous radiocarbon analy- ses on deltaic wood from Born (1972), Benson (1978) pub- lished a new prehistory of the Lake Lahontan surface level through time and, on its basis, pushed the highstand date back to 13.6-11.1 14CKyB.P. The next major advance came with the addition to the data- base of several radiocarbon dates on organic material from pri- marily packrat middens, which in contrast to carbonates depos- ited under water, constrained the maximum depth of the water. When combined with the previous analyses and the addition of new tufa dates, these complementary data led to an even clearer prehistory (Thompson et al., 1986; Benson and Thompson, 1987b; Benson et al., 1990). An exhaustive review on the for- mation history of the several types of tufas in the Lake Lahon- tan area, additional tufa dates, and new dates and lake-level context for two critical chronostratigraphic horizons, namely the Wono and Trego Hot Springs tephra beds, refined this pre- history even further (Benson, 1994; Benson et al., 1995; Ben- son etal., 1997). The current state of the late Pleistocene Lake Lahontan pre- history is represented by the surface level of Lake Lahontan in the Pyramid subbasin, as shown in Figure 11. Although the lake level of the Pyramid Lake subbasin is not a precise proxy for the surface area (the parameter most directly affected by climate) of Lake Lahontan as a whole, modeling by Benson and Paillet (1989, fig. 8) showed that the overall shapes of both quantities over time are quite similar. According to radiocarbon dates on tufas, tufa-coated wood, and organic remains of a camel (see below), low to moderate lake levels were present in this subbasin from ~35 to 23-22 14C Ky B.P., although Lake Lahontan was high enough (1172 m in elevation) for Pyramid Lake to spill over into Winnemuca Lake for a few thousand years -30 14C Ky B.P. (Figure 11). The lake was low enough for a camel to become stuck in the mud near the modern (post-irrigation) shoreline during the "Wizard's Beach Recession" (Dansie et al., 1988), which lasted perhaps as long as 4 Ky from -26 to -22 14C Ky B.P. Further support for this lowstand is provided by coarse-grained alluvial depos- its underlying the Trego Hot Springs tephra in the Black Rock Desert (Ewing, 1996; F. Nials, pers. comm., 1996) and by high values of 818O and 813C measured in tufa of that approximate age (Benson et al., 1996). After the Wizard's Beach recession, abundant data summar- rized in Benson et al. (1990) indicate that the lake rose very quickly at -23-22 14C Ky B.P. to a level high enough to con- nect all of the major subbasins except for Walker Lake (Figure 11). The lake remained stable near this elevation until -16.0 14C Ky B.P. when it underwent a series of high-amplitude oscil- lations ending with the terminal recession to below the Emer- son Pass sill (1207 m) at -9.5 14C Ky B.P. The high-amplitude oscillations included a spike-like, major transgression to the maximum highstand centered somewhere between 14.0 and 13.5 14C Ky B.P. and a return to levels high enough for Pyramid Lake to spill into the Smoke Creek and Black Rock Desert ba- sins at just before 10 14C Ky B.P. This last transgression is sup- ported by dates on shells, tufa, and marl from the Carson Desert subbasin at similar elevations (Elston et al., 1988; Cur- rey, 1990; Benson et al., 1992). The overall changes in the size of Lake Lahontan, as repre- sented in Figure 11, also are recorded in the stable isotope sig- natures of tufas from the Pyramid Lake subbasin (Benson et al., 1996). Benson et al. (1996) found that when Pyramid Lake is a closed lake the values of 818O are dependent upon the balance between input through river discharge, which decreases 818O, and evaporation, which increases 818O. When Pyramid Lake is open 818O is dependent mostly upon the fraction of water input lost to spillover into the next basin; 813C was found to depend upon lake volume indirectly through its affect on the photosyn- thesis-respiration balance. Regarding the lake-level prehistory shown in Figure 11, the major lake-level minima at 26, 15.5, and 12 14C Ky B.P. were found to be associated with relatively heavy S18O and 813C values, and the opposite was found to be true for the major maxima at 14 and 10.5 14C Ky B.P. (Benson et al., 1996). Buoyed by these results, Benson et al. (unpub- lished) derived an even higher-resolution record of change in the level of Pyramid Lake with 818O measurements of the total 34 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1350 1300 1250 Lake Surface Elevation (m) 1200 1150 1100 1050 Younger Dryas- aged transgression Mud Lake Slough Sill o middle Holocene regression? (see Figure 12a) i i i r i i i i *Jet stream returns to the Lahontan Basin on its way north. The already large lake in- creases size to maximum. Darwin Pass Sill Jet stream south of Lahontan Basin. Reduced evaporation from colder temperatures maintains lake size. Emerson Pass Sill Jet stream enters the Lahontan Basin on its way p south. Increased precipitation causes the lake to increase size ~ to level of Darwin Pass sill. t i i i I i i i t I i i i i I i i ( i I i i il i I i i i i I i i i i 0 10 20 14CAge(KyB.P.) 30 FIGURE 11.?High-resolution lake-level prehistory of the Pyramid Lake subbasin for the latest Pleistocene and Holocene after Benson et al. (1995, fig. 11). The various symbols denote the type of subaqueous materials dated. See Benson et al. (1995) for detailed explanation. Abundant, unplotted terrestrial dates, mostly from packrat middens, constrain the maximum lake level from 13 I4C Ky B.P. to present. Interpretation of data in terms of supporting Antevs' jet stream model is after Benson and Thompson (1987a, 1987b) and Benson et al. (1995). inorganic carbon (TIC) fraction from a sediment core sampled at about 200-year intervals. These data, in conjunction with measurements of total organic carbon (TOC), were found to vary through time in high-frequency oscillations with square- wave morphologies that can be correlated in a general way to D-0 cycles and Heinrich events (Benson et al., unpublished). Timing of the Last Lake Lahontan Highstand: The only major point of contention during the last 20 years with the 22-9.5 14C Ky B.P. lake-level record for Lake Lahontan has been the timing of the maximum highstand (Bradbury et al., 1989; Benson, 1991; Grayson, 1993:96). Bradbury et al. (1989) argued that a Lake Lahontan that was high enough to incorpo- rate the Walker Lake subbasin could not have occurred after 23 14C Ky B.P. because a well-dated sequence of sediments from Walker Lake shows this lake to have been a saline lake or a playa for the entire interval from 23 to 5 14C Ky B.P. Further- NUMBER 33 35 more, evidence for a deep lake during this highstand interval is missing from the lake-bottom sediments at a coring location in the Black Rock Desert (Ewing, 1996), although this latter re- cord does indicate a spike-like interval of freshwater conditions at the slightly older age of 14.5 I4C Ky B.P. An attempt to resolve this controversy with a core from the Carson Sink proved to be inconclusive because of inconsistent radiocarbon dates (Thompson et al., 1990). Benson (1991), in an article that focused on replying to Bradbury et al. (1989), made a strong case for deflation as a means to remove the one meter or less thickness of lacustrine sediment he estimated to have been deposited in the Walker Lake subbasin during the approximately 200-year duration of the Lake Lahontan high- stand in this location. Although the Black Rock Desert would have had highstand water in it for a few hundred more years, deflation also could have removed the record of the highstand from the sediments studied by Ewing (1996). The lack of evidence from lake-bottom sediments from only two locations can be explained by deflation or even, perhaps, by erosion from underwater currents or slumping; however, if a highstand lake did indeed exist at this time, then there must be a lake-bottom sediment record preserved somewhere in the Lahontan subbasins. Recent studies in the Carson Desert sub- basin by graduate students from the University of Nevada, Reno (UNR), appear to have found exactly this type of evi- dence. One of these students (Cupp, 1998) found evidence for a surface elevation of the lake at -1280 m and dated it at 12.8 14C Ky B.P. The dated materials were gastropods sampled from well-bedded lacustrine sands interpreted to be a near-shore de- posit. When the Carson Desert subbasin is filled to 1280 m the effective wetness of the entire Lahontan system is approxi- mately 80 percent of maximum (Benson and Paillet, 1989). Al- though this is not high enough for the lake to reach the Adrian Pass spill point (1302 m) into Walker Lake, the lake only reached this elevation during the highstand interval. The work of K. Adams, the other UNR student, provided even more di- rect evidence for the highstand during the time indicated by Benson and others. Working in the Jessup embayment at the north margin of the Carson Desert subbasin, Adams (1998) and Adams and Wesnousky (1998) dated three shorelines, two with gastropod shells and the third with a camel bone preserved in these lakeshore deposits. The older dates, 13.28 ? 0.11 14C Ky B.P. and 13.11 ? 0.11 14C Ky B.P., were on the shells found in shoreline deposits at elevations of 1327 m and 1331 m, respec- tively. The youngest date, 13.07 ? 0.06 14C Ky B.P., was on the camel bone found in shoreline deposits at an elevation of 1338.5 m. These dates and elevations provide strong, indepen- dent support for the 14-12.8 14C Ky B.P. highstand of Benson et al. (1995). Incidentally, further evidence for this highstand is provided by the low values of 518O found by Benson et al. (1996; unpublished) in tufas and lake-bottom sediments with ages of~1414CKy B.P. Testing Paleoclimate Records with the Late Pleistocene Lake Lahontan Record: The high-resolution Lahontan lake-level record from the Pyramid Lake subbasin has been used to test models of late Pleistocene climate change. Benson et al. (1995: 25-26) demonstrated that, in response to the deflection of the jet stream during the last glacial advance, "Lake Lahontan should have experienced two lake-level maxima corresponding to the pre- and post-maximum glacial passages of the jet stream core over the Lahontan basin." The timing of these lake-level maxima are indicated by the dashed vertical lines in Figure 11. The corresponding glacial maximum was at ~20?18 14C Ky B.P., at which time the jet stream would have been de- flected to its southernmost position. In this scenario, the lake would have maintained a high level while the jet stream was south of the Lahontan area because of reduced evaporation as- sociated with lower temperatures and the probable occurrence of ice cover during the winter. This is consistent with the mod- eling results of Benson (1986), Hostetler and Benson (1990), and Hostetler (1991). The larger pluvial maximum occurred as the jet stream returned north after the glacial maximum be- cause the initial size of the lake was larger than it was when the jet stream passed through the Lahontan area on its way south (Benson et al., 1995). Thus, to a first order, the size of Lake Lahontan through time supports Antevs' jet stream migration model. The Lake Lahontan record is sufficiently detailed that it also has been used to test higher-resolution models of climate change. For example, Benson et al. (1992) used the abrupt transgression at -10 14C Ky B.P. as evidence in support of the Younger Dryas climatic interval as a global climate event. Furthermore, Benson et al. (unpublished) used the higher-res- olution 618O record from the Pyramid Lake core to show that relatively dry periods coincided with the HI, H2, and possi- bly H3 Heinrich events. This observation supports the de- scriptive model that Heinrich events represent global climate change. LAKE MODOC.?The latest Pleistocene age control in the Tulelake core is provided by the 23.2 ? 0.3 14C Ky B.P. Trego Hot Springs tephra (Bradbury, 1991; Reick et al., 1992; Ben- son et al., 1997) and a complementary date of 25.06 ? 0.32 14C Ky B.P. from slightly lower in the section. Above the Trego Hot Springs tephra is a several meter gap and then sediments containing a series of three precise but inverted radiocarbon dates in the age range from 14.9 to 9.4 14C Ky B.P. (Bradbury, 1991). From -40 Ka to a few meters above the Trego Hot Springs bed, the Tulelake core is characterized by diatoms that are in- dicative of "somewhat shallower and more alkaline conditions" (Bradbury, 1991:231). Because the Trego Hot Springs bed is exposed in littoral sediments in an outcrop 15m higher than the position of that tephra in the Tulelake core, the lake must have been -15 m deep at that time, 5 m deeper than the historical maximum (Bradbury, 1991). After that time, presumably dur- ing a time interval corresponding to OIS 2, the climate in the Lake Modoc region was dry overall but was punctuated with rapid and large fluctuations in moisture. 36 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1400 o I 1350 I Co 1300 1250 1400 1350 - 1300 - 1250 B -?-?-(1 Freidel(1993)> Various (summarized in Freidel, 1994) 10 15 20 Chewaucan 25 30 " highest shoreline J. Licciardi (unpublished) X"^ im ? X? ? basin bottom .... i .... i .. Allison (1982) " \ ' t& Negrini and Davis (1992) / ; ? ? . 1 ? . . . 1 . . . . 1 . . . . " 10 15 20 25 30 Q. H?-O- archeological index of prevalent lake level relative to sill weighted sum of dates on archeological sites in Fort Rock basin dates on high shorelines small very small standing water in northern Alkali basin 'C Ky B.P. FIGURE 13.?a, Holocene lake-level prehistory for the Pyramid Lake subbasin of the Lahontan pluvial system. Published data are from Born (1972), Benson and Thompson (1987b), Benson et al. (1990), Hattori and Tuohy (1993), Ben- son et al. (1995), and Tuohy and Dansie (1997). Evidence for the middle Holocene 1153 m lowstand is provided by a recent radiocarbon date (5.91 ? 0.125 14C Ky B.P.) on a nearly complete human skeleton with a partial cranium coated in tufa (A. Dansie and D. Tuohy, in litt, 1998). The archeological index corresponds to a qualitative estimate of lake level by Hattori and Tuohy (1993), which was based upon the frequency of occupation in the Winnemucca Caves. These caves would have been abandoned except when lake levels were above the mud lake sill and, thus, Winnemuca Lake held water. A = lake levels above the sill, B = below, and V = variable, b, Holocene lake-level prehistory for Ore- gon pluvial lakes. C = the date on a shoreline corresponding to a relatively high early Holocene Lake Chewaucan (S. Robinson, pers. comm. to Gehr, 1980); M = dates of a relatively high Lake Malheur (Gehr, 1980). Relevant data for early Holocene standing water in the northern Alkali basin is from Willig (1988) and Pinson (MS.). An index of relative availability of water on the sur- face of Lake Fort Rock basin is indicated by the weighted sum of radiocarbon dates published by Mehringer and Cannon (1994). Diamond Pond interpreta- tion is from Wigand (1987). NUMBER 33 39 Selected data for Holocene lake levels in the northwestern Great Basin are summarized in Figure 13. LAKE LAHONTAN.?Because Holocene lake-level changes were too small in magnitude to reach most of the sills between major subbasins, the case for each basin is presented sepa- rately. Holocene lake-level minima in the Pyramid Lake subba- sin are constrained by radiocarbon dates on deltaic wood and tufas; maxima are constrained by organic materials from pack- rat middens and archeological sites (Figure 13a). The lake- level curve shown in Figure 1 la is considered to be smoothed because it reflects general trends in lake levels as indicated by, for example, the occupation of archeological sites. The reader should bear in mind that there are several excursions within these general trends to higher or lower lakes that probably rep- resent higher-frequency lake-level changes. For a related dis- cussion, see Hattori and Tuohy (1993) and Mehringer and Can- non (1994). The Holocene record in the Pyramid Lake subbasin begins with the recession that terminated the relative highstand associ- ated with the Younger Dryas stade (Benson et al., 1992; Ben- son et al., 1995). The lowest lake level during this recession (-1153 m) is indicated by two radiocarbon dates. One of these (9.66 ? 0.13 14C Ky B.P.) was on a piece of fishing line made from sagebrush bark found at an archeological site on the northwestern shoreline of Pyramid Lake (Tuohy, 1988). The other (9.5-9.1 14C Ky B.P.) was on a partial human skeleton found at approximately the same elevation at the north end of Pyramid Lake (Tuohy and Dansie, 1997). Thereafter, the lake probably rose to a level above 1168 m, as indicated by an age of 8.80 ? 0.09 14C Ky B.P. on a piece of wood from sediments interpreted to have been deposited on a delta slope (Born, 1972). The occupation of Winnemuca Caves from -8.4 to 7.8 14C Ky B.P. indicates that water levels high enough for Pyramid Lake to spill into Winnemuca Lake were probably sustained, at least intermittently, until ~8 14C Ky B.P. (Hattori, 1982; Hattori and Tuohy, 1993). Except for a few radiocarbon dates from the Winnemuca Caves (Hattori and Tuohy, 1993), no evidence has been found for a long-lived lake above historical levels for the time inter- val between -8 and 5 14C Ky B.P. Davis (1982) and Benson and Thompson (1987b) argued that such lack of radiocarbon evidence does not indicate a desiccation of Pyramid Lake. They noted that elevations of the lake far below the historical (before modern agricultural water diversion) lowstand level of -1177 m would have killed two species of endemic fish, the cui-ui (Chasmistes cujus Cope) and the Lahontan cutthroat trout (Oncorhynchus clarki henshawi Richarson). This argu- ment is supported by the failure of these fish to survive during the present century without help from hatcheries after the lake was lowered artificially to below 1160 m by irrigation-related diversion of the Truckee River. The lowest possible level for the survival of these fish is, of course, constrained by the 1153 m regression at 9.66 14C Ky B.P., as discussed in the previous paragraph. The lake reached this low level at least one other occasion, in the middle Holocene. Evidence for this 1153 m lowstand is provided by a recent radiocarbon date (5.91 ? 0.125 14C Ky B.P.) on a nearly complete human skeleton with a partial cranium coated in tufa found at Wizard's Beach at Pyra- mid Lake (A. Dansie and D. Tuohy, in litt., 1998). A rise to levels as high as 1185 m during the late Holocene is indicated by abundant dates for inorganic carbonates and wood depos- ited near lake-surface level in a fluvio-deltaic environment (Born, 1972; Benson et al., 1995). The only direct information on Holocene lake levels with quantitative age control in the Carson Sink area is provided by the tephrochronology of Davis (1978). A Mazama-age lake level was inferred by the identification of the Mazama tephra (6.845 ? 0.05 14C Ky B.P.) and its precursor, the Tsoyawata bed (7.015 ? 0.045 14C Ky B.P.), at two localities in this subbasin. The two tephra were found in the alluvium in the West Hum- boldt Range at an elevation of 1204 m. They were also identi- fied in the shallow lake or marsh sediments described by Morrison as Section 22a (Morrison, 1964:128) east of Fallon, Nevada, near the Stillwater Reservoir at an elevation of ?1193 m, which is only one or two meters above the modern low point in the entire Carson Sink subbasin (Davis, 1978; Elston et al., 1988). Thus, in the few hundred years of time spanned by the volcanic eruptions associated with these two tephra, the lake in the Carson Desert subbasin was less than 10 m deep (Davis, 1982). In contrast, sometime after the eruption that cre- ated the Tarupah Flat bed (1500-600 years old), the lake had risen to a level of 1207 m, as evidenced by the wave-cut plat- forms crosscutting this ash layer at two locations near Salt Wells, Nevada (Davis, 1978). Walker Lake was interpreted to have been an ephemeral playa lake during the early Holocene (Bradbury et al., 1989; Benson et al., 1991). This interpretation was based upon the as- semblage of microfossils (diatoms, ostracodes, aquatic pollen), lithology, and an abrupt change in sedimentation rates. After -5.0-4.5 14C Ky B.P. the lake quickly filled to a depth as great as 90 m. This rapid transgression has been attributed to the di- version of the Walker River into the Walker Lake subbasin from its previous destination, the Carson Desert subbasin (King, 1978, 1993; Davis, 1982; Morrison and Davis, 1984a, 1984b; Bradbury et al., 1989; Benson et al., 1991). Based upon similar evidence to that mentioned above and also on the stable isotope analyses of Benson et al. (1991), a coherent lake-level prehistory of this subbasin has been determined for the latest Holocene. The coherence of this prehistory with that of nearby Mono Lake (Stine, 1990) gives confidence that late Holocene lake levels in the Walker Lake subbasin accurately reflect cli- mate change (Bradbury et al., 1989; Benson et al., 1991). There is some controversy as to whether an extreme low- stand of the lake from -2.5 to 2.0 14C Ky B.P. was caused by climate change or by another diversion of the Walker River. Bradbury et al. (1989) contended that climate change is the more likely cause because the very existence of even a shallow lake would have required some recharge from the Walker River 40 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES into the basin, where evaporation greatly exceeded precipita- tion, and because climate records from nearby localities in the Great Basin also indicate drier climates during this time inter- val. They further noted that Walker Lake filled more slowly af- ter 2.0 14C Ky B.P., as evidenced by step-like increases in key ostracode and diatom taxa, than it did after the previous desic- cation. Such a slow transgression is more consistent with cli- mate change than with the sudden diversion of a major river, like the Walker River, back into this subbasin. Benson et al. (1991), on the other hand, preferred river diver- sion over climate change as a cause for this near-desiccation event because (1) the usually coherent discharge of the Truckee and Walker Rivers would not have allowed flow in the Walker River to almost cease while the Truckee was flowing enough to keep Pyramid Lake above the level required for the spawning of cui-ui and emerald trout, (2) the Walker River must have continued to flow somewhere to have allowed the survival of the cutthroat trout while the lake-dwelling cui-ui population was wiped out as Walker Lake nearly dried out, and (3) cur- rently, discharge from local streams is sufficient to maintain Walker Lake at very shallow levels (< 1 m) even in the absence of the Walker River. LAKE MODOC.?Despite the poor age control in the Ho- locene part of the Tulelake core, Bradbury (1991:233) asserted convincingly that "the strong fluctuations between planktonic and benthic diatoms reflect real lake level changes that are im- portant for resolving paleoclimatic and archaeological ques- tions." Clearly, this record will become a valuable one with improved age control. Even as it stands, it provides important information surrounding the Tsoyawata tephra (7.015 ? 0.045 14C Ky B.P.), which was identified in the Tulelake core (Reick et al., 1992). Two complementary 14C dates were obtained within ~1 m below this tephra in the correct stratigraphic or- der, thereby confirming the ash correlation. The dominance of Aulacoseira ambigua in the vicinity of the Tsoyawata tephra bed suggests the presence of a relatively deep lake at this time, perhaps the 20 m lake suggested by an archeological study at Nightfire Island (Sampson, 1985; Bradbury, 1991). Above the Tsoyawata tephra, A. ambigua is replaced by Fragilaria sp. This indicates a lower lake in the middle Holocene. The return of A. ambigua at the top of the core marks a deepening lake at a poorly constrained time, perhaps by ~3 Ka (Bradbury, 1991). OTHER OREGON LAKES.?Most of the evidence for Ho- locene lake-level changes in Oregon lakes has been reviewed by Freidel (1993, 1994), McDowell (1992), Grayson (1993), McDowell and Dugas (1993), and Aikens and Jenkins (1994). A summary of these reviews with a minor amount of supple- mental material follows. The early Holocene appears to have been characterized by pervasive shallow lakes throughout the Oregon portion of the Great Basin (Figure 136). Four very shallow lakes were re- ported to have come and gone during the early Holocene within the confines of pluvial Lake Alkali (Figure 1). The first of these lakes would have been ~0.5 m deep in the Dietz sub- basin, located in the northern part of the Alkali Lake basin. The shore of this lake, named Lake Koko, was occupied by a Clovis-age culture. This conclusion was based upon fluted- point assemblages that were found distributed along a linear trend, which was interpreted to be a shoreline (Fagan, 1988; Willig, 1988; Willig and Aikens, 1988). Another assemblage of artifacts, which was characterized by large, stemmed, shoul- dered, and lanceolate points, was found distributed around a slightly higher shoreline and has been attributed (by Willig, 1988) to the Western Stemmed Point culture. Willig (1988) in- terpreted this association as a later culture that occupied the shores of the younger, slightly deeper (~2 m) Sand Ridge Lake. Thus, she concluded that Lake Koko existed sometime during the time interval attributed to the Clovis-age culture (11.5-11.0 14C Ky B.P.) and that Sand Ridge Lake existed dur- ing the time interval attributed to Western Stemmed Point tra- dition (10.8-7.0 14C Ky B.P.). Although the overlap in age of these two cultures may be larger than implied herein (e.g., Beck and Jones, 1997:196-197), the observation of Lake Koko shorelines covered by Sand Ridge Lake sediments strongly supports an older age for Lake Koko. Charcoal frag- ments associated with Sand Ridge Lake were radiocarbon- dated (Willig, 1988) at 9.61 ? 0.1 14C Ky B.P., providing fur- ther support for the occupation of Sand Ridge Lake by a West- ern Stemmed Point culture. Willig (1988) postulated the exist- ence of two younger lakes, an intermediate-sized Lake Delainie and a deeper (~4 m) Big Cut Lake, based upon the su- perposition of shorelines and lake sediments. The ages of these two younger lakes are not constrained. The Dietz site lake-level history has since been reinterpreted by Pinson (MS.). In this later study, the stratigraphy was inter- preted to consist of deep-lake sediments covered by an eolian deposit, which in turn was covered by a clay-silt layer that was interpreted to have been deposited in a marsh, based upon its internal distribution of gastropods and organic material. Over this last unit is another eolian deposit from which Willig's char- coal date was obtained. A second charcoal date of 9.43 ?0.08 14C Ky B.P. from this horizon confirmed Willig's date (Pinson (MS.)). This eolian unit is also the one from which most of the buried, early Holocene archeological materials were derived. Notably, the top of this unit appears to have a nascent soil hori- zon directly over which lies the Mazama pumice. Pinson (MS.) argued convincingly that this stratigraphy can be interpreted as the drying up of the terminal, deep Pleistocene lake at -10 14C Ky B.P. and that thereafter only a shallow marsh existed. This marsh, in turn, dried completely by ~9 14C Ky B.P., and there has been no standing water in the Dietz subbasin since that time. Fort Rock Basin has been the location of several important late Pleistocene and early Holocene archeological sites (Bed- well, 1973; Grayson, 1993; Aikens and Jenkins, 1994; Beck and NUMBER 33 41 Jones, 1997), some of which reveal a great deal about paleo- lake levels (e.g., Figure 136). Early interpretations, based upon evidence from the Connley Caves excavation, pointed to con- tinuous human occupation for the time from -11 to 3 14C Ky B.P., with the notable exception of the interval from 7.2 to 4.4 14C Ky B.P. (Bedwell, 1973). The apparent disappearance of human occupation was initially interpreted, at least in part, to have been the result of a catastrophic impact on the population caused by the eruption of Mount Mazama (Bedwell, 1973); however, it was later attributed to the shrinkage of large, nearby bodies of water, as evidenced by the near disappearance of waterfowl remains in the archeological middens of these caves (Grayson, 1979, 1993). Since the work at Connley Caves, Aikens and Jenkins (1994) conducted a new, basin-wide project. They concluded that when more sites are considered the hypothesis for the mid- dle Holocene abandonment of the Fort Rock Basin begins to fall apart. For example, Mehringer and Cannon's (1994) inves- tigation of "Kelly's Site" in the Fort Rock Basin indicated that people not only occupied the basin on occasion during that time interval, but they were eating waterfowl. Analyses of ra- diocarbon dates from several additional studies done as part of this basin-wide project, however, revealed shorter-term inter- vals of relatively lower population density at the bottom of Fort Rock Basin that were presumably related to shorter intervals of lower lake levels and dried-up marshes (Figure 136). These pe- riods are roughly 7-5 I4C Ky B.P. (Jenkins, 1994; Mehringer and Cannon, 1994) and 3-2 14C Ky B.P. (Jenkins, 1994). Al- though much more work has to be done to precisely define these dry/wet intervals, it is interesting to note that the onset of wetter conditions at ~5 14C Ky B.P. implied by the above chro- nology coincides almost exactly with the transition from ephemeral marshes to open water at Diamond Pond/Malheur Maar in eastern Oregon south of pluvial Lake Malheur (Wi- gand, 1987). In addition to dispelling the notion of a 3000-year long mid- dle Holocene abandonment in the Fort Rock Basin (Bedwell, 1973), new work by P. McDowell and J. Benjamin suggests that the Neopluvial (post-altithermal) lake postulated by Alli- son (1979) also did not exist in the Fort Rock basin. Their con- clusions were based upon observations of "extensive dunes with elevations low enough and dates early enough that they would have been destroyed or badly eroded had they been in- undated by a basin-wide Neopluvial lake" (Greenspan, 1994: 488; P. McDowell, pers. comm. with Greenspan, 1994). The subbasin remnants of Lake Chewaucan also appear to have been significantly higher during the early Holocene than during modern times. The evidence for this includes a personal communication cited in Gehr (1980) from S. Robinson of the USGS in which Robinson reported a radiocarbon age of 9.39 ? 0.045 14C Ky B.P. on a tufa-coated beach gravel found 20-22 m above the modern level of Lake Abert. Summer Lake rose in the late Holocene, as evidenced by a shoreline 12 m above the bottom of the Summer Lake basin. This shoreline cuts across sand dunes consisting of pumice from the Mount Mazama eruption (Allison, 1982). There is no age control constraining the minimum age of this feature. There is also evidence that Lake Abert rose again, probably in the late Holocene. This evi- dence consists of a higher shoreline (18 m above the modern lake) that truncates presumably young colluvium (Allison, 1982). This feature has not been dated. Pluvial Lake Malheur rose to levels high enough to join all of the subbasins into one lake and perhaps to overflow into the Columbia River system at least three times during the early Holocene (Figure 136). Two of these episodes are dated at 9.62 + 0.05 14C Ky B.P. and 8.68 ? 0.55 14C Ky B.P. (Gehr, 1980). No record for the surface levels of this lake exists for the remainder of the Holocene. TESTING HOLOCENE PALEOCLIMATE MODELS WITH NORTH- WESTERN GREAT BASIN PLUVIAL RECORDS.?The balance of evidence from Figure 13 indicates that pluvial lake levels were slightly higher in the northwestern Great Basin during the early and late Holocene than they were in the middle Holocene. This set of observations, at first glance, appears to support Antevs' Neothermal climate model (Figure 2a), although the observa- tions call for a slightly older Altithermal age than was antici- pated by Antevs (Grayson, 1993). As more evidence has accu- mulated, however, it has become increasingly apparent that exceptions to the rule are abundant and that data must be spa- tially and temporally averaged to conform to the Neothermal model (Grayson, 1993; Mehringer and Cannon, 1994). In par- ticular, the Holocene appears to be characterized by short-pe- riod fluctuations in climate of significant magnitude superim- posed on the long-term averages (e.g., Benson et al., 1991; Mehringer and Cannon, 1994). Such variability is consistent with the high-frequency variations in solar luminosity inferred by cosmogenic isotope concentrations (Figure 2); however, there is no obvious correlation between the observed climate change presented in Figure 13 and those predicted by the lumi- nosity proxies shown in Figure Id. In addition, there is no evi- dence in northwestern Great Basin pluvial records, as of yet, for the early Holocene "cooling event" at 8.4-8.0 cal Ky B.P. (Figure 2c). The lack of correlation between northwestern Great Basin pluvial records and the predictions of Holocene climate models probably results from deficiencies in the collective data set and the lack of specific climate changes inferred from the models. The sampling of Holocene paleoclimate data from individual pluvial lake basins in the northwestern Great Basin is usually too sparse for the detection of high-frequency climate change. Furthermore, little modeling has been done to predict the ef- fects of low-amplitude changes in Holocene climate on pluvial lake levels in this region. For example, one recent study on the geochemistry of Great Barrier Reef corals (Gagan et al., 1998) has demonstrated that tropical oceans in the Pacific were ~1? C warmer -5.35 cal Ky B.P. than they are today and that they 42 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES were enriched in 18O, implying increased levels of evapotrans- piration. This intriguing result is consistent with intensified El Nino events during the middle Holocene, which in turn could explain the apparent increased variability in lake levels during this time (e.g., Grayson, 1993; Mehringer and Cannon, 1994). Before such a hypothesis can be anything more than specula- tive, atmospheric circulation models with the ability to predict the resulting low-amplitude changes in lake levels need to be constructed, and extremely high-resolution pluvial lake records need to be available to test the applicability of this idea to the northwestern Great Basin. Summary and Future Trends SUMMARY OF MAJOR RESULTS Since the time of Antevs, many contributions have been made toward the understanding of climate change from studies of pluvial lakes in the northwestern Great Basin. The following list is a sample of these contributions. 1. Both of the global climate events proposed for the Plio/ Pleistocene boundary at ~2.5 Ma and -1.65 Ma have been ob- served in several paleoclimate indicators from the Tulelake core (Adam et al., 1989). 2. Variations in the diatom stratigraphy of the Tulelake core, which have been interpreted to represent cycles between gla- cial and interglacial climates, change frequency in accordance with the evolution through time of the spectral character of ma- rine oxygen isotope curves and other marine paleoclimate proxies (Bradbury, 1991, 1992). Specifically, periods of 400 Ky and 41 Ky appear to dominate the climate variations prior to ~900 Ka; thereafter, 100 Ky variations dominate. 3. Evidence for an extremely large lake in the Lahontan and surrounding basins during the largest excursion toward glacial oxygen isotope values in the Quaternary (OIS 16) supports a connection between the sizes of ice sheets and the sizes of plu- vial lakes in the northwestern Great Basin (Reheis, 2002). 4. Milankovitch-like periodicities were reported in the spec- trum of a middle to late Pleistocene PSV record for Lake Chewaucan (Negrini and Verosub, 1987; Negrini et al., 1988). Subsequent improvements on the age control of the Chewau- can sediments have disproved Milankovitch-like periodicities as the origin of these PSVs (Negrini et al., 1994). 5. Several lake-level proxies in a 250-20 Ka paleoclimate re- cord from Lake Chewaucan support an in-phase relationship between increases/decreases in lake levels and advances/re- treats of the North American ice sheet as inferred from marine OIS 7-3 (Davis, 1985; Berger and Davis, 1992; Negrini et al., 1994; Erbes, 1996). This represents a significant test of An- tevs' jet stream migration model because the age control of the Lake Chewaucan record is entirely independent of a priori as- sumptions of correlations with climate change. 6. Bradbury (1991, 1992) explained the unusual occurrence of a pluvial diatom assemblage during OIS 5 as being the result of a tectonic or geomorphologic change that affected the local Tule Lake subbasin. In this review paper, late Pleistocene Tule- lake stratigraphy has been reinterpreted, resulting in a signifi- cantly younger age for the Pumice Castle tephra (Figure 10). Thus, the diatom assemblage is determined to be a likely ex- pression of OIS 4. 7. OIS 4 is expressed strongly in the northwestern Great Ba- sin by the presence of Pumice Castle set tephra layers in plu- vial-lake sediments from three lake basins (Lake Modoc, Lake Chewaucan, Lake Malheur). 8. An exceptionally precise and accurate lake-level prehis- tory for Lake Lahontan during the latest Pleistocene is consis- tent with increases in lake size both when the jet stream passes by the Lake Lahontan area on its way south during OIS 2 ice sheet advance and also when it passes by on its way north dur- ing ice sheet retreat (Benson and Thompson, 1987a, 1987b; Benson etal., 1995). 9. Higher-frequency climate-change events inferred by stable isotope and total organic carbon analyses of Pyramid Lake bot- tom sediments have the same square-wave type morphology and duration as the D-0 cycles and Heinrich events observed in paleoclimate records from marine sediments and ice cores (Benson et al., unpublished). 10. The observation by Allison (1945) of several meters of Lake Chewaucan sediments above the Mount Mazama ash was a key piece of data that inspired Antevs' theory of a moist Anathermal age. Improvements in tephrochronology and abso- lute-age dating demonstrated that this ash was, instead, the much older Mount St. Helens Cy tephra (Allison, 1966a; Davis, 1985). 11. Bedwell (1973) noted the lack of middle Holocene arche- ological remains in the Connley Caves of the Lake Fort Rock basin. An analysis of the associated archeofauna revealed the near disappearance of waterfowl remains in the area, prompt- ing Grayson (1979, 1993) to hypothesize a middle Holocene human abandonment of this basin because of lowered water levels in a nearby lake/marsh complex. Subsequent work on dune sites in the area demonstrated that Lake Fort Rock basin was indeed occupied during the middle Holocene and that the presence offish and waterfowl remains ... challenge notions of a long uninterrupted mid-Holocene drought" (Mehringer and Cannon, 1994:321). 12. As more records of Holocene lake-level changes have ac- cumulated, it is becoming more and more apparent that high- frequency change characterized this entire period of time, as suggested by Mehringer and Cannon (1994). Against this back- ground, some long-term trends have emerged (e.g., lower lake levels during the middle Holocene) that loosely fit Antevs' Neothermal age divisions. Neither the long-term changes nor the high-frequency changes appear to correlate unambiguously with newer models of Holocene climate change, such as vari- ability in solar luminosity. NUMBER 33 43 REMAINING PROBLEMS Despite the progress that has been made during the last 50 years, the potential for understanding climate change from northwestern Great Basin pluvial lake records has hardly been tapped. For example, long-term records from pluvial lakes other than Lake Modoc need to be obtained to determine if the climate record from the Tulelake core is representative of the rest of the northwestern Great Basin, considering its proximity to the Pacific Ocean and the fact that Lake Modoc drained into the Pacific after reaching only modest depths. The source vents and correlative proximal volcanic rocks have yet to be identified for many of the water-lain tephra beds in the northwestern Great Basin. Further work of this nature will improve the overall chronology of the region because it is inevitably easier to apply radiometric- and radiogenic-age methods to proximal volcanic rocks than to water-lain ashes. In addition, promising dating techniques, such as luminescence methods, should be used more frequently on the water-lain ashes. The only well-constrained pluvial lake-size record in the northwestern Great Basin is that of the Pyramid Lake subbasin of Lake Lahontan for the time interval from -30 to 9 14C Ky B.P.; the Walker Lake subbasin may be well-constrained for the late Holocene. The provocative results from these prehistories need to be tested with records of similar resolution and accu- racy from other basins. The present state of the composite Holocene record does not allow accurate testing of models for global climate change dur- ing the Holocene. This problem can only be alleviated with well-dated, high-resolution, continuous Holocene records from several basins within the northwestern Great Basin. THE NEED FOR COMPREHENSIVE CORING PROGRAMS Most of the problems mentioned above could be solved with a comprehensive program of coring bottom sediments throughout the northwestern Great Basin. Although core stud- ies can never completely replace field-based surface studies, at this stage in the study of the northwestern Great Basin an ex- tensive coring program would most efficiently provide a wealth of hypotheses. These hypotheses could be tested later by well-focused studies on the extensive exposures of lake de- posits in this region, which represent much of the Quaternary period. There are several indicators that point to the success of such an extensive coring program, many of which are discussed in this review. The work of Benson and others (see "Latest Pleis- tocene," above) has demonstrated that geochemical climate proxies in cores provide estimates of lake sizes through time that are consistent with independently derived lake-size prehis- tories from surface materials and that these proxies can be ac- quired relatively rapidly with high resolution. The consistency of paleoclimate results from geochemical measurements with other bottom-sediment proxies, such as the distribution of fau- nal and floral taxa, lithostratigraphy, granulometry, and envi- ronmental magnetism, points to a variety of useful techniques that may be applicable for each core. Unprecedented age control for terrestrial records can be pro- vided by an already rich, tephrochronological database, which will only improve through iteration as results from additional cores are added to a composite data set. Under ideal condi- tions, the promising and complementary application of paleo- magnetic secular variation often can provide nearly continu- ous relative age control, which conceivably can be tied directly to marine paleoclimate records. This method, too, will grow in usefulness as a reference PSV record continues to de- velop. Advances in luminescence and radiocarbon methods have contributed greatly to our understanding of northwestern Great Basin pluvial lakes and will continue to do so in the fu- ture. Coring programs should be preceded by careful consider- ation of coring sites optimally selected on the basis of sound sedimentological models and geophysical studies. Several lo- calities around most basins should be cored in order to sample and compare the many varieties of depositional environments observed within each basin. At least two cores should be taken at each coring site to guard against localized synsedimentary deformations, loss of core samples during withdrawal, etc. Cores should be oriented azimuthally whenever possible to en- able the accurate measurement of magnetic declination for use in PSV studies. Finally, and perhaps most important, a full ar- ray of lithological, paleontological, geochemical, tephrochro- nological, and magnetic studies should be done on these cores in a timely fashion and with careful management of data and archival of core splits. Acknowledgments Athough the content and interpretations expressed herein are ultimately the responsibility of the author, this work benefited greatly from conversations or written correspondences with D. Adam, K. Adams, L. Benson, P. Bradbury, A. Cohen, A. Dan- sie, J.O. Davis, R. Elston, D. Erbes, F. Foit, R. Forester, D. Fre- idel, K. Gobalet, T. Jones, T. Lowenstein, S. Lund, F. Nials, M. Palacios-Fest, A. Sarna-Wojcicki, R. Thompson, and P. Wigand. Also, K. Edgett, P. Wigand, and M. Sutton provided copies of key pieces of literature cited in this paper. I am grate- ful for the use of unpublished data provided by J. Licciardi and S. Lund. M. Quilliam assisted with library research and proof- reading. The final version of this paper benefited greatly from careful review by A. Cohen, T. Lowenstein, J. Luhr, and an anonymous reviewer. 44 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Literature Cited Adam, D.P. 1993. Long Quaternary Records from the Northwestern Basin and Range Province, California and Oregon. In L.V. Benson, editor, Proceed- ings of the Workshop "Ongoing Paleoclimatic Studies in the North- ern Great Basin," Reno, Nevada, May 1993. United States Geological Survey Circular, 1119:7?8. Adam, D.P., A.M. Sarna-Wojcicki, H.J. Rieck, J.P. Bradbury, W.E. Dean, and R.M. Forester 1989. Tulelake, California: The Last 3 Million Years. Palaeogeography, Palaeoclimatology, Palaeoecology, 72:89-103. Adams, K.D. 1998. Late Quaternary Pluvial History, Isostatic Rebound, and Active Faulting in the Lake Lahontan Basin, Nevada and California. 169 pages. Doctoral dissertation, Department of Geology, University of Nevada, Reno. Adams, K.D., and S.G. Wesnousky In press. Shoreline Landforms and Processes and the Age of the Lake Lahontan Highstand in the Jessup Embayment, Nevada. Bulletin of the Geological Society of America. Aikens, CM. 1983. Environmental Archaeology in the Western United States. In H.E. Wright, Jr., editor, Late-Quaternary Environments of the United States, Volume 2: The Holocene, pages 239-251. Minneapolis: Uni- versity of Minnesota Press. Aikens, CM., and D.L. Jenkins 1994. Environment, Climate, Subsistence, and Settlement: 11,000 Years of Change in the Fort Rock Basin, Oregon. In CM. Aikens and D.L. Jenkins, editors, Archaeological Researches in the Northern Great Basin: Fort Rock Archeology since Cressman. Anthropological Pa- pers, 50:1-20. Eugene: University of Oregon Department of An- thropology and State Museum of Anthropology. Alley, R.B., PA. Mayewski, T. Sowers, M. Stuiver, K.C Taylor, and P.U. Clark 1997. Holocene Climatic Instability: A Prominent Widespread Event 8200 Yr Ago. Geology, 25:483-486. Allison, I.S. 1945. Pumice Beds at Summer Lake, Oregon. Bulletin of the Geological Society of America, 56:798-808. 1946. Early Man in Oregon: Pluvial Lakes and Pumice. Scientific Monthly, 62:63-65. 1966a. Pumice Beds at Summer Lake, Oregon?A Correction. Bulletin of the Geological Society of America, 77:239-330. 1966b. Fossil Lake, Oregon: Its Geology and Fossil Faunas. Oregon State Monographs, Studies in Geology, 9: 48 pages. Oregon State Univer- sity Press. 1979. Pluvial Fort Rock Lake, Lake County, Oregon. State of Oregon, De- partment of Geology and Mineral Industries Special Paper, 7: 72 pages. 1982. Geology of Pluvial Lake Chewaucan, Lake County, Oregon. Oregon State Monographs, Studies in Geology, 11: 79 pages. Corvallis, Ore- gon: Oregon State University Press. Andrews, J.T. 1987. The Late Wisconsin Glaciation and Deglaciation of the Laurentide Ice Sheet. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, volume K-3:13-38. Boulder, Colorado: Geological Society of America. Antevs, E. 1925. On the Pleistocene History of the Great Basin. Publications of the Carnegie Institution of Washington, 352:51-114. 1938. Postpluvial Climatic Variations in the Southwest. Bulletin of the American Meteorological Society, 19:190-193. 1948. Climatic Changes and Pre-White Man. Bulletin of the University of Utah, 38(20):168-191. 1952. Cenozoic Climates of the Great Basin. Geologische Rundschau, 40: 94-108. 1955. Geologic-Climatic Dating in the West. American Antiquity, 20(4): 317-335. 1962. Transatlantic Climatic Agreement versus C-14 Dates. Journal of Geology, 70:194-205. Bacon, C.R. 1983. Eruptive History of Mount Mazama and Crater Lake Caldera, Cas- cade Range, U.S.A. Journal of Volcanology and Geothermal Re- search, 18:57-115. Baksi, A.K., V. Hsu, M.O. McWilliams, and E. Farrar 1992. 40Ar/39Ar Dating of the Brunhes-Matuyama Geomagnetic Reversal. Science, 256:356-357. Banfield, J.F., B.F. Jones, and D.R. Veblen 1991 a. An AEM-TEM Study of Weathering and Diagenesis, Abert Lake, Oregon; I: Weathering Reactions in the Volcanics. Geochimica et Cosmochimica Acta, 55:2781-2793. 1991b. An AEM-TEM Study of Weathering and Diagenesis, Abert Lake, Oregon; II: Diagenetic Modification of the Sedimentary Assem- blage. Geochimica et Cosmochimica Acta, 55:2795-2810. Baumhoff, M.A., and R.F. Heitzer 1965. Postglacial Climate and Archaeology in the Desert West. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 697-708. Princeton, New Jersey: Princeton University Press. Beck, C, and G.T. Jones 1997. The Terminal Pleistocene/Early Holocene Archaeology of the Great Basin. Journal of World Prehistory, 11(2): 161?236. Bedwell, S.F. 1973. Fort Rock Basin Prehistory and Environment. 189 pages. Eugene: University of Oregon Books. Benson, L.V. 1978. Fluctuation in the Level of Pluvial Lake Lahontan during the Last 40,000 Years. Quaternary Research, 9:300-318. 1986. The Sensitivity of Evaporation Rate to Climate Change?Results of an Energy-Balance Approach. United States Geological Survey Wa- ter-Resources Investigations Report, 86-4148: 40 pages. 1991. Timing of the Last Highstand of Lake Lahontan. Journal ofPaleo- limnology, 5:115-126. 1994. Carbonate Deposition, Pyramid Lake Subbasin, Nevada; 1: Se- quence of Formation and Elevational Distribution of Carbonate De- posits (Tufas). Palaeogeography, Palaeoclimatology, Palaeo- ecology, 109:55-87. Benson, L.V, D.R. Currey, R.I. Dorn, K.R. Lajoie, C.G. Oviatt, S.W. Robin- son, G.I. Smith, and S. Stine 1990. Chronology of Expansion and Contraction of Four Great Basin Lake Systems during the Past 35,000 Years. Palaeogeography, Palaeocli- matology, Palaeoecology, 78:241-286. Benson, L.V, D. Currey, Y. Lao, and S. Hostetler 1992. Lake-Size Variations in the Lahontan and Bonneville Basins be- tween 13,000 and 9,000 14C Yr B.P. Palaeogeography, Palaeoclima- tology, Palaeoecology, 95:19-32. Benson, L.V, M. Kashgarian, and M. Rubin 1995. Carbonate Deposition, Pyramid Lake Subbasin, Nevada; 2: Lake Levels and Polar Jet Stream Positions Reconstructed from Radiocar- bon Ages and Elevations of Carbonates (Tufas) Deposited in the Lahontan Basin. Palaeogeography, Palaeoclimatology, Palaeo- ecology, 117:1-30. Benson, L.V, S.P. Lund, J.W. Burdett, M. Kashgarian, T.P. Rose, J.P. Smoot, and M. Schwartz 1998. Correlation of Late-Pleistocene Lake-Level Oscillations in Mono NUMBER 33 45 Lake, California, with North Atlantic Climate Events. Quaternary Research, 49:1-10. Benson, L.V., P.A. Meyers, and R.J. Spencer 1991. Change in the Size of Walker Lake during the Past 5000 Years. Palaeogeography, Palaeoclimatology, Palaeoecology, 81:189-214. Benson, L.V., and F.L. Paillet 1989. The Use of Total Lake-Surface Area as an Indicator of Climatic Change: Examples from the Lahontan Basin. Quaternary Research, 32:262-275. Benson, L.V., and Z. Peterman 1995. Carbonate Deposition, Pyramid Lake Subbasin, Nevada; 3: The Use of 87Sr Values in Carbonate Deposits (Tufas) to Determine the Hy- drologic State of Paleolake Systems. Palaeogeography, Palaeocli- matology, Palaeoecology, 119:201-213. Benson, L.V., J.P. Smoot, M. Kashgarian, A. Sarna-Wojcicki, and J.W. Burdett 1997. Radiocarbon Ages and Environments of Deposition of the Wono and Trego Hot Springs Tephra Layers in the Pyramid Lake Sub Ba- sin, Nevada. Quaternary Research, 47:251-260. Benson, L.V., and R.S. Thompson 1987a. The Physical Record of Lakes in the Great Basin. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, vol- ume K-3:241-260. Boulder, Colorado: Geological Society of Amer- ica. 1987b. Lake-Level Variation in the Lahontan Basin for the Past 50,000 Years. Quaternary Research, 28:69-85. Benson, L.V., L. White, and R. Rye 1996. Carbonate Deposition, Pyramid Lake Subbasin, Nevada; 4: Com- parison of the Stable Isotope Values of Carbonate Deposits (Tufas) and the Lahontan Lake-Level Record. Palaeogeography, Palaeocli- matology, Palaeoecology, 122:45-76. Berger, A.L., M.F. Loutre, and J. Laskar 1992. Stability of the Astronomical Frequencies over the Earth's History for Paleoclimate Studies. Science, 255:560-566. Berger, G.W. 1988. Dating Quaternary Events by Luminescence. In D.J. Easterbrook, editor, Dating Quaternary Sediments. Special Papers of the Geolog- ical Society of America, 227:13-50. 1991. The Use of Glass for Dating Volcanic Ash by Thermoluminescence. Journal of Geophysical Research, 96:19,705 -19,720. Berger, G.W., and A.J. Busacca 1995. Thermoluminescence Dating of Late Pleistocene Loess and Tephra from Eastern Washington and Southern Oregon and Implications for the Eruptive History of Mount St. Helens. Journal of Geophysical Research, 100:22,361-22,374. Berger, G.W., and J.O. Davis 1992. Dating Volcanic Ash by Thermoluminescence: Test and Applica- tion. Quaternary International, 13/14:127-130. Bond, G., W. Broecker, S. Johnsen, J. McManus, L. Labeyrie, J. Jouzel, and G. Bonani 1993. Correlations between Climate Records from North Atlantic Sedi- ments and Greenland Ice. Nature, 365:143-147. Bond, G.C., and R. Lotti 1995. Iceberg Discharges into the North Atlantic on Millennial Time Scales during the Last Glaciation. Science, 267:1005-1010. Booth, D.B. 1987. Timing and Processes of Deglaciation along the Southern Margin of the Cordilleran Ice Sheet. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Degla- ciation. The Geology of North America, volume K-3:13-38. Boul- der, Colorado: Geological Society of America. Born, S.M. 1972. Late Quaternary History, Deltaic Sedimentation and Mud-Lump Formation at Pyramid Lake, Nevada. 97 pages. Reno, Nevada: Cen- ter for Water Resources Research, Desert Research Institute. Botkin, S., and K.R. Carembelas 1992a. Chapter 7: 35HA1911. In C. Raven and R.G. Elston, editors, Land and Life at Malheur Lake: Preliminary Geomorphological and Ar- chaeological Investigations. Cultural Resource Series, 8:57-86. Portland, Oregon: U.S. Department of the Interior, Fish and Wildlife Service, Region 1. 1992b. Chapter 8: 35HA1028. In C. Raven and R.G. Elston, editors, Land and Life at Malheur Lake: Preliminary Geomorphological and Ar- chaeological Investigations. Cultural Resource Series, 8:87-122. Portland, Oregon: U.S. Department of the Interior, Fish and Wildlife Service, Region 1. Bradbury, J.P. 1991. The Late Cenozoic Diatom Stratigraphy and Paleolimnology of Tulelake, Siskiyou Co., California. Journal of Paleolimnology, 6: 205-255. 1992. Late Cenozoic Lacustrine and Climatic Environments at Tulelake, Northern Great Basin, USA. Climate Dynamics, 6:275-285. 1997. A Diatom Record of Climate and Hydrology for the Past 200 Ka from Owens Lake, California with Comparison to Other Great Ba- sin Records. Quaternary Science Reviews, 16:203-219. Bradbury, J.P., R.M. Forester, and R.S. Thompson 1989. Late Quaternary Paleolimnology of Walker Lake, Nevada. Journal of Paleolimnology, 1:249-267. Broecker, W.S. 1994. Massive Iceberg Discharges as Triggers for Global Climate Change. Science, 372:421-424. Broecker, W.S., G. Bond, M. Klas, G. Bonani, and W. Wolfli 1990. A Salt Oscillator in the Glacial Atlantic? 1: The Concept. Pale- oceanography, 5:469-477. Broecker, W.S., and A. Kaufman 1965. Radiocarbon Chronology of Lake Lahontan and Lake Bonneville, II: Great Basin. Bulletin of the Geological Society of America, 76: 537-566. Broecker, W.S., and P.C. Orr 1958. Radiocarbon Chronology of Lake Lahontan and Lake Bonneville. Bulletin of the Geological Society of America, 69:1009-1032. Brown, S.G. 1957. Occurrence of Ground Water near Ana Springs, Summer Lake Ba- sin, Lake County, Oregon. [Unreviewed Open-File Report prepared in cooperation with the Oregon State Engineer.] 24 pages. United States Geological Survey. [Unpublished.] Bryan, A.T., and R. Gruhn 1964. Problems Relating to the Neothermal Climatic Sequence. American Antiquity, 29:307-315. COHMAP Members 1988. Climatic Changes of the Last 18,000 Years: Observations and Model Simulations. Science, 241:1043-1052. Conrad, C.F. 1953. Geology of the Ana River Section, Summer Lake, Oregon. 92 pages. Master's thesis, Department of Geology, Oregon State Uni- versity, Corvallis. Crandell, D.R. 1965. The Glacial History of Western Washington and Oregon. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 341-354. Princeton, New Jersey: Princeton University Press. 1987. Deposits of Pre-1980 Pyroclastic Flows and Lahars from Mount St. Helens Volcano, Washington. United States Geological Survey Pro- fessional Papers, 1444: 92 pages. Cupp, K.C. 1998. Late Pleistocene Lake Lahontan Paleohydrology and Fernley Sink Subbasin Isostatic Rebound, Churchill, Lyon, and Washoe Counties, Nevada. 128 pages. Master's thesis, Department of Geology, Uni- versity of Nevada, Reno. 46 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Currey, D.R. 1990. Quaternary Palaeolakes in the Evolution of Semidesert Basins, with Special Emphasis on Lake Bonneville and the Great Basin, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology, 76:189 ? 214. 1994a. Hemiarid Lake Basins: Hydrographic Patterns. In A.D. Abrahams and A.J. Parsons, editors, Geomorphology of Desert Environments, pages 405-421. New York: Chapman and Hall. 1994b. Hemiarid Lake Basins: Geomorphic Patterns. In A.D. Abrahams and A.J. Parsons, editors, Geomorphology of Desert Environments, pages 422-444. New York: Chapman and Hall. Damon, P.E., S. Cheng, and T.W. Linick 1989. Fine and Hyperfine Structure in the Spectrum of Secular Variations of Atmospheric 14C. Radiocarbon, 31:704-718. Damon, P.E., and J.L. Jirikowic 1992. Solar Forcing of Global Climate Change? In R.E. Taylor, A. Long, and R.S. Kra, editors, Radiocarbon after Four Decades: An Inter- disciplinary Perspective, pages 117-129. New York: Springer- Verlag. Dansgaard, W., S.J. Johnsen, H.B. Clausen, D. Dahl-Jensen, N.S. Gundestrup, C.U. Hammer, D.S. Hvidberg, J.P. Steffensen, A.E. Sveinbjornsdottir, J. Jouzel, and G. Bond 1993. Evidence for General Instability of Past Climate from a 250-Kyr Ice-Core Record. Nature, 364:218-220. Dansgaard, W., S.J. Johnsen, H.B. Clausen, and C.C. Langway, Jr. 1982. Climatic Record Revealed by the Camp Century Ice Core. In K.K. Turekian, editor, The Late Cenozoic Glacial Ages, pages 37-56. New Haven, Connecticut: Yale University Press. Dansie, A.J., J.O. Davis, and T.W. Stafford, Jr. 1988. The Wizard's Beach Recession: Farmdalian (25,500 Yr B.P.) Verte- brate Fossils Co-Occur with Early Holocene Artifacts. In J.A. Wil- lig, CM. Aikens, and J.L. Fagan, editors, Early Human Occupation in Far Western North America: The Clovis-Archaic Interface. An- thropological Papers, 21:153-200. Carson City: Nevada State Mu- seum. Davis, J.O. 1978. Quaternary Tephrochronology of the Lake Lahontan Area, Nevada and California. Research Papers of the Nevada Archeological Sur- vey, 7: 137 pages. 1982. Bits and Pieces: The Last 35,000 Years in the Lake Lahontan Area. In D.G. Madsen and J.F. O'Connell, editors, Man and Environment in the Great Basin. Anthropological Papers, 2:53-75. Society for American Archeology. 1985. Correlation of Late Quaternary Tephra Layers in a Long Pluvial Se- quence near Summer Lake, Oregon. Quaternary Research, 23: 38-53. De Deckker, P., and R.M. Forester 1988. The Use of Ostracodes to Reconstruct Paleoenvironmental Records. In P. DeDeckker, J.P. Colin, and J.P. Peypouquet, Ostracoda in the Earth Sciences, pages 175-200. Amsterdam, The Netherlands: Elsevier. Denton, G.H., and W. Karlen 1973. Holocene Climatic Variations?Their Pattern and Possible Cause. Quaternary Research, 3:15 5 -205. Dicken, S.N. 1980. Pluvial Lake Modoc, Klamath County, Oregon, and Modoc and Siskiyou Counties, California. Oregon Geology, 42:179-187. Eddy, J.A. 1977. Climate and the Changing Sun. Climate Change, 1:173-190. Edwards, R.L., H. Cheng, M.T. Murrell, and S.J. Goldstein 1997. Protactinium-231 Dating of Carbonates by Thermal Ionization Mass Spectrometry: Implications for Quaternary Climate Change. Sci- ence, 276:782-786. Elston, R.G., K.L. Katzer, and D.R. Currey 1988. Chronological Summary. In C. Raven and R.G. Elston, editors, Pre- liminary Investigations in Stillwater Marsh: Human Prehistory and Geochronology. Cultural Resource Series, 1:367-378. Portland, Or- egon: United States Department of the Interior, Fish and Wildlife Service, Region 1. Elston, R.G., C. Raven, and D.R Dugas 1992. Conclusions: Directions for Future Research. In C. Raven and R.G. Elston, editors, Land and Life at Malheur Lake: Prelimi- nary Geomorphological and Archaeological Investigations. Cul- tural Resource Series, 8:131-137. Portland, Oregon: United States Department of the Interior, Fish and Wildlife Service, Re- gion 1. Erbes, D.B. 1996. Late Pleistocene Lithostratigraphy of Pluvial Lake Chewaucan, Ore- gon: Implications for Past Climate Variation. 110 pages. Master's thesis, Department of Physics and Geology, California State Univer- sity, Bakersfield. Ewing, I. 1996. Paleomagnetism and Paleoecology of Late Pleistocene Lacustrine Sediments from the Black Rock Desert, Humboldt County, Nevada. 73 pages. Master's thesis, Department of Physics and Geology, Cal- ifornia State University, Bakersfield. Fagan, J.L. 1988. Clovis and Western Pluvial Lakes Tradition Lithic Technologies at the Dietz Site in South-Central Oregon. In J.A. Willig, CM. Aikens, and J.L. Fagan, editors, Early Human Occupation in Far Western North America: The Clovis-Archaic Interface. Anthropological Pa- pers, 21:389-416. Carson City: Nevada State Museum. Ferguson, G.J., and W.F. Libby 1962. UCLA Radiocarbon Dates, I. Radiocarbon, 4:109-114. Forester, R.M. 1987. Late Quaternary Paleoclimate Records from Lacustrine Ostracodes. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, volume K-3:261-276. Boulder, Colorado: Geologi- cal Society of America. Freidel, D.E. 1993. Chronology and Climatic Controls of Late Quaternary Lake-Level Fluctuations in Chewaucan, Fort Rock, and Alkali Basins, South- Central Oregon. 244 pages. Doctoral dissertation, Department of Geography, University of Oregon, Eugene. 1994. Paleolake Shorelines and Lake-Level Chronology of the Fort Rock Basin, Oregon. In CM. Aikens and D.L. Jenkins, editors, Archaeo- logical Researches in the Northern Great Basin: Fort Rock Archeol- ogy since Cressman. Anthropological Papers, 50:21-40. Eugene: University of Oregon Department of Anthropology and State Mu- seum of Anthropology. Friis-Christensen, E., and K. Lassen 1991. Length of the Solar Cycle: An Indicator of Solar Activity Closely Associated with Climate. Science, 254:698-700. Frye, J.C, H.B. Willman, and R.F. Black 1965. Outline of Glacial Geology of Illinois and Wisconsin. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 43-62. Princeton, New Jersey: Princeton University Press. Gagan, M.K., L.K. Ayliffe, D. Hopley, J.A. Cali, G.E. Mortimer, J. Chappell, M.T. McCulloch, and M.J. Head 1998. Temperature and Surface-Ocean Water Balance of the Mid-Ho- locene Tropical Western Pacific. Science, 279:1014-1018. Gehr, K.D. 1980. Late Pleistocene and Recent Archaeology and Geomorphology of the South Shore of Harney Lake, Oregon. 170 pages. Master's the- sis, Department of Anthropology, Portland State University, Port- land, Oregon. Gobalet, K.W., and R.M. Negrini 1992. Evidence for Endemism in Fossil Tui Chub, Gila bicolor, from NUMBER 33 47 Pleistocene Lake Chewaucan, Oregon. Copeia, 1992:539-544. Goldthwait, R.P., A. Dreimanis, J.L. Forsyth, P.F. Karrow, and G.W. White 1965. Pleistocene Deposits of the Erie Lobe. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 85-98. Princeton, New Jersey: Princeton University Press. Grayson, D.K. 1979. Mount Mazama, Climatic Change, and Fort Rock Basin Archaeo- faunas. In P.D. Sheets and D.K. Grayson, editors, Volcanic Activ- ity and Human Ecology, pages 427-457. New York: Academic Press. 1993. The Deserts Past: A Natural Prehistory of the Great Basin. 356 pages. Washington, D.C.: Smithsonian Institution Press. Greenspan, R. 1994. Archaeological Fish Remains in the Fort Rock Basin. In CM. Aik- ens and D.L. Jenkins, editors, Archaeological Researches in the Northern Great Basin: Fort Rock Archeology since Cressman. An- thropological Papers, 50:485-504. Eugene: University of Oregon Department of Anthropology and State Museum of Anthropology. Guyodo, Y., and J.-P. Valet 1996. Relative Variations in Paleointensity from Sedimentary Records: The Past 200,000 Years. Earth and Planetary Science Letters, 143: 23-36. Hamilton, T.D., and R.M. Thorson 1983. The Cordilleran Ice Sheet in Alaska. In H.E. Wright, Jr., and S.C. Porter, editors, Late-Quaternary Environments of the United States, Volume I: The Late Pleistocene, pages 38-52. Minneapolis: Univer- sity of Minnesota Press. Hammer, C.U., H.B. Clausen, and W. Dansgaard 1980. Greenland Ice Sheet Evidence of Post-Glacial Volcanism and Its Climatic Impact. Nature, 288:230-235. Hansen, J., A. Lacis, R. Ruedy, and M. Sato 1992. Potential Climate Impact of Mt. Pinatubo Eruption. Geophysical Re- search Letters, 19:215-218. Hattori, E.M. 1982. The Archaeology of Falcon Hill, Winnemuca Lake, Washoe County, Nevada. Anthropological Papers, 19: 208 pages. Carson City: Ne- vada State Museum. Hattori, E.M., and D.R. Tuohy 1993. Prehistoric Human Occupation and Changing Lake Levels at Pyra- mid and Winnemucca Lakes, Nevada. In L.V. Benson, editor, Pro- ceedings of the Workshop "Ongoing Paleoclimatic Studies in the Northern Great Basin," Reno, Nevada, May 1993. United States Geological Survey Circular, 1119:31-34. Hays, J.D., J. Imbrie, and N.J. Shackleton 1976. Variations in the Earth's Orbit: Pacemaker of the Ice Ages. Science, 194:1121-1132. Heinrich, H. 1988. Origin and Consequences of Cyclic Ice Rafting in the Northeast At- lantic Ocean during the Past 130,000 Years. Quaternary Research, 29:142-152. Henyey, S., S.P. Lund, M. Schwarz, and L. Keigwin 1995. Paleomagnetic Secular Variation Record from Deep-Sea Sediments of the Blake/Bahama Outer Ridge (North Atlantic Ocean) during Oxygen-Isotope Stages 5 and 6 (10-190 ka)?Further Evidence for the Relationship between Excursion and "Normal" Secular Varia- tion. EOS {Transactions of the American Geophysical Union), 76: F195. Herrero-Bervera, E., C.E. Helsley, A.M. Sarna-Wojcicki, K.R. Lajoie, C.E. Meyer, B.E. Turin, J.M. Donelly-Nolan, M.O. McWHliams, R.M. Negrini, and J.C. Liddicoat 1994. Age and Correlation of a Paleomagnetic Episode in the Western United States by 40Ar/39Ar Dating and Tephrochronology: The Ja- maica, Blake, or a New Polarity Episode? Journal of Geophysical Research, 99:24,091-24,103. Hostetler, S.W. 1991. Simulation of Lake Ice and Its Effect on the Late-Pleistocene Evap- oration Rate of Lake Lahontan. Climate Dynamics, 6:43-48. Hostetler, S.W., and P.J. Bartlein 1990. Simulation of Lake Evaporation with Application to Modeling Lake Level Variations of Harney-Malheur Lake, Oregon. Water Re- sources Research, 26:2603-2612. Hostetler, S.W., and L.V. Benson 1990. Paleoclimatic Implications of the Highstand of Lake Lahontan De- rived from Models of Evaporation and Lake Level. Climate Dynam- ics, 4:207-217. Houghton, J.T., L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg, and K. Maskell 1996. Climate Change 1995: The Science of Climate Change; Contribu- tion of Working Group I. 572 pages. New York: Cambridge Univer- sity Press. Hubbs, C.L., and R.R. Miller 1948. The Great Basin with Emphasis of Glacial and Postglacial Times: The Zoological Evidence. Bulletin of the University of Utah, 38(20):18-166. Imbrie, J. 1985. A Theoretical Framework for the Pleistocene Ice Ages. Journal of the Geological Society of London, 142:417-432. Imbrie, J., J.D. Hays, D.G Martinson, A. Mclntyre, A.C. Mix, J.J. Morley, N.G. Pisias, W.L. Prell, and N.J. Shackleton 1984. The Orbital Theory of Pleistocene Climate: Support From a Revised Chronology of the Marine 818O Record. In A.L. Berger, editor, Milankovitch and Climate, pages 269-305. Boston: Reidel. Imbrie, J., and K.P. Imbrie 1979. Ice Ages. 224 pages. Cambridge, Massachusetts: Harvard University Press. Imbrie, J., and N.G. Kipp 1971. A New Micropaleontological Method for Quantitative Paleoclima- tology: Application to a Late Pleistocene Caribbean Core. In K.K. Turekian, editor, Late Cenozoic Glacial Ages, pages 71-81. New Haven, Connecticut: Yale University Press. Izett, G.A. 1981. Volcanic Ash Beds: Recorders of Upper Cenozoic Silicic Pyroclas- tic Volcanism in the Western United States. Journal of Geophysical Research, 86:10,200-10,222. Izett, G.A., and J.D. Obradovich 1994. 40Ar/39Ar Age Constraints for the Jaramillo Normal Subchron and the Matuyama-Brunhes Geomagnetic Boundary. Journal of Geo- physical Research, 99:2925-2934. Izett, G.A., K.L. Pierce, and C. Jaworski 1992. Isotopic Dating of Lava Creek B Tephra in Terrace Deposits along the Wind River, Wyoming: Implications for Post 0.6 Ma Uplift of the Yellowstone Hotspot. Geological Society of America Abstracts with Programs, 24: A102. Jenkins, D.L. 1994. Settlement-Subsistence Patterns in the Fort Rock Basin: A Cultural- Ecological Perspective on Human Responses to Fluctuating Wet- lands Resources of the Last 5000 Years. In CM. Aikens and D.L. Jenkins, editors, Archaeological Researches in the Northern Great Basin: Fort Rock Archeology since Cressman. Anthropological Pa- pers, 50:599-628. Eugene: University of Oregon Department of Anthropology and State Museum of Anthropology. Johnson, R.G. 1982. Brunhes-Matuyama Magnetic Reversal Dated at 790,000 Yr B.p. by Marine-Astronomical Correlations. Quaternary Research, 17: 135-147. Kaufman, A., and W. Broecker 1965. Comparison of the 230Th and 14C Ages for Carbonate Materials from Lakes Lahontan and Bonneville. Journal of Geophysical Research, 70:4039-4054. 48 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Kelly, P.M., and T.M.L. Wigley 1992. Solar Cycle Length, Greenhouse Forcing and Global Climate. Na- ture, 360:328-330. Kerr, R.A. 1997. Second Clock Supports Orbital Pacing of the Ice Ages. Science, 276:680-682. King, G.Q. 1978. The Late Quaternary History of Adrian Valley, Lyon County, Ne- vada. 88 pages. Master's thesis, Department of Geography, Univer- sity of Utah, Salt Lake City. 1993. Late Quaternary History of the Lower Walker River and Its Implica- tions for the Lahontan Paleolake System. Physical Geography, 14: 81-96. Klamath Core Study Team 1995. Status Report on Cores Taken in 1991-1992 in the Upper Klamath Basin and Vicinity, Oregon and California. In D.P. Adam, J.P. Brad- bury, W.E. Dean, J.V. Gardner, and A.M. Sarna-Wojcicki, editors, Report of the 1994 Workshop on the Correlation of Marine and Ter- restrial Records of Climate Changes in the Western United States. United States Geological Survey Open-File Report, 95-34:13-29. Kutzbach, J.E. 1983. Modeling of Holocene Climates. In H.E. Wright, Jr., editor, Late- Quaternary Environments of the United States, Volume 2: The Ho- locene, pages 271-277, Minneapolis: University of Minnesota Press. 1987. Model Simulations of the Climatic Patterns during the Deglaciation of North America. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, volume K-3:425-446. Boulder, Colorado: Geological Society of America. Kutzbach, J.E., and P.J. Guetter 1986. The Influence of Changing Orbital Parameters and Surface Bound- ary Conditions on Climate Simulations for the Past 18,000 Years: Journal of Atmospheric Sciences, 43:1726-1759. Kutzbach, J.E., and W.F. Ruddiman 1993. Model Description, External Forcing, and Surface Boundary Condi- tions. In H.E. Wright, Jr., J.E. Kutzbach, T. Webb III, W.F. Ruddi- man, R.A. Street-Perrott, and P.J. Bartlein, editors, Global Climates since the Last Glacial Maximum, pages 12?23. Minneapolis: Uni- versity of Minnesota Press. Kutzbach, J.E., and H.E. Wright, Jr. 1985. Simulation of the Climate of 18,000 Yr BP: Results for the North American/North Atlantic/European Sector. Quaternary Science Re- views ,4:147-187'. Lacis, A.A., and B.E. Carlson 1992. Keeping the Sun in Proportion. Nature, 360:297. Lao, Y, and L. Benson 1988. Uranium-Series Age Estimates and Paleoclimatic Significance of Pleistocene Tufas from the Lahontan Basin, California and Nevada. Quaternary Research, 30:165-176. Lehman, B., C. Laj, C. Kissel, A. Mazaud, M. Paterne, and L. Labeyrie 1996. Relative Changes of the Geomagnetic Field Intensity during the Last 280 Kyr from Piston Cores in the Acores Area. Physics of the Earth and Planetary Interiors, 93:269-284. Lemke, R.W., W.M. Laird, M.J. Tipton, and R.M. Lindvall 1965. Quaternary of Northern Great Plains. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 15-28. Princeton, New Jersey: Princeton University Press. Liddicoat, J.C, and R.S. Coe 1997. Paleomagnetic Investigation of Lake Lahontan Sediments and Its Application for Dating Pluvial Events in the Northwestern Great Basin. Quaternary Research, 47:45?53. Lowenstein, T.K. 2002. Pleistocene Lakes and Paleoclimates (0 to 200 Ka) in Death Valley, California. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:109-120. Lund, S.P. 1996. A Comparison of Holocene Paleomagnetic Secular Variation Records from North America. Journal of Geophysical Research, 101:8007-8024. Lund, S.P., J.C. Liddicoat, K.L. Lajoie, T.L. Henyey, and S. Robinson 1988. Paleomagnetic Evidence for Long-Term (104 Year) Memory and Pe- riodic Behavior of the Earth's Core Dynamo Process. Geophysical Research Letters, 15:1101-1104. Madsen, D.B. 2002. Great Basin Peoples and Late Quaternary Aquatic History. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:387-405. Magny, M. 1993. Solar Influences on Holocene Climatic Changes Illustrated by Cor- relations between Past Lake-Level Fluctuations and the Atmo- spheric 14C Record. Quaternary Research, 40:1-9. Manabe, S., and A.J. Broccoli 1985. The Influence of Continental Ice Sheets on the Climate of an Ice Age. Journal of Geophysical Research, 90:2167-2190. Martinson, D.G., N.G. Pisias, J.D. Hays, J. Imbrie, T.C. Moore, Jr., and N.J. Shackleton 1987. Age Dating and the Orbital Theory of the Ice Ages: Development of a High-Resolution 0 to 300,000-Year Chronostratigraphy. Quater- nary Research, 27:1-29. McCave, I.N., and J.P.M. Syvitski 1992. Principles and Methods of Geological Particle-Size Analysis. In J.P.M. Syvitski, editor, Principles, Methods and Application of Par- ticle Size Analysis, pages 3-21. New York: Cambridge University Press. McDonald, E.V., and A.J. Busacca 1992. Late Quaternary Stratigraphy of Loess in the Channeled Scabland and Palouse of Washington State. Quaternary Research, 38: 141-156. McDowell, P.F. 1992. An Overview of Hamey Basin Geomorphic History, Climate, and Hydrology. In C. Raven and R.G. Elston, editors, Land and Life at Malheur Lake: Preliminary Geomorphological and Archaeological Investigations. Cultural Resource Series, 8:13-34. Portland, Ore- gon: U.S. Department of the Interior, Fish and Wildlife Service, Re- gion 1. McDowell, P.F., and D.P. Dugas 1993. Holocene Lake-Level Variations and Eolian Activity in the Oregon Great Basin. In L. V. Benson, editor, Proceedings of the Workshop "Ongoing Paleoclimatic Studies in the Northern Great Basin," Reno, Nevada, May 1993. United States Geological Survey Circu- lar, 1119:53-54. McWilliams, M. 1995. Global Correlation of the 223 Ka Pringle Falls Event. EOS {Trans- actions of the American Geophysical Union), 76:S99. Mehringer, P.J., Jr. 1977. Great Basin Late Quaternary Environments and Chronology. In D.D. Fowler, editor, Models and Great Basin Prehistory: A Sympo- sium. Desert Research Institute Publication, 12:113-167. Reno, Ne- vada: Desert Research Institute. 1986. Prehistoric Environments. In W.C. Sturtevant, editor, Handbook of North American Indians, volume 11:31-50. Washington, D.C.: Smithsonian Institution Press. Mehringer, P.J., Jr., and W.J. Cannon 1994. Volcaniclastic Dunes of the Fort Rock Valley, Oregon: Stratigraphy, Chronology, and Archaeology. In CM. Aikens and D.L. Jenkins, editors, Archaeological Researches in the Northern Great Basin: Fort Rock Archeology since Cressman. Anthropological Papers, NUMBER 33 49 50:283-327. Eugene: University of Oregon Department of Anthro- pology and State Museum of Anthropology. Meyer, C.E., A.M. Sarna-Wojcicki, J.W. Hillhouse, M.J. Wodward, J.L. Slate, and D.H. Sorg 1991. Fission-Track Age (400,000 yr) of the Rockland Tephra, Based on Inclusion of Zircon Grains Lacking Fossil Fission Tracks. Quater- nary Research, 35:367-382. Mickelson, D.M., L. Clayton, D.S. Fullerton, and H.W. Borns, Jr. 1983. The Late Wisconsin Glacial Record of the Laurentide Ice Sheet in the United States. In H.E. Wright, Jr., and S.C. Porter, editors, Late- Quaternary Environments of the United States, Volume I: The Late Pleistocene, pages 3-37. Minneapolis: University of Minnesota Press. Mifflin, M.D., and M.M. Wheat 1979. Pluvial Lakes and Estimated Pluvial Climates of Nevada. Bulletin of the Nevada Bureau of Mines and Geology, 94: 57 pages. Reno: Mackay School of Mines, University of Nevada. Mix, A.C. 1987. The Oxygen-Isotope Record of Glaciation. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans dur- ing the Last Deglaciation. The Geology of North America, volume K-3:111-136. Boulder, Colorado: Geological Society of America. Morrison, R.B. 1964. Lake Lahontan: Geology of Southern Carson Desert, Nevada. United States Geological Survey Professional Papers, 401: 156 pages. 1965. Quaternary Geology of the Great Basin. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 265-285. Princeton, New Jersey: Princeton University Press. 1991a. Introduction. In R.B. Morrison, editor, Quaternary Nonglacial Geol- ogy: Conterminous U.S. The Geology of North America, volume K- 2:1-12. Boulder, Colorado: Geological Society of America. 1991b. Quaternary Stratigraphic, Hydrologic, and Climatic History of the Great Basin, with Emphasis on Lakes Lahontan, Bonneville, and Tecopa. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. The Geology of North America, volume K-2: 283-320. Boulder, Colorado: Geological Society of America. Morrison, R.B., and J.O. Davis 1984a. Supplementary Guidebook for Field Trip 13, Quaternary Stratigra- phy and Archaeology of the Lake Lahontan Area. 1984 Annual Meeting of the Geological Society of America, Reno, Nevada. Pages 252-281. Reno, Nevada: Desert Research Institute, Univer- sity of Nevada. 1984b. Supplementary Guidebook for Field Trip 13, Quaternary Stratigra- phy and Archaeology of the Lake Lahontan Area: A Re-assesssment. 1984 Annual Meeting of the Geological Society of America, Reno, Nevada. Second edition, 38 pages. Reno, Nevada: Desert Research Institute, University of Nevada. Morrison, R.B., and J.C. Frye 1965. Correlation of the Middle and Late Quaternary Successions of the Lake Lahontan, Lake Bonneville, Rocky Mountain (Wasatch Range), Southern Great Plains, and Eastern Midwest Areas. Report of the Nevada Bureau of Mines, 9: 45 pages. Reno: Mackay School of Mines, University of Nevada. Muller, E.H. 1965. Quaternary of New York. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 99-112. Princeton, New Jersey: Princeton University Press. Mullineaux, D.R. 1986. Summary of Pre-1980 Tephra-Fall Deposits Erupted from Mount St. Helens, Washington State, U.S.A. Bulletin of Volcanology, 48: 17-26. Naeser, C.W., and N.D. Naeser 1987. Fission-Track Dating of Quaternary Events. In D.J. Easterbrook, ed- itor, Dating Quaternary Sediments. Special Papers of the Geological Society of America, 111:\?\ 1. Negrini, R.M. 1993. Intrabasinal and Extrabasinal Correlations of Nonmarine High-Res- olution Paleomagnetic Records of the Middle to Late Pleistocene from South-Central Oregon. In L.V. Benson, editor, Proceedings of the Workshop "Ongoing Paleoclimatic Studies in the Northern Great Basin," Reno, Nevada, May 1993. United States Geological Survey Circular, 1119:57. 1994. Replication of Complex VGP Paths for Two Pleistocene Excursions in Sedimentary Records: A Foundation for Comparing Excursions with Full Reversals. EOS {Transactions of the American Geophysi- cal Union), 75:191. Negrini, R., R. Darke, and B. Hoare 1987. Climate-Induced Variations in the Geomagnetic Field: Possible Sup- porting Evidence from a Pleistocene Paleomagnetic Record. EOS {Transactions of the American Geophysical Union), 68:1249. Negrini, R.M., and J.O. Davis 1992. Dating Late Pleistocene Pluvial Events and Tephras by Correlating Paleomagnetic Secular Variation Records from the Western Great Basin. Quaternary Research, 38:46-59. Negrini, R.M., J.O. Davis, and K.L. Verosub 1984. Mono Lake Geomagnetic Excursion found at Summer Lake, Ore- gon. Geology, 12:643-646. Negrini, R.M., D.B. Erbes, A.P. Roberts, K.L. Verosub, A.M. Sarna-Wojcicki, and C. Meyer 1994. Repeating Waveform Initiated by a 180-190 Ka Geomagnetic Ex- cursion in Western North America: Implications for Field Behavior during Polarity Transitions and Subsequent Secular Variation. Jour- nal of Geophysical Research, 99:24,105 -24,119. Negrini, R.M., D.B. Erbes, P. Wigand, M. Hemphill, A. Cohen, and M. Palacios-Fest 1993. Paleomagnetic Secular Variation Records as Correlation Tools for Quaternary Climate Records from Lacustrine Sediments. EOS {Transactions of the American Geophysical Union), 74:228. Negrini, R.M., and K.L. Verosub 1987. Paleomagnetism and the Pleistocene Geology of Pluvial Lakes in the Northwestern Region of the Great Basin: Work in Progress. In J.O. Davis, editor, Introduction to the Pleistocene Geology of North- western Lake Lahontan, Nevada: Field Guide for the 1987 Friends of the Pleistocene, Pacific Cell, Outing. Desert Research Institute Technical Report, 53: 56 pages. Reno, Nevada: Desert Research In- stitute, University of Nevada. Negrini, R.M., K.L. Verosub, and J.O. Davis 1987. Long-Term Nongeocentric Axial Dipole Directions and a Geomag- netic Excursion from the Middle Pleistocene Sediments of the Hum- boldt River Canyon, Pershing County, Nevada. Journal of Geo- physical Research, 92:10,617-10,627. 1988. The Middle to Late Pleistocene Geomagnetic Field Recorded in Fine-Grained Sediments from Summer Lake, Oregon, and Double Hot Springs, Nevada, U.S.A. Earth and Planetary Science Letters, 87:173-192. North American Commission on Stratigraphic Nomenclature 1983. North American Stratigraphic Code: Bulletin of the American Asso- ciation of Petroleum Geologists, 67:841-875. O'Brien, S.R., P.A. Mayewski, L.D. Meeker, D.A. Meese, M.S. Twickler, and S.I. Whitlow 1995. Complexity of Holocene Climate as Reconstructed from a Green- land Ice Core. Science, 270:1962-1964. Oetting, A.C. 1994. Chronology and Time Markers in the Northwestern Great Basin: The Chewaucan Basin Cultural Chronology. In CM. Aikens and D.L. Jenkins, editors, Archaeological Researches in the Northern Great Basin: Fort Rock Archeology since Cressman. Anthropologi- cal Papers, 50:41-62. Eugene: University of Oregon Department of Anthropology and State Museum of Anthropology. 50 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Oviatt, C.G. 1997. Lake Bonneville Fluctuations and Global Climate Change. Geology, 25:155-158. Oviatt, C.G., D.R. Currey, and D. Sack 1992. Radiocarbon Chronology of Lake Bonneville, Eastern Great Basin, USA. Palaeogeography, Palaeoclimatology, Palaeoecology, 99: 225-241. Palacios-Fest, M.R., A.S. Cohen, J. Ruiz, and B. Blank 1993. Cpmparative Paleoclimatic Interpretations from Nonmarine Ostra- codes Using Faunal Assemblages, Trace Elements, Shell Chemistry, and Stable Isotope Data. In P. Swart, K. Lohman, J. McKenzie, and S. Slavin, editors, Climate Change in Continental Isotopic Records. Geophysical Monographs, 78:179-190. Washington, D.C.: Ameri- can Geophysical Union. Parker, K.E. 1993. Magnetostratigraphy of Plio-Pleistocene Sediments from the Hum- boldt River Canyon, Pershing County, Nevada. 91 pages. Master's thesis, Department of Physics and Geology, California State Univer- sity, Bakersfield. Pewe, T.L., D.M. Hopkins, and J.L. Giddings 1965. The Quaternary Geology and Archeology of Alaska. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 355-376. Princeton, New Jersey: Princeton University Press. Pezzopane, S.K. 1993. Active Faults and Earthquake Ground Motions in Oregon. 208 pages. Doctoral dissertation, Department of Geology, University of Oregon, Eugene. Phillips, K.N., and A.S. Van Denburgh 1971. Hydrology and Geochemistry of Abert, Summer, and Goose Lakes, and Other Closed-Basin Lakes in South-Central Oregon. United States Geological Survey Professional Papers, 502-B: 86 pages. Pinson, A.O. MS. Landscape Change and the Early Holocene Occupation of Dry Ba- sin Floors: The Results of Recent Excavations in Dietz Basin, Lake County, Oregon. 7 pages. [Manuscript presented at the 62nd Annual Meeting, Society for American Archaeology, Nashville, Tennessee, 1996.] Reheis, M.C., A.M. Sarna-Wojcicki, R.L. Reynolds, C.A. Repenning, and M.D. Mifflin 2002. Pliocene to Middle Pleistocene Lakes in the Western Great Basin: Ages and Connections. In R. Hershler, D.B. Madsen, and D.R. Cur- rey, editors, Great Basin Aquatic Systems History. Smithsonian Con- tributions to the Earth Sciences, 33:53-108. Reick, H.J., A.M. Sarna-Wojcicki, C.E. Meyer, and D.P. Adam 1992. Magnetostratigraphy and Tephrochronology of an Upper Pliocene to Holocene Record in Lake Sediments at Tulelake, Northern Califor- nia. Bulletin of the Geological Society of America, 104:409-428. Reid, G.C. 1987. Influence of Solar Variability on Global Sea-Surface Temperatures. Nature, 329:142-143. Reynolds, R.L. 1977. Paleomagnetism of Welded Tuffs of the Yellowstone Group. Jour- nal of Geophysical Research, 82:3677-3693. 1979. Comparison of the TRM of the Yellowstone Group and the DRM of Some Pearlette Ash Beds. Journal of Geophysical Research, 84: 4525-4532. Roberts, A.P., R.L. Reynolds, K.L. Verosub, and D.P. Adam 1996. Environmental Magnetic Implications of Greigite (Fe3S4) Forma- tion in a 3 M.Y. Lake Sediment Record from Butte Valley, Northern California. Geophysical Research Letters, 23:2859-2862. Roberts, A.P., K.L. Verosub, and R.M. Negrini 1994. Middle Late Pleistocene Relative Palaeointensity of the Geomag- netic Field from Lacustrine Sediments, Lake Chewacuan, Western United States. Geophysical Journal International, 23:2859-2862. Rosenbaum, J.G., R.L. Reynolds, D.P. Adam, J. Drexler, A.M. Sarna-Wojcicki, and G.C. Whitney 1996. Record of Middle Pleistocene Climate Change from Buck Lake, Cascade Range, Southern Oregon?Evidence from Sediment Mag- netism, Trace-Element Geochemistry, and Pollen. Bulletin of the Geological Society of America, 108:1328-1341. Rosholt, J.N., S.M. Colman, M. Stuiver, P.E. Damon, C.W. Naeser, N.D. Naeser, B.J. Szabo, and D.R. Muhs 1991. Dating Methods Applicable to the Quaternary. In R.B. Morrison, ed- itor, Quaternary Nonglacial Geology. Conterminous U.S. The Geol- ogy of North America, volume K-2'.45-74. Boulder, Colorado: Geological Society of America. Ruddiman, W.F. 1987. Synthesis: The Ocean/Ice Sheet Record. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans dur- ing the Last Deglaciation. The Geology of North America, volume K-3:463-478. Boulder, Colorado: Geological Society of America. Ruddiman, W.F., and M. Raymo 1988. Northern Hemisphere Climate Regimes during the Past 3 Ma: Possi- ble Tectonic Connections. Philosophical Transactions of the Royal Society of London, B-318:411-430. Ruddiman, W.F., M. Raymo, and A. Mclntyre 1986. Matuyama 41,000-Year Cycles: North Atlantic Ocean and Northern Hemisphere Ice Sheets. Earth and Planetary Science Letters, 80: 117-129. Ruddiman, W.F., and H.E. Wright, Jr. 1987. North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, volume K-3: 501 pages. Boulder, Colorado: Geological Society of America. Russell, I.C. 1895. Present and Extinct Lakes of Nevada. National Geographic Mono- graphs, 1(4): 32 pages. American Book Company. Sampson, C.G. 1985. Nightfire Island: Later Holocene Lakemarsh Adaption on the West- ern Edge of the Great Basin. Anthropological Papers, 33: 553 pages. Eugene: University of Oregon. Sarna-Wojcicki, A.M., and J.O. Davis 1991. Quaternary Tephrochronology. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. The Geology of North America, volume K-2:93-116. Boulder, Colorado: Geological Soci- ety of America. Sarna-Wojcicki, A.M., K.R. Lajoie, C.E. Meyer, and D.P. Adam 1991. Tephrochronologic Correlation of Upper Neogene Sediments along the Pacific Margin, Conterminous United States. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. The Ge- ology of North America, volume K-2:l 17-140. Boulder, Colorado: Geological Society of America. Sarna-Wojcicki, A.M., C.E. Meyer, D.P. Adam, and J.D. Sims 1988. Correlations and Age Estimates of Ash Beds in Late Quaternary Sediments of Clear Lake, California. In J.D. Sims, editor, Late Qua- ternary Climate, Tectonism, and Sedimentation in Clear Lake, Northern California Coast Ranges. Special Papers of the Geological Society of A merica, 214:141-150. Sarna-Wojcicki, A.M., C.E. Meyer, H.R. Bowman, N.T. Hall, P.C. Russell, M.J. Woodward, and J.L. Slate 1985. Correlation of the Rockland Ash Bed, a 400,000-Year Old Strati- graphic Marker in Northern California and Western Nevada, and Implications for Middle Pleistocene Paleogeography of Central Cal- ifornia. Quaternary Research, 23:236-257. Sarna-Wojcicki, A.M., C.E. Meyer, and Elmira Wan 1997. Correlations and Age Estimates of Ash Beds in Late Quaternary Sediments of Owens Lake, California. In G.I. Smith and J.L. Bis- choff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Society of America, 317:79?90. NUMBER 33 51 Sarna-Wojcicki, A.M., S.D. Morrison, C.E. Meyer, and J.W. Hillhouse 1987. Correlation of Upper Cenozoic Tephra Layers between Sediments of the Western United States and Eastern Pacific Ocean, and Com- parison with Biostratigraphic and Magnetostratigraphic Age Data. Bulletin of the Geological Society of America, 98:207-223. Sarna-Wojcicki, A.M., and M.S. Pringle, Jr. 1992. Laser-Fusion 40Ar/39Ar Ages of the Tuff of Taylor Canyon and Bishop Tuff, E. California-W. Nevada. EOS (Transactions of the American Geophysical Union), 73:633. Schafer, J.P., and J.H. Hartshorn 1965. The Quaternary of New England. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 113-128. Princeton, New Jersey: Princeton University Press. Schlesinger, M.E., and N. Ramankutty 1992. Implications for Global Warming of Intercycle Solar Irradiance Variations. Nature, 360:330-333. Scuderi, L.A. 1993. A 2000-Year Tree Ring Record of Annual Temperatures in the Si- erra Nevada Mountains. Science, 259:1433-1436. Shackleton, N.J. 1967. Oxygen Isotope Analyses and Pleistocene Temperatures Re-Ad- dressed. Nature, 215:15-17. Simpson, G.D. 1990. Late Quaternary Tectonic Development of the Northwestern Part of Summer Lake Basin, South-Central Oregon. 121 pages, Master's thesis, Department of Geology, Humboldt State University, Arcata, California. Smith, G.I., and F.A. Street-Perrott 1983. Pluvial Lakes of the Western United States. In H.E. Wright, Jr., and S.C. Porter, editors, Late-Quaternary Environments of the United States, Volume I: The Late Pleistocene, pages 190-214. Minneapo- lis: University of Minnesota Press. Sonett, C.P., and S.A. Finney 1990. The Spectrum of Radiocarbon. Philosophical Transactions of the Royal Society of London, A330:413-426. Spaulding, W.G. 1991. Pluvial Climatic Episodes in North America and North Africa: Types and Correlation with Global Climate. Palaeogeography, Palaeoclimatology, Palaeoecology, 84:217-227. Stine, S. 1990. Late Holocene Fluctuations of Mono Lake, Eastern California. Palaeogeography, Palaeoclimatology, Palaeoecology, 78:333-381. Stuiver, M. 1965. Carbon-14 Content of the 18th- and 19th-century Wood: Variations Correlated with Sunspot Activity. Science, 149:533-535. 1970. Oxygen and Carbon Isotope Ratios of Fresh Water Carbonates as Climatic Indicators. Journal of Geophysical Research, 75: 5247-5257. 1980. Solar Variability and Climatic Change during the Current Millen- nium. Nature, 286:868-871. Stuiver, M., and T.F. Braziunas 1993. Sun, Ocean, Climate and Atmospheric 14CO2: An Evaluation of Causal and Spectral Relationships. The Holocene, 3:289-305. Stuiver, M., T.F. Braziunas, B. Becker, and B. Kromer 1991. Climatic, Solar, Oceanic, and Geomagnetic Influences on Late-Gla- cial and Holocene Atmospheric 14C/I2C Change. Quaternary Re- search, 35:1-24. Stuiver, M., T.F. Braziunas, and P.M. Grootes 1997. Is there Evidence for Solar Forcing of Climate in the GISP2 Oxygen Isotope Record? Quaternary Research, 48:259-266. Stuiver, M., P.M. Grootes, and T.F. Braziunas 1995. The GISP2 818O Climate Record of the Past 16,500 Years and the Role of the Sun, Ocean, and Volcanoes. Quaternary Research, 44: 341-354. Stuiver, M., and H.A. Pollach 1977. Discussion-Reporting of 14C Data. Radiocarbon, 19:355-363. Stuiver, M., and P.D. Quay 1980. Changes in Atmospheric Carbon-14 Attributed to a Variable Sun. Science, 207:11-19. Szabo, B.J., C.A. Bush, and L.V. Benson 1996. Uranium-Series Dating of Carbonate (Tufa) Deposits Associated with Quaternary Fluctuations of Pyramid Lake, Nevada. Quaternary Research, 45:271-281. Tauxe, L. 1993. Sedimentary Records of Relative Paleointensity of the Geomagnetic Field: Theory and Practice. Reviews of Geophysics and Space Phys- ics, 31:319-354. Tauxe, L., and N.J. Shackleton 1994. Relative Paleointensity Records from the Ontong-Java Plateau. Geophysics Journal International, 117:769 -782. Thompson, R., and F. Oldfield 1986. Environmental Magnetism. 227 pages. Boston: Allen and Unwin. Thompson, R.S., L.V. Benson, and E.M. Hattori 1986. A Revised Chronology for the Last Pleistocene Lake Cycle in the Central Lahontan Basin. Quaternary Research, 25:1-9. Thompson, R.S., L.J. Toolin, R.M. Forester, and R.J. Spencer 1990. Accelerator-Mass Spectrometer (AMS) Radiocarbon Dating of Pleistocene Lake Sediments in the Great Basin. Palaeogeography, Palaeoclimatology, Palaeoecology, 78:301?313. Thompson, R.S., C. Whitlock, P.J. Bartlein, S.P. Harrison, and W.G. Spaulding 1993. Climate Changes in the Western United States since 18,000 yr B.P. In H.E. Wright, Jr., J.E. Kutzbach, T. Webb III, W.F. Ruddiman, R.A. Street-Perrott, and P.J. Bartlein, editors, Global Climates since the Last Glacial Maximum, pages 12-23. Minneapolis: University of Minnesota Press. Trie, E., C. Laj, J.-P. Valet, P. Tucholka, M. Paterne, and F. Guichard 1991. The Blake Geomagnetic Event: Transition Geometry, Dynamical Characteristics and Geomagnetic Significance. Earth and Planetary Science Letters, 102:1 -13. Tuohy, D. 1988. Artifacts from the Northwestern Pyramid Lake Shoreline. In J.A. Willig, CM. Aikens, and J.L. Fagan, editors, Early Human Occupa- tion in Far Western North America: The Clovis-Archaic Interface. Anthropological Papers, 21:201-216. Carson City: Nevada State Museum. Tuohy, D.R., and A. Dansie 1997. New Information Regarding Early Holocene Manifestations in the Western Great Basin. Nevada Historical Society Quarterly, 40: 24-53. Van Denburgh, A.S. 1975. Solute Balance at Abert and Summer Lakes, South-Central Oregon. United States Geological Survey Professional Papers, 502-C: 29 pages. Verosub, K.L. 1988. Geomagnetic Secular Variation and the Dating of Quaternary Sedi- ments. In D.J. Easterbrook, editor, Dating Quaternary Sediments. Special Papers of the Geological Society of America, 227:123-138. Verosub, K.L., and A.P. Roberts 1995. Environmental Magnetism: Past, Present, and Future. Journal of Geophysical Research, 100:2175-2192. Waitt, R.B., Jr., and R.M. Thorson 1983. The Cordilleran Ice Sheet in Washington, Idaho, and Montana. In H.E. Wright, Jr., and S.C. Porter, editors, Late-Quaternary Environ- ments of the United States, Volume I: The Late Pleistocene, pages 53-70. Minneapolis: University of Minnesota Press. Wayne, W.J., and J.H. Zumberge 1965. Pleistocene Geology of Indiana and Michigan. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 63-84. Princeton, New Jersey. Princeton University Press. 52 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Weeks, R.J., C. Laj, L. Endignous, A. Mazaud, L. Labeyrie, A.P. Roberts, C. Kissel, and E. Blanchard 1995. Normalised Natural Remanent Magnetisation Intensity during the Last 240,000 Years in Piston Cores from the Central North Atlantic Ocean: Geomagnetic Intensity or Environmental Signal? Physics of the Earth and Planetary Interiors, 87:213-229. Wigand, P.E. 1987. Diamond Pond, Harney County, Oregon: Vegetation History and Water Table in the Eastern Oregon Desert. Great Basin Naturalist, 47:427-458. Wigand, P.E., M.L. Hemphill, S. Sharpe, and S. Patra 1995. Eagle Lake Basin, Northern California, Paleoecological Study: Semi-arid Woodland and Montane Forest Dynamics during the Late Quaternary in the Northern Great Basin and Adjacent Sierra. 51 pages. Project Report #CSA-5-93-63-024 to the Lassen National Forest, Quaternary Science Center of the Desert Research Institute, Reno, Nevada. Wigand, P.E., and D. Rhode 2002. Great Basin Vegetation History and Aquatic Systems: The Last 150,000 Years. In R. Hershler, D.B. Madsen, and D.R. Currey, edi- tors, Great Basin Aquatic Systems History. Smithsonian Contribu- tions to the Earth Sciences, 33:309?367. Willig, J.A. 1988. Paleo-Archaic Adaptations and Lakeside Settlement Patterns in the Northern Alkali Basin, Oregon. In J.A. Willig, CM. Aikens, and J.L. Fagan, editors, Early Human Occupation in Far Western North America: The Clovis-Archaic Interface. Anthropological Papers, 21:417-482. Carson City: Nevada State Museum. Willig, J.A., and CM. Aikens 1988. The Clovis-Archaic Interface in Far Western North America. In J.A. Willig, CM. Aikens, and J.L. Fagan, editors, Early Human Occupa- tion in Far Western North America: The Clovis-Archaic Interface. Anthropological Papers, 21:1-40. Carson City: Nevada State Mu- seum. Winograd, I.J., T.B. Coplen, J.M. Landwehr, A.C Riggs, K.R. Ludwig, B.J. Szabo, P.T. Kolesar, and K.M. Revesz 1992. Continuous 500,000-Year Climate Record from Vein Calcite in Devil's Hole, Nevada. Science, 258:255-260. Winograd, I.J., J.M. Landwehr, K.R. Ludwig, T.B. Coplen, and A.C. Riggs 1997. Duration and Structure of the Past Four Interglaciations. Quaternary Research, 48:141-154. Wright, H.E., Jr., and R.V. Ruhe 1965. Glaciation of Minnesota and Iowa. In H.E. Wright, Jr., and J.C. Frye, editors, The Quaternary of the United States, pages 29?42. Princeton, New Jersey: Princeton University Press. Yamazaki, T., and N. Ioka 1994. Long-Term Secular Variation of the Geomagnetic Field during the Last 200 Kyr Recorded in Sediments Cores from the Western Equa- torial Pacific. Earth and Planetary Science Letters, 128:527-544. Zielinski, G.A., P.A. Mayewski, L.D. Meeker, S. Whitlow, M.S. Twickler, M. Morrison, D.A. Meese, A.J. Gow, and R.B. Alley 1994. Record of Volcanism since 7000 B.C. from the GISP2 Greenland Ice Core and Implications for the Volcano-Climate System. Science, 264:948-952. Pliocene to Middle Pleistocene Lakes in the Western Great Basin: Ages and Connections Marith C. Reheis, Andrei M. Sarna-Wojcicki, RichardL. Reynolds, Charles A. Repenning, and Martin D. Mifflin ABSTRACT During the Pliocene to middle Pleistocene, pluvial lakes in the western Great Basin repeatedly rose to levels much higher than those well-documented levels of the late Pleistocene. Sedimento- logic, geomorphic, and chronologic evidence indicates that Lakes Lahontan and Columbus-Rennie were as much as 70 m higher at about 650 thousand years ago (Ka) than during their late Pleis- tocene highstands. At its former altitude of 1400 m, Lake Lahon- tan submerged present-day Reno, Carson City, and Battle Mountain in Nevada and flooded other now-dry basins. To the east, Lakes Jonathan (new name), Diamond, Newark, and Hubbs reached highstands during the early-middle(?) Pleistocene that were 25?40 m above their late Pleistocene shorelines; at these very high levels, the lakes became temporary or permanent tributaries to the Humboldt River and, hence, to Lake Lahontan. This sce- nario presents a startling vision of a western Great Basin in which climatic and tectonic effects conspired to integrate presently iso- lated basins into one enormous drainage system during at least one brief period during the Quaternary. Such a temporary connection could have permitted fish to migrate from the Humboldt River southward into the presently isolated Newark Valley and from Lake Lahontan into Fairview Valley. The timing of drainage inte- gration provides suggested maximum ages for fish to populate the basins of Lake Diamond and Lake Jonathan. Marith C. Reheis and Richard L. Reynolds, United States Geological Survey, MS-980, Federal Center, P.O. Box 25046, Denver, Colorado 80225, USA. Andrei M. Sarna-Wojcicki, United States Geological Sur- vey, MS-975, 345 Middlefield Road, Menlo Park, California 94025, USA. Charles A. Repenning, 9010 West 5th Place, Lakewood, Colo- rado 80226, USA. Martin D. Mifflin, Mifflin and Associates, HCR38, Box 126, Las Vegas, Nevada 89124, USA. Reviewers: Kenneth L. Adams, Desert Research Institute, 7010 Dan- dini Boulevard, Reno, Nevada 89512, USA; M.A. Chan, Department of Geology and Geophysics, University of Utah, Salt Lake City, Utah 84112, USA; Robert Hershler, Division ofMollusks, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560- 0118, USA; D.E. Wilkins, University of Utah, Salt Lake City, Utah 84112, USA. The Walker Lake subbasin of Lake Lahontan held deep lakes at least four times during the early Pleistocene; these were at least 15-25 m above the Sehoo highstand. One site on the Humboldt River in the northeastern Lahontan Basin displays exposed, faulted lake sediments of early Pleistocene age that are now 40-55 m above the Sehoo limit. These sediments suggest that the northern and Walker Lake subbasins were contiguous parts of Lake Lahon- tan by about one million years ago (Ma), but fish fossils indicate that drainage connections existed before 1 Ma. For example, high beach deposits suggest that Lake Russell, now a tributary to the Owens River, previously overflowed to the Walker River, poten- tially providing a migration route for fish southward along the east side of the Sierra Nevada into the Owens River system. Deposits of Pliocene lakes suggest connections existed among proto-Lakes Lahontan, Columbus-Rennie, Clayton, and possibly Tonopah. Tephrochronologic ages of these Pliocene lakes provide con- straints on the availability of aquatic migration routes at different times. Reconstructing and dating these lake levels also has important implications for paleoclimate, tectonics, and drainage evolution in the western Great Basin. For example, shorelines in several basins form a stair-step sequence downward with time from the highest levels, thought to have formed at about 650 Ka, to the lowest, formed during the late Pleistocene. This descending sequence indi- cates progressive drying of pluvial periods, possibly caused by uplift of the Sierra Nevada and other western ranges relative to the western Great Basin. These effects, however, cannot account for the extremely high lake levels during the early-middle Pleistocene; rather, these high levels were probably caused by a combination of increased effective moisture and changes in the size of the Lake Lahontan drainage basin. Introduction Pluvial lakes in the western Great Basin have been the focus of many classic studies on paleoclimate and on biogeographic distributions of aquatic species, such as fish and snails. Until recently, these studies have focused on sedimentary records of the late Pleistocene and on the distributions of modern species (especially fish), because few pre-late Pleistocene deposits of 53 54 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES pluvial lakes were known and because fossil localities were scarce and commonly not well dated. In their seminal work, Hubbs and Miller (1948) argued for relatively rapid evolution of fish in now-isolated drainage basins on the assumption that, in most cases, the last aquatic connections between the fish- bearing basins were during the last pluvial period, now known to have ended about 12 thousand years ago (Ka). They recog- nized that in some cases modern sills between basins were far too high to have been exceeded by late Pleistocene lake levels; they called upon transfers by interdivide stream diversions or "pre-pluvial" connections to explain these fish distributions. As more fossil localities were identified and dated, research in the last 20 years has shown that many modern species of fish and springsnails are very similar to ancestral species that were widely distributed in the western Great Basin and areas to the north and south during the Pliocene (Miller and Smith, 1981; Taylor and Smith, 1981; Taylor, 1985; Minckley et al., 1986; Gobalet and Negrini, 1992; papers in this volume). As a result, less credence has been placed in rapid evolution, and modern aquatic-species distributions are more commonly attributed to inheritance from widespread Tertiary populations followed by extinctions and intermittent population-mixing caused by wet-dry climatic cycles and tectonic disruptions during the Pleistocene. The possibility that modern species-distributions could be produced in part by aquatic migration routes older than late Pleistocene in age but younger than Tertiary in age has been given little consideration in biogeographic studies for the very good reason that such intermediate-age migration routes were largely unknown. This paper documents connections among several large lakes in the western Great Basin during part of the early or middle Pleistocene but that probably have been severed for the past one-half million years or more. Arguments for the evolution of modern species from widely distributed ancestral ones that suggested rapid evolutionary changes took place dur- ing the last few thousand years (e.g., Hubbs and Miller, 1948) have been challenged recently (Smith et al., 2002) as being not representative of the much slower rates observed in the fossil record during the past several million years. We suggest that if the observed divergence between some species can be averaged over a half a million years rather than over a few thousand, the apparent discrepancy in evolutionary rates may be greatly re- duced. This new information at least provides alternative hy- potheses about rates of evolution from independent evidence, as urged by Smith et al. (2002). In some cases, these older Pleistocene lakes may require revision of previous interpreta- tions about ages and relations among fossil fish (e.g., Miller and Smith, 1981) that were based on incorrect age assignments. In addition to far-reaching implications for biogeographic studies, the new lacustrine records hold significance for studies of paleoclimate, tectonics, and drainage evolution in the west- ern oreat Basin. The existance of enormously expanded lakes at about 650 Ka and earlier imply that much wetter conditions prevailed at that time than are indicated in global climate records, such as oxygen-isotope data from marine cores; how- ever, sudden and dramatic integrations of previously discon- nected basins with the Lahontan Basin may help to account for the large size of Lake Lahontan in the early-middle Pleis- tocene. Progressively shrinking lake highstands from the mid- dle to late Pleistocene indicate different regional climatic con- ditions. Some Pliocene and younger basins were probably isolated by tectonic processes, such as compressional folding and extensional faulting along the Walker Lane tectonic belt (Stewart, 1988). This paper focuses on pluvial lakes in the Lahontan Basin and the basins of Columbus Salt Marsh and Fish Lake Valley, where most of the field work in the western Great Basin has been concentrated. Detailed descriptions of the stratigraphy, chronology, geomorphology, and geology of pre-late Pleis- tocene lacustrine deposits in the Lahontan Basin are presented first, followed by a correlation of the deposits from sampled lo- calities. Next, data from reconnaissance field work for this study and from previously published studies are discussed briefly for Lake Russell (Mono Basin) in western Nevada and for lakes in central Nevada, including those in the basins of Kobeh (Lake Jonathan)-Diamond, Long (Lake Hubbs), New- ark, and Pine Valleys. These lakes are described in chronologic order, beginning with Pliocene localities. A discussion of the ages of the pre-late Pleistocene lakes in the study area is fol- lowed by an examination of the possible causes of these very large lakes. The paper concludes by discussing the implications of very high lake stands and prior connections among drainage basins on the migration of aquatic species. PREVIOUS WORK During its well-studied late-middle (Eetza) and late Pleis- tocene (Sehoo) highstands, Lake Lahontan reached a maximum altitude of about 1330 m and extended more than 350 km, from the Nevada-Oregon border to beyond the southern end of Walker Lake (Figure 1; Russell, 1885; Benson et al., 1990; Morrison, 1991). Many smaller, isolated lakes also occupied other basins in the western Great Basin (Mifflin and Wheat, 1979), but their late Pleistocene highstands are generally not well dated, and little information has been available on older pluvial lakes and their high shorelines. Previous workers have observed that the Eetza (~180-130 Ka, but possibly in part as old as 300 Ka; Morrison, 1991) and Sehoo (~35?12 Ka; Benson et al., 1990) shorelines in un- FIGURE 1 (opposite).?Regional map showing pluvial lakes of northern Nevada and vicinity, late Pleistocene shorelines, and areas of maximum middle Pleistocene (where known) and late Pleistocene lakes. See Table 1 for shoreline altitudes. Triangles show study sites (data for selected sites is in Table 1); arrows on drainages show possible links between basins during highest lake stands. Localities: CD = Carson Desert, CLV = Clayton Valley, CV = Campbell Valley, FLV = Fish Lake Valley, HL = Honey Lake,' LCR = Lake Columbus-Rennie, LM = Lone Mountain, LT = Lone Tree Hill,' MW = McGee Wash, PV = Pine Valley, RM = Rhodes Salt Marsh, RP = Rye Patch Dam, SA = Sunshine Amphitheater, SC = Smoke Creek Desert, TB = Thorne Bar, WL = Walker Lake. Towns: B = Battle Mountain,' C = Carson City, E = Elko, R = Reno, W = Wei Is, WI = Winnemucca. NUMBER 33 121? 55 117? 42 Area of late Pleistocene lakes (not all are shown) Maximum known area of middle Pleistocene lakes f Possible additional area of middle Pleistocene lakes Early and middle Pleistocene lake deposit locality & Pliocene lake deposit locality Late Pleistocene overflow Inferred earlier overflow and modern sill height (m) above late Pleistocene shoreline Late Pleistocene boundary of Lahontan basin Possible additional area of basin in middle Pleistocene Major drainages 56 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES faulted positions throughout the Lahontan Basin lie at about 1330 ? 12 m (Mifflin and Wheat, 1979; Benson, 1978). Iso- static rebound has caused shoreline altitudes to range from about 1318 m at the northeastern margin of the lake to as much as 1341 m in the Carson Desert (Adams, 1996, 1997). The highest vertical faulting rates of late Pleistocene shorelines are along the Wassuk fault, which bounds the west side of the Walker Lake subbasin (Figures 1, 2). Demsey (1987) estimated a Holocene slip rate of 0.4-0.5 mm/yr and total offset of 6-7 m of the Sehoo shoreline along this fault. Morrison (1991) summarized the previously known ages and altitudes of early-middle and early Pleistocene nearshore and deltaic deposits of Lake Lahontan. Four sites were in outcrops below the Sehoo shoreline where incision exposed older depos- its buried by Sehoo- and Eetza-aged sediment. The fifth site, Thorne Bar, is described below. The pre-Eetza lake cycles identified by Morrison (1991) are the Rye Patch Alloformation, from about 700 to 600 Ka, and two sets of unnamed deposits containing and underlying the Glass Mountain tephra, which were erupted about one million years ago (Ma) (only found in the Walker Lake subbasin). He believed that the 0.76 My old Bishop ash bed was deposited during a long period of alluvial 119?15' 119?00' 118?45' 118?30' - Adrian -4T Valley EXPLANATION Area of late Pleistocene lake with 1330-m shoreline j 1 Area of middle Pleistocene lake with 1400-m shoreline Probable additional area of middle Pleistocene lake accounting for sedimentation ? Locality of pre-late Pleistocene MW lacustrine sediment or shoreline FIGURE 2.?Map of Walker Lake basin and southernmost Carson Desert (Raw- hide Flats) showing study sites and lake areas. Area of lake at 1400 m is drawn using modern topography, acknowledging that this is only an approximation of the topography in the middle Pleistocene because of subsequent isostatic rebound, faulting, erosion, and sedimentation. From north to south, letters indi- cate the following: SA = Sunshine Amphitheater, CV = Campbell Valley, WR = Weber Reservoir, MW = McGee Wash, TW = Twentymile Beach, TB = Thorne Bar, SSV = Soda Springs Valley. and eolian sedimentation (the Lovelock Alloformation) during which no deep lakes existed in the Lahontan Basin; however, the Bishop ash bed is contained within lacustrine deposits in basins to the south, including pluvial Lakes Rennie (Fish Lake Valley; Reheis et al., 1993), Owens (Smith et al., 1997), Tecopa (Sarna-Wojcicki et al., 1987), and Manley (Death Valley; Knott et al., 1996). Also, an ash bed with Bishop-like chemical com- position was identified in beach gravel exposed in a fault scarp below the Sehoo shoreline south of Schurz, Nevada (Davis, 1978). Only in the Walker Lake subbasin of Lake Lahontan are there previously known outcrops of beach gravel at altitudes higher than the Sehoo and Eetza shorelines. W.J. McGee ob- served that the eastern end of the "crest of the gravel embank- ment separating Walker Lake Valley from the... Walker River ... is now fully 200 feet above its original position" (that is, the Sehoo shoreline) (Russell, 1885:142, pi. 28). This locality, which apparently had not been revisited until the time of the present study, is southeast of Weber Reservoir (Figure 2) and is informally named McGee Wash (Morrison and Davis, 1984). The Thorne Bar site, at the southeast edge of Walker Lake (Figures 2, 3) was first reported by King (1978) and Mif- flin and Wheat (1979) to have pre-Eetza "shore gravel with layers of lake-deposited carbonate, preserved in a much- eroded V-bar complex...up to 1443 m altitude" (Morrison, 1991:296). At the McGee Wash site, Russell (1885) thought the gravels had been folded into an anticline; mapping for this study (see Figure 7) and that of Morrison and Davis (1984) confirms the presence of numerous faults and tilted strata. The beach ridge at Thome Bar has been attributed either to faulting (King, 1978:67-68) or to northward tilting of the entire Lahontan Basin (Mifflin and Wheat, 1979; Mifflin, 1984; Morrison, 1991). Detailed leveling surveys by Demsey (1987), however, revealed no evidence of Quaternary faults or defor- mation either at Thorne Bar or northward along the Gillis Range, in contrast to several localities of superelevated beach gravel along the footwall block of the Wassuk Range fault (King, 1978; R. Bucknam, U.S. Geological Survey, in litt. to M.C.R., 1996). Little work has been done on other pluvial lakes in western Nevada (Figure 1). Mifflin and Wheat (1979) mapped late Pleistocene shorelines for all of the pluvial lake basins of Ne- vada. These authors also briefly noted higher, older shoreline features of Lakes Diamond, Newark, and Hubbs; Mifflin (1984) thought that these features supported his hypothesis of the northward tilting of northwestern Nevada. Tackman (1993) obtained ages on the late Pleistocene shorelines of Lake Dia- mond. He also located many pre-late Pleistocene shorelines and discussed Lake Diamond's history as controlled by over- flow from Kobeh Valley (Lake Jonathan) and the incision of sills. Davis (1981) investigated Tertiary lacustrine deposits in Clayton Valley. Gilbert et al. (1968) mapped Pliocene lacus- trine deposits interbedded with basalt and rhyolite flows on the NUMBER 33 57 118? 40' 118? 39' 118 118? 37' Qfh Qfy Qly -Qfm- 23 m omitted North of Thome Bar Section 1 Pebbles lagged on bedrock slope up to altitude 1398 m (4585 feet) i? 5 1?0 Variable amounts of tufa cement Fishy bones EL21-WL(Bishop ash?) Bwk--20 cm !] Bw?--15cm Btk-30 cm Matrix-poor Btk-50 cm Btk-90 cm Btk--30 cm EXPLANATION Sandstone and silty sandstone, locally conglomerate beds Sandy pebble con- glomerate, well rounded clasts (near- shore deposits) Sandy cobble conglomerate, well rounded clasts (beach deposits) Boulder conglomerate, poorly bedded (shore-face deposit) o o"< 0 0 0 O 0 < 0 0 0 A -4 ''A XXX EL21-WL Sandstone and con- glomerate, poorly to moderately sorted Volcanic rocks Steep dips (backset beds) Calcareous cement, locally tufa Tephra layer and sample number Buried soil and diagnostic horizon thickness 2 KILOMETERS 1 MILE CONTOUR INTERVAL: 40 FEET; 200 FEET IN MOUNTAINS MAP EXPLANATION Fan deposits (Holocene); includes some Holocene lake deposits Fan deposits (late Pleistocene) Deposits of late Pleistocene (Sehoo) lake (unit QI7T) Fan deposits (middle Pleistocene) Fan deposits (middle and early? Pleistocene) Qlo \/ A A ' /TVS. 7 Nearshore deposits and beach gravel of middle to early Pleistocene lakes (units QI1T to QI6T) Volcanic and sedimentary rocks (Pliocene and Miocene) Late Pleistocene shoreline Middle Pleistocene shorelines; dashed where inferred Tephra locality and measured section Soil pit (Adams and Wesnousky, 1999) FIGURE 3.?Geologic map of area around Thorne Bar (Figure 2, Table 1) and measured section north of Thome Bar (triangle in northern map area). 58 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 1.?Information for representative sites (see Figure 1; within each basin, sites are in order beginning in the south and proceeding clockwise or northward). Site General location Latitude (N) Longitude (W) Altitude range of nearshore features (m) Type of deposit or shoreline feature Lake Lahontan (late Pleistocene shoreline, 1330 ? 12 m) US-95 roadcut Thorne Bar area 20-Mile Beach McGee Wash Campbell Valley Sunshine Amphi- theater Clifton site 1-80 roadcut (?) Secret Valley cutoff Burro Mountain Freds Field Iron Point Treaty Hill area Lone Tree Hill Drumm Summit Rhodes Salt Marsh Southeastern Walker Lake valley West-central Walker Lake valley Northern Walker Lake valley Northern Walker Lake valley Northern Walker Lake valley Carson River canyon Truckee River canyon North of Honey Lake valley Smoke Creek Desert Smoke Creek Desert Humboldt River Humboldt River Humboldt River Fairview Valley 38?14.7' 38?39'-44' 38?46.6' 38?45.1' 39?01'-03.5' 39?04-O7' 39?08.6'-09.4' 39? 18.1-18.4' 39?30.8' 40?25.0' 40?32.5' 40?44.7'-45' 40?56.6'-59.5' 40?50.3'-58.0' 40?49.4'-50.5' 39? 18-18.5' 118?03.3' 118?37.3'-39' 118?45.4' 118?46.5' 118?47.5'-50.7' 118o59'-119?01.2' 118?59.5'-119?00.7' 119?27.4'-28.0' 119?38.7' 120?16.9' 119?50.6'-51.4' 119?31.4-32.2' 117?17.2-19.5' 117?04'-12' 117?12.0'-12.7' 118?07.9' 1402-1423 1360-1400, 1350-1365 1330-1350 Depends on footwall location; <1475 1420-1433, 1393-1402 1347-1365 1378-1390,1356-1378 1341-1426 1333-1380 1360 1370-1372 1396-1402, 1373 1378-1400, 1350-1370 1353-1368 1344-1362, 1390-1414 1326-1414 1355-1384 Sinter mound Borrow pit Rock Hill Coaldale Junction Fish Lake Valley Southern Columbus Salt Marsh Northern Columbus Salt Marsh Lake Columbus-Rennie (late Pleistocene shoreline, 1402 m) 1442-1460 38?00'-00.9' 38?08.4'-09' 118?00' 118?01.2'-02.2' 118?56.9-57.8' Southeastern Columbus 38?00.7'-01.5' 117?52.6'-53.8' Salt Marsh 1442-1452, 1426-1433, 1417 1390-1417 1422-1430 Lake Russell (Mono Basin) (late Pleistocene shoreline, 2155 m) Southern outlet area Rancheria Gulch Kirkwood Spring Andesite sill US-50 site Slough Creek Borrow pit on county road Eastern Mono Basin Northern Mono Basin Northern Mono Basin Northeastern Mono Basin Eastern Kobeh Valley Southwestern side Lone Mountain Northeastern Kobeh Valley 38?00'-04' 38?05.5'-06' 38?08.7'-09' 38?18.5' H8?44'_47.5' 119?05'-06' 118?58'-59' 118?46.7' 2207-2277 <2268,2195-2225 2205-2252 2237 Lake Jonathan (Kobeh Valley) (no late Pleistocene lak 39?34.1'-35.3' 39?33.5'-34' 39?42.0' 116? 16.6-17.4' 1871-1920 1856-1890 1914-1925 Nearshore and sublacustrine deltaic deposits Tufa-cemented beach gravel and sand; 3 shoreline levels, 6 ages of deposits Tufa-cemented beach gravel and sand; 2 shoreline levels, 2-4 ages of deposits Lake mud, silt, sand, and gravel; 3 shoreline levels, 5 ages of deposits Lake mud, silt, sand, and gravel; shoreline angle; 2 shoreline levels, 3 ages of deposits Lake mud, silt, sand, and gravel; 3 shoreline levels, 5 ages of deposits Beach or fluvial gravel lag on bedrock Deltaic(?) and fluvial deposits Large botryoidal tufa heads, locally attached to well-rounded clasts; scattered patches of well-rounded pebbles and cobbles Beach gravel lag and tufa; 2 different levels Rounded, water-polished gravel; 2 different levels Beach or fluvial gravel lag Sand and gravel in extensive delta plain to north; upper level is berm on Treaty Hill Lake mud, silt, sand, and gravel; 1 shoreline level, 3 or 4 ages of deposits Lake mud and silt (Pliocene?) overlain by beach(?) gravel of uncertain age Sinter-cemented silt and sand; Bishop ash (Reheis et al., 1993) Lake silt, sand, and gravel; shoreline angle; 3 shoreline levels, 4 ages of deposits Beach sand and gravel in moderately pre- served berm; pit exposes buried lacustrine deposits containing Bishop ash Beach sand and gravel in eroded berms Lake silt, sand, gravel, and beach rock; at least 2 ages of deposits, oldest is late Pliocene; upper limit defined by beach gravel lagged on bedrock Beach gravel; 2 shoreline levels, lower marked by berm Scattered beach cobbles lagged on bedrock Bedrock sill bearing scattered rounded clasts Lake silt, sand, and gravel; 2 ages of deposits Lake clay, silt, sand, and gravel; 3 ages of de- posits Lake sand and gravel; 3 ages of deposits sep- arated by paleosols; well-preserved berms in vicinity NUMBER 33 59 TABLE 1.?Continued. Site General location Latitude (N) Longitude (W) Altitude range of nearshore features (m) Type of deposit or shoreline feature Lake Newark (late Pleistocene shoreline, 1847 m) 39?27'-27.4' 115?39.6'-40.6' 1856-1859, 1850-1855 Beach sand and gravel; 2 shoreline levels, West of Beck Pass Stage Road Maple Syrup well Long Valley Canyon Elko-White Pine County line Independence interchange Wells Angel Island Emigrant Pass section The Gap Redlich Summit Southern Newark Valley Eastern Newark Valley 39?36' Northern Newark Valley 115?39.0' 39?52.7'-53.5' 115?39.2'-40.4' more eroded upward 1850-1855 Beach sand and gravel in moderately pre- served berm 1887-1890, 1868-1875, Beach gravel in berms and as lag on bedrock; 1856-1859, 1850-1855 3 or 4 shoreline levels (possibly one re- peated by fault), more eroded upward Lake Hubbs (Long Valley) (late Pleistocene shoreline, 1920 m) Southern Long Valley 39?32.2'-34' 115?25'-28' 1935-1940, 1920-1927, Beach sand and gravel; 3 different shoreline Eastern Long Valley 39?35.3' 115?20.8' 1907-1912 1930-1940?, 1920-1927 levels, more eroded upward Beach gravel in moderately preserved berm; lag gravel above Lake Gale (late Pleistocene shoreline, 1915 m?revised from 1905 m as stated by Mifflin and Wheat, 1979) Northeastern Butte Valley Independence Valley 40?03.5'-07.5' 115?01'-04' 1917-1922 Lake Clover (late Pleistocene shoreline, 1730 m) 41?05.5'-06.3' 114?42.2'-43' 1754-1763 Humboldt River head- 41?06.4'-07.4' 114?58.5'-59.1' waters 1737-1750 Eastern Clayton Valley 37?40M4' Pliocene localities 117?32'-35' 1340-1440 Northeastern Fish Lake Valley Columbus-Fish Lake sill Rhodes-Columbus divide area 117?55' 37?53.4' 37?54.5'-57' 117?57'-59' 1475-1560 1432-1463 38?12.5'-14.5' 117o59'-118?01.5' 1450-1585 Beach sand and gravel in subdued berms above well-defined late Pleistocene shore- line features Beach (?) gravel exposed in 1-80 borrow pits North-dipping delta foreset beds of gravel, sand, and silt; locally tufa-cemented; lo- cally overlain by marl beds Indurated lake clay, silt, sand, and gravel; at least 2 ages of deposits, many tephra lay- ers; Pliocene (Davis, 1981) Indurated lake clay, silt, and sand; at least 2 ages of deposits, many tephra layers; Pliocene (Reheis et al., 1993) Indurated lake clay, silt, and sand; at least 2 ages of deposits, many tephra layers; Pliocene Indurated lake clay, silt, sand, and gravel; at least 2 ages of deposits, many tephra lay- ers; Miocene to late Pliocene east side of Mono Lake. Reheis et al. (1993) documented sev- eral Pliocene to middle Pleistocene cycles of pluvial Lake Ren- nie in Fish Lake Valley (Figure 1) and proposed that Lake Ren- nie merged with Lake Columbus to the north about 760 Ka at a shoreline altitude of about 1460 m. METHODS The vastness of the Lahontan Basin and the large number of other pluvial lakes in Nevada precluded a systematic search for shorelines older and higher than those of late Pleistocene age. This search relied heavily upon (1) extrapolation from known localities in the Walker Lake subbasin and in Newark, Long, and Diamond Valleys (Figure 1), (2) Mifflin's recollections of possible shoreline localities from field work in the 1970s, and (3) observation from roads. Geologic maps were searched for outcrops of late Tertiary or Quaternary fine-grained deposits; several sites in the northern Walker Lake subbasin, for exam- ple, had been mapped as Miocene to Pliocene sedimentary rocks (Greene et al., 1991) but in fact are Quaternary lacustrine deposits. Potential sites were examined on l:24,000-scale topo- graphic maps and (or) aerial photographs for relatively flat benches above late Pleistocene shorelines. Dozens of potential sites were checked in the field, but only those with evidence of lacustrine deposition are listed in Table 1. The altitudes and ages of shorelines (Table 1) were estimated from shoreline features and from sedimentologic evidence. Classic shoreline morphologies, such as wave-cut cliffs and 60 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES beach berms, become increasingly difficult to discern with age and are rarely preserved after half a million years or more of erosion. Thus, the highest altitude of beach-gravel beds or lag on bedrock slopes was taken as a minimum shoreline altitude. Where beach morphology was discernible, the preservation of berm features and the degree of development of relict soils per- mitted assignment of relative ages. Fine-grained beds and even some gravels locally contained tephra layers; these were identi- fied by chemical correlation with tephra of known age by the United States Geological Survey (USGS) Tephrochronology Laboratory. Ages of some units also were constrained by mag- netostratigraphy and by vertebrate biostratigraphy. Glass shards from tephra layers found within lacustrine se- quences during mapping were analyzed by electron microprobe for major-oxide composition (Table 2) using methods de- scribed by Sarna-Wojcicki et al. (1984) and Sarna-Wojcicki (2000). Sample compositions were compared with composi- tions in a data base of previously analyzed tephra layers using statistical programs (Table 3). Correlations were based on simi- larity in chemical composition, petrographic characteristics (such as shard morphology, phenocryst mineralogy), and strati- graphic data. Values for glass compositions in Tables 2 and 3 were normalized to 100 percent to correct for the variable amounts of hydration of the volcanic glass, as reflected by the original totals. In addition to hydration, some postdepositional alkali exchange had occurred in many of the samples. Variabil- ity in sodium and potassium contents was particularly apparent for the older tephra layers. Samples used for paleomagnetic directions were collected as large (typically 500-1000 cm3) blocks of sediment that were oriented in the outcrop using a magnetic compass and a level to mark the horizontal plane. Many samples were sufficiently in- durated to carve from the outcrop; some friable samples were hardened using sodium silicate before extraction. In the labora- tory, smaller samples were shaped from block samples to fit into plastic cubes (3.2 cm3 volume). We estimate that the final orientation of each sample was within ?10? of the magnetic field declination and within 10? of the original horizontal atti- tude. Such orientation is adequate for polarity determinations at these sites. We corrected for tectonic tilting when bedding dips exceeded 5?. The natural remanent magnetization of the ori- ented samples was measured using a 90 Hz spinner magnetom- eter having a sensitivity better than 10~5 amperes per meter (A/m). Samples were demagnetized by alternating-field (AF) demagnetization with peak fields of 5, 10, 20, 40, 60, and 80 milliteslas (mT). We evaluated the paleomagnetic reliability of each sample from an examination of orthogonal demagnetiza- tion diagrams that show the change in remanent direction and magnitude in response to each demagnetization step (Zijder- veld, 1967). Samples that showed unstable behavior (lack of systematic decay in magnitude and direction toward the origin of the diagram over most of the demagnetization range) were not considered further and are represented by a question mark on the diagrams of stratigraphic sections. Polarity was deter- mined for stably magnetized samples: normal, when the rema- nent vector was northward (clockwise between 60? and 300?) with a moderate to steep, positive inclination (between 20? and 80?); reversed, when the remanent vector was southward (clockwise between 120? and 240?) with a moderate to steep, negative inclination. The expected inclination for the latitude of 39?N through the study region is about 58?. Several samples exhibited magnetic behavior that reflected the removal of a normal polarity overprint in response to low AF inductions to isolate a reversed component of magnetization. Although incli- nations in these samples may not stabilize in moderately steep angles, the magnetization clearly is reversed. Such samples are listed as having a reversed component (R comp; see Figure lie for sample P10). Samples that showed stable magnetization are listed as questionably normal ("N?") or questionably reversed ("R?") if they had moderate to steep positive or negative incli- nation, respectively, with a declination beyond ranges for a clearly defined polarity. ACKNOWLEDGMENTS This project was initially funded as a Gilbert Fellowship by the USGS; Martin D. Mifflin was the coproposer with Reheis and was the behind-the-scenes advisor, freely sharing his inspi- ration and vast store of Lahontan Basin knowledge. Roger B. Morrison reviewed early versions of the proposal and kindly led us to several areas with "old" Lake Lahontan deposits. Ken- neth D. Adams (University of Nevada, Reno) made observa- tions at the Campbell Valley and Sunshine Amphitheater sites and subsequently made the sites available to us for study. Scott Stine (California State University at Hayward) participated in the initial, startling finds of very high beach gravels in the Mono Basin. Bruce Braginton (Newmont Gold Corporation) graciously arranged and shepherded our field work in the Lone Tree Mine and gave us access to one of the long cores. James Honey (University of Colorado) and Robert R. Miller (Univer- sity of Michigan) kindly identified some of the vertebrate fos- sils. Richard M. Forester and J. Platt Bradbury (USGS) identi- fied and interpreted ostracode and diatom samples, respectively. Charles E. Meyer (USGS) performed much of the tephrochronologic analyses. Joseph G. Rosenbaum and Will- iam C. Rivers (USGS) instructed Reheis on the initial analyses of paleomagnetic samples and advised her on their interpreta- tion. Jonathan Rademaekers and Michael O'Donnell (USGS) produced the color figure using Geographic Information Sys- tems (GIS) techniques; George Garcia (USGS) provided some of the computer drafting. Finally, Reheis thanks her field assis- tants, who worked very hard in bad weather and rough camping conditions: Andrea Lonner, Nadine Calis, Cassandra Fenton, Rebecca Gagliardi, Kathleen Johnson, Marketta McGuire, and Joanna Redwine. NUMBER 33 61 Evidence for Old Pluvial Lakes in the Western Great Basin We have adopted several conventions for ordering the infor- mation presented and for naming the lacustrine units so as to accomodate the complex geographic, stratigraphic, and chro- nologic relationships among the many sites discussed in this paper. Because most of the work to date has focused upon the Lahontan Basin and upon the Walker Lake subbasin in particu- lar, we begin by describing these primary study sites from south to north, followed by evidence from supporting locali- ties, also from south to north. Other pluvial-lake records in western and central Nevada are described in chronological or- der by lake basin, working roughly from southwest to north- east. At each study site, the lacustrine units were numbered in order from oldest to youngest, followed by a subscript letter designating the site: for example, the youngest lake unit at Thome Bar is Q17T. As a result of differential deposition and preservation, this convention can produce same-numbered units of different ages at different study sites; however, it avoids bias in the nomenclature and in the correlation of some lake units that are undated or whose ages are only relative. Based upon current evidence and best judgment, we then corre- late the units from site to site and assign names to deposits of similar age; names were chosen for localities where a unit is well exposed. For example, deposits associated with the 1400 m highstand of Lake Lahontan were named "lake deposits of McGee," for the well-preserved arcuate beach ridge above Mc- Gee Wash that was deposited during this highstand. THE LAHONTAN BASIN The Lahontan Basin consists of a sprawling complex of sub- basins (Figure 1), many of which have independent hydro- graphic histories (Davis, 1982; Reheis and Morrison, 1997) controlled not only by climate but also by interbasin river di- versions and changes in threshold altitudes. Of the four princi- pal rivers?the Truckee, Carson, and Walker, draining the Si- erra Nevada, and the Humboldt, draining northern and eastern Nevada?all but the Carson River have been diverted among different subbasins, some more than once. Thus, lake levels re- corded in some subbasins may not apply to other subbasins, es- pecially at moderate to low water levels. EVIDENCE FROM THE WALKER LAKE SUBBASIN.?The Walker Lake subbasin (Figures 1, 2) contains much of the stratigraphic evidence of early and middle Pleistocene lakes with high shorelines in the Lahontan Basin. This evidence must be interpreted with caution because the area is tectonically ac- tive. The Walker Lake subbasin is bounded on the west by the Wassuk Range and the active Wassuk normal fault and is bounded on the east by the Gillis Range. At the north end of the Gillis Range, faults of the strike-slip Walker Lane belt (Stew- art, 1988) intersect the basin. The Thorne Bar area lies in a tec- tonically quiescent area of the basin (Demsey, 1987), but other sites discussed below have been affected by faulting. The Walker River, the only major source of water, arises in the Si- erra Nevada and presently feeds Walker Lake. At times in the past, however, the Walker River flowed north into the Carson Desert via Adrian Valley (Morrison and Davis, 1984; Bradbury etal., 1989; King, 1993). Thorne Bar Area: The beach ridge, or bar, at Thorne Bar is a large, V-shaped barrier complex of shore gravel built at the mouth of a large canyon that drains the Gillis Range (Figures 2, 3). The part of this complex above the Sehoo shoreline was previously described as pre-Sehoo lacustrine gravel that was elevated by northward tilting of the Lahontan Basin (King, 1978; Mifflin and Wheat, 1979; Morrison, 1991). Detailed lev- eling studies by Demsey (1987) and interpretation (this study) of aerial photographs revealed no evidence of Quaternary de- formation on this side of the Walker Lake subbasin. We inter- pret the altitudes of the bar and its morphologic subdivisions to represent essentially undeformed shorelines. The bar consists mainly of pebble- to cobble-sized gravel that is locally tufa-cemented; the bedding ranges from horizon- tal to angled at 25? or more. The bar can be divided into four morphologic units (Figures 3, 4A). The lowest and youngest unit is the group of historic shorelines below 1250 m in alti- tude, formed during the last 100 years; this unit is not sepa- rately numbered. The next higher group of shorelines is distin- guished by sharp, well-preserved barriers (unit Q17T) that are marked at the top by the Sehoo highstand shoreline at an alti- tude of about 1330 m (~4360 ft; Figure 3). A small loop bar bears a soil profile with a thin vesicular A horizon and a slightly oxidized Bw horizon with a few thin clay films and lit- tle pedogenic CaCO3 (Table 3, profile WL-5c of Adams and Wesnousky, 1999). This degree of development is typical of Sehoo-aged or younger soils (Chadwick and Davis, 1990). Between 1330 m and about 1370 m (4360-4500 ft), the bar complex consists of two nested, eroded V-shaped barriers; the lower (unit Q16T, marked by vehicles in Figure 4A) is better preserved than the upper (unit Q15T). Soil pits on the eroded surfaces of these barriers show increasing thicknesses of Av and Bw horizons and increasing cementation by pedogenic CaCO3 and silica of subsurface horizons relative to barrier alti- tude and to the Sehoo-aged soil, although clay and silt content remain low (Adams and Wesnousky, 1999). These soil features indicate that the barriers increase in age upward and are older than the Sehoo highstand (Chadwick and Davis, 1990), but the soils are not adequately preserved for a more accurate age as- sessment. Between 1370 m and 1402 m (4500-4600 ft) the complex has no preserved morphology, but deep arroyos have exposed well-bedded, cemented shore gravel of unit Q14T, and beach pebbles locally occur as lag on basalt outcrops. In summary, the degree of morphologic preservation and the degree of soil development in three reconnaissance soil pits (Ad- ams and Wesnousky, 1999) indicate the presence of at least four lacustrine units above the historic limit at 1250 m: the Sehoo- aged bar, unit Q17T, extending upward from the historic limit to 62 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ? J: J: S S ?3_P ? BIS * >> Soil 2 e Long 2 03 Total O ~~^ Tota n O z nm O o 2 o 2 rnQ LL. o n -|_z dy site Stu eno.o. c3oo ON ON ON ON CO ON inSO rr rrON co ?? ? rT ? SO ?? ro? ? m in ? CN fN rTin r-p- CO co? 00 p- rT 00 CO t Marsh esSa South o UJ P-? ? ? so CN? ? co? ? CN? ? s ? in ? rTso ? -fi OO ~^ s o o sorT ON ON in ro rr ? ? roCO ? 00 ?? rr? ? CN SO ? 00 00 fN SO SO so P-. 00 in 17? SO in p- ro Valley Lake u. The Gai \UJ rr? o S ? ro ? ? s ? ?? SO? ? ? rr? +1 00 IIB O? o? p-ON ro ON CO00 in ?in ro so? ? ONin ? ? ? ? ? SO in ? ro CN SO so ooin 17? so inin r-^ co Valley Lake on U. The Gai u- rTK UJ fN ? SOCN ? ? ? ro? ? fN ?? ? ? SO? ? P- ? CO ? +1 SO ~^ IIB ?? ?? P- coON P-rr co ON fN in ? fN? ON rT ? ? infN ? CO fN P-ON rT OO ON CN ?in 00 SO CN? ON ro A er Lake Walk j="1 McGee n i/i UJ ? 1?1 O g ? in? ? CN ?? CN? ? ? ? 00 ? ?rT ? er Da X> (WD M in IIs o? ?? ?? rr in rr so ?? rr? ? ? ro? ? so p^ ? ro00 CN inm so so o? 19? p~ in? ON ro S x> ?> alker La all x> a. ? er Lake1 Was McGee Q ON ??( '-, UJ r- ? rT ? ?? ? ? ? CN ?? CN? ? CN ? r- ? 00 CN ? ? IIB o sofN ON 00SO rr ro ?? ? rrrr ? s ? CO? ? 00 p- ? oo 00 CN 00r- so p- ON ONrr 18? in CN? ON co X> er Lake Walk McGee Q >? ? CN '?, UJ o ? 00o ? ? ? CN? ? CN ?? ? ? OO ? ? oCN ? p~ in ? +1 ON *?'II B ?? OO00 rr ON ONSO rT s ? CNrr ? ? ? ? ? rot? ? on? ro in so r- ro noCO 18? p^ CN 00 ro B basi ?c ca r, Walke ea CQo pj North o >?,?, fN 'iUJ CO? ? no ? ? ? ? ro? ? CN ?? rr oo ?? rr ? soo ;?s rTII B CNO ?? in? ON ON? SO r-~ co p- ? ONin ? ? ? so? ? p- *"* soin fN no in inp- CN rr? 18? in infN 00 co ley ca Springs Soda ca"B East of rrCN \ UJ CN ? CN ? g ? rT? ? g ? ? ? ? ? rT ? ?rr ? -fi ? CN IIs ? ?? rrCO ON 00 rr CN00 rT s ? fNro ? ? ro? ? SO in ? 00ON fN ON CN SOP- co ro? 18? r-rr oo ro t Marsh esSa -% South o sors 1 PL) ON ? ON ? CN ?? ro? ? fN ?? ? ? s ? sofN ? inso ? oo ^~II s ? - p-CN co ON ?so in 00 SO ro s ? roCO ? 00 ?? in? ? ? 00 CN SO SO p- p- ON in 17? ON co? 00 co Summ Redlich Pi ON CN'*, UJ rrfN ? fN CN ? g ? CN? ? fN ?? ? ? SO? ? fN ? ?in ? +\ so ^*II s ?? ?? oCN ON o00 SO p- 00 ro in ? ONco ? ? ? ON? ? rrCN ?" CNin ro rr00 rop- ^_ ? 00 in ro 00 ro Summ Redlich 4 '-. UJ COrr ? 00 CN ? ? ? fN? ? ro ?? ? ? so? ? inro ? rT00 ? +1 OO ^- IIB ? ON ON ON ONrT ON 00 p-ro SO in fN fN NO ? roin ? in ?? in? ? ro SO ? ON in fN r- roP- P~ rT inrr 18? in ,_ fN 00 co B eaCQ oo 2 o , North anyonO ring MudSp oo in CO UJ so? rTin ? CN ?? in? ? (N ?? ? ? in ?? 00 CM ? 00 ? oo +1ON *-? IIS 00 ON ON ON ON ON SO in rTCO co so ?? 00rr ? rr ?? rT? ? 00 in ? roON CN ? ON SOP- rT in rr 00 in ,?, CN 00 ro B t/5 eaCQ o o 2o , North anyonU ring MudSp oo min ro\ UJ oSO ? ?in ? g ? ? g ? ? ? rT? ? SOCN ? SO00 ? cn +1 ON ^* IIE ? no ON ofN in COin co p- ?? norr ? SO ?? in? ? ON in ? p- fN rop- P- rT in 00 in ,_ fN 00 CO B mCQ O o f, , o , North inyonO ring MudSp 00 soCO \ UJ soCO ? fNCN ? ? ? ? ? co? ? ? ? in? ? 00? ? 00 SO ? ;?s p- IIE ? - P^ ON fNfN ro in soin ro ? ON 00 ? CO ? p-fN ? P- so MgO 61 U. o > 2 C S E en ?? '> g a a p T3 -a S ? CB a. a. a.E E E 64 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 3.?List of tephra layers analyzed in this study and their compositionally closest matches, based on electron- microprobe analysis of volcanic glass shards. All samples were from lacustrine deposits unless otherwise noted; lo- calities are given in Table 2. Analyses and methods are given in Table 2. Closest matches are based on calculations of similarity coefficients using oxides of the elements Si, Al, Fe, Ca, Na, and K, except as noted below, and data not used in the calculations are shaded in gray. Closely matched samples that did not provide age constraints or strati- graphic context were omitted from the table. Output no. Field sample no. SiO2 A12O3 Fe2O3 MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer in roadcut south of Rhodes Salt Marsh (1). Closest matches are with the Teels Lake tephra sequence (2-4, 11), the tephra layers in lake sediments of the Mono Basin (7, 9; 2.3 Ma; R. Fleck, in litt. to A.S.-W., 1997), and the tephra layer at Willow Wash, Nevada (8; <2.57 Ma, >2.22 Ma; Sarna-Wojcicki et al., In review). 1 2 3 4 7 8 9 11 1 2 3 4 5 6 8 9 10 1 4 5 6 9 EL-2-RM TLN-1 TLN-3 TLN-4 EL-54A-ML FLV-36-WW EL-54B-ML TLN-2 77.54 77.28 77.20 77.26 77.38 77.47 77.53 77.06 12.72 12.76 12.87 12.83 12.76 12.83 12.64 12.96 0.55 0.56 0.54 0.58 0.55 0.58 0.53 0.58 0.03 0.04 0,04 0.04 0.04 0.05 0.03 0.05 0.06 0.08 0.09 0.08 0.07 0.09 0.07 0.09 0.44 0.44 0.43 0.45 0.40 0.43 0.41 0.42 0,06 0.08 0.09 0.08 0.08 0.11 0.07 0.09 3.94 4.09 4.03 3.95 3.83 3.66 3.82 4.22 4.65 4.69 4.70 4.74 4.91 4.76 4.91 4.54 99.99 100.02 99.99 100.01 100.02 99.98 100.01 100.01 Tephra layer at The Gap (1). Closest matches are with the tuff of Taylor Canyon tephra layers and related units (2-6, 8-10; Sarna-Wojcicki et al., EL-3-FL FLV-150-CS PC-C WAC-6 CH88004 FLV-151-CS TTC-6 TTC-19(2) DG-4 76.66 77.26 76.76 77.20 76.67 77.34 77.20 77.54 76.96 In review). 12.88 12.71 13.14 12.74 13.20 12.70 12.93 12.57 13.03 0.62 0.61 0.60 0.60 0.64 0.60 0.58 0.59 0.56 0.04 0.03 0.03 0.04 0.05 0.04 0.05 0.05 0.04 0.08 0.09 0.08 0.06 0.09 0.09 0.09 0.07 0.09 0.33 0.34 0.32 0.34 0.34 0.34 0.33 0.33 0.33 0.07 0.05 0.07 0.09 0.08 0.05 0.09 0.06 0.08 4.13 3.96 3.90 3.98 3.80 4.09 4.08 3.91 4.00 5.19 4.96 5.10 4.94 5.13 4.74 4.64 4.89 4.92 100.00 100.01 100.00 99.99 100.00 99.99 99.99 100.01 100.01 Tephra layer at The Gap (1). Closest matches are with the tuff near Mount Jackson (5), the tuff (surface) in Clayton Valley (4), and the tephra layers in Mud Spring Canyon north of Mono Basin (this study; EL-4-FL D77-27B MJ-1A EL-37-MS EL-35B-MS 76.21 77.05 76.95 76.17 76.90 13.14 13.00 13.05 13.64 12.93 0.56 0.58 0.56 0.60 0.58 0.04 0.04 0.04 0.05 0.04 0.07 0.04 0.07 0.05 0.04 0.59 0.53 0.50 0.52 0.48 0.06 0.09 0.08 0.10 0.06 6,9). 3.50 3.45 3.26 3.47 3.34 5.83 5.22 5.49 5.40 5.61 100.00 100.00 100.00 100.00 99.98 1.0000 0.9884 0.9850 0.9821 0.9705 0.9703 0.9676 0.9647 1.0000 0.9747 0.9739 0.9727 0.9706 0.9699 0.9678 0.9675 0.9674 1.0000 0.9539 0.9507 0.9492 0.9451 Tephra layer in alluvial deposits at McGee Wash (1). The only reasonably close match is to the tephra layer at Weber Reservoir, in Lovelock Formation(?) (2; J.O. Davis, unpublished data); possibly the same sample locality. Note: 8- element comparison, excludes Mn. 1 EL-5-WD 71.98 14.97 2.31 0.25 0.04 1.49 0.20 5.29 3.47 100.00 2 DR-96 72.64 14.93 2.33 0.24 0.05 1.46 0.24 4.71 3.41 100.01 Tephra layer at Campbell Valley (1). Closest matches are with the Bishop ash bed in Owens Lake core OL-92 (3, 5, 9, 10), at Lake Tecopa, California (7), and in Adobe Valley (6); and with the ~1 Ma Glass Mountain tephra at Mc- Gee Wash (8). 12.83 0.76 0.03 0.05 0.44 0.06 12.91 0.74 0.04 0.03 0.44 0.07 13.05 0.73 0.03 0.04 0.44 0.02 13.23 0.76 0.03 0.05 0.44 0.07 12.89 0.75 0.04 0.03 0.43 0.06 12.88 0.78 0.03 0.04 0.44 0.06 12.95 0.74 0.03 0.03 0.45 0.07 12.97 0.73 0.04 0.04 0.45 0.06 Tephra layer in alluvial deposits at McGee Wash (1). Closest matches are with the upper Miocene tephra layers (3-5; 11.81 ?0.03 Ma correlated age; A.M. Sarna-Wojcicki, unpublished correlated age to date in Perkins et al., 1995). Note: 7-element comparison, includes Ti. 1 EL-19-WD 75.88 12.23 2.02 0.09 0.03 0.70 0.25 3 DSDP173-21-4 (27-28 cm) 76.40 12.16 2.14 0,06 0.03 0.69 0.26 4 ALVES-1 76.33 12.12 2.08 0.05 0.03 0.65 0.23 5 DSDP173-21-4 (32-35 cm) 76.47 12.08 2.08 0.06 0.05 0.69 0.23 1 3 5 6 7 8 9 10 EL-15-CA OL92-1030 OL-92-2 (97) EL-38-AV TECO-28B EL-20-WD OL92-1022 OL92-1028 (major) 76.55 76.94 76.88 76.69 77.00 76.78 76.95 76.89 1.0000 0.9532 4.14 3.90 4.12 3.81 3.80 4.31 3.99 3.97 5.14 4.93 4.70 4.93 5.00 4.68 4.79 4.85 100.00 100.00 100.01 100.01 100.00 100.00 100.00 100.00 1.0000 0.9773 0.9748 0.9746 0.9740 0.9731 0.9721 0.9709 3.36 3.12 3.24 2.99 5.42 5.14 5.28 5.36 99.98 100.00 100.01 100.01 1.0000 0.9651 0.9633 0.9622 NUMBER 33 65 TABLE 3.?Continued. Output no. Field sample no. 1 7 9 10 11 SiO2 A12O3 Fe2O3 MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer at McGee Wash (1). Closest matches are with the Bishop ash bed (11, 12, 14), with the younger tephra layers of Glass Mountain (7; 1.2-0.8 Ma), and with the tephra layers near the Thorne Bar (9) and in Campbell Val- ley (10). 76.78 12.88 0.78 0.03 0.04 0.44 0.06 4.31 4.68 100.00 77.29 12.83 0.79 0.03 0.04 0.45 0.06 4.03 4.48 100.00 76.52 13.08 0.73 0.03 0.03 0.42 0.06 4.44 4.69 100.00 76.55 12.83 0.76 0.03 0.05 0.44 0.06 4.14 5.14 100.00 77.20 12.83 0.76 0.04 0.08 0.42 0.06 4.20 4.42 100.01 76.52 12.90 0.80 0.04 0.05 0.44 0.06 4.10 5.11 100.02 77.40 12.81 0.74 0.03 0.04 0.44 0.06 4.20 4.27 99.99 EL-20-WD PICO-73 EL-21-WL EL-15-CA OL92-1019 12 OL92-1027 (minor) 14 JC-239 Tephra layer at site north of Thorne Bar (1). Closest matches are with the Bishop ash bed (2, 6, 10, 14), with the younger tephra layers of Glass Mountain (5; 1.2-0.8 Ma), and with the tephra layer at McGee Wash (3). 1 EL-21-WL 76.52 13.08 0.73 0,03 0.03 0.42 0,06 4.44 4.69 100.00 2 OL-92-2(97) 76.88 13.05 0.73 0.O3 0.04 0.44 0.02 4.12 4.70 100.01 3 EL-20-WD 76.78 12.88 0.78 0.03 0.04 0.44 0,06 4.31 4.68 100.00 5 PICO-78 (2) (100-200 cm) 76.33 12.93 0.73 0.03 0.04 0.41 0.06 4.24 4.24 99.01 6 OL92-1019 77.20 12.83 0.76 0.04 0.08 0.42 0.06 4.20 4.42 100.01 10 OL92-1029 76.92 13.06 0.71 0.03 0.04 0.42 0.05 3.92 4.85 100.00 14 OL92-1028 (minor) 76.73 13.16 0.71 0.04 0.01 0.45 0.07 4.08 4.75 100.00 Tephra layer east of Mina (1). No good specific matches. Generally similar to upper Miocene and lower Pliocene tephra layers of the Snake River Plain eruptive source areas, Idaho (2), and with the tephra layers at Drumm Summit (3,4). 1.0000 0.9745 0.9734 0.9731 0.9731 0.9729 0.9704 1.0000 0.9789 0.9734 0.9702 0.9702 0.9693 0.9672 1 2 3 4 1 3 5 9 10 11 12 13 14 1 3 11 1 2 3 4 6 1 6 8 12 13 EL-24-M F8-90-G21 (39-41 cm) (high Fe, low Na) EL-66-FV EL-59-FV (2) 75.58 76.32 76.28 76.52 12.56 12.37 12.22 11.99 1.77 1.67 1.86 1.88 0.06 0,07 0.04 0.04 0.03 0.04 0,02 0.03 0.59 0.59 0.60 0.58 0.17 0,17 0.18 0,17 3.17 2.71 2.94 2.73 6.09 6.06 5.87 6.06 100.02 100.00 100.01 100.00 Tephra layer in roadcut south of Rhodes Salt Marsh (overlies EL-2-RM) (1). Closest matches are with the tuff of the Taylor Canyon tephra layers (3, 5, 9-14; 2.22-2.14 Ma; Sarna-Wojcicki et al., In review). EL-26-RM DG-4 FLV-SP-11 DG-3A TTC-6 FLV-43-WW P-169Q PC-C TTC-19 (2) Tephra layer at 76.29 76.96 77.07 76.88 77.20 78.00 77.59 76.76 77.54 12.98 13.03 13.07 12.92 12.93 12.32 12.55 13.14 12.57 0.56 0.56 0.56 0.63 0.58 0.56 0.58 0.60 0.59 0.03 0.04 0.03 0.02 0.05 0.03 0.04 0.03 0,05 : Redlich Summit (1). Closest matches are -1.95-1.79 Ma; Sarna-Wojcicki et al., In review). EL-29-RM CH88014 TTC-23 76.61 76.84 77.58 12.81 12.91 12.62 Tephra layer at Redlich Summit, dated 5 the tephra at Willow Wash, California (6; EL-34-RM FLV-83-HT FLV-80-HT EL-65-RM FLV-105-WW 73.84 74.11 74.55 73.50 74.57 13.52 13.89 13.45 14.26 13.81 0.76 0.76 0.71 .9Ma(l). 0,05 0.02 0,03 o.n0.09 0.09 0.09 0.09 0,08 0.10 0.08 0.07 with the 0.08 0.07 0.03 0.32 0.33 0.33 0.31 0.33 0.32 0.33 0.32 0.33 0.06 0.08 0.07 0.05 0.09 0,06 0.06 0.07 0.06 older Glass Mountain 0.33 0.32 0.32 0,09 0.06 0.03 4.82 4.00 3.89 4.45 4.08 3.79 3.87 3.90 3.91 tephra 3.68 3.71 3.58 4.84 4.92 4.89 4.63 4.64 4.84 4.87 5.10 4.89 layers (3, 5.60 5.31 5.10 Closest matches are with the tuffs in Horsethief Canyon (2 100.01 100.01 100.00 99.98 99.99 100.00 99.99 100.00 100.01 11; 100.01 100.00 100.00 ,3), Reheis et al., 1991), and the overlying tephra layer at Redlich Summit (4). 1.24 1.19 1.17 1.11 1.21 0.09 0.10 0.07 0.09 0,09 Tephra layer in alluvial deposits at Mud Spring Canyon (V Wash (6, 8, 12, EL-35A-MS FLV-34-WW FLV-121-WW FLV-33-WW FLV-25-WW 13; <2.83, >2.57 Ma; Sarna-Wojcicki et al., 77.37 76.85 76.61 76.87 77.60 12.59 12.93 12.95 12.95 12.45 0.63 0.62 0.65 0.64 0.61 0.05 0.08 0.04 0.07 0.07 0.10 0.09 0.10 0.10 0.10 0.39 0.37 0.37 0.38 0.38 0.15 0.19 0.19 0.19 0.20 3.87 3.84 3.58 3.71 4.35 6.80 6.21 6.52 6.66 5.28 100.00 99.99 100.00 100.00 99.99 ). Closest matches are with the tephra layers at Willow In review) 0.05 0.06 0.06 0.07 0.07 0.53 0.62 0.57 0.62 0.56 0.06 o.to 0.07 0.12 0.10 2.25 2.52 2.85 2.69 2.89 6.37 6.22 6.20 5.96 5.66 99.90 100.00 100.00 99.99 100.01 1.0000 0.9669 0.9621 0.9604 1.0000 0.9618 0.9583 0.9542 0.9541 0.9523 0.9470 0.9455 0.9454 1.0000 0.9832 0.9600 1.0000 0.9639 0.9602 0.9585 0.9309 1.0000 0.9459 0.9374 0.9295 0.9279 66 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 3.?Continued. Output no. Field sample no. SiO2 A12O3 Fe2O3 MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer in alluvial deposits at Mud Spring Canyon (1). Closest matches are with the tuffs near Mount Jackson (2, 5; 2.9 ? 0.16 Ma, S. Weiss, in litt. to M.C.R., 1993), an overlying tephra layer (3), and the tuff (surface) in Clay- ton Valley, Nevada (7). EL-35B-MS 76.90 12.93 0.58 0.04 0.04 0.48 0.06 3.34 5.61 99.98 MJ-1A 76.95 13.05 0.56 0.04 0.07 0.50 0.0S 3.26 5.49 100.00 EL-36-MS 77.31 12.71 0.59 0.05 0.06 0.48 0.07 3.53 5.20 100.00 MJ-1B 76.90 13.00 0.59 0.04 0.08 0.49 0.07 3.62 5.20 99.99 D77-27B 77.05 13.00 0.58 0.04 0.04 0.53 0.09 3.45 5.22 100.00 Tephra layer in alluvial deposits at Mud Spring Canyon (1). Closest matches are with the (2; see reference above), an underlying tephra layer (3), in southeastern Mono Basin (6); Valley, Nevada (4), a similar tuff at Willow Wash (5; 11.3 Ma; Reheis and Sawyer, 1997), at Willow Wash (10), -2.57 Ma (Sarna-Wojcicki et al., In review). tuff near Mount Jackson with the tuff in Clayton and with the tephra layer Tephra layer at Clayton Valley (1). Closest matches are with the ash flow at Mount Adams, Washington (2, 3; ~510 Ka, W. Hildreth, unpub. data), with the tuff above Huckleberry Ridge ash bed at Confidence Hills, Death Valley, California (4, 5; -2.04 Ma, Sarna-Wojcicki et al., In review), and with the tuff in Burmeister core, Great Salt Lake, Utah (6, 10; -3 Ma., Williams, 1993). 1 2 4 5 6 7 8 EL-42-CV #1 (major) SSP-2 C-83-6A ASW-613 84-27 ASW-61286-14 BUR-915 FLV-WW-52 72.20 72.25 72.50 72.98 73.51 72.02 72.48 5.65 5.68 5.39 5.40 5.53 5.66 6.10 3.22 3.30 3.36 3.17 3.24 3.63 3.19 100.01 99.99 100.01 100.01 100.00 100.01 100.00 1.0000 0.9912 0.9788 0.9765 0.9758 0.9743 0.9697 Tephra layer at Clayton Valley (1). Closest matches are with the La Salida tuff (2, 4, 5, 6), the tephra layer from the Burmeister core in the Great Salt Lake, Utah (7; 3.06 Ma; Williams, 1993), and the tephra layer in Willow Wash, California (8). 14.67 2.67 0.27 0.13 0.89 0.31 14.56 2.66 0.24 0.08 0.90 0.32 14.47 2.73 0.27 0.11 0.89 0.29 14.17 2.66 0.26 0.12 0.92 0.33 13.59 2.60 0.27 0.08 0.89 0.29 14.46 2.67 026 0.11 0.87 0.33 13.92 2.70 0,27 0.10 0.92 0,32 Tephra layer at Clayton Valley (1). Closest matches are with the Spearhead Member of Stonewall Flat tuff (2; 7.61 ? 0.03 Ma; and 3; 4.02 ? 0.08 Ma) as well as with similar tuffs in southern Fish Lake Valley (4, 5; Reheis and Sawwyer, 1997), but Fe and K are slightly different. Also similar to the tephra in Willow Wash (6; <2.81, >2.57 Ma). 0.01 0.06 0.31 0.14 0.01 0.07 0.31 0,12 0.00 0.07 0.32 0.11 0.00 0.08 0.31 0.09 0.00 0.07 0.31 0.10 0.02 0.05 0.35 0.15 Tephra layer at The Gap (1); underlies EL-3-FL. Closest matches are with the tuff of Taylor Canyon, 2.22-2.14 Ma (Sarna-Wojcicki et al., In review), from Emigrant Pass area, Nevada (3, 8), from Waucoba lake beds (11), and from Cowan Pumice Mine, California (9, 13). 1 2 3 4 5 6 EL-43-CV 61990-1 B85-195-A FLV-50-CM FLV-5-CC FLV-31-WW 76.61 76.42 76.07 76.86 76.56 77.47 12.17 12.05 12.35 11.75 11.89 12.02 1.97 1.85 1.86 1.87 1.91 2.01 1 3 8 9 11 13 EL-44-FL FLV-183-EP FLV-180-EP TTC-20 W2A TTC-21 1.0000 0.9784 0.9723 0.9678 0.9662 1 2 3 4 5 6 10 EL-36-MS MJ-1B EL-53-ML D77-27B FL-KA-3 EL-35B-MS FLV-35-WW 77.31 76.90 77.27 77.05 77.54 76.90 77.58 12.71 13.00 12.69 13.00 12.27 12.93 12.69 0.59 0.59 0.57 0.58 0.62 0.58 0.60 0.05 0.04 0.04 0.04 0.07 0.04 0.06 0.06 0.08 0.06 0.04 0.04 0,04 0.07 0.48 0.49 0.45 0.53 0.47 0.48 0.49 0.07 0.07 0.08 0.09 0.09 0.06 0.10 3.53 3.62 3.63 3.45 3.47 3.34 3.76 5.20 5.20 5.22 5.22 5.43 5.61 4.65 100.00 99.99 100.01 100.00 100.00 99.98 100.00 1.0000 0.9879 0.9784 0.9728 0.9723 0.9723 0.9652 1 2 3 4 5 6 EL-42-CV #2 (high K) MA-360 (A) MA-360 (B) CH88008 TL88001 BUR-915 (high Fe) 71.48 71.49 71.62 70.92 70.94 73.05 14.53 14.44 14.48 15.64 15.62 14.13 2.73 2.72 2.86 2.71 2.67 2.75 0,26 0.19 0.20 0.2S 0.27 0.25 0.09 0.07 0.08 0.09 0.09 0J0 0.92 0.87 0.91 0.96 0.96 0.86 0.36 0.33 032 0.29 0.30 0.32 5.06 5.12 4.80 5.06 5.01 4.52 4.57 4.77 4.73 4.06 4.15 4.01 100.00 100.00 100.00 100.01 100.01 99.99 1.0000 0.9803 0.9755 0.9601 0.9595 0.9415 4.24 4.24 4.24 3.95 3.91 3.42 4.49 4.94 4.98 5.10 5.25 4.51 100.00 100.01 100.00 100.01 100.00 100.00 1.0000 0.9726 0.9655 0.9539 0.9539 0.9408 76.63 77.20 77.14 77.70 77.35 77.68 13.04 12.70 12.89 12.58 12.48 12.57 0.58 0.56 0.59 0.58 0.60 0.60 0.03 0.05 0.04 0.05 0.03 0.05 0.12 O.U 0.11 0.06 0.08 0.07 0.32 0.31 0.32 0.31 0.34 0.32 0.07 0.07 0.10 0.05 0.08 0.06 3.43 3.23 2.99 3.38 3.30 3.33 5.77 5.78 5.82 5.29 5.74 5.32 99.99 100.01 100.00 100.00 100.00 100.00 1.0000 0.9735 0.9713 0.9703 0.9687 0.9683 NUMBER 33 67 TABLE 3.?Continued. Output no. Field sample no. SiO, Al9 Fe-, MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer at Redlich Summit (1). Closest matches are the Wono ash bed of Davis (1978) (27.2 Ka) (Benson et al., 1997), from a core off the Oregon coast (6), and from Nixon, Nevada (9, 12, 17, 18). 1 6 9 12 17 18 EL-50-RM F8-90-G21 (37-39 cm) LD-30 LD-30 NIXON-1 (minor) NIXON-2 (major) 74.03 73.98 74.08 74.10 74.60 74.54 13.93 14.26 14.03 14.03 13.77 13.81 2.05 2.03 2.11 2.11 2.06 2.10 0.32 0.27 0.30 0.30 0.31 0.30 0.05 0.05 0.06 0,06 0.05 0.06 1.43 1.41 1.38 1.38 1.33 1.34 0.28 0.31 0.32 0.32 0.32 0.30 4.30 4.53 4.51 4.53 4.46 4.46 3.61 3.17 3.21 3.17 3.11 3.10 100.00 100.01 100.00 100.00 100.01 100.01 1.0000 0.9633 0.9619 0.9593 0.9553 0.9534 Tephra layer in southeastern Mono Basin (1). Closest matches are with the tephra layers in Willow Wash section (3, 11; 5.9-5.4 Ma; and 5; <2.57, >2.22 Ma) and various tephra layers (7, 8, 12, 14) of indeterminate age, probably late Miocene and early Pliocene age. 0.58 0,04 0.07 0.43 0.09 0.56 0.04 0.09 0.42 0.09 0.55 0.06 0.08 0.41 0.06 0.57 0.04 0.08 0.56 0.07 0.58 0.05 0.08 0.39 0.08 0.53 0.06 0.08 0.39 0.09 0.54 0.O5 0.07 0.43 0.08 0.58 0.05 0.06 0.40 0.11 Tephra layer in southeastern Mono Basin (1). Closest matches are with a tephra layer in northeastern Fish Lake Val- ley (2; <2.57, >2.22 Ma by correlation), a tephra layer at Mud Spring Canyon (3), the tuff near Mount Jackson (5), and the tephra layer in Yellowjacket Canyon, California (7; <2.57, >2.22 Ma, Sarna-Wojcicki et al., In review). 1 3 5 7 8 11 12 14 EL-53-ML (high K) FLV-138-WW FLV-29-WW 86LP39 OT-2 FLV-28-WW D3SW-1 BE-62 76.86 76.77 76.85 76.57 78.22 77.46 77.93 77.36 12.60 12.96 12.65 12.90 12.23 12.55 11.76 12.59 2.47 2.65 2.20 2.47 2.33 2.17 2.80 1.81 6.85 6.42 7.13 6.75 6.05 6.67 6.33 7.04 99.99 100.00 99.99 100.01 100.01 100.00 99.99 100.00 1.0000 0.9638 0.9582 0.9515 0.9478 0.9436 0.9428 0.9381 1 2 3 5 7 EL-53-ML#l FLV-63-CS EL-36-MS MJ-1B YJC-4-87 77.27 76.87 77.31 76.90 77.17 12.69 12.94 12.71 13.00 12.92 0.57 0.58 0.59 0.59 0.56 0.04 0.04 0.05 0,04 0.04 0.06 0.09 0.06 0.08 0.10 0.45 0.43 0.48 0.49 0.42 0,08 0.06 0.07 0,07 0.07 3.63 3.77 3.53 3.62 3.45 5.22 5.23 5.20 5.20 5.28 100.01 100.01 100.00 99.99 100.01 1.0000 0.9791 0.9784 0.9749 0.9726 Tephra layer in southeastern Mono Basin (1). Closest matches are with the adjacent tephra layer (2; 2.3 Ma, R. Fleck, in litt. to A.S.-W., 1997), with the tephra layer at Teels Lake, Nevada (5), and with the tephra layer south of Rhodes Salt Marsh (7). 1 EL-54B-ML 77.53 12.64 0.53 0.03 0.07 0.41 0.07 3.82 4.91 100.01 1.0000 2 EL-54A-ML 77.38 12.76 0.55 0.04 0.07 0.40 0.08 3.83 4.91 100.02 0.9875 5 TLN-3 77.20 12.87 0.54 0.04 0.09 0.43 0.09 4.03 4.70 99.99 0.9697 7 EL-2-RM 77.54 12.72 0.55 0.03 0.06 0.44 0,06 3.94 4.65 99.99 0.9676 Tephra layer in southeastern Mono Basin (1), dated 2.3 Ma (R. Fleck, in litt. to A.S.-W., 1997). Closest matches are with the adjacent tephra layer (2) and with the tephra layers south of Rhodes Salt Marsh (6) and at Teels Lake (9). 1 EL-54A-ML 77.38 12.76 0.55 0.04 0,07 0.40 0.08 3.83 4.91 100.02 1.0000 2 EL-54B-ML 77.53 12.64 0.53 0.03 0.07 0.41 0.07 3.82 4.91 100.01 0.9875 6 EL-2-RM 77.54 12.72 0.55 0.03 0.06 0.44 0.06 3.94 4.65 99.99 0.9705 9 TLN-3 77.20 12.87 0.54 0.04 0.09 0.43 0,09 4.03 4.70 99.99 0.9681 Tephra layer in alluvial deposits at McGee Wash (1). Closest matches are with the ash flows in the Silver Peak Range (5, 8, 20) in the McAfee Creek area (tuff of Cottonwood Springs, 6.1 Ma; R. Fleck, in litt. to A.S.-W., 1997) and in the Volcanic Hills (22), and with a tephra layer in the Cave Springs Wash area (14; 1.2-0.8 Ma, upper tuffs of Glass Mountain), all in the Fish Lake Valley area of California and Nevada. 1 5 8 14 20 22 1 4 6 10 16 EL-57-WD FLV-SP-1 FLV-SP-2 FLV-55-CS FLV-SP-3 FLV-VH-5B 76.73 76.87 76.79 77.07 77.18 77.29 12.97 12.99 13.07 12.89 12.76 12.69 0.69 0.71 0.71 0.73 0.73 0.70 0.05 0.04 0.05 0.03 0.07 0.05 0.07 0.08 0.07 0.07 0.07 0.07 0.43 0.41 0.44 0.41 0.44 0.43 0.09 0-10 0.10 0.07 0.09 0.10 3.65 3.67 3.73 3.66 3.71 3.85 5.33 5.13 5.02 5.07 4.95 4.82 Tephra layer at McGee Wash (1). Closest matches are with the younger tephra layers of Glass Mountain (4, 6; 0.8 Ma) and similar tephra layers in Fish Lake Valley, Nevada (16), and with the Bishop ash bed (10; 0.76 Ma) EL-58-WD(l) BT-14 BT-IC 63CJ-26 (2) FLV-162-CS 77.65 77.41 77.84 77.93 77.36 12.76 12.77 12.73 12.45 12.89 0.78 0.77 0.77 0.72 0.77 0.04 0.04 0.03 0.03 0.03 0.05 0.04 0.05 0.05 0.07 0.45 0.45 0.42 0.45 0.46 0.07 0.07 0.06 0.06 0.07 3.52 3.53 3.52 3.53 3.40 4.68 4.92 4.57 4.77 4.95 100.01 100.00 99.98 100.00 100.00 100.00 1.2- 100.00 100.00 99.99 99.99 100.00 1.0000 0.9798 0.9768 0.9728 0.9688 0.9682 1.0000 0.9886 0.9820 0.9789 0.9772 68 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 3.?Continued. Output no. Field sample no. SiO2 A12O3 Fe2O3 MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer at McGee Wash (1). Closest matches are with the younger tephra layers of Glass Mountain (3; 1.2-0.8 Ma) and with the Bishop ash bed (4, 6, 7, 11, 12; 0.76 Ma). 1 3 4 6 7 11 12 EL-58-WD (2) BT-2 FLV-69-PA FLV-4-WP FLV-S2 (also FLV-148A-CS) FRIANT 5A BT-8 (2) (bulk) 77.99 77.64 78.04 77.93 77.81 77.73 77.75 12.31 12.67 12.31 12.49 12.55 12.41 12.48 0.74 0.74 0.74 0.74 0.72 0.75 0.73 0.03 0.03 0.02 0.05 0.03 0,05 0.03 0.03 0.03 0.06 0.04 0.03 0.04 0.0S 0.44 0.43 0.44 0.44 0.43 0.46 0.44 0.07 0.07 0.05 0.06 0.07 0.09 0.05 3.65 3.66 3.77 3.73 3.66 3.66 3.82 4.73 4.73 4.55 4.53 4.70 4.81 4.65 99.99 100.00 99.98 100.01 100.00 100.00 100.00 1.0000 0.9903 0.9882 0.9868 0.9866 0.9854 0.9847 Tephra layer at Drumm Summit in Fairview Valley (1). Closest matches are with the upper Miocene to lower Pliocene tephra layers (2, 4, 6) and with the Lava Creek B ash bed, except for a high-iron mode found in the Bur- meister core (5; Williams, 1993; correlated age, 0.65 Ma). Note: 7-element comparison, includes Ti. 1 2 4 5 6 EL-59-FV (2) VW-CIMA (2) EL-59-FV BUR-281.9 (high Fe) VW-CIMA-1 76.52 76.59 76.47 76.59 76.79 11.99 11.91 12.25 12.10 11.87 1.88 1.88 1.92 1.98 1.84 0.04 0.0(5 0.04 0.02 0.06 0.03 0.04 0.02 0.05 0.04 0.58 0.54 0.58 0.59 0.55 0.17 0.17 0.19 0.16 0.19 2.73 2.70 2.63 2.79 2.69 6.06 6.12 5.91 5.72 5.98 100.00 100.01 100.01 100.00 100.01 1.0000 0.9861 0.9701 0.9694 0.9686 Tephra layer at Drumm Summit, Fairview Valley (1). Replicate of EL-59-FV, above; same comments apply. Note: 7-element comparison, includes Ti. 1 EL-59-FV 76.47 2 VW-CIMA-1 76.79 5 EL-59-FV (2) 76.52 6 EL-66-FV 76.28 7 083089-A 76.95 8 08309-A (2) 77.05 Tephra layers in Truckee River canyon (1,2). Samples 1 and 2 are indistinguishable at analytical precision levels. Similar to Ishi tuff (4, 10; ~2.5 Ma) and to the tephra layer at 1-80 at 320 mile marker (14; J.D. Davis, unpub. data); age unknown. Note: 8-element comparison, excludes Mn. 12.25 11.87 11.99 12.22 12.15 12.20 .92 1.84 .88 .86 .84 .84 0.04 0.06 0.04 0.04 0.05 0.05 0.02 0.04 0.03 0.02 0.03 0.03 0.58 0.55 0.58 0.60 0.59 0.62 0.19 0.19 0.17 0.18 0.21 0.19 2.63 2.69 2.73 2.94 2.60 2.46 5.91 5.98 6.06 5.87 5.58 5.55 100.01 100.01 100.00 100.01 100.00 99.99 1.0000 0.9768 0.9701 0.9665 0.9664 0.9652 1 2 4 10 14 EL-60-TC EL-61-TC RS-2 (3) RS-2 (2) GS-85 77.35 77.34 77.97 77.51 77.68 13.29 13.34 12.48 12.82 12.63 0.83 0.81 0.85 0.81 0.83 0.15 0.14 0.14 0.16 0.15 0.08 0.07 0.05 0.02 0.04 0.69 0.69 0.78 0.74 0.69 0.13 0.11 0.13 0.15 0.16 2.47 2.49 2.81 3.39 3.71 5.02 5.01 4.78 4.40 4.11 100.01 100.00 99.99 100.00 100.00 1.0000 0.9677 0.9446 0.9100 0.9054 Tephra layer at Drumm Summit (1). Matches samples 2, 4, and 5 (this study); generally similar to upper Miocene and lower Pliocene tephra from Snake River Plain sources, Idaho (8, 9), and with the tephra layer from gravels in Las Vegas Wash in pre-Chemehuevi Formation (14). Note: 7-element comparison, includes Ti. 1 2 4 5 8 9 14 EL-66-FV EL-59-FV (2) EL-59-FV EL-24-M VW-CIMA-1 2-8-16J 083089-A (2) 76.28 76.52 76.47 75.58 76.79 76.32 77.05 12.22 11.99 12.25 12.56 11.87 12.28 12.20 .86 .88 .92 .77 .84 .82 .84 0.04 0.04 0.04 0.06 0.06 0.06 0.05 0.02 0.03 0.02 0.03 0.04 0.03 0.03 0.60 0.58 0.58 0.59 0.55 0.62 0.62 0.18 0.17 0.19 0.17 0.19 0.21 0.19 2.94 2.73 2.63 3.17 2.69 2.71 2.46 5.87 6.06 5.91 6.09 5.98 5.94 5.55 100.01 100.00 100.01 100.02 100.01 99.99 99.99 1.0000 0.9680 0.9665 0.9621 0.9592 0.9583 0.9536 Tephra layer at Lone Mountain, Kobeh Valley (1). Matches with the Rye Patch Dam ash bed at Tulelake, California (3), with other samples of this ash bed at Kobeh Valley (4, 10), and with samples at the type locality of this ash bed at Rye Patch Dam, Nevada (12, 14). Similar to Desert Springs tuff (ash-flow) in the Bend, Oregon, area (15), which is probably a near-source site of the Rye Patch Dam ash bed. Note: 6-element comparison, excludes alkalies and Mn. 0.37 0.06 0.37 0.08 0.36 0.07 0.36 0.09 0.34 0.06 0.39 0.09 0.39 0.09 1 3 4 10 12 14 15 EL-69-KV TULELAKE-321 EL-70B-KV AA-04-JOD JOD-18-7-78B(l) JOD-18-7-78A (1,2) DST-3 71.76 70.85 71.81 71.37 71.25 71.34 71.31 14.86 14.80 14.79 14.34 14.82 14.37 14.58 3.32 3.35 3.33 3.35 3.22 3.37 3.19 1.41 1.46 1.39 1.44 1.39 1.45 1.48 0.43 0.43 0.44 0.42 0.43 0.44 0.44 4,35 5.16 4.34 5.01 5.00 5.08 4.40 3.44 3.51 3.46 3.60 3.49 3.47 4.12 100.00 100.01 99.99 99.98 100.00 100.00 100.00 1.0000 0.9900 0.9879 0.9799 0.9775 0.9741 0.9691 NUMBER 33 69 TABLE 3.?Continued. Output no. Field sample no. SiO, Al,02^3 Fe9 MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer at Lone Mountain, Kobeh Valley (1); low-Fe fraction. Similar to the tephra layers at Pringle Falls, Ore- gon, about -180-220 Ka (Herrero-Berver et al., 1994) (2, 3, 6, 7, 8), to the Summer Lake bed NN, Oregon (4) (Davis, 1985), and to the Columbia Canal ash-flow tuff near Bend, Oregon (10). Note: 8-element comparison, ex- cludes Mn. 1 2 3 4 6 7 8 10 EL-72-KV#l (lowFe) PF-88-0 (mafic) PF-88-0 (silicic-ct) DR-70 PF-88-P PF-88-Q (silicic) PF-88-0 (silicic-b) BPT-CC 69.86 69.21 68.64 69.70 69.95 68.38 69.36 69.34 15.46 15.37 15.68 15.07 15.17 15.51 15.27 15.31 3.68 3.81 3.85 3.79 3.87 3.85 3.91 3.70 0.79 0.73 0.80 0.67 0.69 0.79 0.74 0.72 0.10 0.12 0.12 0.09 0.11 0.12 0.14 0.13 2.14 2.14 2.27 2.16 2.12 2.23 2.11 2.12 0.70 0.69 0.71 0.68 0.61 0.73 0.65 0.67 4.71 5.40 5.56 5.12 5.00 5.94 5.44 5.70 2.55 2.53 2.38 2.71 2.47 2.37 2.38 2.30 99.99 100.00 100.01 99.99 99.99 99.92 100.00 99.99 1.0000 0.9656 0.9526 0.9518 0.9471 0.9465 0.9465 0.9456 Tephra layer at Lone Mountain, Kobeh Valley (1); high-Fe fraction. Similar to the tephra layer from Honey Lake, California, stratigraphically close above Huckleberry Ridge ash bed at Honey Lake, California (2); other similar samples are the upper Miocene tephra from Balcolm Canyon, southern California (3), and the upper Pleistocene tephra layer from the Tulelake area, California (4). Note: 8-element comparison, excludes Mn. 1 EL-72-KV #2 (high Fe) 61.09 16.78 6.84 2.28 0.14 5.04 1.28 5.02 2 CDGD-1 60.70 16.14 7.71 2.27 0.15 4.92 1.36 5.08 3 MOD-13 61.37 16.03 6.81 2.47 0.15 5.29 1.21 4.85 4 AS W-61085-14 (mafic) 63.69 14.75 7.13 1.84 0.10 4.86 1.26 4.87 Tephra layer in roadcut section, Lone Tree Mine (1). Most similar to the younger Glass Mountain tephra layers, in- cluding Bailey ash bed, Ventura, California (4), and tephra layer beneath Bishop ash bed, northern Fish Lake Valley, Nevada (5). Also similar to the Bishop ash bed at Friant, California (6-8). 1.53 1.68 1.82 1.50 100.00 100.01 100.00 100.00 1.0000 0.9568 0.9468 0.9380 1 4 5 6 7 8 EL-76A-LT PICO-5 (2) FLV-160-CS FRIANT 4A FRIANT-1 FRIANT-2 77.18 77.24 77.40 77.23 77.54 77.53 12.97 12.73 12.82 12.90 12.73 12.86 0.69 0.71 0.73 0.71 0.73 0.68 0,03 0.03 0.01 0.03 0.03 0.03 0.05 0.03 0.06 0,03 0.03 0.03 0.49 0.49 0.50 0.51 0.48 0.47 0.07 0.06 0.07 0.05 0.06 0.06 3.59 3.56 3.50 3.70 3.56 3.40 4.92 5.15 4.91 4.83 4.83 4.94 99.99 100.00 100.00 99.99 99.99 100.00 1.0000 0.9833 0.9806 0.9798 0.9792 0.9791 Tephra layer in roadcut section, Lone Tree Mine (1). Most similar to the younger Glass Mountain tephra layer at Lake Tecopa, California (3), and to the Bishop ash bed at the southern end of volcanic tableland north of Bishop, California (11). 1 EL-78-LT 77.29 12.90 0.71 0.03 0.05 0.48 0.05 3.82 4.66 99.99 3 TECO-26 77.61 12.73 0.72 0.03 0,03 0.48 0.06 3.70 4.63 99.99 11 BT-8(2) 77.48 12.68 0.71 0.03 0.02 0.45 0,06 3.85 4.72 100.00 Tephra layer at Lone Mountain, Kobeh Valley (1). Most similar to the Huckleberry Ridge ash bed near Ventura, Cal- ifornia (2, 4), and in Meade County, Kansas (5, 6); somewhat less similar to the Lava Creek B ash bed from Butte Valley, California (7), Lake Tecopa, California (8), and Onion Creek, Utah (9). Note: 7-element comparison, in- cludes Ti. 0.02 0.04 0.60 0.03 0.03 0.60 0.Q2 0.05 0.60 0.02 0.03 0.58 0.02 0.04 0.58 0.03 0.03 0.55 0.02 0,05 0.53 0.02 0.03 0.54 Tephra layer on southwest side of Lone Mountain, Kobeh Valley (1). Closest matches are with the tephra layers of Cowan Pumice Mine, Blind Springs Valley, California (4, 8), from northeastern Fish Lake Valley, Nevada (3, 6), and from stratigraphically beneath the Huckleberry Ridge ash bed at Ventura, California (7). 1 2 4 5 6 7 8 9 EL-80-KV PICO 158 ASW-3-23-75-1 68W98 68W98 #4078 RM880904 67W104 76.73 76.71 76.62 76.65 76.99 76.64 76.63 77.00 12.32 12.25 12.13 12.53 12.30 12.22 12.19 12.23 1.67 1.71 1.83 1.69 1.66 1.65 1.65 1.51 1.0000 0.9885 0.9829 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 3.54 3.62 3.63 3.32 3.22 3.51 3.55 3.58 4.98 4.95 5.02 5.08 5.09 5.27 5.28 4.99 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 1.0000 0.9918 0.9804 0.9793 0.9777 0.9760 0.9714 0.9686 1 3 4 6 7 8 EL-81-KV FLV-13-CS TTC-19 (2) FLV-151-CS P-169Q TTC-9A 77.63 77.48 77.54 77.34 77.59 77.51 12.59 12.82 12.57 12.70 12.55 12.83 0.60 0.60 0.59 0.60 0.58 0.59 0.04 0.04 0.05 0.04 0.04 0.05 0.08 0.08 0.07 0.09 0.10 0.10 0.33 0.33 0.33 0.34 0.33 0.34 o.oa0.04 0.06 0.05 0,06 0.07 3.95 3.85 3.91 4.09 3.87 3.87 4.69 4.75 4.89 4.74 4.87 4.65 99.99 99.99 100.01 99.99 99.99 100.01 1.0000 0.9904 0.9883 0.9856 0.9843 0.9841 70 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 3.?Continued. Output no. 1 7 9 11 15 Field sample no. SiO, A12O3 Fe2O3 MgO MnO CaO TiO2 Na2O K2O Total (R) Similarity coefficient Tephra layer in core at Lone Tree Mine (1). Closest matches are with the Lava Creek B ash bed near Oreana, Nevada (2; J.O. Davis, Desert Research Institute, Nevada), the Huckleberry Ridge ash bed in Meade County, Kansas (3, 5, 7; Izett and Wilcox, 1983), and with the Lava Creek B ash bed, Lake Tecopa, California (4), and Ventura, California (8). 0.02 0.03 0.56 0.03 0.03 0.56 0.02 0.04 0.58 0.03 0.04 0.56 0.02 0.03 0.58 0.02 0.05 0.58 0.02 0,00 0.56 Tephra layer near U.S. Highway 50 in Kobeh Valley, Nevada (1). Closest matches are with the Rye Patch Dam ash bed at Rye Patch Dam, Nevada (4, 7), at a nearby site in Kobeh Valley (15), and in a core from Tulelake, California (5); and with Desert Springs tuff, Bend, Oregon (13). Note: 8-element comparison, excludes Mn. 1 2 3 4 5 7 8 EL-92-LT LM-24 68W98 T89-5/6/9 #1 68W98 68W98 PICO-133C 76.89 76.70 76.99 76.88 76.65 76.67 75.71 12.31 12.62 12.30 12.16 12.53 12.28 12.37 1.66 1.67 1.66 1.67 1.69 1.71 1.59 0.12 0.08 0.10 0.13 0.10 0.12 0.11 3.30 3.30 3.22 3.49 3.32 3.35 3.40 5.12 5.01 5.09 5.05 5.08 5.22 5.25 100.01 100.00 100.00 100.01 100.00 100.00 99.01 1.0000 0.9909 0.9889 0.9856 0.9855 0.9828 0.9806 1 4 5 7 13 15 AA-04 JOD LD-61 TULELAKE-321 JOD-18-7-78A (1,2) DST-1 EL-69-KV 71.37 71.07 70.85 71.34 72.69 71.76 14.34 14.70 14.80 14.37 13.39 14.86 3.35 3.32 3.35 3.37 3.21 3.32 0.36 0.37 0.37 0.39 0.38 0.37 0.09 0.09 0.08 0.09 0.07 0.06 1.44 1.45 .46 1.45 1.47 1.41 0.42 0.45 0.43 0.44 0.43 0.43 5.01 4.99 5.16 5.08 4.75 4.35 3.60 3.56 3.51 3.47 3.61 3.44 99.98 100.00 100.01 100.00 100.00 100.00 1.0000 0.9808 0.9804 0.9765 0.9654 0.9629 1 3 7 12 16 D77-2D 63CJ-26(1) YJC-1-87 PICO-141 BT-8 (2) 77.39 77.37 77.28 77.29 77.48 3.72 3.60 3.74 3.68 3.85 4.89 4.99 4.75 4.91 4.72 99.99 100.00 100.01 99.99 100.00 1.0000 0.9908 0.9872 0.9839 0.9836 Tephra layer at depth of ~175 m in a core from Clayton Valley, Nevada (1). Best matches are with the reworked Bishop ash at Friant, California (3), the airfall Bishop ash bed at Yellowjacket Canyon near Benton, California (7), the Bishop ash bed in the Ventura basin, California (12), and the airfall Bishop ash bed at Insulating Aggregates Quarry north of Bishop, California (16). 12.66 0.73 0.03 0.05 0.45 0.07 12.69 0.73 0.04 0.05 0.45 0.08 12.91 0.73 0.04 0.05 0.44 0.07 12.84 0.73 0.03 0.02 0.42 0.07 12.68 0.71 0.03 0,02 0.45 0.06 Tephra layer in large pit 9.5 m below surface at Columbus Marsh, Nevada (1). Best matches are with the airfall Bishop ash bed near Westgard Pass, California (7), the Glass Mountain D tephra at Chalk Bluffs north of Bishop, California, -0.9 Ma (9), the airfall Bishop ash bed at Friant, California (11), and the airfall Bishop ash bed at Insu- lating Aggregates Quarry north of Bishop, California (15). 29068501 78.06 12.27 0.76 0.04 0,04 0.44 0.07 FLV-4-WP 77.93 12.49 0.74 0.05 0.04 0.44 0.06 BT-2 77.64 12.67 0.74 0.03 0.03 0.43 0.07 FRIANT 5A 77.73 12.41 0.75 0.05 0.04 0.46 0.09 BT-8 (2) bulk 77.75 12.48 0.73 0.03 0.05 0.44 0.05 3.64 3.73 3.66 3.66 3.82 4.67 4.53 4.73 4.81 4.65 99.99 100.01 100.00 100.00 100.00 1.0000 0.9834 0.9826 0.9822 0.9814 about 1330 m; two higher, older barriers, units Q16T and Q15T, representing different lake cycles that reached altitudes be- tween about 1330-1350 m and 1350-1370 m, respectively; and at least one old bar, unit Q14T, that reached a minimum alti- tude of about 1402 m. An excellent small outcrop about 4 km north of Thome Bar (Figures 2, 3) exposes four sequences of lacustrine gravel and sand that overlie bedrock (Figures 3, 4B), all lying well above the Sehoo shoreline. The four sequences are separated by pa- leosols and alluvium. The second youngest lacustrine unit here, Q13T, contains a tephra (EL-21-WL, Table 2) that is chemically correlative with the Bishop-Glass Mountain family of tephra erupted from Long Valley, California. Paleomagnetic measure- ments on the silty sand layers containing the tephra are not de- finitive in their polarity; thus, the tephra may be either the Bishop ash bed dated from 0.76 Ma (normal polarity) or one of the tuffs of Glass Mountain dated from 1.2 to 0.8 Ma (reversed polarity; Saraa-Wojcicki et al., 1991). Small fragments offish bones were found in the ashy layers (G.R. Smith, University of Michigan, in litt. to M.C.R., 1998). The youngest unit at this site, Q14T, consists of about 38 m of tufa-cemented, well-bed- ded pebble and cobble shore gravel that is commonly steeply bedded (backset beds) and that rises to an altitude of about 1393 m. Beach pebbles continue upward as lag on bedrock to an altitude of about 1398 m. Because these outcrops and the outcrops of the oldest morphologic unit at the Thorne Bar site are similar in cementation and preservation and because they rise to about the same altitude, we infer that they are equivalent. The Bishop ash bed, identified by chemical analyses con- ducted by Davis (1978) and by paleomagnetic data from the NUMBER 33 71 vehicles FIGURE 4.?Photographs of Thome Bar area, A, View south toward Thorne Bar (Figure 3); bedrock of the Gillis Range is on left. Small white dots are vehicles parked on pre-late Pleistocene berm at about 1353 m, 20 m above the Sehoo shoreline. B, View of outcrop 4 km north of Thorne Bar (Figure 3); people are for scale. East-dipping gravels in upper part are tufa-cemented backset beds of unit Q14T. Line of coarse boulders at top of trench are shoreface clasts at base of unit Q12T. Trench exposes paleosols, fan deposits, and lake unit Ql 1T just above basalt outcrop. present study (data not shown), is present in beach gravel be- neath Sehoo-aged deposits about 4 km south of Schurz, Ne- vada (Figure 2). This outcrop indicates that a lake of uncertain size existed in the Walker Lake subbasin at about 0.76 Ma. We tentatively assign the tephra-bearing lake deposits of unit Q13T north of Thorne Bar to this period, but we recognize that they may be as old as 1 My. Four of the six paleosols formed in units Q11T and Q12T and the intervening fan deposits (Fig- ures 3, 4B) have moderately developed to well-developed argillic horizons that are 20-90 cm thick. This degree of de- velopment indicates lengthy (>50 thousand years (Ky)) peri- ods of nondeposition (Reheis et al., 1995) and thus a rela- tively long stratigraphic record prior to deposition of unit Q13T. 72 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES The stratigraphy in the Thorne Bar area is shown schemati- cally in Figure 5. The oldest and highest shoreline in this area, as revealed by the outcrop north of Thorne Bar, is underlain by deposits of four different lakes: two lakes (units Q11T and Q12T) predated deposition of the tephra layer; one lake (QBT) existed during deposition of the tephra layer; and one lake (Q14T) was younger than the tephra layer. Units Q11T, Q12T, and Q13T reached minimum altitudes of about 1355 m, whereas Q14T rose to a minimum altitude of about 1400 m. In addition, barriers with preserved morphology at the Thorne Bar site indicate that three younger lakes (units Q15j, Q16j, and Q17j) successively reached altitudes of about 1365 m, 1350 m, and 1330 m (the Sehoo highstand). 20-Mile Beach: At least two localities preserve beach grav- els along the Wassuk Range front (Figure 2). These localities are at much higher altitudes than most other sites in the Walker Lake subbasin. The first site is north of 20-Mile Beach and has two levels of tufa-cemented beach gravel. The lower level is a moderately preserved wave-cut bench overlain by 1-2 m of tufa-cemented beach rock; it extends up to an altitude of about 1400 m (Figure 6A). The upper level is an eroded remnant of beach rock with foreset beds overlying sheared granitic rocks; the highest outcrop of beach rock is at 1452 m. The second site is about 3 km to the south of the first. It has four exposed beds of beach gravel interbedded with fan gravel above the Sehoo shoreline; the highest bed caps a ridge at 1475 m (Figure 6B). The super-elevated beach gravels at these two sites are inter- preted to reflect absolute uplift of the footwall block of the Was- suk fault. Demsey (1987) showed that this active normal fault had about 7 m of displacement during the Holocene, of which about 20%-25% was the result of uplift of the Wassuk Range. McGee Wash: The study site at McGee Wash (Figure 2) ex- tends from the Walker River just below Weber Reservoir east- ward to U.S. Highway 95 (Figure 7). Exposures indicate that sediments there are faulted and tilted (Morrison and Davis, 1984); only the larger faults are shown in Figure 7. Dips are generally greatest on the eastern side near the highway and de- crease toward the west. Younger units are progressively less de- 22 O 1 s Q- Lone Tree Hill Thorne Bar area McGee Wash area Campbell Valley Sunshine Amphitheater QI3L (beach gravel as high as -1415 m) (inset beach gravel as high as -1350 m) "\_/-\_/-\_^ QI5T (inset beach gravel as high as -1365 m) ~\-/ QI4T (beach gravel as high as -1402 m) (inset beach gravel as high as -1365 m) (beach gravel as high as -1402 m) Bishop ash N?XXXXXBishop QI3T ash? "5T R XXXXX 5 R Glass Mtn. Q|-|o U. --?- L o R ash Huckleberry ?_. Ridgeash(?) u" Tg XXXXX QTI XXXXX Tt QI3C (inset beach gravel as high as -1356 m, possibly to 1378 m) '\_/-\_/-\_^' QI2C (beach gravel as high as -1378 m, probably to 1390 m) R R XXXXX R Glass Mtn. QI1C ash QI4S (inset beach gravel as high as -1378 m) QI3S (beach gravel as high as -1414 m) (beach gravel QI2S crops out to-1396 m) cm QTI R (<2.6 Ma) Lake-unit names Younger lake deposits of Thorne Older lake deposits of Thorne Lake deposits of McGee Lake deposits of Lone Tree Lake deposits of Campbell Lake deposits of Sunshine 1111 /\x\ XXX NorR EXPLANATION Buried soil Unconformity Tephra layer and correlation Paleomagnetic polarity Correlation between units FIGURE 5.?Schematic strati graphic correlations in the Walker Lake subbasin. Units and symbols are the same as in Fig- ures 3, 6, 8, and 11. Deposits of Sehoo-aged lake not shown (Q17T in Thorne Bar area, grouped with fan deposits as unit Qlfy in other site maps). Wavy lines represent unconformities; attached vertical lines are paleosols. N and R indicate normal and reversed polarities, respectively, of sampled horizons, not shown to scale. NUMBER 33 73 "Jnr B beach ^ ?/-- ? beach fa\J_. ??^C? "????"?? Tan besfch Sehoo berm FIGURE 6.?Photographs of very high beach gravels on the footwall block of the Wassuk fault (see Figure 2 and Table 1). A, View south from gravel- and tufa-covered bench at 1399 m to sloping, eroded remnants of tufa- cemented beach rock on skyline as high as 1452 m. Bush on right is about 1 m high. B, View west of beach gravel interbedded with fan gravel above Sehoo shoreline at about 1330 m. Beach gravels crop out as high as 1475 m at top of hill. formed. The overall structure appears to be that of a block or horst tilted progressively westward with time. This block was periodically partly or entirely covered by lakes that prograded eastward and by alluvial and eolian deposits that prograded westward. The stratigraphic units at this site are shown sche- matically in Figure 5. The oldest units are exposed near the highway and consist of a conformable sequence (Figures 7, 8) of alluvial deposits (unit Tt) overlain by about 50 m of well-bedded lacustrine mudstone, siltstone, and sandstone capped by beach gravel (unit QT1), which in turn is overlain by a very thick (unmeasured, but more than 50 m) sequence of alluvium and eolian deposits (unit QTt) containing numerous carbonate-enriched paleosols. All three of these units contain tephra layers, but only one, EL-19-WD in alluvial unit Tt, has been identified; it is tentatively correlated with an upper Miocene to lower Pliocene tephra erupted from 74 118?51r 118?50' SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 118?49' 118?48' 0 0 1 MILE CONTOUR INTERVAL: 40 FEET EXPLANATION Qes Eolian sand (Holocene) Qfy Qlfy Qce Qfo QI3? Fan deposits (Holocene and late Pleistocene) Lacustrine and fan deposits, undifferen- tiated (Holocene and late Pleistocene) Colluvial and eolian deposits (Holocene and late Pleistocene) Deposits of early late(?) Pleistocene lake Fan deposits (middle Pleistocene) Deposits of middle Pleistocene lake; pattern shows beach gravel Colluvial deposits (middle Pleistocene) Deposits of middle? and early Pleistocene lakes; pattern shows possibly correlative beach gravel Qlo QTt EQTE Tvs," 58-WD Lacustrine deposits, undifferentiated (middle and early Pleistocene) Alluvial, eolian, and colluvial deposits (early Pleistocene and Pliocene) Deposits of early Pleistocene? or Pliocene lake Alluvial deposits (Pliocene and Miocene) Volcanic and sedimentary rocks (Pliocene and Miocene) Fault; bar and ball on downthrown side; dashed where inferred, dotted where concealed Shoreline scarp Tephra locality and sample number of sites not within measured sections (Fig. 8) Line of measured section (see Fig. 8) Strike and dip of beds FIGURE 7.?Geologic map of area around McGee Wash (see Figure 2, Table 1), east of Weber Reservoir. NUMBER 33 75 0_X LU CD O CO T3 2 CD T3 "5 WCD to ?e-o CD C ?P co COCO T3i j2 2k ? CD ? %> CO CO CD "O o V- CD sli 15115o o ? 82s Cn ~ -ao CD O "DQ--D - CDCD E r^ CO -p E S 0 0 C ) O 0o o c i 0 On n ( >?oOc ^ 3 ? Q CO co -g is CO Q_ 9- co .CD CO LU _0) Q. E CO CO o "CD ^" O) CO1" ? CO CO Q_ CD CDo inI COCD E CO Q CD 1?s toCO LU C o"T3 CD C/D Green, smectite- rich, well bedded ,5 -5 -i 211818 ?1 11 "D^ II a?i U ?r- c | | ij (S M 3 " a .2 o cO .5 E 5 2 76 the Snake River Plain (Table 3). In the south-central part of the map area (Figure 7), erosion between the time of deposition of alluvial units Tt and QTt has removed lacustrine unit QT1. Unit QTt apparently extends westward to underlie younger lacus- trine units around McGee Wash. Three tephra layers crop out in unit QTt in McGee Wash. Sample EL-57-WD may be of Pliocene age (Table 3); two paleomagnetic samples (P3, Figure 8) indicate that EL-5-WD has normal polarity. Sample EL-6- WD has not been correlated with tephra of known age. Four pre-Sehoo lacustrine units separated by unconformities overlie unit QTt in McGee Wash. The oldest, Q11M, is a thin (1-2 m) bed of pebble gravel overlain by silty sand (Figure 8); this unit is preserved in only a few small outcrops. Unit Q12M is locally much thicker and crops out around the head of McGee Wash (Figure 7) to a maximum altitude of about 1390 m (4560 ft). This unit consists of a basal bed of sand and pebble gravel that grades up into well-bedded silt and sand containing at least one tephra layer. Two samples (EL-20-WD, EL-58-WD; Tables 2, 3) have been identified as Bishop- or Glass Mountain-like tephra layers. Paleomagnetic data show a sequence of stable re- versed magnetization in sediments below the tephra bed, stable normal magnetization in silt overlying the tephra bed, and sta- ble reversed magnetization in overlying mudstone (Figures 8, 9). A couple of normal-magnetization samples near the tephra bed lie in between samples of reversed magnetization. We have not assessed magnetic mineralogy to determine whether these normal directions represent some type of remagnetization dur- ing a later normal-polarity interval. Nevertheless, the pattern of reversed-normal-reversed magnetizations and ash chemistry in- dicate that the ash is a Glass Mountain tephra that erupted at about 1 Ma, just before the Jaramillo Normal Subchron (Sarna- Wojcicki et al., 1991). If a beach-pebble facies of unit Q12M had been deposited, it has mostly been removed by erosion, but remnants of beach gravel that are probably older than that of unit Q13M are locally preserved capping units QT1 and QTt near the highway at an altitude of about 1430 m (4700 ft; Figure 7). Unit Q13M unconformably overlies silt and sand of unit Q12M around the head of McGee Wash, and it consists mainly of loose sand and gravel (Figures 7, 8). A buried soil and collu- vium is locally preserved at the unconformity. Unit Q13M is poorly exposed, but in aerial photographs it has arcuate mor- phology resembling that of a shoreline berm. The berm rises to an altitude of 1402 m (4600 ft), the same altitude as the highest beach gravel at Thorne Bar (Figure 3). The youngest of the four pre-Sehoo lacustrine sequences, unit Q14M, is inset into all older units (Figure 7). It consists mostly of pebbly sand; in McGee Wash the pebbly sand over- lies a gray clay layer that Morrison and Davis (1984) assigned to the Eetza Alloformation. The isolated remnants of this unit lie at altitudes of 1350-1365 m and may represent faulted de- posits of one lake cycle or deposits of more than one lake cycle. Sunshine Amphitheater: Sunshine Amphitheater is a very large exposure of lacustrine deposits of five different ages (Figures 2, 10; stratigraphic summary in Figure 5). It is re- markable for the abundance of vertebrate fossils and the quality SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES A N (UP) W (Hor) S (down) B N (Up) 0.1 A/mW(Hor)-}??i? E q80 P45 S (down) N(Up) 5mT NRM W (Hor) 20 NRM H I hE 0.001 A/m P42 S (down) FIGURE 9.?Orthogonal demagnetization diagrams (method of Zijderveld, 1967) shown in stratigraphic order for specimens from unit Q12M, section 8b, McGee Wash. The closed symbols represent projections of the magnetic vector onto the vertical plane, whereas the open symbols represent projections of the magnetic vector onto the horizontal plane. Remanent magnetizations are given in amperes/meter (A/m). Demagnetization steps are in milliTesla (mT). A, Reversed magnetization in sample P49; B, normal magnetization in sample P45; c, reversed magnetization in sample P42, from 26 cm above tephra bed EL-20-WD. NUMBER 33 77 119"01'W 119?00'30" 119?00' 118?59'30" Qaf Qlfy Qfu EXPLANATION Alluvium of the Walker River and fan deposits, undifferentiated (Holocene) Lacustrine and fan deposits, undifferentiated (Holocene and late Pleistocene) Fan deposits, undiff. (Holocene and Pleistocene) |o ooj Beach gravel of late middle(?) l? o s i Pleistocene lake Qfo Old fan deposits (middle Pleistocene) 'Qfo+/ ?rrv 1 Contour interval 1.0 miles 40 feet Section 14 Qf3 Unit QI3S 12 m omitted Section 15 ^ P34-R Pale marke /// ^ P33-R? Unit QI2S /P19-R / Common silt beds 5 m omitted , Fossil bones P17-R? /Pale; marker silt bed Pale; locally interbedded with sand and pebbly sand -P16-N Abundant bones ~-3 Papery; fossil bones Intermixed old fan deposits and beach gravel of early middle(?) Pleistocene lake Deposits of two early Pleistocene lakes sepa- rated by unconformity (dotted line); pattern shows transgressive basaltic beach gravel Alluvial eolian, and colluvial deposits (early Pleistocene) Deposits of Pliocene or early Pleistocene lake Volcanic and sedimentary rocks, mostly basalt (Pliocene) Late Pleistocene shoreline Upper limit of beach gravel Fault; bar and ball on downthrown side Line of measured section Strike and dip of beds Fossil sample locality P11-N Abundant fossil bones lagged on surface Unconformity; slight dip change P10-Rcomp. UnitQI1s Green -P13-R ? Section steps north Basaltic gravel, Section 12 Top of bench at -4520 feet break in slope Unit QI4S P31-R? Unit QTI e? P30-R i? P29-R E? P8-? Pale, massive =? P7-R W B ? 0 Tan grading up to brown paper shale 1?P6-R Brown paper shale 1? P5-R EXPLANATION Wi Mudstone ~3 Interbedded mudstone =3 and sandstone =3 Siltstone and sandy ^ siltstone Sandstone and silty sandstone, locally conglomerate beds Sandy pebble conglomerate, well rounded clasts (nearshore deposits) Sandstone and conglomerate, poorly to moderately sorted Fanglomerate, poorly sorted and bedded Steep dips (foreset beds) Buried soil Unconformity P9-R Paleomagnetic sample and polarity (N, normal; R, reversed) Unit QTt FIGURE 10.?Geologic map of Sunshine Amphitheater (Figure 2, Table 1) and selected measured sections. The portion of section 14 above the closely spaced contours (top of unit Q12S) was not measured and is schematic. Question mark (?) in place of N or R indicates polarity uncertain. 78 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES of their preservation; thus far, identified remains include those of giant sloth, bird, horse, camel, mammoth, and fish. The abundance of fossil remains at Sunshine Amphitheater may be the result of its location adjacent to the Walker River near the point where the river bends in a U-turn (Figure 2). Floods coming down the north-flowing reach of the river would have been forced to slow and drop part of their loads; the natural deposition site would have been on the outside of the bend. Small faults are exposed, but the units do not appear to be greatly offset, in contrast to the McGee Wash and Campbell Valley sites. The oldest non-bedrock unit, QT1, consists mainly of brown, finely laminated mudstone capped by pebbly sandstone; the mudstone has reversed magnetization (Figures 10, 11D). It is overlain by unit QTt, which has alluvial and eolian deposits containing carbonate-enriched paleosols. Unit QTt is conform- ably overlain by well-bedded pebbly sandstone, siltstone, and mudstone of unit Qlls, which also has reversed magnetization (Figure lie). A distinctive horse bone (Equus excelsus(!) Leidy, 1858) from unit QUS appears to constrain the age to less than 1.4 Ma. The jaw of a Lahontan cutthroat trout (Oncorhyn- chus clarki (Richardson, 1836)) and a vertebra of a sucker N(Up) 0.1 A/m-- W (Hor) P34 S (down) 5 mT NRM B (Up) W (Hor) 0.1 A/m-- S (down) N (Up) so W (Hor) 0.001 A/m 20 P10 0.02 A/m - 5 mT w (Hor) NRM S (down) S (down) FIGURE 11.?Orthogonal demagnetization diagrams (method of Zijderveld, 1967) shown in stratigraphic order for specimens from Sunshine Amphitheater sections. Symbols are the same as in Figure 9. A, Reversed magneti- zation in sample P34 in unit Q12S, section 15; B, normal magnetization in sample PI 1, lower part of unit Q12S, section 15; c, reversed magnetization component in sample P10 in unit Qlls, section 15 is indicated by the isola- tion of southward, negative-inclination (up) vector after the removal of a northward, positive-inclination (down) vector following demagnetization at and above 40 mT; D, reversed magnetization in sample P30 in unit QT1, sec- tion 12. NUMBER 33 79 {Catostomidae sp.; G.R. Smith, University of Michigan, in litt. to M.C.R., 1996) also were collected from this unit. This age range is consistent with the assignment of the reversely mag- netized sediments to the Matuyama Reversed Chron (2.6-0.8 Ma; magnetic-polarity ages of Cande and Kent, 1995). An un- conformity, locally marked by thin colluvium and a paleosol, separates units Qlls and Q12S (Figures 10, 12A). Many of the fossils are associated with this stratigraphic interval, including mammoth bones (Mammuthus sp., also less than 1.4 Ma) and the astragalus of a sloth identified as post-Blancan in age (<2.6 Ma; J. Honey, University of Colorado, pers. comm. to M.C.R., 1995). Unit Q12S extends to an altitude of 1396 m and consists of poorly indurated siltstone and sandstone that coarsen up- ward into beach gravel. These deposits also contain fossilized horse and bird bones. Paleomagnetic measurements indicate that this unit has reversed magnetization except for an interval near the base, which has normal magnetization (Figures 10, 11A,B). These data and the presence of a horse bone beneath the unconformity indicate that the normal polarity interval probably correlates with the Jaramillo Normal Subchron from about 1 Ma. Thus, the three oldest lacustrine units at Sunshine Amphitheater lie within the Matuyama Reversed Chron; unit Qlls was deposited between 1.4 and 1.0 Ma, and unit Q12S was deposited beginning at about 1 Ma (Figure 5). Unit Q12S at the top of the bluff (section 14, Figures 10, 12A) bears a well-developed argillic paleosol (Figure 12B) overlain by fan gravel that forms a sloping bench. Exploratory soil pits on this bench indicate that the fan gravel is capped by a paleo- sol and is buried in turn by beach sand and gravel of unit Q13S, another paleosol, and a younger fan gravel; unit Q13S rises to an altitude of about 1414 m. Unit Q14S consists of sandy beach gravel inset into all of the older units; it rises to an altitude of about 1378 m. Campbell Valley and Vicinity: The interpretation of out- crops in Campbell Valley (Figures 2, 13) is complicated by a network of northwest-striking strike-slip faults of the Walker Lane. Displacement along these faults has juxtaposed lake sed- iments of different ages, making it difficult to correlate depos- its of beach gravel that mark shorelines with the finer-grained sediment that contains age-related information. Three lacus- trine units are locally exposed (Figures 13, 14). The oldest, Qllc, consists mainly of mudstone that grades up into beach sand and gravel; four samples from the mudstone have re- versed magnetization. Deposits in another outcrop, herein as- signed to the same unit as the mudstone but possibly belonging to a younger lake cycle, consist of shoreface gravel that rapidly grades laterally into fine-grained beds lapping onto subjacent basalt. A tephra layer, sample EL-15-CA, drapes this facies change from shoreface into deep-water deposits (Table 2; Fig- ure 14A). The tephra, a sample from which has normal magne- tization, is correlated with the Bishop?Glass Mountain family of tephra (Table 3). The reversed magnetization of sediment just above and below the tephra layer, however, indicates that the tephra cannot be Bishop ash on the basis of enveloping re- versed magnetizations; thus, the tephra is probably a Glass Mountain tephra from about 1 Ma. Unit Qllc is unconformably overlain by beach gravel of unit Q12C; a paleosol and colluvium are locally preserved at the contact. Unit Q12C rises to at least 1378 m, and beach pebbles that probably represent this unit are present as lag on basalt up to an altitude of 1390 m. The peb- bles are found up to an abrupt change in slope that resembles a preserved shoreline angle cut on a small island, or stack (Figure 14B). Beach gravel of unit Q13C is apparently inset into the older units and forms a berm at an altitude of 1356 m. Another site in Campbell Valley, between Sunshine Amphi- theater and Weber Reservoir on the east side of the Walker River (Figure 2), records shorelines cut into two basalt hills. Subdued breaks in slope occur on the west sides of these hills, with the most prominent slope break occurring at an altitude of 1365 m. From a distance these slope breaks resemble flow con- tacts, but they are actually degraded shoreline angles. Rounded beach pebbles locally lie below the most prominent slope break, and they are identical to pebbles that abundantly mark the Sehoo shoreline-cut on the same hills at 1330 m. Additional small remnants of slope breaks may also record shoreline an- gles at altitudes of 1353 m and 1396 m. A high shoreline is fur- ther supported by an outcrop on the east side of one of the hills. There, rounded clasts of rhyolite and basalt overlie a rhyolite flow weathering into angular clasts; the rounded clasts, inter- preted as beach gravel, crop out as high as 1390 m. EVIDENCE FROM THE NORTHEASTERN LAHONTAN BASIN (HUMBOLDT RIVER).?During the Sehoo highstand, the north- ern subbasins of Lake Lahontan were filled to an altitude of -1330 m. According to Davis (1982), Lake Lahontan could have been connected to the Walker Lake subbasin only by the narrow Adrian Valley (Figures 1, 2). King (1993), however, thought that Adrian Valley did not exist during the early-mid- dle(?) Pleistocene. If his theory is correct, high lakes in the Walker Lake subbasin would have been isolated from Lake Lahontan during this time. Yet in the foregoing discussion, we have documented a lake in the Walker Lake subbasin that rose to an altitude of about 1400 m. This lake could have spilled over a pair of sills at 1370 m into the northern subbasins of Lake Lahontan through Rawhide Flats (Figure 2), if the sills leading into Rawhide Flats were not much higher then than they are now. If such a connection existed, evidence of high shorelines north of the Desert Mountains would be required to establish that the very high shorelines found in the Walker Lake subbasin were coextensive with Lake Lahontan as a whole. No such evidence had been previously reported. Rob- erts (1965), however, suggested that the region around Lone Tree Hill contained evidence of a lake, but because the sites he described were above the Sehoo highstand, he inferred that this lake was separate from Lake Lahontan and must have been im- pounded by a dam in the Humboldt River canyon east of Winnemucca. Most of Roberts' sites, except that at Treaty Hill, consist of fluvial terrace deposits of the Humboldt River. 80 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Figure Section 12B 14 Top of section 15 FIGURE 12.?Photographs of central part of Sunshine Amphitheater. A, View north showing area of sections 14 and 15 (see Figure 10); fence posts show scale. B, Paleosols and fan deposits separating units Q12S and Q13S at the top of section 14, exposed at top of bluff in A. Trowel indicates top of unit Q12S. NUMBER 33 81 o I- x LU 5 w CD 73 CD CO = = Q) CO CO _Q ? Q-g ofjfi CO ? -^ CDS C o oo. co .2 CD > O SJ919LU IT) ?? ell \ Q. ?03 O ion CD oCM O Unit o co T3 ?= > O O c Q. CO ?-2 ? E o ?- a"5 ?i .Q CO O) (0 ??'.???.??.???.??. DC CD CM CL Q. o below E DC LO CM Q_ cg CD(/) m itted! O "o T3CD CO CO CD COco lably J_Io CO ins tep is tconCO d I (Glass rf _i LU CO O XI O E o a: 4 CM Q. DC E CO CD CD ^ CM 60C .2 (-a S ipsy T3 d 1 O S u E T3 Kt CL JUI 01 i?(UJ u 3 hra SJ a -88 2 ?3 a. 82 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES B FIGURE 14.?Photographs of Campbell Valley area; person is for scale, A, View west of outcrop of tephra EL-15- CA, showing drape of tephra from shoreface position atop basalt into deeper-water sediments. See Figure 13 for location. B, View north of shoreline angle cut on outcrop marking shoreline (-1390 m) reached by lake during deposition of unit Q12C. Hill is just behind highest point visible in A. Work began in 1997 at a new site on the Humboldt River be- tween Winnemucca and Battle Mountain (Figure 1), within the area studied by Roberts (1965). This site has both spectacular exposures and geomorphic evidence of pre-late Pleistocene lakes. The outcrops are within large open pits of the Lone Tree gold mine, currently operated by Newmont Gold Corporation; as a result of mining operations, the exposures are excellent but change constantly. Our work has focused on a less fre- quently remodeled roadcut within the mine on the south flank of Lone Tree Hill (Table 1) and on samples from a core drilled southeast of the hill (Figure 15). The data from these sites are supported by observations and tephra samples from other, now-destroyed pit walls. Premining topographic maps show a prominent berm as high as 1410 m extending south from Lone Tree Hill. The roadcut sequence (Figures 15, 16) is within this berm. Well-rounded cobbles are lagged on slopes above the berm as high as about 1415 m. Outcrops of lacustrine deposits form the upper 5-50 m of the walls of all of the open pits south and west of the hill, providing a three-dimensional view of kilometers of outcrop. The lacustrine deposits and older units are cut by three normal faults that bound the hill on the east, south, and west; thus, Lone Tree Hill is a horst. From outcrops and core data, the base of the lake deposits appears to be offset at least 40 m by the eastern fault, 8 m by the southern fault, and 50 m by the west- ern fault. The amount of absolute uplift of the horst is un- known, but it could be considerable in such an active tectonic setting. A younger, lower berm at about 1370 m in altitude was present on the northwest side of the hill, but the berm is now obscured by construction. Outcrops in Lone Tree Mine and core samples are tentatively interpreted to represent four lake units separated by unconfor- mities and paleosols. In the roadcut exposure, unit Q11L is re- versely magnetized where the sediments are fine-grained; these sediments interfinger with coarser sand and gravel with steep foreset beds (Figures 15, 16) containing a tephra layer (sample NUMBER 33 83 Roadcut sequence, northeast mine area (highest elevation 1410 m) QI3L .o.o. o o Mn-cemented beds -Diatomite;S11-LT . EL78-LT; Bishop (?) ash, 0.76 Ma QI2L S13-LT (projected; approx. stratigraphic location) S9-LT EL76-LT; Glass Mtn. ash (?), ~1 Ma Core GT-2 (collar altitude 1366 m) EXPLANATION Beach gravel? Greenish-white; floating pebbles ^ GT2-28.2-LT ? unconformity FIGURE 16.?Photographs of central portion of roadcut sequence in Lone Tree Mine. A, East-dipping foreset gravel beds and diatomite in unit Q12L overlie a horizontal unconformity and paleosol that formed on west-dip- ping foreset beds and tephra EL-76A-LT (Table 3) in unit Q11L. B, Close-up of fossiliferous sand layer above tephra (pencil for scale); sand contains abundant fish bones {Catostomus sp.; G.R. Smith, University of Michi- gan, in litt. to M.C.R., 1998) and ostracode shells. EL-76A-LT, Tables 2, 3) with Bishop-Glass Mountain-type chemistry. A sandy layer above the tephra layer (sample S9-LT, Figures 15, 16B) contains abundant ostracodes and small bones of Catostomus sp. (sucker; G.R. Smith, University of Michi- gan, in litt. to M.C.R., 1998). Unit Q12L in the roadcut consists entirely of sandy gravel, which is commonly steeply bedded, and aim thick diatomite bed (Figure 16A); the unit overlies an unconformity and a paleosol. In two places, gravel clasts of unit Q12L are embedded in a matrix of tephra (sample EL-78- LT), which also has Bishop-Glass Mountain-type chemistry. These characteristics indicate that units Q11L and Q12L are from about 1 Ma and 0.76 Ma, respectively. It is possible that the NUMBER 33 85 lower tephra could be the Bishop ash because the sediment en- closing the tephra was too coarse-grained for paleomagnetic determination, but this is unlikely because this coarse sediment interfingers laterally with fine sediment that is reversely mag- netized, as noted above. Also, the presence of a paleosol sepa- rating the units indicates that the upper tephra was probably not reworked from the lower tephra. The youngest and highest lake unit in the roadcut is undated. Correlation of the roadcut exposures to the core samples is problematic because of differences in grain size, fault displace- ment between the core and the roadcut, and alteration in the core sediments that prevented analysis of most of the tephra samples. However, the core should contain a more complete re- cord of lacustrine deposition because of its location on a sub- siding fault-block. A tephra layer (sample EL-92-LT) from the lowest lake unit in the core contains shards that are chemically similar to the Lava Creek B ash bed and to the Huckleberry Ridge ash bed, both of which erupted from the Yellowstone area (Izett, 1981) 0.66 Ma and 2.1 Ma, respectively (Pluhar et al., 1992; Sarna-Wojcicki et al., In review). Ostracodes that be- came extinct by about 0.6 Ma (Elkocythereis bramletti (Dickin- son and Swain, 1967); R. Forester, USGS, in litt. to M.C.R., 1998) are present throughout the core, indicating the older age is likely correct. The older lake unit (from 2.1 Ma) in the core is not present in the roadcut (Figure 15). Environmental conditions interpreted from ostracodes throughout most of the core and from unit Q11L in the roadcut indicate this area had shallow to moderate water levels, rela- tively saline alkaline water, and cool-season runoff as well as warm summers (R. Forester, USGS, in litt. to M.C.R., 1998). An unusual suite of ostracodes in sample S-13-LT from unit Q12L (Figure 15) includes several highly ornate, endemic spe- cies that are thought to have evolved in a long-lived, stable lake (discussed in Forester, 1991) and a species that can live in a rel- atively deep-water lake (Candona aff. C. caudata Kaufmann, 1900). We speculate that the endemic species, which clearly did not evolve locally because they are not present in any other samples, could have arrived during a sudden influx of water from a previously undrained lake basin upstream. The Can- dona shells suggest deeper water prevailed during the time of deposition of unit Q12L than during that of Q11L. A few samples from the core and the roadcut exposures contained ostracodes and diatoms (J. Bradbury, USGS, in litt. to M.C.R., 1998) that live in springs. This is consistent with sediments in the core and outcrops that are interpreted as spring deposits and with the presence of modern springs at Lone Tree Hill. Other localities near Lone Tree Hill provide additional evi- dence of pre-late Pleistocene highstands. On the north side of the Humboldt River, a saddle between the two summits of Treaty Hill (Figure 1, Table 1) has the form of a beach berm and consists of very well-rounded pebble to cobble gravel of diverse lithologies. The saddle is at an altitude of 1402 m; peb- ble lag extends up the south hillslope to as high as 1414 m. This gravel is unlikely to be a terrace deposit of the Humboldt River because the shape and orientation of the berm seem to preclude deposition by the river, and because the river does not flow in an antecedent course across a rising block at this loca- tion. However, some tectonic uplift of Treaty Hill, like that in- ferred at Lone Tree Hill, is certainly possible. A lower ero- sional bench on the east side of Treaty Hill also may represent a shoreline at about 1360 m, as noted by Roberts (1965). West of Lone Tree Hill, at Iron Point on the Humboldt River, scattered, very well-rounded pebbles lie on gently sloping basalt surfaces at altitudes from 1350 m to 1370 m (Table 1). Northeast of Treaty Hill is a broad, very flat surface that slopes gently westward for several kilometers from an altitude of 1362 m to 1344 m (Table 1) and is underlain by well- rounded gravel and sand. This surface is interrupted by a sub- dued break in slope at 1356 m. A drainage ditch exposes a well-developed argillic soil that formed on these gravel depos- its and that is buried beneath nearly 2 m of loess. Although ex- posures are few and poor, the evidence suggests that this sur- face and the underlying deposits may represent pre-late Pleistocene deltaic deposits of the Humboldt River graded to a lake level that reached as high as about 1356 m. EVIDENCE FROM OTHER SUBBASINS OF LAKE LAHONTAN.? Soda Springs Valley (Figures 1, 2) extends southeast of Walker Lake. Modern elevations of the valley floor are below 1400 m; small basins in the valley are as low as 1330-1352 m, and sills are between 1377 m and 1396 m. It is likely that valley-floor altitudes are higher at present than in the past because of allu- vial-fan sedimentation. These sediments accumulate because rates of vertical displacement along the active strike-slip faults in this part of the Walker Lane belt are low (Yount et al., 1993). Thus, the 1400 m high lake-level recorded at several sites in the Walker Lake subbasin indicates that the pre-late Pleistocene lake should have flooded Soda Springs Valley. Lacustrine de- posits of Miocene to late Pliocene age are preserved along the range fronts on both sides of Soda Springs Valley and south of Rhodes Salt Marsh (discussed below); however, younger lacus- trine deposits have not been found. Several sites along the Truckee and Carson Rivers in can- yons below Reno and Carson City, respectively (Figure 1, Ta- ble 1), consist mainly of rounded, exotic clasts overlying bed- rock. The clasts may represent times when rivers were graded to lake levels far above the Sehoo shoreline, or they may repre- sent terraces stranded by uplift of the mountain blocks through which the rivers flow. Arcuate geomorphic features with rounded cobbles occur in the Truckee River canyon at an alti- tude of about 1400 m, and one roadcut exposes sediment that may be lacustrine or deltaic at about 1355 m. At one locality on the Carson River, the rounded clasts are found on slopes pro- tected from river erosion and above a bedrock sill where the Sehoo-aged lake spilled northward out of the canyon. The clasts at this site are more likely to reflect high lake-stands rather than fluvial gravels. Good evidence of high, old shorelines exists at three sites in the Smoke Creek Desert (Figure 1, Table 1). These old shore- 86 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES lines were first observed by Mifflin while conducting field work during the 1970s. The southern site is near Burro Moun- tain in the Smoke Creek Desert and lies 1-2 km west of an ac- tive Quaternary fault that bounds the western edge of the basin (Weick, 1990). Two subtle benches and risers are underlain by basalt and rhyolite flows (Figure 17A). The upper bench termi- nates upward at a slope break at an altitude of about 1400 m, and well-rounded small pebbles, many flattened and discoidal in shape, are scattered over the bench at altitudes between 1395 and 1400 m. The lower bench, further east and closer to the fault, is at about 1375 m and is blanketed with abundant well- rounded pebbles and a few fragments of lacustrine tufa (Figure 17B). A few kilometers to the north, another site also has abun- dant well-rounded pebbles of similar shape at about 1390 m. Alluvial interbeds that could have weathered to provide rounded gravel were not observed within the volcanic flows at these sites. It is possible that the well-rounded pebbles at these sites could represent deposition by an ancient stream prior to uplift along the Bonham Ranch fault zone (Weick, 1990); three other nearby sites that were field-checked by M.C.R. were so interpreted on the basis of their lithologic composition and broad range of clast sizes, which were similar to those of the upper bench K^.v^EJ-.-^ FIGURE 17.?Beach-gravel sites in the Smoke Creek Desert (see Figure 1, Table 1). A, East flank of Burro Moun- tain. View is to the west from the lower shoreline bench at 1375 m to the upper bench at 1400 m. B Typical scat- ter of beach pebbles at Burro Mountain site; remnants of lacustrine tufa also present on lower bench. Lens cap is for scale. C, View northwest to Freds Field site. Subtle benches are cut on basalt flows above the Sehoo shoreline on photo right. D, Well-rounded and polished basalt clasts make up the majority of block streams on the shoreline benches of the Freds Field site. Rock hammer is for scale. NUMBER 33 87 local modern streams, namely, Smoke Creek and Buffalo Creek. The eastern site in the Smoke Creek Desert (Figure 1; Table 1, Freds Field) consists of two benches and risers underlain by basalt flows (Figure 17c), forming subdued topography above a steep cliff on which the Sehoo shorelines are cut. The benches are laterally continuous and are covered with basalt clasts closely fitted together in pavements interspersed with rougher surfaces. The risers are also covered with basalt clasts that form matrix-poor stone stripes. More than 50 percent of the clasts on the benches and risers are polished and rounded (Figure 1 7D) rather than angular or subangular, as is typical of clasts in colluvial block fields. This polishing and rounding re- quires vigorous transport in water. The upper bench at this site rises gradually from an altitude of about 1370 m to a subtle slope break at about 1400 m; the lower bench rises from a flat surface at 1360 m to a distinct slope break at about 1365 m. The Freds Field site appears to be unaffected by faulting, and it is located well away from possible locations of ancestral streams that might have carried rounded gravel. The Sehoo shoreline at Freds Field bears clasts of more varied lithologies than do the upper benches. This difference possibly results from the topographic setting; east-flowing longshore currents in the inferred higher lakes might have been channeled along the cliff bearing the Sehoo shoreline but below the higher benches (Figure 17c). Summary: Very High Shorelines of Lake Lahontan Several outcrops in the Walker Lake subbasin, one locality in the northeastern Lahontan Basin, and two localities in the Smoke Creek Desert subbasin document highstands of a lake during the early and middle Pleistocene that rose far above the Sehoo and Eetza(?) shorelines. A conservative correlation among all the sites in the Walker Lake subbasin yields a total of six or seven lacustrine units whose shorelines exceeded that of the Sehoo lake. Three of the units date from about 1 Ma to less than 2.6 Ma; one tentatively contains the Bishop ash bed; and three are probably younger than the Bishop ash bed, but they definitely date from less than 1 Ma. At Lone Tree Hill, the old- est lake unit is from about 2 Ma, two middle units are probably from about 1 Ma and 0.76 Ma, and the highest unit was depos- ited less than 0.76 Ma. Other lake units may be present in the core. In the area north of Lone Tree Hill, erosional features and possible deltaic deposits are interpreted to indicate an even younger and lower highstand. Based upon these correlations, we assign informal names to the deposits that comprise a lake unit?that is, the fine- and coarse-grained lacustrine sediments deposited during the rise and fall of lakes during a pluvial climatic phase. Naming the lake units permits easy revision should a new pluvial phase be identified, and it allows changes in the correlation of lake units at a site if new age information is obtained. This is particularly likely where correlations are made chiefly on the basis of alti- tude. Variations in altitudes of such old landforms and deposits, especially as they represent minimum altitudes of lacustrine highstands, are expectable in the tectonically dynamic land- scape of the Basin and Range. Without independent dating of beach gravels at different sites, we cannot be confident that they represent the same highstand. Stratigraphic relations among four sites (Figure 5) permit a conservative correlation of the two youngest pre-Sehoo units; these units always crop out as beach gravel inset into all older units. The maximum altitude of the youngest pre-Sehoo unit (Q16T) at Thorne Bar is about 1350 m; it is characterized by a relatively well-preserved barrier and is henceforth referred to as "younger lake deposits of Thorne." The delta plain between 1360 m and 1345 m in altitude, north of the Humboldt River near Treaty Hill, is provisionally correlated with this lake unit. Maximum altitudes of the next older lake units range from 1356 m to 1378 m at the three tectonically deformed sites (units Q14M, Q13C, and Q14S). Unit Q15T at the Thorne Bar site has not been deformed, and its maximum altitude is about 1370 m, which corresponds with the shoreline angle cut on basalt hills north of Weber Reservoir (Figure 2). We infer that these units represent a highstand at about 1370 m and are herein termed "older lake deposits of Thorne." The lower of the two shoreline benches at two sites in the Smoke Creek Desert sub- basin is correlated with these deposits. The beach-gravel units overlying older units containing tephra at the sections north of Thorne Bar and at McGee Wash, Campbell Valley, and Lone Tree Hill are also those that rise to the highest altitudes at each site: about 1400 m at Thorne Bar and McGee Wash (units Q14T and Q13M); 1390 m at Campbell Valley (unit Q12C); and 1414 m at Lone Tree Hill. These data suggest, but do not prove, that these units represent the same highstand at about 1400 m altitude. We also tentatively assign unit Q13S (altitude 1414 m) at Sunshine Amphitheater to this period, as well as the higher of the two shoreline benches at sites in the Smoke Creek Desert subbasin. This unit is named "lake deposits of McGee" for the arcuate berm composed of beach deposits at that locality. The tephra layer within unit Q13T north of Thorne Bar is ten- tatively correlated with the Bishop ash bed. The correlation is based upon the chemical composition (Table 3) and strati- graphic position of the tephra layer (Figure 3). The tephra-ma- trix bed in unit Q12L at Lone Tree Hill (Figures 15, 16) has the chemical composition of the Bishop ash and overlies unit Q11L containing probable Glass Mountain tephra from 1 Ma. This stratigraphic relation strengthens the interpretations that (1) the tephra layers near Thorne Bar and Schurz (discussed above) are the Bishop ash, and (2) this unit represents a sepa- rate lake cycle, herein termed "lake deposits of Lone Tree" (Figure 5). Unit Qllc at Campbell Valley, unit Q12M at McGee Wash, and unit Q11L at Lone Tree Hill contain a Glass Mountain tephra that erupted at about the time of the Jaramillo Subchron, which occurred about 1 Ma. That would correlate this unit with SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Morrison's (1991) younger pre-Eetza unit west of Weber Res- ervoir (Figure 2). Unit Q12S at Sunshine Amphitheater also was deposited during and after the Jaramillo Subchron. We name this unit "lake deposits of Campbell" for the tephra in lacus- trine sediments at this locality. Paleomagnetic data from Sunshine Amphitheater show that the two oldest lacustrine units there lie within the Matuyama Reversed Chron. These units may be correlative to older units at Thorne Bar, McGee Wash, and Lone Tree Hill, and to Morri- son's (1991) older pre-Eetza unit west of Weber Reservoir. Unit Qlls is from between 1 and 1.4 Ma, based on paleomag- netic data and fossil ages, and is herein named "lake deposits of Sunshine." Unit QT1 is from more than 1.4 Ma but less than 2.6 Ma (timescale of Cande and Kent, 1995). The lowest lake unit in the core from Lone Tree Mine (Figure 15) may be equivalent to unit QT1 in the Walker Lake subbasin. Because unit QT1 seems to represent different environmental conditions than the younger lakes (paper shale at Sunshine Amphitheater, gypsifer- ous mudstone at McGee Wash) and may include deposits of different ages, we do not assign it a pluvial-lake name. The combined shoreline sites can be used to infer a partial history of connections among the Lahontan subbasins (Figure 1). Shoreline evidence from Freds Field in the Smoke Creek subbasin and from Thorne Bar in the Walker Lake subbasin, as well as deltaic deposits of the Humboldt River northeast of Treaty Hill, indicate the presence of highstands at altitudes of about 1400 m, 1370 m, and 1350 m (Figure 18), implying that the northern and southern subbasins have periodically merged into a single lake since the early-middle Pleistocene. In the Lone Tree area, faulted deposits of the highest shoreline, as- signed to the lake deposits of McGee, are as high as 1414 m in altitude on the footwall block. The lake deposits of Lone Tree and Campbell represent minimum highstands at about 1355 m in the Walker Lake subbasin and at 1405 m and 1390 m, re- spectively, on the footwall block of Lone Tree Hill. Total dis- placement in the last one million years is at least 40 m on two of the faults at Lone Tree Hill, and the orientation of the faults suggests that uplift of Lone Tree Hill could account for a sig- nificant portion of the displacement. Thus, the discordance in altitudes still permits the possibility of a contiguous Lake Lahontan from the middle Humboldt River to Walker Lake as early as 1 Ma. The presence of bones of the Lahontan cutthroat trout (R.R. Miller, University of Michigan, in litt. to M.C.R., 1996) in unit Qlls at Sunshine Amphitheater (Figure 10) requires a drainage connection (but not necessarily a continuous lake) between the Walker Lake subbasin and the northern subbasins at some time prior to 1 Ma. A continuous lake at 0.76 Ma is supported by the appearance of Candona caudata, which can live in colder and thus deeper water than the ostracodes that were preserved in older deposits at Lone Tree Mine (R. Forester, USGS, in litt. to M.C.R., 1998). Possibly these older deposits (>0.76 Ma) reflect damming of the Humboldt River above its canyon east of Win- nemucca, as suggested by Roberts (1965). These data do not refute King's (1993) interpretation that the Walker River flowed into an isolated Walker Lake subbasin in the early Pleistocene, but they may push back the timing of integration to before 1 Ma. The general correspondence of shoreline altitudes of the lake deposits of McGee and the older and younger lake deposits of Thorne in the Walker Lake subbasin and in the northern subba- sins argues for little or no regional northward tilting during the middle Pleistocene, although a few meters of tilt during the late Pleistocene (Mifflin, 1984) are possible. The shorelines, how- ever, have been significantly affected by local faulting, tilting, and isostatic rebound. For example, the proposed correlations among Lahontan sites (Figure 5) imply that the -1400 m shoreline altitude recorded near Thorne Bar has been raised about 14 m at Sunshine Amphitheater, Lone Tree Hill, and Treaty Hill and lowered about 10 m in Campbell Valley; the sense of these displacements corresponds with those of mapped faults at the four localities (Reheis and Morrison, 1997). Very high beach gravels are preserved at two sites on the west side of Walker Lake (Table 1, Figures 2, 6) on the footwall block of the Wassuk normal fault. If the highest rem- nant of these gravels corresponds to the 1400 m shoreline, it records 50-75 m of absolute uplift of the Wassuk Range since the early-middle Pleistocene. OTHER BASINS IN WESTERN AND CENTRAL NEVADA Previous workers inferred the existence of ancient drainage connections or contiguous lakes extending southeastward from the Walker Lake subbasin to Fish Lake Valley (Figure 1). Such "early pluvial" connections were first suggested by Hubbs and Miller (1948) on the basis of remnant populations of Lahontan fish, including Siphateles bicolor obesus (Girard, 1856) (tui chub) in Fish Lake Valley and S. b. obesus and Rhinichthys os- culus (Girard, 1856) (speckled dace) in the basin of Lake Toiy- abe in northern Big Smoky Valley. The lowest outlet to Lake Toiyabe at present is south toward Lake Tonopah through an al- luvium-filled valley about 60 m above the late Pleistocene shoreline of Lake Toiyabe. Mifflin and Wheat (1979) specu- lated that Rhodes and Columbus Salt Marshes and Clayton Val- ley once acted as overflow basins for drainage spilling south from the Lahontan Basin because they contain anomalous con- centrations of various salts relative to their drainage basins. Pre-Pleistocene Lakes LAKES SOUTH OF WALKER LAKE SUBBASIN.?We infer that the 1400 m level of Lake Lahontan extended 60 km southeast from Walker Lake through Soda Springs Valley to the south end of Rhodes Salt Marsh (Figure 1). No definite deposits of middle Pleistocene age have been identified in these valleys. Older lacustrine deposits crop out along the range fronts both east and west of Mina, Nevada, and at one site these deposits contain a probable late Miocene tephra layer (EL-24-M, Table NUMBER 33 89 | g o o E o O O o w U681-~ ujseq UBIUOU.B~| OJUj P|OLJS9JL|} UJ9P0UJ 9t?6l-~ P|0L|S9JLj} ujapooi meters Ititude de in d altitu CO a)c relio .cCO \1 7 CO CO CJ 73 CO g CD ?*-? CD 1 T3CD o nlcc? Q. > CO CO ; deposit ^J or R is strine or lacu olf c o < iGJar. S o event amill tes tep ?. is Jar 3 COis ] Jai- an oQ. C "IS 3 unceiO) CO ?^_ o CDO) cCO tr <- C\J CO ? .5 i- TJ c "-P ~ == 1 and s its s ine depu- to3 o _es (U So ON ? "a rie et ta fromca T3 "2o u u ,g so .S2 o o o o *c3 ?cQ c oo o from .22 j^ oo o "S-a i?T ii c" c o las: O IIo exo x: CQ IIII CQ CO '-.o 1?60 CnJ icate Arrow o o ? O C a 5 I o oHZ flat tion oca "3 o -as o ickle O3 PH :t(Su ?51 "limat<-> H o O c"Z ?? riabi tifie o tters ? t-~ o>"1 Sr ^r H o 08 OH Rye o,eJ ^" a CO C3 S-s^S i o6e sjeaA jo 90 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 3) and late Tertiary diatoms (J. Bradbury, in litt. to M.C.R., 1996). Younger deposits, probably dated at about 2.5-2.2 Ma based upon tephra correlations (layers EL-2-RM and EL-26- RM, Table 3) and diatoms, crop out along the highway south of Rhodes Salt Marsh up to an altitude of about 1423 m (Figures 1, 19A). These deformed, latest Pliocene lake deposits may ei- ther lap onto or underlie old fan gravel that blankets the hills to the south (Figures 1 9B, 20) that, in turn, lie above a thick se- quence of lacustrine sediment beneath Redlich Summit, the present-day divide between Rhodes and Columbus Salt Marshes (Reheis and Morrison, 1997). Lacustrine deposits in- vestigated by J. Liddicoat (Columbia University) at the south end of Teels Marsh, west of Rhodes Salt Marsh, contain a tephra layer identical to tephra EL-2-RM in the roadcut at the south end of Rhodes Salt Marsh (Table 3). Lacustrine deposits occupy a wide area north of Redlich Summit (Figures 19B, 20). They had been previously mapped (Stewart et al., 1994) but had not been studied in detail. Recon- naissance mapping and several measured sections for this study indicate the presence of two or more lacustrine units separated by unconformities or periods of desiccation. Outcrops north- east of the highway expose an upward sequence of deltaic sandstone, diatomaceous siltstone, green massive mudstone, and nearshore sandstone with oscillatory ripple marks. Chemi- cal analyses of many tephra layers indicate correlations with tephra ranging in age from late Miocene to late Pliocene (e.g., EL-29-RM, EL-34-RM, and EL-50-RM, Table 2). Tephra lay- ers near the top of one section where lacustrine deposits are in- terbedded with fan deposits exhibit inverse grading of pumice clasts, reflecting a beaching line (shoreline; Sarna-Wojcicki FIGURE 19.?Photographs of lacustrine deposits south of Rhodes Salt Marsh (see Figure 1, Table 1). A, Roadcut exposes two sets of late Pliocene lacustrine sediments containing tephra layers (Tables 2, 3) separated by fluvial or deltaic deposits. B, View northeast near Redlich Summit of late Miocene to late Pliocene lacustrine deposits filling divide between Rhodes and Columbus Salt Marshes (power poles included for scale). NUMBER 33 91 118? orw 118? OO1 117? Section 5 Section 7 Section 6 15 m omitted 15 m omitted Unit Qfo Reddish; cross- bedded, some debris flows Unit Tim Common thin, 40 m omitted altered tephra layers Unit Tim Dip change; ? minimum thick- nesses below / EL32-RM/ Green; massive, saline^'many small folds' and faults Reddish; cross- bedded, some debris flows EL31-RM Gray; thin-bedded, common oscillatory ripple marks Unit TIs Coarsens upward Probably same as tephra EL47-RM Green Probably same as tephra EL46-RM Green EL51-RM Reddish; mudcracks Green; ripple marks EL33-RM EL50-RM Unit Tim EL49-RM EL48-RM Green, Unit Tld diatomaceous ***.'EL45-RM EXPLANATION p~-^| Mudstone Igidgij Interbedded mudstone N=^=H and sandstone Siltstone and sandy siltstone Sandstone and silty sandstone, locally conglomerate beds Sandstone and con- glomerate, poorly to moderately sorted Fanglomerate, poorly sorted and bedded | xxx| Tephra layer and EL31-RM sample number & ? 5 Qfy 38?14'N Qfm Tu Ta TIs 38? 13' EXPLANATION Fan deposits (Holocene) Fan deposits (late and middle Pleistocene) Fan deposits (middle and early Pleistocene) Tertiary sediments, undiffer- entiated (late Pliocene) Tertiary alluvial deposits (late Pliocene) Tertiary lacustrine deposits (late Miocene? and Pliocene). Locally divided into: Nearshore sandstone and conglomerate Diatomaceous sandstone and siltstone Mudstone Deltaic sandstone and conglomerate Stratigraphic relation of surface and underlying unit Tephra locality Line of measured section Fault; bar and ball on downthrown side Strike and dip of beds 1 2 KMJ 0 1 MILE CONTOUR INTERVAL: 80 FEET FIGURE 20.?Preliminary geologic map and mea- sured sections of area northwest of Redlich Summit between Rhodes and Columbus Salt Marshes (see Figure 1, Table 1). 92 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES and Davis, 1991; Reheis et al., 1993). These deposits indicate that a lake intermittently occupied the present divide between Rhodes and Columbus Salt Marshes up until about 2 Ma. Fish Lake Valley is located south of Columbus Salt Marsh and presently is part of its drainage area (Figure 1). Pluvial Lake Rennie existed in Fish Lake Valley during at least four in- tervals from the late Pliocene to the middle Pleistocene (Reheis et al., 1993). Lacustrine deposits in the valley include mud- stone, siltstone, sandstone, and conglomerate. Numerous tephra layers (e.g., EL-3-FL, EL-4-FL, and EL-44-FL, Table 3) within the lacustrine deposits document lake episodes from about 3.4 to 2.8 Ma and from about 2.2 to 2.0 Ma. Deposits of similar age do not crop out in Columbus Salt Marsh. Davis (1981) and Kunasz (1970) used data obtained during field work and data from well records from a lithium-extraction operation to show that Clayton Valley (east of Fish Lake Val- ley, Figure 1) was periodically filled with fresh to brackish lakes. Exposures on the east side of the valley indicate at least two such intervals occurred (Davis, 1981). A recent analysis of tephra layers (samples EL-42-CV and EL-43-CV, Table 3) from the lower interval suggests the interval occurred between 3 and 2 Ma. Thus, the record of Pliocene lakes in Clayton Val- ley is in accord with those in Rhodes Salt Marsh and Fish Lake Valley (Figure 18). The parallel histories of Pliocene lakes south of Rhodes Salt Marsh and in Clayton Valley and Fish Lake Valley, combined with the lack of sills or the presence of very low sills between Columbus Salt Marsh and Fish Lake, Clayton, and Big Smoky Valleys (Figure 18), permits the interpretation that the lakes in these basins were interconnected during the Pliocene. In addi- tion, we infer that a Miocene to Pliocene lake basin that at least included Rhodes and Columbus Salt Marshes was disrupted by faulting and (or) compression along the northeast-striking Huntoon Valley fault system (Stewart, 1988; dePolo et al., 1993) that lifted the area of Redlich Summit, which is presently about 35 m above the highest known shoreline altitude in Co- lumbus Salt Marsh (discussed below) and about 95 m above the 1400 m high shoreline of Lake Lahontan. Early to Middle Pleistocene Lakes PLUVIAL LAKE COLUMBUS-RENNIE.?Reheis et al. (1993) showed that a large lake, informally named Lake Rennie, peri- odically existed in Fish Lake Valley during the Pliocene (dis- cussed above) to middle Pleistocene. They noted that thick, shallow-water deposits of Bishop ash indicate the existence of a large lake at about 760 Ka with a highstand of about 1460 m in elevation; a lake of uncertain size probably also existed at about 1 Ma. These outcrops of lacustrine sediment are sup- ported by data recorded in drilling logs for privately owned wa- ter wells. The logs, copies of which are stored at Carson City, Nevada, by the Water Resources Division of the U.S. Geologi- cal Survey, record the occurrence of numerous thick beds of clay and sandy clay that are inferred to be deep-water lacus- trine deposits. Reheis et al. (1993) speculated that Lake Rennie and Lake Columbus at times were contiguous through The Gap, the present-day drainage connection between Fish Lake Valley and Columbus Salt Marsh. New discoveries in Columbus Salt Marsh (Figure 1) indicate the presence of at least four lacustrine units that are higher and older than the late Pleistocene shoreline (altitude 1400 m; Mif- flin and Wheat, 1979; Reheis et al., 1993) of Lake Columbus. No Quaternary faults have been mapped in this subbasin (Dohrenwend, 1982a; Stewart et al., 1994). Outcrops in a small hill on the south side of the subbasin expose three units com- posed of beach gravel, lacustrine sand, and silt; each unit is separated by paleosols that formed in alluvial or eolian deposits (Figures 21, 22). Paleomagnetic measurements yielded mainly indeterminate polarity as a result of diagenetic and pedogenic alteration, but one sample from unit Q12CM has reversed mag- netization. The uppermost unit in this section, Q13CM, consists of tufa-cemented gravelly foreset beds that underlie a promi- nent berm, which protrudes from the south side of the hill at an altitude of about 1435 m (Table 1). This altitude and the well- preserved morphology are similar to those of beach berms ini- tially identified by J.O. Davis (Desert Research Institute, in litt. to M.C.R., 1989) on the southeast side of the subbasin at Coaldale Junction (Reheis and Morrison, 1997). Very well- rounded gravel clasts (in part reworked from a nearby Tertiary conglomerate) and tufa are present as lag on the slope above the berm to the top of the hill at an altitude of 1452 m. The clasts and tufa suggest a yet higher lake stand that could be equivalent to unit Q13CM or to one of the buried lacustrine units. A similar gravel-covered bench on the west side of Columbus Salt Marsh extends to an altitude of 1457 m. These altitudes are very similar to the shoreline altitude of pluvial Lake Rennie (1460 m), and we infer that they record the highest shoreline of the combined lakes at about 760 Ka. Evidence of the youngest pre-late Pleistocene lacustrine unit here has been destroyed by gravel operations, but in 1991, excavations exposed gravelly foreset beds in a berm at an altitude of about 1417 m. This alti- tude corresponds to those of small berms on the north side of the subbasin (Figure 1, Table 1); surface soils on these berms are moderately developed with argillic horizons that are not present on the late Pleistocene deposits of Lake Columbus. A deep pit on the north side of the subbasin exposes lacus- trine sand and marl containing probable Bishop ash (sample 29068501, Tables 2, 3; collected by J. Whitney and P. Glancey, USGS) that are stratigraphically beneath late Pleistocene beach deposits of Lake Columbus (Figure 21A). The tephra layer is impure, and no unconformities or buried soils lie between it and the overlying late Pleistocene deposits. Thus, the tephra may have been reworked into much younger deposits and may not truly represent the age of the older lake unit. In summary, deposits and berms at several locations around Columbus Salt Marsh indicate there had been at least four lake cycles that were older and higher than the late Pleis- tocene shoreline: the two oldest lakes rose in altitude to at least 1420 m, and one of these two rose to 1457 m; an inter- mediate-aged lake rose to about 1435 m; and a younger lake NUMBER 33 93 LakePletelocene ^JE^ deposits e FIGURE 21.?Pre-late Pleistocene lacustrine features in Columbus Salt Marsh (see Figure 1). A, View of west side of deep pit on north side of valley; small dog and 1 m pick below tephra are for scale. Late Pleistocene lake gravel at top of pit overlies a sequence of debris-flow deposits that, in turn, overlie middle Pleistocene lake gravel containing probable Bishop ash (Tables 2, 3, no. 29068501) at bottom of pit. B, View north from borrow pit on southwest side of valley (see Figure 1) to shorelines (arrows) cut on basalt outcrops. rose to 1417 m. One of the two oldest lakes probably occurred about 760 Ka. PLUVIAL LAKE CLAYTON.?The floor of Clayton Valley has been cored numerous times by the USGS (Pantea et al, 1981) and by mining companies (reported in Kunasz, 1970; Davis, 1981). No exposures of lacustrine deposits of Pleistocene age have been found. Most of the wells bottomed in the so-called "ash aquifer," a thick (as much as 10 m), coarse-grained tephra layer that lies at depths between about 100 and 300 m below the playa surface. This tephra layer contains dissolved lithium salts in a large brine pool, and it is these salts that are exploited by the mining companies. Sediments in the cores consist of alter- nating beds of sand, mud, gypsum, and halite, and include sev- eral coarse tephra layers and glass-rich intervals. The mud is in- terpreted to reflect deeper, freshwater conditions in the lake, whereas the gypsum and halite beds are interpreted to reflect periods of brackish water and desiccation (Davis, 1981). In one core, Kunasz (1970) interpreted the absence of gyspum and ha- lite below about 70 m to indicate freshwater conditions. In this core, this depth approximately corresponds to the interval in- cluding and below the ash aquifer. A sample of the ash aquifer obtained by Davis (1981) from a depth of 175 m in another core hole is tentatively identified as the Bishop ash (sample D77-2D, Table 3). In summary, sediment beneath Clayton Valley records a long history of intermittent lacustrine deposition, including one relatively freshwater lake at about 760 Ka. PLUVIAL LAKE RUSSELL (MONO BASIN).?The Mono Basin is a structural depression that probably began forming about 4-3 Ma (Gilbert et al., 1968). Freshwater lacustrine sediments interbedded with volcanic flows deposited during this time 94 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Columbus Salt Marsh Section 3 EXPLANATION Unit Q'3CM Colluvium Top of bench at 1435 m; beach pebbles and tufa to 1452 m Moderate soil; gypsum in matrix P39-? P38-? P37--R Strong argillic soil P35-? P36--? ? 0 P37--R Siltstone and sandy siltstone Sandstone and silty sandstone, locally conglomerate beds Sandy pebble conglomerate, well rounded clasts (nearshore deposits) Sandy cobble conglomerate, well rounded clasts (beach deposits) Sandstone and con- glomerate, poorly to moderately sorted Steep primary dips (foreset beds) Calcareous cement, locally tufa Buried soil Unconformity Paleomagnetic sample and polarity (?, polarity uncertain) FIGURE 22.?Measured section above borrow pit on southwest side of Colum- bus Salt Marsh (see Figure 1, Table 1), above late Pleistocene shoreline. crop out on the east side of the basin. These deposits are com- posed of diatomaceous sand and silt that contain abundant fos- sils, including molluscs and fish. Lake Russell has a well- known late Pleistocene history, as established by studies per- formed by Russell (1889), Putnam (1949), and Lajoie (1968). The highstand shoreline during the late Wisconsin was at an al- titude of 2155 m; however, the last overflow of Lake Russell southward into Adobe and Owens Valleys (Figure 1) occurred at a highstand of about 2188 m. This lake cycle is thought to be equivalent to the Tahoe glaciation. Reheis et al. (2002) dis- cussed the shoreline altitudes and changing drainage direction of Lake Russell since the late Pliocene. Miller and Smith (1981) and Hubbs and Miller (1948) used fossils from the Mono Basin and modern fish populations in other areas (no fish presently live in Mono Lake) to show that fish had migrated southward into the Mono Basin from the Lahontan Basin, probably during the Pliocene. Fish later mi- grated southward from the Mono Basin into Owens Valley and Death Valley, probably using the overflow route through Adobe Valley (Figure 1). Russell (1889) hinted, and Hubbs and Miller (1948) proposed, that the earlier connection between the Lahontan and Mono Basins had been via the East Walker River and Aurora Valley over a sill at the northeast corner of Lake Russell that is presently at an altitude of 2237 m. Recent discoveries around the north and east sides of the Mono Basin (Figure 1, Table 1) include (1) lacustrine deposits near the former southern outlet of Lake Russell, representing at least one lake cycle that greatly exceeded the 2188 m high shoreline, (2) beach gravels on the north side at altitudes as high as 2268 m at one site and 2252 m at another site, and (3) an abandoned valley, now partially filled by a basalt flow, that lies north of the putative northern sill. The lacustrine deposits near the southern outlet lie at an alti- tude of about 2205-2220 m and include two tephra layers within a coarsening-upward sequence of silt, sand, and beach gravel. Chemical analysis indicated that these tephra (samples EL-53-ML and EL-54-ML) are latest Pliocene in age (2.5-2.0 Ma; Table 3). A lake that at one time stood much higher than these deposits, however, is documented by beautifully rounded beach pebbles, large chunks of tufa-cemented beach rock, and snail coquina that are found on slopes as high as 2243 m in alti- tude directly above the tephra-bearing deposits. Several north- east-striking faults displace these beach-gravel remnants; lo- cally the maximum altitude of the beach pebbles is 2277 m. Beach pebbles within a similar range of altitudes in areas with little evidence of Quaternary faulting (Dohrenwend, 1982b) on the north side of the Mono Basin (Figure 1, Table 1) corrobo- rate the prior occurrence of very high levels of Lake Russell. In addition, the similarity of the beach pebbles and their outcrop appearance (including a preserved berm at one locality) to those of the last two highstands indicates that these deposits are Pleistocene in age. The landscape must have been very differ- ent when Lake Russell reached these altitudes because there is no present-day topographic barrier high enough to enclose a lake for several kilometers east and west of the outcrops near the southern outlet. Gilbert et al. (1968) reported numerous, closely spaced faults that have cut basalt flows from 3 Ma, and examination (this study) of aerial photographs has revealed ev- idence of recent faulting activity. It is likely that faulting acted to lower the terrain in this area, which eventually permitted Lake Russell to drain southward. These old lacustrine deposits around the Mono Basin repre- sent a lake that was high enough to have overflowed northward across the sill in the northeastern corner at its present altitude of 2237 m. It is difficult if not impossible to reconstruct prefault- ing altitudes in this tectonically active part of the basin (Dohren- wend, 1982b); however, the andesite-cored sill locally bears patches of subrounded pebble to cobble gravel, suggesting a former drainage way. A sample of this andesite (EL-KA-2) was dated by conventional K-Ar techniques at 1.44 ? 0.06 Ma (E. McKee, USGS, in litt. to M.C.R., 1995) (Reheis et al., 2002). North of the sill, the drainage is northward into the west- draining Mud Spring Canyon, which is a tributary to the East Walker River (Figure 1). Mud Spring Canyon appears to be an abandoned, underfit drainage; it carries water only intermit- tently, but its floor is as much as 300 m wide and is incised as much as 50 m below old alluvial surfaces. The alluvial depos- its exposed in Mud Spring Canyon contain several thick, coarse tephra beds. Chemical analyses of glass in these beds (samples EL-35A-MS, EL-35B-MS, and EL-36-MS) indicate ages of 2.9-2.5 My (Tables 2, 3). The upper part of the canyon, near NUMBER 33 95 the intersection of the northward drainage from the andesite sill, has been filled twice by basaltic eruptions and subse- quently reincised by discharge from ancient Lake Russell dur- ing the early Pleistocene (Reheis et al., 2002). The high beach gravels, drainage relations, and ages of lake sediments and volcanic flows on the east side of Mono Basin support the following scenario: an ancestral Lake Russell reached higher altitudes than its last overflow level (2188 m) during the late Pliocene and early Pleistocene. At least once during this time, the lake reached an altitude between 2237 m (the elevation of the sill) and 2277 m (the highest preserved beach gravel) and overflowed to the north and thence west down a precursor of Mud Spring Canyon into the East Walker River. Fish fossils (Miller and Smith, 1981) imply that a con- nection to the Lahontan Basin existed during the late Pliocene. The age of the andesite underlying the sill suggests that north- ward overflow also occurred after about 1.5 Ma. Thus, the Mono Basin periodically may have been a tributary to the Walker Lake basin in the late Pliocene and early to middle(?) Pleistocene (Reheis et al., 2002). PLUVIAL LAKE JONATHAN (KOBEH VALLEY).?The large modern drainage basin of Lake Diamond includes several val- leys to the southwest (Figure 1). When Lake Diamond fills and overflows, this drainage basin becomes a tributary to the Hum- boldt River and thus to Lake Lahontan (Tackman, 1993). Most of the drainage basin feeds first into Kobeh Valley, which lies just upstream of Diamond Valley and is separated from it by a very narrow, steep bedrock canyon called Devils Gate (Figure 23A). Lake sediments in Kobeh Valley (Figures 23B, 24) repre- sent an internally drained lake that breached its sill at Devils Gate and became a tributary to Lake Diamond sometime after 670 Ka (Davis, 1987). We propose to name this former lake "Lake Jonathan" for Jonathan O. Davis, who discovered it two years before his untimely death. Lacustrine deposits in Kobeh Valley are extensively exposed in the eastern part of the valley near Lone Mountain, which formed an island in Lake Jonathan (Figures 1, 23B). Reconnaissance field work combined with data from Davis (1987) indicates a long and complex lacustrine history, which is worthy of a far more detailed investigation. Two measured sections expose deposits of at least three lake units. In section 16, normally magnetized sediments of two of these lake units are separated by an unconformity and collu- vium (Figures 24, 25). The upper lake unit contains the Rye Patch Dam ash bed (dated at 670 Ka) (AA-04, Tables 2, 3), which crops out in several other localities near Lone Mountain (samples EL-69-KV, shown in Figure 23c, and EL-70-KV, not in tables). Ostracodes in the ash bed are indicative of alkaline, moderately deep to deep lakes (R. Forester, USGS, in litt. to M.C.R., 1998). This lake unit was the last deposited before the sill was breached, and it is represented by several relatively well-preserved, carbonate-cemented berms (e.g., Figure 23B) in eastern Kobeh Valley. The lower lake unit contains an al- tered tephra layer that we believe is the same as that identified as a Bishop-like ash by Davis (1987); correlation to the Bishop ash is supported by the paleomagnetic data. In section 18, three lake units are separated by unconformi- ties. The two oldest units contain ostracodes that are indicative of alkaline, shallow to moderately deep lakes (R. Forester, USGS, in litt. to M.R.C., 1998). Of these two, the oldest unit is reversely magnetized, whereas the next younger, or middle, unit contains reversely magnetized samples near the base and normally magnetized samples above (Figure 25), suggesting that the unit spans the Matuyama-Brunhes boundary. Sample EL-81-KV taken from a tephra layer in the middle unit, how- ever, is best correlated, on the basis of glass chemistry (Table 3), with one of the several tuffs of Taylor Canyon layers that erupted from Long Valley caldera 2.2-2.1 Ma (Sarna- Wojcicki et al., In review). This age assignment is supported by deposits that are unconformably beneath the youngest lake unit (dated at 0.67 Ma; see section 16, Figure 24) on the east side of Lone Mountain. This lake unit contains a tephra (sam- ple EL-80-KV, Tables 2, 3) that correlates with the Huckle- berry Ridge ash bed (dated at 2.1 Ma). If this correlation is ac- cepted, then (1) the ~14 m of normal-magnetization sediment in the middle lake unit may have been deposited during the Reunion Normal Subchrons within the Matuyama Chron; (2) the lower lake unit in section 16 also may represent this time period, so the correlation to the Bishop ash bed is incorrect; and (3) the unconformity between these deposits and the youngest lake unit represents about 1.5 million years (My) rather than 100 Ky. A third outcrop north of Lone Mountain (Figure 1, Table 1) exposes three deposits of beach gravel separated by unconfor- mities and paleosols; the uppermost gravel belongs to the lake unit dated at 670 Ka. The development of paleosols on the lower two beach gravels is moderate (weak argillic horizons and stage II?III carbonate), and together they probably can ac- count for no more than 200 Ky of subaerial exposure. These re- lationships suggest that these two lake units are no older than early Pleistocene, and that two or more lake units may be miss- ing from section 18. OTHER BASINS IN CENTRAL NEVADA.?Two independent basins in east-central Nevada held Lake Hubbs and Lake New- ark during the latest Pleistocene at shoreline altitudes of about 1902 and 1848 m, respectively (Figure 1, Table 1). Mifflin and Wheat (1979) observed higher shorelines at the south ends of both lakes and attributed them to regional northward tilting; however, berms at similar altitudes on the north and east sides of both basins were located during field reconnaissance for this study in 1997, so northward tilting can be ruled out. Three lev- els of beach gravels that are increasingly indistinct, hence older, with altitude are preserved above the late Pleistocene shorelines of both lakes (Table 1). The highest levels are repre- sented only by rounded beach gravel lagged on bedrock slopes. Relative preservation of berm morphology and properties of relict soils (Figure 26) indicate significant age differences among the lowest three shorelines (including that of the late Pleistocene) with preserved berm crests on the southeast side of Lake Newark. Soil properties, including solum thickness and pedogenic CaCO3 content, suggest that the penultimate 96 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Devils Gate Lone Mountain Location of photo C FIGURE 23.?Photographs of lacustrine depos- its of Lake Jonathan in Kobeh Valley, A, View southeast from top of bench in B on Lone Mountain showing Devils Gate, the bedrock channel that formed when Lake Davis over- topped the divide between it and Lake Dia- mond. B, View northwest showing apron of pale-colored lake sediment capped by beach gravel as high as 1920 m flanking the "island" of Lone Mountain. C, Close-up of these sedi- ments showing thick tephra layer with soft-sed- iment deformation (shovel is about 40 cm long); this layer (EL-69-KV) is identical in appearance to one in another outcrop identified as 670 Ka Rye Patch Dam ash (Tables 2, 3, no. AA-04 JOD) by Davis (1987). shoreline probably occurred at least 100-50 Ka and the next older likely occurred at least 200 Ka by comparison to soils of known age in the Basin and Range area (e.g., Reheis et al., 1995). The soil chronosequence that formed on shorelines of different ages in the Newark Basin has been studied in detail by Redwine (In prep.). Several levels of well-preserved, nested shoreline features in Diamond Valley indicate highstands of pluvial Lake Diamond NUMBER 33 97 Section 16 (U.S. Hwy 50) Section 18 (SW Lone Mtn. EXPLANATION =:^=*? P77--N AA04-JOD; Rye Patch/dam ash bed/~670 ka -N, S5-KV -KV P65B--N? P63A,B~N, S3-KV P60A.B--N, S2-KV P61A--N (Bishop ash? Davis, 1987) S1-KV P79--N P78--N, S10-KV EL-81-KV;tuffof Taylor Canyon, ca. 2.2 Ma P76--R, S9-KV P75--R P74--R, S8-KV Yellowish-gray; shallow lacustrine EL72-KV P73--R P72--R S7-KV; P-71-? Yellowish-gray; cut-and-fill, fluvial S6-KV Mudstone Interbedded mudstone and sandstone Siltstone and sandy siltstone Sandstone and silty sandstone, locally conglomerate beds Sandy pebble conglomerate, well rounded clasts (nearshore deposits) Sandstone and con- glomerate, poorly to moderately sorted -?- _,_ I Calcareous cement, ^1 locally tufa Buried soil and unconformity x x x IEL81-KV Tephra layer and ' sample number -5 -L 0 FIGURE 24.?Measured sections of lacustrine deposits of Lake Jonathan in Kobeh Valley (see Figure 1, Table 1). Symbols are the same as in Figure 9. Question mark (?) after N indicates polarity uncertain. as much as 25 m above its late Pleistocene shoreline (Figure 1; Tackman, 1993). Relict soils indicate that the highest of these is likely several hundred thousand years old (Mifflin and Wheat, 1979); however, the decrease with time in the levels of pluvial Lake Diamond is likely the result of progressive ero- sion of the sill between Diamond Valley and Huntington Creek, a tributary to the Humboldt River. Pine Valley, northwest of Diamond Valley, was an internally drained, fault-controlled basin with a long record of lacustrine sedimentation (Smith and Ketner, 1976; Gordon, 1990). Dur- ing much of its history, probably spanning the late Pliocene and early Pleistocene, the lake was alternately alkaline and rela- tively fresh. Sometime after the deposition of the Lava Creek B ash bed (660 Ka), which lies near the top of the lake beds, the northern sill was breached and the valley began draining into the Humboldt River via Pine Valley Creek. Gordon (1990) in- ferred that Lake Pine was never a deep freshwater lake. His opinion was based upon the alkaline character of the sediments and the reported absence of beach gravels. In field reconnais- sance for the present paper, however, an approximately 10 m thick bed of well-rounded beach gravel, locally steeply bedded, was found at the top of the lacustrine sequence in the northeast- ern corner of the valley. We suggest that Pine Valley, like Kobeh Valley, was breached and integrated with the Humboldt River drainage when an exceptionally large lake spilled over a sill composed of easily eroded Tertiary tuff and sedimentary rocks at, or shortly after, 660 Ka. Discussion AGES OF MIDDLE PLEISTOCENE SHORELINES IN THE WESTERN GREAT BASIN The ages of deposits that mark the very high shorelines through- out the Lahontan Basin and in the basins of Lakes Columbus, Hubbs, and Newark can only be estimated because the units every- where consist of gravel and sand that have not been directly dated. 98 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES C N (UP) W (Hor) W (Hor) S (down) B N (Up) S (down) NRM 5mT W (Hor) NRM W (Hor) 0.002 A/m - - 0.01 A/m P76 S (down) S (down) NRM FIGURE 25.?A-D, Orthogonal demagnetization diagrams (method of Zijderveld, 1967) shown in stratigraphic order for specimens from section 18 in Kobeh Valley. A, reversed magnetization in sample P73 from the oldest exposed lake unit; B, reversed magnetization in sample P76 from the lower part of the middle lake unit; C, normal magnetization in sample P77 from the middle lake unit; D, normal magnetization in sample P78 from the upper part of the middle lake unit. This sample is from just above tephra EL-81-KV, correlated with -2.2 Ma tephra of the tuff of Taylor Canyon (Table 3). The lake deposits of McGee overlie the lake deposits of Campbell deposited about 1 Ma at three sites in the Walker Lake subbasin, and they overlie the lake deposits of Lone Tree deposited about 0.76 Ma (Figure 5). At Sunshine Amphithe- ater, two layers of fan gravel and three argillic paleosols inter- vene between the lake deposits of Campbell and the highest beach gravel, an interval likely representing at least 200 Ky. These relationships and the very high lake level represented by the lake deposits of McGee suggest a correlation to marine ox- ygen-isotope stage (OIS) 16, one of the largest deviations to- ward glacial (and hence pluvial) conditions (Figure 18). We in- terpret the unit to be equivalent to the Rye Patch Alloformation NUMBER 33 99 FIGURE 26.?Photographs of soils formed on beach gravels on south- east side of Lake Newark (see Figure 1, Table 1). A, Late Pleistocene berm at 1845 m (trowel is 28 cm long). B, Berm at 1853 m (divisions on tape = 10 cm intervals). C, Highest berm at 1858 m. Soils are sig- nificantly more developed with age: late Pleistocene soil contains lit- tle pedogenic carbonate; soil formed on next higher berm has stage III carbonate horizon below 55 cm; soil formed on highest berm has strongly cemented Bkm horizon below 60 cm. 100 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES deposited about 0.7-0.6 Ma (Reheis and Morrison, 1997) along the Humboldt River. At this site, sublacustrine deltaic de- posits containing the Rye Patch Dam (-0.67 Ma; A. Sarna- Wojcicki, in litt. to M.C.R., 1997) and Lava Creek B (0.66 Ma; Izett et al., 1992) ash beds crop out at an altitude of about 1285 m (Morrison, 1991). If the lake deposits of McGee (at an alti- tude of 1400 m) correlate with the Rye Patch Alloformation, the deltaic deposits at Rye Patch Dam must represent low-level phases of the lake cycle. The correlation to OIS 16 is bolstered by lake sediments in Kobeh and Pine Valleys (Figure 1). The youngest lake sediments and highest beach gravels of Lake Jonathan in Kobeh Valley are dated at 0.67 Ma by the Rye Patch Dam ash (Figures 23c, 24; Davis, 1987). The youngest lake sediments and highest beach gravels in Pine Valley are dated at 0.66 Ma by the Lava Creek ash bed (Smith and Ketner, 1976; Gordon, 1980; and this study). We infer, based upon the rare occurrence and poor preserva- tion of lag beach gravel on bedrock, that the highest shoreline deposits of Lakes Columbus, Hubbs, and Newark also repre- sent OIS 16 (Figure 18). From its degree of soil development, the highest shoreline of Lake Diamond (Mifflin and Wheat, 1979) also may have formed at this time, possibly as a result of overflow from Lake Jonathan (Tackman, 1993). Beach deposits of the -1370 m and -1350 m high shorelines of Lake Lahontan (the upper and lower lake deposits of Thome, respectively; Figure 5) and of the two intermediate shorelines of Lakes Columbus, Hubbs, and Newark, locally retain the forms of berms and wave-cut benches. In the Walker Lake sub- basin, these beach deposits are inset into older lake units. The relationships of these beach deposits and their morphologic preservation cause us to conclude that the upper and lower lake deposits of Thorne represent units that are younger than those correlated to the Rye Patch Alloformation but are as old or older than the last highstand represented by the upper part of the Eetza Alloformation, which is thought to be equivalent to OIS 6 (Morrison, 1991). Thus, these two sets of shorelines probably represent pluvial episodes corresponding to OIS 14 to perhaps as young as OIS 6 (-530-130 Ka). The upper lake de- posits of Thorne and the higher of the two intermediate shore- lines of the other lakes may be equivalent to OIS 12 (-430 Ka), a relatively large glacial episode in the marine record (Figure 18). This age assignment is consistent with the morphology of relict soils formed on these shorelines at the Thorne Bar site (Adams and Wesnousky, 1999) and at the south end of Lake Newark (Figure 26c). The -1350 m high shoreline is only represented in the Lahontan Basin by the lower lake deposits of Thorne (Figures 5, 18) at Thorne Bar and possibly along the Humboldt River, but other lake basins contain numerous remnants of a lower in- termediate shoreline 7?17 m above the late Pleistocene high- stand (Figure 1, Table 1). Properties of relict soils at Thorne Bar (Adams and Wesnousky, 1999) and in the basins of Lakes Columbus and Newark (Figure 26B) indicate ages from 100 to 50 Ka or earlier. Thus, these shorelines may represent pluvial lakes associated with OIS 6 (an intense glacial episode at about 150-130 Ka) or possibly OIS 8 (a less intense glacial episode at about 250 Ka), but they are not likely to be as old as OIS 10 (-340 Ka; Figure 18). If the lower lake deposits of Thorne are as young as OIS 6, it seems that more remnants of this shore- line should have been found in the Lahontan Basin. Assign- ment of the 1350 m high shoreline of Lake Lahontan to OIS 6 also conflicts with Morrison's (1964, 1991) evidence that the highest Eetza shoreline is at about the same altitude as the 1330 m Sehoo shoreline; however, soil data of Adams and Wes- nousky (1999) suggests that Lake Lahontan shorelines around the 1330 m level may all be Sehoo in age. Recently published uranium-series ages of tufa from the Pyramid Lake area of Lake Lahontan show that the lake achieved at least an inter- mediate level (-1207 m in altitude) almost continuously from about 400 to 170 Ka (Szabo and Bush, 1996) and may have in- termittently stood much higher, but the timing and altitude of the highstands are unknown. CAUSES OF VERY HIGH SHORELINES DURING THE EARLY AND MIDDLE PLEISTOCENE The records of lakes in Nevada described herein stand in contrast to records presently available for lakes in the south- western Great Basin and for Lake Bonneville in the eastern Great Basin; however, the records in the different regions are based upon different types of evidence. For example, interpre- tations of the records of Lakes Owens, Searles, and Bonneville depend largely upon stratigraphic evidence from cores and to a lesser extent from outcrops, whereas those of the present study depend upon shoreline deposits. The records of Lakes Owens and Searles show that these were fresh, overflowing lakes dur- ing periods that correspond with OIS 16, 12, 10, and 6 (Jannik et al., 1991; Smith et al., 1997). Flow-through lakes, however, do not leave adequately detailed evidence of changes in lake size because lake size is controlled by sill height. Oviatt et al. (1999) reinterpreted the middle Pleistocene record of the Bonne- ville Basin (east of the present study) using data from cores and concluded that four freshwater lakes were present during the Brunhes Chron. The oldest of these lakes just preceded the Lava Creek ash (660 Ka), and the younger three were inferred to have occurred during OIS 12, 6, and 2 (Oviatt et al., 1999). In contrast to Pleistocene lakes in northern Nevada, the largest of the Bonneville lakes formed during OIS 2. The large size of this late Pleistocene lake is not directly related to climate; it was probably caused by an increase in water volume from the capture of the Bear River drainage (Bouchard et al., 1998). The descending bathtub-ring character of several Nevada shoreline sequences of this study suggests that pluvial periods became increasingly drier from the early-middle to the late Pleistocene. Climate, tectonics, drainage changes, and changes in moisture sources have probably interacted to produce the ob- served changes in pluvial lake levels in the western Great Ba- sin, as discussed below, but their relative importance is difficult to determine and probably has changed with time. NUMBER 33 101 Changes in the position of the jet stream induced by chang- ing configurations of the Laurentide ice sheet could affect both precipitation and temperature in the western United States. Global circulation models show that the ice sheet induces split- ting of the jet stream, and the movement of the two branches and their associated storm tracks appears to account for pluvial- lake fluctuations in the Basin and Range during the late Pleis- tocene (e.g., Thompson et al., 1993). Previous configurations (especially altitudes) of pre-late Pleistocene ice sheets are too poorly known to estimate their effects, and they may have dif- fered little from that of the late Wisconsin (except in the area west of the Mississippi River; summarized in Hallberg, 1986). Geologists have long believed that uplift of the Sierra Ne- vada since the late Pliocene, estimated to be about 1 km since 3 Ma and 300 m since 0.8 Ma (Huber, 1981; Winograd et al., 1985), has produced an increasing rain shadow to the east through time. In contrast, recent studies have suggested that the range was elevated before the middle Miocene and has since been decreasing in average elevation (Small and Anderson, 1995; Weraicke et al., 1996). Nevertheless, several lines of evi- dence indicate progressive warming and drying of at least the southern Great Basin during the last two million years. These include the depletion in deuterium content and the increase in 818O content of ground water (Winograd et al., 1985, 1997), the decline in the water table (Winograd and Szabo, 1988), and the decrease in pluvial-lake size (this study). In addition, Whitney and Harrington (1993) used relict colluvial deposits in southern Nevada to infer that winter temperatures in the southern Great Basin were at least 1-3? C colder in the early and middle Pleis- tocene than in the late Pleistocene. The apparent discrepancy between the two views of Sierran uplift may be resolved by Small and Anderson's (1995) model. Their model suggests that geomorphic, rather than tectonic, uplift of the Sierra Nevada permitted 30-200 m of uplift per one million years during the Quaternary and that the uplift was in the maximum (as opposed to the average) altitude of slowly eroding parts of the Sierran crest. In addition, there is strong evidence that the Transverse and Cascade Ranges, to the south and north of the Sierra Ne- vada, respectively, also have increased in altitude during this period (Priest, 1990; Hildreth and Lanphere, 1994; Page et al., 1998; Spotila et al., 1998). Nevertheless, relative uplift of the western Cordillera does not account for the extremely high lake levels that we correlate with OIS 16, but it may help to explain the progressive decrease in lake levels since then. Regional-scale lithospheric subsidence may have decreased the average altitudes in the northern Great Basin following up- lift associated with emplacement of the plume head of the Yel- lowstone hot spot centered in northern Nevada (Parsons et al., 1994). This possible effect seems contradicted by recent inter- pretations of fossil flora that suggest elevations in Nevada were much higher at about 15 Ma but had decreased to about the same as the present by about 12 Ma (Wolfe et al., 1997). More moisture may have crossed the Sierra Nevada in the early and early-middle Pleistocene because of the presence of Lake Clyde, a tectonically dammed lake that filled the Great Valley of California (Sarna-Wojcicki, 1995). Pacific air masses that gave up moisture while passing over the Coast Ranges could have been recharged by water that evaporated from this large lake (30,000-50,000 km2). The added moisture could have increased the snow pack on the Sierra Nevada and on ranges in the Great Basin farther east. Lake Clyde overtopped its sill, rapidly incised its outlet, and drained shortly after 660 Ka (Sarna-Wojcicki, 1995). This moisture effect helps explain the high lake levels of OIS 16, but it does not account for the intermediate shoreline altitudes of pluvial Lakes Lahontan, Co- lumbus, Newark, Hubbs, and possibly Gale that formed be- tween OIS 16 and 2. Drainage changes have altered the size of the Lahontan drainage basin through time. For example, Lake Russell prob- ably discharged into the Walker River in the late Pliocene and early Pleistocene (Figure 1; Reheis and Morrison, 1997; Reheis et al., 2002). The addition of water from Lake Russell could help to account for high lake levels during the early Pleistocene in the Walker Lake subbasin, if it were not integrated with the Lahontan Basin to the north during part of this period. Other lake basins that are now tributaries to Lake Lahontan were previously isolated. Smith and Ketner (1976) showed that Pine Valley (PV, Figure 1), presently a tributary to the Hum- boldt River, was internally drained until sometime after about 660 Ka (following deposition of the Lava Creek B ash bed). Davis (1987) suggested that the lake in Kobeh Valley (Lake Jonathan) was internally drained until sometime after about 670 Ka (following deposition of the Rye Patch Dam ash bed). Recent field work (this study) confirms this proposal and sug- gests that Lakes Jonathan and Diamond (Figure 1) became inte- grated with each other and with the Humboldt River when Lake Jonathan spilled over its eastern sill shortly after 670 Ka. The sudden appearance of endemic ostracodes (sample S13-LT, Figure 15) in lake deposits of Lone Tree at Lone Tree Mine could represent a rapid influx of water from a long-lived, stable lake not previously part of the drainage basin (R. Forester, USGS, in litt. to M.C.R., 1998). The newly discovered highest shorelines of Lakes Hubbs and Newark (Table 1) are very close to modern sill levels between these two basins and between the Newark and Lahontan Basins (Figure 18); thus, it is possible that these lakes were temporarily tributaries to Lake Lahontan during this old (OIS 16?) lake cycle. Beach gravel at elevations 3-7 m higher than the late Pleistocene shoreline of Lake Gale, and at the same altitude as its present northern sill (Figure 1, Table 1), indicate that this lake once discharged northward into the Lake Franklin basin, as previously suggested by Hubbs et al. (1974). In addition, other pluvial lakes in east-central Ne- vada that were isolated during the late Pleistocene, such as Lakes Gilbert, Franklin, and Clover (Figure 1; Mifflin and Wheat, 1979), could have potentially discharged into the Lahontan Basin if their water levels were about 10?40 m above their late Pleistocene levels (estimated from modern topogra- phy and not accounting for changes caused by faulting and sed- imentation). These three basins as well as the Lake Gale basin presently contain relict fish populations that indicate former 102 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES drainage connections with the Lahontan Basin via the Hum- boldt River, as discussed below (Hubbs and Miller, 1948). Gravel pits north of the northeastern corner of Lake Clover ex- pose thick deposits of probable beach gravel that indicate the lake level was as much as 32 m above the late Pleistocene high- stand (Table 1). In addition, gravel pits on the west side of Wells, Nevada, expose a 5-10 m thick sequence of deltaic gravel and sand beds, including steeply dipping foreset beds. These beds lie directly north of the lowest sill around the Lake Clover basin and are interpreted to indicate overflow of the Lake Gale-Lake Franklin-Lake Clover chain of lakes into a small basin around Wells that is now the headwaters of the Humboldt River (Reheis, 1999). Other overflows and basin connections are also possible (Figure 1). IMPLICATIONS FOR MIGRATION OF AQUATIC SPECIES IN THE WESTERN GREAT BASIN The evidence presented in this paper for connections be- tween presently disjunct lake basins permits geological and chronological constraints to be placed on the Pliocene and Qua- ternary dispersion of fish and molluscs in the western Great Ba- sin. The pioneering work of Hubbs and Miller (1948) and sub- sequent refinements by Hubbs et al. (1974), Miller and Smith (1981), and Smith et al. (2002) established that many of these isolated basins contain (or once contained) species of relict fish that are related to each other, to species in the modern Lahontan Basin, and to species in the Columbia River basin to the north and the Owens Lake-Death Valley area to the south. Similar studies of springsnails (Hershler and Sada, 2002) and leeches (Hovingh, 1997) have suggested similar connections, although many of the migrations are probably much older than the pe- riod of time discussed in this paper. Research conducted in the last 20 years on material from fossil localities has shown that many modern species of fish and springsnails are very similar to ancestral species that were widely distributed in western North America during the Pliocene. As a result, speciation pre- viously attributed to rapid evolution since about 12 Ka (Hubbs and Miller, 1948) is now more commonly attributed to inherit- ance from widespread Tertiary populations followed by extinc- tions and intermittent population-mixing caused by wet-dry climatic cycles and tectonic disruptions during the Pleistocene (e.g., Miller and Smith, 1981; Taylor and Smith, 1981; Taylor, 1985; Minckley et al., 1986; Hershler and Sada, 2002; Smith et al., 2002). This paper documents large, previously unknown lakes in the western Great Basin that were intermittently connected during the early and middle Pleistocene. This long lacustrine history may allow the observed divergence between some species to be averaged over half a million years rather than over several thousand years, thereby greatly reducing the apparent discrep- ancy in evolutionary rates. At the least, this new information provides alternative hypotheses about rates of evolution from independent evidence, as urged by Smith et al. (2002). In some cases, these older Pleistocene lakes may require revision of pre- vious conclusions that were based upon incorrect age assign- ments. For example, Taylor and Smith (1981) reported on fossil snails and fish from an outcrop in the Mopung Hills in the Lahontan Basin. They noted that the fish were probably ances- tral to Lahontan forms but were unlike known Pliocene fossils from the Honey Lake area; therefore, they postulated that a bar- rier had existed between the two subbasins. These lacustrine deposits in the Mopung Hills had been assigned a Pliocene age by Willden and Speed (1974); lacking chronologic dating, these authors assumed incorrectly that indurated lacustrine de- posits older than the late Pleistocene Lake Lahontan sediments must be Pliocene in age! Similar deposits previously mapped as Pliocene at McGee Wash and Sunshine Amphitheater (Greene et al., 1991) are now known to represent early to middle Pleis- tocene deposits of Lake Lahontan. PLIOCENE AND EARLY PLEISTOCENE CONNECTIONS.?Late Miocene to latest Pliocene lacustrine deposits that fill the present divide between Rhodes and Columbus Salt Marshes (Figures 1, 20) reflect the presence of a former waterway. This connection could have allowed fish to migrate from the Lahon- tan Basin into Fish Lake Valley, where a relict population of a Lahontan tui chub, Siphateles bicolor obesus, survived until historic settlement of the valley (Hubbs and Miller, 1948). Thus, this fish could have migrated into Fish Lake Valley no later than about 2 Ma, after which movement along the Huntoon Valley fault system disrupted the ancient lake basin. Siphateles b. obesus and Rhinichthys osculus lariversi Lugaski, 1972 (a speckled dace related to the Lahontan dace; Smith et al., 2002), are found in the basin of Lake Toiyabe, in northern Big Smoky Valley, and may have arrived by the same route. Al- ternatively, Hubbs and Miller (1948) postulated that these fish could have migrated southward into Big Smoky Valley from Grass Valley (pluvial Lake Gilbert, Figure 1) by diversion of tributary streams in the divide area between these two valleys. The modern distribution of a springsnail (Pyrgulopsis gibba Hershler, 1995) in the Lahontan Basin, Grass Valley, and Big Smoky Valley supports the latter route (Hershler and Sada, 2002). The combination of high remnants of lacustrine deposits con- taining Lake Lahontan-related fish fossils (Miller and Smith, 1981), drainage relationships, and ages of sediments and flows on the east side of the Mono Basin support the interpretation that the basin was a tributary northward to the Walker Lake ba- sin (Figure 1) in the late Pliocene and early to middle(?) Pleis- tocene (Reheis et al., 2002). Sometime after 1.3 Ma (the age of basalt filling a channel that carried lake discharge northward), this outlet was abandoned and highstands of Lake Russell be- gan to overflow to the south, permitting fish to move into the Owens River drainage. The creation of the new outlet probably was caused by intense faulting in the area of the southern out- let, as evidenced by mapping studies conducted during this study. The relatively youthful appearance of many of the fault scarps suggests that this faulting may have followed the erup- tion of the Bishop ash (760 Ka) (Sarna-Wojcicki et al., 1991) from Long Valley caldera (Bailey et al., 1976); several studies NUMBER 33 103 have concluded that the rates of vertical slip along fault zones in the area increased markedly after this eruption (dePolo, 1989; Gillespie, 1991; Reheis and Sawyer, 1997). A deep core taken on the floor of Owens (dry) Lake indicates that Siphate- les bicolor, Prosopium sp. (whitefish), and Oncorhynchus cf. clarki (trout) were present in Owens Lake by 730 Ka (Firby et al., 1997); however, fossils identified as Chasmistes (lake sucker), similar to C. batrachops Cope, 1883, from localities in Oregon and the Snake River Plain, were reported in sediments of early Pleistocene age near China Lake, south of Owens Lake (Miller and Smith, 1981). These fossils suggest that some fish were spread widely along the east slope of the Sierra Nevada by the late Pliocene (Reheis et al., 2002). Three species of springsnails in Owens Valley are most closely related to a Lahontan species (Hershler and Sada, 2002); they may have ar- rived by migration via the Mono Basin, as discussed above. Another species {Pyrgulopsis wongi Hershler, 1995) lives in several isolated basins along the eastern side of the Sierra Ne- vada south of the Carson River (Figure 1), suggesting the oc- currence of a widespread pre-Quaternary population that was subsequently isolated by tectonic disruptions. MIDDLE PLEISTOCENE CONNECTIONS.?The implications of Lake Lahontan having a shoreline at an altitude of 1400 m, herein estimated to have occurred about 650 Ka (within OIS 16), are staggering (Figure 1). For example, such a lake would have (1) inundated Granite Springs and Kumiva Valleys, (2) backed up the Truckee and Carson Rivers to submerge all or part of present-day Reno and Carson City, Nevada, (3) backed up the Humboldt River at least to present-day Battle Mountain, and (4) extended 60 km southeast from Walker Lake to Rhodes Salt Marsh. If allowances are made for local tectonics and for sedimentation in the last half million years, the lake also would have flooded into Dixie and Fairview Valleys and would have inundated many other flat-lying valleys whose floors are now only 5-30 m above the 1400 m level. Such projections allow specific hypotheses to be proposed and tested. For example, Hubbs and Miller (1948) postulated an "early pluvial" connection between Lake Dixie and Lake Lahontan (Figure 1) on the basis of a subspecies of S. b. obesus that inhabits a present-day spring in Pleasant Valley, between Dixie Valley and Grass Valley (an arm of Lake Lahontan). The modern distribution of the springsnail P. gibba also supports this connection (Hershler and Sada, 2002). As shown in Figure 1, the 1400 m level of Lake Lahontan probably would have flooded Dixie, Fairview, and Pleasant Valleys (the present low- est sill height is only 8 m above this lake level; Reheis and Morrison, 1997), providing easy access for migrating fish. An enigmatic outcrop in eastern Fairview Valley provides support- ing evidence for such a connection. At this site, possible beach gravel extends up to an altitude of 1384 m; this gravel rests un- conformably on fine-grained lacustrine deposits containing tephra of probable early Pliocene age (samples EL-59-FV and EL-66-FV, Tables 2, 3). Further searches should be made in Dixie and Fairview Valleys for better sites to confirm their pos- tulated connection to Lake Lahontan. It is also likely that Lake Lahontan, while at its 1400 m high- stand about 650 Ka, flooded Granite Springs and Kumiva Val- leys. These valleys are located within a large "island" in the center of the Lahontan Basin, and each held a small, isolated pluvial lake during the Sehoo highstand of Lake Lahontan (Fig- ure 1; Mifflin and Wheat, 1979). No search for deposits of the 1400 m highstand has been done in this internally drained area; however, pre-late Pleistocene inundation is supported by the existence of leeches (Erpobdella sp.) in present-day springs within the "island" that are identical to species in springs of Ea- gle Lake, Madeline Plains, and Winnemucca Lake (P. Hovingh, Utah Museum of Natural History, in litt. to M.C.R., 1997) within the Lahontan Basin. The springsnail Pyrgulopsis gibba also lives in both the Lahontan and Granite Springs Basins (Hershler and Sada, 2002). Several basins in central Nevada show evidence of much higher lake levels in the past. Results of the present study com- bined with previous research show that two of these basins, Kobeh Valley (Davis, 1987; Tackman, 1993) and Pine Valley (Smith and Ketner, 1976), were internally drained basins con- taining pluvial lakes until at least about 670 and 660 Ka, re- spectively, and that they subsequently integrated with the Lahontan Basin. It is likely that shortly after 660 Ka, very high stands of Lake Jonathan (Kobeh Valley), and perhaps of Lake Pine (Pine Valley), triggered overflows that caused the integra- tion of these drainage basins with the Lahontan Basin. If so, fish presently living in these valleys, including populations of S. b. obesus and R. osculus (Hubbs et al., 1974), may not have arrived until after 660 Ka. Hubbs et al. (1974) proposed that Lakes Hubbs and Newark may have once been connected to the Humboldt River (Figure 1) by overflow from Lake Hubbs to Lake Newark through Beck Pass and thence from Lake Newark northward to Huntington Creek. Their best evidence for overflow of Lake Newark came from the presence of two subspecies of S. bicolor in modern springs of Newark Valley; no fish were found in Long Valley (Lake Hubbs). They observed what they considered to be wave- cut terraces nearly up to the height of the sill (1891 m in alti- tude) at the north end of Newark Valley. Although Mifflin and Wheat (1979) reported no shorelines this high, field work for the present study found definite beach gravel as high as 1887-1890 m in altitude at the north end of Newark Valley, only 1-4 m below the modern divide (Table 1); there is also a large drainage channel leading north into Huntington Creek that begins just below the summit. Beach gravel is preserved at least as high as about 1927 m in Long Valley, which is 25 m above the late Pleistocene shoreline of Lake Hubbs, and possibly as high as 1940 m, which is about 6 m below the modern sill at Beck Pass. The highest shorelines of Lake Hubbs and Lake Newark may have formed at the same time as those of the 1400 m highstand of Lake Lahontan (Figure 1) and of Lake Jonathan and Lake Pine, i.e., at about 650 Ka (Figure 18). The highest shoreline of Lake Newark is so close to the present northern di- vide that it is very likely that the lake overflowed northward into Huntington Creek at this time, providing a migration route 104 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES for the fish presently in Newark Valley. We speculate that Lake Hubbs may have overflowed into Lake Newark at the same time. These hypotheses can be tested by more detailed field work and mapping. We further speculate that other isolated fish-bearing basins in central and eastern Nevada could have been populated by fish and springsnails during pre-late Pleistocene overflows (Figure 1). These basins held Lakes Gilbert (Grass Valley: R. osculus, P. gibba), Gale (Butte Valley: Relictus solitarius Hubbs and Miller, 1972), Franklin (Ruby Valley: R. solitarius, Pyrgulopsis leporina Hershler, 1998), and Clover {R. osculus and S. bi- color) (Hubbs et al., 1974; Hershler and Sada, 2002; Smith et al., 2002). We acknowledge, however, that some of the fish species?especially R. solitarius?are so different from their nearest relatives in the Lahontan Basin that they could have re- sulted from divergence from fish that were widely distributed across Nevada in the Pliocene, when conditions were much wetter and drainage basins were configured differently (For- ester, 1991). The modern divide between Grass Valley and Crescent Valley (part of the Lahontan drainage basin) is 36 m above the late Pleistocene shoreline of Lake Gilbert (Figure 1; Mifflin and Wheat, 1979). As noted by Hubbs et al. (1974), the altitude of this divide has probably been affected by faulting. Lakes Gale, Franklin, and Clover form a chain of lakes with very low divides that are less than 15m above their late Pleis- tocene shorelines (Figure 1). Both Hubbs et al. (1974) and Mif- flin and Wheat (1979) noted the presence of a channel leading north from the present divide between Butte and Ruby Valleys; reconnaissance for the present study found beach gravels at al- titudes at and below the altitude of the lip of this channel in northern Butte Valley (Figure 1, Table 1). These authors ex- plained that overflow from Lake Gale to Lake Franklin in pre- late Pleistocene time enabled the same fish to inhabit these two lakes. The divide between Ruby and Clover Valleys is even lower, and it seems very likely that Lake Franklin also over- flowed into Lake Clover, although Hubbs et al. (1974) found no evidence of a connecting channel and thought that such an overflow had not occurred because of the disparity in fish spe- cies between Ruby and Clover Valleys. Nevertheless, R. soli- tarius, which apparently evolved from a Lake Lahontan-type dace (Hubbs et al., 1974), had to have reached Ruby and Butte Valleys by some means, and the obvious route was by discharge from Lake Franklin to Lake Clover and thence northward into the Humboldt River. Recent discovery of probable beach gravel that lies 32 m above the late Pleistocene shoreline of Lake Clo- ver as well as deltaic deposits at Wells, just north of the present lowest sill of Lake Clover (Table 1), suggests that a discharge had occurred. Such a connection is also supported by the occur- rence of the springsnail P. leporina in both Ruby Valley and the upper Humboldt River drainage (Hershler and Sada, 2002). Summary The large study area and the amount of detail presented in this paper is dauntingly complex. In the following summary of results, we present a speculative but coherent framework of hy- potheses that provides context for the results and a springboard for future integrative studies to test the hypotheses. We encour- age other researchers to pursue these tests. During the Pliocene to middle Pleistocene, pluvial lakes in the western Great Basin repeatedly rose to levels much higher than those of the well-documented later Pleistocene. Early mi- grations of aquatic species were facilitated by connections among Pliocene lakes, which were subsequently disrupted by tectonic movements. The geologic evidence for these Pliocene connections supports recent work summarized by Smith et al. (2002) and by Hershler and Sada (2002), indicating that many of the modern distributions of aquatic species result from pre- Quaternary dispersions. The parallel histories of Pliocene lakes in Rhodes Salt Marsh, Clayton Valley, and Fish Lake Valley, combined with the lack of sills or very low sills between Columbus Salt Marsh and Clayton, Fish Lake, and Big Smoky Valleys, support the inter- pretation that these lakes were interconnected during the Pliocene. In addition, we infer that a Miocene to Pliocene lake basin that at least included Rhodes and Columbus Salt Marshes was disrupted by faulting and (or) compression. The timing of these connections indicates that fish and possibly springsnails must have migrated into Fish Lake Valley from the north before about 2 Ma. The populations of fish and springsnails in north- ern Big Smoky Valley may have arrived by the same route or across a higher divide from Grass Valley (Lake Gilbert) to the north (Hubbs and Miller, 1948), or they may be relict from older, more widespread ancestral populations (e.g., Taylor, 1985;Minckleyetal., 1986). The combination of high remnants of lacustrine deposits con- taining Lake Lahontan-related fish fossils (Miller and Smith, 1981), drainage relationships, and ages of sediments and flows on the east side of the Mono Basin indicate that the Mono Basin periodically was a tributary to the Walker River in the late Pliocene and early Pleistocene, permitting Lake Lahontan fish to enter the Mono Basin. Some of these fish then migrated southward into the Owens River drainage when a new, lower outlet opened sometime after 1.3 Ma and possibly as recently as after 760 Ka. Some fish, however, may have already reached the area south of Owens Lake before the early Pleistocene (Miller and Smith, 1981). Several outcrops in the Lahontan Basin document highstands of a lake that rose far above the Sehoo and Eetza(?) shorelines during the early and middle Pleistocene. A conservative corre- lation among all the known sites yields a total of seven lacus- trine units whose shorelines exceeded that of the Sehoo lake; three of the units postdate the lake unit containing a tephra pro- visionally correlated to the Bishop ash bed dated from 0.76 Ma, and three predate the tephra-containing unit but are younger than 2.6 Ma. The Walker Lake subbasin became integrated with the northern Lahontan subbasins probably by 1 Ma (lake depos- its of Campbell) but possibly as recently as about 650 Ka (lake deposits of McGee). The youngest three units represent middle Pleistocene shorelines stepped progressively downward with time. The lake deposits of McGee, representing the highest Lake Lahontan shoreline preserved at an altitude of 1400 m, is correlated with the Rye Patch Dam Alloformation from 670 to 630? Ka (Morrison, 1991) and with marine OIS 16. The younger two shorelines, at altitudes of about 1370 and 1350 m, are tentatively correlated with OIS 12 (-430 Ka) and OIS 10 (-340 Ka) or younger glacial stages, respectively. NUMBER 33 105 Similar records of undated shorelines descending from an early-middle Pleistocene highstand are preserved in the basins of Lakes Columbus-Rennie, Hubbs, and Newark. Very high stands also were achieved shortly after about 660 Ka by Lake Jonathan and Lake Pine. The geographic distribution of these lakes indicate that the causes of the very high stands were not limited in their effects to western Nevada; how far east the higher lake levels of middle Pleistocene age can be found is un- known, but they have not been reported in the Bonneville Basin of Utah (Oviatt et al., 1999). Complex interactions among cli- mate, tectonics, drainage changes, and moisture sources are re- sponsible for the observed differences in pluvial lake levels in the western Great Basin. These interactions could be tested by global circulation models. At about 650 Ka, pluvial lakes in Nevada reached such high levels that numerous interbasin thresholds were likely over- topped. In the cases of Lake Jonathan and Lake Pine, these overflows caused the basins to become permanently or periodi- cally part of the Lahontan drainage basin. In other cases, such as Granite Springs and Dixie Valleys, Lakes Hubbs and New- ark, and perhaps simultaneously Lakes Gilbert, Gale, Franklin, and Clover, the overflows were temporary. These connections provided potential migration routes for relict populations of Lake Lahontan-type fishes and springsnails that presently re- side in internally drained basins, especially those species and subspecies that apparently have diverged little from their mod- ern relatives in the Lahontan Basin. This scenario presents a startling vision of a western Great Basin in which climatic and tectonic effects conspired to integrate presently isolated basins into one enormous drainage system during at least one brief pe- riod during the Quaternary. Much work remains to be done to test this hypothetical vision. Literature Cited Adams, K.D. 1996. Field Trip Guide. In K.D. Adams and S.A. Fontaine, editors, Quater- nary History, Isostatic Rebound and Active Faulting in the Lake Lahontan Basin, Nevada and California, pages 5-56. Reno, Nevada: Friends of the Pleistocene, Pacific Cell, 1996 Fall Field Trip Guide- book. 1997. Late Quaternary History, Isostatic Rebound, and Active Faulting in the Lake Lahontan Basin, Nevada and California. 169 pages. Doc- toral dissertation, University of Nevada, Reno. Adams, K.D., and S.G. Wesnousky 1999. The Lake Lahontan Highstand: Age, Surficial Characteristics, Soil Development, and Regional Shoreline Correlation: Geomorphology, 30:357-392. Bailey, R.A., G.B. Dalrymple, and M.A. Lanphere 1976. Volcanism, Structure, and Geochronology of Long Valley Caldera, Mono County, California. Journal of Geophysical Research, 81(5): 725-744. Bassinot, F.C., L.C. Labeyrie, E. Vincent, X. Quidelluer, N.J. Shackleton, and Y. Lancelot 1994. The Astronomical Theory of Climate and the Age of the Brunhes- Matuyama Magnetic Reversal. Earth and Planetary Science Letters, 126:91-108. Benson, L.V. 1978. Fluctuation in the Level of Pluvial Lake Lahontan during the Last 40,000 Years. Quaternary Research, 9:300-318. Benson, L.V, D.R. Currey, R.I. Dorn, K.R. Lajoie, C.G. Oviatt, S.W. Robinson, G.I. Smith, and S. Stine 1990. Chronology of Expansion and Contraction of Four Great Basin Lake Systems during the Past 35,000 Years. Palaeogeography, Palaeocli- matology, Palaeocology, 78:241-286. Benson, L.V., J.P. Smoot, M. Kashgarian, A. Sarna-Wojcicki, and J.W. Burdett 1997. Radiocarbon Ages and Environments of Deposition of the Wono and Trego Hot Springs Tephra Layers in the Pyramid Lake Subbasin, Nevada. Quaternary Research, 47:251-260. Bouchard, D.P., D.S. Kaufman, A. Hochberg, and J. Quade 1998. Quaternary History of the Thatcher Basin, Idaho, Reconstructed from the 87Sr/86Sr and Amino Acid Composition of Lacustrine Fos- sils: Implications for the Diversion of the Bear River into the Bonne- ville Basin. Paleogeography, Paleoclimatology, Palaeoecology, 141:95-114. Bradbury, J.P., R.M. Forester, and R.S. Thompson 1989. Late Quaternary Paleolimnology of Walker Lake, Nevada. Journal ofPaleolimnology, 1:249-267. Cande, S.C., and D.V. Kent 1995. Revised Calibration of the Geomagnetic Polarity Timescale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research, 100(B4):6093-6095. Chadwick, O.A., and J.O. Davis 1990. Soil-Forming Intervals Caused by Eolian Sediment Pulses in the Lahontan Basin, Northwestern Nevada. Geology, 18:243-246. Davis, J.O. 1978. Tephrochronology of the Lake Lahontan Area, Nevada and Califor- nia. Nevada Archaeological Research Papers, 7:137. 1982. Bits and Pieces: The Last 35,000 Years in the Lahontan Area. In D.B. Madsen and J.F. O'Connell, editors. Man and Environment in the Great Basin, volume 2:53-75. Washington, D.C.: Society for American Archeology. 1985. Correlation of Late Quaternary Tephra Layers in a Long Pluvial Se- quence near Summer Lake, Oregon. Quaternary Research, 23: 38-53. 1987. Tephrochronology of Middle Pleistocene Lake Sediments in Kobeh Valley, Nevada: Implications for Neotectonics and Drainage History of the Central Great Basin. Geological Society of America Abstracts with Programs, 19(6):370. Davis, J.R. 1981. Late Cenozoic Geology of Clayton Valley, Nevada, and the Genesis of a Lithium-Enriched Brine. 233 pages. Doctoral dissertation, Uni- versity of Texas at Austin. Demsey, K. 1987. Holocene Faulting and Tectonic Geomorphology along the Wassuk Range, West-Central Nevada. 64 pages. Master's thesis, University of Arizona, Tucson. dePolo, CM. 1989. Seismotectonics of the White Mountains Fault System, East-Central California and West-Central Nevada. 354 pages. Master's thesis, University of Nevada, Reno. dePolo, CM., W.A. Peppin, and P.A. Johnson 1993. Contemporary Tectonics, Seismicity, and Potential Earthquake Sources in the White Mountains Seismic Gap, West-Central Nevada and East-Central California. Tectonophysics, 225:271-299. 106 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Dohrenwend, J.C. 1982a. Preliminary Surficial Geologic Map of the Excelsior Mountains Area, West-Central Nevada. United States Geological Survey Mis- cellaneous Field Studies, Map MF-I372. 1982b. Map Showing Late Cenozoic Faults in the Walker Lake 1? by 2? Quadrangle, Nevada-California. United States Geological Survey Miscellaneous Field Studies, Map MF-13 82-D. Firby, J.R., S.E. Sharpe, J.F. Whelan, G.R. Smith, and W.G. Spaulding 1997. Paleobiotic and Isotopic Analysis of Mollusks, Fish, and Plants from Core OL-92: Indicators for an Open or Closed Lake System. In G.I. Smith, and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Spe- cial Papers of the Geological Society of America, 317:121-126. Forester, R.M. 1991. Pliocene-Climate History of the Western United States Derived from Lacustrine Ostracodes. Quaternary Science Reviews, 10:133?146. Gilbert, CM., M.N. Christensen, Y. Al-Rawi, and K.R. Lajoie 1968. Structural and Volcanic History of Mono Basin, California-Nevada. Memoirs of the Geological Society of America, 16:275?329. Gillespie, A.R. 1991. Quaternary Subsidence of Owens Valley, California. In C.A.J. Hall, V. Doyle-Jones, and B. Widawski, editors, Natural History of East- ern California and High-Altitude Research. Proceedings of the White Mountain Research Station Symposium, 3:356-382. Gobalet, K.W., and R.M. Negrini 1992. Middle Pleistocene Fish Fossils from the Lahontan Basin. [Ab- stract.] Paleobios, 14 (supplement to no. l):2-3. Gordon, I.R. 1990. Sedimentation and Structure of the Pine Valley Basin, Northeast Ne- vada. 133 pages. Master's thesis, University of Wyoming, Laramie. Greene, R.C., J.H. Stewart, D.A. John, R.F. Hardyman, NJ. Silberling, and M.L. Sorenson 1991. Geologic Map of the Reno 1? by 2? Quadrangle, Nevada and Cali- fornia. United States Geological Survey Miscellaneous Field Studies Map, MF-2154-A. Hallberg, G.R. 1986. Pre-Wisconsin Glacial Stratigraphy of the Central Plains Region in Iowa, Nebraska, Kansas, and Missouri. Quaternary Science Re- views, 5:11-15. Herrero-Bervera, E., C.E. Helsley, A.M. Sarna-Wojcicki, K.R. Lajoie, C.E. Meyer, M.O. McWilliams, R.M. Negrini, B.D. Turrin, J.M. Donnelly- Nolan, and J.C. Liddicoat 1994. Age and Correlation of a Paleomagnetic Episode in the Western United States by 40Ar/39Ar Dating and Tephrochronology: The Ja- maica, Blake, or a New Polarity Episode? Journal of Geophysical Research, 99(B12):24,091-24,103. Hershler, R., and D.W. Sada 2002. Biogeography of Great Basin Aquatic Snails of the Genus Pyrgulop- sis. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Ba- sin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:255-276. Hildreth, W., and M.A. Lanphere 1994. Potassium-Argon Geochronology of a Basalt-andesite-dacite Arc System: The Mount Adams Volcanic Field, Cascade Range of Southern Washington. Bulletin of the Geological Society of America, 106:1413-1429. Hovingh, P.W. 1997. Paleozoological and Zoogeographical Studies of Leeches (Hirudinea) in the Great Basin Region of Western United States. [Abstract.] 4 unnumbered pages. Symposium on Great Basin Aquatic Systems History, Salt Lake City, Utah, 1997. Hubbs, C.L., and R.R. Miller 1948. The Great Basin, II: The Zoological Evidence. Bulletin of the Uni- versity of Utah, 38:17-166. Hubbs, C.L., R.R. Miller, and L.C. Hubbs 1974. Hydrographic History and Relict Fishes of the North-Central Great Basin. Memoirs of the California Academy of Sciences, 7:259. Huber, N.K. 1981. Amount and Timing of Late Cenozoic Uplift and Tilt of the Sierra Nevada, California?Evidence from the Upper San Joaquin River Basin. United States Geological Survey Professional Paper, 1197:1-28. Imbrie, J., A. Berger, E.A. Boyle, S.C. Clemens, A. Duffy, W.R. Howard, G. Kukla, J. Kutzbach, D.G. Martinson, A. Mclntyre, A.C. Mix, B. Molfino, J.J. Morley, L.C. Peterson, N.G. Pisias, W.L. Prell, M.E. Raymo, N.J. Shackleton, and J.R. Toggweiler. 1993. SPECMAP Archive #4, IGPB PAGES/World Data Center-A for Pal, Data Contribution Series 93-031. NOAA/NGDC Paleoclimatology Program, Boulder, Colorado. [This is a World Wide Web database located at ftp.://ftp.ngdc.noaa.gov/paleo/paleocean/specmap/ specmap4/specmap4.readme; accessed 14 July 1997.] Izett, G.A. 1981. Volcanic Ash Beds: Recorders of Upper Cenozoic Silicic Pyroclastic Volcanism in the Western United States. Journal of Geophysical Re- search, 86(B11): 10,200-10,222. Izett, G.A., K.L. Pierce, N.D. Naeser, and C. Jaworowski 1992. Isotopic Dating of Lava Creek B Tephra in Terrace Deposits along the Wind River, Wyoming?Implications for Post 0.6 Ma Uplift of the Yellowstone Hotspot. Geological Society of America Abstracts with Programs, 24(7):A102. Izett, G.A., and R.E. Wilcox 1983. Map Showing Localities and Inferred Distributions of the Huckle- berry Ridge, Mesa Falls, and Lava Creek Ash Beds (Pearlette Fam- ily Ash Beds) of Pliocene and Pleistocene Age in the Western United States and Southern Canada. U.S. Geological Survey Miscellaneous Investigations, Map I-1325. Jannik, N.O., F.M. Phillips, G.I. Smith, and D. Elmore 1991. A 36C1 Chronology of Lacustrine Sedimentation in the Pleistocene Owens River System. Bulletin of the Geological Society of America, 103:1146-1159. King, G.Q. 1978. The Late Quaternary History of Adrian Valley, Lyon County, Ne- vada. 88 pages. Master's thesis, University of Utah, Salt Lake City. 1993. Late Quaternary History of the Lower Walker River and Its Implica- tions for the Lahontan Paleolake System. Physical Geography, 14:81-96. Knott, J.R., A.M. Sarna-Wojcicki, C.E. Meyer, J.C.I.I.I. Tinsley, E. Wan, and S.G. Wells 1996. Late Neogene Stratigraphy of the Black Mountains Piedmont, East- ern California: Implications for the Geomorphic and Neotectonic Evolution of Death Valley. Geological Society of America Abstracts with Programs, 28(5):82. Kunasz, I.A. 1970. Geology and Geochemistry of the Lithium Deposit in Clayton Val- ley, Esmeralda County, Nevada. 114 pages. Doctoral dissertation, Pennsylvania State University, State College. Lajoie, K.R. 1968. Late Quaternary Stratigraphic and Geologic History of Mono Basin, Eastern California. 271 pages. Doctoral dissertation, University of California at Berkeley. Mifflin, M. 1984. Paleohydrology of the Lahontan Basin. In J. Lintz, Jr., editor, West- ern Geological Excursions: Guidebook for the 1984 Annual Meeting of the Geological Society of America, volume 1:134-137. Reno, Ne- vada: Geological Society of America and the Department of Geolog- ical Sciences, Mackay School of Mines, University of Nevada, Reno. Mifflin, M.D., and M.M. Wheat 1979. Pluvial Lakes and Estimated Pluvial Climates of Nevada. Bulletin of NUMBER 33 107 the Nevada Bureau of Mines and Geology, 94: 57 pages. Miller, R.R., and G.R. Smith 1981. Distribution and Evolution of Chasmistes (Pisces: Catostomidae) in Western North America. Occasional Papers of the Museum of Zool- ogy, University of Michigan, 696:1-46. Minckley, W.L., D.A. Hendrickson, and C.E. Bond 1986. Geography of Western North American Fishes: Description and Re- lationships to Intracontinental Tectonism. In C.H. Hocutt and E.O. Wiley, editors, Zoogeography of North American Freshwater Fishes, pages 519-613. New York: John Wiley and Sons. Morrison, R.B. 1964. Lake Lahontan: Geology of Southern Carson Desert, Nevada. United States Geological Survey Professional Paper, 401: 156 pages. 1991. Quaternary Stratigraphic, Hydrologic, and Climatic History of the Great Basin, with Emphasis on Lakes Lahontan, Bonneville, and Tecopa. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. Geology of North America, volume K- 2:283-320. Boulder, Colorado: Geological Society of America. Morrison, R.B., and J.O. Davis 1984. Supplementary Guidebook for Field Trip 13, Quaternary Stratigra- phy and Archeology of the Lake Lahontan Area: A Re-assessment. Second edition. Social Sciences Center Technical Report, 41:1?50. Reno, Nevada: Desert Research Institute, University of Nevada. Oviatt, C.G., R.S. Thompson, D.S. Kaufman, J. Bright, and R.M. Forester 1999. Reinterpretation of the Burmester Core, Bonneville Basin, Utah. Quaternary Research, 52:180-184. Page, B.M., G.A. Thompson, and R.G. Coleman 1998. Late Cenozoic Tectonics of the Central and Southern Coast Ranges of California. Geological Society of America Bulletin, 110:846-876. Pantea, M.P., S. Asher-Bolinder, and J.D. Vine 1981. Lithology and Lithium Content of Sediments in Basins Surrounding Clayton Valley, Esmeralda and Nye Counties, Nevada. United States Geological Survey Open-File Report, 81-962: 23 pages. Parsons, T., G.A. Thompson, and N.H. Sleep 1994. Mantle Plume Influence on the Neogene Uplift and Extension of the U.S. Western Cordillera. Geology, 22:83-86. Perkins, M.E., W.P. Nash, F.H. Brown, and R.J. Fleck 1995. Fallout Tuffs of Trapper Creek, Idaho?A Record of Miocene Ex- plosive Volcanism in the Snake River Plain Volcanic Province. Geo- logical Society of America Bulletin, 107:1484-1506. Pluhar, C.J., J.H. Holt, J.L. Kirschvink, K.K. Beratan, and R.W. Adams 1992. Magnetostratigraphy of Plio-Pleistocene Lake Sediments in the Confidence Hills of Southern Death Valley, California. San Bernar- dino County Museum Association Quaterly, 39(2): 12-19. Priest, G.R. 1990. Volcanic and Tectonic Evolution of the Cascade Volcanic Arc, Cen- tral Oregon. Journal of Geophysical Research, 95(B12): 19,583? 19,599. Putnam, W.C. 1949. Quaternary Geology of the June Lake District, California. Bulletin of the Geological Society of America, 60:1281-1302. Redwine, J.L. In prep. Investigation of Quaternary Pluvial History and Paleoclimatic Im- plications in Newark Valley, Nevada. Master's thesis, Humboldt State University, Arcata. Reheis, M.C. 1999. Possible Pleistocene Integration of the Humboldt River, NV. Geo- logical Society of America Abstracts with Programs, 3 l(7):A-444. Reheis, M.C, J.C. Goodmacher, J.W. Harden, L.D. McFadden, T.K. Rockwell, R.R. Shroba, J.M. Sowers, and E.M. Taylor 1995. Quaternary Soils and Dust Deposition in Southern Nevada and Cali- fornia. Bulletin of the Geological Society of America, 107: 1003-1022. Reheis, M.C, and R.B. Morrison 1997. High, Old Pluvial Lakes of Western Nevada. In P.K. Link and BJ. Kowallis, editors, Proterozoic to Recent Stratigraphy, Tectonics, and Volcanology, Utah, Nevada, Southern Idaho and Central Mexico, volume 1:459-492. Provo, Utah: Brigham Young University Geol- ogy Studies. Reheis, M.C, A.M. Sarna-Wojcicki, D.M. Burbank, and C.E. Meyer 1991. The Late Cenozoic Section at Willow Wash, East-Central California: A Tephrochronologic Rosetta Stone. In M.C. Reheis, A.M. Sarna- Wojcicki, C.E. Meyer, E.H. McKee, J.L. Slate, D.M. Burbank, T.L. Sawyer, and E.G. Pendall, Late Cenozoic Stratigraphy and Tectonics of Fish Lake Valley, Nevada and California: Road Log and Contri- butions to the Field Trip Guidebook, 1991 Pacific Cell, Friends of the Pleistocene. U.S. Geological Survey Open-File Report, 91- 290:46-66. Reheis, M.C, and T.L. Sawyer 1997. Late Cenozoic History and Slip Rates of the Fish Lake Valley, Emi- grant Peak, and Deep Springs Fault Zones, Nevada and California. Bulletin of the Geological Society of America, 109:280?299. Reheis, M.C, J.L. Slate, A.M. Sarna-Wojcicki, and C.E. Meyer 1993. A Late Pliocene to Middle Pleistocene Pluvial Lake in Fish Lake Valley, Nevada and California. Bulletin of the Geological Society of America, 105:953-967. Reheis, M.C, S. Stine, and A.M. Sarna-Wojcicki 2002. Drainage Reversals in Mono Basin during the Late Pliocene and Pleistocene. Bulletin of the Geological Society of America, 114: 991-1006. Roberts, R.J. 1965. Stratigraphy and Structure of the Antler Peak Quadrangle, Humboldt and Lander Counties, Nevada. United States Geological Survey Pro- fessional Paper, 459A:1?93. Russell, I.C 1885. Geological History of Lake Lahontan, a Quaternary Lake of North- western Nevada. Monographs of the United States Geological Sur- vey, 11:1-288. 1889. Quaternary History of Mono Valley, California. Annual Report of the United States Geological Survey, 8:261-394. Sarna-Wojcicki, A.M. 1995. Age, Areal Extent, and Paleoclimatic Effects of "Lake Clyde," a Mid-Pleistocene Lake that Formed the Corcoran Clay, Great Valley, Calif. In Alan Gillespie, convenor, Abstracts for Glacial History of the Sierra Nevada, Calif; A Symposium in Memorial to Clyde Wahr- haftig, Sept. 20?22, 1995, White Mountain Research Station, Bishop, California. 10 pages. Bishop, California: White Mountain Research Station. 2000. Tephrochronology. In J.S. Noller, J.M. Sowers, and W.R. Lettis, edi- tors, Quaternary Geochronology, Methods and Applications, pages 357-377. Washington, D.C.: American Geophysical Union. Sarna-Wojcicki, A.M., H.R. Bowman, C.E. Meyer, P.C. Russell, M.J. Wood- ward, G. McCoy, J.J. Rowe, Jr., P.A. Baedecker, F. Asaro, and H. Michael 1984. Chemical Analyses, Correlations, and Ages of Upper Pliocene and Pleistocene Ash Layers of East-Central and Southern California. United State Geological Survey Professional Paper, 1293:1-40. Sarna-Wojcicki, A.M., and J.O. Davis 1991. Quaternary Tephrochronology. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. Geology of North America, volume K-2:93-116. Boulder, Colorado: Geological Society of America. Sarna-Wojcicki, A.M., K.R. Lajoie, C.E. Meyer, D.P. Adam, and H.J. Rieck 1991. Tephrochronologic Correlation of Upper Neogene Sediments along the Pacific Margin, Conterminous United States. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. Geology of North America, volume K-2:117-140. Boulder, Colorado: Geo- logical Society of America. 108 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Sarna-Wojcicki, A.M., S.D. Morrison, C.E. Meyer, and J.W. Hillhouse 1987. Correlation of Upper Cenozoic Tephra Layers between Sediments of the Western United States and Eastern Pacific Ocean, and Compari- son with Biostratigraphic and Magnetostratigraphic Age Data. Bul- letin of the Geological Society of America, 98:207-223. Sarna-Wojcicki, A.M., M.C. Reheis, C.E. Meyer, R.J. Fleck, B. Troxel, J.L. Slate, E. Wan, J.K. Nakata, and D.W. Burbank In review. The Tuff of Taylor Canyon and Related Upper Pliocene and Pleistocene Tephra Layers, California, Nevada, and Utah: Isotopic Ages, Magnetostratigraphy, and Correlation. Shackleton, N.J., A. Berger, and W.R. Peltier 1990. An Alternative Astronomical Calibration of the Lower Pleistocene Timescale Based on ODP Site 677. Transactions of the Royal Soci- ety of Edinburgh, Earth Sciences, 81:251-261. Silberman, M.L., and E.H. McK.ee 1972. A Summary of Radiometric Age Determinations on Tertiary Volca- nic Rocks from Nevada and Eastern California, Part II: Western Ne- vada. Isochron/West, 4:7-28. Small, E.E., and R.S. Anderson 1995. Geomorphically Driven Late Cenozoic Rock Uplift in the Sierra Ne- vada, California. Science, 270:277-280. Smith, G.I., J.L. Bischoff, and J.P. Bradbury 1997. Synthesis of the Paleoclimatic Record from Owens Lake Core OL- 92. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleo- climatic Record from Core OL-92, Owens Lake, Southeast Califor- nia. Special Papers of the Geological Society of America, 317: 143-160. Smith, G.R., T.E. Dowling, K..W. Gobalet, T. Lugaski, D.K. Shiozawa, and R.P. Evans 2002. Biogeography and Timing of Evolutionary Events among Great Ba- sin Fishes. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:175-234. Smith, J.F.J., and K.B. Ketner 1976. Stratigraphy of Post-Paleozoic Rocks and Summary of Resources in the Carlin-Pinon Range Area, Nevada. United States Geological Survey Professional Paper, 867-B:48. Spotila, J., K. Farley, and K. Sieh 1998. Uplift and Erosion of the San Bernardino Mountains Associated with Transpression along the San Andreas Fault, California, as Con- strained by Radiogenic Helium Thermochronometry. Tectonics, 17: 360-378. Stewart, J.H. 1988. Tectonics of the Walker Lane Belt, Western Great Basin: Mesozoic and Cenozoic Deformation in a Zone of Shear. In W.G. Ernst, editor, Metamorphism and Crustal Evolution of the Western United States, Rubey volume 7:683-713. Englewood Cliffs, New Jersey: Prentice Hall. Stewart, J.H., P.C. Kelleher, and E.A. Zorich 1994. Geologic Map of the Monte Cristo Range Area, Esmeralda and Min- eral Counties, Nevada. United States Geological Survey Miscella- neous Field Studies Map, MF-2260. Szabo, B.J., and C.A. Bush 1996. Uranium-Series Dating of Carbonate (Tufa) Deposits Associated with Quaternary Fluctuations of Pyramid Lake, Nevada. Quaternary Research, 45:271-281. Tackman, G. 1993. Middle and Late Pleistocene Hydrologic History of Diamond Valley, Eureka and Elko Counties, Nevada, with Climatic and Isostatic Im- plications. 192 pages. Master's thesis, University of Utah, Salt Lake City. Taylor, D.W. 1985. Evolution of Freshwater Drainages and Molluscs in Western North America. In C.J. Smiley, E.E. Leviton, and M. Berson, editors, Late Cenozoic History of the Pacific Northwest: Interdisciplinary Studies on the Clarkia Fossil Beds of Northern Idaho, pages 265-321. San Francisco: American Association for the Advancement of Science, Pacific Division. Taylor, D.W., and G.R. Smith 1981. Pliocene Molluscs and Fishes from Northeastern California and Northwestern Nevada. Contributions from the Museum of Paleontol- ogy, University of Michigan, 25:339-413. Thompson, R.S., C. Whitlock, P.J. Bartlein, S.P. Harrison, and W.G. Spaulding 1993. Climatic Changes in the Western United States since 18,000 yr B.P. In H.E.J. Wright, J.E. Kutzbach, T. Webb, W.F. Ruddiman, F.A. Street-Perrott, and P.J. Bartlein, editors, Global Climates since the Last Glacial Maximum, pages 468-513. Minneapolis: University of Minnesota Press. Weick, R.J. 1990. Structural, Tectonic, and Quaternary Study of the Eastern Madeline Plains, California and Southwestern Smoke Creek Desert, Nevada. Master's thesis, University of Nevada, Reno. Wernicke, B., R. Clayton, M. Ducea, C.H. Jones, S. Park, S. Ruppert, J. Saleeby, J.K. Snow, L. Squires, M. Fliedner, G. Jiracek, R. Keller, S. Klemperer, J. Luetgert, P. Malin, K. Miller, W. Mooney, H. Oliver, and R. Phinney 1996. Origin of High Mountains in the Continents: The Southern Sierra Nevada. Science, 271:190-193. Whitney, J.W., and CD. Harrington 1993. Relict Colluvial Boulder Deposits as Paleoclimatic Indicators in the Yucca Mountain Region, Southern Nevada. Bulletin of the Geologi- cal Society of America, 105:1008-1018. Willden, R., and R.C. Speed 1974. Geology and Mineral Deposits of Churchill County, Nevada. Bulle- tin of the Nevada Bureau of Mines and Geology, 83:1-95. Williams, S.K. 1993. Tephrochronology and Basinal Correlation of Ash Deposits in the Bonneville Basin, Northwest Utah. 93 pages. Master's thesis, Uni- versity of Utah, Salt Lake City. Winograd, I.J., J.M. Landwehr, K.R. Ludwig, T.B. Coplen, and A.C. Riggs 1997. Duration and Structure of the Past Four Interglaciations. Quaternary Research, 48:141-154. Winograd, I.J., and B.J. Szabo 1988. Water-Table Decline in the South-Central Great Basin during the Quaternary: Implications for Toxic Waste Disposal. In M.D. Carr and J.C. Yount, editors, Geologic and Hydrologic Investigations of a Potential Nuclear Waste Disposal Site at Yucca Mountain, Southern Nevada. Bulletin of the United States Geological Survey, 1790: 147-152. Winograd, I.J., B.J. Szabo, T.B. Coplen, A.C. Riggs, and P.T. Kolesar 1985. Two-Million-Year Record of Deutermium Depletion in Great Basin Ground Waters. Science, 227:519-522. Wolfe, J.A., H.E. Schorn, C.E. Forest, and P. Molnar 1997. Paleobotanical Evidence for High Altitudes in Nevada during the Miocene. Science, 276:1672-1675. Yount, J.C, J.W. Bell, CM. DePolo, A.R. Ramelli, P.H. Cashman, and P.A. Glancy 1993. Neotectonics of the Walker Lane, Pyramid Lake to Tonopah, Ne- vada; Part II: Road Log. In M.M. Lahren, J.H.J. Trexler, and C. Spi- nosa, editors, Crustal Evolution of the Great Basin and the Sierra Nevada, pages 391-408. Reno, Nevada: Mackay School of Mines. Zijderveld, J.D.A. 1967. A.C. Demagnetization of Rocks: Analysis of Results. In D.W. Col- linson, K.M. Creer, and S.K. Runcorn, editors, Methods in Palaeo- magnetism; Proceedings of the NATO Advanced Study Institute on Paleomagnetic Methods, Held in the Physics Department of the Uni- versity of Newcastle upon Tyne, April 1-10, 1964, pages 254-286. Amsterdam: Elsevier. Pleistocene Lakes and Paleoclimates (0 to 200 Ka) in Death Valley, California Tim K. Lowenstein ABSTRACT The existence of an ancient lake in Death Valley, California, named Lake Manly by geologists in the early part of the twentieth century, was suggested on the basis of shoreline features (gravel deposits, erosional terraces, and carbonate tufas) and basin-center muds and evaporites. Radiocarbon dating of black lacustrine muds and uranium-series dates on shoreline tufas later established that permanent lakes occurred in Death Valley during at least two dif- ferent periods: marine oxygen isotope stage 2 (26 thousand to about 10 thousand years ago (Ka)) and marine oxygen isotope stage 5e/6 (>200 to -120 Ka). The sources of inflow water to Death Valley during wet pluvial periods were the Amargosa River, the Owens River system, and the Mojave River. The timing of spillover from the Owens River system via Panamint Basin into Death Valley is still uncertain; inflow from the Mojave River is interpreted to have occurred dur- ing the last 18 Ky, but more precise timing and the significance of the Mojave River inflow is unknown. A borehole core (DV93-1) taken from Badwater Basin, Death Valley, in 1993, is composed of interbedded salts and muds. Core DV93-1 contains a 186 m long, 200 thousand years (Ky) record of closed-basin paleoenvironments and paleoclimates, which is inter- preted on the basis of sedimentology, ostracode assemblages, homogenization temperatures of fluid inclusions in halite. The lacustrine sediments in core DV93-1 correlate with dated shoreline tufa. The 200 Ky paleoclimate record is dominated by two cycles of dry and/or warm conditions followed by wet and cold condi- tions that occurred on a 100 Ky time scale. Of significance are the major lacustrine phases of 35?10 Ka and 186-120 Ka (marine oxygen isotope stages 2 and 5e/6, respectively), which represent dramatically colder and wetter conditions than those of modern Death Valley. Of the two major lake periods, the penultimate gla- cial lakes were deeper and lasted far longer than those of the last glacial. Tim K. Lowenstein, Department of Geological Sciences, State Univer- sity of New York at Binghamton, Binghamton, New York 13902, USA; e-mail: lowenst@binghamton. edu. Reviewers: D.E. Anderson, Center for Environmental Sciences and Education, Northern Arizona University, Flagstaff, Arizona 86011, USA; David B. Madsen, Environmental Sciences, Utah Geological Survey, Salt Lake City, Utah 84114, USA. Previous Work on the Ancient Lakes of Death Valley The first published evidence of an ancient lake in Death Val- ley, California, is from Noble (1926), who briefly described strandlines along a basalt hill in southern Death Valley (Shore- line Butte) and in gravel deposits at Mormon Point (Figure 1). These former shorelines were, according to Noble (1926), 122 to 153 m, respectively, above the low point of Death Valley. The ancient lake in Death Valley was named Lake Manly by Means (1932) for William Manly, who traveled through Death Valley in 1849?and survived. Blackwelder (1933) described further evidence of former Lake Manly, including the mud de- posits underlying the modern surface salt crust in the center of Death Valley, and at least 12 erosional terraces on Shoreline Butte that were up to 95 m above sea level and contained rounded gravels and carbonate coatings. Less pronounced ter- races from several locations on the southwest and eastern mar- gins of Death Valley (Park Village and Mormon Point, respec- tively) also were described by Blackwelder (1933), as well as two localities where small tufas were interpreted as possible shoreline deposits of a former lake. Blackwelder (1933, 1954) speculated that Lake Manly ex- isted during a glacial period. The poor preservation of shoreline features compared with the well-documented features from Lake Bonneville and Mono Lake led him to conclude that Lake Manly existed during the Tahoe glaciation (early Wisconsin, now dated at ~74 to -49 thousand years ago (Ka)) and that a shallower lake also may have existed during the late Wisconsin Tioga glacial (dated at -30 to -12 Ka (Phillips et al., 1990; Phillips et al., 1996)). Blackwelder (1933) suggested that Lake Manly was a relatively short-lived lake because of the relative paucity of tufa and the poor development of erosional shoreline features in Death Valley. Hunt and Mabey (1966) followed previous workers when they concluded that Lake Manly had been synchronous with Wisconsin (Tahoe and Tioga) glaciation in the Sierra Nevada, and they added several more observations to the growing list of shoreline features. The lacustrine shoreline and near-shore 109 110 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 1.?Map of Death Valley, California, and location of core DV93-1 (indicated by the white star). 35?37' 117?30' W Saline pan f.?"??;:'-/--N Mud flat Elevation <305 m, predominantlyalluvial fan and delta sediments features described by Hunt and Mabey (1966) included shin- gled gravels along the north and east sides of Death Valley that were up to 116m above sea level; a large gravel bar composed of well-sorted and crossbedded gravels, near Beatty Junction at 46 m above sea level; terraces near Park Village at elevations between sea level and 46 m above sea level; calcite-cemented gravels along the Black Mountains between Badwater and Mormon Point, most commonly between sea level and 61 m above sea level; and numerous horizontal shoreline features of Shoreline Butte. Hooke (1972) found important new information on the an- cient lakes in Death Valley from borehole cores taken in the center of Badwater Basin and from shoreline carbonate tufas plastered on the Black Mountains. Hooke (1972) was the first to publish quantitative information on the timing of the ancient lakes in Death Valley, which he derived from radiocarbon dat- ing of organic carbon from the Badwater Basin cores. The Badwater Basin cores contained brown mud, 3-15 m thick, overlying 1-2 m of olive green mud, which graded into black mud below. The black muds were interpreted as lake deposits that were never subaerially exposed (Hooke, 1972). Radiocar- bon dates from two samples near the top of the black mud in- terval were 11,900 yr and 12,980 yr. A date of 21,500 yr was obtained from lower in the black mud interval. Hooke (1972), therefore, proposed that a perennial lake existed in Death Val- ley from about 26 to 10.5 Ka. Radiocarbon dates from the ba- sin-center lacustrine muds led Hooke (1972) to interpret the shorelines 90 m above sea level as latest Wisconsin in age, 11-10 Ka, and he named the youngest shoreline features the Blackwelder stand of Lake Manly. Hooke (1972) also studied carbonate tufas from various locations in the Black Mountains. He tentatively inferred the presence of three generations of tu- fas on the basis of their coloration and weathering characteris- tics: latest Wisconsin (Blackwelder stand), early Wisconsin, and pre-Wisconsin. Later work on dating the Black Mountain tufas by Hooke and Lively (1979) and by Hooke and Dorn (1992) has led to new interpretations of the Blackwelder stand. Hooke and Lively (1979) dated 34 tufas by uranium-series (U-series) methods. Although there were problems with the possible loss NUMBER 33 111 of uranium from samples, and also possible 230Th contamina- tion, a majority of the dates fell between 225 and 135 Ka, which led Hooke and Lively (1979) and Hooke and Dorn (1992) to conclude that the Blackwelder stand was much older than previously thought and likely coincided with Illinoian gla- ciations and marine oxygen isotope stage 6. In addition, two tu- fas from the high shorelines of neighboring Panamint Valley (Figure 2) were dated by Hooke and Lively (1979). They, too, were apparently Illinoian in age, 138?7 Ka and 137?4 Ka, al- though loss of uranium and 230Th contamination may have af- fected the samples. Finally, Dora (1989) and Dora et al. (1989) performed cation-ratio dating of desert varnish and concluded that the high Blackwelder stand receded 130-120 Ka. Dora et al. (1990) used radiocarbon dating of rock varnish to estimate the beginning of the recession of the latest Pleistocene lake in Death Valley. They obtained dates of 12,970 yr from Mormon Point (3 m above sea level) and 12,630 yr from Park Service Housing, Park Village (5 m below sea level) (Dorn et al., 1990). These dates were consistent with the radiocarbon dates obtained by Hooke (1972) for the existence of a perennial lake in Death Valley between 26 and 10 Ka. Most recently, Meek (1997) carefully measured the shoreline elevations at Shoreline Butte and Mormon Point. The promi- nent 94.7 m bench at Shoreline Butte was interpreted as an Illi- noian-age (marine oxygen isotope stage 6) shoreline, marking a lake 1606 km2 in area and having a volume of 176 km3. Meek (1997) interpreted the shorelines around 5 m above sea level as late Wisconsin in age (marine oxygen isotope stage 2). The highest shoreline in Death Valley, 135.4 m above sea level at Shoreline Butte, was interpreted by Meek (1997) as being older than marine oxygen isotope stage 6. Meek (1997) discussed the difficulties involved in correlating shorelines from different lo- calities because the timing and magnitude of faulting are un- known. For example, the highest shoreline features at Mormon Point (104-113 m) and the uppermost shoreline at Shoreline Butte (135.4 m) may represent different lake phases or they may represent the same lake phase with faulting of approxi- mately 26 m between locations (Meek, 1997). D.E. Anderson and Wells (1997) and Anderson (1998) inves- tigated late Wisconsin-age lake sediments from 10 shallow cores taken in southern Death Valley. On the basis of sedimen- tary textures and accelerator mass spectrometry (AMS) radio- carbon ages, Anderson (1998) identified three late Pleistocene highstands at -26, -18, and -12 Ka, which are in complete agreement with the work of Hooke (1972). Lake Manly was fed by inflow from the Amargosa River. This river currently drains areas to the east of Death Valley in southwestern Nevada and is a source of inflow water into the southern end of Death Valley. In the Pliocene and Pleistocene, Lake Tecopa, with an estimated area of 235 km2, may have been the terminal lake of the Amargosa River drainage (Figure 2) (Morrison, 1991). Correlation of tephra layers indicates the breaching of Lake Tecopa may have occurred between 170 and 150 Ka (Morrison, 1991). After that time, the Amargosa River flowed directly into Death Valley. Additional inflow sources that may have supplied water to Death Valley during wet pluvial periods were (1) the Owens River system to the west of Death Valley, which receives water from the Sierra Nevada, and (2) the Mojave River from the south, which receives the drainage from the San Bernardino Mountains (Figure 2) (see for example, Blackwelder, 1933, 1954; Smith, 1979, 1984; Smith and Pratt, 1957; Jannik et al., 1991; Brown and Rosen, 1995). The Owens River system de- rived waters from the Sierra Nevada and eventually reached Death Valley via a system of lakes and spillways, namely Owens Lake, China Lake, Searles Lake, Panamint Lake (Lake Gale), and finally Death Valley (Lake Manly) via Wingate Pass in southwestern Death Valley (Gale, 1914; Blackwelder, 1954; Smith and Pratt, 1957; Smith, 1979, 1984; Jannik et al., 1991). The inferred connection between Panamint Lake and Lake Manly at Wingate Pass is controversial because evidence of erosion by spillover flow from Panamint Lake into Death Val- ley is equivocal. Hooke (1972) interpreted four hills composed of gravels as delta deposits, implying they were formed by wa- ter spillover from Panamint Lake. Tufa deposits, along with erosional benches and shallow-water gravel deposits, have been described by R.S.U. Smith (1978) from various locations in Panamint Valley between elevations of 502 and 627 m above sea level. Some of these shoreline features are at or above the elevation of Wingate Pass (605 m), which is strong evidence for spillover into Death Valley, ignoring the effects of faulting. Fitzpatrick and Bischoff (1993) obtained U-series dates on samples from the high shorelines of Panamint Lake but argued that the dates may be unreliable because of post-depositional mobilization of uranium. They concluded that the shoreline features are between 300 and 30 Ka in age but probably do not represent late Pleistocene marine oxygen isotope stage 2 (Fitz- patrick and Bischoff, 1993). Their isochron plots suggest ages between 95 and 55 Ka; therefore, although water clearly en- tered Death Valley via spillover from Panamint Lake, the tim- ing of that spillover is still uncertain. Inflow of water from the Mojave River into the south end of Death Valley was discussed by Brown and Rosen (1995), who concluded that Lake Manly may have received inflow from the Mojave drainage area only in the latest Pleistocene. This con- clusion was based on the age of Pleistocene Lake Mojave (now Silver Lake playa near Baker, California) to the south of Death Valley (younger than 22 Ka) and on the supposition that for water to overflow into Death Valley from the Mohave River drainage area, Lake Mojave itself must have been at or above a spillover level (Figure 2). This interpretation is supported by work on the Manix Basin by Meek (1989; pers. comm., 1996). From the study of the sediments at Afton Canyon, near Lake Manix, Meek (1989; pers. comm., 1996) concluded that Lake Manix was the terminal lake in the Mojave River system for most of the Pleistocene until Afton Canyon was eroded some- time after 18 Ka. At that point, waters from the Mojave River 112 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Present-day river Kilometers 119? 118? 117? 116? 115? FIGURE 2.?Map of pluvial lakes, modern lakes and playas, Pleistocene drainage patterns, and modern drainages of the Owens River, Amargosa River, and Mojave River systems. Modified from Blackwelder (1954), Smith (1984), Jannik et al. (1991), and Brown and Rosen (1995). reached the Silver Lake and Soda Lake playas. During excep- tionally high levels of Silver Lake (Lake Mojave), water may have overflowed into the southern end of Death Valley. There are no features reported from Death Valley indicating inflow of large amounts of water at approximately 18 Ka, but this pe- riod does coincide with a wet period in Death Valley when the basin contained a perennial lake (see "Sedimentology," be- low). Additional information on the timing of inflow waters from the Mojave River into the southern end of Death Valley comes from studies of Lake Dumont conducted by K.C. Anderson and Wells (1996, 1997). The Dumont Basin (10 km2) is a small playa located between Silver Lake playa and the southern arm of the Amargosa River (Figure 2). K.C. Anderson and Wells (1996, 1997) found lacustrine sediments in two 13 m long cores taken in Dumont Basin. They identified Lake Dumont I (25.3-30 Ka) and Lake Dumont II (18-19.8 Ka) on the basis of radiocarbon dates obtained from ostracode-bearing lacustrine muds. K.C. Anderson and Wells (1996) concluded that Lake Dumont II broke through its sill sometime after 18 Ka, which established a direct connection between the Mojave River, the Amargosa River, and Death Valley. That date coincides with the timing of the erosion of Afton Canyon upstream in the Mo- jave River system (Meek, 1989; pers. comm., 1996). It is not known whether Lake Dumont I overflowed into Death Valley between 30 and 25 Ka. Analysis and Interpretation of Data from Core DV93-1 from Death Valley A 6.4 cm diameter core, DV93-1, was drilled to a depth of 185.5 m through Badwater Basin, Death Valley, during April and May of 1993 (Figure 1). The drill site was near the center of the saline pan, 3.2 km northwest of Badwater Springs. The core is composed of interbedded salt and mud. Virtually all salt layers were completely recovered, but retrieval of the mud in- tervals was about 60 percent successful. The core records the 200 Ky record of closed-basin paleoenvironments and paleocli- mates in Death Valley, as interpreted on the basis of sedimen- tology, ostracode species, fluid-inclusion homogenization tem- peratures in halite, saline mineralogy, and correlation with shoreline tufa. NUMBER 33 113 U-series dating (ka) Environment Salt pan and saline mud flat Perennial salinelake Salt pan with ephemeral lakes Salt pan with ephemeral lakes Shallow saline lake 97.1 ?6.8 ar|d salt pan U-series dating (ka) Environment Primary halite Cavity-filling halite Core not recovered 119.6 ?2.6 Mud flat Perennial salinelake 150 160 170 180 Perennial lake Perennial 146 + 6 saline _ lake Perennial lake .3 ?8.5 Perennial saline i lake Salt pan and shallow saline lake Mud flad TpT Vertical dissolution pipes in halite Mud Mud with mudcracks FIGURE 3.?Stratigraphic column, Core DV93-1, with uranium-series ages, sediment types, and interpreted dep- ositional environments. DATING CORE DV93-1.?The core was dated using the 23OTh/234U method, which determines the time uranium was in- corporated into a sample (Luo and Ku, 1991; Ku et al, 1998). The total dissolution isochron technique was used to account for initial detrital 230Th in the samples (Luo and Ku, 1991). Samples from 12 stratigraphic intervals were dated (Figure 3). The samples contained preserved primary halite as well as in- corporated clay, carbonate (predominantly calcite), and sulfate minerals (mainly gypsum or glauberite, CaSO4-Na2SO4). Au- thigenic 23OTh/234U and 234U/238U ratios used in calculating the age of a halite sample were determined from linear plots (isoch- rons) of 23OTh/232Th versus 234U/232Th and of 234U/232Th versus 114 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 238U/232Th of a series of subsamples taken from each of the 12 stratigraphic intervals. These activity ratios were counted using alpha particle spectrometry (Ku et al., 1998). The given age er- rors are one standard deviation, which were derived from counting statistics and the fitting of isochrons. The age errors are less than 11 percent, except for the uppermost sample from 7.7 m, where the large error results from the relatively large amount of detrital 230Th initially in this relatively young sam- ple. The reliability of the dating technique can be seen from the general increase in age down the core. Two samples from around 60 m (Figure 3) with very close ages, but with very dif- ferent amounts of detrital material, and hence different 230Th/ 232Th ratios, supports the validity of the isochron approach. All dates discussed that were not obtained directly from the U-se- ries isochron method were calculated by interpolation, assum- ing constant accumulation rates for each sediment interval. Sedimentation rates averaged over the entire 186 m of core DV93-1 are lm/lKy (thousand years). Salts, predominantly halite, were deposited rapidly, 1.7-3.8 m/lKy, whereas muds accumulated more slowly, at rates of 0.4-1.0 m/1 Ky. Carbonate shoreline tufas from the east side of Death Valley, between Badwater Springs and Mormon Point were carefully chosen to avoid samples with impurities, weathering rinds, ex- tensive cements, or recrystallization textures (see "Shoreline Carbonate Tufas," below). Samples were cleaned and dis- solved, and 238U, 234U, 232Th, and 230Th were counted by alpha particle spectrometry (Ku et al., 1998). The ages of tufas, along with one standard deviation errors derived from counting sta- tistics, are shown in Figure 4c. For most samples, the high ac- tivity ratios of 23OTh/232Th, greater than 15, suggest that the ages of samples should be close to the time of tufa formation (Kuetal., 1998). SEDIMENTOLOGY.?Paleoenvironments present in Death Valley over the past 200 Ky were interpreted by comparing sedimentary structures and petrographic features described from modern closed-basin sediments with those observed in core DV93-1 (Figures 3, 4A) (Hardie et al., 1978; Lowenstein and Hardie, 1985; Smoot and Lowenstein, 1991; Roberts et al., 1994; Li et al., 1996). The cored sediments contain a record of closed-basin paleoenvironments ranging from hot, dry mud- flats, similar to the modern floor of Death Valley, to temperate climate, deep perennial lakes. Disrupted brown muds overlain by a 0.25 m thick surface ha- lite crust (0-7.7 m) indicate Death Valley has been a mudflat and saline pan during the last 10 Ky (Figures 3, 4A). Further details are lacking because most sedimentary structures from this interval were destroyed during coring. A perennial saline lake existed in Death Valley for an -25 Ky period, from 10 to 35 Ka (core depths of 7.7-18 m) (Li et al., 1996). Interbedded mud and halite cycles accumulated during the early perennial lake stage, followed by interlayered the- nardite (Na2SO4) and mud. A cap of chevron halite formed during the latest lake stage. These deposits record a generally wet climate, with fluctuating lake levels and salinities. The lack of synsedimentary dissolution features in evaporite layers indi- cates that accumulated salts were permanently protected from dilute flood waters by the saline lake, which suggests that lakes existed more or less continually, without drying out, for 25 Ky. The mud layers are structureless, contain rare ostracodes, and are black in color but change to gray green after exposure to air. Halite textures include cumulate settle-out layers and bot- tom-grown crusts, which are interlayered with black muds (Li et al., 1996). Thenardite layers from 10 to 13 m are apparently diagenetic in origin but may have first formed as mirabilite (Na2SO4-10H2O), as observed in the winter in Great Salt Lake and Owens Lake (Spencer et al., 1984; Spencer et al., 1985a; Spencer et al., 1985b; Smith et al., 1987). The age of the the- nardite is ~22 to ~16 Ka in core DV93-1. This age coincides with the last glacial maximum and suggests that Na2SO4 crys- tallized as mirabilite during winter cold water conditions, prob- ably at temperatures below -10? C (Smith et al., 1987). Halite from the uppermost saline lake interval (7.7-10 m) contains chevrons, raft textures, and pisoids. These are interpreted to have accumulated in a shallow, perennial lake that ultimately dried out about 10 Ka (Li et al., 1996). Saline-pan evaporites present from 18 to 60 m indicate that Death Valley was commonly desiccated between ~35 to -60 Ka, but it did receive enough inflow water to maintain shallow saline lakes at times and to supply the solutes required to accu- mulate salts rapidly, at rates of 1.7 m/1 Ky (Figures 3, 4A) (Li et al., 1996). Saline-pan halites consist of poorly bedded halite, with little preservation of primary depositional textures. A dis- tinctive feature of halite layers is a vertical pipework fabric ce- mented with halite, which signifies multiple episodes of sub- aerial exposure, dissolution of surficial halite by dilute floodwaters, and later cementation of cavities by shallow groundwater brines. Five thin halite intervals between 43 and 52 m (Figure 4A) (~50-54 Ka), which carry primary features such as chevrons and rafts, indicate that shallow saline lakes existed in Death Valley during this time. Clayey silts, some massive and some with mudcracks and sand-patch textures, which are diagnostic of deposition on dry mudflats, are the dominant sediment type from 60.5 to 109 m (Figures 3, 4A) (60-120 Ka) (Smoot and Lowenstein, 1991; Roberts et al., 1994; Smoot and Castens-Seidell, 1994; Li et al., 1996). Sand patches are wind-blown or water-deposited sedi- ments, which filled depressions in efflorescent crusts on the mudflat surface (Smoot and Castens-Seidell, 1994). Later flooding dissolved the efflorescent salt crusts, which caused collapse of the overlying sediments and produced internal frac- tures, faulting, and rotation of sand patches. Saline minerals consist of crystals of halite, glauberite (CaSO4Na2SO4), and gypsum (CaSO4-2H2O) that formed during early diagenetic displacive growth from groundwater brines. Interbedded rem- nants of saline-pan halite and a total lack of lacustrine fossils, such as diatoms and ostracodes, reinforce the interpretation that these muds were from mudflats, not perennial-lake deposits. One halite interval, interpreted to have been deposited in shal- NUMBER 33 115 AGE (ka) 0 20 40 50 60 80 100 110 120 130 150 170 190 TTiiir 1 I I DEPTH(m) 1 n i n r CO 0 10 20 40 60 80 100 120 j, Deep Lake PerennialSaline Lake Saline PanMudflat 120 140 0 180 (A) "g L. Sappaensis o L. Staplini 2 L. Ceriotuberosa to C. Caudata I I II Illl?II II INI II 1 I III I I I I III Illl II Mill H-H h- -H 1 H-H 1 (B): -166 ?6- High Shoreline Tufa Ages (ka) ^i^178?15 1^,7* 1? 162.5 ?2.5* . 185?15 i-216*15- 15 Gypsum Glauberite Thenardite 50 co 40 10 iiiiiiiiiiiiini mi i iiiiiiiiiiiiiiiniii in iiiiiiini niiii iiiniiiniiiiiiniiii (E)- (F) DEPTH (m) ,9 MCO (D ' O 20 40 60 0 20 40 50 80 100. 120_ 140 160^ 180 (Q) 60 80 100 110 120 130 150 170 -7? 190 AGE (ka) FIGURE 4.?Paleoclimate record, core DV93-1, during the last 200 Ky. A, Lake levels and paleoenvironments, based on relative ranking, from driest to wettest: mudflat, saline pan, perennial saline lake, and deep lake. The deepest lakes are interpreted from the occurrence of ostracode species Candona caudata and Limnocythere ceri- otuberosa (R.M. Forester, pers. comm., 1995). B, Ostracodes. C, Uranium-series ages of high-shoreline (about 90 m above sea level) carbonate tufas, Black Mountains, Death Valley. D, Maximum homogenization temperatures of fluid inclusions in halite (Ty^x )? E, Sulfate minerals, determined by x-ray diffraction analysis of bulk sam- ples. F, Weight percent calcite, determined by x-ray diffraction analysis of the >1 \\rs\ size insoluble fraction of the core. G, Uranium-series ages, interpolated ages, and marine oxygen isotope stages. low saline lake and saline-pan settings, occurs from 78 to 87 m (Figure 3). Uranium-series dates from the top and bottom of the interval are 98.6 Ka and 97.1 Ka, respectively, which sug- gests rapid evaporite accumulation. Perennial-lake sediments were deposited during a 66 Ky pe- riod, between 186 and 120 Ka (Figures 3, 4A). Dark lacustrine muds (core depth of 127-161 m) were overlain by saline-lake halites (core depth of 109-127 m), which were interlayered with muds (Roberts et al., 1994, 1997). The perennial-lake muds were massive to vaguely layered and dark gray when first exposed to air, but they changed color to gray green over time. Halite layers at 152 m (166 Ka) and at 137.5 to 139.5 m (146 Ka) indicate high salinities and probable shallower lake waters occurred during these intervals. Perennial-lake halites are gen- erally clear and lack dense, fluid-inclusion banding; they have sorted cumulate textures produced by crystallization at the air- brine interface as rafts, skeletal crystals, and cubes (Roberts et al., 1994; Roberts and Spencer, 1995). It is likely that lake lev- els and water salinities commonly fluctuated over the entire 66 Ky lacustrine phase (Roberts et al., 1997). The upper 18 m of the perennial-lake succession (109-127 m) consist of interlay- ered subaqueous halite (settle-out cumulates or vertical crystal 116 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES fabrics formed by slow crystallization at the brine bottom) and mud that record two major shallowing events between 120 and 128 Ka, and final desiccation at 120 Ka. The bottom 25 m of the core, 186 to 192 Ka (core depth of 161-186 m) is predominantly layered halite and mud that formed in saline pan and shallow saline lake environments. The halite layers contain primary fluid-inclusion banded-chevron textures and cumulate settle-out layers that crystallized at the bottom and top of a shallow saline lake (Roberts and Spencer, 1995). Vertical pipework fabrics, cemented with halite, indicate exposure of salt layers to dilute floodwaters above the water ta- ble, which produced the diagnostic vertical pipes (Roberts et al., 1994). The lowest one meter of the core is composed of mudcracked silty mud and is interpreted as a mudflat deposit. OSTRACODES.?Ostracodes from core DV93-1 provide in- formation on the lower limits of water salinities of the two long-lived perennial lakes in Death Valley during the last gla- cial, 35-10 Ka, and the penultimate glacial, 186-120 Ka (ma- rine oxygen isotope stages 2 and 5e/6, respectively) (Figure 4B). Perennial-lake muds contain ostracode species Candona caudata Kaufman, 1900, Limnocythere ceriotuberosa De- lorme, 1967, L. sappaensis Staplin, 1963, and L. staplini Gutentag and Benson, 1962, which, although not abundant, in- dicate that the salinity of the Death Valley lakes in which the ostracodes lived was typically below 10,000 parts per million and at times was less than 3000 parts per million for the salinity sensitive species C. caudata (Forester, 1983, 1987). Such lower-salinity lake phases are inferred from ostracode occur- rences at 10, 14, and 18 m (approximately 16, 25, and 34 Ka, respectively) for the last glacial lake and 113-116, 125-126, 138-143, 157, and 162 m (approximately 121-123, 127-129, 146-154, 179, and 186 Ka, respectively) for the long-lived penultimate glacial lake. Unfortunately, information on the last glacial lake is incomplete because muds from depths of 9-14 m (~ 12-25 Ka) were partially lost during coring. FLUID-INCLUSION PALEOTEMPERATURES.?The paleotem- perature component of the Death Valley climate record was ob- tained from homogenization temperatures of fluid inclusions in subaqueously formed halites from core DV93-1, which record brine temperatures during salt precipitation (Goldstein and Reynolds, 1994; Roberts and Spencer, 1995; Roberts et al., 1997; Lowenstein et al., 1998) (Figure 4D). Shallow saline lake temperatures, in turn, correlate closely with air temperatures in modern settings (Carpelan, 1958; Hammer, 1986; Roberts and Spencer, 1995; Roberts et al., 1997; Lowenstein et al., 1998). Crystals of halite with textures diagnostic of crystallization in a lacustrine environment (cumulates, chevrons, bottom-grown crusts) and with initially one-phase liquid inclusions were chilled in a freezer at temperatures of-5? C to -20? C to nucle- ate vapor bubbles without freezing the inclusions (Roberts and Spencer, 1995). These crystals, now with abundant two-phase liquid-vapor inclusions, were slowly heated to the temperature at which the vapor phase disappeared and the fluid inclusions "homogenized" back to one liquid phase. The temperatures of homogenization record the original brine temperatures at which the crystal grew in the Death Valley saline lakes. Maxi- mum homogenization temperatures of fluid inclusions in halite (ThMAx) were used as a record of maximum brine temperature and maximum air temperature during halite crystallization. This method was used by Roberts and Spencer (1995) and Lo- wenstein et al. (1998) for the precipitation of halite from Death Valley in 1993. Halites in 66 stratigraphic intervals of core DV93-1, from depths of 0 to 183.5 m (0-192 Ka) have fluid-in- clusion homogenization temperatures commonly below the modern rhMAX of 34? C (maximum brine and air temperature during halite precipitation, late April and early May, 1993) (Roberts et al., 1997; Lowenstein et al., 1998). Lacustrine ha- lites deposited during the last glacial period, 35-10 Ka, have low fluid-inclusion maximum homogenization temperatures (ThMAX = 19-30? C), which suggests brine temperatures were ~4?-15? C below modern late April and May values. Ephem- eral saline lake halites precipitated from 60 to 35 Ka have rhMAX between 23? C and 28? C, which are 6?-ll? C below the modern T^MAX values. A relatively low temperature inter- val, 10? C below the modern, existed ~50 to -47 Ka. Limited data from fluid inclusions in two halite intervals, -92 and -80 Ka in age, suggest generally cool conditions, 6?-7? C below modern late April to May temperatures. Relatively higher rhMAX values, to 34? C and 35? C in halite dated as -100 Ka and to 32? C in halite dated as 120 Ka, may record climate re- gimes more similar to the modern. These temperature, how- ever, are still below modern mid-summer temperatures in Death Valley (average air temperatures in Death Valley are 39? C in July and 37? C in August; average maximum air tem- peratures are 46? C in July and 45? C in August). Generally colder conditions are recorded in most fluid inclu- sions in halite for the 186-120 Ka perennial-lake sequence, where rhMAX values range from 25? C to 32? C, but only 3 of 22 stratigraphic intervals have T^MAX greater than 30? C (Rob- erts et al., 1997). A combination of homogenization tempera- tures of fluid inclusions in halite and petrographic evidence of possible pseudomorphs of hydrohalite (NaCl-2H2O) suggest that temperatures in the perennial lakes may have dropped be- low 0? C at times and that temperatures during this period probably averaged 10?-15? C below the modern (Roberts et al., 1997). In the bottom 20 m of core DV93-1 (164-184 m, 186-192 Ka), homogenization temperatures are relatively high. Of 19 stratigraphic intervals analyzed, 14 have ThMAX values greater than or equal to 30? C, which is similar to those ob- tained from halite precipitated in modern Death Valley in late April and early May. SALINE MINERALOGY.?Chemically precipitated minerals (sul- fates and carbonates) in the Death Valley core provide informa- tion on Pleistocene water inflow sources and climate (Figure 4E,F) (Li et al., 1997). Abundant glauberite (CaSO4-Na2SO4) and gypsum (CaSO4-2H2O) and relatively small amounts of calcite (13%?35% of the >1 |J,m size insoluble fraction) are as- sociated with dry periods (mudflat deposits of 120-60 Ka and NUMBER 33 117 10-0 Ka and saline-pan sediments of 192-186 Ka). In contrast, scarce CaSO4-bearing minerals but relatively abundant calcite (20%-50% of the >1 |j,m size insoluble fraction) are associated with halite and mud layers from wetter periods (186-120 Ka, 60-10 Ka). Different mixing ratios of inflow waters between wet and dry periods are interpreted to be responsible for the re- lationship between mineral assemblages and climate. Modern Death Valley, which is representative of ancient dry periods, is characterized by Na-Cl-SO4 brines that are pro- duced by mixing two basic types of inflow waters (Li et al., 1997): (1) Na-HCO3-rich and Na-Cl-SO4-HCO3-rich meteoric waters from the Amargosa River, from springs, and from groundwaters from northern and central Death Valley; and (2) Na-Ca-Cl-rich springs and groundwaters from southern Death Valley, possibly related to volcanism, hydrothermal activity, and a 15 km deep magma body (de Voogd et al., 1986; Serpa et al., 1988; Li et al., 1997). During dry periods, relatively abun- dant Ca-rich spring inflow removes HCO3 as calcite during the early stages of brine evolution (Jones, 1966; Hardie et al., 1978; Li et al., 1997). Further evaporative concentration pro- duces gypsum and glauberite from the remaining Ca. During wetter periods, increased discharge of meteoric HCC^-rich Amargosa River water and basin-margin spring waters re- moves most Ca from the brine via precipitation of calcite, and, with low calcium concentrations, CaSO4-minerals are not abundantly formed during further evaporative concentration. Such Ca-poor, Na-Cl-SO4-rich brines precipitate thenardite (Na2SO4) during later stages of brine evolution. SHORELINE CARBONATE TUFAS.?Shoreline carbonate tu- fas provide information on the depths of Pleistocene lakes in Death Valley (Figure 4c). Shoreline tufas, most prominent ap- proximately 90 m above sea level, form horizontal terraces en- crusting bedrock and gravel on the eastern side of Death Valley (Black Mountains, between Badwater Springs and Mormon Point) (see "Previous Work on the Ancient Lakes in Death Val- ley"; see also Hooke, 1972). Tufas were collected from loca- tions previously described by Hooke (1972), including, from north to south, Badwater, Goblet, Westilt, and Mormon Point (see Figure 1 for locations) (Ku et al., 1998). Relatively fresh hand-samples of tufas are porous and fine-grained with mas- sive or radiating fabrics. In thin section, tufas commonly form crusts around tubular molds, which are up to several millime- ters in length and range in diameter from tens of microns to about 1 mm. The tubular molds are surrounded by coatings composed of dark micrite, and then calcite microspar, which may have a radial fabric. These textures probably represent car- bonate precipitation around plant material, such as green algae and moss. A second petrographic texture exhibited by tufas consists of clusters of dark micrite, tens of microns in diameter, organized into millimeter-scale "bundles of grapes" similar in appearance to the "arborescent shrubs" of Demicco and Hardie (1994). Yet a third tufa fabric consists of bladed calcite crys- tals, up to 1 mm long, grown from thin carbonate crusts. The calcite crystals have upward-radiating growth fabrics, with the number of calcite crystals decreasing away from the original growth substrate. This fabric is interpreted to have formed by direct chemical precipitation of calcite. Elemental mapping of samples for Mg and Sr using an electron microprobe indicates that chemical zonation of these elements coincides with tex- tural and mineralogical (calcite and minor aragonite) zones in the tufas, which confirms their unrecrystallized primary origin. The above petrographic and geochemical studies indicate that the tufas selected for dating have not recrystallized nor have they undergone major later cementation, both of which could complicate interpretation of age dates (Benson et al., 1990; Ku etal., 1998). Eleven samples of tufa from the 90 m shoreline were dated by U-series methods. Of the 11 samples, 7 are 185-150 Ka, which closely agrees with the results (200-120 Ka) obtained by Hooke and Lively (1979) and Hooke and Dorn (1992) on sam- ples from the same locations and the same general elevations. The dates also correlate with perennial lake sediments ob- served in core DV93-1 from 186 to 128 Ka; thus they serve as an independent check confirming the reliability of the dating methods. These tufas indicate lake depths of up to 335 m, if faulting is ignored and if 130-160 m of basinal sediments are added to reach equivalent-age lacustrine sediments in core DV93-1. Four high-shoreline tufas, from 73 to 90 m above sea level, dated between 194 and 216 Ka, are older than any sedi- ments from core DV93-1. These oldest tufas suggest the exist- ence of yet a third perennial lake stage in Death Valley at ap- proximately 200 Ka, which is consistent with earlier age-dating (Hooke and Lively, 1979; Hooke and Dorn, 1992). Summary and Conclusions The 200 Ky paleoclimate record of Death Valley from core DV93-1 is dominated by two dry and/or warm and wet/cold cy- cles that occurred on a 100 Ky time scale (Figure 4A,D). These cycles began with mudflat deposits (192 Ka to bottom of core, and 120?60 Ka). Wetter conditions, colder climates, or both, produced greater effective moisture, with saline-pan and ephemeral-saline-lake sediments (192-186 Ka, 60-35 Ka) overlying the mudflat sediments. Eventually, enough water was supplied to Death Valley to sustain perennial lakes, fluctuating from fresh to saline (186-120 Ka, 35-10 Ka, respectively). The perennial-lake phase ended, the lakes desiccated (120 and 10 Ka), and dry conditions led to the accumulation of mudflat de- posits, completing the cycle. Of particular importance are the major lacustrine phases from 35 to 10 Ka and from 186 to 120 Ka (marine oxygen isotope stages 2 and 5e/6, respectively), which represent markedly colder and wetter conditions than those in modern Death Valley. Of the two perennial-lake peri- ods, the penultimate glacial lake was deeper and lasted far longer than the last glacial lake. In summary, the Death Valley climate record is divided into six periods (Figure 4). (1) From 192 to 186 Ka, the climate was dry, dominated by saline pans and shallow ephemeral 118 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES lakes. Inflow and evaporation were sufficient to allow the rapid accumulation of 25 m of salt in 6 Ky. ThMAX values are 30? C or greater for 14 of 19 stratigraphic intervals, indicating that temperatures were similar to the modern late April to early May ThMAX values. (2) The climate from 186 to 120 Ka was relatively cool to cold, with abundant inflow to continually maintain a series of fluctuating perennial lakes, at times greater than 300 m deep and having a salinity less than 3000 parts per million. Temperatures during this period probably av- eraged 10?-15? C below the modern. If the effects of faulting are ignored, the ages of high-shoreline tufa indicate that Death Valley at that time commonly had lakes hundreds of meters deep. During the latest phase of this 66 Ky lacustrine period, from 128 to 120 Ka, reduced inflow or more saline inflow pro- duced hypersaline conditions, which ultimately led to basin desiccation at 120 Ka. Slightly higher fluid-inclusion tempera- tures (ThMAX = 32? C), comparable to the modern late April to early May ThMAX? mark the terminal lake stage at 120 Ka. (3) From 120 to 60 Ka, a dry period dominated by mudflat en- vironments prevailed. Limited data from fluid inclusions in two halite intervals suggest generally cool conditions, 6?-7? C below modern late April to May temperatures, ~92 and ?80 Ka; however, ThMAX from 10 halite intervals around 100 Ka are close to the modern late April to early May values, which suggests a warm climate but still cooler than the modern at 100 Ka. (4) From 60 to 35 Ka, the climate was cold, with moderate aridity; saline pans dominated, but there was enough inflow to accumulate thick salts relatively rapidly and to maintain shal- low saline lakes approximately 54-50 Ka. Homogenization temperatures (ThMAX between 23? C and 28? C) are 6?-ll? C below the modern late April to May values. A relatively low temperature interval, 10? C below the modern, existed -50 to -47 Ka. (5) From 35 to 10 Ka, the climate was cold and wet; plentiful inflow fed a perennial lake that was commonly saline but at times became up to 90 m deep (Dorn et al., 1990) with salinities less than 3000 parts per million. Overall, for this 25 Ky period, fluid inclusion ThMAX data suggest brine tempera- tures were 4?-15? C below modern late April to May values. During the latest lake phase, from 15 to 11 Ka, ThMAX values of halites were only 4? C below modern values, whereas ThMAX values decreased to 7? C below the modern late April to May temperatures in halites from ~11 to ~10 Ka, signifying cooler temperatures during this period. (6) From 10 to 0 Ka, dry arid conditions featured mudflat and/or saline-pan environ- ments, minor inflow of meteoric waters, and a water table nor- mally below the surface. Acknowledgments The study of core DV93-1 was a collaborative project funded by the National Science Foundation Global Change Research Program, grant EAR-9218717. Jianren Li and Chris Brown (State University of New York at Binghamton) and Sheila Rob- erts and Wenbo Yang (University of Calgary, under the direc- tion of Ronald Spencer) studied the core as part of their gradu- ate research. Teh-Lung Ku and Shangde Luo (University of Southern California) dated core DV93-1 and shoreline tufas. Several members of the United States Geological Survey were involved in various phases of the project: Art Clark, drilling; Rick Forester, analysis of ostracodes and modern waters, and interpretation of the paleoclimate record; Platt Bradbury, analy- sis of diatoms; and Zell Peterman, strontium isotope analyses of sulfates and carbonates. To all these colleagues, many thanks! Literature Cited Anderson, D.E. 1998. Late Quaternary Paleohydrology, Lacustrine Stratigraphy, Fluvial Geomorphology, and Modern Hydroclimatology of the Amargosa River/Death Valley Hydrologic System, California and Nevada. 521 pages. Doctoral dissertation, Department of Earth Sciences, Univer- sity of California, Riverside, California. Anderson, D.E., and S.G.Wells 1997. Late Pleistocene Death Valley Lakes: Subsurface Records of Chang- ing Paleoenvironments. In R.E. Reynolds and i. Reynolds, editors, Death Valley: The Amargosa Route. San Bernardino County Mu- seum Association Quarterly, 44(2):89-92. Redlands, California: San Bernardino County Museum Association. Anderson, K.C., and S.G. Wells 1996. Late Pleistocene Lacustrine Record of Lake Dumont and the Rela- tionship to Pluvial Lakes Mojave and Manly. Geological Society of America Abstracts with Programs, 28:A-458. 1997. Late Pleistocene and Holocene Valley-Fill Deposits of Lake Du- mont. In R.E. Reynolds and J. Reynolds, editors, Death Valley: The Amargosa Route. San Bernardino County Museum Association Quarterly, 44(2):29-32. Redlands, California: San Bernardino County Museum Association. Benson, L.V., D.R. Currey, R.I. Dorn, K.R. Lajoie, C.G. Oviatt, S.W. Robin- son, G.I. Smith, and S. Stine 1990. Chronology of Expansion and Contraction of Four Great Basin Lake Systems during the Past 35,000 Years. Palaeogeography, Palaeocli- matology.Palaeoecology, 78:241-286. Blackwelder, E. 1933. Lake Manly: An Extinct Lake of Death Valley. Geographical Re- view, 23:464-471. 1954. Pleistocene Lakes and Drainage in the Mojave Region, Southern California. In R.H. Jahns, editor, Geology of Southern California. Bulletin of the California Division of Mines, 170:35-^10. San Fran- cisco: California Division of Mines and Geology. Brown, W.J., and M.R. Rosen 1995. Was There a Pliocene-Pleistocene Fluvial-Lacustrine Connection between Death Valley and the Colorado River? Quaternary Re- search, 43:286-296. Carpelan, L.H. 1958. The Salton Sea: Physical and Chemical Characteristics. Limnology and Oceanography, 3:373-386. NUMBER 33 119 Demicco, R.V., and L.A. Hardie 1994. Sedimentary Structures and Early Diagenetic Features of Shallow Marine Carbonate Deposits. 1265 pages. Tulsa, Oklahoma: SEPM (Society of Sedimentary Geology; formerly the Society of Eco- nomic Paleontologists and Mineralogists), Atlas Series, de Voogd, B., L. Serpa, L. Brown, E. Hauser, S. Kaufman, J. Oliver, B. Troxel, J. Willemin, and L.A. Wright 1986. Death Valley Bright Spot: A Midcrustal Magma Body in the South- ern Great Basin, California. Geology, 14:64-67. Dorn, R.I. 1989. Cation-Ratio Dating of Rock Varnish: A Geographical Perspective. Progress in Physical Geography, 13:559?596. Dorn, R.I., A.J.T. Ml, D.J. Donahue, T.W. Linick, and L.J. Toolin 1989. Accelerator Mass Spectrometry Radiocarbon Dating of Rock Var- nish. Bulletin of the Geological Society of America, 101:1363-1372. 1990. Latest Pleistocene Lake Shorelines and Glacial Chronology in the Western Basin and Range Province, U.S.A.: Insights from AMS Ra- diocarbon Dating of Rock Varnish and Paleoclimatic Implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 78:315?331. Fitzpatrick, J.A., and J.L. Bischoff 1993. Uranium-Series Dates on Sediments of the High Shoreline of Pana- mint Valley, California. United States Geological Survey, Open-File Report, 93-232:1-15. Forester, R.M. 1983. Relationship of Two Lacustrine Ostracode Species to Solute Com- position and Salinity: Implication for Paleohydrochemistry. Geol- ogy, 11:435-438. 1987. Late Quaternary Paleoclimate Records from Lacustrine Ostracodes. In W.F. Ruddiman and H.E. Wright, editors, North America and Ad- jacent Oceans during the Last Deglaciation. Geology of North America, K-3:261-276. Boulder, Colorado: Geological Society of America. Gale, H.S. 1914. Salines in the Owens, Searles, and Panamint Basins, Southeastern California. Bulletin of the United States Geological Survey, 580-L: 251-323. Goldstein, R.H., and T.J. Reynolds 1994. Systematics of Fluid Inclusions in Diagenetic Minerals. Society of Economic Paleontologists and Mineralogists Short Course 31, 184 pages. Tulsa: Society of Economic Paleontologists and Mineralo- gists. Hammer, U.T. 1986. Saline Lake Ecosystems of the World. 616 pages. Dordrecht: Dr. W. Junk Publishers. Hardie, L.A., J.P. Smoot, and H.P. Eugster 1978. Saline Lakes and Their Deposits: A Sedimentological Approach. In A. Matter and M.E. Tucker, editors, Modern and Ancient Lake Sed- iments. International Association of Sedimentologists Special Publi- cation, 2:7-41. Oxford, United Kingdom: Blackwell Scientific Publications. Hooke, R. LeB. 1972. Geomorphic Evidence for Late-Wisconsin and Holocene Tectonic Deformation, Death Valley, California. Bulletin of the Geological Society of America, 83:2073-2098. Hooke, R. LeB., and R.I. Dorn 1992. Segmentation of Alluvial Fans in Death Valley, California: New In- sights from Surface Exposure Dating and Laboratory Modelling. Earth Surface Processes andLandforms, 17:557-574. Hooke, R. LeB., and R.S. Lively 1979. Dating of Late Quaternary Deposits and Associated Tectonic Events by U/Th Methods, Death Valley, California. Final Report for Na- tional Science Foundation Grant EAR-7919999,21 pages. Washing- ton, D.C. [Unpublished.] Hunt, C.B., and D.R. Mabey 1966. Stratigraphy and Structure, Death Valley, California. United States Geological Survey Professional Paper, 494-A: 1-162. Jannik, N.O., F.M. Phillips, G.I. Smith, and D. Elmore 1991. A Cl-36 Chronology of Lacustrine Sedimentation in the Pleistocene Owens River System. Bulletin of the Geological Society of America, 103:1146-1159. Jones, B.F. 1966. Geochemical Evolution of Closed Basin Water in the Western Great Basin. In J.L. Rau, editor, Second Symposium on Salt, 1:181-200. Cleveland, Ohio: Northern Ohio Geological Society. Ku, T.-L., S. Luo, T.K. Lowenstein, J. Li, and R.J. Spencer 1998. U-Series Chronology of Lacustrine Deposits in Death Valley, Cali- fornia. Quaternary Research, 50:261-275. Li, J., T.K. Lowenstein, and I.R. Blackburn 1997. Responses of Evaporite Mineralogy to Inflow Water Sources and Climate during the Past 100 k.y. in Death Valley, California. Bulletin of the Geological Society of America, 109:1361-1371. Li, J., T.K. Lowenstein, C.B. Brown, T.-L. Ku, and S. Luo 1996. A 100 ka Record of Water Tables and Paleoclimates from Salt Cores, Death Valley, California. Palaeogeography, Palaeoclima- tology, Palaeoecology, 123:179?203. Lowenstein, T. K., and L.A. Hardie 1985. Criteria for the Recognition of Salt-Pan Evaporites. Sedimentology, 32:627-644. Lowenstein, T.K., J. Li, and C.B. Brown 1998. Paleotemperatures from Fluid Inclusions in Halite: Method Verifica- tion and a 100,000 Year Paleotemperature Record, Death Valley, California. Chemical Geology, 150:223-245. Luo, S., and T.-L. Ku 1991. U-Series Isochron Dating: A Generalized Method Employing Total- Sample Dissolution. Geochimica et Cosmochimica Acta, 55: 555-564. Means, T.H. 1932. Death Valley. Sierra Club Bulletin, 17':67-76. Meek, N. 1989. Geomorphic and Hydrologic Implications of the Rapid Incision of Afton Canyon, Mojave Desert, California. Geology, 17:7-10. 1997. The Elevations of Shorelines in Death Valley. In R.E. Reynolds and J. Reynolds, editors, Death Valley: The Amargosa Route. San Ber- nardino County Museum Association Quarterly, 44(2):75-84. Red- lands, California: San Bernardino County Museum Association. Morrison, R.B. 1991. Quaternary Stratigraphy, Hydrologic, and Climate History of the Great Basin, with Emphasis on Lakes Lahonton, Bonneville, and Tecopa. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. Geology of North America, K-2:283-320. Boul- der, Colorado: Geological Society of America. Noble, L.F. 1926. Note on a Colemanite Deposit near Shoshone, California. Bulletin of the United States Geological Survey, 785-D:63-75. Phillips, F.M., M.G. Zreda, L.V. Benson, M.A. Plummer, D. Elmore, and P. Sharma 1996. Chronology for Fluctuations in Late Pleistocene Sierra Nevada Gla- ciers and Lakes. Science, 274:749-751. Phillips, F.M., M.G. Zreda, S.S. Smith, D. Elmore, P.W. Kubik, and P. Sharma 1990. Cosmogenic Chlorine-36 Chronology for Glacial Deposits at Bloody Canyon, Eastern Sierra Nevada. Science, 248:1529-1532. Roberts, S.M., and R.J. Spencer 1995. Paleotemperatures Preserved in Fluid Inclusions in Halite. Geochimica et Cosmochimica Acta, 59:3929?3942. Roberts, S.M., R.J. Spencer, and T.K. Lowenstein 1994. Late Pleistocene Saline Lacustrine Sediments, Badwater Basin, Death Valley, California, In A.J. Lomando, B.C. Schreiber, and P.M. Harris, editors, Lacustrine Reservoirs and Depositional Systems. So- ciety for Sedimentary Geology Core Workshop, 19:61-103. Tulsa, Oklahoma: Society for Sedimentary Geology. 120 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Roberts, S.M., R.J. Spencer, W. Yang, and H.R. Krouse 1997. Deciphering Some Unique Paleotemperature Indicators in Halite- Bearing Saline Lake Deposits from Death Valley, California, USA. Journal of Paleolimnology, 17:101-130. Serpa, L., B. de Voogd, L. Wright, J. Willemin, J. Oliver, E. Hauser, and B. Troxel 1988. Structure of the Central Death Valley Pull-Apart Basin and Vicinity from COCORP Profiles in the Southern Great Basin. Bulletin of the Geological Society of America, 100:1437-1450. Smith, G.I. 1979. Subsurface Stratigraphy and Geochemistry of Late Quaternary Evaporites, Searles Lake, California. United States Geological Sur- vey Professional Paper, 1043:1-130. 1984. Paleohydrologic Regimes in the Southwestern Great Basin, 0-3.2 my Ago Compared to Other Long Records of "Global" Climate. Quaternary Research, 22:1-17. Smith, G.I., V.J. Barczak, G.F. Moulton, and J.C. Liddicoat 1983. Core KM-3, a Surface-to-Bedrock Record of Late Cenozoic Sedi- mentation in Searles Valley, California. United States Geological Survey Professional Paper, 1256:1-24. Smith, G.I., I. Friedman, R.J. McLaughlin 1987. Studies of Quaternary Saline Lakes, HI: Mineral, Chemical, and Iso- topic Evidence of Salt Solution and Crystallization Processes in Owens Lake, California, 1969-1971. Geochimica et Cosmochimica Ada, 51:811-827. Smith, G.I., and W.P. Pratt 1957. Core Logs from Owens, China, Searles, and Panamint Basins Cali- fornia. Bulletin of the United States Geological Survey, 1045-A: 1-62. Smith, R.S.U. 1978. Pluvial History of Panamint Valley, California; A Guidebook for Friends of the Pleistocene, Pacific Cell. 36 pages. Houston: Univer- sity of Houston. Smoot, J.P., and B. Castens-Seidell 1994. Sedimentary Features Produced by Efflorescent Salt Crusts, Saline Valley and Death Valley, California. In R.W. Renaut and W.M. Last, editors, Sedimentology and Geochemistry of Modern and Ancient Saline Lakes. Society of Economic Mineralogists and Paleontolo- gists, Special Publications, 50:73-90. Smoot, J.P., and T.K. Lowenstein 1991. Depositional Environments of Non-Marine Evaporites. In J.L. Melvin, editor, Evaporites, Petroleum, and Mineral Resources. Elsevier Developments in Sedimentology, 50:189?347. Amsterdam: Elsevier. Spencer, R.J., M.J. Baedecker, H.R Eugster, R.M. Forester, M.B. Goldhaber, B.F. Jones, K. Kelts, J. Mckenzie, D.B. Madsen, S.L. Rettig, M. Rubin, and C.J. Bowser 1984. Great Salt Lake, and Precursors, Utah: The Last 30,000 Years. Con- tributions to Mineralogy and Petrology, 86:321-334. Spencer, R.J., H.P. Eugster, and B.F. Jones 1985a. Geochemistry of Great Salt Lake, Utah II: Pleistocene-Holocene Evolution. Geochimica et Cosmochimica Acta, 49:739-747. Spencer, R.J., H.P. Eugster, B.F. Jones, and S.L. Rettig 1985b. Geochemistry of Great Salt Lake, Utah I: Hydrochemistry since 1850. Geochimica et Cosmochimica Acta, 49:727-737. Bonneville Basin Lacustrine History: The Contributions of G.K. Gilbert and Ernst Antevs Charles G. Oviatt ABSTRACT The careful observations and eloquent writings of Grove Karl Gilbert in the late 1800s laid the framework for most subsequent studies of lacustrine history in the Bonneville Basin. By the time of the 1942 symposium on the glacial and postglacial history of the Great Basin, his work had been largely unchallenged by the scientific community. Ernst Antevs, who participated in that sym- posium, reinterpreted some of Gilbert's observations and contrib- uted both a global perspective and a hypothesis involving atmospheric circulation to studies of Lake Bonneville and other Great Basin Quaternary lakes. Fifty years after the 1942 sympo- sium, the contributions of Gilbert and Antevs are still at the core of state-of-the-art research on Great Basin Quaternary lakes. Introduction Systematic studies of ancient lakes in the Bonneville Basin began with the work of Grove Karl Gilbert, who published his monograph on Lake Bonneville in 1890. By the time of the 1942 symposium on the glacial and postglacial history of the Great Basin, Gilbert's reconstruction of the lake's history had been reinterpreted by only a few people (Sack, 1989). Ernst Antevs, who contributed to the 1942 symposium, was one of those few. In subsequent decades, many new chronologies and hypotheses about the lake have been proposed. This paper is a brief review of the knowledge of the Lake Bonneville basin as of 1942, with emphasis on the contributions of G.K. Gilbert and Ernst Antevs, and a critique of those contributions from the per- spective of the late 1990s. ACKNOWLEDGMENTS.?Ellis Yochelson and an anonymous reviewer provided useful suggestions for improving the manu- script. I am grateful to all my friends, colleagues, professors, Charles G. Oviatt, Department of Geology, Thompson Hall, Kansas State University, Manhattan, Kansas 66506-3201, USA. Reviewers: Ellis Yochelson, Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0137, USA; and one anonymous reviewer. and students, from whom I have learned about Lake Bonneville and the Great Basin. I dedicate this paper, however, to the one person who has taught me more than all the others and who continues to inspire me with his intelligence, curiosity, knowl- edge, and good humor?Don Currey. G.K. Gilbert Lakes began forming in the Bonneville Basin during the Miocene with the initiation of tectonic extension in the region (Miller, 1997). Gilbert (1890) noted a few exposures of Tertiary lake deposits in the Bonneville Basin, but very little was known about them at the time. Based on the meager evidence he had available, Gilbert made an insightful comment concerning the Tertiary history of the basin. He thought it was "possible that the Bonneville Basin continuously held a lake which, as the land rose and fell unequally, was expanded and contracted, now in one direction, now in another" (Gilbert, 1890:101). Later studies have confirmed this general statement, but a detailed history of Tertiary (primarily Neogene) lakes is just beginning to emerge (Miller, 1997; Thompson et al., 1997; Davis, 2002). Gilbert's great contributions were the careful study of sedi- mentary deposits and shorelines of Lake Bonneville and the proposal of hypotheses to explain its history. Gilbert (1890) re- ported geomorphic and stratigraphic evidence that suggested Lake Bonneville had risen twice to high levels in the late Pleis- tocene after a long period of low lakes and alluvial-fan deposi- tion on piedmont slopes. His conclusions were based on his in- terpretation of sediments deposited on the bottom of Lake Bonneville (the Bonneville beds). Two of his study sites were along the Old River Bed, and one was near the town of Leam- ington, in west-central Utah (Figure 1). Gilbert defined the stratigraphic units at these sites by measuring the sediment lay- ers, which he then traced or correlated from outcrop to outcrop throughout the basin. Gilbert argued that the shorelines of the ancient lake and the stratigraphy of the offshore deposits to- gether tell the story of rising and falling lake levels. OLD RIVER BED.?The Old River Bed is an abandoned river valley cut into the Bonneville beds. Gilbert demonstrated that it 121 122 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 1.?Map of Lake Bonneville at its highest stage, the Bonneville shore- line. SLC = Salt Lake City. Arrows point to the areas at Old River Bed and Leamington illustrated in Figures 2 and 4, respectively. A heavy line marks the Old River Bed strait, which is a narrow constriction between Keg Mountain (KM) to the west and the Simpson Mountains (SM) to the east. was formed during the regressive (falling) phase of the lake by a river that flowed from the Sevier Basin in the south to the Great Salt Lake basin in the north (Figures 1, 2). The exposures of the Bonneville beds and the shorelines along the Old River Bed are some of the finest in the basin, and Gilbert took advan- tage of this by studying them for a total of about two weeks during several different trips in 1879 and 1880 (Hunt, 1982; Oviatt and McCoy, 1988). Gilbert measured and described stratigraphic sections of the Bonneville beds at two sites at the Old River Bed. The Lower FIGURE 2.?Map of the Old River Bed in the vicinity of Gilbert's Lower and Upper River Bed sections. River Bed section was located approximately where the Over- land Stage, or Pony Express, Road crossed the Old River Bed (Figure 2). At this section, Gilbert defined two stratigraphic units, the Yellow Clay and the White Marl. Because both of these units are fine grained, he interpreted them as deep-water depos- its. The Yellow Clay was the thicker of the two and occurred be- low the White Marl, so he concluded that it was deposited during a relatively long period of deep-lake conditions, called the Yel- low Clay Epoch, prior to the White Marl Epoch (Figure 3). The White Marl was deposited during the most recent deep- water period and included the time of formation of the promi- nent Bonneville, Provo, and Stansbury shorelines. A thin sand unit containing gastropod shells separates the Yellow Clay from the White Marl. Gilbert interpreted the base of this sand as marking an unconformity, or period of time when the lake was below the level of the section and was not depositing sediments there. In his interpretation, the sand and gastropods were depos- ited as the lake rose during the White Marl Epoch. Gilbert re- ferred to the Lower River Bed section as a typical section of the Bonneville beds, which therefore recorded the major events in the history of the lake (Figure 3). Gilbert's Upper River Bed section displayed the Yellow Clay, the White Marl, and several other stratigraphic units. The Yel- low Clay and the White Marl are overlain at this section by an NUMBER 33 123 FEET 1000 850 400 330 YELLOW CLAY\ ALLUVIUM YELLOW CLAY YELLOW CLAY WHITE CLAY ALLUVIUM EROSION AND FIRST GRAVEL UNCONFORM. WHITE MARL & UROSION AND ' LOWER SAND 1 SECOND GRAVEL/ UPPER SAND THIRD GRAVEL, EROSION OF OLD RIVER BED, AND LATER ALLUVIUM WHITE MARL EROSION & ALLUVIUM A B C BONNEVILLE SHORE LINE LEAMINGTON UPPER RIVER BED LOWER RIVER BED WATER OF GREAT SALTLAKE FIGURE 3.?Modification of "Diagram of Lake Oscillations inferred from Deposits and Erosions" from Gilbert (1890, fig. 30). The horizontal axis can be read as time, with a variable scale, or as vertical distance (also with variable scale) in each of the three stratigraphic sections (Lower River Bed, Upper River Bed, and Leamington). The White Clay at Leamington is equivalent to the White Marl. Letters A, B, and C represent the three lake high- stands interpreted by Gilbert (A = Yellow Clay Epoch; B = White Marl Epoch; C = minor regressive-phase fluc- tuation). additional sand unit, termed the Upper Sand, which Gilbert con- sidered as representing a short-lived, minor rise in lake level (Figure 3). Of primary importance in Gilbert's interpretation of the Up- per River Bed section was a gravel unit, which he referred to as the First Gravel. The First Gravel is more than 15 m thick along the east side of the exposures at the Upper River Bed section (Figure 4), but it thins abruptly to the west to less than 1 m, thus leading Gilbert to refer to the gravel as a "wedge" in his field notes (Hunt, 1982:128). Gilbert called the First Gravel a "bar" in his field notes and described large quantities of tufa cementing the gravel (Hunt, 1982:128), thus indicating that he regarded the gravel as having been deposited in a lake environment. In United States Geo- graphical Survey (USGS) Monograph 1, Gilbert (1890:195) wrote that "the form of the [gravel mass] described, the associ- ated tufa, and the preponderance of boulders of local derivation, indicate shore action, but it is possible that an interlacustrine [that is, between the Yellow Clay and White Marl Epochs] river was the agent of transportation." References in other places in his monograph to thick alluvial deposits at this horizon suggest that by the time Monograph 1 was published Gilbert had de- cided that the First Gravel at the Upper River Bed section was alluvial rather that lacustrine gravel, despite the field evidence he had observed that the gravel had a lacustrine origin. Regardless of whether the First Gravel was interpreted as al- luvial or lacustrine, it indicated to Gilbert that the lake fell to a low level relative to the deep-lake intervals of the Yellow Clay and White Marl Epochs (Figure 3). The tops of both River Bed sections represented the final regression of the lake and local deposition of alluvium. LEAMINGTON.?The section at Leamington, Utah, is approx- imately 140 m higher in altitude than the Old River Bed sec- tions, and it is close to the Bonneville shoreline (Figure 1). At Leamington, Gilbert noted stratigraphic units that he correlated with the Yellow Clay and White Marl. They were thicker and sandier than at the Old River Bed. Because Leamington is lo- cated at the mouth of the Sevier River where it enters the Sevier Desert, the river delivered large volumes of fine-grained sediment (silt, sand) to the lake margin in this area (Figure 5). Gilbert interpreted gravel between the Yellow Clay and the White Marl (e.g., White Clay in Figure 3) as alluvium, and he correlated this gravel with the unconformity at the Lower River Bed section and with the First Gravel at the Upper River Bed section. Leamington is too high in altitude to record events af- ter the lake fell from the Bonneville shoreline (Figure 3). GILBERT'S CONCLUSIONS.?By combininsg the stratigraphic evidence from the Old River Bed and Leamington sections with information from the shorelines and other geologic evidence in the basin, Gilbert deduced a lake history illustrated in Figure 6. 124 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 4.?Aerial photograph of Gilbert's Upper River Bed section and surrounding area. A = Yellow Clay and White Marl exposed in the section described by Gilbert (1890); B = lacustrine gravel spit composed of Gilbert's First Gravel (note the eastward-recurved distal end of the spit; longshore current direction was from the north); C = eolian sand dunes composed of Gilbert's Upper Sand. FIGURE 5.?Gilbert's (1890) fig. 28, about which he stated, "Section showing succession of Lacustrine and Allu- vial Deposits at Leamington, Utah. 1. Paleozoic sandstone. 2. The Yellow Clay (Lower Bonneville). 3. Wedge of alluvial gravel. 4. The White Marl [White Clay of Figure 3] (Upper Bonneville). 5. Recent alluvial gravel. 6. Bonneville shore notch, with recent talus." NUMBER 33 125 FIGURE 6.?Modified from Gilbert's (1890) fig. 34, about which he stated, "The upper and lower horizontal lines represent the horizons of the Bonneville shore and the surface of Great Salt Lake. Horizontal distances represent time, counted from left to right. The curve represents the height of the oscillating water surface, and the shaded area indicates ignorance." Gilbert's curve shows the Yellow Clay Epoch (YC), the White Marl Epoch (WM), the Provo shoreline (PS), and the regressive-phase fluctuation he deduced from the Upper Sand at the Upper River Bed section (US). The Yellow Clay Epoch occurred after a long period of shallow lakes and alluvial-fan deposition in the basin. By Gilbert's esti- mate, the Yellow Clay lake did not overflow, and it persisted much longer than the White Marl lake. Gilbert's interpretation of the First Gravel at the Upper River Bed section and the gravel at Leamington as alluvium led him to conclude there had been a profound lake-level lowering between the Yellow Clay and the White Marl Epochs (Figure 6). This idea strongly influ- enced Antevs and most later workers. According to Gilbert, the White Marl Epoch included the rise to overflow at Red Rock Pass, Idaho (Figure 1), and the forma- tion of the Bonneville shoreline. Catastrophic failure of the al- luvial-fan dam at Red Rock Pass resulted in the Bonneville Flood and the subsequent formation of the Provo shoreline. A minor rise in lake level during the regression from the Provo shoreline at about the altitude of the Upper River Bed section deposited the Upper Sand; the lake subsequently dropped to levels near that of the modern Great Salt Lake. Gilbert appar- ently thought the Stansbury shoreline formed during the regres- sive phase of Lake Bonneville, that is, later than the formation of the Provo shoreline, although he did not explicitly state this in his monograph. This reconstructed history was the starting point for the work of Ernst Antevs. Ernst Antevs During the first half of the twentieth century, Ernst Antevs published a number of papers concerning Lake Bonneville and Great Basin Quaternary history. He also contributed to the 1942 symposium. Antevs was convinced that lake fluctuations in the Bonneville Basin were correlated with growth and decay of Northern Hemisphere ice sheets and mountain glaciers. Despite the lack of direct evidence for the ages of either the glacial or lacustrine deposits, he presented careful arguments for ways in which they simultaneously varied in size during the late Pleis- tocene (Figure 7). Evidence from the mouths of Little Cotton- wood and Bells Canyons in the Wasatch Mountains near Salt Lake City indicated that Lake Bonneville reached its highest level, the Bonneville shoreline, shortly after the glaciers in those canyons reached their maximum down-valley extent (An- tevs, 1948), so that correlation with mountain glaciers in the Si- erra Nevada and ice sheets in Canada was likely. Antevs spent less time in the field examining the stratigraphic sections at the Old River Bed and Leamington than did Gilbert, but he reinterpreted Gilbert's descriptions at the three primary sections, focusing especially on those at the Old River Bed. An- tevs thought the Yellow Clay Epoch was correlated with an old period of glaciation in the Rocky Mountains and Sierra Nevada (Bull Lake and Tahoe glaciations, respectively), and he did not include it in his discussion of Lake Bonneville history in his symposium paper (Figure 7) (Antevs, 1948). His interpretation of the depositional environment of the Yellow Clay as having been a deep lake was essentially the same as Gilbert's. Western Ice (Cordilleran, Keewatin) Eastern Ice (Labrador, Patricia) o ? Q. CD FIGURE 7.?Fluctuations of Lake Bonneville (L.B.), Lake Lahontan (L.L.) in western Nevada, and Northern Hemisphere ice sheets (modified from Antevs, 1948:171). D.L. = Dendritic Lake; L.P. = Lake Provo; L.S. = Lake Stansbury; T.L. = Thinolite Lake. Letters B and C represent lake highstands and correla- tions with Gilbert's interpretations of Lake Bonneville history (see Figure 3). 126 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Antevs regarded the regressive-phase lake rise, which Gilbert had deduced from the Upper Sand at the Upper River Bed sec- tion, as a major event in the history of the lake. He suggested that during this event the lake reoccupied the Provo shoreline and overflowed for a second time at Red Rock Pass, Idaho. In addition, Antevs thought the Stansbury shoreline formed during the interval between the two Provo occupations. He then corre- lated fluctuations of Lake Bonneville with fluctuations of Lake Lahontan in western Nevada, and with the Cordilleran and Lau- rentide ice sheets (Figure 7). Antevs (1948) argued that during the late Pleistocene the large North American ice sheets, particularly the Cordilleran ice sheet, attained a size large enough to deflect the polar front far to the south of its mean position in historic time. As a conse- quence, frequent storms carried moisture from the Pacific Ocean across the Sierra Nevada to the Great Basin to create the large pluvial lakes. Thus, the filling and emptying of the lake basins were correlated with the expansion and contraction of the Northern Hemisphere ice sheets. He also thought that the astronomical hypothesis of climate change formulated by Milu- tin Milankovitch was probably valid (Antevs, 1938). Added to this was his fundamental belief that major climatic events had to be global in extent (e.g., Antevs, 1955). Accordingly, he was eager to find correlative climatic events across North America and in northern Europe. Late Twentieth Century Views Machette and Scott (1988) and Sack (1989) have reviewed how interpretations of Lake Bonneville have changed during the second half of the twentieth century. In this paper, I focus only on current interpretations of the work of Gilbert and Antevs. With a few possible exceptions, all the lake sediments ob- served by Gilbert, including the Yellow Clay and the White Marl, are now recognized to have been deposited during a sin- gle major lake cycle between about 30 thousand years (Ky) be- fore present (B.P.) and 10 Ky B.p. (Scott et al., 1983; Oviatt, 1987; Oviatt et al., 1994). This is the Bonneville lake cycle, or Lake Bonneville. Some Pleistocene lake cycles older than 30 Ky B.P. have been given separate names (e.g., Cutler Dam, Lit- tle Valley, Pokes Point; Figure 8). Massive shoreline deposits, erosional shorelines, deltas, and offshore sediments are prominent in the Bonneville Basin. Most of these were almost certainly produced during the Bonneville lake cycle. In a refinement of Gilbert's initial work, prominent shorelines named and mapped by him have been dated and linked to the offshore deposits, and an accurate radio- carbon chronology has been established that can now be used in many ways as a predictive model (Figure 9) (Oviatt et al., 1992; Oviatt, 1997). The major shorelines have been mapped with precision (Currey, 1982), and Currey (1990) has shown how the extensive knowledge about the Bonneville Basin serves as a prime example of the landforms and processes typical of semi- desert basins. In addition to the accumulation of abundant numerical-age data and numerous other advances in understanding the Bonne- ville lake cycle, Gilbert's sections at the Old River Bed and Leamington have been reinterpreted. The Yellow Clay is now recognized to be a shallow lacustrine deposit; it consists essen- tially of fine-grained deltaic sediment deposited early in the rise of Lake Bonneville (Oviatt, 1987). During deposition of the Yellow Clay the Sevier River flowed across the flat floor of the Sevier Desert, through the Old River Bed strait, and transported sand, silt, and clay to the margin of the rising lake in the Great Salt Lake basin ("Main Body" in Figure 1). In light of this in- terpretation, Gilbert's dilemma with explaining why the Yellow Clay is so much thicker and contains so much more silt and sand than the White Marl (Gilbert, 1890:200-209), is re- solved?the Yellow Clay and the White Marl were deposited in different depositional environments and should not be expected to look similar. Although deposits similar to the Yellow Clay are present at many localities in the basin, they represent dis- connected shallow-water depositional environments near the mouths of large rivers rather than a single, widespread, deep- water lake deposit. In contrast, the White Marl is the basin- wide stratigraphic unit that records the deep-water, or offshore, history of Lake Bonneville (Oviatt, 1997). 60 to 8 408 ro 0 A CO ? 8 as A 100 200 300 400 age (ka) 500 600 700 FIGURE 8.?Schematic diagram of middle and late Pleistocene lake cycles in the Bonneville Basin. Modified from Machette and Scott (1988, fig. 3) and Oviatt and Miller (1997, fig. 2). NUMBER 33 1600 .? 1500 cg CO ?. 1400 CDTJ 0 # 1300 CO 127 1200 lake level o carbonate ? wood ? organics A basaltic ash Stansbury Pilot Valley U2U3 ri Bonnevillec Gunnison Gilbert 35 30 25 20 15 age (ka) 10 FIGURE 9.?Chronology of Lake Bonneville fluctuations, modi- fied from Oviatt et al. (1992), Oviatt (1997), and Oviatt and Miller (1997). Elevations have been adjusted for isostatic rebound. Ul, U2, and U3 are transgressive-phase fluctuations (Oviatt, 1997). Radiocarbon ages were obtained for dissemi- nated organic carbon, wood or charcoal, carbonate (shell or tufa), and basaltic ashes. Figure 4 is an aerial photograph of the area around Gilbert's Upper River Bed section. It shows the prominent recurved lacustrine spit composed of Gilbert's First Gravel, which he had decided might be alluvium. The spit was deposited by long- shore currents in Lake Bonneville after deposition of the Yel- low Clay as the lake began to rise, following formation of the Stansbury shoreline (Figure 9); it does not represent river depo- sition during a lowstand between two high lake stands, as Gil- bert suggested in his monograph. At the Lower River Bed sec- tion the unconformity between the Yellow Clay and the White Marl, which Gilbert and Antevs interpreted as representing a long period of erosion and alluvial deposition, represents a short period (possibly less than 500 years) when the lake level fell about 45 m and the Stansbury shoreline formed (Oviatt, 1987; Oviatt et al., 1990). The gravel wedge between the Yel- low Clay and the White Marl at Leamington, which Gilbert in- terpreted as alluvium, is now interpreted as lacustrine gravel deposited during a fluctuation in lake level during the trans- gressive phase of the lake (U2 in Figure 9; Oviatt et al., 1994). Figure 4 also shows well-developed eolian sand dunes com- posed of Gilbert's Upper Sand. Although Gilbert recognized that the surface of this sand "may have gained or lost [thick- ness] by the action of the wind, which throws its surface into waves, and has caused it to bury at the north the exposure of the lower formations" (Gilbert, 1890:196), he thought that the lower part of the sand was lacustrine. Because no features of lacustrine origin have been found in the sand, his interpretation of a minor lake rise, and Antevs' suggestion of a second trans- gression to the Provo shoreline, are unlikely. Despite these changes in interpretation of Gilbert's strati- graphic sections, which are the results of building on his work and the work of others and the availability of superior technol- ogy for field and laboratory investigations, Gilbert laid the groundwork for all subsequent research on Lake Bonneville. Most of his observations from his horseback trips into the desert were highly accurate, even if some of his interpretations have been replaced by alternative hypotheses. It is clear that he thoroughly understood the lake beyond the context of his own century, and many of his conclusions have not been improved upon even after 100 years of subsequent research. One of Gil- bert's many long-lasting contributions to science was to dem- onstrate that paleoclimatic information can be obtained from studies of ancient lake deposits. Antevs also made his mark on Great Basin paleoclimate re- search. Modern research has demonstrated empirically the gen- eral validity of Antevs' conviction that the major climate changes that resulted in large pluvial lakes were global in scope. Although the details of his correlations with Lake Lahontan and with glacial events have not survived scientific tests, the urge to find global connections persists today. With refinements in glacial chronologies and modeling, including the realization that the Laurentide ice sheet was even more influen- tial than he thought, Antevs' general hypothesis that ice sheets diverted storm tracks into the Great Basin to feed the large plu- vial lakes is well supported by computer simulations and field data (COHMAP members, 1988; Kutzbach et al., 1993; Thompson et al., 1993). Future Work Since the days of Gilbert and Antevs, scientific advance- ments have resulted in major changes in interpretations, but the solid foundation laid by Gilbert and the refinements made by Antevs half a century later are still at the heart of our knowl- edge of Lake Bonneville and late Pleistocene climate change in the Great Basin (e.g., Oviatt, 1997; Benson, 1997). Despite many advances in understanding the history of Lake Bonneville and Great Salt Lake, a number of important questions remain only partially answered. These include questions about the tim- ing and duration of regression from the Provo shoreline and the 128 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES age of the Provo shoreline itself; the timing and character of the initial rise of Lake Bonneville between about 30 and 20 Ky B.P.; the chronology of fluctuations of Great Salt Lake during the last 10 Ky; the precise relationships between Lake Bonne- ville and Great Salt Lake and climate, including the relative im- portance of temperature and precipitation or other variables in determining lake size; and the details of pre-Bonneville lake cycles and their connections to global climate. These questions and others are being pursued in the Bonneville Basin, and, thanks in part to the powerful early contributions of Gilbert and Antevs, reliable answers are likely to be forthcoming in the near future. Literature Cited Antevs, E. 1938. Climatic Variations during the Last Glaciation in North America. Bulletin of the American Meteorological Society, 19:12-16. 1948. Climatic Changes and Pre-White Man. In The Great Basin, with Em- phasis on Glacial and Postglacial Times. Bulletin of the University of Utah, 38(20): 167-191. 1955. Geologic-Climatic Dating in the West. American Antiquity, 20:317- 335. Benson, L.V. 1997. Keynote Address: Global Change Modulation of Great Basin Aquatic History. [Paper presented at "GBASH 97: Great Basin Aquatic System History: Fifty Years of Geologic, Biologic, and Hy- droclimatic Progress in Late Cenozoic Aquatic System History," September 17-20, 1997, University Park Hotel, Salt Lake City, Utah.] [Unpublished.] COHMAP members 1998. Climatic Changes of the Last 18,000 Years: Observations and Model Simulations. Science, 241:1043-1052. Currey, D.R. 1982. Lake Bonneville: Selected Features of Relevance to Neotectonic Analysis. United States Geological Survey Open-File Report, 82- 1070: 30 pages. 1990. Quaternary Paleolakes in the Evolution of Semidesert Basins, with Special Emphasis on Lake Bonneville and the Great Basin, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology, 76:189- 214. Davis, O.K. 2002. Late Neogene Environmental History of the Northern Bonneville Basin: A Review of Palynological Studies. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:295? 307. Gilbert, G.K. 1890. Lake Bonneville. Monographs of the United States Geological Sur- vey, 1: 438 pages, figures 28, 30, 34. Hunt, C.B. 1982. Pleistocene Lake Bonneville, Ancestral Great Salt Lake, as De- scribed in the Notebooks of G.K. Gilbert. Brigham Young University Geology Studies, 29(1): 225 pages. Kutzbach, J.E., P.J. Guetter, P.J. Behling, and R. Selin 1993. Simulated Climatic Changes: Results of the COHMAP Climate- Model Experiments. In H.E. Wright, Jr., J.E. Kutzbach, T. Webb III, W.F. Ruddiman, F.A. Street-Perrott, and P.J. Bartlein, editors, Global Climates since the Last Glacial Maximum, pages 24-93. Minneapo- lis: University of Minnesota Press. Machette, M.N., and W.E. Scott 1988. Field Trip Introduction: A Brief Review of Research on Lake Cycles and Neotectonics of the Basin and Range Province. Miscellaneous Publications of the Utah Geological and Mineral Survey, 88-1: 7-14. Miller, D.M. 1997. Cretaceous to Quaternary Geologic Evolution of the Great Basin. [Paper presented at "GBASH 97: Great Basin Aquatic System His- tory: Fifty Years of Geologic, Biologic, and Hydroclimatic Progress in Late Cenozoic Aquatic System History," September 17-20, 1997, University Park Hotel, Salt Lake City, Utah.] [Unpublished.] Oviatt, C.G. 1987. Lake Bonneville Stratigraphy at the Old River Bed. American Jour- nal of Science, 287:383-398. 1997. Lake Bonneville Fluctuations and Global Climate Change. Geology, 25:155-158. Oviatt, C.G., D.R. Currey, and D.M. Miller 1990. Age and Paleoclimatic Significance of the Stansbury Shoreline of Lake Bonneville, Northeastern Great Basin. Quaternary Research, 33:291-305. Oviatt, C.G., D.R. Currey, and D. Sack 1992. Radiocarbon Chronology of Lake Bonneville, Eastern Great Basin. Palaeogeography, Palaeoclimatology, Palaeoecology, 99:225?241. Oviatt, C.G., and W.D. McCoy 1988. The Old River Bed. Miscellaneous Publications of the Utah Geolog- ical and Mineral Survey, 88-1:60-65. Oviatt, C.G., W.D. McCoy, and W.P. Nash 1994. Sequence Stratigraphy of Lacustrine Deposits: A Quaternary Exam- ple from the Bonneville Basin, Utah. Bulletin of the Geological Soci- ety of America, 106:133-144. Oviatt, C.G., and D.M. Miller 1997. New Explorations along the Northern Shores of Lake Bonneville. In P.K. Link and B.J. Kowallis, editors, Mesozoic to Recent Geology of Utah. Brigham Young University Geology Studies, 42(2): 345-371, figure 2. Sack, D. 1989. Reconstructing the Chronology of Lake Bonneville: An Historical Overview. In K.J. Tinkler, editor, History of Geomorphology, pages 223-256. London: Unwin Hyman. Scott, W.E., W.D. McCoy, R.R. Shroba, and M. Rubin 1983. Reinterpretation of the Exposed Record of the Last Two Cycles of Lake Bonneville, Western United States. Quaternary Research, 20: 261-285. Thompson, R.S., C.G. Oviatt, and K.H. Anderson 1997. Paleoclimates and Great Basin Aquatic History. [Paper presented at "GBASH 97: Great Basin Aquatic System History: Fifty Years of Geologic, Biologic, and Hydroclimatic Progress in Late Cenozoic Aquatic System History," September 17-20, 1997, University Park Hotel, Salt Lake City, Utah.] [Unpublished.] Thompson, R.S., C. Whitlock, P.J. Bartlein, S.P. Harrison, and W.G. Spaulding 1993. Climatic Changes in the Western United States since 18,000 yr B.P. In H.E. Wright, Jr., J.E. Kutzbach, T. Webb III, W.F. Ruddiman, F.A. Street-Perrott, and P.J. Bartlein, editors, Global Climates since the Last Glacial Maximum, pages 18-153. Minneapolis: University of Minnesota Press. Fluvial Linkages in Lake Bonneville Subbasin Integration Dorothy Sack "... and so lakes developed in nearly all the basins. Of these lakes the great majority over- flowed into adjacent basins ..." E.Blackwelder (1948:12) ABSTRACT Several late Pleistocene lakes in the Great Basin, including Lake Bonneville, consisted of component subbasins that were united into a single large lake only when water levels were high enough to breach the subbasins' thresholds. During the last deep-lake cycle, but before and after their period of integration with the major water body, isolated closed-basin lakes existed in many of those subbasins. This study focuses on two of Lake Bonneville's seven major subbasins, Puddle Valley and Tule Valley, because they display distinctive geomorphic and sedimentologic evidence of the subbasin integration process. In both cases, Lake Bonneville burst through material at the dividing threshold, and water poured into the adjoining subbasin as a fluvial current of high discharge. Where the jet spread out onto the subbasin's piedmont, it deposited a gravel expansion bar over which gravel was transported in large-scale subaqueous bedforms. Fluvial braid bars in their initial stage of development provide a good analogue for the general form of the inflow bars and help identify the fluvial processes of inflow-bar formation. Details of the geomorphology and sedimen- tology of the inflow bars reflect the site-specific conditions under which they formed. The ability to recognize inflow bars in the geo- morphic and stratigraphic record will help researchers identify subbasin integration events and will, therefore, contribute to improved accuracy in reconstructing the geomorphic and paleocli- matic history of other paleolake basins. Introduction For more than a century, scientists have periodically pub- lished comprehensive inventories of late Pleistocene lakes in western North America (e.g., Russell, 1885, 1896; Meinzer, Dorothy Sack, Department of Geography, 122 Clippinger Labs, Ohio University, Athens, Ohio 45701, USA. Reviewers: B.G. Bills, National Aeronautics and Space Administra- tion, Goddard Space Flight Center, Greenbelt, Maryland 20771, USA; R.M. McCoy, University of Utah, Salt Lake City, Utah 84112, USA. 1922; Hubbs and Miller, 1948; Snyder et al, 1964; Mifflin and Wheat, 1979; Williams and Bedinger, 1984). Although these in- ventories have been compiled for various specific physio- graphic or political boundaries, each included all or a large part of the Great Basin. Information provided about Great Basin paleolakes in that literature ranged from their approximate loca- tions and relative extent, as delineated on small-scale maps; to measurements of various lake attributes, such as depth and area; to observations regarding possible overflow episodes for the predominantly closed-basin lakes. Not surprisingly, those con- tributions emphasized the maximum extent of the paleolakes and any associated overflow that made the lakes tributary to lower closed-basin lakes, or to the sea, via long, connecting stream courses. The emphasis on maximum lake size, however, obscured the fact that, like most large modern lakes (Sly, 1978), several late Pleistocene lakes in the Great Basin were assem- blages of subbasins united into a single lake only during periods of relatively high water levels (Table 1). Many of the lacustrine TABLE 1.?Examples of Great Basin paleolakes that consisted of component subbasins. Paleolake Lake Bonneville Lake Lahontan Lake Chewaucan Fort Rock Lake Lake Alkali Lake Warner Lake Malheur Lake Alvord Lake Surprise Lake Searles Lake Panamint Little Mojave Lake Lake Mojave Location UT, NV, ID NV OR OR OR OR OR OR CA,NV CA CA CA CA Open-basin interval Yes No No ? No ? No ? No No* No* No No Reference Gilbert, 1890 Russell, 1885 Waring, 1908; Allison, 1945 Waring, 1908; Allison, 1979 Waring, 1908 Waring, 1908, 1909; Weide, 1974 Russell, 1903; Waring, 1909 Waring, 1909 Russell, 1884 Gale, 1915 Smith, 1976 Thompson, 1929 Thompson, 1929 "Overflowed to another closed-basin lake by means of a long fluvial connec- tion. 129 130 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES subbasins that constituted arms or embayments of a major lake at its higher levels contained independent closed-basin lakes before and after their interval of integration. In order to reconstruct accurate paleohydrologic and paleo- climatic models for such a region of closed-basin lakes, it is important to distinguish those water-level fluctuations con- trolled primarily by geomorphic factors from those induced more directly by climatic variables. A shoreline that is well represented in the geomorphic or stratigraphic record, for ex- ample, is not necessarily the result of static climatic condi- tions; nor does a change in lake level always indicate a climati- cally induced shift in effective moisture (Sack, 1994c). In addition to climatic factors that affect lake hydrology, geomor- phic factors influencing lake hydrography are also reflected in the lake's shoreline record. Variations in lake size that create multiple shorelines may be the result of changes in drainage basin area caused by such factors as tectonism, volcanism, mass movement, stream diversion, or overflow of upstream lake basins. If the geomorphically induced water-level change is rapid, a vertical zone mostly devoid of shoreline evidence, i.e., a null zone, may result (Sack, 1990, 1995). It is also possi- ble that a well-developed shoreline in a closed-basin lake marks the elevation of an intrabasin threshold rather than an in- terval of climatic equilibrium (Eardley et al., 1957; Benson, 1978; Currey and Oviatt, 1985; Benson and Paillet, 1989). For example, when, under conditions of increasing effective mois- ture, the water level in one subbasin reaches the elevation of an intra-basin sill, it may remain essentially at that elevation until overflow fills the adjoining closed basin to the threshold level and a single integrated closed-basin lake is formed. The coa- lescing of subbasin lakes into a single water body by succes- sive overflows from one subbasin to another can create in the confluent lake basin a complex set of shoreline evidence and can result in the amalgamated lake basin having not just a sin- gle lake-level chronology but, rather, a complicated assem- blage of subbasin chronologies (Davis, 1982). Properly allo- cating the contributions to paleolake-level changes influenced primarily by geomorphic factors versus those responding di- rectly to climatic variables leads to a more accurate reconstruc- tion of a paleolake basin's climatic and geomorphic history. At its maximum extent, Lake Bonneville was the largest of the standing water bodies that accumulated in the Great Basin during the late Pleistocene (Figure 1). The Bonneville lacustral cycle lasted from about 30 thousand years ago (Ka) to 12 Ka (Figure 2) (Oviatt et al., 1992), but considerable shoreline and sediment evidence of it remains visible in the landscape (Sack, 1992). The white marl, for example, is a distinctive Lake Bonneville marker bed that was deposited widely around the basin during the deeper phases of Lake Bonneville spanning an interval of at least 8 thousand years (Ky) (Gilbert, 1890; Oviatt, 1987). The most prominent of the relict Lake Bonneville shore- lines are the Stansbury, Bonneville, and Provo. The Stansbury shoreline was formed between about 22 and 20 Ka, during the transgressive phase of Lake Bonneville (Currey et al., 1983; -45?N -40 >^ -35? / ? Great Basin ? Lake Bonneville D Great Salt Lake 0 300 km 120? i M 115? 110? W I l FIGURE 1.?Map showing Lake Bonneville at its greatest extent relative to the Great Basin. 1550 1450 1350 1250 PS - BS BF SS - ^??-r Provo shoreline Bonneville shoreline Bonneville flood Stansbury shoreline , SS / J B5 / RF PS K \ 30 25 20 Age (Ka) 15 10 FIGURE 2.?Generalized hydrograph of Lake Bonneville (after Burr and Cur- rey, 1988; Oviatt etal., 1992). Oviatt et al., 1990). After Stansbury shoreline time, the lake continued rising until it reached the lowest point on its drainage divide, perhaps as late as 15 Ka (Oviatt et al., 1992). With its upper limit constrained by this external threshold, the lake con- structed the Bonneville shoreline (Gilbert, 1890). Bonneville shoreline time ended about 14.5 Ka when failure of material at the threshold triggered the Bonneville Flood (Gilbert, 1890; Malde, 1968; Oviatt et al., 1992; O'Connor, 1993). The lake re- NUMBER 33 131 stabilized at the level of a bedrock sill about 110 m below the Bonneville shoreline, and the conspicuous Provo shoreline was created during a period of a few hundred years while the lake was controlled by this lower external threshold (Gilbert, 1890; Burr and Currey, 1988; Oviatt et al., 1992). Closed-basin con- ditions finally resumed when Lake Bonneville regressed from the Provo shoreline because of climatic reasons, with the lake cycle ending roughly 2 Ky later (Oviatt et al., 1992). The Bonneville lake basin consists of seven major closed subbasins each of which was integrated into the megalake for a different period of time (Figure 3; Table 2). Lake Bonneville originated in the lowest subbasin, that of Great Salt Lake. Rush Valley was part of the Great Salt Lake basin at the start of the Bonneville lacustral cycle but was isolated from it after Bonne- ville shoreline time because of the barricading effect of the Stockton Bar complex, which Lake Bonneville had built across their connecting strait (Gilbert, 1890). Integration of the Great TABLE 2.?The major Lake Bonneville subbasins and selected characteristics. Subbasin Great Salt Lake Great Salt Lake Desert Puddle Valley Sevier Desert Tule Valley Cedar Valley Rush Valley Approximate modern threshold elevation (m) Lowest basin 1285 1362 1399 1446 1507 1585 Mode of integration Origination basin Spillover from Lake Bonneville Spillover from Lake Bonneville Overflow to Lake Bonneville Spillover from Lake Bonneville Overflow to Lake Bonneville (?) Origination basin, isolated later Salt Lake Desert, Puddle Valley, and Tule Valley occurred when transgressing Lake Bonneville spilled over successively higher thresholds into those adjacent subbasins (Currey, 1980; Currey and Oviatt, 1985; Sack, 1990). Although Currey and Oviatt (1985) maintained that Cedar Valley was also integrated by spillover from Lake Bonneville, overflow may actually have 113?W 112?W -42?N Idaho to Utah Puddle Valley Great Salt \p Lake _Wyqming Salt Lake City Great Salt Lake Desert ?: / ?-Rush ? / .? Valley / ', CedarValley Sevier Desert Maximum Extent of Lake Bonneville 60 km 40 mi Tule Valley 38?N- FlGURE 3.?Map showing the seven major subbasins of Lake Bonneville. 132 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES been from the high Cedar Valley lake basin, which has a rela- tively small capacity, into Lake Bonneville. On the other hand, it seems clear that the Sevier Desert basin initially drained northward over the lowest point on its divide into Lake Bonne- ville by means of a river channel tens of kilometers long (Cur- rey et al., 1983; Oviatt and McCoy, 1988). Lake Bonneville eventually transgressed well above that threshold, converting the connecting river bed into a strait and the Sevier Desert trib- utary basin into its Sevier Desert arm. Regardless of the direction of overflow, spillover from one subbasin to an adjacent one is essentially a fluvial process. Even if both basins initially contain independent closed lakes, differences in such variables as elevation, relief, hypsometry, precipitation, surface runoff, and ground-water budgets will cause the water bodies to rise at different rates (Benson, 1978; Benson and Mifflin, 1986). The first one to overtop their con- necting sill will discharge a current of water flowing down- slope by the force of gravity toward the base level of the re- ceiving basin. Depending upon specific geomorphic and hydrologic circumstances, this fluvial current may occupy a long or short channel, flow on a continuous or intermittent ba- sis, and have a total existence of brief or extensive duration. In both the Puddle Valley and Tule Valley subbasins of Lake Bonneville, a well-exposed example of a distinctive deposi- tional landform created during integrating inflow is found adja- cent to the subbasin threshold (Currey, 1980; Currey and Oviatt, 1985; Sack, 1990). Combined with the results of valley- wide investigations, a comprehensive analysis of the geomor- phology, sedimentology, and morphostratigraphy of these fea- tures, referred to in this paper as inflow bars, provides insights into the nature of the subbasin integration process and the tim- ing of Lake Bonneville's integration of the two valleys. Results of this investigation, moreover, are useful beyond the Bonne- ville Basin. Recognizing similar evidence in the geomorphic and stratigraphic record of other Quaternary and perhaps even pre-Quaternary lake basins will help researchers to identify subbasin integration events and, thereby, to reconstruct more accurately the geomorphology, chronology, and paleoclimate of other paleolake basins. Previous Work The fact that climate is not the only variable affecting the size of closed-basin lakes has been recognized in several previ- ous paleolake studies. More than a century ago Gilbert (1890:262) emphasized the significance of "topographic con- figuration" as well as "meteorologic conditions" with regard to changing lake levels. Stream diversion is probably the most frequently mentioned nonclimatic influence on closed-basin lake levels (e.g., Gilbert, 1890; Bright, 1963; King, 1978; Davis, 1982; Benson and Paillet, 1989). In their report on late Pleistocene lakes in Nevada, however, Mifflin and Wheat (1979) stressed the importance of recognizing episodes of ex- ternal drainage in a lake's history because pluvial hydrologic indexes calculated for shorelines controlled by external thresh- olds are only minimum estimates, which limits their usefulness in making interbasin hydrologic comparisons. Smith and Street-Perrott (1983:192) noted that the size, shape, and hydro- logic regime of a lake basin, and therefore the position of its water plane, may change as the result of various nonclimatic processes, including tectonism, volcanism, landsliding, sedi- ment infilling, stream capture, and the cutting of an external outlet. Such alterations can make interpreting the paleoclimatic signal from closed-basin lakes especially challenging (Street-Perrott and Harrison, 1985). The Great Basin paleolake literature contains some discus- sion of the nature of constituent subbasins and intrabasin sills and their role in lake history. The second largest late Pleis- tocene lake in the Great Basin, Lake Lahontan, has thus far been studied more intensively from the subbasin perspective than any other Great Basin paleolake, probably because even when portrayed at its largest extent it maintains the strong ap- pearance of discrete water bodies separated by narrow straits. The importance of the subbasin nature of Lake Lahontan was first identified by Russell (1885), who recognized that various component basins would have been united or isolated at differ- ent lake levels. In another early work, Jones (1925:37) ob- served that the Lake Lahontan intrabasin sills consist of allu- vium, and he examined precipitation differences among the "subsidiary basins." Later, those subbasin lakes that would ex- pand mainly by direct inflow from perennial tributary streams were distinguished from those without perennial streams, which would rise instead by spillover from adjoining subbasins (Benson, 1978). More recently, Benson and Paillet (1989) used seven major Lake Lahontan subbasins to demonstrate the influ- ence of topography on lake level. They concluded that total lake-surface area is a better indicator of climate change than lake level or volume. Since the late 1970s, subbasin events and sill elevations have figured prominently in efforts to recon- struct the chronology of Lake Lahontan (Benson, 1978, 1991; King, 1978; Davis, 1982, 1983; Thompson et al., 1986; Benson and Thompson, 1987; Benson et al., 1992). It was in this period that Davis (1982:54) summarized so well the importance of component basins to reconstructing the chronology of a paleo- lake: "there is no 'Lahontan chronology'; there is a complex family of chronologies, which are fewer when the water is high, and more numerous when it is low." Gilbert (1890) briefly mentioned constituent subbasins and intrabasin sills in his extensive monograph on Lake Bonneville. Gilbert (1890:20) observed that, because of coalescence of ex- panding lakes, the number of independent endoreic drainage basins in the entire Great Basin decreased during the late Pleis- tocene lake epoch, with Lake Bonneville becoming the largest of the "confluent lakes." Desiccation at the end of the lake cy- cle left the Lake Bonneville basin redivided into numerous, in- dependent closed basins (Gilbert, 1890). According to Gilbert (1890:186), the strongly developed geomorphic characteristics of one of the lake's most conspicuous closed-basin-phase NUMBER 33 133 shorelines, the Stansbury, suggest that it was controlled "by an outflow or its equivalent" rather than by climatic equilibrium. He considered the possibility that sill-controlled inflow to a component subbasin "might have served as a regulator, causing the oscillating lake to linger at a particular level each time it rose" (Gilbert, 1890:186-187). In a 1957 paper, Eardley et al. (pages 1155-1164) presented some hypotheses regarding Lake Bonneville subbasins and shorelines. Those authors emphasized the notion that the ex- pansion of Lake Bonneville into a subbasin increased the lake's surface area, resulting in greater evaporation from the lake. If precipitation surplus over evaporation was modest during the lake's transgressive phase, which they believed was generally the case, expansion in surface area by subbasin integration would create a temporary balance between precipitation and evaporation, and a water-level stillstand would result. The longest stillstand, and hence the best shoreline evidence, would be associated with the largest and most arid receiving basins. During the regressive phase, with precipitation slightly less than evaporation, abandonment of a subbasin would cause a sudden decrease in evaporation from the lake surface. The re- sulting temporary balance between precipitation and evapora- tion would again cause a stillstand in lake level. Eardley et al. (1957) used sill and shoreline elevation data to reject Gilbert's (1890) suggestion that the Stansbury shoreline of Lake Bonne- ville was controlled by an intrabasin sill. On the other hand, they hypothesized that the stillstand that created the Gilbert shoreline of Great Salt Lake was caused by Great Salt Lake spilling over its threshold into the Great Salt Lake Desert. More recent publications that mention Lake Bonneville sub- basins and interior sills identify the landforms created by in- flow. In a study of the coastal geomorphology near Great Salt Lake, Currey (1980) suggested that during the regressive phase of the last deep-lake cycle inflow from the desiccating Lake Bonneville briefly maintained a lake in the Puddle Valley sub- basin. According to Currey (1980:76), a low-level shoreline marks the small Lake Puddle, and a "lobate, steep-fronted delta" indicates the point of high-velocity inflow. A subsequent analysis of shoreline evidence, however, describes rapid filling of Puddle Valley by Lake Bonneville inflow during the trans- gressive phase of the lake cycle and concludes that the low-level Puddle Valley shoreline was formed by an indepen- dent Lake Puddle that existed during the regressive phase of the lake cycle after the subbasin lost its surface connection with Lake Bonneville (Sack, 1995). Shoreline evidence supporting the existence of an isolated Lake Puddle during the transgres- sive phase of the last lake cycle, before Lake Bonneville inte- gration, has not been reported. In their outline of the fluctuations of Lake Bonneville, Cur- rey and Oviatt (1985) included probable transgressive-phase stillstands caused by the lake spilling into the Cedar Valley and Tule Valley subbasins. Those authors attributed a scoured zone close to the Cedar Valley threshold to high-velocity inflow, but they were unable to identify a shoreline in Lake Bonneville that resulted from that inflow event. Currey and Oviatt (1985:11) found inflow into Tule Valley to be "clearly marked by a bed-load delta," and they tentatively assigned a particular, un- named Lake Bonneville shoreline to the associated stillstand. Sack (1990, 1994a, 1994b) later suggested that the Tule Valley inflow "delta" may be a compound feature, and she attributed the 29 m high Tule Valley null zone, a basinwide sector of the piedmont that lacks substantial shoreline evidence, to rapid in- flow from Lake Bonneville. She also reported on the existence during the last lake cycle of an isolated Lake Tule both before and after the subbasin's interval of integration with Lake Bonneville. Data and Interpretation PUDDLE VALLEY GENERAL SETTING.?Puddle Valley is located about 100 km west of Salt Lake City, Utah (Figure 3). It is bordered by the Lakeside Mountains to the east, Cedar Mountains to the south, Grassy Mountains to the west, and by smaller ridges and knolls to the north (Figure 4). The 400 km2 closed drainage basin is entirely surrounded by (inset into) the Bonneville Basin. Spe- cifically, it adjoins the Great Salt Lake subbasin, which lies to threshold and inflow area subbasin divide FIGURE 4.?Regional map of Puddle Valley. 134 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the north, east, and southeast, and the Great Salt Lake Desert subbasin, which lies to the west and southwest. Elevations in the subbasin range from 1316 m on the valley floor to 2019 m at the highest peak in the Lakeside Mountains. There is no nat- urally occurring perennial or intermittent surface water in Pud- dle Valley today. The lowest point on the Puddle Valley drainage divide, i.e., its threshold, lies in an unnamed pass in the northeastern part of the valley (Figure 4). A paved highway to Lakeside, Utah, ex- tends through the pass. The pass has an approximate elevation of 1362 m, which is below the modern local elevations of the Bonneville (1625 m), Provo (1501 m), and Stansbury (1380 m) shorelines and above the post-Bonneville Lake Puddle level (1338 m) (Currey, 1982; Sack, 1995). On the Puddle Valley (south) side of the threshold, the pass area consists of a north- south-trending, trough-shaped valley, which is about 1.2 km long and 0.5 km wide (Figure 5). The sides of the trough con- ca\vy ^nave\>een Teworkeo. into alluvial fans; eolian, fluvial, and FIGURE 5.?Aerial photograph of the Puddle Valley threshold area, including the unnamed pass (P), Lake Bonneville shoreline deposits (S), the trough-like depression (T), and the inflow bar (IB). alluvial-fan sediments dominate the floor of the trough. The floor of the confined trough slopes about 0.7? toward Puddle Valley, but the modern topography is locally irregular primarily because of eolian action. INTEGRATION INTERVAL.?The approximate duration of Puddle Valley's integration into the Bonneville lacustrine sys- tem can be estimated by plotting the rebound-corrected eleva- tion of the Puddle Valley threshold on a detailed hydrograph of Lake Bonneville. The following formula is used to convert modern Bonneville Basin elevations to their values prior to hy- dro-isostatic rebound (Currey and Oviatt, 1985): Za = Zr-{Zr~l200\zb-\552),(Zfc-1200) (Equation 1) where Za is the adjusted rebound-free (estimated original) ele- vation, Zr is a given rebounded (present) elevation, Zb is the local modern elevation of the Bonneville shoreline, 1552 is 1200 is an empirical constant (all elevations are in meters). For the Puddle Valley threshold, 1362 m and 1625 m are used m Equation 1 for Zr and Zb, respectively. Plotting the result- ing Za (1334 m) on the Lake Bonneville hydrograph of Oviatt et al. (1992) yielded an estimate of 25.8 to 12.2 Ka for the in- terval of Puddle Valley integration (Figure 6). The latter value is a minimum-limiting age estimate because some hydro-iso- static rebound had already occurred by this stage of Lake Bonneville (Currey and Burr, 1988); the elevation of the pass at the time of re-isolation would have been somewhat higher than 1334 m. GEOMORPHOLOGY AND SEDIMENTOLOGY OF THE INFLOW BAR.?The depositional landform that marks the inflow of Lake Bonneville into Puddle Valley (Currey, 1980) lies imme- diately south of the Puddle Valley threshold area, just beyond the southern end of the trough-shaped valley (Figures 4, 5, 7). Field work and the interpretation of aerial photographs reveal that the inflow landform is a large-scale, tongue-shaped bar on 1550 1450 ??g 1350 UJ PS Provo shoreline BS Bonneville shoreline BF Bonneville flood SS Stansbury shoreline 30 10 FIGURE 6.- Bonneville. -Estimated interval of Puddle Valley's integration with Lake NUMBER 33 135 which are found linguiod, flow-transverse bedforms, i.e., three-dimensional subaqueous dunes (Figure 8) (Ashley, 1990). Both the bar and the subaqueous dunes are composed of gravel. The feature ranges in elevation from 1341 m at the dis- tal (southern) base of the bar to 1357 m at the crest of the high- est bedform. The bar, which is about 1.5 km long and up to 1.2 km wide, consists of a stoss-side pan bordered by a continuous lateral and lee-side rim. Along the lee side the rim has a maxi- mum lakeward slope of 15?. The bar complex is mantled with Lake Bonneville marl, clastic lacustrine fine-grained sedi- ments, and postlacustrine eolian sandy silt. The maximum measured thickness of this post-inflow sequence is 1.9 m. Over the ridges of the linguoid bedforms, where the post-inflow se- quence is thinner, the surface is gravelly. Postlacustrine eolian fines thicken dramatically toward the northeastern portion of the feature; yellowish gray eolian sandy silt found along the northern end of the left lateral rim is at least 1.0 m thick. A measured cross section (Figure 9) shows that the three- dimensional dunes are moderately asymmetrical and com- A-A' location of cross section I I bedforms swales bedform crest 0 200 400 600 m FIGURE 7.?Map-view diagram of the Puddle Valley inflow bar, showing line of measured section, A-A'. FIGURE 8.?Large-scale subaqueous gravel dunes overlain by light-colored lacustrine and eolian fines exposed in a road cut along the measured section in Puddle Valley. pound, with medium-scale dunes superposed on large-scale dunes, which, in turn, are superposed on very large dunes (Ash- ley, 1990). Measured lengths (L) range from 6.1 to 143.3 m, heights (H) from 0.2 to 3.2 m, and the vertical-form index {LI H) from 17.8 to 62.2. The equation that expresses the height versus length relationship of the Puddle Valley subaqueous dunes, H = 0.0445 L0.90 (Equation 2) B1353 m n 1351 m- 1349 m- [ | Post-inflow fines [ ? :'."l Inflow gravel B' V.E.=10x 20 40 60 80 100 120 m FIGURE 9.?Measured cross section through gravel dunes on the Puddle Valley inflow bar. Location of measured section is shown on Figure 7. (V.E. = vertical exaggeration.) 136 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 3.?Metric height (H) vs dunes. Equation H = 0.074L0-77 H = 0.0052L1 5 H = 0.0677L0 898 H = 0.0445L090 H = 0.0409L113 . length (L) relationships for subaqueous Source Allen (1968) Baker(1973) Ashley (1990) Puddle Valley (this paper) Tule Valley (this paper) is similar to those derived from "giant current ripples" by Allen (1968) and Baker (1973:50) and is close to the equation deter- mined from almost 1500 bedforms, reported in Ashley (1990) (Table 3). The subaqueous dunes on the Puddle Valley inflow bar con- sist of cross beds that, at the measured section, dip 9? to 24? to the east-southeast. Orientation data collected from the 10 larg- est clasts at 14 different sediment-sample sites along A-A' also show a current direction to the east-southeast (Figure 10). This direction of dip for beds and clasts reflects the position of the line of section to the east of the main axis of flow (Figure 7). East-southeast is perpendicular to the crestal trend of the sub- aqueous dunes along A-A'. Sampled clasts were predominantly rounded or subrounded, and the 10 largest had an average longest axis (A-axis) of 19 cm. The overall 10 largest clasts observed anywhere along the length of the cross section had a mean A-axis of 35 cm. In ad- dition to the local Paleozoic limestones, dolomites, cherts, and sandstones, gravel-size clasts of tufa-cemented conglomeratic beachrock, tufa, and laminated marl were noted in the sample pits and elsewhere along the cross section. A gravel-size rip-up clast of marl that was collected from one of the sediment-sam- ple pits contained an ostracode assemblage that appears to be of pre-Bonneville age (R.M. Forester, pers. coram., 1992). FIGURE 10.?Representative sediment-sample pit dug in the Puddle Valley gravel dunes. Note size of clasts and dip of beds. Shovel handle is approxi- mately 0.5 m long. INFLOW DISCHARGE.?In order to reconstruct the paleodis- charge during inflow from Lake Bonneville to Puddle Valley several variables have to be determined. Depth of the inflow current can be estimated from the following regression equa- tion derived for particle diameters between 5 and 330 cm (Knox, 1987): d = 0.0001DL215 ?'57 (Equation 3) where d is the competent flow depth (in meters), D is the inter- mediate axis (B-axis) of the coarsest particles moved (in mm), and S is the dimensionless energy slope. For the Puddle Valley data, D = 133.0 mm, which is the mean of the B-axes of the 10 largest clasts sampled from the A-A' cross section. S, the hy- draulic slope, is estimated with 0.0000747, the value derived from field measurement of the topographic slope just upflow from the gravel dunes. When substituted into Equation 3 these values yield a flow depth, d, of 8.36 m. From flow depth and from the representative particle diameter, a roughness value, n, is calculated with Limerinos' (1970) equation, 0.113d1/6 (Equation 4) with D now expressed in meters. ?or relatively wide and shal- low channels it is generally considered acceptable to substitute flow depth for hydraulic radius in the Manning equation (Le- opold etal., 1964), ,2/3cl/2 a S (Equation 5) where v is flow velocity. Using the depth and slope values from above and using n = 0.034, as calculated from Equation 4, ve- locity of flow during dune formation would have been 1.05 m/s. Paleodischarge, Q, is calculated as Q = vwd, (Equation 6) where w is the channel width. Setting approximate channel width to 1200 m, which is the width of the bar at its proximal end, and using the values of d and v calculated above, dis- charge becomes 10,500 m3/s (=370,000 ft3/s). If the maximum height of the subaqueous dunes, 3.2 m, is substituted for depth in Equations 4 through 6, a discharge of 2000 m3/s (~73,000 ft3/s) results. This latter value is considered a minimum esti- mate of discharge during inflow. TIMING AND NATURE OF THE INFLOW EVENT.?The white marl is a prominent Lake Bonneville marker bed that was de- posited during the deeper levels of Lake Bonneville (Gilbert, 1890; Oviatt, 1987). Its location on top of the inflow bar indi- cates the bar was not created during the regressive phase of Lake Bonneville, as was previously implied by Currey (1980). There is no evidence of subaerial alteration of the dunes, and the contact between the inflow gravel and marl appears con- formable (Figure 9); therefore, it is concluded that the inflow bar was created when the transgressing Lake Bonneville spilled NUMBER 33 137 into Puddle Valley about 25.8 Ka, rather than during a previous lake cycle. The dimensions of the bedforms, the size of constituent clasts, and the presence of rip-up fragments of marl, tufa, and beachrock, together with the discharge estimate, indicate that spillover was a powerful event. It seems likely that most of the material transported in the inflowing current was eroded from the threshold area. From the rounded to subrounded shape of the gravels and the present lack of bedrock outcrop at the threshold it is inferred that just prior to inflow the pass con- sisted mostly of unconsolidated or semiconsolidated deposits rather than coherent bedrock. This unconsolidated material may have had alluvial and colluvial as well as lacustrine com- ponents. No geomorphic or stratigraphic evidence has yet been found to indicate the presence of an isolated water body in Puddle Valley just before inflow from Lake Bonneville. If there was one it must have been quite small, occupying only the lowest part of the subbasin. In the absence of pre-integration Lake Puddle shorelines near the inflow landform, there is no reason to conclude that the landform was created by anything other than the strictly fluvial processes of the inflow current per se, shortly before being submerged under the rising Puddle Valley arm of Lake Bonneville. If inflow discharge was indeed within the 2000 to 10,500 m3/s estimated range, the volume of Puddle Valley below the sill (=4 km3) would have filled in roughly four to 23 days. TULE VALLEY GENERAL SETTING.?Tule Valley lies between the House and Confusion Ranges in the southwestern part of the Bonne- ville Basin, approximately 210 km southwest of Salt Lake City (Figures 3, 11). Like Puddle Valley, it is currently an indepen- dent desert basin of interior drainage with only ephemeral streamflow. Several springs, however, are found on the Tule Valley floor and along its piedmont junction. Tule Valley has greater relief (1340-2950 m) and a larger area (1425 km2) than Puddle Valley. Climatic conditions are essentially the same on the two subbasin floors, but because of elevation differences, somewhat cooler and wetter conditions exist in the ranges adja- cent to Tule Valley compared with those bordering Puddle Val- ley (Stevens and Brough, 1987; Sack, 1990). The lowest point on the drainage divide between Tule Valley and the main body of Lake Bonneville is located at Sand Pass in northeastern Tule Valley (Figures 11, 12). Sand Pass has an elevation of 1446 m, which is approximately 21m below the local elevation of the Provo shoreline. A channel-like depres- sion slopes about 1.2? from the threshold to the Tule Valley in- flow bar. The Sand Pass threshold area consists of a compli- cated assemblage of primarily coastal sediments deposited near the elevation of the Provo shoreline. In addition to Provo shoreline deposits, white marl, coastal gravel from the trans- gression of Lake Bonneville, and indurated remnants of a par- \ 113.5 W/ \\ p. I ) Great Salt YU \ V Lake Desert JV I ) VVNVK Snake | / Valley I Tule ) Valley / ( Confusion J / / Range / JP ' inflow subbasin divide 0 5 mi 0 10 km ( ^\^ Sand Pass / \ Juab Co. S MillardCo.1 \ House ^-v Range ^i 1 /^ Sevier V. Desert \ 1 39?N- \ FIGURE 11.?Regional map of Tule Valley. tially eroded sand and gravel coastal landform, interpreted as an oxygen isotope stage (OIS) 6 barrier complex, are found in the threshold area. Remnants of the barrier complex consist of in-place as well as large, partially collapsed blocks of ce- mented, cross-bedded sandstone and conglomerate beachrock exposed on both the east and west sides of the channel-like de- pression. The central portion of the barrier, representing most of its original length, has been completely eroded away. Lacus- trine tufa deposited after the barrier was breached fills tension cracks between semicollapsed blocks of the cross-bedded sand- stone and conglomerate (Figure 13). Interpretation of the bar- rier complex as an OIS 6 feature is supported by amino-acid analyses (conducted by the Amino Acid Geochronology Labo- ratory (AGL), University of Massachusetts) of gastropod shells (AGL2083, AGL2084, AGL2085) collected from indurated coastal deposits inferred to be topsets of the barrier. In addition, along the west side of the channel the beachrock lies beneath 138 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Tufa FIGURE 12.?Aerial photograph of the Sand Pass area showing Sand Pass (SP), the channel-like depression extending from the pass (C), remnants of cement- ed, cross-bedded sandstone and conglomerate beachrock interpreted as pre-Bonneville coastal deposits (B), and the Tule Valley inflow bar (IB). Cross-bedded sandstone and conglomerate beachrock also crop out along the west side of the channel-like depression. Fkil'Kr 13.?Lacustrine tufa deposited in a tension crack separating a col- lapsed block of the cross-bedded conglomeratic beachrock from an intact por- tion of the beachrock. 26795, 13C adjusted; 17,300 ? 320 yr B.P., unadjusted), deter- mined on gastropod shells collected from the top of a Lake Bonneville gravel spit located 45 km south-southwest of Sand Pass, shows that Tule Valley was integrated with Lake Bonne- ville at least by 17.7 Ka (Sack, 1990). The 13.5 Ka estimated age of reisolation of Tule Valley from Lake Bonneville is sup- ported by a radiocarbon age of 13,430 ? 130 yr B.P. (Beta-47754, 13C adjusted; 13,020 ? 130 yr B.P., unadjusted) determined from gastropod shells found in fine-grained calcar- eous sediments overlying the Tule Valley inflow bar. The ap- proximately 10 cm thick fine-grained unit containing those gastropod shells and the overlying 60 cm thick gravel are inter- preted as sediments of independent Lake Tule deposited after Tule Valley's reisolation from Lake Bonneville. GEOMORPHOLOGY AND SEDIMENTOLOGY OF THE INFLOW BAR.?Like the Puddle Valley example, the inflow landform in Tule Valley consists of a large-scale linguoid bar that displays large transverse bedforms, a stoss-side pan, and a lateral and lee-side rim (Figures 12, 15). The Tule Valley inflow bar is ap- the transgressive Lake Bonneville sequence of coastal gravel and white marl. INTEGRATION INTERVAL.?As with Puddle Valley, the inter- val of Tule Valley's integration with Lake Bonneville can be estimated by plotting the rebound-corrected elevation of the threshold on the Lake Bonneville hydrograph. Using the mod- ern elevation of Sand Pass, 1446 m, as an approximation of the pre-integration threshold elevation (Zr) and 1583 m for the lo- cal elevation of the Bonneville shoreline (Zb), Equation 1 yields a rebound-corrected elevation (Za) of 1426 m. Accord- ing to the hydrograph of Oviatt et al. (1992), the Bonneville Basin landscape at 1426 m was submerged by Lake Bonneville from about 19.4 to 13.5 Ka (Figure 14). These estimates are consistent with stratigraphic observations and numeric ages ob- tained in association with Tule Valley lake sediments. A radio- carbon age of 17,710 ? 320 yr before present (B.P.) (Beta- 1550 1450 ?| 1350 LU 1250 PS Provo shoreline BS Bonneville shoreline BF Bonneville flood SS Stansbury shoreline BS 30 BF PS Tule Valley integration interval 25 20 Age (Ka) 10 FIGURE 14.?Estimated interval of Tule Valley's integration with Lake Bonne- ville (a.s.l. = above sea level). NUMBER 33 139 proximately 2 km long and 1.5 km wide, with elevations rang- ing from 1391 m at the foot of the distal edge to at least 1426 m on the right lateral rim. The lakeward slope off the lee-side rim is about 4.5?. In contrast to the arcuate distal end of the Puddle Valley feature (Figure 7), the Tule Valley inflow bar displays in planimetric view a much straighter lee-side rim. The Tule Val- ley bar is overlain by Bonneville white marl, by coastal sedi- ments of the post-Bonneville independent Lake Tule (Sack, 1990), and in some places by significant accumulations of post- lacustrine subaerial deposits. Compared with the Puddle Valley bar, morphologic details of the Tule Valley bar are more ob- scured by postlacustrine fluvial, eolian, and mass-wasting pro- cesses. Nevertheless, the gravel dunes in Tule Valley also ap- pear to be three-dimensional. A cross section was measured along the right lateral margin of the bar where a gravel pit south of Sand Pass Road and a stream cut north of the road provided good exposures (Figures 15, 16). The representative portion of the measured cross sec- tion presented in Figure 16 shows subaqueous dunes overlain by marl and capped by coastal deposits of regressive Lake Tule. The lower portion of the marl sequence is Lake Bonne- ville white marl, but the upper part of the unit is interpreted as Bonneville marl reworked by Lake Tule after Tule Valley's re- isolation from Lake Bonneville. Dunes from the entire measured section are large (Ashley, 1990), ranging in length (L) from 15.9 to 61.6 m, in height (H) from 0.9 to 4.8 m, and in vertical-form index (L/H) from 10.0 to 30.2. The relationship between length and height is de- scribed by the equation A-C location of cross section ? bedform & rim crests [-.. j eolian deposits I | pan 400 m 2000 ft H = 0.0409L1.13 (Equation 7) FIGURE 15.?Map-view diagram of the Tule Valley inflow bar, including the location of the measured cross section (A-C). Although the exponent in this equation is somewhat larger than that reported by Ashley (1990) for a very large data set, it is smaller than the exponent that Baker (1973:50) found for "giant current ripples" created by the Lake Missoula Flood (Table 3). The subaqueous gravel dunes studied in detail along the right margin of the inflow bar (Figure 15) consist of cross beds that dip 12? to 31? to the southwest. An exposure at the southeast corner of the bar, however, exhibits gravel bedform cross beds that dip 15? to 20? to the southeast. The 10 largest clasts were investigated from eight sediment- sample sites along the entire right-lateral exposure. Measured clasts had an average A-axis of 18 cm and showed dips to the 1415 ? c 1410H "CD LLJ 1405- B Regressive Lake Tule deposits White and reworked marl Inflow gravel V.E. = 7x 50 I 100 I 150 200 Distance (m) FIGURE 16.?Cross-sectional diagram of large-scale subaqueous dunes exposed along the southwest edge of the Tule Valley inflow bar. See Figure 15 for location of B-C. (V.E. = vertical exaggeration.) 140 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES southwest or west. Most of the sampled gravels were sub- rounded or subangular. Numerous boulders were noted in dune gravel along the distal edge of the bar (Figure 17). A-axis mea- surements of the 10 largest clasts observed along the entire length of the lee-side rim ranged from 100 to 200 cm, with a mean length of 120 cm. Dune and bar gravels consist primarily of limestones, dolomites, and quartzites derived from the adja- cent mountain ranges, but clasts of conglomerate and sandstone beachrock identical to the indurated remnants of the breached barrier found adjacent to Sand Pass are common. The 10 larg- est clasts of sandstone and conglomerate beachrock found in the inflow bar, up to 2.5 km downflow from the breached bar- rier, had an average A-axis of 100 cm. INFLOW DISCHARGE.?The large subaqueous dunes and large size of constituent gravel reveal that inflow from Lake Bonneville was a high-velocity event. Paleodischarge can be reconstructed using the same technique as was employed with the Puddle Valley data (Equations 3-6). With a representative clast diameter, D, of 115 mm, which is the average B-axis of the 10 largest sampled clasts from the Tule Valley gravel dune sediment pits, and an approximate slope near the pan of FIGURE 17.?Boulder-size clast of conglomerate beachrock like that found in outcrop near Sand Pass (Figure 12) but located at the distal edge of the Tule Valley inflow bar. Lens cap has a 50 mm diameter. 0.0036, depth is estimated to be 0.77 m. These values yield a roughness coefficient of 0.038 and a velocity of 1.33 m/s. Us- ing the 1450 m maximum width of the bar for channel width results in an estimated discharge of 1500 m3/s (* 52,000 ft3/s). The depth of 0.77 m calculated from Knox's (1987) empirical equation, however, is much smaller than the maximum bed- form height, 4.8 m. Substituting 4.8 m for flow depth yields a velocity estimate of 5.17 m/s and a discharge estimate of 36,000 m3/s (?1,271,000 ft3/s). TIMING AND NATURE OF THE INFLOW EVENT.?The con- formable sequence of white marl overlying preserved subaque- ous dunes verifies that the Tule Valley inflow bar is of trans- gressive Lake Bonneville age and not from a previous lake cycle. According to the above estimates of the Tule Valley in- terval of integration into Lake Bonneville, the inflow bar was deposited before 17.7 Ka and probably about 19.4 Ka. The common occurrence within the subaqueous dunes of conglomerate and sandstone beachrock clasts derived from the breached barrier located near the divide helps support the no- tion that the barrier was created during a pre-Bonneville lake cycle and eroded during the spillover of Lake Bonneville into Tule Valley. Apparently, the relict pre-Bonneville barrier that crossed the Sand Pass threshold area failed suddenly either when the hydraulic load from Lake Bonneville exceeded the barrier's strength, thereby initiating inflow, or shortly after the lake began to spill over the barrier into Tule Valley. Tule Valley definitely contained an independent, transgress- ing lake, called Lake Tule, when Lake Bonneville spilled over Sand Pass (Sack, 1990). Shorelines created by Lake Tule are well exposed around the valley. Their pre-integration relative age is determined from the conformable relationship between the shoreline gravel and the Bonneville white marl. The highest Lake Tule shoreline lies at an approximate elevation of 1416 m. Above that level, shoreline evidence is lacking in a vertical null zone stretching to about 1446 m, the elevation of the Sand Pass threshold. A continuous series of Lake Bonneville shore- lines extends from about 1446 m to the Bonneville shoreline at approximately 1583 m. When Lake Bonneville stood at the brink of spilling into Tule Valley, it occupied the highest eleva- tion it had yet achieved in the lake cycle. At that same time, Lake Tule was probably also at its highest level; therefore, the 1416 m Lake Tule shoreline should mark the position of the in- dependent water body when Lake Bonneville began flowing over Sand Pass (Sack, 1990). Exposures of this highest Lake Tule shoreline are found adjacent to the lee-side rim on both the east and west sides of the Tule Valley inflow bar. The distal end of the Tule Valley inflow bar was apparently deposited at the margin of the water body in Tule Valley. At the estimated inflow discharge of 1500 to 36,000 m3/s, the water level in Tule Valley would have risen the 30 m from the Tule Valley shoreline to Sand Pass (<22.5 km3) in roughly seven to 174 days. NUMBER 33 141 Discussion The linguoid bars created by inflow of Lake Bonneville into Puddle and Tule Valleys appear very similar to the bars that Rundle (1985a, 1985b) observed in the first stage of fluvial braid-bar formation (Figure 18). According to Rundle, that ini- tial tongue-shaped braided stream deposit, referred to in this paper as the "Rundle tongue," is an ephemeral, constructional form deposited downstream from a comparatively deep, con- fined channel reach. It is created under conditions of radiating flow where water discharging from the confinement mixes with calmer water beyond the orifice. Turbulent mixing along the contact between the stream and quiet water "saps the mo- mentum of the stream and reduces its velocity progressively toward the core" (Rundle, 1985a:9-10). Flow is retarded along the edges of the spreading jet relative to its axis, causing the flow to turn and spread away from the axis. As flow velocity decreases, the load is deposited downstream from the orifice of the confined channel. The resulting deposit, a kind of ex- pansion bar, is linguoid, consists of an upstream pan and a lat- eral and downstream rim, and displays a steep lee-side slope. In addition, Rundle (1985b) observed that some bars have multiple concentric rims, formed by successive sheets of gravel being transported over the pre-existing pan surface. The complete Rundle tongue is a transient feature in braided streams because it is soon dissected by chutes at falling- and low-water stages. This dissection of the tongue is, according to Rundle, the mechanism of braiding. Remnants of the Rundle tongue preserved between the chutes constitute longitudinal braid bars. In many respects the Puddle Valley and Tule Valley inflow bars are analogous to the initial stage of the braid bar. Like the Rundle tongue, the inflow bars extend downcurrent from the orifice of a confined channel, which is the sloping, trough-like depression leading from each of the thresholds. The Rundle x X 11 i \\ \< ^?-?-? rim I\ rim pan ^**^ orifice __ orifice X' 1 X1 FIGURE 18.?Plan and profile views of the ideal tongue-shaped bar that forms where water flowing as a fast jet leaves an orifice and spreads out into a region of quieter water (after Rundle, 1985b). The Puddle Valley and Tule Valley inflow bars closely resemble this model. tongue and the inflow bars consist of a pan and a well-devel- oped rim. The subaqueous dunes of the inflow bars are similar to Rundle's (1985b) multiple concentric rims, and both show that bedload was being transported across the top of the lin- guoid bars. A major chute, like those that dissect the initial braided-stream tongue structure, has breached the Puddle Val- ley inflow bar on its western side (Figure 7). Although the Lake Bonneville inflow bars and the Rundle tongue are similar, they do have some differences. The pan on the Puddle Valley example appears smaller than that on the ideal Rundle tongue because of the presence of numerous sub- aqueous dunes on the bar (Figures 7, 18). Unlike the Rundle tongue, which forms where a jet spreads out into quieter water, Lake Bonneville's inflow bars are deposits of the rapid and ex- panding inflow current alone. Deposition of the Lake Bonne- ville inflow bars occurred where the inflow current spread out on the piedmont and was not caused by the presence of lake water in the receiving basins. Deposition of the Tule Valley in- flow bar, in fact, ended bluntly where the inflowing current struck the rising Tule Valley water body. The water level in Tule Valley rose so rapidly that inflow current expansion ceased virtually at the moment of contact. The presence of a large receiving lake in Tule Valley and the absence of one in Puddle Valley is regarded as responsible for the difference in planimetric shape of the distal end of the bars (Figures 7, 15). The flattened distal end of the Tule Valley bar reflects a sudden decrease in velocity that occurred where the inflowing jet hit Lake Tule. In Puddle Valley there is no evi- dence to indicate that the inflow current interacted with a pre-existing water body. The Puddle Valley inflow bar was de- posited by the jet of water spreading out onto the piedmont of an initially dry subbasin, or, perhaps more likely, onto the slope well above a very small Puddle Valley lake or marsh. By not meeting rising lake waters until after the bar was com- pletely formed, the Puddle Valley current expanded symmetri- cally outward from the axis of flow, giving the resulting de- posit a more arcuate shape at its distal end. As a result, compared with the Tule Valley example, the Puddle Valley bar is regarded as having the more ideal inflow-bar shape. In contrast to the Rundle tongue of braided fluvial environ- ments, which is dissected after its formation by falling water levels (Rundle, 1985a, 1985b), both Lake Bonneville inflow bars are mostly preserved because their formation was fol- lowed by rising instead of falling water levels. As the expand- ing water bodies in the subbasins rose rapidly to equilibrate with the main body of Lake Bonneville, no major reworking of the features occurred; however, for 13.6 and 5.9 Ky after their respective formation, the Puddle Valley and Tule Valley inflow bars existed at the bottom of a water column that was as much as 275 m and 183 m deep, respectively. During the sublacus- trine period the inflow bars were blanketed with pelagic white marl. 142 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Conclusions The Rundle tongue reported from a braided stream system (Rundle, 1985a, 1985b) provides a good analogue for the gen- eral form of the studied Lake Bonneville inflow bars; however, details of the geomorphology and sedimentology of the Lake Bonneville inflow bars reflect the site-specific physical condi- tions under which they formed. The large size of transported clasts, the presence of three-dimensional subaqueous dunes, and the paleohydraulic reconstructions support the notion that in both studied subbasins Lake Bonneville spillover was a high- discharge event, probably resulting from an outburst through material at the threshold. Field evidence in Tule Valley indicates that Lake Bonneville inflow burst through a relict OIS 6 barrier complex at Sand Pass. Because such relict "divide barriers" (in- cluding tombolos) created during the Bonneville lake cycle are common in the landscape today in passes throughout the lake basin, and considering the nature of the dune gravel, it is proba- ble that before integration the unnamed pass separating Puddle Valley from the main body of Lake Bonneville was also crossed by some kind of divide barrier. Both of the studied inflow bars date from the transgressive phase of Lake Bonneville and are relatively well preserved because the subbasin water level rose rapidly over them on the transgression and fell relatively quickly past them during the regressive phase of the lake cycle. Their original surface form was subdued under a mantle of lacustrine fine-grained sediments, regressive coastal deposits, and later subaerial sediments. The geomorphic and hydrologic circumstances within Puddle Valley and Tule Valley at the time of Lake Bonneville spillover were such that the integrating fluvial current was a very short-lived but continuous flow, although some initial minor spillover may have occurred. Water flowed through a fairly confined region adjacent to the threshold as a fast jet. Very coarse sediment provided by threshold failure and spillway ero- sion was transported by that current. Where the jet emerged from the orifice of the confined zone, it spread out. Flow diver- gence just outside of the orifice formed the pan. Beyond the pan, bedload was transported in the form of large-scale sub- aqueous dunes. The spreading current caused a comparative de- crease in flow velocity especially near the margins of the flow, resulting in deposition of a fluvial bar. The tongue shape of the bar reflects this velocity decrease along the edges of the spread- ing jet relative to its axis, which effected a rotational motion in the flow (Bates, 1953). Considering the fluvial processes by which they formed, the inflow landforms in Puddle and Tule Valleys are most appropri- ately referred to as bars. The ideal, linguoid bar form was more effectively developed in Puddle Valley than in Tule Valley be- cause in the latter subbasin expansion of the fluvial current ceased abruptly where it struck the Tule Valley lake waters. Rather than being responsible for formation of the inflow land- form, interaction of the inflow current with the Tule Valley lake actually terminated the process of bar formation. For this rea- son, and because no evidence has been found to indicate the ex- istence of a transgressive-phase Lake Puddle anywhere near the Puddle Valley inflow landform at the time of its deposition, the term delta, although originally applied to these landforms (Cur- rey, 1980; Currey and Oviatt, 1985; Sack, 1990), is no longer considered the most appropriate term. Studying Lake Bonneville as a complex system of compo- nent subbasins focuses attention on aspects of its geomorphol- ogy, sedimentology, and chronology that have been generally overlooked by studying the lake as one simple, coherent unit. This subsystem approach may be, and to some extent already has been (e.g., Benson, 1978), profitably applied to other large paleolake basins, which typically consist of several subbasins. Because lakes and lacustrine deposits are sensitive indicators of climate change, paleolake investigations can make important contributions to reconstructing paleoclimates. The accurate modelling of paleoclimates, moreover, is essential for the re- finement of general climate models. Lake-level changes, how- ever, are influenced by tectonism, drainage-basin changes, and subbasin integration as well as by climate change. Because lacustrine inflow bars are evidence of subbasin integration, rec- ognizing them will help paleoenvironmental researchers to ac- curately reconstruct the hydrographic history of a paleolake, thereby contributing to an improved understanding of the true paleoclimatic signal. Acknowledgments Acknowledgment is made to the Donors of The Petroleum Research Fund, administered by the American Chemical Soci- ety, for the partial support of this research. The research that led to this paper was also partly funded by the National Geographic Society and the Graduate School of the University of Wiscon- sin-Madison. I am grateful to D.R. Currey, G.C. Nadon, B.G. Bills, R.M. McCoy, and C.G. Oviatt for suggestions, helpful discussions, and comments on earlier drafts of this manuscript. I appreciate the many hours of field assistance provided by T. Burr, S. Kaplan, J. Killian, G. Nadon, A. Reid, and T. Sabin. Literature Cited Allen, J.R.L. 1968. Current Ripples: Their Relation to Patterns of Water and Sediment Motion. 433 pages. Amsterdam: North-Holland Publishing Co. Allison, I.S. 1945. Pumice Beds at Summer Lake, Oregon. Bulletin of the Geological Society of America, 56:789-807. 1979. Pluvial Fort Rock Lake, Lake County, Oregon. State of Oregon De- partment of Geology and Mineral Industries Special Paper, 7: 72 pages. Ashley, G.M. 1990. Classification of Large-Scale Subaqueous Bedforms: A New Look at an Old Problem. Journal of Sedimentary Petrology, 60:160-172. NUMBER 33 143 Baker, V.R. 1973. Paleohydrology and Sedimentology of Lake Missoula Flooding in Eastern Washington. Geological Society of America Special Paper, 144: 79 pages. Bates, C.C. 1953. Rational Theory of Delta Formation. Bulletin of the American Asso- ciation of Petroleum Geologists, 37:2119-2162. Benson, L.V. 1978. Fluctuation in the Level of Pluvial Lake Lahontan during the Last 40,000 Years. Quaternary Research, 9:300-318. 1991. Timing of the Last Highstand of Lake Lahontan. Journal of Paleo- limnology, 5:115-126. Benson, L.V., D.R. Currey, Y. Lao, and S. Hostetler 1992. Lake-Size Variation in the Lahontan and Bonneville Basins between 13,000 and 9000 14C yr B.P. Paleogeography, Paleoclimatology, Palaeoecology, 95:19-32. Benson L.V., and M.D. Mifflin 1986. Reconnaissance Bathymetry of Basins Occupied by Pleistocene Lake Lahontan. United States Geological Survey Water-Resources Investigation Report, 85-4262: 14 pages. Benson, L.V., and F.L. Paillet 1989. The Use of Total Lake-Surface Area as an Indicator of Climate Change: Examples from the Lahontan Basin. Quaternary Research, 32:262-275. Benson, L.V., and R.S. Thompson 1987. Lake-Level Variation in the Lahontan Basin for the Past 50,000 Years. Quaternary Research, 28:69-85. Blackwelder, E. 1948. The Geological Background. Bulletin of the University of Utah, 38: 3-16. Bright, R.C. 1963. Pleistocene Lakes Thatcher and Bonneville, Southeastern Idaho. 292 pages. Doctoral dissertation, University of Minnesota, Minne- apolis, Minnesota. Burr, T.N., and D.R. Currey 1988. The Stockton Bar. Miscellaneous Publications of the Utah Geologi- cal and Mineral Survey, 88-1:66?73. Currey, D.R. 1980. Coastal Geomorphology of Great Salt Lake and Vicinity. Bulletin of the Utah Geological and Mineral Survey, 116:69-82. 1982. Lake Bonneville: Selected Features of Relevance to Neotectonic Analysis. United States Geological Survey Open-File Report, 82-1070: 31 pages. Currey, D.R., and T.N. Burr 1988. Linear Model of Threshold-Controlled Shorelines of Lake Bonne- ville. Miscellaneous Publications of the Utah Geological and Min- eral Survey, 88-1:104-110. Currey, D.R., and C.G. Oviatt 1985. Durations, Average Rates, and Probable Causes of Lake Bonneville Expansions, Stillstands, and Contractions during the Last Deep-Lake Cycle, 32,000 to 10,000 Years Ago. In P.A. Kay and H.F. Diaz, editors, Problems of and Prospects for Predicting Great Salt Lake Levels, pages 9-24. Salt Lake City: University of Utah Center for Public Affairs and Administration. Currey, D.R., C.G. Oviatt, and G.B. Plyler 1983. Lake Bonneville Stratigraphy, Geomorphology, and Isostatic Defor- mation in West-Central Utah. Utah Geological and Mineral Survey Special Studies, 62:63-82. Davis, J.O. 1982. Bits and Pieces: The Last 35,000 Years in the Lahontan Area. Soci- ety for American Archaeology Papers, 2:53?75. 1983. Level of Lake Lahontan during Deposition of the Trego Hot Springs Tephra about 22,400 Years Ago. Quaternary Research, 19:312? 324. Eardley, A.J., V. Gvosdetsky, and R.E. Marsell 1957. Hydrology of Lake Bonneville and Sediments and Soils of Its Basin. Bulletin of the Geological Society of America, 68:1141?1201. Gale, H.S. 1915. Salines in the Owens, Searles, and Panamint Basins, Southeastern California. Bulletin of the United States Geological Survey, 580: 251-323. Gilbert, G.K. 1890. Lake Bonneville. Monographs of the United States Geological Sur- vey, 1: 438 pages. Hubbs, C.L., and R.R. Miller 1948. The Zoological Evidence. Bulletin of the University of Utah, 38: 17-166. Jones, JC. 1925. The Geologic History of Lake Lahontan. Publications of the Carn- egie Institution of Washington, 352:1?50. King, G.Q. 1978. The Late Quaternary History of Adrian Valley, Lyon County, Ne- vada. 88 pages. Master's thesis, University of Utah, Salt Lake City, Utah. Knox, J.C. 1987. Stratigraphic Evidence of Large Floods in the Upper Mississippi Valley. In L. Mayer and D. Nash, editors, Catastrophic Flooding, pages 155-180. Boston: Allen and Unwin. Leopold, L.B., M.G. Wolman, and M.G. Miller 1964. Fluvial Processes in Geomorphology. 522 pages. San Francisco: W.H. Freeman and Co. Limerinos, J.T. 1970. Determination of the Manning Coefficient from Measured Bed Roughness in Natural Channels. United States Geological Survey Water-Supply Paper, 1898B: 47 pages. Malde, H.E. 1968. The Catastrophic Late Pleistocene Bonneville Flood in the Snake River Plain, Idaho. United States Geological Survey Professional Paper, 596: 52 pages. Meinzer, O.E. 1922. Map of the Pleistocene Lakes of the Basin-and-Range Province and Its Significance. Bulletin of the Geological Society of America, 33: 541-552. Mifflin, M.D., and M.M. Wheat 1979. Pluvial Lakes and Estimated Pluvial Climates of Nevada. Bulletin of the Nevada Bureau of Mines and Geology, 94: 57 pages. O'Connor, J.E. 1993. Hydrology, Hydraulics, and Geomorphology of the Bonneville Flood. Geological Society of America Special Paper, 274: 83 pages. Oviatt, C.G. 1987. Lake Bonneville Stratigraphy at the Old River Bed, Utah. American Journal of Science, 287:383-398. Oviatt, C.G., D.R. Currey, and D.M. Miller 1990. Age and Paleoclimatic Significance of the Stansbury Shoreline of Lake Bonneville, Northeastern Great Basin. Quaternary Research, 33:291-305. Oviatt, C.G., D.R. Currey, and D. Sack 1992. Radiocarbon Chronology of Lake Bonneville, Eastern Great Basin, USA. Palaeogeography, Palaeoclimatology, Palaeoecology, 99: 225-241. Oviatt, C.G., and W.D. McCoy 1988. The Old River Bed. Miscellaneous Publications of the Utah Geo- logical and Mineral Survey, 88-1:60?65. Rundle, A. 1985a. The Mechanism of Braiding. Zeitschrift fur Geomorphologie Sup- plement, 55:1?13. 1985b. Braid Morphology and the Formation of Multiple Channels, the Ra- kaia, New Zealand. Zeitschrift fur Geomorphologie Supplement, 55: 15-37. Russell, I.C. 1884. A Geological Reconnaissance in Southern Oregon. Fourth Annual Report of the United States Geological Survey, pages 431-464. 144 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1885. 1896. 1903. Sack, D. 1990. 1992. 1994a. 1994b. 1994c. 1995. Sly, P.G. 1978. Smith, G.I 1983. Geological History of Lake Lahontan, a Quaternary Lake of North- western Nevada. Monographs of the United States Geological Sur- vey, 11: 288 pages. Present and Extinct Lakes in Nevada. In The Physiography of the United States. National Geographic Society Monographs, pages 101-132. Notes on the Geology of Southwestern Idaho and Southeastern Or- egon. Bulletin of the United States Geological Survey, 217: 83 pages. Quaternary Geology of Tule Valley, West-Central Utah. Utah Geo- logical and Mineral Survey Map, 124: 26 pages. Obliteration of Surficial Paleolake Evidence in the Tule Valley Sub- basin of Lake Bonneville. Society for Sedimentary Geology {SEPM) Special Publication, 48:427-433. Geologic Map of the Coyote Knolls Quadrangle, Millard County, Utah. Utah Geological Survey Map, 162: 16 pages. Geologic Map of the Swasey Peak NW Quadrangle, Millard County, Utah. Utah Geological Survey Map, 163: 16 pages. Geomorphic Evidence of Climate Change from Desert-Basin Palae- olakes. In A.D. Abrahams and A.J. Parsons, editors, Geomorphol- ogy of Desert Environments, pages 617-630. London: Chapman and Hall. The Shoreline Preservation Index as a Relative-Age Dating Tool for Late Pleistocene Shorelines: An Example from the Bonneville Basin, USA. Earth Surface Processes and Landforms, 20:363? 377. Sedimentary Processes in Lakes. In A. Lerman, editor, Lakes: Chemistry, Geology, Physics, pages 65-89. New York: Springer- Verlag. ., and F.A. Street-Perrott Pluvial Lakes of the Western United States. In S.C. Porter, editor, Late Quaternary Environments of the United States, the Late Pleis- tocene, pages 190-212. Minneapolis: University of Minnesota Press. Smith, R.S.U. 1976. Late-Quaternary Pluvial and Tectonic History of Panamint Valley, Inyo and San Bernardino Counties, California. 295 pages. Doc- toral dissertation, California Institute of Technology, Pasadena, California. Snyder, C.T., G. Hardman, and F.F. Zdenek 1964. Pleistocene Lakes in the Great Basin. United States Geological Sur- vey Miscellaneous Geological Investigations Map, 1-416. Stevens, D.J., and R.C. Brough 1987. Western Utah Climatography. Publications of the Utah Geological Association, 16:31?57. Street-Perrott, S.A., and S.P. Harrison 1985. Lake Levels and Climate Reconstruction. In A.D. Hecht, editor, Pa- leoclimate Analysis and Modeling, pages 291-340. New York: John Wiley and Sons. Thompson, D.G. 1929. The Mohave Desert Region California. United States Geological Survey Water-Supply Paper, 578: 759 pages. Thompson, R.S., L.V. Benson, and E.M. Hattori 1986. A Revised Chronology for the Last Pleistocene Lake Cycle in Cen- tral Lahontan Basin. Quaternary Research, 25:1?9. Waring, G.A. 1908. Geology and Water Resources of a Portion of South-Central Ore- gon. United States Geological Survey Water-Supply Paper, 220: 86 pages. 1909. Geology and Water Resources of the Harney Basin Region, Oregon. United States Geological Survey Water-Supply Paper, 231: 93 pages. Weide, D.L. 1974. Postglacial Geomorphology and Environments of the Warner Val- ley-Hart Mountain Area, Oregon. 293 pages. Doctoral dissertation, University of California, Los Angeles. Williams, T.R., and M.S. Bedinger 1984. Selected Geologic and Hydrologic Characteristics of the Basin and Range Province, Western United States. United States Geological Survey Miscellaneous Investigations Map, 1-1522-D. Environment and Paleolimnology of Owens Lake, California: A Record of Climate and Hydrology for the Last 50,000 Years J. Platt Bradbury andR.M. Forester ABSTRACT Diatoms and ostracodes from lake-sediment cores beneath Owens Lake playa in southeastern California document an almost continuous paleolimnological record of limnological, hydrologi- cal, and climatological changes for the last 50 thousand years (Ky), based upon an extrapolated radiocarbon chronology. Close- interval sampling enabled chronological resolution at a century- scale or less. Modern ecologic studies of diatoms and ostracodes, as well as climatic and hydrologic processes affecting the Owens Lake basin, were used to track environmental changes. The middle Wisconsin to Holocene record indicates that climates wetter than those of modern times existed 80 percent of the time, but that changes between wet and dry climates occurred rapidly?within 300 years to a century or less in some instances. The ostracode record indicates that polar air masses permanently resided at the latitude of Owens Lake during parts of the full-glacial period, whereas the diatom record and heavy-mineral analyses imply that extensive alpine glaciers (on the scale of the full-glacial period) did not exist in the Owens Lake drainage before 22 thousand years ago (Ka) and that the effects of glaciation on sedimentation in this area ended by 13 Ka. During oxygen-isotope stage 3 (-50-25 Ka) at Owens Lake, summer-season freshwater planktonic diatoms correlate with geochemical evidence related to increased winter precipitation in marine cores at similar latitudes on the California margin. These data document southward shifts and increased strength of the Aleutian low-pressure system (Aleutian Low) in the northeast Pacific Ocean and its role in determining the hydrologic balance of J. Platt Bradbury, 5784 Horseradish Gulch, Golden, Colorado 80403, USA. R.M. Forester, United States Geological Survey, P.O. Box 25046, Federal Center, Denver, Colorado 80225, USA. Reviewers: Larry V. Benson, United States Geological Survey, P.O. Box 25046, Federal Center, Denver, Colorado 80225, USA; S. Fritz, Lehigh University, Bethlehem, Pennsylvania 18015; Robert Her shier, Division of Mollusks, National Museum of Natural History, Smithson- ian Institution, Washington, D.C. 20560-0118, USA; G.I. Smith, United States Geological Survey, MenloPark, California 94025, USA. western North America as well as its effect on the California Cur- rent. Such correlations indicate that the Owens Lake paleolimno- logical record responded to the major dynamics of climate variation in the northern hemisphere. Introduction Thick, fossiliferous lacustrine deposits underlying Owens Lake (Smith and Pratt, 1957) in southeastern California have been cored and analyzed to extract millennium-scale pale- olimnologic and paleoclimatic information for the Owens Lake basin and surrounding region spanning the last 800 thou- sand years (Ky) (Bradbury, 1997a; Smith and Bischoff, 1997). In this paper, we interpret climate-controlled paleolimnologi- cal changes, documented in diatom and ostracode records at relatively high chronological resolution, between the middle Wisconsin (-50 thousand years ago (Ka)) and the late Ho- locene (0.27 Ka). Sample resolution averages about 250 years for diatoms and about 75 years for ostracodes. In addition, we synthesize ecological, hydrological, limnological, and clima- tological data for the lake and its drainage to provide a back- ground for interpreting the fossil record. We compare the Owens Lake paleolimnological history to local packrat mid- den records of past vegetation and to the glacial records in the Sierra Nevada. Our data provide an independent chronology for the timing and magnitude of Sierra Nevada glaciation that augments conclusions of Bischoff et al. (1997b). Finally, we evaluate the middle Wisconsin paleolimnology of Owens Lake in terms of paleoceanographic and paleoclimatic changes registered in marine cores off the California coast spanning the same time interval as at Owens Lake. This detailed, multi- proxy record of past climate and hydrology from Owens Lake provides important insights to the timing, magnitude, and character of expected future climate changes at Yucca Moun- tain, a proposed nuclear waste repository site 140 km east of Owens Lake. 145 146 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Modern Owens Lake Basin SETTING Owens Lake (playa floor elevation, 1080 m; area, 182 km2) occupies a graben immediately east of the highest part of the Sierra Nevada (Mt. Whitney, 4418 m), near the intersection of coordinates 36.5?N and 118.0?W (Figure 1). Drainage from the eastern flank of the Sierra Nevada contributes the major vol- ume of water to the Owens River. Although streams from the White and Inyo Mountains east of Owens Lake also recharge groundwater to the basin, the streams rarely reach the Owens River (Hollett et al., 1991). Today, most of the water in the Owens River is diverted to Los Angeles, and the lake basin is almost dry. The Owens Lake playa is surrounded by desert saltbush (A triplex) and shadscale (Sarcobatus) plant communities. Salt grass (Distichlis) grows on alkaline soils near the playa, whereas grasses (Poaceae), sedges (Scirpus), and reeds (Phrag- mites), along with riparian communities of cottonwoods (Popu- lus) and willows (Salix), characterize the valley floor vegeta- tion close to the Owens River (Hollermann, 1973). Shrub communities of creosote bush (Larrea) grow south of the playa lake, whereas sagebrush (Artemisia) and blackbrush (Coleog- yne) dominate upstream and to the north of the basin. Scattered stands of Utah juniper (Juniperus osteosperma) and single-leaf pinyon (Pinus monophylla), as well as sagebrush, occupy sur- rounding mountain slopes as low as 1600 m elevation (Litwin et al., 1997). Along the Owens Valley, these conifers can form open pygmy conifer forests above 2100 m elevation (Koehler and Anderson, 1994). Montane and subalpine forests of pine (Pinus) and fir (Abies) occupy elevations between 2500 m and the tree line at 3600 m (Hollermann, 1973). 38>o 37C Owens Lake 36U - present playas and lakes Pleistocene lakes ? Pleistocene drainage packrat middens coresOL-90/1+2 & OL 84/B marine records 119 FIGURE 1 ?Index map showing Owens Lake and Owens River drainage system in southeastern California and locations of marine cores along the California margin. B = Bishop, C = Crowley Reservoir, HC - Hot Creek, I - Independence, LP = Lone Pine, LV = Long Valley, T = Tinemaha. NUMBER 33 147 CLIMATE Owens Lake has a wet-winter climate regime. Moisture is delivered to the west coast of California by the southward mi- gration of the Polar Front (and the associated Aleutian Low) as Arctic high-pressure systems expand and coalesce in the win- ter. As the Subtropical (Hawaiian) High weakens and moves southwestward, westerly air flow drives storms toward the Si- erra Nevada where heavy snowfall accumulates; however, be- cause Owens Lake lies to the east of the Sierra Nevada, a pro- nounced rain shadow affects the basin. Mean annual pre- cipitation is only about 14 cm; about 80 percent of which falls in the winter (Smith and Bischoff, 1997). Annual temperatures average 15.6? C. Lake evaporation rates are high (1.5 m/yr), and arid conditions prevail, especially during the summer when temperatures are high (Figure 2). West of Owens Lake, up to the crest of the Sierra Nevada, precipitation, mostly as snow, accumulates rapidly to local annual amounts greater than 150 cm. Today, annual precipitation increases at a rate of about 23 cm per 1000 m elevation in this area (Smith and Bischoff, 1997). Eighty percent of the variation in California precipitation re- lates to the direction and speed of the Polar Front jet stream over the California coast (Bell and Basist, 1994). Interannual variation of precipitation near Owens Lake occurs as irregular, circa-decadal cycles (Figure 3). These cycles may reflect un- stable air-sea interactions in the North Pacific that ultimately govern the speed and direction of the jet stream (Latif and Bar- nett, 1994). Heavy precipitation years may or may not co- incide with the El Nino-Southern Oscillation pattern. Substan- tial summer precipitation at Owens Lake occurs mostly when tropical cyclones bring moisture from the Gulf of California max. temperature min. temperature IV V VI VII VIII IX X XI XII MONTH precipitation FIGURE 2.?Thirty-year (1961-1990), monthly normals (mean values) of pre- cipitation, and maximum and minimum temperatures at Independence, Califor- nia. Data are from the National Weather Service as contained on the World Weather Disk, 1994 edition, version 1.4. and the adjacent Pacific Ocean. Such storms often characterize those years when the previous winter precipitation in the area was low. HYDROLOGY SURFACE WATER.?In today's climate, precipitation at Owens Lake is limited and has minor relevance to hydrology 1980 1970 1960 1950 1940 1930 1920 1910 1900 1890 i " " i " " i " " i 10 20 30 40 precipitation (cm) 50 FIGURE 3.?Variation of annual precipitation at Independence, California (water years 1899-1980). A water year begins on 1 October and ends on 30 September of the following year. Silhouette with horizontal lines represents summer (May to October) precipitation. Data are from the State of California, Department of Water Resources. 148 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES and limnology. Nevertheless, wet winters in the Owens Valley are positively correlated with heavy snowfall in the Sierra Ne- vada (Figure 4), and snowmelt runoff dominates flow in the Owens River. Consequently, there is a positive relationship be- tween winter precipitation and upper Owens River discharge 40 if. o 8c o pend ?ac oop CDCM O = ? X n (cmo as +->'Q. 'oa> 35 30 25 20 15 10 ': ; : ? : V : Ml?r i ?-1 'l- ? _i|l i ? i _ 1 ? ? ? ? 31-1969 0 10 20 30 40 50 60 70 80 90 precipitation (cm) in snow maximum yearly value (1926-80) Rock Creek, Bishop, Calif. FIGURE 4.?Relationship between annual maximum water content of snow averaged for three Rock Creek snow course stations (elevations of 3049 m, 2759 m, and 2652 m) 30 km WNW of Bishop, California (37?27.0'N, 118?44.5'W), and winter precipitation [December (XII) to March (III)] at Inde- pendence, California (elevation 1195 m). Snow data are from the State of Cali- fornia Department of Water Resources, California Cooperative Snow Surveys. Precipitation data sets span 1926-1980. The 1969 data point identifies data for that year. (Figure 5). Before diversion of water from the Owens River to Los Angeles, the level of Owens Lake was related to Sierra Ne- vada winter precipitation and the resulting discharge of snow- melt down the Owens River. For example, the abundant local precipitation and snowfall during the winter of 1968-1969 (Figures 3-5) so greatly exceeded the capacity of the Los An- geles Aqueduct that the excess runoff was delivered to Owens Lake, raising the level of the almost dry playa to a depth of 2.4 m (Friedman et al., 1976). Two years later, Owens Lake had al- most desiccated again, and dissolved salts reprecipitated over the playa floor. The unusually abundant precipitation that occurred during the winter of 1968-1969 provided a perspective for evaluating the capability of surface water to fill and maintain Owens Lake by indicating how sensitive this basin is to increases in effec- tive moisture. Runoff from snowmelt and precipitation at the town of Lone Pine (19 km upstream from Owens Lake) reached 26.8 cm (more than twice the normal) that season and added 390 hm3 to Owens Lake (Friedman et al., 1976). At that time, the water diverted to Los Angeles equaled 350 hm3; if this water had not been diverted, it would have provided a po- tential total discharge to Owens Lake of 740 hm3 (Smith and Street-Perrot, 1983). This potential total discharge was 87 per- cent of that required to maintain a full and overflowing Owens Lake at present temperatures and evaporation rates. Smith and Street-Perrot (1983) also calculated that the annual discharge of Owens River would have to be 210 percent of historic values (410 hm3) to fill the basin to overflowing. The relative dis- charge of the Owens River and its tributaries during water year 1969 (October 1968 through September 1969) was 180 percent of the historic discharge. After the wet winter (January and February) and spring of 1969, evaporative loss exceeded sur- face-water input, and Owens Lake desiccated by 1971. Never- reO 8c 0)?o c2. Q)75 C ,/yrE co cL 'o2 Q. 50 40 30 ? 20 10 - ? ??? ? ? ? ? ? ? m , ? ? ?? 100 -1969 150 200 250 300 350 400 450 Upper Owens River discharge (hm /yr) FIGURE 5.?Relationship between annual precipitation at Independence, California, and full natural flow dis- charge on the upper Owens River (Long Valley to Tinemaha) (see Figure 1). Data span 1962-1992. Discharge data are from the State of California Department of Water Resources, Division of Flood Management. The 1969 data point identifies all data for that year. NUMBER 33 149 theless, the magnitude of the anomalous precipitation during the winter of 1968-1969 was almost the amount needed to fill the lake, especially if mean annual temperatures had been sev- eral degrees lower (Smith and Street-Perrot, 1983, tbl. 10.2). The major dissolved ions in high-elevation water from the Sierra Nevada and White Mountains along the Owens River drainage are Ca and HCO3 (Table 1). The ratio of total alkalin- ity (mostly HCO3, in meq/L) to Ca is usually <2 (see Table 1). Sodium (Na), Mg, and Cl are present in low concentrations compared with the other major ions, except in White Mountain drainages that have somewhat higher Mg concentrations de- rived from dolomite dissolution. Concentration of total dis- solved solids (TDS) in high-elevation water is low, mostly less than 200 mg/L. As high-elevation surface water flows down the Owens Valley, it becomes enriched in Na and total alkalin- ity (HCO3) that the river gains from groundwater input, which is modified by interaction with volcanic rock. Locally, water from highly alkaline, thermal springs (Table 1) enters tributar- ies to the Owens River, making water in the river more alka- line. GROUNDWATER.?The input of groundwater via springs and seeps to the immediate Owens Lake groundwater basin (39 km in length) was estimated to be 28.6 hm3 (Lopes, 1987). Hollett et al. (1991) estimated spring and seep discharge along the Owens Valley aquifer north of Owens Lake (108 km in length) to be 40.7 hm3. If these figures are approximately correct, they indicate that proportionally more spring discharge occurs in the immediate Owens Lake groundwater basin than along the Owens Valley. The combined discharge, about 69 hm3, is about. 20 percent of the average natural discharge of the Owens River (370 hm3) at Lone Pine (Gale, 1914). Accordingly, groundwa- ter input could have an important effect in determining the vol- ume of Owens Lake were it not for withdrawal of water to the Los Angeles Aqueduct. Springwater and groundwater chemistries along the Owens Valley have two distinct categories: low-salinity water domi- nated by Ca and HCO3 and alkalinity-enriched water of modest to high salinity (Hollett et al., 1991). Groundwater from rela- tively shallow, permeable, hydrogeologic units that is chiefly recharged through alluvial fans by Sierra Nevada runoff char- acterizes the low-salinity water. This type of groundwater is typical of the west side of Owens Lake and of the aquifer asso- ciated with the Owens River delta at Owens Lake. Deep and presumably old groundwater associated with active volcanic processes or with ancient deposits from Owens Lake is saline and high in Na, HCO3, and Cl. This alkalinity-enriched groundwater characterizes the east side of Owens Lake where it rises along permeable fracture zones and mixes to variable degrees with fresher, shallow groundwater sources associated with alluvial fans (Font, 1995). Large, active springs, such as Casa Diablo Hot Springs and others along Hot Creek, discharge into the Owens River system in the Long Valley caldera area (Figure 1) and can supply large amounts of dissolved solids to surface water that ultimately flows to Owens Lake (Smith, 1976a; Menking et al., 1997). Crowley Reservoir, on the Owens River, receives a mixture of alkaline-enriched springwater and relatively low-salinity, cal- cium bicarbonate water from high-elevation Sierra Nevada sources. We estimate that about 10 percent of Crowley Reser- voir water input comes from springs and seeps in the Long Val- ley caldera region. This estimate was based upon the percent of Cl in the Hot Creek thermal springs (Smith, 1976b) and in the high-elevation Sierra Nevada water sources (Table 1). The input of alkaline-enriched groundwater via springs and seeps is important in the hydrochemistry of Owens Lake, espe- cially during times of low surface-water discharge. Low-eleva- tion water derived from spring discharge is dominated by Na, HCO3, and Cl. Sulfate is variable, and Ca and Mg are at low concentrations. The alkalinity to Ca ratio (alk/Ca) is large, generally ?2, and the TDS are high, often greater than 500 mg/L (Table 1). During evaporation, the low-elevation, alka- line-enriched springwater and seep water becomes increasingly more concentrated with solutions of Na and HCO3. This hydro- chemistry characterized Owens Lake in the late nineteenth cen- tury (Gale, 1914), because surface-water input was low and al- kaline-enriched springwater input was proportionally high during arid climate conditions. Owens Lake hydrochemistry is predisposed toward Na, HCO3, and Cl enrichment from the chemical character of its groundwater input, and dissolved sol- ids in the lake water can undergo considerable concentration before saline minerals begin to precipitate and change the basic hydrochemistry. In such a system, Ca entering the lake via the Owens River would soon precipitate as CaCO3. LIMNOLOGY The earliest limnological observations concerning Owens Lake are from the 1861 Nevada-California border survey (Woodward, 1961). The cited description (on 27 March 1861) indicated an approximately oval or pear-shaped lake about 26 km long and 15 km wide, with an extensive, shallow littoral area. A white alkali efflorescence bordered the lake margin; this was replaced farther upslope by a belt of salt grass (Dis- tichlisl) and rushes (probably Scirpus). High winds made the water turbulent and turbid. In the middle to late nineteenth cen- tury, Owens Lake was a closed basin about 15m deep and alka- line. Between 1872 and 1905, the lake received only about 25 percent of its natural input because of water divergence for irri- gation in the Owens Valley (Gale, 1914). Nevertheless, the in- flow was sufficient to maintain a perennial, if saline, lake about 6 m deep. The construction of the Los Angeles Aqueduct in 1913 removed the principal source of water to the lake, and by 1923 it desiccated (Smith, 1976b; Tyler et al., 1997). If filled to its spill point today (1145 m), Owens Lake would be about 65 m deep with an area of 622 km2 and a volume of 22,100 hm3 (BischoffetaL, 1997a). In its low-stage, Holocene phase, Owens Lake would have had a seasonal (early summer) input of freshwater that would 150 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ^7 Dd S S "a 2 5? tr ? s <=? ?22 a ?; T3 Ji Alk/Ca 6* Q H ___ oon d1 K 2 03 13 [u T3 oo ,-; i Lai_o 03 > ide Ek '5b de Loi3 Latil /?, ~60 3 3 a ^ i UI) j B, ?i 60 _J B, ^ ^ ?^ ? ~Sb ~ebUI) 33a. ? > ality c1 ? NO ON ? m CN _ ? ON ? ro ^? ro ? ? ? ? ? CNP~ NO CN cnr- CN ?? ON ?NO 37. Mammoth o ? NO ON ? NO ? 00 ? CN '?' in o CN ? ? P- CN ON in NO CN fNro ON oo roNO r- ro Jxi a Sherwin C; ? ONro oo CN OO ? ? ?< fN ? ro ? p- ??> ? fN _: ?. ? 00 ro sr- 00ro in ? ? ON 37. M Deadman Cree oo t?1 p- 109 fN ?i ? ON ^. NO inro ^? in ?CN CN ON P- 37. Q 5 Bridge 1 w *?"' rN? -aB CN CN~? CN ro00 00 NONO ro 60 a. bek Hot Creek ? ? CNfN ? ? ^^ P~ ? ?* Tj" ON ro ON ro ? ?ro ON 00 p- 00 00^^ CN ? CN OO 00p~ ro (WM) Marble Crf ? ^o NO 16. ON ? ? in ? in ? CN |n __ CN ro _H ro ? ?fN 00 ?in ^-< oo NOP- ro00 ?NO ro (WM) Millner Cr ? ?ro fN ON ro ? ? in NO NO ? NO ?n p- fN ? NO fN Tf tN ? ^ ro ^~ NO00 NOCN ONro 00 ooNO 37. 5 lant (Wl 0. Moss Powt ? ? p~ 23. fN ? P^ m ?ro ? ro Tf" NO 00 ? ro 00 ? in ? ON CNrf OO ?r- ro ONP- ro ro 00 ? 37. Creek ( 1 Silver Can; ? ? ?ON ro 1,080 CN CN o ? CNro NO in in ro fN ? 00r<1 m ON ^J- ON ? ? NOin in NO P- roON p*- ? ON? ?oo in NO ro slough)- 'er( Owens Riv E T3B ?o fN CN ON _ ON roro CN fN ro NO CN in m in ?CN p^ ?NO P- _ CN ?ro "O B ro *?i CN roCN 00 ON p- ro oo? ON erl Owens Riv ? ?aB 330 ON ? ro CN 00 ro i-~ NO CN CN p~ __ CN ON ro ON _' ro ? NO CN *?< 00 inON CN 00 ? roCN 18. NOo 37. J:CS it Tinem?3 o Owens Riv V) "S Reserv ?oa Lakes o 00 fN ? in ? NO Tj- CN ? i?i ON in ??i n NO ? P^ ? in ? oo ro ? 00 ? ? ? NO , ! in ON in 00 p- ro Q s 03 2 a ? ONro P- ro SO fNro fN ? in CN in ?CN in ? _' NO ? ON in CN tNNO ^^ CN NOro ON in ^-* 00ro a ike WRJ B Upper Twi ? ?in 00 CN rN cn NO ro ? ^- NOCN NO ? 00 00 ? ro t~~ NO P- NOin oo in"- 1 CN ro ON NO *-? 00ro Q akeWR B Lower Twi ? ? fNON 210 r^ roro ? NO 00 8.2 *?' ? tN^^ NO inCN NO 00 CN ?ON CN _ ro NO 00 00 ON ?NO _" ro ON? 00(? 00 NO ro ce south Lai Big Alkali ? NO r~"1 CN00 784 NO CN ? ON r~ CN 00 ? in NO <*3- ? Little Lake B ?oB ro ? p- 00 ro ? ?CN p-00 ri CNCN CN 00 ??NO ONro 00in ro in ? NO in "" T3B CN CN fN ro 00 00 NO NO ro rings ing8 5. B. Hot Creek -aB TJB ? ? >n o CN00 ro00 fN *~ in fN p~00 ?p- in ^- CNro _< NO ro ? m fN '?"' "OB CN CN"~ CO 00 00 NO NO p- ro rser8?-, u60 Hot Creek NUMBER 33 151 on CQ a 22 o 00 QH U o 00 m ^^200 g/L TDS as a result of short-term climate variation and increasing use of water for irrigation. The lake water was a so- dium bicarbonate plus carbonate brine that supported abundant brine flies (Ephydra), brine shrimp (Artemia), and assorted microorganisms (infusoria) (Lee, 1906). Fishes, such as the Owens sucker (Catostomus fumeiventris Miller), now living in the Owens River might have survived in the lake during peri- ods of freshwater inflow and stratified lake conditions until high salinities eliminated the populations, as documented by Smith etal. (1987). The lake could have stratified thermally, at least for short pe- riods of calm, warm weather (Friedman et al., 1976), but at its shallow-depth stages, the water column usually would have mixed completely, as reported in 1861. Most of the time, there- fore, bottom-water temperature would have been close to the mean monthly air temperature (Smith et al., 1987). Possibly, short-lived periods of chemical stratification also existed dur- ing the runoff season when less dense, freshwater could float on top of dense, saline water. Owens Lake was almost certainly a warm, polymictic system?a lake mixing throughout the year without ice cover (Lewis, 1983). Pyramid and Walker Lakes in Nevada, 250 and 400 km north of Owens Lake, respectively, do not freeze today, although they are fresher and deeper than the historic Owens Lake. At high stages (e.g., 30+ m in depth), Owens Lake probably would have resembled the warm mo- nomictic Pyramid and Walker Lakes, which circulate in the winter and thermally stratify in the summer. DIATOM ECOLOGY Diatoms certainly lived in Owens Lake in the late nineteenth and early twentieth centuries, even when lake salinity was al- most 100 g/L, but apparently no information exists from those times. The modern diatom flora of Mono Lake may be repre- sentative of the diatom communities of the historic Owens Lake, at least in part. Historically, Mono and Owens Lakes had similar chemistries and high alk/Ca ratios (Lee, 1906). The present Mono Lake salinity of about 100 g/L (Table 1) is of the same order of magnitude as the salinity for historic Owens Lake before the diversion of water from the Owens River, al- though considerable variation in salinity characterized both lakes in the early twentieth century (Lee, 1906; Gale, 1914). In littoral samples between 0.15 and 10 m depths, Kociolek and Herbst (1992) listed Anomoeoneis sphaerophora (Ehrenberg) 152 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Pfitzer, Navicula crucialis (O. Miiller) Frenguelli, Nitzschia frustulum (Kutzing) Grunow, Nitzschia latens Hustedt, and a new species, Nitzschia monoensis Kociolek and Herbst, as the numerically important diatoms in Mono Lake. The same, or closely related, species occur in the Holocene sediments of Owens Lake. Diatoms in the Owens Lake sediments may be grouped into taxa having ecological associations with freshwater (<1 g/L TDS) and saline water (> 1 g/L TDS). Many species found in the cores overlap in their salinity tolerances, so these associa- tions need to be used carefully. For example, although most of the taxa considered to live in freshwater have optima well be- low 1 g/L TDS, they may inhabit waters two to three times more saline. Furthermore, saline taxa, which often tolerate (or even prosper in) a broad range of salinities above 5 g/L TDS, may be present in waters with less than 1 g/L TDS. Diatoms occupy different physical habitats in the photic zone of lakes: planktonic diatoms remain in open water and are sus- pended by turbulence; benthic semiattached diatoms are loosely attached to substrates, such as submerged vascular plants; benthic motile diatoms are capable of moving over and through the mud in illuminated areas of the lake bottom; and benthic ad- nate diatoms are firmly attached by stalks or pads of mucilage to substrates in the shallowest parts of lakes and can withstand strong currents. In modern lakes, these habitats are more or less distributed along a water depth gradient (deeper to shallower) controlled by the photic zone and the effect of currents. As with salinity, diatoms overlap in habitat, especially sea- sonally. Semiattached species may become planktonic when there is sufficient turbulence, as do some benthic motile species of Campylodiscus, Nitzschia, and Surirella. Planktonic species often settle to the lake bottom until favorable conditions return to support them in the open-water photic zone. These criteria and habitat options characterize three principal ecological groups that have more or less coherent ecological distributions (Figures 6-8). 1. Saline diatoms (Figure 6). Benthic species: Amphora coffaeiformis (Agardh) Kiitzing, Anomoeoneis sphaerophora, Campylodiscus clypeus Ehrenberg, Navicula crucialis, Navic- ula pygmaea Kiitzing, Nitzschia frustulum, Nitzschia monoen- sis, Nitzschia pus ilia Grunow, Rhopalodia acuminata Kram- mer, Surirella hoefleri Hustedt, Surirella ovalis Brebisson, Surirella striatula Turpin. Planktonic species: Chaetoceros muelleri Lemmermann, Cyclostephanos sp., Cyclotella me- neghiniana Kiitzing, Cyclotella quillensis Bailey. 2. Freshwater benthic diatoms (Figure 7). Attached spe- cies: Achnanthes lanceolata (Brebisson) Grunow, Amphora ovalis (Kutzing) Kutzing, Amphoraperpusilla Grunow, Cocco- neis placentula Ehrenberg, Cymbella muelleri Hustedt, Epi- themia adnata (Kutzing) Brebisson, Epithemia turgida (Ehren- berg) Kiitzing, Fragilaria capucina Desmazieres, Rhoicosphenia curvata (Kutzing) Grunow, Rhopalodia gibba (Ehrenberg) O. Muller. Semiattached species: Fragilaria vaucheriae (Kutzing) Peterson, Pseudostaumsira brevistriata (Grunow) Williams & Round, Staurosira construens (Ehren- berg) Williams & Round, Staurosirella pinnata (Ehrenberg) Williams & Round, Synedra rumpens Kutzing, Synedra ulna (Nitzsch) Ehrenberg. 3. Freshwater planktonic diatoms (Figure 8). All species of Aulacoseira and Stephanodiscus; Asterionella formosa Has- sall, Cyclotella bodanica Grunow, Cyclotella ocellata Pantoc- sek, some Cyclostephanos spp., Fragilaria crotonensis Kitton, Tabellaria flocculosa (Roth) Kutzing. Because of the great variety and abundance of freshwater planktonic diatoms in the Owens Lake cores, the ecological dy- namics and limnoclimatic relationships of these species are im- portant in the interpretations of past environments in Owens Lake. Stephanodiscus species generally characterize phytoplankton blooms in the spring or fall (Kilham et al., 1986; Bradbury, 1988). In the Owens Lake drainage, Stephanodiscus oregoni- cus (Ehrenberg) Hakansson and the small Stephanodiscus spe- cies (S. minutulus (Kutzing) Round and S. parvus Stoermer and Hakansson) are indicators of major hydrologic input in the spring, probably associated with melting of the Sierra Nevada snowpack and discharge into the Owens River. In March 1996, S. oregonicus was numerically dominant within the phy- toplankton communities in Klamath Lake, Oregon (Figure 8). Stephanodiscus niagarae Ehrenberg is probably a fall species because it was numerically dominant in the September 1994 bloom at Crowley Reservoir, upstream from Owens Lake (J.P.B., unpublished data). Except for Stephanodiscus niagarae, the Stephanodiscus species common in the Owens Lake cores cannot be perfectly related to existing taxa in the Western United States. In general, the fossil Owens Lake species are more robust and more heavily silicified than modern relatives, perhaps as a result of distinctive nutrient conditions that existed in ancient Owens Lake (Theriot and Stoermer, 1981). Descriptions of some Qua- ternary fossil Stephanodiscus taxa, however, fit the morpholog- ical characteristics of the fossil Owens Lake diatoms fairly well, although much variation exists in the assemblages. For example, the fossil taxa Stephanodiscus subtransylvanicus Gasse and S. asteroides Gasse (Gasse, 1980) are morphologi- cally close to S. oregonicus (and to each other), but they tend to be larger, heavier, and more coarsely structured than modern representatives of S. oregonicus in the western United States. Because S. asteroides compares favorably with the Owens Lake fossil Stephanodiscus taxa, this epithet is used for this species complex in the core studies. Because of its general sim- ilarities to S. oregonicus, S. asteroides was assumed to occupy a similar ecological niche as S. oregonicus, i.e., blooming in the spring shortly after ice breakup when circulation and river inflow provide abundant nutrients. Asterionella formosa blooms in low-conductivity lakes when phosphorus to silicon ratios are low, usually in the early sum- mer or early fall (Kilham and Kilham, 1978). Its presence in Owens Lake would indicate conditions of comparatively dilute freshwater and seasonal nutrient supplies governed by phy- toplankton succession and lake stratification. Asterionella for- NUMBER 33 153 Site TDS (mg/L) ,55- *&f y 118 632 247 133 128 Nitzschia monoensis Navicula crucial is Nitzschia pusilla Nitzschia frustulum Amphora coffaeiformis Anomoeoneis sphaerophora Surirella spp. aff. S. ovalis Surirella striatula Campylodiscus clypeus Rhopalodia acuminata PLANKTONIC DIATOMS Chaetoceros muelleri Cyclotella quillensis Cyclostephanos sp. Cyclotella meneghiniana FIGURE 6.?Estimated abundance of saline-habitat diatoms from modern aquatic environments in the western Great Basin. Sites along the abscissa are arranged in order of decreasing salinity (see Table 1). mosa was common in phytoplankton collections taken in June 1996 from Sierra Nevada lakes and reservoirs above an eleva- tion of 2000 m (J.P.B., unpublished data). Little is known about the ecology and nutrient requirements of the Cyclostephanos species, although some prosper in highly eutrophic and often turbid freshwater systems (Hakansson and Kling, 1990). The species found in the sediments of Owens Lake presently lives in Utah Lake (500-1300 mg/L TDS) dur- ing the winter (under ice) and early spring (Squires and Rush- forth, 1986; they initially identified this diatom as Stephanodis- cus dubius). In the Owens Lake system, this taxon also probably represented winter or very early spring diatom pro- duction. A similar diatom lives in hypereutrophic Upper Cow Lake in Malheur County, Oregon, and in Diaz Lake, about 4 km northwest of the Owens Lake playa (J.P.B., unpublished data). Aulacoseira granulata (Ehrenberg) Simonsen and A. am- bigua (Grunow) Simonsen bloom when wind-driven mixing of the water and a deepening of the thermocline supply nutrients, generally phosphorus and silica, from the hypolimnion and when turbulence maintains the diatoms in the photic zone. 154 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Site TDS (mg/L) <2000 m) in the Sierra Nevada. Limnocythere sappaensis lives in saline springs around Owens Lake, such as Dirty Socks Spring and the Keeler springs on the eastern margin of Owens Lake, and in wetlands and shallow lakes having high alk/Ca ra- tios (>7) along the Owens River (e.g., Big Alkali Lake and Lit- tle Lake). Limnocythere ceriotuberosa lives in Diaz Lake, which varies seasonally in water volume and amount of TDS (as a result of controlled input from the Los Angeles Aqueduct) and has a moderately elevated alk/Ca ratio of just more than 3. This species has not been collected elsewhere in the Owens River drainage. Cytherissa lacustris does not live in the Owens River drain- age nor in any lake south of Canada that was sampled during this study, except for rare populations in the Great Lakes and in Yellowstone Lake. In Canada and Alaska, C. lacustris lives in dilute, oligotrophic, cold-water lakes of the boreal forest. De- lorme (1989) reported the upper TDS tolerance of C. lacustris at 300 mg/L. The level of TDS alone, however, does not ac- count for the distribution of C. lacustris. This species also re- quires lakes that have interseasonal chemical stability and low, or at least predictable, thermal variation. The biogeographic distribution, temperature, and chemical preferences of these four ostracode species provide a means of understanding the probable climate and limnological history of Owens Lake. The presence of Cytherissa lacustris in core sedi- ments implies there once was an almost continuous residence of polar air in the Owens Lake region. Today, this air mass re- mains in the boreal regions of Canada and southern Alaska. The presence of polar air in the Owens Lake drainage during past climates reflects the continental glaciation that existed in northern North America, which maintained polar air south of the ice-sheet margin. The subsequent disappearance of C. lacustris from the fossil record implies the retreat of polar air northward and the start of deglaciation. From a limnological standpoint, the presence of C. lacustris implies, by analogy, that a dimictic lake with a long ice-cover season existed in the Owens Lake basin. Candona caudata may be present with Cytherissa lacustris, but Candona caudata has a wider distribution in lakes and streams well south of Canada and Alaska. The southernmost modern record of Candona caudata is in southern Nevada, and its upper TDS limit is about 3 g/L (Bradbury et al., 1989). At Owens Lake, the presence in the fossil record of Candona cau- data without Cytherissa lacustris implies the generally fresh- water conditions and temperature regimes that are characteris- tic of the Great Basin today also occurred in the past. To maintain freshwater habitats in the shallow lake basin, flow in the Owens River needed to have been sufficient on a seasonal basis to offset potentially high valley-bottom evaporation rates; therefore, the presence of Candona caudata in the fossil record indicates a positive hydrological balance that, at least, partly reflects greater precipitation at that time. Limnocythere ceriotuberosa usually lives in lakes that have strong seasonality in volume and associated chemistry. It has been collected alive in lakes that have a TDS level as low as 400 mg/L and as high as 12 g/L. Lakes such as Pyramid and Walker Lakes, which seasonally freshen and increase in size from surface flow but then shrink and become more saline once surface flow decreases, are common settings for this ostracode. Often these limnological changes in such lakes represent a sea- sonal riverine hydrological shift from surface flow derived from snowmelt to base flow derived from shallow aquifers. Generally, such lakes and rivers are in semiarid, temperate set- tings. Increased abundance of L. ceriotuberosa in Owens Lake deposits suggests surface-water discharge into the Owens River and then to Owens Lake varied seasonally. Limnocythere ceriotuberosa often appeared in the Owens Lake fossil record during transitions between glacial and interglacial climates. Limnocythere sappaensis commonly lives in wetlands and lakes that are saline or that maintain a saline, alkaline-enriched solute composition during the year (Forester, 1983, 1986). The species lives in wetlands and lakes that have salinities as low as 650 mg/L and as high as 110 g/L. The solute and salinity char- acteristics of its habitat indicate that L. sappaensis would have lived in Owens Lake during arid interglacial periods when al- kaline base flow predominated in the Owens River, which re- sulted in a generally low-volume lake with alkaline, saline hy- drochemistry. Present-day Owens River water, if combined with alkaline spring discharge and allowed to enter Owens Lake, could support L. sappaensis. Like diatoms, ostracodes disperse rapidly and easily, and they proliferate quickly. Because of their larger size, most adult and juvenile ostracode transport is by birds. Ostracode eggs, some of which are sticky and require desiccation to hatch, may be transported by fish, birds, insects, and other terrestrial ani- mals, or perhaps by wind. Ostracode eggs apparently survive transport through the digestive tracts of most transport animals. Fossil Record from Owens Lake Basin METHODS The paleolimnological history of Owens Lake is based upon three continuous sediment cores. Core OL-84-B was raised in 1984 from near the center of the Owens Lake playa (Figure 1) (36?23.91'N, 117?56.142'W) using a modified Livingstone piston corer (Lund et al., 1996; Benson et al., 1997). It bot- tomed at a depth of 11.6 m. Cores OL-90-1 and OL-90-2 were NUMBER 33 157 obtained about 3 km southwest of core OL-84-B using a truck- mounted, hydraulic, split-spoon sampler (Benson et al, 1996). Cores OL-90-1 and OL-90-2 bottomed at depths of 32.8 and 28.2 m, respectively. For diatom and ostracode analyses, a 34 m composite section (OL-90-1+2) containing the upper 28.16 m of core OL-90-2 and the basal 5.91 m of core OL-90-1 was constructed from a magnetic susceptibility correlation at 3 cm intervals (Benson et al., 1996). One-half cubic centimeter splits of magnetic susceptibility samples from core OL-84-B and from the upper 16 m of core OL-90-2 provided material for diatom studies. The remaining samples from cores OL-90-2 and OL-90-1 were excised with a spatula and placed in labeled, air-tight plastic bags. Approxi- mately 0.5 cm3 subsamples were processed for diatoms by us- ing concentrated HNO3 in a water bath at 100? C. After the res- idue was washed free of acid and dissolved solids, it was settled onto coverslips and mounted in naphrax (Bradbury, 1997a). Diatom data were presented as relative abundance. In samples that had diatom valve counts less than 100, an artifi- cial count of 100 was used to calculate percentages, thereby un- derweighting relative diatom abundances in samples of low di- atom concentration. Silt-sized heavy-mineral grains of magnetite, zircon, and monazite were tallied from the diatom microslides on the basis of color, transparency, and crystal form. Heavy-mineral grain concentration is presented as the number of grains per millimeter of microscope transect at xlOOO magnification. Heavy-mineral grain identifications were verified by petrography. Estimated abundance of diatoms in modern samples are de- noted as follows: dominant = more than 10 specimens of a taxon in every field of view examined; abundant = taxon oc- curred in every field of view; common = taxon present in every fourth field of view; few = taxon present in one of 10 fields of view; rare = taxon present in one of 20 fields of view; and absent = not seen. Sections of about 3 cm thickness were cut from core OL-84- B between 1.63 m and 9.11 m to obtain ostracode samples. Sections of about 2 cm thickness were cut from cores OL-90- 1+2. The sediment was weighed, disaggregated, and washed through sieves to concentrate the ostracodes, which were re- ported as number of adult valves per gram of sediment. CHRONOLOGY Polynomial-age models were based upon 40 radiocarbon dates from the total organic fraction using accelerator mass spectroscopy (AMS) (Benson et al., 1996; Benson et al., 1997). These provided the chronologies for cores OL-84-B and OL- 90-1+2. The radiocarbon dates indicate hiatuses or periods of very slow deposition at 2.25 m and 9.13 m in core OL-84-B. A hiatus at 9.13 m in core OL-84-B correlates with one at 6.0 m in core OL-90-2. In both cores, this hiatus separates lithostrati- graphic units C and D of Lund et al. (1996) and represents about 2000 years of missing time (and sediment?) between about 15.5 and 13.5 14C Ka in core OL-90-2 (Benson et al., 1996). In core OL-84-B, the C-D transition occurs 30 cm above a radiocarbon date of 15.02 ?0.1 Ka, indicating that, in this core, the hiatus may span only from 14.7 to ~13.5 Ka. Un- calibrated radiocarbon dates used in this paper are reported as thousands of years ago (Ka). According to these chronologies, samples from cores OL-90-1+2 date from 52.5 to 12.7 Ka. Samples from core OL-84-B date between 17.3 and 0.27 Ka. CORES OL-90-1+2 DIATOMS.?Freshwater planktonic diatoms numerically dominate most of the OL-90-1+2 record. Saline diatoms only appear near the base of the core at an extrapolated age of about 52.5 Ka (Figure 9). The saline diatom assemblage is an indica- tor of environmental conditions, including low lake levels, im- portant hydrologic contributions from alkaline, mildly saline springs along the Owens River, and seasonal but reduced input of surface water from the Owens River and tributaries marginal to Owens Lake. Cy do Stephanos spp., Stephanodiscus asteroides, S. niaga- rae, and Aulacoseira ambigua rapidly replaced the saline dia- tom assemblage by 52.0 Ka and continued to characterize Owens Lake until about 41 Ka when Stephanodiscus sp. cf. S. carconensis Grunow became the predominant diatom in the lake. Benthic freshwater diatoms (Figure 10) also formed an important part of the assemblage. Pseudostaurosira, Stauro- sira, and Staurosirella species, hereafter referred to as "Fragilaria" species because they once resided in that genus, fluctuated but increased in abundance from 52 Ka until 41 Ka. This trend indicates a progressively greater representation of shallow-water habitats through time and may imply either a gradual shallowing of the lake caused by sedimentation or the establishment of marsh and shallow, freshwater habitats around Owens Lake and upstream along the Owens River. Other benthic diatom ecological groups had less regular fluctuations. These fluctuations in benthic freshwater diatoms, as well as the fluctuations of the planktonic freshwater diatoms, partly reflect percentage control when one or another species, usually a planktonic diatom, became especially abundant, causing the other taxa to decrease to proportionally lower percentages. Nevertheless, high percentages of both groups document lim- nological changes, which were possibly climate mediated, at comparatively short intervals. Aulacoseira spp. and Stephano- discus niagarae assemblages alternated with minor, but ap- proximately regular, increases in S. asteroides (Figure 9), per- haps indicating alternations of greater and lesser Sierra Nevada precipitation, respectively. In particular, the five regular pulses of S. asteroides between 50 and 43 Ka (Figure 9) indicate a cy- clicity of 1.4 Ky, a millennial-scale rhythm (within the limits of chronological control) that has also been observed in marine cores that reflect discharge of cold, ice-bearing water into the North Atlantic Ocean (Bond et al., 1997). Aulacoseira ambigua, A. granulata, and Stephanodiscus nia- garae bloom today in the summer or early fall in intermediate- 158 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES IftilfiitiitLdtlhWlin^^ 3 2 ?2 a -g a o o rj- O 100 o oCD ent 2CD Q. iato ?u nico c "5. a ;ntriio is, major diatorr _c"I ?S3 oK 55-s: d ^5 oj ^> wens Lai 4. This short-term (500 year) saline event ended by 10.4 Ka when freshwater planktonic diatoms returned to predominance for about 800 years (Figures 14, 15). This freshwater phase in Owens Lake may correlate with sedi- mentological evidence for an early Holocene lake at Silver Lake playa, about 200 km southeast of Owens Lake, that ex- isted just before 9.3 Ka (Enzel et al., 1989). After 9.6 Ka, the record of saline diatoms in Owens Lake be- comes more continuous (Figure 15). A short interval (9.4-8.5 Ka) of mixed freshwater and saline diatom assemblages, includ- ing large numbers of Chaetoceros sp., documents a transition to 166 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES higher lake salinity and indicates arid climate. Today, Chaetoc- eros is an open-water planktonic diatom that is common in Pyr- amid and Walker Lakes (Figure 6). Its wide salinity tolerance (Fritz et al., 1993) may indicate that Owens Lake remained fairly deep during this short interval. The freshwater planktonic diatom {Stephano discus spp.) component of this assemblage (Figure 14) could have prospered in the lake seasonally when spring runoff sufficiently diluted salinities or during short peri- ods when the lake spilled and became fresh. After 8.5 Ka, benthic saline species, such as Campylodiscus clypeus, persisted in the extensive shallow-water habitats cre- ated by low lake levels. By 8 Ka, saline Nitzschia species (Nitz- schia latens, N. pusilla, N. monoensis), which presently inhabit very saline, shallow-water habitats in Mono Lake and similar environments (Kociolek and Herbst, 1992; Blinn, 1993), be- came prominent in response to a substantial decrease in Owens River flow, a proportional increase in alkaline spring discharge, and high evaporation rates. Owens Lake remained saline and shallow until about 6.5 Ka when, according to the radiocarbon chronology, the lake desiccated and clastic sedimentation ended (Benson et al., 1997). Sedimentation of oolitic sand re- sumed after 4.3 Ka, and freshwater and saline diatoms are sparsely represented in core OL-84-B by fragments of corroded valves (Figures 13-15). The presence of freshwater planktonic and benthic diatoms in samples from after about 8 Ka may have resulted from the reworking of diatomaceous highstand deposits left behind as Owens Lake shrank. These reworked di- atoms do, however, indicate episodic fluxes of water into the basin that may reflect flooding during very wet years (Smith et al., 1987). Saline benthic diatoms, including the Nitzschia spe- cies of very saline habitats, returned between 4 and 3 Ka. In-situ diatom productivity certainly occurred throughout the late Holocene in Owens Lake, but, generally, high lake alkalin- ities and salinities combined with episodic desiccation likely destroyed sedimented diatoms by dissolution and breakage. OSTRACODES.?Except for isolated, presumably reworked valves of Cytherissa lacustris from highstand, full-glacial de- posits (Figure 17), core OL-84-B has no substantial late-glacial (15?12.5 Ka) ostracode record. Decreased inflow from the Owens River and a climate regime of increased summer insola- tion would have promoted lake stratification, and seasonal an- oxia may have eliminated ostracode habitats in the central re- gions of the Owens Lake basin. The presence of ostracodes, however, would have indicated that shallow and saline lake conditions existed between 15.5 and 13.5 Ka, but those depos- its could have been removed by erosion. At 11.5 Ka, Limnocythere ceriotuberosa appeared (Figure 17); its presence indicates this was a time of variable lacustrine conditions associated with seasonal filling and freshening of Owens Lake that was followed by lake evaporation, concentra- tion, and shrinkage. Larger pulses of L. ceriotuberosa, accom- panied by Candona caudata, occurred from 11.0 to 10.5 Ka and from 9.9 to 9.5 Ka. These ostracode pulses approximately coincided with the shallow-water, moderately saline lake envi- ronments evidenced by the diatoms, particularly "Fragilaria," Campylodiscus clypeus, Surirella spp., and Cyclotella spp. (Figures 13-15). The maximum salinity tolerance of Candona caudata (3 g/L) and the low (or absent) representation of Lim- nocythere sappaensis also indicate Owens Lake had only mod- est salinity at those times, although the diatoms and L. ceriotu- brosa can tolerate considerably higher salinities. These pulses in the fossil record probably record episodes when the Owens River flooded the basin to produce short-lived, reasonably fresh, shallow lake systems. Near the mouth of the Owens River, at an elevation of 1097 m, coarse deltaic sands that have a freshwater molluscan fauna, including the river clam An- odonta californiensis Lea (dated at 11.4 Ka), also indicate riv- erine environments occurred at the north end of Owens Lake (Koehler, 1995). The large number of Limnocythere sappaensis by 9.4 Ka in- dicates increased salinity and a high alk/Ca ratio. The lake may have become shallower, although this ostracode can live in deep lakes as well. The absence or low abundance of ostra- codes between 9.4 and 8.8 Ka in the section of core having abundant Chaetoceros may indicate a deeper lake was present at that time, perhaps with seasonally anoxic bottom water. Progressively larger pulses of Limnocythere ceriotuberosa characterized the millennium between 8.8 and 7.8 Ka (Figure 17), documenting a return to lacustrine environments that had variable depths and salinities. Large quantities of Heterocypris appeared near the end of this episode. Today, this species re- sides in springs and seep-fed marsh environments. Its presence between 8.8 and 7.8 Ka implies that Owens Lake had shrunk considerably and was surrounded, at least near the mouth of the Owens River, by marsh communities. The marshes probably were dominated by Scirpus, which presently grows abundantly along the Owens River, within the valley, and along the mar- gins of playa springs (Hollermann, 1973). Limnocythere sappaensis dominated the ostracode assem- blage from 7.7 to 6.5 Ka. Occasional Cytherissa lacustris in this interval resulted from the reworking of highstand lake sed- iments deposited during the last glacial maximum (Carter, 1997). During this time, saline, alkaline springwater issuing along the Owens River and around Owens Lake probably had a major effect on the hydrochemistry of the lake, although sur- face water from Sierra Nevada snowmelt continued seasonally to enter the basin. Nitzschia pusilla and N. monoensis (Figure 15), which tolerate highly saline environments, such as at Mono Lake today, coexisted with L. sappaensis. Limnocythere sappaensis remained prominent after the hia- tus in sedimentation (6.1-4.3 Ka) (Figure 17), as did benthic saline diatoms and Nitzschia spp. (Figure 15). These saline lake assemblages persisted until about 3 Ka, when diatoms occurred only as rare, reworked fragments and ostracodes disappeared from the Owens Lake record; however, in core OL-92-3, oolitic sediment, radiocarbon dated between 5.3 and 5.0 Ka (Bischoff et al., 1997c), had abundant L. sappaensis, which indicates NUMBER 33 167 Y*? hiatus 50 0 10 0 1 0200 0 10 0 valves / gram FIGURE 17.?Concentration of ostracode valves versus age in core OL-84-B. 300 some sedimentation occurred in Owens Lake during the middle Holocene drawdown, at least at the site of core OL-92-3. Discussion The diatom and ostracode records from the middle Wiscon- sin to the late Holocene at Owens Lake provide the most con- tinuous paleontological evidence for changing limnological and climatic environments in the southern Great Basin. Be- cause this important record relates to the glacial and vegeta- tional history in this region, it can provide a chronological framework for local glaciation and for isolated vegetation as- semblages found in packrat middens. Conversely, the nonla- custrine evidence of climate change from packrat middens and glacial records can refine climate interpretations from lacus- trine microfossils. RECORD OF GLACIATION.?Owens Lake cores OL-90-1+2 and OL-84-B record the timing and intensity of Sierra Nevada alpine glaciation in the Owens River basin. The substantial quantity of heavy-mineral fragments (magnetite and zircon + monazite) in diatom slides (Figure 11) indicates that glacial erosion of the Sierra Nevada granitic plutons occurred. These minerals were incorporated into the base load of the Owens River. The middle Wisconsin period (~54 to -25 Ka), or oxy- gen isotope stage (OIS) 3, for the most part, had small, isolated heavy-mineral peaks that may represent minor high-elevation glaciation at 52-51, 44, 40, 35, and 26 Ka. The absence of the boreal ostracode Cytherissa lacustris during these periods, however, indicates that severe glacial climates did not affect Owens Lake. In fact, the presence of the warm-season diatoms Aulacoseira spp. during most of these periods seems inconsis- tent with extensive glaciation. These minor peaks in heavy- 168 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES mineral concentration may simply have recorded years or de- cades of exceptional runoff down the Sierra Nevada drainages. Nevertheless, geochemical evidence of glacial-rock flour from Owens Lake core OL-92-1 (Bischoff et al., 1997b), taken in the same area of the basin, approximately correlates with most of the heavy-mineral peaks shown in Figure 11 and supports the interpretation that these peaks represent minor Sierra Nevada glaciations. The heavy-mineral peaks (Figure 11) approximately corre- late with the magnetic-susceptibility and organic-carbon evi- dence of glaciation recorded by Benson et al. (1996). The mag- netic-susceptibility and organic-carbon records, however, contained more peaks, which were interpreted by Benson et al. (1996) as glacial advances, than are present in the heavy-min- eral record. Uncertain equivalencies between radiocarbon and calendar years make correlations to 36Cl-dated moraines (Phil- lips et al., 1996) problematic; however, the potential glacial ad- vance at about 51 Ka, which perhaps correlates to the younger Tahoe moraines, also may correlate with the double peak in heavy minerals between 52 and 51 Ka (Figure 11). The Tioga 1 stage moraines, which were dated to 32-30 calendar Ka (Phil- lips et al., 1996), may correspond to heavy-mineral peaks at 27-25 Ka in core OL-90-1+2, assuming an approximate 4 Ky correction to equate the radiocarbon and calendar chronologies. The episodes defined by high concentrations of heavy minerals generally correspond to perennial lake stages in Searles Lake. The mineralogical evidence of a 5? C cooling in Searles Lake at 26-25 Ka (Smith, 1979:89) might corroborate the occur- rence of a glacial period at that time. According to the heavy-mineral record from cores OL-90- 1+2 (Figure 11), glaciation during OIS 2 (the last major glacial period) began by 23 Ka. The heavy minerals continued to in- crease, but with sharp fluctuations, until about 18 Ka when their values decreased, again with strong fluctuations between greater and lesser quantities. In contrast to the Tioga 1 mo- raines, the Tioga 2 and 3 (glacial maximum) moraine dates (Phillips et al., 1996) correspond closely to the heavy-mineral record of glaciation. The large and persistent peak in magnetite grains centered on 22 Ka probably corresponds to the extensive Tioga 2 moraines identified and dated by Phillips et al. (1996). The several sharp fluctuations in heavy-mineral concentrations afterward may provide much more detail than the Tioga 3 mo- raine sequences, but transportation of heavy minerals may have occurred during episodes of glacial advance and retreat. The sharp fluctuations in heavy-mineral concentrations correlative to the Tioga 3 glacial deposits generally resemble and partly correlate with the sharp fluctuations in the magnetic-suscepti- bility record reported by Benson et al. (1996). The Tioga 4 gla- cial advance at about 15 Ka does not stand out as an especially distinctive peak of large concentrations of heavy minerals in cores OL-90-1+2, although such a peak exists at this time in core OL-84-B (Figure 11). CLIMATE AND LIMNOLOGICAL EVIDENCE FROM PACKRAT MIDDENS.?Radiocarbon-dated packrat middens near Owens Lake document paleoclimate records from fossil vegetation as- semblages. The closest midden to Owens Lake is from the northeast side of the lake (Figure 1), about 10 m above the highstand elevation (1145 m). Juniperus osteosperma domi- nates the full-glacial levels of this midden (23.0-14.5 Ka), but J. scopulorum (Rocky Mountain juniper) and Pinus mono- phylla Torrey and Fremont form large components of the as- semblage at midden levels corresponding to 17 Ka (Koehler and Anderson, 1994). Juniperus scopulorum Sargent now lives 2100-2900 m above Owens Lake (Koehler and Anderson, 1994), but it lived at least 900 m lower during the full-glacial period. The pres- ence oiP. monophylla and J. osteosperma at lower elevations indicates that 17 Ka precipitation was greater by a factor of two to three and temperature was about 7? C lower than today, ac- cording to modern lapse rates (Smith and Bischoff, 1997). To- day in the Great Basin, pinyon-juniper woodland vegetation does not characterize regions where mean January tempera- tures fall below -1? C (Thompson, 1990). Modern mean Janu- ary temperature at Independence, California, 100 m above Owens Lake, is about 5? C (Figure 2), so a 7? C decrease would be somewhat below the tolerance of pinyon-juniper woodland. A full and overflowing Lake Owens, however, could have served as a local thermal buffer against very low (and high) temperatures and thereby may have created suitable conditions for the establishment and growth of P. monophylla and J. osteosperma during the full-glacial period. Packrat midden sites in the Alabama Hills, at elevations of 1264-1460 m and about 17 km northwest of Owens Lake (Fig- ure 1), extend the fossil vegetation record of the Owens Valley to 31.4 Ka (Koehler and Anderson, 1995). Full-glacial middens in the Alabama Hills do not contain P. monophylla, perhaps be- cause of the elevation of this locality. Nevertheless, the pres- ence of Yucca brevifolia (Joshua tree) and abundant J. osteo- sperma at 31.4, 25.7, and 21.1-17.8 Ka indicates winter temperatures generally were above freezing. This association now grows on the east side of Walker Pass, about 90 km south- west of Owens Lake, at an elevation of about 1500 m. As sug- gested above, perhaps the large body of water in the Owens Lake basin during this episode helped mitigate severe winter temperatures in the Alabama Hills. The cold temperature epi- sode from 24 to 21 Ka, implied by the presence of the boreal ostracode Cytherissa lacustris (Figure 12), fell between epi- sodes of mild winter temperatures, as evidenced by Y. brevifo- lia in the Alabama Hills packrat middens. The principal climate change documented by macrofossils from packrat middens in the Alabama Hills occurred at 13.3 Ka when most elements of the modern Sarcobatus scrub plant community appeared (Koehler and Anderson, 1995). Juniperus osteosperma, however, persisted in the midden sequence, indi- cating cooler and wetter conditions, compared with modern en- vironments, prevailed until 9.5 Ka. Diatoms and ostracodes in the late-glacial and early Holocene Owens Lake indicate that freshwater and presumed overflowing lake conditions existed NUMBER 33 169 until 11 Ka. Afterward, lake levels and salinity fluctuated, re- flecting variable, but decreasing, input from the Owens River. The first major appearance of Limnocythere sappaensis, an in- dicator of shallow and saline aquatic habitats, coincided with the disappearance of J. osteosperma and with the complete dominance of desert shrubs and herbs in the middens in Ala- bama Hills at 9.5 Ka. MIDDLE WISCONSIN LINKAGES TO MARINE RECORDS ON THE CALIFORNIA MARGIN.?Isotopic, magnetic, and geochem- ical records from Owens Lake have been tentatively correlated with climate changes recorded in North Atlantic Ocean marine sediments (Benson et al., 1996) and Greenland ice cores (Ben- son et al., 1997). These correlations could imply that hemi- spheric-scale climate changes affected far-separated regions, but the interactive climatic and oceanographic mechanisms that might link changes in Owens Lake to changes in the North At- lantic Ocean and in Greenland were unspecified. Gardner et al. (1988) proposed a marine-terrestrial paleocli- mate linkage for the past 20 Ky between the lacustrine record at Clear Lake in California (latitude 39?N) and an adjacent ma- rine record (core V1-80-P3) on the continental slope off north- ern California (latitude 38.43?N) (Figure 1). Benthic coastal and reworked freshwater diatoms in the marine core from the full-glacial period document lower sea levels and greatly de- creased deep-water upwelling than today. These marine condi- tions correlate with cooler climates in coastal California where Pinus was the dominate vegetation. An episodic, fluctuating transition to marine planktonic diatoms indicates that up- welling off the California coast characterized the establishment of Holocene conditions. On land, warming temperatures fa- vored abundant growth of Quercus (oak) and the local presence of Sequoia (redwood). Coastal fogs produced by upwelling may have been responsible for providing suitable conditions for Sequoia. Upwelling of nutrient-rich marine water on the California margin relates to the strength and persistence of the Hawaiian or Subtropical High, whereas increased winter and spring precipi- tation that causes runoff to California coastal regions reflects the predominance of the Aleutian Low. These two pressure sys- tems affect present-day terrestrial climates and ocean dynamics, although their effects vary from north to south (Gardner et al., 1997). Their important role in glacial, Holocene, and modern marine and terrestrial climate records in western North America implies that similar variations in the strength of the Aleutian Low and the Hawaiian High may also account for earlier pale- oceanographical and paleolimnological changes. The middle Wisconsin, or OIS 3, is represented by the ma- rine core ODP-893 (latitude 34.29?N) from the Santa Barbara Basin (Kennett and Venz, 1995) and by core F2-92-P34 (lati- tude 35.03?N) from the California margin off Point Conception (Gardner et al., 1997) (Figure 1). Both cores have sufficient chronological control for comparison to the Owens Lake pale- olimnological record between 55 and 20 Ka. Perhaps the most distinctive paleolimnological evidence for climate change dur- ing this interval in the Owens Lake core is the distribution of the summer-season planktonic diatoms Aulacoseira granulata and A. ambigua. The most logical explanation for the presence of these species relates to past episodes of exceptional snowfall in the Sierra Nevada. During these episodes, high-volume run- off down the Owens River would have persisted into late sum- mer or fall; this would have provided suitable freshwater lim- netic conditions for Aulacoseira blooms. Intervals in which these diatoms were uncommon or absent represent, by this rea- soning, episodes in which precipitation was less but still suffi- cient to maintain freshwater in Owens Lake, at least during spring and early summer. Using extrapolations from uncorrected radiocarbon chronol- ogies for marine core F2-92-P34 (Gardner et al., 1997) and for Owens Lake, a correlation between small fluctuations in the percent of aluminum present and the abundance of summer Aulacoseira species is evident (Figure 18). Increased concen- trations of aluminum and other geochemical evidence of detri- tal input within the coastal marine cores indicate greater runoff from the Coast Ranges and the Sierra Nevada (Dean et al., 1997), and the same process may explain the approximately co- eval proliferation of Aulacoseira in Owens Lake. Gardner et al. (1997) suggested that increased fluxes of detri- tal material to the coastal sediments during the full-glacial pe- riod were caused by the same moisture source that supplied the glaciers in the Sierra Nevada and the lakes in the Great Basin. An intensification and southerly displacement of the Aleutian Low and the storms associated with the Polar Front jet stream, probably forced by a southward expansion of polar air masses, are the most probable reasons for this moisture source. The same climatic mechanisms can be used to explain similar, but less intense, variations during OIS 3. In core F2-92-P34, intervals that are not characterized by in- creased percentages of aluminum instead contain evidence of increased anoxia, such as high values of molybdenum and or- ganic matter (Dean et al., 1997). In other presumably correla- tive marine cores, anoxic conditions excluded burrowing ma- rine organisms, and zones of laminated sediments resulted. Increased productivity (upwelling) of the California Current apparently provided sufficient marine productivity and organic material that, once deposited, promoted the formation and maintenance of anoxic environments. The history of upwelling intensity of the California Current also can be deduced by the fluctuating presence of the plank- tonic foraminifera Globigerina bulloides D'Orbigny in a core from ODP-893 in the Santa Barbara Basin (Kennett and Venz, 1995). According to the oxygen-isotope chronology developed from this core, fluctuations of G. bulloides (Figure 18) were antithetical to the aforementioned evidence of increased pre- cipitation {Aulacoseira and Al), suggesting that the same cli- matic conditions that increased upwelling also decreased win- ter precipitation along the California coast. In this case, the strength and persistence of the Hawaiian High was the likely candidate because clockwise winds generated around this high- 170 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 20- Aulacoseira (summer) % Globigerina bulloides 20 55 55 Point Conception F2-92-P34 Owens Lake OL-90/1+2 0 50 Santa Barbara Basin ODP #893 FIGURE 18.?Chronological comparison of middle Wisconsin episodes of abundant Aulacoseira species from Owens Lake cores OL-90-1+2, percentage fluctuations of Al from core F2-92-P34 on the California margin west of Point Conception, and percentage of Globigerina bulloides from ODP site #893 in the Santa Barbara basin (Figure 1). Aluminum data are from W.E. Dean (USGS). Foraminifera data are from Kennett and Venz (1995). pressure center would have driven the California Current to the south, which is what happens today. Because this atmospheric circulation pattern would have blocked moisture from the trop- ical Pacific Ocean, it also would have reduced the annual pre- cipitation reaching the Sierra Nevada and, consequently, Owens Lake. Admittedly, correlations of the paleoceanographic and lacus- trine evidence for past climate change in cores from coastal California and from Owens Lake are not perfect, and they rely on chronologies of uncertain accuracy. Nevertheless, this inter- pretation is based upon operationally consistent climatic, ter- restrial, and oceanographic systems and processes that logi- cally force large changes between glacial and interglacial environments. Conclusions High-resolution diatom and ostracode records from cores from Owens Lake have revealed a sensitive and comparatively simple record of Sierra Nevada precipitation for the last 55 Ky. At one extreme, high precipitation, combined with cold tem- peratures associated with the southward expansion of polar air masses, produced glaciation in the Sierra Nevada (Hostetler and Clark, 1997). The Owens Lake basin filled with detrital sediment containing the remains of freshwater diatoms and bo- real ostracodes. At the other extreme, scant winter precipitation failed to offset high evaporation rates, and the Owens Lake ba- sin desiccated to playa or near playa conditions. Intermediate limnological and climatic conditions existed at other times. These conditions included episodes of saline lakes of shallow to moderate depths and regimes of greater than modern precip- itation under temperatures warm enough to preclude extensive glaciation. During the last 50 Ky, the predominant paleocli- matic scenario at Owens Lake included greater precipitation than at present and probably cooler than modern temperatures. During this time, overflow conditions existed for about 40 Ky (50-10 Ka), and only during the last 10 Ky of the Holocene did Owens Lake remain below its spill level for long periods of NUMBER 33 171 time. The paleolimnological record of the last 50 Ky is consis- tent with the longer record of Owens Lake reported by Brad- bury (1997a, 1997b), which documents overflow conditions for about 80 percent of the last 800 Ky. The high-resolution study of limnological and climatic vari- ations at Owens Lake also indicates rapid changes between positive and negative hydrologic balance. For example, the change from closed- to open-basin conditions about 52 Ka oc- curred in less than 300 years. Changes in diatom and ostracode assemblages document the step-like progression of this shift. Holocene changes from fresh to saline diatom floras about 10 Ka indicate that a shift in the hydrologic balance occurred within a century or less. Assuming a reasonable chronological accuracy, this rate of change in salinity was faster than what Bischoff et al. (1997a, fig. 5) predicted for simple evaporative concentration of Owens River water during current climates; thus, saline springwater discharge entering Owens Lake must have had an important hydrochemical effect during arid cli- mate periods. The apparent correlation of the Owens Lake re- cord with the paleoceanographic records indicates that climatic conditions at Owens Lake were not just local, but were part of a global system that involved major reorganizations of atmo- spheric and oceanic circulation. The Owens Lake fossil record supports the contention that climate change can be rapid and that future climate in the Owens Lake area could be wetter rather than drier. Acknowledgments S.P. Lund (University of Southern California) is particularly thanked for providing access to the cores used in this study. L.V. Benson (United States Geological Survey (USGS)) pro- vided much of the core chronology, and he and S.P. Lund of- fered important insights into the history of limnologic and cli- matic changes in the Owens Lake basin. R.S. Thompson, J.L. Bischoff, W.E. Dean, G.I. Smith, J.P. Smoot, R.L. Reynolds (all USGS), and Vera Markgraf (University of Colorado) pro- vided helpful discussion of points raised in this paper. M.G. Paquette and K.L. Conrad (both USGS) prepared the samples for study, and Claire Carter (USGS) identified and counted the ostracodes. This study is part of the USGS Yucca Mountain Fu- ture Climates Project and the USGS Climate History Program. Literature Cited Bell, G.D., and A.N. Basist 1994. The Global Climate of December 1992-February, 1993, Part I: Warm ENSO Conditions Continue in the Tropical Pacific; Califor- nia Drought Abates. Journal of Climate, 7:1581-1605. Benson, L.V., J.W. Burdett, M. Kashgarian, S.P. Lund, F.M. Phillips, and R.O. Rye 1996. Climatic and Hydrologic Oscillations in the Owens Lake Basin and Adjacent Sierra Nevada, California. Science, 274:746-749. Benson, L., J. Burdett, S. Lund, M. Kashgarian, and S. Mensing 1997. Nearly Synchronous Climate Change in the Northern Hemisphere during the Last Glacial Termination. Nature, 388:263-265. Benson, L.V., D.R. Currey, R.I. Dorn, K.R. Lajoie, C.G. Oviatt, S.W Robinson, G.I. Smith, and S. Stine 1990. Chronology of Expansion of Four Great Basin Lake Systems during the Past 35,000 Years. Paleogeography, Paleoclimatology, Paleo- ecology, 78:241-286. Bischoff, J.L., J.P. Fitts, and J.A. Fitzpatrick 1997a. Responses of Sediment Geochemistry to Climate Change in Owens Lake Sediment: An 800 Ky Record of Saline/ Fresh Cycles in Core OL-92. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core 01-92, Owens Lake, Southeast Cal- ifornia. Special Papers of the Geological Society of America, 317: 37-47. Bischoff, J.L., K.M. Menking, J.P. Fitts, and J.A. Fitzpatrick 1997b. Climatic Oscillations 10,000-155,000 Yr B.P. at Owens Lake, Cali- fornia, Reflected in Glacial Rock Flour Abundance and Lake Salin- ity in Core OL-92. Quaternary Research, 48:313-325. Bischoff, J.L., T.W. Stafford, and M. Rubin 1997c. A Time-Depth Scale for Owens Lake Sediments of Core OL-92; Ra- diocarbon Dates and Constant Mass-Accumulation Rate. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Spe- cial Papers of the Geological Society of America, 317:91?98. Blinn, D.W. 1993. Diatom Community Structure along Physiochemical Gradients in Saline Lakes. Ecology, 74:1246-1263. Bond, G., W. Showers, M. Cheseby, R. Lotti, P. Almasi, P. deMenocal, P. Pri- ore, H. Cullen, I. Hajdas, and G. Bonani 1997. A Pervasive Millennial-Scale Cycle in North Atlantic and Holocene Climates. Science, 278:1257-1266. Bradbury, J.P. 1987. Late Holocene Diatom Paleolimnology of Walker Lake, Nevada. Archivfiir Hydrobiologie Supplement, 79( 1): 1-27. 1988. A Climatic-Limnologic Model of Diatom Succession for Paleolim- nological Interpretation of Varved Sediments at Elk Lake, Minne- sota. Journal of Paleolimnology, 1:115-131. 1991. The Late Cenozoic Diatom Stratigraphy and Paleolimnology of Tule Lake, Siskiyou Co., California. Journal of Paleolimnology, 6: 205-255. 1997a. A Diatom-Based Paleohydrologic Record of Climate Change for the Past 800 k.y. from Owens Lake, California. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Society of America, 317:99-112. 1997b. A Diatom Record of Climate and Hydrology for the Past 200 ka from Owens Lake, California, with Comparison to Other Great Ba- sin Records. Quaternary Science Reviews, 16:203-219. Bradbury, J.P., Ye. Bezrukova, G.P. Chernyaeva, S.M. Colman, G. Khursevich, J.W. King, and Ye.V. Likoshway 1994. A Synthesis of Post-Glacial Diatom Records from Lake Baikal. Journal of Paleolimnology, 10:213-252. Bradbury, J.P., R.M. Forester, and R.S. Thompson 1989. Late Quaternary Paleolimnology of Walker Lake, Nevada. Journal of Paleolimnology, 1:249-267. Brugam, R.B. 1983. The Relationship between Fossil Diatom Assemblages and Limno- 172 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES logical Conditions. Hydrobiologia, 98:223-235. Carter, C. 1997. Ostracodes in Core OL-92: Alternation of Saline and Freshwater Forms through Time. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core 01-92, Owens Lake, Southeast California. Special Papers of the Geological Society of America, 317:113-119. Cumming, B.F., S.E. Wilson, R.I. Hall, and J.R Smol 1995. Diatoms from British Columbia (Canada) Lakes and Their Relation- ship to Salinity, Nutrients, and Other Limnological Variables. Bib- liotheca Diatomologica, 31:1-207. Berlin: J. Cramer. Dean, W.E., J.V. Gardner, and D.Z. Piper 1997. Inorganic Geochemical Indicators of Glacial-Interglacial Changes in Productivity and Anoxia on the California Continental Margin. Geo- chemica et Cosmochemica Ada, 61:4507-4518. Delorme, L.D. 1989. Methods in Quaternary Ecology 7: Fresh-Water Ostracodes. Geo- science Canada, 16:85-90. Enzel, Y., D.R. Cayan, R.Y. Anderson, and S.G. Wells 1989. Atmospheric Circulation during Holocene Lake Stands in the Mo- jave Desert: Evidence for Regional Climate Change. Nature, 341: 44-47. Font, K.R. 1995. Geochemical and Isotopic Evidence for Hydrologic Processes at Owens Lake, California. 219 pages. Master's thesis, University of Nevada, Reno. Forester, R.M. 1983. Relationship of Two Lacustrine Ostracode Species to Solute Com- position and Salinity: Implications for Paleohydrochemistry. Geol- ogy, 11:435^*38. 1985. Limnocythere bradburyi n. sp.: A Modern Ostracode from Central Mexico and a Possible Quaternary Paleoclimatic Indicator. Journal of Paleontology, 59:8-20. 1986. Determination of the Dissolved Anion Composition of Ancient Lakes from Fossil Ostracodes. Geology, 14:796-798. 1987. Late Quaternary Paleoclimate Records from Lacustrine Ostracodes. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, K-3:261-276. Boulder, Colorado: Geological Soci- ety of America. 1991. Ostracode Assemblages from Springs in the Western United States: Implications for Paleohydrology. Memoirs of the Entomological So- ciety of Canada, 155:181-201. Forester, R.M., S.M. Colman, R.L. Reynolds, and L.D. Keigwin 1994. Lake Michigan's Late Quaternary Limnological and Climate His- tory from Ostracode, Oxygen Isotope, and Magnetic Susceptibility Records. Journal of Great Lakes Research, 20:93-107. Friedman, I., G.I. Smith, and K.G. Hardcastle 1976. Studies of Quaternary Saline Lakes, II: Isotopic and Compositional Changes during Desiccation of the Brines in Owens Lake, Califor- nia, 1969-1971. Geochemica and Cosmochemica Acta, 40:501-511. Fritz, S.C., S. Juggins, and R.W. Battarbee 1993. Diatom Assemblages and Ionic Characterization of Lakes of the Northern Great Plains, North America: A Tool for Reconstructing Past Salinity and Climate Fluctuations. Canadian Journal of Fisher- ies and Aquatic Sciences, 50:1844-1856. Gale, H.S. 1914. Salines in the Owens, Searles, and Panamint Basins, Southeastern California. Bulletin of the United States Geological Survey, 580-L: 251-323. Gardner, J.V., W.E. Dean, and P. Dartnell 1997. Biogenic Sedimentation beneath the California Current System for the Past 30 kyr and Its Paleoceanographic Significance. Paleocean- ography, 12:207-225. Gardner, J.V., L.E. Heusser, P.J. Quinterno, S.M. Stone, J.A. Barron, and R.Z. Poore 1988. Clear Lake Record vs. the Adjacent Marine Record: A Correlation of Their Past 20,000 Years of Paleoclimate and Paleoceanographic Responses. In J.D. Sims, editor, Late Quaternary climate, Tec- tonism, and Sedimentation in Clear Lake, Northern California Coast Ranges. Special Papers of the Geological Society of America, 214: 171-182. Gasse, F. 1980. Les Diatomees Lacustres Plio-Pleistocenes du Gadeb (Ethiopie). Revue Algologique Memoire, special article, 3:1-249. Hakansson, H., and H. Kling 1990. The Current Status of Some Very Small Freshwater Diatoms of the Genera Stephanodiscus and Cyclostephanos. Diatom Research, 5: 273-287. Heinrich, H. 1988. Origin and Consequences of Cyclic Ice Rafting in the Northeast At- lantic Ocean during the Past 130,000 Years. Quaternary Research, 29:142-152. Hollermann, P. 1973. Some Aspects of the Geoecology of the Basin and Range Province (California Section). Arctic and Alpine Research, 5:A85-A98. Hollett, K.J., W.R. Danskin, W.F. McCaffrey, and C.L. Walti 1991. Geology and Water Resources of Owens Valley, California; Chapter B. United States Geological Survey Water-Supply Paper, 2370: 77 pages. Hostetler, S.W., and P.U. Clark 1997. Climate Controls of Western U.S. Glaciation at the Last Glacial Maximum. Quaternary Science Reviews, 16:505-511. Kennett J.P., and K. Venz 1995. Late Quaternary Climatically Related Planktonic Foraminiferal As- semblage Changes: Hole 893A, Santa Barbara Basin, California. In J.P. Kennett, J.G. Baldauf, and M. Lyle, editors, Proceedings of the Ocean Drilling Program, Scientific Results, 146(2):281-293. Kilham, S.S., and P. Kilham 1978. Natural Community Bioassays: Predictions of Results Based on Nu- trient Physiology and Competition. Verhandlungen des Internation- alen Vereins fur Limnologie, 20:68?74. Kilham, P., S.S. Kilham, and R.E. Hecky 1986. Hypothesized Resource Relationships among African Planktonic Diatoms. Limnology and Oceanography, 31:1169-1181. Kociolek, J.P., and D.B. Herbst 1992. Taxonomy and Distribution of Benthic Diatoms from Mono Lake, California, U.S.A. Transactions of the American Microscopical So- ciety, 111:338-355. Koehler, P.A. 1995. Late Pleistocene Mollusks from the Dolomite Site, Owens Lake Playa, California. The Veliger, 38:312-318. Koehler, P.A., and R.S. Anderson 1994. Full-Glacial Shoreline Vegetation during the Maximum Highstand at Owens Lake, California. Great Basin Naturalist, 54:142-149. 1995. Thirty Thousand Years of Vegetation Changes in the Alabama Hills, Owens Valley, California. Quaternary Research, 43:238-248. Latif, M., and T.P. Barnett 1994. Causes of Decadal Climate Variability over the North Pacific and North America. Science, 266:634-637. Lee, W.T. 1906. Geology and Water Resources of Owens Valley, California. United States Geological Survey Water Supply and Irrigation Paper, 181: 28 pages. Lewis, W.M., Jr. 1983. A Revised Classification of Lakes based on Mixing. Canadian Journal of Fisheries and Aquatic Sciences, 40:1779-1787. Litwin, R.J., D.R Adam, N.O. Frederiksen, and W.B. Woolfenden 1997. A Palynomorph Record of Owens Lake Sediments of Core OL-92. NUMBER 33 173 Preliminary Analyses. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Society of America, 317:127-142. Lopes, T.J. 1987. Hydrology and Water Budget of Owens Lake, California. 128 pages. Master's thesis, University of Nevada, Reno. Lowenstein, T.K., J. Li, C. Brown, S.M. Roberts, R.J. Spencer, W. Yang, T.-L. Ku, S. Luo, and R.M. Forester 1997. Death Valley Salt Core: 200,000 Year Paleoclimate Record from Sedimentary Structures, Saline Mineralogy, Fluid Inclusions in Ha- lite, and Ostracodes [Abstract], In Great Basin Aquatic System His- tory, Program, Abstracts and Field Guide, Smithsonian Institution. [Unpaginated.] Lund, S.P., M.S. Newton, D.E. Hammond, O.K. Davis, and J.P. Bradbury 1996. Late Quaternary Stratigraphy of Owens Lake, California. In L.V. Benson, editor, Proceedings of the Workshop "Ongoing Paleocli- matic Studies in the Northern Great Basin," Reno, Nevada, May, 1993. United States Geological Survey Circular, 1119:47-52. Mason, D.T. 1967. Limnology of Mono Lake, California. University of California Pub- lications in Zoology, 83: 110 pages. McLaughlin, R.B., and J.L. Stone 1986. Some Late Pleistocene Diatoms of the Kenai Peninsula, Alaska. Beihefte zur Nova Hedwigia, 82: 148 pages. Menking, K.M., J.L. Bischoff, J.A. Fitzpatrick, J.W. Burdett, and R.O. Rye 1997. Climatic/Hydrologic Oscillations since 155,000 yr B.P. at Owens Lake, California, Reflected in Abundance and Stable Isotope Com- position of Sediment Carbonate. Quaternary Research, 48:58-68. Peterson, C.G. 1996. Response of Benthic Algal Communities to Natural Physical Distur- bance. In R.J. Stevenson, M.L. Bothwell, and R.L. Lowe, editors, Algal Ecology?Freshwater Benthic Ecosystems, pages 375-402. San Diego: Academic Press. Phillips, F.M., M.G. Zreda, L.V. Benson, M.A. Plummer, D. Elmore, and P. Sharma 1996. Chronology for Fluctuations in Late Pleistocene Sierra Nevada Gla- ciers and Lakes. Science, 274:749-751. Smith, A.J., and R.M. Forester 1994. Estimating Past Precipitation and Temperature from Fossil Ostra- codes. In Proceedings of the American Nuclear Society, 5th Interna- tional High-Level Radioactive Waste Management Conference and Exposition, Las Vegas, Nevada, pages 2545-2552. American Soci- ety of Civil Engineers and American Nuclear Society. Smith, G.I. 1976a. Origin of Lithium and Other Components in the Searles Lake Evaporites. In J.D. Vine, editor, Lithium Resources and Require- ments by the Year 2000. United States Geological Survey Profes- sional Paper, 1005:92-103. 1976b. Paleoclimatic Record in the Upper Quaternary Sediments of Searles Lake, California, U.S.A. In S. Horie, editor, Paleolimnology of Lake Biwa and the Japanese Pleistocene, 4:577-604. Japan: Kyoto Uni- versity. 1979. Subsurface Stratigraphy and Geochemistry of Late Quaternary Evaporites, Searles Lake, California. United States Geological Sur- vey Professional Paper, 1043: 130 pages. 1997. Stratigraphy, Lithologies, and Sedimentary Structures of Owens Lake Core OL-92, Southeast California. In G.I Smith and J.L. Bis- choff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Society of America, 317:9-23. Smith, G.I., and J.L. Bischoff 1997. Core OL-92 from Owens Lake: Project Rationale, Geologic Setting, Drilling Procedures, and Summary. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Society of America, 317:1-8. Smith, G.I., I. Friedman, and R.J. McLaughlin 1987. Studies of Quaternary Saline Lakes, HI: Mineral, Chemical, and Iso- topic Evidence of Salt Solution and Crystallization Processes in Owens Lake, California, 1969-1971. Geochemica and Cosmochem- icaActa, 51:811-827. Smith, G.I., and W.P. Pratt 1957. Core Logs from Owens, China, Searles, and Panamint Basins, Cali- fornia. Bulletin of the United States Geological Survey, 1045-A: 62 pages. Smith, G.I., and F.A. Street-Perrot 1983. Chapter 10: Pluvial Lakes of the Western United States. In H.E Wright, editor, Late Quaternary Environments of the United States, pages 190-212. Minneapolis: University of Minnesota Press. Sorensen, S.K., and S.E. Schwarzbach 1991. Reconnaissance Investigation of Water Quality, Bottom Sediment, and Biota Associated with Irrigation Drainage in the Klamath Basin, California and Oregon, 1988-89. United States Geological Survey, Water-Resources Investigations Report, 90-4203: 64 pages. Squires, L.E., and S.R. Rushforth 1986. Winter Phytoplankton Communities of Utah Lake, Utah, USA. Hy- drobiologia, 131:235-248. Theriot, E.C., and E.F. Stoermer 1981. Some Aspects of Morphological Variation in Stephanodiscus niaga- rae (Bacillariophyceae). Journal of Phycology, 17:64-72. Thompson, R.S. 1990. Chapter 10: Late Quaternary Vegetation and Climate in the Great Basin. In J.L. Betancourt, T.L. Van Devender, and P.S. Martin, edi- tors, Packrat Middens, pages 200-239. Tucson: University of Ari- zona Press. Tyler, S.W., S. Kranz, M.B. Parlange, J. Albertson, G.G. Katul, G.F. Cochran, B.A. Lyles, and G. Holder 1997. Estimation of Groundwater Evaporation and Salt Flux from Owens Lake, California, USA. Journal of Hydrology, 200:110-135. Wilcox, L.V. 1946. Boron in the Los Angeles (Owens River) Aqueduct. U.S. Depart- ment of Agriculture, Bureau of Plant Industry, Soils, and Agricul- tural Engineering, Research Report, 78: 66 pages. Woodward, A. 1961. Camels and Surveyors in Death Valley: The Nevada-California Border Survey of 1861. 73 pages. Palm Desert, California: Desert Printers, Inc., and Death Valley '49ers, Inc. Woolfenden, W.B. 1996. Late Quaternary Vegetation History of the Southern Owens Valley Region, Inyo County, California. 347 pages. Doctoral dissertation, University of Arizona, Tuscon, Arizona. Biogeography and Timing of Evolutionary Events among Great Basin Fishes G.R. Smith, T.E. Bowling, K. W. Gobalet, T. Lugaski, D.K. Shiozawa, andR.P. Evans ABSTRACT Hubbs and Miller's Great Basin works are model studies of the evolution of isolated endemic aquatic organisms and of hydro- graphic history. They investigated spatial and temporal barriers that permitted evolutionary divergence of fishes. They also used fish evidence to identify the locations and possible timings of past aquatic connections among basins. Hubbs and Miller focused on morphological differences among fish populations in related basins and hypothesized that their evolutionary divergence occurred over only hundreds to thousands of years. Their conclu- sions depended upon the assumption that divergence began after the most recent time of homogeneous ancestral populations, which was assumed to be the time of closure of the most recent hydro- graphic connection. This approach and its key assumptions are standard in studies of evolutionary rates, but they require reexami- nation. Our data suggest that rates of morphological change in oligo- genic and ecophenotypic traits of Great Basin fishes may be rapid in the short term (thousands of years), but Miocene to Pleistocene fossils of western North American fishes indicate slow changes over millions of years. Tests of hypotheses about evolutionary rates require independent estimates of both time and degree of divergence. Our evolutionary-rate estimates are based upon regression of percent sequence divergence (of mitochondrial DNA) between pairs of lineages against millions of years of diver- G.R. Smith, Museum of Zoology, University of Michigan, Ann Arbor, Michigan 48109, USA. T.E. Dowling, Department of Zoology, Arizona State University, Tempe, Arizona 85287, USA. K. W. Gobalet, Depart- ment of Biology, California State University, Bakersfield, California 93311, USA. T. Lugaski, deceased, formerly of the Mackay School of Mines, University of Nevada, Reno, Nevada 89557, USA. D.K. Shiozawa, Department of Zoology, Brigham Young University, Provo, Utah 84602, USA. R.P. Evans, Department of Zoology, Brigham Young University, Provo, Utah 84602, USA. Reviewers: Robert Hershler, Division of Mollusks, National Museum of Natural History, Smithsonian Institution, Washington, D.C., 20560- 0118, USA; Edward O. Wiley III, Department of Ecology and Evolu- tionary Biology, Division of Biological Sciences, University of Kan- sas, Lawrence, Kansas 66045, USA; and one anonymous reviewer. gence of sister clades, estimated from dates of first appearances in the fossil record. The results indicate rates of molecular evolution in the range of approximately one-half percent (salmonids) to one percent (cyprinids and cyprinodonts) sequence divergence per one million years. The ages of some forms thought to represent ten thousand years of postpluvial evolution are more likely to be hun- dreds of thousands of years old, according to molecular/fossil esti- mates. Estimated divergence times of fishes, based upon DNA, provide unique data on the nature and timing of barriers and aquatic connections among basins. Great Basin fish faunas exhibit high endemism, but they have lower diversity than predicted by pluvial cycles of isolation and by rapid evolution in allopatry. We hypothesize that uplift causes high stream gradients and, therefore, decreased local habitat diversity because of rapid stream runoff. Small and isolated populations on the basin floors are prone to allopatric differentiation, but they also are prone to high extinction rates; therefore, isolation and stabiliz- ing selection in desert basins rarely produced new species that sur- vived extinction events. Dedication This paper is dedicated to Carl L. Hubbs, Robert R. Miller, and Dwight W. Taylor, whose studies of aquatic organisms in western North American desert basins set the twentieth century standard for aquatic biogeography and conservation. Introduction to Great Basin Ichthyology Carl L. Hubbs began his studies of fishes in the Great Basin in 1915 while on an expedition to the Bonneville Basin as a field assistant to J.O. Snyder (Miller and Shor, 1997). Prior to the Bonneville trip, Snyder had surveyed fishes in the Lahontan Basin and in the Oregon Lakes, Owens River, and Mohave River basins (Snyder, 1908, 1912, 1914, 1917a, 1917b, 1918, 1919, 1924). Hubbs continued to conduct exhaustive explora- tions of the Great Basin in 1934, 1938, and 1942; with his wife, Laura, and their children, Earl, Frances, and Clark, he collected and identified more than 118,000 specimens. Hubbs was joined by his student, Robert Miller, in 1938, and between them they 175 176 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES produced about three dozen papers on Great Basin fishes. Miller's efforts resulted in numerous papers on fishes of the Death Valley system (e.g., 1946a, 1946b, 1950, 1972) and on broader issues relating to evolution and conservation of Great Basin fishes (1959, 1961a, 1961b, 1965, 1981). Miller contin- ued conducting this field work with his wife, Frances Hubbs Miller, and their five children, and he extended the project to include a survey of the fishes of Mexico. The landmark synthesis of the systematics and biogeography of fishes of the Great Basin was published by Carl Hubbs and Robert Miller in 1948(b), which was expanded from a sympo- sium paper read in 1942. The monograph described the geo- graphical limits of more than 100 hydrographic units, the eleva- tions of ancient lake levels, and the biogeographic relationships of fishes in each basin. The authors' goal was to show how convincingly physiographic deductions are supported by knowledge of the distribution of fishes in the Great Basin. The Hubbs-Miller team and Elliot Blackwelder (1933) both used fish distributions to infer past aquatic connections, which were based upon the principles of Jordan (1905, chapters 16, 17). The interaction of time and barriers framed the questions in the Hubbs-Miller inventory of Great Basin fish evolution. Their descriptions and classification of those fishes were in- tended to answer questions about hydrographic connections among basins, barriers between basins, and rapidity of evolu- tion. Although they chose not to speculate about evolutionary mechanisms (1908b: 119), they offered frequent hypotheses about amounts and rates of evolution. They later expanded upon their ideas about fish speciation (Hubbs, 1940, 1941a, 1941b; Miller, 1946a, 1950). Hubbs and Miller's most exciting idea was declared in the introduction to their 1948 monograph (Hubbs and Miller, 1948b:24): Some of the isolation and speciation that is dated from Pleistocene time may be much more recent. Now disconnected waters within many of the basins of western North America may have been connected and their fish fauna made uniform not many centuries ago. From a consideration of the evidences on cli- matic changes ... it seems likely that such connections may have been estab- lished several times during the last three millennia. We believe, however, that most of the local differentiation and endemism among the fishes of the isolated desert waters dates from the end of the Glacial period. Hutchinson (1939) con- cluded that the similar differentiation of now isolated fish populations in Kash- mir and Indian Tibet dates from the period of maximum glaciation. ... In correlation with Post-pluvial isolation, a limited degree of speciation has oc- curred here and there, but data are too meager to warrant any attempt to use the rate of species evolution in dating the Pluvial waters. ... A considerable body of observations and mathematical data indicates that small, closely inbreeding populations may speciate at a much more rapid rate than was previously thought possible (Wright, 1940a, 1940b, 1942; Huxley, 1943). The authors treated their geological and ichthyological data in two ways: they used geological data to infer rates offish dif- ferentiation, and they used degree of morphological divergence to infer the times of geological barriers. The usual assumption was that fish populations were uniform at the end of the most recent connection between aquatic habitats. This led to the de- duction that rapid speciation occurred in the postpluvial inter- val of isolation, about 11,000 years. Carl Hubbs, Robert Miller, and Laura C. Hubbs (1974) ex- panded upon their previous work and the work of others in their subsequent monograph on hydrographic history and relict fishes of the North-Central Great Basin. Laura Hubbs mea- sured hundreds of fish specimens from different basins and cal- culated the mean and standard deviation for the populations in each basin, providing replicated comparisons of times and amounts of divergence. Hubbs and Miller (1948a, 1972) de- scribed many additional localized cyprinid species and subspe- cies, which provided further evidence for postpluvial diver- gence in geographic isolation (see Appendix). Hubbs and Miller sought to answer a question that still moti- vates biogeographers (e.g., Ricklefs and Schluter, 1993); that is, What are the appropriate scales of the spatial and temporal interactions that bring about species production? Hubbs and Miller (1948b) emphasized differences among and within ba- sins, namely, different taxonomic levels, numbers of endemic forms, and local diversity in relation to regional diversity as in- dicators of evolutionary history: "The Lahontan fauna ... com- prises, we estimate, about 30 forms, as compared with perhaps 28 in the Bonneville and about 24 in the Death Valley system. The relatively high number of total forms is correlated with the isolation of local populations, for the Lahontan system has be- come greatly disrupted since Pluvial times" (Hubbs and Miller, 1948b:32). Hubbs and Miller (1948b:41) also recognized an important role for hybridization in evolution: "Since the waters have dwindled, and the lake and creek types have been forced into cohabitation, the two have hybridized very extensively. We have referred to this specific breakdown in connection with our analysis of a similar case of hybridization in the Mohave Desert (Hubbs and Miller, 1943)." The effects of hybridization in reducing diversity and erasing evidence of evolutionary his- tory are peculiarly prominent among fishes in the western United States (Smith, 1992a; Dowling and Secor, 1997). Our goal is to continue the comparison of fish and hydro- graphic information so as to answer questions about timing and origins of fish diversity. Evidence presented herein for timing of fish divergence comes from the fossil record (G.R. Smith, 1981a, 1981b; Smith et al., 1982) and from molecular studies (Dowling et al., 2002; Shiozawa and Evans, unpublished). In this paper, we summarize gene sequence divergence data in the context of the fossil record of Great Basin fishes. We seek to shed light on the rates of evolutionary changes in Great Basin drainages in the spirit of the other Great Basin naturalists of 50 years ago (Hall, 1946:59-67; Durrant, 1952:490-518; LaRiv- ers, 1962; and the vast body of literature published by Hubbs and Miller). Our understanding of freshwater fish distributions benefits from the fact that freshwater fish are more geographically con- strained than most other organisms (Cope, 1868; Jordan, 1905). The principles thought to influence distribution and evolution NUMBER 33 177 of fishes have remained rather constant during the last 130 years, but estimates of timing of geological and evolutionary events vary by five orders of magnitude (compare Cope, 1868, with Hubbs et al., 1974). The Great Basin offers a unique op- portunity to study the time scales of evolutionary processes be- cause of its numerous, isolated basins that have replicated fish habitats and histories. In addition, well-documented evidence exists for environmental gradients, pluvial cycles, and ages of sediments that preserve fossil ancestors of Holocene fishes (e.g., Morrison, 1965, 1991; Benson and Thompson, 1987; Benson et al., 1990; Oviatt et al., 1992; Reheis and Morrison, 1997; Smith and Bischoff, 1997). These factors provide the ba- sis for Cope's (1883) and Hubbs and Miller's (1948b) assump- tion that fish evidence can contribute much to the study of hy- drographic history. Our attempts to fulfill this promise have been guided further by Dwight Taylor's methods and his prin- ciple (1960) that hydrologic, topographic, and climatic features change more rapidly than do lineages of organisms. The Late Cenozoic fossil record in the Great Basin reflects the climatic and faunal revolution between the Paleogene and the Neogene. For example, the temperate Miocene to Holocene fish faunas of the Great Basin (as now known) share only one family (Catostomidae) with the subtropical faunas of the west- ern North American Eocene in Wyoming, Utah, and Colorado (Grande, 1984). Transitional Great Basin fish records are Cope's (1872) Amyzon mentalis and Trichophanes hians from the Osino, Nevada, Oligocene coal beds and Early Tertiary fos- sils similar to Diplomystus in Elko County, Nevada. Our dis- cussion herein, however, will be limited to the late Cenozoic record. Publications on fossil Empetrichthys, fundulids, and cyprinodontids by Miller (1945b) and Uyeno and Miller (1962) included reviews of many pre-Pleistocene records of modern lineages in the southwestern Great Basin. LaRivers (1962) compiled the first account of endemic and introduced fishes in Nevada, including what was known at that time about ecology and fossil fishes. Newberry (1871, and references therein) was the first to study Great Basin fishes from a geological context, but he was soon followed by Jordan (1878a, 1878b, 1891), Cope (1883), Jordan and Evermann (1896), and Tanner (1936). J.O. Snyder and Hubbs and Miller, in works cited above, were the next to dominate the field. Uyeno and Miller (1962), Miller (1981), M.L. Smith (1981), and G.R. Smith (1966, 1978, 1981a, 1981b) discussed fossils and described patterns of differentia- tion, species density, and zoogeography of western North American fishes among basins. Minckley et al. (1986) dis- cussed this history in detail, then compiled current geological literature on the Cenozoic tectonics of western North America and hypothesized broad aspects of zoogeographic history of western fishes. They looked for congruence between area cla- dograms and fish cladograms then available and concluded that the modern fauna occupied western tectonic units in the Oli- gocene before they were separated from Atlantic Ocean tribu- taries by the Miocene cycle of tectonic uplift. Minckley and Deacon (1991) described critical conservation issues and ex- tinction threats to the modern fauna. Behnke (1992) provided original analysis and summarized the available information on recent trouts. Sigler and Sigler (1987, 1996) summarized the modern fauna, including introduced fishes. The fishes discussed in the present paper inhabit or inhabited the hydrographic Great Basin, as defined, for example, by Hubbs and Miller (1948b) as being the region of Holocene in- ternal drainage bounded by barriers shared with the present drainages of the Colorado, Sacramento, Klamath, and Colum- bia Rivers (Figure 1). The Great Basin is not static; rather, it is a mosaic whose boundaries and external connections have shifted with tectonic, volcanic, and climatic changes across time scales ranging from 10 thousand years (Ky) to 10 million years (My) during the last 17 My (Minckley et al., 1986; and citations below). Many hydrographic connections have been hypothesized. For example, in the middle Miocene, there were large lakes in northern Nevada and Utah that drained through the Snake River Plain southwest across the Sierra Nevada to the Pacific coast drainage (Stokes, 1979). Later, the Lahontan Basin (which still drained to the Pacific Ocean) was connected to the Snake River Plain. At different times, the White River (Hubbs and Miller, 1948b), Las Vegas Creek (Haynes, 1967), Death Valley-Owens River systems (Miller, 1946a, 1948), and perhaps even the Bonneville Basin (Ives, 1948) and upper Snake River were variously connected to the Bouse Embay- ment of the lower Colorado River. As the Sierra Nevada was uplifted, it is possible that a north-to-south drainage system in- cluded basins in Oregon and possibly Washington (Smith et al., 2000). There is good evidence that 2.5 million years ago (Ma) the Snake River's course changed with its capture by the Co- lumbia River through Hell's Canyon (Wheeler and Cook, 1954). At undetermined times in the late Cenozoic, there were connections from the Mono Lake basin to the Lahontan Basin, from Owens River west to the Pacific coast drainage, and later, south to Mexico. In addition, Death Valley was connected to the lower Colorado River as well as to the Basin and Range Province in Mexico (Minckley et al., 1986; Hale, 1985). Fif- teen to 20 thousand years ago (Ka) the basins of internal drain- age were probably restricted mainly to northern and eastern Nevada and sometimes to the Death Valley system. At about that time, the Bonneville Basin and the basins of Catlow and Fort Rock Valleys and of Harney and Malheur Lakes (hereaf- ter, Harney-Malheur) were sometimes tributaries to the Snake River (Taylor and Bright, 1987; Bisson and Bond, 1971; Minckley et al., 1986), whereas the White River was occasion- ally connected to the lower Colorado River (Taylor, 1983). The sizes and connections of basins fluctuated with climatic mois- ture through the late Pliocene (Smith et al., 1982) and the Pleis- tocene (Morrison, 1991; DiGuiseppi and Bartley, 1991). The hypotheses presented in this paper do not assume the above hydrographic outline, but they suggest that connections and disconnections, such as outlined above, were more frequent than commonly assumed. Our approach is to use fish fossils, 178 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 37' 121* 119' 117* 115 ? Miocene A. Pliocene # Pleistocene 1 ' i -35" FIGURE 1.?Fossil fish localities in the Great Basin and adjacent drainages. Ages of fossil fish are indicated by symbols defined on the figure. Fossils are listed in the Appendix, where they are indicated by a cross. Heavy lines demarcate major drainage divides. phylogenetic analysis, biogeography, and genetic distances as primary evidence for the times and locations of hydrographic change. Material and Methods Our hypothesized drainage connections are based upon dis- tribution ranges of fish clades, species, or genetically similar populations, especially species and subspecies that span drain- age divides that are no longer passable. We assume that ancient fishes dispersed only through their habitats, because this method is supported by testable generalizations, in contrast to ad hoc hypotheses for dispersal offish eggs, such as on birds' feet (e.g., Brown and Rosen, 1995; Spencer and Patchett, 1997). Likely geological mechanisms for transfer include lake spillover (Hubbs and Miller, 1948b), river capture, erosional beheading of streams, diversion of channels by erosion, and di- version of drainage by uplift or volcanic eruptions (Bishop, 1995). The degree of isolation of basins constrains dispersal, which leads to predictions that are testable by geographic, geo- NUMBER 33 179 logic, genetic, and cladistic analyses. The large number of adja- cent basins with similar habitats permits analysis of replicated histories of similar local faunas in addition to isolated case his- tories. It is important that we attempt to use fish data that are independent of geological assumptions, in the spirit of Hubbs and Miller (1948b) and Taylor (1960), so as to contribute non- circular conclusions useful to geology. Records of Holocene fishes from the Great Basin were com- piled primarily from the records of Hubbs (and his students and successors), which have been deposited in the Division of Fishes, Museum of Zoology, University of Michigan, and from the literature. Late Cenozoic fossil fish data (Figure 1; Table 1) were compiled from collections in the Museum of Paleontol- ogy and the Museum of Zoology, University of Michigan; the W.M. Keck Museum, Mackay School of Mines, University of Nevada, Reno; the Museum of Vertebrate Paleontology, Uni- versity of California at Berkeley; the Department of Biology, California State University at Bakersfield; and the paleontolog- ical literature. (Lists of comparative material are available from authors G.R.S. and K.W.G.) Records of new distributions are noted in the species accounts in the Appendix. These accounts include information on specimen characteristics, collecting lo- calities, and names of the collectors when the specimens were not deposited at the above-listed institutions (Appendix; Table 5). Ancestral-descendent pairs are recognized by shared apo- morphies that are not shared with the next closest sister group. The probability of discovering ancestral-descendant pairs is thought to be low in marine and terrestrial systems (Foote, 1996), but in the Great Basin it is greatly enhanced by short geological time frame, highly constrained geography, the small number of taxa, and the salient osteological characteristics among fish taxa. Cladistic methods used to estimate branching sequences for cyprinid fishes are from Dowling et al. (2002), Simons and Mayden (1997, 1998), and Cavender and Coburn (1992); for Catostomidae, Smith (1992b); for Prosopium, Smith and Todd (1993); for trout, Shiozawa and Evans (unpublished); and for Cyprinodontidae, Echelle and Dowling (1992). Molecular methods follow those used by Echelle and Dowling (1992), Shiozawa and Evans (1995), and Dowling and Naylor (1997). Cladograms were calculated from mitochondrial DNA (mtDNA) sequence data, using the parsimony algorithm "Phy- logenetic Analysis Using Parsimony" (PAUP*; Swofford, 1990), version 4.0 d64 by D. Swofford. Informative characters among the species of Cyprinidae are the shared base pairs of the mitochondrial cytochrome b gene, relative to outgroups. North American and Eurasian cyprinids were used as out- groups for the analysis of western North American Cyprinidae. Informative cladistic characters among Oncorhynchus clarki and among Cyprinodon are shared restriction-site differences in mtDNA, relative to outgroups. Oncorhynchus mykiss (rain- bow trout) is the outgroup for the cladistic analysis of subspe- cies of Oncorhynchus clarki (cutthroat trout). Cyprinodon mac- ularius and several species of southwestern Cyprinodon from outside the Great Basin served as outgroups for the cladistic analysis of Cyprinodon species of the Death Valley system. DATA USED TO ESTIMATE THE RATE OF EVOLUTIONARY CHANGE PER UNIT TIME.?The relative homogeneity of molec- ular distances (Tables 2-4) among species with a common an- cestor suggested a source of independent evidence to break the logical circularity inherent in using geological events to date evolution and in using evolutionary differentiation to date geo- logical events. Previous workers carefully contrasted geologi- cal and ichthyological data, but it has been difficult to find in- dependent data to constrain the time of geographic isolation. The traditional method (Hutchinson, 1939) defines a relevant interval of evolutionary time as beginning with the last possible period of hydrographic connection. This method assumes that one panmictic and homogeneous form at the time of connec- tion evolved into two isolated and differentiated forms in the present. For example, it is assumed that a uniform Pliocene population before the Panamanian land bridge evolved into Pa- cific Ocean and Atlantic Ocean populations isolated by the Panamanian isthmus during the last 3.5 My (Bermingham et al., 1997). This method is limited to investigation of only one possible time for the origin of divergence, with bias toward the most recent. It has been deemed parsimonious to choose the time of most recent connection as the beginning of the divergence because to choose an earlier time seemed to add an unnecessary ad hoc hypothesis, i.e., more time than is minimally necessary. Dis- covery of the amount of time required for the observed diver- gence is, however, the original question, and to assume the shortest of several possible times does not test the hypothesis. It is important to test the alternative possibility, namely, that multiple populations were present prior to the time of last con- nection and that the forms observed in the Holocene have been accumulating genetic differences for a time period longer than since the last connection. This alternative is important for stud- ies such as the Panamanian separation of Atlantic and Pacific marine faunas, as well as for freshwater fishes in the Great Ba- sin. Divergence times of clades must ultimately be based upon ages determined from the stratigraphic context of fossils?usu- ally some combination of mammalian biostratigraphy (Wood- burne, 1987), potassium-argon, 40argon-39argon, or other ra- diogenic dates (e.g., Baksi et al.,1992), magnetostratigraphy (Negrini et al., 1987), fission-track ages (Kimmel, 1982), and field studies of geomorphology and geochemistry (e.g., Alli- son, 1940, 1982; Hubbs and Miller, 1948b; Bright, 1966, 1967; Benson and Thompson, 1987; Benson et al., 1990; Curry, 1990; Morrison, 1991; Reheis and Morrison, 1997). We compared data on divergence of mtDNA molecules among fishes (Kocher and Carleton, 1997) with dates of rele- vant fossils to calculate rates of molecular evolution (Collins et al., 1996). Molecular differences among homologous mito- chondrial genes in these taxa are expressed as pairwise percent sequence divergence (Gillespie, 1991). The difference values are proportions of the specified part of the genome that differ in 180 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 1.?Late Cenozoic fossil fish localities in the Great Basin. Numbers refer to indexed taxa mentioned in the Appendix; a question mark (?) = uncertain identification or age. Locality, by state Nevada Bullfrog Mine near Beatty, Nye County, 23 Ma? Esmeralda, Esmeralda County Siebert Tuff, Nye County, 17 Ma Buffalo Canyon, Churchill County, 15 Ma Middlegate Formation, Churchill County Stewart Valley, Mineral County, 12 Ma Virgin Valley, Humboldt County Truckee Formation, Churchill County Red Rock locality, Churchill County Rabbit Hole, Pershing County Sahwave Mountains, Pershing County Humboldt Formation, Elko County, 9.5 Ma Coal Valley, Mineral County Alum Mine, Nye County Jersey Valley, Pershing County Brady Pocket, Churchill County, Hemphillian? Carson Valley, Douglas County Oregon Trout Creek, Harney County Utah Cache Valley Formation, Salt Lake group California Secret Valley, Lassen County Madeline Plains, Lassen County White Hills, Airport Lake, Inyo County Mono Basin Honey Lake Idaho Salt Lake group at Georgetown Nevada Carson Valley, Douglas County, early Blancan Carson Valley, Douglas County, late Blancan Pine Nut Ridge, Douglas County Duck Valley, Washoe County Carson Sink, Churchill County Mopung Hills, Churchill County Fish Cave, Churchill County Oregon Fort Rock Lake, Lake County, Blancan Utah Junction Hills, Salt Lake group, Hemphillian Park Valley, Box Elder County, Hemphillian? California Manix Lake, San Bernardino County Searles Lake, San Bernardino County China Lake, San Bernardino County Lake Mohave or Soda Lake, San Bernardino County Owens Lake Core, Inyo County Nevada Humboldt River Canyon, Pershing County near Fallon, Churchill County Carson Valley, Irvingtonian Sunshine Amphitheater, northwestern Walker Lake, pre-Lahontan Lahontan Cave deposits Smith Creek Cave Indexed taxa MIOCENE 117 5,27,28,58,88, 120 122 6,88 88 27,88 6,58, 124 8,29,80,89, 128 87 88 116, 128 123,130 6,88 88 8 56 15,30,88 129 3 PLIOCENE 1,30,69,91, 124 1,25,267,30,91 12,62 1,13, 14,30,63,74 9,25,60,65,68,70,91 3,59 2, 15,30,64,75 2, 15,64,92 6 10,62 92 9,72,93, 124 124 7, 59, 62, 90 3,23,26,30,57,61,69,81 125 PLEISTOCENE 12,129 12 12,62 12 13,73,82,92 10,92 60,72 15,30,64 72,92 124 23,99 Source1 Lugaski, UNRM LaRivers, 1962, UNRM Lugaski, 1978, UNRM T. Lugaski, UNRM D.I. Axelrod, UCD Lugaski, UNRM Bell, UMMP J.R. Alcorn, UMMP LaRivers, UNRM Lugaski, UNRM T. Kelly, MVP LaRivers, UNRM Lugaski, UNRM Charles Dailey, MVP, UMMP C. Arnold, UMMP J.S. Williams, USU UMMZ, Wagner et al., 1997, MVP Wagner etal., 1993, MVP Gobalet; UMMP Miller and Smith, 1981; Gobalet; UMMP Taylor and Smith, 1981, UMMP S.A. Bilby, ISU T. Kelly, Gobalet, MVP T. Kelly, Gobalet, MVP T.Kelly, 1994, MVP Taylor and Smith, 1981, UMMP R. Miller, UMMP Allison and Bond, 1983, OSU P. McClellan, MVP T. Cavender, UMMP Buwalda, 1914; Reynolds, 1991, 1994 UMMP Leatham, UMMP G.I. Smith, UMMP Gobalet, MVP J.R. Alcorn, UMMP T. Kelly, Gobalet, MVP M. Reheis, USGS Jordan, 1924, MVP Mead etal., 1982 NUMBER 33 181 TABLE 1.?Continued. Locality, by state Oregon Chewaucan, Lake County Utah Hot Springs, Salt Lake County Black Rock, Salt Lake County Black Rock, Tooele County Homestead Cave Thatcher Basin Old River Bed Nevada Nevada AEC Test Site, Nye County, Miocene? California Indexed taxa 11 23, 66, 84, 85, 125, 126 23,84,85,98, 125, 126 23 23, 66, 84, 85, 87?, 98, 126 61 62, 82, 98 STRATIGRAPHY UNCERTAIN 119 Titus Canyon, Pliocene? (stratigraphy uncertain) 111, 117 Titus Canyon Formation, Oligocene? Furnace Creek, Death Valley, Pliocene? 117 118 Source' Gobalet and Negrini, 1992 Smith etal., 1968, UMMP Smith etal., 1968, UMMP J.H. Madsen, Jr., UMMP Broughton, UU Bright, UMMP Oviatt, UU Richard Wyman, UMMP Miller, 1945b Lugaski, UNRM Miller, 1945b 'Published source and/or collector and location of specimens. Abbreviations: ISU = Idaho State University Museum, Pocatello; MVP = Museum of Vertebrate Paleontology, University of California, Berkeley; OSU = Oregon State University, Corvallis; UCD = University of California, Davis; UMMP = University of Michigan Museum of Paleontology, Ann Arbor; UMMZ = University of Michigan Museum of Zoology; UNRM = Mackay School of Mines, University of Nevada, Reno; USGS = United States Geological Survey; USU = Department of Geology, Utah State University; UU = University of Utah. accumulated mutations, corrected for back mutation (Tamura and Nei, 1993). We were encouraged to pursue this method by the homogeneity of values of pairwise comparisons of sets of lineages, each derived from a common ancestor (Tables 2-4). This homogeneity suggested that the rate of molecular evolu- tion was roughly constant for a given gene system in a given clade. The relative-rate test of Sarich and Wilson (1967) was used to evaluate the precision of the time estimate. It is based upon the assumption that all taxa derived from one node (branch point) on a cladogram have evolved while accumulat- ing mostly neutral mutations for the same amount of time (e.g., Figure 2; Table 2), regardless of the number of intervening branch points; therefore, all the distances (in a homologous molecule) from the sister clade to all the members of the clade in question should be identical. This degree of similarity is a measure of the constancy of the rate of molecular change and, therefore, an estimate of the precision of a molecular estima- tion of geological time. In other words, the homogeneity of the measured distances between sister branches within a clade is an indication of how constant the rate of molecular change has been. It is important to stress that these distance values were tested in conjunction with ages of fossil fish that we interpret as ancestors; distance values were not converted to rates of mo- lecular change based upon molecular clock assumptions from mammals or other organisms. If we accept the accumulated differences (genetic distance) as numerators in a rate estimate for these fishes, we then need an independent denominator measured in millions of years to yield the amount of molecular change per one million years (Collins et al., 1996). The denominator, evolutionary time, is estimated herein as millions of years since the time from first appearance of a fish lineage to the present. Relative times of lineage origins, recognized by possession (in fossils) of evolu- tionarily derived states that diagnose appropriate groups (Lundberg, 1993), are underestimated because occurrences of fossils are usually minimum estimates of the age of their lin- eage (Minckley et al., 1986). This problem exists because the first members of a lineage are not only rare, they are also not easily recognized (assuming that apomorphies accumulate gradually at many points along branches between nodes). Where fossils are limited to infrequently preserved deposi- tional environments, the sample sizes are not adequate to eval- uate statistical error around the estimated age of the first occur- rence of a lineage. The denominator error from underestimated times of first appearance results in overestimated rates of change (rate = percent divergence/time). Errors in the numera- tors, e.g., from back mutations, yield underestimated rates, un- less genetic distances are adequately corrected (Tamura and Nei, 1993). The potential consequences of these errors are readily apparent from inspection of Figure 14, in which numer- ator error is corrected, but denominator error is not; therefore, our rates of molecular change (0.5%-1%/My) are likely to be overestimates, approximating upper bounds. If the oldest of these lineages are as old as the Oligocene, as suggested by Minckley et al. (1986), correct rate estimates could be 0.3%- 0.7%/1 My. The pairwise difference values for lineages in a clade, such as the Cyprinodontidae, Salmonidae, or Cyprinidae, are re- gressed against geological divergence-time estimates (Figure 14). The regression coefficient then gives the percent sequence 182 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES *5 11 Qi o c ?? ?s ? 2 s?z 3 c A .a roro in ro in ro in cin ro oo o CO ^N,ro BOO ro _i Oro *0fN ?fN tN 00Tf fN O Tf tN . TftN OTf tN ft roCN en rotN Cro tN CuTf CN c? TffN TffN Tf fN ? fN ON O ON Tf in Cu Taxon or population t Phoxinus neogaeus cu ooON ?II >? d fN" fN ?II ?q in tN IIi> o 3 H0"3 ca E CU Q ?a "o ster "to 1 TfCN Mylopharodon conocephalus in ^^ 00 ? inCN Lavinia exilicauda ? ( ro ? 00 ?Tf vOCN Hesperoleucus symmetricus Tf ^^ VO ll ^> d Tf" CN? II ?qtn ONTf ||u G.3 ?5c ca ro? voro ON Tf CO? ?ro CN Siphateles bicolor, NV ON ON II *?fN SII ? t?H ,_i .. II CN i fN? ? ?P- ON? OtN ?ro Tf 00 00 Tf in ?Tf fN Gila seminuda m fN i tN ON? fNON ? ino ONON ? VOCN fNfN ? Ti- roro Tf roCN Gila purpurea Cu fN Tf tN II II > d ON" ?|| ?q ? 00 p- ?|| o 3 Tf fN| gTf CN Q cjT3 cao \-o .22"to 1 ? ? CNON ?ON 00? in? CN ? ?? ro? 00 oo vo? fNON Gila atraria, N. Bonneville e tN i ON? ? ONVO ? ON OO? VO 00? ino VO 00? ro? ? ON ON 00 *~ CNON Gila atraria, Sevier R. CO tN ( Tf fN? Tf fN? *O? VOr- ?ro P-? sCN ? inON p-?^ ON fN ? CN Gila atraria, Bear R. x> CN i Tf tN? f- i?i ?,?, i?i ? in ? 00 00? in 00? CN in? ?p- ? CN? ? in ON? OO? CN Gila cypha o fN i s Tf fN? ro ?p- ?? ro ? 00 00? in 00? ON s?? p?ON ?? ? ON _ ?CN Gilajordani ?, CN i inr- ? inr- ? fNP- ? fNr? ?Tf ? tN ?H [-^ ^M ro ?CN roTf CNTf Tf 00 VOVO 00ro CN Gila orcutti o CN '?' fN II >d ON" ? || C/3 o"f- ?|| r- cato '?Bc cag -2*-~ *0;_ ?^o o 4) ca ~ ?H Lo 1 *-* "" roVO ? p-~VO ? VOin ? 00 VO? invo ? VO 00? ON? 00 00? ?CN ? Tf fN ONro in Tf in Tf roCN Gila nigrescens i co? in 00? ON ? in? ? ? 00 Tf? 00 Tf? p^VO ? VOVO ?co VO? in? 00 00? ro? Tf ? Tf ?CN p- ?CN Gila pandora so -o CN CN in ?fN II. d 910 ll ?q oo" p- ?II c3 V) J- CB -2 Vjo tu lado Ui "3 1 VOVO ? in 00? fNON ? VOvo ? ONVO ? fN VO? CNNO ?ro vo? ?00 ?r- ON? 00 ON? VO? s ? 00 ? fN ?ro CNro CN Acrocheilus alutaceus CN i Tf VO? VO? Tf 00? ^^^"* VOVO ? CNr- ?(?. in? invo ?in VO? ON ? CNON ? 00 00? p~? 2 TfON VO? ON VO ? VO ON Klamathella coerulea CN fN ON tN|| > d fN ?II -o CN" t? ?II CD s idis ca g infN 1g fN i ?" ? in ?oo VO? Tf oo? in ON? r--VO ? ( .? fNVD ?Tf VO? VO? roON ? ON ON? VO? ? ? in? in ?~ <^ P^ 00 CN 00 00 ?< Ptychocheilus lucius 301 i 00ON ? ON ON? CNON ? VOON ? CN ?? ^^ ro Tf ON? roON ?ro 00? 00 ON? ONo CN ^^ p- ^-? i?i VO ? Tf ?i CN"?' P- tN NO P^ VO ro CNCN Ptychocheilus grandis ? ro ^*00 II o** rsJo II ,s.dp- fN II O 3 ?5 g 301:| gTf tN 1 ?ro ??i Tf Tf roro ? in Tf ^^ r-^" VOin ,?, in ^? in roTf CN in ~* ON Tf ONin ^~ Tf ro 00ro in ro00 ?" inro 00 Tf ? in Tf ONro ON inCN Eremichthys acros en 1 00 Tf ?ro inCN roro in ^^??' Tf ^^ ONro voro inCN ONfN ro ^?Tf fN?< ON tN 00 tN Tf Tf Tf Tf ? Tf P?in ^^ in Tf ro00 VO p~ inCN Relictus solitarius ro ? 00 ~? II >'d ? CN ? II "1 II O 1"O cea CL> E 1 CN^ in?? ON Tf VO? vo? P-- ON? ON ON? ON ?ro inro ,?,ON ? ON? Tf ON? Tf ?Tf ? 00? Tf CN 00 CN ON? Tf ?? 00 ? ? ONin ON i?i fN Ptychocheilus oregonensis 30o VO VD'?' II > ro ^>II ?q 87, < o ?3 ?5cca E ? p- ?ro CN ON CNCN ,? 00 ? ONON ro00 00 00 ? TfON iCN ?CN ,?,VO ,?,VO 00VO p- NO Tf in? roON ro? CN Tf ?tN p^? in 00 -" NO 00 in ON in fN in fNfN 00 ro i?i CN Orthodon microlepidotus O in ro roro ro ^m enro 00 CN ON tN roro ^_ en , 6 092s in Tf ? p-fN ^_ fN fN p- ro inro ON roro P- 00 CN ON? cos rooo inro ON inro ONro ro invo roON roro ^^Tf rooo Tf roON inro p-m roP- roro ?VO roro roro ?Tf co CNro r?1 in ro roCN in ro P- roro TfTf ro Tf en NOro Lepidomeda albivalis en i in s 7 090(S VOTf? CN VO (?^ VOro Tf ro ^H roro ^H 00 CN VOON CN ONON CN rom roTf Tfro Tfro roON inro roro ro 00CN co inro rnfN fNVO roCN Tfco p-in ONCN ? ON ?ON ON Tf fN 1?1 tNro ? roTf CNco 00 VOro ro? roro 00 CN in? rop- ?ro p-ON fN Tf roro i?i tNro p^ I?. CN ON fN i?i 00 tN in 00 tN VO 00 fN ON i?i roVO inro 00 Tfro Tf ONCN Tf Tf ro p- ON f?00 CN CNfN ro ONfN ro CN00 fN CO CNro Pogonichthys macrolepidotus Qcu 283 roin CN p-- 00CN ONfN rom Tf CN ONCO ON ? VOCN f^ ro ro vo 141 ON roON Tfro ?VO CO ON Tfro in p^ro 00 inro ro00 roro inro in 00ro ooin ro? O\ro ooCN Tf 00p- roro 00 CO ro ro.?, ?Tf rop- rop^ p^ro ?P- ro00 p**ro ?OO ro? Tf roVO CNTf roON VO CO ONVO ro ONTf roON i-H Tf Rhinichthys osculus ? fN Tf 216 316 VO? ro roro rofN ro VO? Tf in 00 ro? ? ON ro ?vo ?p- mCN Tf CN Tfp- ON 103 Tf in roro CNro Tf VO? Tf CN? Tf i?i ONro ,_00 en 00 inro ^H P--en CNCN Tf roON enVO ?Tf ro Tf p- *?i Tf ^^ 00 v-4 Tf ON Tf ?CN Tf ?Tf fNTf TfVO ro? voro s 194 Tf 00 Tf VO 00ro Rhinichthys atratulus ea Tf Tfin Tf ?ro ro ON ro in CO rovo roro p-ro roCN Tf ^H ^H Tf CN OO in 3 ? CN Tf in 00 ro00 CNro ^ roin i?i Tf Tfen 00ro to ? ?o ON 00ro 00 ?? ro ON? Tf roVO Tf O\ CO ro ro? inro fNvo roTf roro ONin roCN 1?ro 00ro Tfro Tfro rovo in ?? VOro ON 813 en Tfvo ro O ?o Ctenopharyngodon idella U NUMBER 33 183 i i i i i i r 35 30 25 20 15 10 5 % sequence divergence Phoxinus neogaeus, Eastern N.A. Mylopharodon, Sacramento Dr. Lavinia, Sacramento Dr. Hesperoleucus, Sacramento Dr. Siphateles obesus, Lahontan Dr. Siphateles o. pectinifer, Pyramid L. Gila (Moapa) coriacea, White Dr. Gila robusta, Colorado Dr. Gila elegans, Colorado Dr. Gila seminuda, Virgin Dr. Gila purpurea, RioYaqui Gila atraria, Bonneville-Snake Dr. Gila cypha, Colorado Dr. Gilajordani, White Dr. Gila orcutti, Coastal S California Gila nigrescens, Guzman Basin Gila pandora, Rio Grande Acrocheilus, Columbia Dr. - Klamathella coerulea, Klamath Dr. Ptychocheilus lucius, Colorado Dr. Ptychocheilus grandis, Sacramento Dr. Eremichthys, Lahontan Dr. Relictus, L Franklin-L Waring, NV Ptych. oregonensis, Columbia Dr. - Orthodon, Sacramento Dr. Snyderichth. S Bonneville Dr. Snyderichth. Snake-Bear R. Lepidomeda molispinus, Virgin Dr. Lepidomeda vittatus, Little Colorado Dr. Rhinichthys osculus, western U.S. Mylocheilus, Columbia Dr. - Pogonichthys, Sacramento Dr. lotichthys, Bonneville Richardsonius balteatus, Bonneville- Columbia Dr. FIGURE 2.?Mitochondrial DNA phylogeny of western Cyp- rinidae, based upon parsimony cladistic analysis of the cyto- chrome b gene. This tree is extracted from a larger analysis of 88 cyprinid taxa (Dowling, unpublished). The relative rates test (see "Methods") is based on the expectation that all of the sister species to any outgroup should have the same percent sequence divergence. The variation in those distances (Tables 2-4) is a measure of the variation in relative rates of change and the lack of precision in any inference. The depth to each node is scaled to age in millions of years, assuming that 1 My corresponds to 1% average distance between clades (Table 2). Rank order of cladistic nodes and distances is preserved within 1%, except that the Relictus-Eremichthys distance is 15%. divergence per one million years for that gene and that clade (Figure 14), with an indication of the reliability of that estimate (Table 2). Once we have estimated the rate of change for a gene in a taxon, we can return to the cladogram and can estimate the amount of absolute time elapsed between sister clades in differ- ent basins, as estimated by genetic differences. Cladograms (without time) can then be recast as phylogenies (with estimates of time), as in Figures 2, 10, and 11. The estimated rate depends upon the correction factor for back mutations at all base pair sites. For example, the Tamura-Nei correction, used herein, gives 50 percent higher sequence-divergence values and older branch times (for the older cyprinids) than p distances or the Kimura two-parameter correction (Tamura and Nei, 1993). In the analyses below, both morphological and molecular cladistic estimates of rates of change are presented for the Cyp- rinidae (Figures 2, 3), but the molecular cladistic estimate is used to frame the relative-rate test. Correlations between the node sequences of the molecular trees and percent sequence di- vergence are positive and significant; correlations between node sequences of the morphological trees and percent se- quence divergence are not significant. Different kinds of hydrographic connections between basins affect molecular-rate estimates differently. For example, lacus- trine spillovers of lakes into rivers that transferred low-eleva- tion suckers, minnows, pupfishes, lacustrine trouts, and white- fishes occurred in sometimes datable pluvial times during glacial cycles (e.g., Lake Bonneville's spillover, McCoy, 1987), but headwater stream captures that transferred cold-wa- ter, high-gradient mountain suckers, salmonids, and sculpins occurred more frequently (e.g., between the Bear, Snake, and Green Rivers). Therefore, large, lowland fishes will generally contribute to older time estimates, whereas headwater fishes contribute to younger estimates. This is independent of the question of whether headwater fishes evolve more rapidly (Smith, 1992b). There are two alternative general hypotheses for times of di- vergence of Great Basin fishes. The first alternative is that a population was devoid of genetic structure prior to the most re- cent connection. Alternatively, differentiation may date not from the most recent connection and disconnection, but from a previous period of high water in the Pleistocene or Pliocene (see discussions of Whitehorse and Bonneville cutthroat trout and Death Valley pupfishes, below). Consideration of a broader range of possible isolation phenomena, along with dated fossils and molecular data, enable us to test alternative hypotheses about timing of divergence. Several additional problems may detract from the generality of these methods. In particular, fish fossils are limited, and some possess few apomorphic characters. More complete mo- lecular cladograms for more taxa are also needed. In addition, mtDNA data may have suffered introgressive loss of evolution- ary information in basins that frequently changed connections 184 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Richardsonius Phoxinus, etc. Hesperoleucus Lavinia Orthodon Mylocheilus Acrocheilus Pogonichthys Mylopharodon Ptychocheilus Gila atraria lotichthys Siphateles Relictus Eremichthys Gila Plagopterins Gila (Moapa) Rhinichthys FIGURE 3.?Morphological phylogeny of western Cyprinidae after Coburn and Cavender (1992). This tree was extracted from a consensus tree, which was based upon 100 equally parsimonious trees representing a larger sample of cyprinid fishes. among allopatric populations (Smith, 1992a; Echelle and Dowling, 1992; Dowling and Secor, 1997). Cladistic patterns and estimated amounts of molecular diver- gence are presented in the "Results." Distinct population units are listed and indexed in the Appendix, with their diagnostic characters, location, fossil ages, and brief history. Statements about the kinds of characters and the qualitative degree of dif- ferentiation seen between sister units are included to enable comparisons with our molecular- and fossil-based hypotheses about rates of evolution. Authors of recent species and subspecies names are given in the Appendix, following Robins et al. (1991). Citations and au- thors of species names of fossils are given in the Appendix. In- stitutional abbreviatons are defined in Table 1. Acknowledgments Robert Miller and his family and students collected thou- sands of fossil fish bones. Dwight W. Taylor was important in field and theoretical development of this project, but our in- debtedness to him does not imply his agreement with all of our methods or conclusions. The late Carl L. Hubbs and Ira LaRiv- ers collected fossils and compiled information. Ray Alcorn, Clyde Barbour, Robert Behnke, Sue Ann Bilby, Beverly Black, Anthony Boudreau, Bob Bright, Jack Broughton, Lee Carlson, David Catania, Charles Dailey, Jonathan Davis, R.L. Elder, John Fortuna, Kallie Hardin, Geoff Hoetker, Pat Holroyd, Tho- mas J. Kelly, Britt Leatham, Susan Lindstrom, David Madsen, James Madsen, Rose Malamma, Carl Mann, Todd Martin, Kelly McGuire, Carrie Muerer, Robert Negrini, Doug Nelson, Kevin Padian, Phil Pister, Marith Reheis, Bill Robin, Mark Roeder, Gary Scoppettone, Keith Smith, Ray Smith, Ralph Stearley, Scott Stine, Steve Stone, Joanna Strain and family, Dwight Taylor, Paul Wagner, David Whistler, David White, Russ Wickwire, Richard N. Williams, Todd Yeoman, and many other family members, students, and friends collected speci- mens or provided comparative material and information. Shane Webb provided goodeid mtDNA data, examined fundulid fos- sils for goodeid characters, and helped with computations. Donald Buth provided fish DNA for molecular studies. We are especially grateful to Robert Miller for granting us permission to mention fossils that he collected and for directing us to use- ful literature. The California Academy of Science and the Mu- seum of Paleontology, University of California at Berkeley, loaned fossils for study. Research was supported by grants from the National Science Foundation (NSF) and the U.S. Bu- reau of Reclamation to Dowling. The U.S. Fish and Wildlife Service, Utah Division of Wildlife Resources, Nevada Depart- ment of Wildlife, U.S. Forest Service, and the Wyoming De- partment of Game and Fish supported research by Shiozawa. California State University, Bakersfield University Research Council, and the Far Western Anthropological Research Group, Inc., Davis, California, supported research by Gobalet. Bonnie Miljour created the figures. Robert Miller, Catherine Badgley, Robert Hershler, Ed Wiley, and anonymous reviewers read and improved the manuscript. Results Fifty-two recent species?19 Cyprinidae, 12 Catostomidae, eight Salmonidae, five Cyprinodontidae, four Goodeidae, and four Cottidae?comprise the Holocene Great Basin fish fauna. Twenty-seven of these species are endemic. Four species among the cyprinids, goodeids, and cottids have been driven to extinction, and two species or subspecies of salmonids have been extirpated from the Great Basin as a result of habitat de- struction caused by human activities (Minckley and Deacon, 1991). Among the 52 species (especially cyprinids, salmonids, and cyprinodonts), more than 50 additional isolated popula- tions have been recognized as separate subspecies. Sixty-four fossil forms, from the late Miocene to the present, represent many recent lineages and several clades that are now locally extinct (for example, Gasterosteidae and Centrarchidae). Fossil NUMBER 33 185 fishes are now known from 55 Great Basin localities (Figure 1; Table 1); several hundred additional fossil localities occur at the northern edge of the Great Basin, on the Snake River Plain. Extensive analyses of mtDNA divergence have been carried out on recent representatives of these fossil forms: Dowling et al. (2002) on Cyprinidae (minnows), Echelle and Dowling (1992) on Cyprinodontidae (pupfishes), and Shiozawa and Evans (1995) on Salmonidae (trouts). CYPRINIDAE.?Minnows, family Cyprinidae, are the most diverse and widespread fishes in the Great Basin; the 19 Ho- locene lineages include 10 endemic species and more than 20 endemic subspecies. Three species of minnows, Gila atraria, Siphateles bicolor, and Rhinichthys osculus, have been espe- cially persistent and have differentiated into many distinct sub- species and local populations. Unfortunately, small minnows are under-represented in the Great Basin fossil record, so they contribute little to our estimates of rates of molecular evolu- tion. At least five lineages, Ptychocheilus, Mylocheilus in- flexus, M. robustus, Mylopharodon, and Lavinia (exclusive of Hesperoleucus), lived in the Great Basin in the Pliocene and Miocene but are absent there today. Siphateles is the best repre- sented genus in the fossil record, occurring in the Miocene to Pleistocene of the Lahontan Basin and Mohave River drain- ages. Fossil and living Cyprinidae in the Great Basin are herein as- signed to three ancient clades, which are based on our molecu- lar analyses (Figure 2; see also Simons and Mayden, 1997, 1998). These three clades are older than the Great Basin, and molecular evidence indicates that they probably date back to the Oligocene (Table 2; Appendix). The first of the three clades (Figure 2) in the Great Basin (including fossils) is the large western North American chub group that includes Mylopharo- don, Lavinia, Hesperoleucus, Siphateles, Gila, Acrocheilus, Klamathella, Ptychocheilus, Eremichthys, and Relictus (Ap- pendix, numbers 1-32). These are related to Orthodon in the Sacramento drainage system and to Phoxinus in eastern North America. This "Western Chub clade" has four subgroups in the Great Basin: (1) the Mylopharodon-Lavinia-Hesperoleucus and Siphateles clade (Appendix, numbers 1-22); (2) the Gila-Acrocheilus-Klamathella clade, with Ptychocheilus lu- cius as its sister group (Appendix, numbers 23-29); (3) the Relictus-Eremichthys clade, including Ptychocheilus grandis (Appendix, numbers 31-32); and (4) the Ptychocheilus oregon- ensis clade (Figure 2; Appendix, number 30). Whether or not Ptychocheilus is actually paraphyletic, as indicated by Figure 2, is problematical. Ptychocheilus cf. oregonensis is one of the oldest cyprinids in the region (Smith et al., 2000). The second major cyprinid group in the Great Basin is the Snyderichthys-Lepidomeda (spinedace) clade (Appendix, numbers 33-35). It is related to the other spinedaces, which in- habit tributaries of the Bouse Embayment (see below). Sister groups of this clade include Margariscus and Couesius of northern and eastern North America. The third major minnow clade includes western American outliers of the huge eastern North American shiner clade. West- ern forms include Iotichthys (Appendix, number 36), Richard- sonius (Appendix, numbers 37-39), Rhinichthys (Appendix, numbers 40-55), and Mylocheilus (Figure 2; Appendix, num- bers 56-58). This group also includes Pogonichthys of the Sac- ramento drainage system. An alternate cladistic hypothesis (Figure 3), based upon morphological analysis, shows different interpretations of relationship (Coburn and Cavender, 1992). Great Basin cyprinids are listed (Tables 1, 2), numbered, and reviewed in the Appendix in approximately the sequence given in Figure 2. Typical distribution patterns of minnows (illus- trated by Gila and Siphateles) are shown in Figure 4. Fossil minnows in the Great Basin date back to the middle Miocene or earlier (Cavender, 1986; Reynolds, 1994). Many cyprinid genera are early members of the Great Basin fauna (Appendix, numbers 5, 14, 15, 26-28, 30, 56-58). The Gila group is widespread and especially diverse in all of the Great Basin as well as in the Colorado River drainage, the Rio Grande, and northern Mexico. Siphateles is centered in the Lahontan Basin and Sacramento River drainages, and Gila is in the drainages surrounding the Lahontan Basin. The Great Basin part (White River) of the lower Colorado River (Bouse Embay- ment) drainage is inhabited by Lepidomeda, Moapa, and Gila, as well as by Catostomus clarki and the empetrichthyine good- eids. Snyderichthys copei of the Bonneville Basin and upper Snake River drainages is cladistically part of the Plagopterin group and provides evidence of Pliocene and Pleistocene drain- age connections (Dowling et al., 2002). The southern Califor- nia coastal form, Gila orcutti, is the sister group to Gila atraria of the Great Basin and to G. jordani and G. cypha of the Colo- rado River drainage. Electrophoretic variation in Gila atraria of the Bonneville Basin was studied by Rosenfeld (1991), who concluded that ac- cumulated differences were influenced more by natural selec- tion in local springs, lakes, and streams than by vicariance and time. He estimated that 14,500 years was too short an interval for the accrual of qualitative differences between hierarchical groups of vicariant populations. Populations of Gila atraria from northern and southern Bonneville basins and from the Bear River have diverged extensively in their mtDNA; they have not remained uniform during the last 2 My (sequence divergence = 0.009-0.024) despite recent and probably fre- quent sympatry. The northern Bonneville form of G. atraria shares an unexpectedly high proportion of its mtDNA with G. cypha and G. jordani of the Colorado River (percent sequence divergence = 1.1% and 0.7%, respectively; Table 2), support- ing the hypothesis that a connection existed from Lake Bonne- ville through Meadow Valley Wash, Nevada, to the Colorado River drainage (Ives, 1948). A separate biogeographic pattern encompasses the northern Great Basin and the Columbia River and Snake River drain- ages. This pattern is marked by the distributions of Kla- mathella, Siphateles, Acrocheilus, Lavinia, Mylopharodon, and 186 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 121? ?119- 117* 115 Miocene Pliocene Pleistocene -35s FIGURE 4.?Distribution of Holocene and fossil Siphateles, Gila, and certain relatives in the Great Basin. Map shading indicates different species or subspecies ranges. Ages of fossil fish are indicated by symbols defined on the figure. Numbers refer to species or subspecies listed in the Appendix. Mylocheilus in a Columbia-northern Great Basin-Pit River track. At a lower cladistic hierarchical level, this biogeographic pattern broadens to include the Sacramento River drainage, with Ptychocheilus grandis, Hesperoleucus, and Orthodon (Figures 2, 3), and the Colorado River drainage, with Pty- chocheilus lucius. The several Great Basin records of My- locheilus and Mylopharodon are evidence of a late Miocene connection between west-central and northwestern Nevada and the Snake River Plain. Mylocheilus caurinus (Table 2, number 58) is the closest living relative to Mylocheilus robustus of the Mio-Pliocene Snake River Plain (Appendix, number 58) and to Mylocheilus heterodon from the Ringold Formation of Wash- ington (Smith et al., 2000). Pairwise distances among clades that are based upon a parsi- mony cladistic analysis enable evaluation of relative rates of molecular change (Table 2). The homogeneity of these rates among minnows can be evaluated by comparing the standard deviation (s.d.) around the mean distances. The 26 species esti- mated to share a common ancestor with Phoxinus have a mean distance between each of them and Phoxinus of 0.22 (s.d. = 0.022). The 66 combinations of sister species to Mylopharo- don, Lavinia, and Hesperoleucus have a mean distance of 0.15 NUMBER 33 187 (s.d. = 0.02); the 38 combinations of Siphateles sister species have a mean distance of 0.12 (s.d. = 0.02); the 30 species of the Gila orcutti-G. atraria group have a mean distance of 0.08 (s.d. = 0.02); the 26 sister species to Acrocheilus and Kla- mathella have a mean distance of 0.08 (s.d. = 0.02); the 216 combinations of the spinedaces have a mean distance of 0.34 (s.d. = 0.03); and the 210 combinations of the western repre- sentatives of the shiner clade have a mean distance of 0.35 (s.d. = 0.05). With the rate of divergence estimated at about 1% per one million years, the standard deviations of group or bar- rier ages is usually on the order of 10%, but rarely more than 25%, of the estimated age (Table 2). The distances and inferred timing of cyprinid patterns span the time from Miocene to the last postpluvial period. CATOSTOMIDAE.?Catostomids are the most ancient of the fish groups present in the Great Basin (Cope, 1872; Smith, 1966, 1992b; Minckley et al., 1986). They are also among the largest and hardiest aquatic survivors of the Great Basin's his- torically variable climate and ecology. Catostomidae are the second most diverse and abundant family in the Great Basin, with 12 species, but the maximum number of sympatric forms is three or four. Great Basin suckers are part of a primarily western American monophyletic group (Figure 5), the Catosto- mini, whose sister group (and outgroup in the analysis) is the 78 Catostomus clarki, Virgin-White 77 C. discobolus, Colorado-Snake-Bear C. santaanae, Los Angeles 79 C. platyrhynchus, Great Basin-Columbia C. plebeius, Rio Grande C. leopoldi, Yaqui C. wigginsi, Sonora C. microps, Pit C. rimiculus, Rogue, Klamath 76 C. columbianus, Columbia C. catostomus, Missouri-Columbia 72 C. tahoensis, Lahontan 73 C. fumeiventris, Owens C. snyderi, Klamath 66 C. ardens, Bonneville-U Snake 67 C. macrocheilus, Columbia-L Snake C. latipinnis, Colorado C. occidentalis, Great Valley C. bernardini, Yaqui C. insignis, Lower Colorado C. commersoni, Mississippi Xyrauchen texanus, Colorado 59 Deltistes luxatus, Klamath Chasmistes brevirostris, Klamath 60 Chasmistes cujus, Lahontan 61 Chasmistes liorus, Bonneville FIGURE 5.?Estimate of the phylogenetic tree for Chasmistes, Deltistes, and Catostomus in the Great Basin and their relatives in the Catostomini (from Smith, 1992b). Not all species of Chasmistes were included in the cladis- tic analysis. Numbers refer to the species accounts in the Appendix. 188 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Moxostomatini of the eastern United States and Mexico (Smith, 1992b). The Catostomini range from medium-sized, benthic, head- water species in the subgenus Pantosteus (genus Catostomus) (Figure 7), to large-sized benthic predators in low-gradient streams (other Catostomus, Figure 6), to large-sized plankti- vores in large lakes (Chasmistes, Figure 7). The fossil record of catostomids is diverse; it includes plesiomorphic forms as well as forms that were morphologically more specialized than present-day species. Fossil Catostomini are known from the Miocene (10 Ma) to the Holocene, with Great Basin diversity reaching a peak of four species in the northern Bonneville Ba- sin (and upper Snake River and upper Colorado River basins) in the late Pleistocene (see Appendix, numbers 59-79). Ca- tostomids are tetraploid fishes that show abundant hybridiza- tion and extreme variation, but no indication of rapid speciation (Smith, 1992b). There are two clades of Catostomini in the Great Basin (Smith, 1992b). The first is the genus Chasmistes, the lake suck- ers. Four Holocene species of Chasmistes are localized in Utah ^37' miles 113' ? Miocene A Pliocene # Pleistocene ' ' iin* -35' FIGURE 6.?Native distribution of Catostomus ardens (Utah, 66), C. macrocheilus (Idaho, 67), C. fumeiventris (southern California, 73) and C. tahoensis (Nevada, 72) in the Great Basin and some of their relatives outside the Great Basin. Shading indicates drainage basins occupied by species (as numbered in the Appendix). Fossil records and Holocene occurrences are indexed to the numbered species accounts in the Appendix. Ages of fossil fish are indicated by symbols defined on the figure. These are low-gradient fishes that required a major river or lake connection to cross a drainage divide. NUMBER 33 189 76 columbianus 77 discobolus 78 clarki gg?| 79 platyrhyncus | | -351 121" FIGURE 7.?Distribution of the mountain suckers: Catostomus platyrhynchus (Utah, Idaho, Wyoming, Nevada, 79), C. columbianus (Idaho, Oregon, 76), C. clarki (southern Nevada, 78), and C. discobolus (Utah, Wyoming, Idaho, Colorado, 77) in the Great Basin and of their relatives (C. rimiculus, Oregon; C. santaanae, California) outside the Great Basin. Ages of fossil Chasmistes and Deltistes in the Great Basin are indicated by symbols defined on the figure. Numbers refer to the species accounts in the Appendix. and Pyramid Lakes, in Klamath River (Figure 7), and in the headwaters of the Snake River. There are also numerous Mio- cene to Holocene fossil Chasmistes specimens in the Great Ba- sin (Miller and Smith, 1981; Figure 7). Silicified gill-raker plates with fimbriate distal ends have been found in the Mono Lake basin. Fimbriate distal ends are an apomorphy for Chas- mistes. Their presence in Mono Lake deposits indicates this ad- aptation to planktivory occurred 4-3 Ma (Gobalet, unpublished data). The sister genera to Chasmistes are Xyrauchen and Del- tistes (Appendix, number 59) of the Colorado River and the Kla- math system, respectively. The modern lake sucker in Utah Lake, Chasmistes liorus mictus Miller and Smith (1981), shows intense introgression from Catostomus ardens. The introgres- sive event is remarkable because Chasmistes and Catostomus diverged 10-4 Ma (both genera are morphologically distinct in Pliocene sediments of the Snake River Plain; Smith et al., 1982). The second clade is the genus Catostomus, which includes two ecological groups that were formerly classified as subgenera 190 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES (Smith, 1966; Smith and Koehn, 1971). As now understood, the subgenus Pantosteus (mountain suckers) (Appendix, num- bers 76-79) is monophyletic, but recognition of Pantosteus makes the subgenus Catostomus paraphyletic (Smith, 1992b). The more plesiomorphic species of catostomids are usually re- stricted to lowlands in basins (Figure 6), where they inhabit low-gradient rivers or lakes in a vicariant pattern. These spe- cies are C. ardens in the Bonneville Basin, C. macrocheilus in the Harney-Malheur Basin (and Columbia River basin), C. warnerensis in the Warner Lakes basin, C. tahoensis in the Lahontan Basin, C. fumeiventris in the Owens River basin, and C. shoshonensis in Wall Canyon of Surprise Valley (Ap- pendix, numbers 66-75; Figures 5, 6). Fossil Catostomus (10 Ma to Holocene) are known from the north edge of the Great Basin, which is along the southern margin of the Snake River Plain. The mountain suckers (C. columbianus, C. platyrhynchus, C. discobolus, C. clarki) are morphologically derived (Smith, 1992b). They inhabit high-gradient streams and stream head- waters, and they have been broadly dispersed by many headwa- ter transfers (Figure 7; Smith, 1966). Catostomus platyrhyn- chus has been transferred from the Bonneville Basin to the Lahontan Basin and to the drainages of the Snake/Columbia, Colorado, and Missouri Rivers. Catostomus discobolus is na- tive to the upper Colorado River, including the Grand Canyon, and has been transferred among the Colorado River and Snake River drainages to the Bear and Weber Rivers in the northern Bonneville Basin. Catostomus clarki is found in the drainages of the Gila, Virgin, and White Rivers, but not the Colorado River in and above the Grand Canyon, and it helps define the Pliocene drainage of the Bouse Embayment of the Gulf of Cal- ifornia during the time when it might have been blocked from its connection with the upper Colorado River drainage (Taylor, 1983). Catostomus santaanae, in the Los Angeles area (Figure 7), is the sister species to the Colorado River species, C. disco- bolus and C. clarki. Catostomus santaanae represents the pop- ulation that lived in the lowest Colorado Paleoriver when it flowed to its Los Angeles delta, prior to 9 Ma. Catostomus co- lumbianus inhabits both the Harney-Malheur Basin and the Columbia River drainage and helps to define the overflow through the South Fork of the Malheur River to the Snake River across the Voltage lava flow (Piper et al., 1939), which occurred as recently as about 9000 years B.P. (Gehr and New- man, 1978). ICTALURIDAE.?North American catfishes of the genus Ameiurus were rare in the Great Basin. Two Miocene forms are known (Appendix, numbers 80, 81), one from the Truckee For- mation in Nevada and one from formations in the Salt Lake group in northern Utah. Their relatives include fossils in the Pliocene Columbia River, Miocene and Pliocene Snake River, and Miocene to Holocene rivers in eastern North America (Lundberg, 1992; Smith et al., 2000). These occurrences pro- vide additional evidence for former Great Basin connections to the Snake River Plain and to south-central Washington before the formation of the modern Columbia drainage. The Miocene form from northern Utah is most similar to bullheads of the Snake River Plain, supporting a Hemphillian or older connec- tion between the Salt Lake group and Lake Idaho (Taylor, 1966; Stokes, 1979). No Ameiurus are known from the south- western United States or Mexico, despite the diversity of other genera of catfishes in Mexico (Lundberg, 1992). SALMONIDAE.?Eight native species of salmonids apparently inhabited the Great Basin in recent times: rainbow trout (Onco- rhynchus mykiss), cutthroat trout {Oncorhynchus clarki, with many subspecies), bull trout (Salvelinus confluentus), four whitefishes (Prosopium spp.), and possibly the Arctic grayling (Thymallus arcticus). Additional trouts and char are represented in the fossil record. Salmonids are cool-water fishes that inhabit mountain streams and deep lakes. Their distribution patterns in- dicate that they dispersed among basins via headwater-stream captures as well as by discharge from pluvial lakes. During in- terglacial times (such as the present), desiccation and warming of aquatic habitats restricts these fishes to small populations in a few cool lakes, rivers, and headwaters. During pluvial stages of the last 3 My, however, there were numerous, large popula- tions (see Appendix) in large, deep lakes and large rivers. Salmonids from the Miocene, Pliocene, and Pleistocene are known to have occurred in lake basins, such as the Salt Lake group, Fort Rock Lake basin, Lahontan Basin, Honey Lake ba- sin, Bonneville Basin, and Owens Basin. Holocene salmonids are found in drainages west, north, and east of the Great Basin, and in a few drainages at high elevations in Mexico. COREGONINAE: There are six recent species of whitefishes in the genus Prosopium of the subfamily Coregoninae. Four live in the Great Basin; three of these are endemic to Bear Lake in Utah and Idaho (Appendix, numbers 82-85). The other two, P. cylindraceum and P. coulteri, inhabit eastern to north- western North America and Siberia (Smith and Todd, 1993). At least one fossil species, P. prolixus Smith (1975), is com- mon in Pliocene sediments of the Snake River Plain. The mor- phologically based, cladistic estimate of relationships of the species in the group is shown on a distribution map (Figure 8). Unpublished molecular data (Shiozawa) indicates a more com- plicated pattern than that depicted in Figure 8. Prosopium illus- trates intralacustrine as well as allopatric speciation in Bear Lake and in subbasins of the Bonneville-Bear system. As pre- dicted by the intralacustrine speciation hypothesis, the differ- entiated characters among the species of Prosopium are mostly associated with feeding: the shape of the mouth and jaw and the numbers of teeth, gill-rakers, and pyloric caeca (Smith and Todd, 1993). The phylogenetic pattern indicates that the most- derived forms in the genus evolved in the Bonneville Basin and Bear Lake. THYMALLINAE: Grayling {Thymallus) have not been re- garded as part of the Great Basin fauna. There are no fossil or Holocene specimens documenting their presence. Inclusion in the present list is based upon the apparently knowledgeable identifications in the journal of John Townsend (1978), who recorded collecting grayling in the Bear River in 1834 (Appen- dix, number 86). NUMBER 33 191 123? 121* -43 prolixum, Pliocene, Idaho coulteri, North, L. Superior cylindraceum, North williamsoni, West spilonotum, Bear Lake abyssicola, Bear Lake gemmifer, Bear Lake -35 109? 107' FIGURE 8.?Native distribution of Prosopium (shown with shading) in the Great Basin and nearby drainages, with estimate of the phylogenetic tree (from Smith and Todd, 1993). Shading indicates drainage basins occupied by species (as numbered in the Appendix). Symbols representing fossil ages are defined in Figure 1. SALMONINAE: Settlers in the nineteenth century found abundant trout in many northern Great Basin lakes and streams (map, Figure 9), but these were quickly overexploited (Behnke, 1992). By the 1930s, introduced hatchery fish began to cloud the possibilities for interpreting salmonid and hydro- graphic history (Behnke, 1981, 1992), and unfortunately mu- seum specimens that predate these introductions are sparse. Modern protein and mtDNA studies are, however, discrimi- nating native from introduced stocks and are adding consider- able historic information (Leary et al., 1984, 1987). Trouts are adapted to cold lakes and streams and to high gradients. Stream forms survived arid times at high elevations and gla- cial times in large lakes. There are indications that, in addition to trouts (Oncorhynchus), at least two char (Salvelinus) were native to the northern Great Basin; this evidence is reviewed in the Appendix (number 87). The fossil record of Salvelinus is older than the fossil record of Oncorhynchus. Salvelinus was in the Great Basin (central Nevada) and northern Idaho in the middle Miocene and in the Snake River Plain in the late Miocene. The late Miocene Oncorhynchus of the Truckee Formation differs in many characters from modern species of Oncorhyn- chus. It possesses synapomorphies that are shared by O. clarki, indicating it is more advanced than the basal Mexican golden 192 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 123? 12T 119* 117" 121? ?119* 117* 115 Miocene Pliocene Pleistocene ' i -35' FIGURE 9.?Distribution of Holocene and fossil species and subspecies of rainbow trout {Oncorhynchus mykiss, stipple pattern) and cutthroat trout (O. clarki, shaded patterns) in the Great Basin and adjacent drain- ages. Shading indicates drainage basins occupied by species and subspecies (as numbered in the Appendix). Squares = Miocene; triangles = Pliocene; bullets = Pleistocene. trout, O. chrysogaster. The characteristics of the Truckee trout indicate that it existed before the differentiation of rainbow trout and cutthroat trout (Stearley and Smith, 1993), while its habitat was connected to the Pacific Ocean. As early as late Miocene and throughout the Pliocene, the trout, ^Oncorhyn- chus ("Rhabdofario") lacustris (Cope), of the lower Snake River drainage and northwestern Nevada, shows a composite of O. clarki (cutthroat trout) and O. mykiss (rainbow trout) characters, e.g., O. clarki dentaries, intermediate maxillae, and O. mykiss premaxiilae. These characters diagnose either the ancestor prior to the clarki-mykiss divergence or the ancestral redband trout {Oncorhynchus mykiss gairdneri). The distinc- tion cannot be made at this time because intermediates be- tween O. clarki and O. mykiss still exist, especially in the Co- lumbia Basin. Characters of these trout subsequently sorted into O. mykiss and O. clarki elsewhere, but they continue to be introgressed or unsorted in redband trout throughout their range in the lower Columbia River, Klamath River, and Sacra- mento River drainages (Leary et al., 1984, 1987; Berg, 1987; Allendorf and Leary, 1988), possibly throughout much of the last 3 My. A sample of hatchery-produced O. mykiss differs from O. clarki by 3.7% mtDNA (Table 3). NUMBER 33 193 ?? pi 'S u < B Q "=E I g)VJ U of > ? ?a S I? O O 60 > .? U o 5 w '+3 W D. O O O O o o o o ?pop o d d d d 00 00 00 00 00 o o o o o o ~H t^ i^ r~ t~* t~-o o o o ? o p p p p p p a c> d d <^ d cO ^^ ON ON ON ON ON ocic>cicicid -HroTfoooor^vovo pppppppp odd d od d do ? m t oo oo h vo ? pppppppp oododddo'dd p p p p p p p p p poddddddddo'd ^?fOOOONCNCN-^^-i ? ? ~ ? OO ? ? ? ?' ?pppppppppp oodo'dddddddd r--r^oooovoTj-cNoo oo't~~ p^ r~ pp pppppppppp oddddodo'ddddo r~r-oooovo-^-rNoooor-r~r~ OO?C ?' ? -H?I-H ? ? ? ?PPppppppPppO ooddoddodddodd cNcNoooooooovoTtrnoooor-r~t^ OOOO ? ? ? ~-c ? ? ?< ?. ? ?pppppppp pppppp ooodooooooooodo TrfnmoNONOooovo^moooot~-c-^r? OOOOO ? ? ? ? ? ? ? ? ? -Hpppppop pppppppp oo'doddddo'dddddod in in ro ro ^5 ON r^- r*^ vo vo P^ ^5 o ON ON ON pppppppp pppppppp a>cicid>c>c>c>c> minncNrnrommmm^l-Tf'a-TtTj-rnfnmmm pop pppppp pppppop p COOOTt^CNCNCNCNOOOOOOONOfNCNrOmmr-4O ? ? ? ? ? ? ? ? -H^^CNCNCNCNCNCN pppppop pppppp pppppodo'oddo'o'ddo'oddo'oddd ^ts^H\otsrvffifjoot^^orvM(NfS'?< *?'ro in r*^ ^" c> en co ro ro ^r ^" ro ro ro ^}" ^J" ^f TJ* ^" ppppppppppppppppppp mykis out Cr o s PA C/5 Pi Cr u h a; Q on CQ - _ Pi U ^ X Cr u 3 is 13 e uu? 0H Oncorhynchus clarki, the cutthroat trout, has been in the Great Basin longer than proposed by Behnke (1981, 1992), ac- cording to the Miocene fossil record. Molecular data indicate that interbasin differentiation of O. clarki subspecies goes back nearly 5 My, into the Pliocene (Table 3). The Great Basin trouts form three clades. The Bonneville Basin clade is cladistically basal, being the sister group to the remainder of the species (Figure 10). The Snake River clade is the next-most basal group, and it includes the Bear River form. Although the Bear River is now a tributary to the northern Bonneville Basin, it was part of the Snake River drainage prior to 20-15 Ka (McCoy, 1987; Curry, 1990; Bouchard et al., 1998). The Lahontan Basin and southern Oregon subspecies are a sister group to the westslope trout subspecies, O. c. lewisi. Coastal cutthroat trouts, O. c. clarki, are an intermediate clade between the Colorado River plus Snake River groups and the Westslope plus Oregon-Lahontan groups (Figures 9, 10). Fort Rock Valley, west of the Harney-Malheur Basin in Lake County, Oregon, was the site of Fort Rock Lake, also called Fossil Lake, which was a tributary to the Deschutes and Co- lumbia Rivers prior to blockage of the outlet by postpluvial lava flows (Allison, 1940, 1982). Pliocene and Pleistocene sed- iments in the Fort Rock Valley contain trouts, minnows, and suckers. A Pliocene salmon from diatomaceous sediments of the Fort Rock Formation, Oncorhynchus sp. (either nerka, keta, or gorbuscha), was reported by Cavender and Miller (1972). The trout specimens in this basin (Oncorhynchus sp.; Appen- dix, number 90) are osteologically similar to redband trout and were considered to date from the late Pleistocene by Allison and Bond (1983). They based their conclusion upon shared characters of the palatine, maxilla, hyomandibular, and orbito- sphenoid. More than 50 dentaries from Fossil Lake were exam- ined for the present study. Forty-seven maxillae have a mixture of modern O. clarki (from Pyramid Lake before introductions) features and features of Pliocene O. lacustris (pre-clarki and pre-mykiss) of the Snake River Plain, whereas seven maxillae resemble O. m. gairdneri (redband trout). One premaxilla re- sembles that of O. clarki and seven resemble that of O. mykiss, but it is not clear whether the differences represent variability in one species. Palatines from Fossil Lake are deep, like that of O. lacustris, and the hyomandibular is angled posteroventrally, also like that of O. lacustris. The Pliocene fish fauna of Secret Valley, Modoc County, California (Oncorhynchus (Appendix, number 91), Kla- mathella, Ptychocheilus, Lavinia, Chasmistes, and possibly Ac- rocheilus and Cottus), and the Pliocene fish fauna of the Snake River Plain are very similar; this suggests the two drainages were connected prior to the Pliocene (Wheeler and Cook, 1954; Miller and Smith, 1981; Wagner et al., 1997). A similar fauna from the nearby Alturas Basin, Modoc County, California, con- tains Klamathella, Ptychocheilus, Chasmistes, and Catostomus (Wagner et al., 1997). These similarities suggest former hydro- graphic connections (not necessarily contemporaneous) among the Pit River, the Snake River, and the Klamath Basin. 194 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES mykiss c. subsp. (99) c. subsp. (99) c. Utah (98) c. subsp. (100) c. bouvieri c. bouvieri c. bouvieri c. pleuriticus c. pleuriticus c. clarki c. clarki c. lewisi c. lewisi c. lewisi c. henshawi (93) c. subsp. (95) c. seliniris (94) c. subsp. (96) c. subsp. (97) Trout Cr Deaf Smith Sevier Dr. Bear R Yellowstone Snake R Sedge Cr Current Cr Sheep Cr Mack Flynn FishL Wing Cougar Washburn Humboldt Silver King Willow Cr Whitehorse Cr w s T 04 03 02 01 0 % sequence divergence FIGURE 10.?Estimate of the phylogenetic tree of Oncorhynchus clarki (cutthroat trout) and its outgroup, Oncorhynchus mykiss, based on restriction fragment analysis of mtDNA. Branch lengths may be read as esti- mated time if divided by the calculated rate of 0.5% sequence divergence per 1 My. Abbreviations: BONNEV - Bonneville Basin, SNAKE = Snake River drainage, COLO = Colorado River drainage, COAST = Coastal drainages, WEST = west slope drainages, LAH = Lahontan Basin, OR - Oregon Lakes drainages, southeast Oregon. CYPRINODONTIDAE.?The Great Basin pupfishes are the showpiece of Great Basin ichthyology. They were the focus of R.R. Miller's doctoral research, the results of which were pub- lished in his 1948 monograph. This monumental systematic and experimental work illustrated the relationship between isolation and differentiation in relict waters. Pupfishes of the Great Basin's Death Valley system (Miller, 1946a) have been the subject of extensive ecological and physiological studies (Brown and Feldmeth, 1971; Soltz and Naiman, 1978; Feld- meth, 1981), behavioral observations (Liu in Turner, 1974), biochemical comparisons (Turner, 1984; Echelle and Echelle, 1993), and molecular analyses (Echelle and Dowling, 1992; Parker and Kornfield, 1995). Relationships of these species are now understood to be more complex than formerly thought. Cyprinodon radiosus (Appendix, number 101) of the Owens River is the sister species to C. macularius of the lower Colorado River, and these two species are the sister group to the clade that includes C. salinus, C. nevadensis, C. diabolis, and relatives, in the Salt River, the Amargosa River, Devil's Hole, and Ash Meadows (Figure 11; Appendix, numbers 102-110). This Death Valley and Ash Meadows clade possibly includes Cyprinodon fontinalis of northern Chihuahua, Mex- ico (Figure 11), but this is not consistent with geography or scale morphology (Miller and Smith, 1986; Minckley et al., 1986). Hydrographic connections among basins and with drainages in Mexico are indicated by Cyprinodon mtDNA analysis (Figure 11; Table 4; Echelle and Dowling, 1992; Ech- elle and Echelle, 1993) and molluscan data (Hershler and Pratt, 1990). Study of the data usually regarded as indicative of the relationship between isolation and divergence reveals interesting conflicts?molecular, biochemical, and morpho- logical divergence and reproductive isolation are incongruent NUMBER 33 195 C. eximius C. variegatus C. meeki Whitefin west Whitefin east Palomas C. m. eremus C. m. macularius C. radiosus C. fontinalis C. salinus C. diabolis C. n. mionectes C. n. shoshone C. n. amargosae C. nevadensis sbsp Rio Grande Gulf Coast, Miss. RMezquital,Mex. RYaqui,Mex Guzman Dr. Mex. Palomas, Chih.,Mex. Organ Pipe N.M., Ariz. Colorado R., Calif. Owens R., Calif. Chihuahua, Mex. Salt R., Death V. Devils Hole, Death V. Ash Meadows, Death V. Shoshone Spr., Death V. Amargosa R., Death V. Ash Meadows, Death V. 08 07 06 05 04 03 02 % sequence divergence 01 FIGURE 11.?Phylogenetic relationships (strict consensus tree) of certain western North American Cyprinodon, based upon parsimony analysis of mtDNA restriction site data (Data and trees from Echelle and Dowling, 1992). The Death Valley species are C. C. nevadensis, C. salinus, and C. diabolis; C. milleri has C. salinus mtDNA. Cyprinodon radiosus is the Owens River species, and C. macularius is in the lower Colorado River and vicinity. The other species are primarily from Mexico (Echelle and Dowling, 1992). The depth to each node may be scaled to age in millions of years, assuming that 1 My corresponds to 1% of the average percent sequence diver- gence between clades. with each other and with evidence for timing of divergence (Turner, 1974; Echelle and Dowling, 1992; Echelle and Ech- elle, 1993). Cladistically analysed molecular data indicate that Cyprinodon has been in the Death Valley system for about 3 My. Biogeographic relationships of Cyprinodon (and good- eids) of the Great Basin are to the south, in Mexico, especially if C. fontinalis, from Chihuahua, is related to the Death Valley clade (Figure 11). The diversity of Cyprinodon species within the Great Basin, taken with its sister species, C. macularius Baird and Girard, displays a geographic pattern that extends from Death Valley and the Owens River to across the Gila River and the lower Colorado River drainages (Minckley et al, 1986). Pliocene connections from the lower Colorado River to the Owens, Mohave, and Amargosa Rivers occurred about 3 Ma, according to molecular data (Figure 11; Table 4). The connec- tions possibly involved China Lake, Searles Lake, Panamint Valley, and Wingate Pass to Lake Manly, now Death Valley, according to fish and other evidence (Miller, 1946a; Hubbs and Miller, 1948a; Benson et al., 1990). The Pliocene Owens-Mo- have drainage connected to the lower Colorado River drainage (Blackwelder, 1933, 1954; Hale, 1985) near the region of Lud- low Pass (at 594 m), Bristol Lake, and Cadiz Lake. Brown and Rosen (1995) found no evidence of a through-flowing connec- tion across the present topography, and they suggested alterna- tive dispersal processes, routes, or times; however, rounded boulders in a large paleochannel, which are evidence of a through-flowing stream, are present near Ludlow (Hale, 1985; G.I. Smith, pers. comm. to G.R.S., 1999). Encroachment of the Gulf of California toward Death Valley (Durham and Allison, 1960) and the presence of foraminifera correlated with the Bouse Formation in cores taken from the dry lakes of Danby and Cadiz in southeastern San Bernardino County, California (P.B. Smith, 1960, 1970), suggest a connection existed to the estuary of the lower Colorado River. The most compelling fish evidence for a connection between the Mohave and Colorado Rivers is the distribution of Cyprinodon macularius of the lower Colorado drainage and C. radiosus of the Owens River. Mitochondrial DNA of these two species indicates they had a common ancestry between 2 and 3 Ma (Table 4). 196 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES "S. s a S ? "S ? to O 1 c (J cj Species or (3 amargos cu a ? olis-o ?SSj ?S"3 a 5?2 1<*> .3'C -2s o1 3 erem, | a. pup ^?' a.3 CL ? 5 1 egi ? (MS .5 subspecies ? :imius ? ? ? iriegatus NJ ? oo ? ? ? ? eekiS U ? ON ? ? OOVO ?? ? ? ? V3(LI ^. :efin pupfish (v ? ?? ON ? OO ? ? ? ? ?3 :fin pupfish (ea'Js ? fN^5 ?? fN? ?? OO ? ? ? ? ON ? ? mas pupfish CO OH ? fN? ? fNfN ?? ON ?? ON ? ? ? ? ? ? . eremusS ? oo ?? fN? ? ON ?? ? ? ? ? ? ? . maculariusS ? 00fN ??' r- fN? ? fN? ? OO fN? ? OO fN? ? fN00 ? ? S? t- ? ? 1 ? ? ? fN? ?" f tN? ? ?? IT) ?? If} ?? OO VO ?? ? ? ? ? ntinalis ? fN? ? ONs ?? ??' NO ?? fN ?? f^l f**lO ? f*N ?? 00 ?? fN ? fN ? ? | ? fN ? ? ON ?? ? ? NO ?? ?? VOfN ? ? 00 fN? ? 00 tN? ? ? 00 ?? ? ? wo ? ? abolis O ? ?? ? ? ? ?fN ?? 00 ? o VO CO? ?? tNen ?? fN? ? ON fN? ? ON fN? ? r- ?? ?00 ?? ? ? mionectes a ? VD ??' in ?? ? ? ? 00 fN? ? ? ? m? ?' 00 fN? ? ?fN ? ? fNtN ?? fNfN ?? ? ? tNOO ?? fN O ? shoshone s: ? ,003? VO oo oo ? ? ? fN? ? in ro O? j?m ?? tN ?? fN? ? VOfN ?? VO fN? ? ? ? i?r- ?? ? ? a.o o amargosae (T a U 0.002 002 ? ?? ?' ?? ? fN o ? tNtN ??' tN ? ? ?? ? ?? fNfN ?? fN? ? tN? ? 00 ? ? oor- ?? ? ? ^i '.vadensis subsja G00DEIDAE.?The poolfishes and springfishes, Empetrich- thys and Crenichthys, respectively (Appendix, numbers 112-115), are Great Basin endemics in the subfamily Empet- richthyinae (Webb, 1998) of the family Goodeidae (Parenti, 1981; Grant and Riddle, 1995). Empetrichthys and Crenichthys are oviparous and were formerly classified as cyprinodontids, but Parenti's (1981) phylogenetic evidence demonstrated their relationship to the live-bearing goodeids of the Mexican Pla- teau. Empetrichthyines resemble fundulids in osteology, num- bers of fin rays, and tooth form, which distinguish them from the cyprinodontids, but unlike fundulids, they lack pelvic fins. Fossil Empetrichthys were discovered from the Pliocene near Los Angeles, California, and from the Pleistocene in southern Nevada (see Appendix, number 112). Paleontological studies by M.L. Smith (in Miller and Smith, 1986) and molecular stud- ies by Webb (1998) indicate middle to late Miocene goodeid vicariance between the White River and the lower Colorado River drainages and the Mexican Plateau, probably involving tributaries of the Bouse Embayment and the Gulf of California. The distribution of fossil and Holocene Empetrichthys, Fundu- lus, and Gasterosteus, along with species of Pyrgu I ops is (hy- drobiid snails), provide evidence that late Miocene lowland Pa- cific coast tributaries flowed westward from west-central Nevada and Death Valley across a lower, proto-Sierra Nevada (Bell, 1979; Hershler and Pratt, 1990) about 6 Ma, perhaps through the (reversed) Mohave River valley (Weldon, 1982). FUNDULIDAE.?Most species of killifish in the family Fun- dulidae live in eastern North America, but there are representa- tives in Mexico and California. Several fossil forms were de- scribed (Miller, 1945b) before it was recognized (Parenti, 1981) that the Goodeidae of the Mexican Plateau, relatives of Fundulidae, are represented in the Great Basin. The Nevada fossils (Appendix, numbers 116-123) have some observable characters that suggest they could be related to the Empetrich- thyinae or the Goodeidae (Miller, 1948:100), except that none of the Nevada fossils lack pelvic fins, as do the empetrichthy- ines. The ancient diversity of fundulids in the southwestern Great Basin is remarkable, considering the paucity of other genera (except Siphateles and Gasterosteus) in the fossil record from there (M.L. Smith, 1981) and the absence of Fundulus in the Great Basin today. The presence of fundulids from northeast (Humboldt Formation) and west-central (Truckee Formation) Nevada and Death Valley through most of the Miocene and Pliocene establishes a benchmark of antiquity that is important, given that the ages of origin of Great Basin fish groups are un- derestimated (see "Material and Methods" and the Appendix). COTTIDAE.?The Cottidae, like the Salmonidae, are low- temperature, usually high-gradient fishes. They occupy moun- tain streams and cold lakes. Their fossil record in the Great Ba- sin goes back to the Pliocene and Pleistocene in the Lahontan Basin (C. beldingi) and to the late Pleistocene in the Bonneville Basin (C. bairdi, C. extensus). Their phylogenetic relationships are to the north and west. More species of northwestern NUMBER 33 197 sculpins occur in the northern Bonneville Basin and Bear River drainages than are currently acknowledged in the literature, ac- cording to new discoveries by Shiozawa and others (unpub- lished). Sculpins are not found with cyprinodontoids in western North America, unlike their occasional sympatry in the eastern United States, reflecting the more extreme habitat gradients (temperature and stream profiles) in the west. GASTEROSTEIDAE.?The Miocene three-spined stickleback {Gasterosteus doryssus; Appendix, number 128) occurs in the Great Basin from the Truckee Formation in west-central Ne- vada from Hazen (Bell, 1974) to the Sahwave Mountains (LaRivers, 1964). It represents a fauna that colonized the basin prior to final elimination of the drainage outlet to the Pacific coast by the Sierra Nevada uplift (Bell et al., 1985). The Ne- vada Gasterosteus is the most significant fossil fish lineage in the Great Basin, if not in Cenozoic ichthyology in general, be- cause of the abundant, beautifully preserved samples in annual laminae of diatomite and the detailed studies of these samples by Bell and his students (Bell and Haglund, 1982; Bell, 1994, and references therein). In brief, the study of more than 8000 specimens from a 110,000-year time sequence has shown rap- idly fluctuating but gradual change, with no stasis (Figure 12). Bell (1994) concluded that much of the change can be attrib- uted to selection, but some seems uncorrelated with environ- mental change. Gasterosteus aculeatus aculeatus (Appendix, number 129), the modern, anadromous, fully plated, three-spined relative, was reported from the late Miocene Monterey Formation at Lompoc, California, by Bell (1977), strengthening the hypothe- sis that inland freshwater populations of sticklebacks were in- dependently derived from a widespread anadromous form (Bell, 1979). Comparison of these sticklebacks demonstrates that new species can arise repeatedly, as restricted peripatric and paraphyletic derivative populations, without effects on the ecology, morphology, or long-term history of the widespread, ancestral species. Bell's point demonstrates the logical difficul- ties attending Hennig's rule (1966), which requires changing a 0 10 20 30 40 50 60 70 80 90 100 110 Time ( x 1000 years) FIGURE 12.?Variation in mean number of dorsal spines in Gasterosteus dorysus. Variation shows rapid short-term changes and slow long-term trend (after Bell et al., 1985). species name with each branching event in a clade's history, if the parent species did not change while daughter lineages were produced (Bell, 1979). CENTRARCHIDAE.?Sunfish occupied lakes and streams in western North America from Idaho, Washington, and Oregon to Nevada and northern Utah (Salt Lake group) during the Miocene and Pliocene. The Nevada specimens suggest that the climate during the middle to late Miocene was warm and that the basin floor was at low elevations, but floristic data indicate the floor was at higher elevations (Wolfe et al., 1997). The combination suggests local topographic extremes. There was more diversity during the Miocene than is recognized in the currently defined genera Archoplites and Plioplarchus. An undescribed form ex- isted in the Miocene of Oregon (Appendix, number 130). Mod- ern centrarchids are restricted to Archoplites of the Great Valley, California, and to a radiation of nine genera and more than 30 species in the Atlantic coastal drainages of North America. Discussion FISH BIOGEOGRAPHY AND GEOLOGICAL TIME RELATIONSHIP BETWEEN PHYSIOGRAPHIC AND BIOLOGICAL EVIDENCE.?Hubbs and Miller's search for congruence be- tween physiographic and ichthyological histories is evident throughout their 1948b monograph. They noted, for example, the local morphological distinctiveness of Siphateles, Rhinich- thys, and Cyprinodon as evidence of postpluvial separation of previously uniform gene pools (Hubbs and Miller, 1948b:66, 83, 91). The seemingly parsimonious assumption that faunas were uniform at the time of the last habitat connection has rarely been questioned (Kocher and Stepien, 1997), but even in 1948, it was in conflict with the fossil record of modern fishes, which, when known, extended back beyond Pleistocene events to the Pliocene or Miocene (Hubbs and Miller, 1948b:25-27). Taylor (1960) noted the antiquity of many modern lineages of mollusks and fishes and suggested that they are much older than modern physiography, ice ages, and pluvials. The corol- lary of this suggestion is that the tectonic, volcanic, and glacial changes in the landscape occur more rapidly than the evolving changes in lineages on that landscape. Taylor's principle sug- gests that we cannot assume that the observed amount of spe- cies and subspecies evolution automatically occurred in the in- terval since the most recent possible geological changes and vicariance events. Multiple possible vicariant times should be among the tested alternatives. For example, the assumption that cladogenesis of Holocene birds originated in the Pleistocene was refuted by Klicka and Zink (1997), who demonstrated that numerous supposedly Pleistocene species date back to the Pliocene. Our interpretations of fossil fish specimens combined with DN A analyses and phylogenetic analyses of Holocene species indicate that the ages of clades and vicariance barriers have usually been underestimated, thereby challenging the fre- quent assumption that the most recent possible connection was the operative one. 198 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Phylogenetic estimates for the Cyprinidae (Figures 2, 3), Ca- tostomidae (Figure 5), Coregoninae (Figure 8), Oncorhynchus (Figure 10), and Cyprinodontidae (Figure 11) are presented in the figures with key words for taxon geography to provide background for examining the relative timing of historic hy- drographic connections among basins. Traditional studies (Hubbs and Miller, 1948b; Smith, 1966, 1978; Hubbs et al., 1974; Minckley et al., 1986) and the new phylogenies enable the formulation of hypotheses concerning the relative ages of barriers. According to the phylogenies, moderately old and prominent barriers separated the Great Basin from its neighboring drain- ages to the east and the Sacramento Basin to the west. Phyloge- netic indicators of the barrier between the Great Basin and the Colorado Basin, for example, are the relative antiquity of the nodes between the Gila atraria group and the G. robusta group (Figures 2, 3), between Xyrauchen and Chasmistes, between Catostomus ardens and C. latipinnis, between C. platyrhyn- chus and C. discobolus (Figure 5), and between Oncorhynchus clarki Utah and O. c. pleuriticus (Figures 9, 10). The impor- tance of the Sierra Nevada as a barrier is indicated by the sepa- ration of Catostomus occidentalis and C. microps from C. tahoensis (Figure 5), and by Oncorhynchus mykiss from O. clarki (Figures 9, 10). A strong barrier separated the Bonne- ville and Lahontan Basins along the original line of separation in the Great Basin, as illustrated by the relative distance be- tween nodes, for example, between nodes for Gila and Siphate- les (Figures 2, 3), C. ardens and C. tahoensis (Figure 5), and O. c. Utah and O. c. henshawi (Figures 9, 10). The moderately strong separation of the Lahontan Basin and Colorado Basin drainages is illustrated by the nodal distances between Gila and Siphateles (Figures 2, 3), C. tahoensis and C. latipinnis (Figure 5), and O. c. henshawi and O. c. pleuriticus (Figures 9, 10). The weakest, most recent barriers separated the Lahontan Ba- sin and Death Valley systems, and the subdrainages within the Bonneville, Lahontan, White River, and Death Valley systems (most examples are similar species or subspecies not shown on the phylogenies), but all of these barriers have been breached by headwater-stream captures, indicated by some of the exam- ples below and in Figures 2, 3, 5, and especially Figure 10. Fish distributions suggest that much of the drainage from the area of the Great Basin flowed to the Pacific Ocean at times during the Miocene, Pliocene, and Pleistocene, even though the Basin and Range topography began forming more than 15 Ma. The fish data presented herein indicate that these hydrologic confluences probably had an episodic history that began much earlier than the most recent known connection. MIOCENE.?Geological and paleontological studies indicate that at times during the Miocene, the Death Valley, Lahontan, and Bonneville systems drained westward to the Pacific Ocean (Bell, 1974, 1994; Stokes, 1979; Taylor, 1985; Minckley et al., 1986; G.I. Smith, pers. comm. to G.R.S., 1999). Some Mio- cene biogeographic patterns, however, were dominated by north-south systems (Taylor, 1985) that were congruent with the north-south structures that resulted from convergence of the west coast of North America with the Pacific Plate (Figure 13). Early hydrographic connections occurred between the Humboldt lacustrine system and the western Snake River Plain and southeastern Oregon, according to Stokes (1979) and evi- dence from the distribution of Centrarchidae. Later in the Hemphillian, the Salt Lake group, with a relatively rich fish fauna in the northern Bonneville Basin, was probably con- nected to the eastern Snake River Plain (McClellan, 1977). The fish fauna was related to, but specifically different from, that of the western Snake River Plain (McClellan and G.R. Smith, In press). The presence of Lavinia in the lower Salt Lake group indicates there were earlier connections to the western Snake River Plain and the Pacific coastal drainages. Lavinia is now restricted to the Sacramento drainage. The mixture of lowland fish faunas and upland floras in Nevada during the Miocene (Wolfe et al., 1997) suggests that evolution of regional topo- graphic relief took place subsequently (Orr, 1982; Pierce and Morgan, 1992). LATE MIOCENE TO PLIOCENE.?The southern Lahontan Ba- sin fossils from the late Miocene and Pliocene are indicative of low elevations, but the northern Lahontan Basin fossils are in- dicative of higher elevations. Elevations of the Snake River Plain were low in the late Miocene, but uplift of the eastern Snake River Plain at this time occurred in the path of the Yel- lowstone hot spot (Pierce and Morgan, 1992). A connection from the Mono Lake area of the Great Basin to the San Joaquin River ended when basalt dammed the V- shaped channel near Deadman Pass in the Devil's Postpile quadrangle (Huber and Rinehart, 1965, 1967), causing cessa- tion of river flow 3.2 Ma (Dalrymple, 1964; Huber, 1981). The channel is now at about 8400 ft (2580 m) elevation, having been raised 3120 ft (950 m) by uplift of the Sierra Nevada (Dalrymple, 1964; Huber, 1981), so the estimated elevation of the active channel was 5280 ft (1620 m; G.I. Smith et al., 1983; G.I. Smith, pers. comm. to G.R.S., 1999). The outlet possibly continued from the Great Valley across the Diablo Uplift to de- posit the Monterey Fan (Cole and Armentrout, 1979). Cyprinid and gasterosteid fish connections require lower elevations dur- ing the late Miocene, based upon their present elevational re- strictions. PLIOCENE.?Our phylogenetic data suggest the occurrence of some ancient drainage connections that had not been previ- ously noted (Figure 13). For example, the similarity of mor- phology and DNA (see below) of Lepidomeda from the Virgin River drainage and Snyderichthys from the upper Snake River and the Bear River drainages suggests there were connections from the upper Snake River to the Virgin River-White River drainage (Dowling et al., 2002). Keyes (1917), in a discussion of the origins of Lake Bonneville, proposed that the headwaters of the Snake River were formerly a tributary to the Virgin River. Ives (1948) presented evidence that Lake Bonneville drained from its Escalante arm to Meadow Valley Wash (a trib- utary to the Virgin River) prior to its drainage through Red NUMBER 33 199 35 >?? ? ? ? Miocene '.??' Pliocene -35' 121* ?119* ? Pleistocene in* FIGURE 13.?Paleogeographic map showing major Miocene, Pliocene, and Pleistocene hydrographic con- nections, based upon our current interpretations of age and location and supplemented by the conclusions of Cole and Armentrout (1979), Taylor (1985), Taylor and Bright (1987), Ives (1948), and G.I. Smith (pers. comm. to G.R.S., 1999). Heavy dashed lines and arrows represent captures, overflows, or rivers. See "Discus- sion" and Appendix for the estimated timing of some of the barriers represented by lightweight lines. Rock Pass to the upper Snake River. Hubbs and Miller (1948b) rejected these hypotheses as unsupported by fish evidence. Possible connections require serious consideration now be- cause of new cyprinid and molluscan evidence. Miocene and Pliocene distributions of freshwater snails, namely, Valvata idahoensis Taylor, Lutralimnea dineana (Taylor), L. gentilis Taylor, and Tryoniaprotea (Gould), led Taylor (1985:285, 296, 317) to propose late Miocene and possibly Pliocene connec- tions (not necessarily contemporaneous) from the upper Snake River, through the Bonneville area, to the lower Colorado River drainage. The White River canyon in southern Nevada is surrounded by possible drainage connections, as indicated generally by Taylor (1985, figs. 15, 25), leading to the Bouse Embayment of the Gulf of California in the late Miocene or early Pliocene (Shaffiqullah et al., 1980; Buising, 1990) and in the Pleis- tocene (DiGuiseppi and Bartley, 1991). The Colorado Paleo- river was diverted from its delta in the Los Angeles Basin to 200 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the Salton trough about 9 Ma (Howard, 1996). The deposi- tional elevation of the Bouse Formation (near the Arizona-Ne- vada border) is usually cited as at sea level, constraining the time of uplift of the Colorado Plateau, but this is currently in dispute. Bouse Formation sediments and shells show ratios of strontium isotopes intermediate between values expected for marine and freshwater, indicating this area had a mixture of water sources with different strontium isotopes, according to Spencer and Patchett (1997). These authors' contention that the formation was deposited in a lake, possibly at an elevation of 400-700 m, is rejected on the basis of fossil and biogeo- graphic data; the Bouse Formation contained a fauna of fora- minifera, barnacles, brackish-water fishes, and marine mol- lusks (Metzger, 1968; Smith, 1970) that could not have been supported for 3 My by dispersal on birds' feet, as suggested by Spencer and Patchett (1997). The hypothesis that the Bouse Formation was lacustrine (Spencer and Patchett, 1997) has been tested and, by inference, rejected by the absence of any such marine faunas in saline lakes today (Kidwell in Spencer and Patchett, 1997). Pliocene fluvial connections between the northern pre- Lahontan Basin and the Alvord Basin of Oregon are indicated by geological data (Reheis and Morrison, 1997) and fish dis- tributions (Appendix, numbers 9, 22, 91, 96, 97). DNA diver- gence data (below; Table 3) suggest that reconnections 0.2 Ma allowed trout populations in these basins to once again mix. PLIO-PLEISTOCENE.?Connections among the Mono Lake, Owens River, Death Valley, Mohave River, and lower Colo- rado River drainages are indicated by cyprinodontid fish distri- butions (Miller, 1948) and geological data (Gale, 1915; Morri- son, 1965; G.I. Smith, 1978; Benson et al., 1990). The connection between the Mohave River and the Colorado River basins through Bristol, Cadiz, and Danby Basins lacks support from geomorphology, according to Brown and Rosen (1995), but the connection suggested by present Cyprinodon distribu- tion is supported by sedimentary data (Hale, 1985) and by ma- rine or saline-lake foraminifera from cores in Danby and Cadiz dry lakes (Smith, 1960). The foraminifera occur in green clays that are currently within a few hundred feet above or below sea level. PLEISTOCENE.?Pleistocene connections occurred on a land- scape more like that of the present (Figure 13). Additional evi- dence for exchange between the Lake Bonneville and Virgin River drainages, possibly in the Pleistocene, is seen in Catosto- mus clarki from the headwaters of the Virgin River. These fish have unique predorsal scale and gill-raker counts that increase clinally toward their relatives across the divide in the Bonne- ville Basin (Smith, 1966). Numerous possible stream captures are associated with the major north-to-south valley west of Hurricane Fault (Averitt, 1964; Anderson and Mehnert, 1979) in southern Utah (along Highway 91). The drainage connection between the Bonneville Basin and the Colorado River drainage is also suggested by intergrades between Catostomus platy- rhynchus and C. clarki in Shoal Creek (Smith and Koehn, 1971), which is in the southernmost part of the Bonneville Ba- sin near the outlet proposed by Ives (1948). Small headwater-stream captures are required to explain the broad northern distributions of headwater inhabitants, such as Catostomus platyrhynchus, C. discobolus (Figure 7), Onco- rhynchus clarki subspecies (Figure 9), and Coitus spp. (Appen- dix, numbers 124, 125). Numerous examples of northern con- nections can be recognized on topographic maps and in the field. For example, in Wyoming, two such possible connections may be related to the distribution of Catostomus platyrhynchus, C. discobolus, and Coitus bairdi in the Green, Bear, and Snake Rivers. Willow Creek, formerly tributary to Muddy Creek of the Green River (T.I3 N., R.I 19 W., Ogden sheet, United States Geological Survey (USGS) 1:250,000 series), was cap- tured by the headwaters of the Bear River 11 mi (17 km) south- east of Evanston, Wyoming. Twin Creek, of the Bear River be- tween Fossil Butte and Kemmerer, Wyoming (T.21 N., R.I 17 W., Ogden sheet), apparently captured the head of Hams Fork of the Green River. We hypothesize that these captures could have carried Catostomus discobolus from the Green River drainage to the Bear River, from whence they could have trans- ferred to the Snake River at Soda Springs (Bright, 1967) and to the Weber River in upper Chalk Creek (Hansen, 1969, 1985). MLTOCHONDRIAL DNA EVIDENCE FOR THE TIMING OF PHYLOGENETIC AND HYDROGRAPHIC EVENTS Analysis of DNA divergence data from fishes offers a means of constraining ages of barriers and uplift in the Great Basin. Estimated rates of evolution of mtDNA genes in Cyprinidae, Salmonidae, and Cyprinodontidae (Figure 14) vary from about 0.5% sequence divergence per one million years in salmonids to about 1 % per one million years in cyprinids and cyprinodon- tids. These values are similar to those reported for snails of the marine gastropod genus Nucella (see Collins et al., 1996) and for goodeid fishes (Webb, 1998) but differ from the estimate of Martin and Palumbi (1993). The relative-rate tests (e.g., Table 2) show standard deviations of 10%?25% of the mean values of percent sequence divergence, or roughly 2-5 My for mean ages of 7-38 My (Table 2) for basal branches in the cyprinid data. Standard deviations are 15%?30% in the salmonid data, allow- ing a rough idea of the imprecision of the estimates. These are overestimates of rates of molecular change because the fossil ages in the denominators of the rate calculations are underesti- mates. If the rates are underestimated by 10%?30%, elapsed times inferred from rates of molecular change are expected to be underestimates (by as much as 0.5-1.2 My in 4 My for spinedaces, for example). This means that the connection indi- cated by the Snake River Snyderichthys relationship to the Vir- gin River Lepidomeda (percent sequence divergence = 4%) probably occurred within the range of 4-5 Ma. In general, divergence times based upon the rates calculated herein are old, as predicted by Minckley et al. (1986). Great Basin cyprinids in the sister clade to Phoxinus of eastern North America differ from Phoxinus by an average of 22% NUMBER 33 201 .50 -I 8 .40c CD .30- CD .20- CD CD CO y = 0.011x + 0.05 R2 = 0.28 ? ? Cyprinidae 10 15 Million years before present CD y = 0.0047X + 0.007 R2 = 0.65 4 6 8 10 Million years before present 12 14 CDO c CD > .15 i CD -if CD CT CD CO .05- y = 0.097x + .036 R2 = 0.69 Cyprinodontidae 5 10 15 Million years before present 20 FIGURE 14.?Rates of molecular change in Great Basin and Snake River plain Cyprinidae, Salmonidae, and Cyprinodontoidei. Estimated ages of fossils are given in the Appendix and Table 5. DNA sequence divergences are from Tables 2-5; R2 is an estimate of the percent of the variance explained by the regression. sequence divergence (Table 2), implying that separation oc- curred 22 Ma, or in the early Miocene. Rhinichthys osculus from the Virgin River and R. atratulus of eastern North Amer- ica differ by 22%. Snyderichthys and the Lepidomeda group differ from Semotilus of eastern North America by 20%. Closer relatives across the continental divide generally differ by less. For example, Clinostomus and Richardsonius differ by 12% sequence divergence, and eastern and northern North 202 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 5.?Fossil cyprinid, salmonid, and cyprinodontoid taxa and ages of samples of first appearances used in the estimations for the rates of molecular change in Figure 14. Fossil taxon numbers correspond to num- bers in the Appendix; unnumbered taxa are from outside of the Great Basin. Ages in parentheses are addi- tional fossil records for that taxon. (Fm. = formation.) Fossil taxon CYPRINIDAE 1. Lavinia 2. Lavinia 3. Lavinia 5. Mylopharodon 5. Mylopharodon 5. Mylopharodon? 6. Siphateles 6. Siphateles? 7'. Siphateles 8. Siphateles 9. Siphateles 10. Siphateles bicolor obesa 14. Siphateles sp. 15. Siphateles sp. 23. Gila a. atraria 25. Klamathella 26. Acrocheilus 26. Acrocheilus 27. Gila 28. Gila 29. Gila Gila spp. 30. Ptychocheilus 37. Richardsonius 56. Mylocheilus 57. Mylocheilus 58. Mylocheilus SALMONINAE 87. Salvelinus 88. Oncorhynchus 89. O. clarki (Truckee Fm.) 90. 0. clarki (Ft. Rock Basin) 91.0. clarki (Honey Lake) 92. O. clarki (Carson Sink) 92. O. clarki (Mopung Hills) 92. O. clarki (Humboldt River Canyon) 97. O. clarki (Whitehorse Basin) 97. O. clarki (Sunshine Amphitheater) 98. O. clarki (Sevier Basin) 98. O. clarki (Trout Creek) O. nerka1 O. keta1 O. rastrosus^ O. keta[ CYPRINODONTOIDEI Chapalichthys encaustus2 Ameca splendens2 Alloophorus robustus1 Goodea atripinnis2 Girardinichthys multirad2 Tapatia2 111. Cyprinodon breviradius 113. Empetrichthys erdisi 122. Fundulus lariversi Sister taxon Mylopharodon Mylopharodon Mylopharodon Siphateles Lavinia Eastern minnows Gila Eastern minnows Gila Gila Gila Mylopharodon Gila Gila atraria subsp. Gila Klamathella Klamathella Siphateles Siphateles Siphateles Ptychocheilus Gila Clinostomus Richardsonius Richardsonius Richardsonius Hucho O. clarki + O. mykiss O. mykiss O. mykiss O. mykiss O. clarki O. clarki O. clarki (Snake River) (Sevier River) O. gorbuscha O. gorbuscha O. nerka O. nerka Chapalichthys pardalis Xiphophorus variatus Ameca, etc. Ataeniobus. toweri G. viviparus other Girardinichthyini C. spp. Crenichthys Goodeidae Age (My) 6(3.5,3) 5(3) 6 12 4.5 12 6(6) 15 3 6(6) 3(3) 2 (0.6) 8 5(3,1) 0.01 6 6 7 12 12 12 6 15(6,6,8) 3.5 7 6 12 14 14 6 3.5 3.5 3.5 3 0.7 0.15 1 0.01 0.01 6 6 12 6 0.25 0.25 0.25 0.25 5.0 6.0 5.5 4.0 17 % seq. div. 8.1 8.1 8.1 15.0 8.1 42 11.5 36 11.5 11.5 11.5 3.1 15.0 11.5 0.9 6.2 6.4 6.4 11.5 11.5 11.5 10.3 10.4 12.0 21 21 21 4.8 10 4 4 4 1 1 0.1 0.1 0.1 1.3 1.0 5 3 5 6 0.6 4.7 6.2 9.1 7.1 12.2 2.9 7.0 22 'Data from G.R. Smith (1992a). 2Data from M.L. Smith (1980) and Webb (1998). NUMBER 33 203 American Couesius differ from Lepidomeda by 16% (Dowl- ing et al., 2002), implying that separations occurred in the early to middle Miocene. Great Basin Gila differ from Colo- rado River and Rio Grande Gila by 4%-9% (except for north- ern Gila atraria, G. jordani, and G. cypha, which differ by 0.4%?1.1% sequence divergence, implying there have been relatively recent introgressive transfers of mtDNA; Table 2). Similarly, Great Basin Gila and their relatives in the Colorado River, Rio Grande, Rio Yaqui, and in waterways in nearby valleys in the Basin and Range Province of northern Mexico differ among themselves by 5%?9%, implying lower eleva- tions and more connections existed to the south 5-9 Ma (Ta- ble 2; Dowling, unpublished data). Great Basin Gila differ from Columbia-Snake Acrocheilus and Klamathella by 6% sequence divergence, but Great Basin Siphateles differ from the same taxa by 10%-ll%, implying separation occurred well back in the Miocene. Differences between conspecific populations of Gila and Snyderichthys within the Bonneville Basin range up to 2.4%, indicating subspecific divergence be- gan in the late Pliocene. Populations of Oncorhynchus clarki (cutthroat trout, Table 3) differ from O. mykiss (rainbow trout and redband trout) by about 4% sequence divergence, implying separation occurred about 8 Ma (assuming rates of molecular evolution of about 0.5% sequence divergence per 1 My). Populations of Onco- rhynchus clarki in the Great Basin (except for Deaf Smith Creek and Bear River) differ from populations of the west slope (Columbia River) and the Snake-Yellowstone-Bear Riv- ers by l%-2% sequence divergence, implying separation oc- curred about 4-2 Ma. Bonneville Basin populations are hetero- geneous. The Deaf Smith Creek (northern Bonneville) population differs from all other trouts by about 2% more than do other Bonneville trout populations (indicating rapid change occurred in the Deaf Smith Creek population because it is not cladistically basal, Figure 10). The Bear River (most northern Bonneville) population differs from the upper Snake River population by only 0.3%-0.4% sequence divergence, which implies separation took place about 0.7 Ma, which is roughly consistent with part of the Bear River chronology of Bouchard et al. (1998). Populations of the Bonneville and Lahontan Ba- sins differ from each other by an average of 2% sequence di- vergence, implying they have been separated for about 4 My. Outside the Bonneville Basin, variation within basins is usually less than 0.1%. Lahontan Basin populations differ from those from the Oregon Lakes by only 0.1% sequence divergence, im- plying they have been separated for about 200 Ky. Great Basin pupfish populations are related to Cyprinodon macularius and its relatives of the Colorado River and Mexico (Echelle and Dowling, 1992; Echelle and Echelle, 1993; Figure 11, Table 4). The Owens pupfish, C. radiosus, is a member of the C. macularius clade but not the Death Valley clade; C. ra- diosus differs from C. macularius by 2.7%-2.8% sequence di- vergence. Death Valley pup fishes differ from C. macularius and C. radiosus by about 3% sequence divergence, implying separation occurred about 3 Ma. Cyprinodon fontinalis of Mex- ico may be a member of the Death Valley clade. These relation- ships require multiple connections between the Death Valley system and drainages in Mexico, through the Colorado River drainage in the Plio-Pleistocene. Cyprinodon salinus of the Salt River in Death Valley differs from other Death Valley pupfishes by 1.3% sequence divergence, implying separation took place a little more than 1 Ma. Cyprinodon diabolis and several subspecies of C. nevadensis form a clade whose popu- lations differ from each other by about 0.2%-0.6% sequence divergence, implying separations occurred 0.2-0.6 Ma. Miller (1948:85-86) concluded that the divergences of C. salinus, C. diabolis, and C. nevadensis date from about the middle Pleis- tocene, which is consistent with the molecular differences given above (Table 4). For divergence of some Death Valley populations to have been as rapid as hypothesized by Hubbs and Miller (1948b:82)?as short as 10 Ky?would require un- realistically high rates of sequence divergence. SPECIES DIVERSITY Fishes of the Great Basin are not diverse, except in endemic species and subspecies (Hubbs and Miller, 1948b). Most of these taxa are explainable as relicts of repeated cycles of ex- tinction or introgression during dry periods between Pleis- tocene pluvials. The subspecies are the product of rapid change in morphology in isolated, small populations in rapidly chang- ing environments (Hubbs and Miller, 1948b). But low diversity in a landscape so divided by barriers (permitting divergence and recolonization following divergence) is not consistent with most hypotheses about geographic and climatic control of spe- ciation processes (Hubbs and Miller, 1948b; Mayr, 1963; Vrba, 1993). Great Basin fossil and Holocene fish data indicate, how- ever, that periods of aridity and insularity, coupled with high barriers that prevent recolonization after local extinctions, have eliminated more species than were produced by the numerous stages of isolation; therefore, the processes that cause species divergence have not matched extinction at the spatial and tem- poral scales of isolation that have occurred in the Great Basin. High extinction rates have been predicted for small, isolated populations (MacArthur and Wilson, 1967; Jablonski, 1980). To determine whether speciation has been influenced by these factors, we considered the processes that have operated in Great Basin fish populations in relation to isolation by pluvial cycles. Two unusual kinds of data lend themselves to empirical anal- yses of these two problems. First, the Basin and Range Physio- graphic Province is divided into dozens of pseudo-replicated fish habitats isolated by mountain ranges, making a highly con- strained physiographic region. This permitted us to trace lin- eages with rather well-controlled histories. Second, the region has a more detailed fossil fish record (Figure 1) than is avail- able for most modern fish faunas, and this record is supported 204 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES by a reasonably well-known stratigraphic context (e.g., Benson et al., 1990; Morrison, 1991; Oviatt et al., 1992) and by mam- malian stratigraphy (e.g., Stirton, 1936; Lindsey et al., 1984; Lundelius et al., 1987; Repenning, 1987; Tedford et al., 1987; Woodburne, 1987). These circumstances enabled us to begin to address some of the most difficult problems in evolutionary biogeography: Over what time scales do speciation processes occur? How long does it take for differentiation of races, sub- species, and species to occur in isolation? Are the rates of dif- ferentiation interval-dependent (Gingerich, 1986), i.e., is the concept of a general rate inappropriate (Bookstein et al., 1978; Bookstein, 1986; Bell, 1994)? Consideration of questions about species and their origins requires a species concept that is open to testable hypotheses concerning processes of origina- tion. SPECIES CONCEPTS AND SPECIATION HYPOTHESES.?Three species concepts stand out as most relevant to models of fresh- water fish speciation in the Great Basin: phylogenetic, biologi- cal, and evolutionary. The phylogenetic species concept (Crac- raft, 1983) focuses on diagnosable, monophyletic, and irreducible units that maximize the recognition of biotic diver- sity. Many more phylogenetic species could be recognized than are currently acknowledged in the Great Basin. The phyloge- netic species concept is often rejected because its lower limits are not well constrained (Mishler and Donoghue, 1982) and be- cause many species are not monophyletic. Local demes, such as the subspecies and races of Cyprinodon nevadensis, Siphate- les bicolor, and Rhinichthys osculus, could be recognized as species under this model. Moritz et al. (1995) argued for use of mutually and reciprocally unique differences as species criteria. This concept requires no information about time, monophyly, or reproductive isolation, so it is independent of evolution of reproductive isolation. The addition of reproductive and histor- ical tests to proposed species enables more complete evaluation of speciation models. The biological species concept (Mayr, 1963) recognizes groups of populations that are reproductively isolated from other such groups. The focus on reproductive isolation makes this concept central to most speciation process models. Species so defined often contain geographically variant populations, so that in the Great Basin the concept is realistic and objective but is operational only by inference. This definition is sometimes applied too restrictively to accommodate some Cyprinidae, Ca- tostomidae, and Salmonidae, for which there is evidence of gene flow between historically well-differentiated lineages that diverged several million years ago (Smith, 1992a). The biolog- ical species concept is improved for our purposes by integra- tion with the ecological species concept of Van Valen (1976, 1988), that is, accommodation of limited gene exchange be- tween long-term, ecologically independent lineages. The evolutionary species concept (Simpson, 1951; Wiley, 1981; Mayden and Wood, 1995) requires evidence that lin- eages have separate evolutionary histories, usually diagnosed in a cladistic context. This concept does not require (or rule out) evidence for reproductive isolation as evidence for lineage independence. The evolutionary species concept is applicable to Great Basin fish lineages to the extent that introgressed but still-separate lineages are not excluded because of non-mono- phyly or temporary lapses in independence. The time required for acquisition of genetically based ecolog- ical or reproductive isolation as a guarantor of lineage indepen- dence is a matter of special concern in the study of speciation in the Great Basin because of the large number of diagnosable, but possibly ephemeral, allopatric populations. Western North American fishes show a high degree of allopatry, but they also provide unusually frequent examples of gene exchange be- tween species of long-standing independence (Hubbs, 1955). Great Basin species of minnows, suckers, trouts, pupfishes, and sculpins are especially consistent with species concepts and origination models that require reproductive independence, but not necessarily isolation. Modest introgressive influence does not contradict this definition, if the morphologically and eco- logically different lineages are not amalgamated through hy- brids with high fitness. Reproductive independence and repro- ductive isolation are often assumed to be synonymous and mutually required for species recognition, but in many species, it is worth making a distinction. Evidence for introgression is usually taken to be evidence against independence (by zoolo- gists, not botanists; Wagner, 1983; Rieseberg and Wendel, 1993), but if forms are ecologically different enough to be dis- ruptively selected, introgression does not necessarily reduce two lineages to one. For example, the ecologically distinct pairs Catostomus ardens and Chasmistes liorus of Utah Lake belong to clades that diverged 3-6 Ma. Introgression occurred in the Provo River at the time of the drought in the early 1930s (Miller and Smith, 1981), but morphological stability of C. ardens and eco- logical persistence of (modified) C. liorus indicate sufficient reproductive separateness between these forms to warrant spe- cies rank. In contrast, some pairs of populations of Siphateles bicolor obesus and S. b. pectinifer in the Lahontan Basin failed to stay separate through recent times (Hubbs and Miller, 1948b). These different outcomes are ecology dependent (Van Valen, 1976, 1988) in that the reproductive isolation is main- tained only by spawning-site selection based upon ecological characteristics?depth, current, light, temperature, and sub- strate. Evolution of lineage independence, therefore, is partly a function of ecological context (Smith, 1992a; Smith et al., 1995). The introgression hypotheses of Hubbs and Miller (1948b) and Smith (1992a) suggest a Pleistocene history of al- ternation between habitat-rich pluvials and habitat-poor, arid interpluvials, which could have resulted in cyclic adaptation to special habitats in wet times, followed by selection of morpho- logically generalized phenotypes during times of reduced habi- tat diversity. The result is introgressive destruction of some (but not all) independent lineages. We therefore define species as lineages whose separate his- tory can be reasonably inferred from diagnostic morphological, NUMBER 33 205 ecological, genetic, or behavioral apomorphies. This definition recognizes that species are individuals (Ghiselin, 1974, 1997), and it is meant to be consistent with Ghiselin's general defini- tion (1997:99): "Biological species are populations within which there is, but between which there is not, sufficient cohe- sive capacity to preclude indefinite divergence." Our wording acknowledges extrinsic properties and emphasizes history. It might also be peculiarly ichthyological?D.S. Jordan (quoted by Joseph Grinnell, 1924:225) said that "a species is not merely a group of individuals distinguished from other groups by definable features;" it is "a kind of animal or plant which has run the gauntlet of the ages and has persisted." The intrin- sic bases for maintaining historical lineage separateness may be ecological (different habitats), reproductive (different times or places of spawning), or developmental (different genes for growth), but they are ultimately genetic and controlled by long- term selection (Fisher and Ford, 1947; Wright, 1948). The traits that we focus upon herein are associated with reduced hy- brid fitness because of high mortality or low fertility (Dowling et al., In prep.). Tests of intrinsic reproductive separation are usually inferred from (1) the presence or absence of traits unique to sympatric relatives, (2) apomorphic courtship dis- plays, (3) assortative mating, (4) failure to produce fertile or vi- able offspring, or (5) reduced fitness of hybrid products. Tests of long-term persistence require paleontological or molecular biogeographic data that provide a time context in which to evaluate persistence or stability for a period longer than a post- glacial fragment of a glacial cycle. We suggest that molecular divergence measures can contribute to this test, although a sharp, defining demarcation in time or sequence divergence is not suggested. The criteria presented above provide appropriate evidence for accepting or rejecting species assignment under both the evolutionary and the biological species concepts. These considerations provide a context for analyzing the tem- poral scale of morphological and genetic differentiation in rela- tion to speciation. The above definition and the patterns of Great Basin fishes through geologic time suggest avoidance of the chronospecies concept, which is popular in paleontology but incompatible with cladistic, evolutionary, or biological views of species (Krishtalka and Stucky, 1985; Ghiselin, 1997). Great Basin species are interpretable as individuals that may change through time, sometimes anagenetically, and sometimes through recombination of genes from related populations. In ei- ther case, the individuality of ancestral-descendant lineages through time is paramount (Ghiselin, 1974, 1997), and it seems unhelpful to the study of speciation to change lineage names to indicate anagenetic changes. To name chronological species recognizes only stratigraphic diversity, not species diversity. Similarly, it seems unhelpful to change the name of a sister taxon that has given rise to a peripatric daughter taxon without any change in the morphology of the original population (Bell, 1979). SPECIES FORMATION IN THE GREAT BASIN.?The Great Ba- sin fishes are a product of cycles of expanded and contracted habitats, resulting in cycles of differentiation of local popula- tions but only modest persistence of lineages through time. Here the speciation process involves three steps (Smith et al., 1995): (1) separation of populations by hydrographic or habitat barriers (or microhabitat difference in lakes); (2) accumulation of genetic differences in growth, development, physiology, ecology, or behavior; and (3) accumulation of some minimum combination of genetically based differences sufficient to cause hybrids to suffer reduced fitness. Operationally, the products of these process are recognized when lineage independence seems reasonably supported by reciprocally nonoverlapping molecu- lar, morphological, or ecological apomorphies that indicate an intrinsic basis for lineage separateness. Confidence in and pre- dictions from these hypotheses are subject to tests by paleonto- logical or biogeographic circumstances that indicate persis- tence through more than one environmental cycle. The Cenozoic tectonic and Pleistocene hydrographic history of the Great Basin created hundreds of allopatric isolates with the potential to become species, as demonstrated by Hubbs and Miller (1948b) and Hubbs et al. (1974). We conclude, however, that the same conditions (small isolates) and processes (long isolation and short cycles of habitat fluctuation) that caused the initial differentiation also caused extinction of the majority of potential species. Evidence of this extinction is seen in the de- clining late Cenozoic fish diversity, depauperate local faunas, and absence of the vast species numbers expected from scores of so-called "species-pump" cycles. Examples of surviving vicariant pairs in the Great Basin are nevertheless notable. They include Lepidomeda albivallis and L. altivelis in the White River Valley and in Pahranagat Valley, Nevada; Richardsonius egregius and R. balteatus in the Lahon- tan and the Bonneville-Columbia Basins; Catostomus tahoen- sis and C. fumeiventris in the Lahontan and Owens Basins; Cyprinodon salinus, C. milleri, C. nevadensis, and C. diabolis in the Death Valley system; Crenichthys baileyi and C. nevadae in the White River valley and in Railroad Valley; Empetrich- thys latos and E. merriami in Pahrump Valley and Ash Mead- ows; and Cottus extensus and C. echinatus in Bear Lake and Utah Lake. Most of the other species in the Great Basin have sister taxa outside the basin or have complex distribution pat- terns requiring dispersal. Examples of stages of differentiation similar to the above are found among the subspecies and races of Siphateles bicolor, Rhinichthys osculus, Gila atraria, and Oncorhynchus clarki. These forms are early and perhaps tem- porary examples of vicariant speciation, as described above. Our DNA divergence data and the hypothesis that speciation usually involves the long-term accumulation of multiple ge- netic differences suggest that, on average, species-level differ- entiation (beyond the subspecies and races of Hubbs and Miller) requires a longer time than the duration of the Pleis- tocene. 206 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Lacustrine speciation requires special explanation. Examples include the species of Prosopium in Bear Lake and the Bonne- ville Basin, and Siphateles bicolor pectinifer in the Lahontan Basin. The trophic and lacustrine adaptations of these sympat- ric forms indicate that they might have differentiated in intrala- custrine micro-allopatry (Smith and Todd, 1984) involving evolution of different habitats and spawning times. Indepen- dent evidence suggests that intralacustrine speciation is much more rapid than speciation in most other circumstances (Mayr, 1984; McCune et al., 1984; Smith et al., 1995), possibly be- cause it is driven by competitive selection (Rosenzweig, 1978). Variability also might be enhanced by relaxed selection during lake transgression and population expansion (McCune, 1990). There is a widely cited hypothesis that lacustrine speciation is driven by cyclic allopatry, with low stages that separate allo- patric populations, which diverge, and high stages, which fos- ter dispersal and widespread sympatry (e.g., Mayr, 1963; Mc- Cune et al., 1984). The history of the Great Basin includes numerous pluvial cycles (dozens, if pluvial cycles tracked gla- cial cycles) during the last 3 My (Imbrie and Imbrie, 1979; Benson et al., unpublished data; G.I. Smith et al., unpublished data). The pattern of species abundance and distribution in Great Basin fishes does not conform to the lacustrine fluctua- tion model of lacustrine speciation (Mayr, 1984; Smith and Todd, 1984) because extinction, not speciation, has dominated Great Basin history (Smith, 1978), perhaps because the arid parts of the cycle have been so extreme. The fossil and molecular evidence for an ancient presence (a criterion suggested by Ricklefs and Schluter, 1993) of most of the species in the Great Basin also suggests that extinction has been more frequent than species production. The abundance of races and subspecies in the Great Basin suggests that frequent origin and extinction jointly act on populations with limited ranges. Species with stable habitats or large geographic ranges, by contrast, tend to be long-term survivors (Jablonski, 1980). The pattern of morphological change through time, if viewed on the million-year time scale, would appear to be consistent with the hypothesis that most morphological changes in species occurred at the time of initial divergence (Stanley, 1975; Gould and Eldredge, 1977). Viewed on the thousand-year time scale, we see early morphological response to changing environ- ments, but slow accumulation of characters that may be capa- ble of sustaining long-term individuality. TEMPORAL SCALE OF SPECIES DIFFERENTIATION IN THE GREAT BASIN.?These cases suggest a model in which Great Basin fish differentiation was distributed broadly across pluvial cycles of the last 3 My. Through time, some fraction of the new forms went extinct, and some species and subspecies intro- gressed into relatives during extinction episodes. Some of these diverged again during pluvial and early postpluvial episodes. The net result is abundant geographic variation cycling through time, as seen in the fossil record, combined with low species diversity, as seen in Great Basin fishes today (Miller, 1959). The fossil record indicates that throughout the million-year time scale, net changes have been slow. Geographic variation shows that during the thousand-year time scale, some morpho- logical changes in oligogenic and ecophenotypic traits have been rapid (Hubbs and Miller, 1948b; Bell, 1994). It is possible that bony characters seen in the fossil record are multigenic and evolve more slowly than oligogenic meristic, color, and shape characters that differentiate subspecies. It is also likely that body shape, scale, and fin characters fluctuate, perhaps ecophe- notypically, which is consistent with the random-walk hypothe- sis (Bookstein et al., 1978). Fluctuating variations observed during a time scale of hundreds of years (Bell and Haglund, 1982; Bell et al., 1985; Bell, 1994) and decades support this hypothesis (Figure 12). Small, rapid changes are expected in short-term studies, in contrast to slow changes expected in long-term studies (Gingerich, 1986), in part because genetic variation is consumed by selection (Lewontin, 1974). These differences in amounts of change as a function of time are observed in empirical studies in the laboratory as well as in comparisons of paleontological rates (Gingerich, 1986). Ran- dom fluctuation was not one of the hypotheses considered by early students of Great Basin fishes: Hubbs (1940:211) saw "orderly adjustment under the control of environment" not "aimless wandering of genes through the organic world." Stud- ies of morphological fluctuations in the fossil record (Figure 12; Bookstein et al., 1978; Chernoff, 1981; Bookstein, 1986; Bell, 1994) and complex biogeographic patterns in the Great Basin suggest a model of rapid adaptive response to changing environments but minimal long-term trends because of the high amplitude and periodicity of the fluctuations. Conclusions The Great Basin fish fauna is the most depauperate ichthyo- fauna in North America, despite being situated on a landscape ideal for allopatric speciation. Comparison of morphological subspecies differences (Hubbs and Miller, 1948b; Hubbs et al., 1974) with paleontological studies of populations through time indicates that abundant, newly differentiated forms usually fail to become long-term members of Great Basin biodiversity. This fauna has not significantly diversified despite repeated Pleistocene fluctuations in lake levels. The absence of the di- versity predicted by repeated environmental change leads to two conclusions: (1) extinction of small populations is a more potent long-term force than speciation in fluctuating environ- ments, such as occurred in the Pleistocene Great Basin; and (2) speciation processes are favored in small populations (Wright, 1940a; Mayr, 1963; Gould and Eldredge, 1977), but during a time scale of thousands of years, most of this diversity be- comes extinct if individuals in small populations cannot dis- perse among drainages. A related aspect of these fishes is that western salmonids, ca- tostomids, cyprinids, and cyprinodontids are unexpectedly lacking in intrinsic reproductive isolation among morphologi- NUMBER 33 207 cally and ecologically differentiated lineages, some of which have separate evolutionary histories going back several mil- lions of years (Smith, 1992a). Fluctuations in abundance have caused frequent hybridization (Hubbs and Miller, 1943; Hubbs, 1955). The frequent exchange of genes through introgression appears to have maintained some genetic compatibility, even as it provided variation for adaptations to environmental change. The extreme isolated allopatry and resulting divergence of Great Basin fishes permit some calibration of the chronology of hydrologic history of basins, as sought by Hubbs and Miller. Molecular divergence estimates, in conjunction with ages of fossil ancestors, enable the calculation of DNA divergence-rate estimates in salmonids (0.5% per 1 My) and in cyprinids and cyprinodontids (1% per 1 My). These results permit percent se- quence divergence to be used to calculate a rough chronology of hydrologic barriers between individual basins, groups of ba- sins, and drainages neighboring the Great Basin. The Rocky Mountain barrier between Great Basin and eastern North Amer- ican fishes was crossed at low elevations multiple times be- tween 22 and 12 Ma, according to cyprinid sequence diver- gences. Major connections among Great Basin, Colorado River, Rio Grande, Rio Yaqui, and adjacent drainages in the Basin and Range Province of northern Mexico date from 9 to 5 Ma, based upon cyprinid divergences. Great Basin and Columbia-Snake divergences of cyprinids imply that separations occurred early, 11-6 Ma, but that many reconnections occurred during the in- terim. Two levels of differences between lowland and upland taxa of the Lahontan and Bonneville Basins imply that major confluences existed prior to 12-6 Ma and minor headwater con- nections occurred 4-0.2 Ma. Estimates based upon trout DNA sequence divergences imply that the Lahontan Basin headwa- ters were connected to the Columbia River drainage about 0.2 Ma, but Lahontan Basin populations were connected to each other more recently than 0.2 Ma. The Bear River populations were connected to the Snake River populations 0.3 Ma, but the trouts of the Bonneville and Lahontan Basins were connected 4-2 Ma. Cyprinodontoids of the Death Valley system were con- nected to the lower Colorado River basin and to Mexico several times in the last 2-3 My and were among basins in the Death Valley system during the last 0.1-1.3 My. In summary, paleon- tological and molecular evidence indicates that most Great Ba- sin fish lineages are older than previously thought. Appendix The fish species and subspecies of the hydrographic Great Basin, both fossil and Holocene, are listed below in approxi- mate phylogenetic sequence and are numbered to correspond to references to them in the text, tables, and figures. Forty-one forms known only from fossils are indicated by a dagger (|). An asterisk (*) indicates that the account of the Holocene lin- eage contains a pertinent fossil record. CYPRINIDAE Western Chub Group Mylopharodon-Lavinia-Hesperoleucus Subgroup 1. ^Lavinia sp. (Secret Valley hitch). This small species is diagnosed by pharyngeal teeth with corrugated grinding sur- faces and by dentaries with flared lateral edges (Smith, 1975, fig. 19B). Idadon of the Snake River Plain is herein interpreted as a synonym of Lavinia of the Great Valley ("1" in Table 2). Idadon and Lavinia share the unique features of corrugated pharyngeal teeth and flared dentaries, although these features are more extremely developed in Idadon. Flared dentaries col- lected from elsewhere on the Madeline Plains, northeastern California, have been identified as Lavinia and Acrocheilus by Wagner et al. (1997). The Secret Valley dentaries are more flared than those from the Snake River Plain. Although highly variable, the Secret Valley dentaries are more similar to Acro- cheilus than to Lavinia (Wagner et al., 1997), but they lack the abrupt angles of Acrocheilus. {Acrocheilus teeth are not present among UMMZ specimens from Secret Valley.) The Secret Val- ley fossils are assigned an age of ~3.5 Ma; this assignment was based on their correlation with specimens from the Snake River Plain. Lavinia + Hesperoleucus is the sister group to Mylopharodon in the cladogram (Figure 2). Lavinia and Mylopharodon are known from apparent first appearances on the Snake River Plain about 6 Ma and 4 Ma, respectively; these ages correlate with the sequence divergence value of 8.1% (Tables 2, 5) in Figure 14. Molecular evidence (Table 2) indicates a sequence diver- gence of 15% between Siphateles bicolor and Lavinia. (See number 14, below.) 2. ^Lavinia sp. (Carson Valley hitch). A partial left den- tary from late Blancan sediments (~3 Ma) and three caudal ver- tebrae from early Blancan sediments (~5 Ma) were collected in Carson Valley by T. Kelly. The dentary is 10.5 mm long and represents a fish about 18 cm in standard length (SL). The specimen is diagnosed as Lavinia by the flared gnathal ramus, coronoid process at midpoint of the bone, mesially curved ante- rior end, and position of eight lateralis pores. The caudal verte- brae are dorsoventrally compressed, with a thick, single longi- tudinal ridge on the lateral surface of the centrum. 3. ^Lavinia sp. (Cache Valley hitch). Specimens collected from Cache Valley near Logan, Utah, by J. Stewart Williams of Utah State University, appear to be large Lavinia. This conclu- sion is based upon the slightly corrugated grinding surfaces on 208 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the pharyngeal teeth, the flared dentaries, and the high fin-ray counts. The distinct morphology and the absence of Pliocene indicators with these fossils suggest a pre-Blancan, late Mio- cene age (~6 Ma). McClellan (1977) reported Idadon (herein treated as a synonym of Lavinia) from the Junction Hills fauna of the Salt Lake group, near Fielding, Box Elder County, Utah. The fossils appear to be Hemphillian in age by correlation with adjacent faunas. Cyprinids similar to Lavinia or Mylopharo- don, but not to Gila, have been collected from the Bear River valley near Georgetown, Idaho, by Sue Ann Bilby and have been deposited in the collections of the Idaho State Museum of Natural History. This record supports the age estimate of 6 Ma assigned to the sequence divergence value of 8.1% in Figure 14. 4. Hesperoleucus symmetricus (Baird and Girard) (Califor- nia roach). This species lives in the Great Basin only in the Warner Lakes system, near Adel, Lake County, Oregon. It is most similar to Hesperoleucus sp. in Goose Lake and the Pit River drainage (from where it might have been introduced (Miller, 1959)). Although sometimes placed in Lavinia because of protein and molecular similarities, the morphological differ- ences and fossil record suggest that Lavinia and Hesperoleucus are long-different lineages that share molecular similarity (they differ by only 1.3% sequence divergence) probably because of introgression. 5. *Mylopharodon Ayres (western chubs). Chubs with molariform teeth occurred in the Great Basin and in the Snake River Plain from the middle Miocene through the Pliocene, and they remain in the Great Valley of California. Fossils of this ge- nus are rare in the Great Basin. |Mylopharodon{l) doliolus Smith and Kimmel was a large chub with widely spaced, rounded molariform teeth (Smith et al., 1982, fig. 4E). It was placed in Mylopharodon because of its molariform teeth, elon- gate posterodorsal process of the pharyngeal arch, and general shape of the pharyngeal arch. This form is from the Esmeralda Formation in Silver Peak Quadrangle, Nevada, and may be as old as 12 + 2 My. It lacks the derived characters of all My- locheilus and Mylopharodon (of the Snake River Plain, Great Basin, Sacramento Basin, northeastern California, and south- eastern Oregon) and, therefore, could represent the stem group leading to Mylopharodon and Siphateles. Mylopharodon lived in the Honey Lake basin (collected by Todd Yeoman of Sierra College) and the Snake River Plain during the Pliocene. ^Mylo- pharodon hagermanensis Uyeno first appeared at Hagerman, Idaho, on the Snake River Plain about 4.5 Ma. Mylopharodon conocephalus (Baird and Girard) of the Great Valley (number 5 in Table 2) and Siphateles (of the Lahontan Basin) differ by 10.2% sequence divergence (Figure 2; Table 2). The Miocene fossil could also be the earliest western relative of eastern cyp- rinids, from which Mylopharodon conocephalus differs by 42% sequence divergence (Table 2; Dowling, unpublished data). Siphateles Subgroup 6. ^Siphateles sp. (chub). Chubs with a single row of pha- ryngeal teeth on each arch are widespread in the late Miocene to Holocene fossil record of Nevada. Whole fossils and numer- ous separate bones similar to Siphateles were discovered in Barstovian (middle Miocene, -15 Ma) beds of Buffalo Canyon, Churchill County, Nevada, by Ron and Zachary Jones. These are the oldest known cyprinid bones in the Great Basin, so they provide an estimate (underestimate) of the time of vicariance with the eastern clade of North American shiners, from which Siphateles differs by more than 30% sequence divergence (Ta- ble 2; Dowling, unpublished data). A complete but undescribed specimen similar to Siphateles is known from the Nevada Test Site in southern Nye County, Nevada. It is believed to be Mio- cene in age. Several complete specimens and numerous bones are known from the late Miocene beds of the Virgin Valley For- mation, Humboldt County, Nevada. Unidentified cyprinid fos- sils are known from Pine Nut Ridge (Douglas County, late Hemphillian to Blancan) and from the Coal Valley Formation. The late Hemphillian fossils are estimated to be 6 My old and may indicate the branch point with Gila, which differs from Siphateles by an average of 11.5% (s.d. = 2.2%) sequence di- vergence (Figure 14; Table 2). 7. f Siphateles breviarchus (Cope) (Fort Rock chub). This is the common cyprinid in the Pliocene sediments of Fossil Lake, Lake County, Oregon. It has robust pharyngeals that are similar to those of Mylopharodon. It represents a stage of evo- lution prior to diversification of the Siphateles bicolor com- plex. The date of its divergence from Mylopharodon is prob- ably about 10 Ma based on sequence divergence (Figure 14; Table 2). 8. tSiphateles traini (Lugaski) (chub). This fossil form, from the late Miocene lacustrine beds in Jersey Valley, Persh- ing County, Nevada (Lugaski, 1979), is morphologically somewhat similar to S. bicolor. The zeolite beds in which S. traini occurs suggest that the fish lived in a shallow, saline, al- kaline lake of pH 9, much like Pyramid Lake, Washoe County, Nevada, in which S. b. obesus currently lives (Lugaski, 1979). Michael Bell and others have collected a large cyprinid from the late Miocene part of the Truckee For- mation, near Hazen, Nevada (see number 29, below), with ro- bust proportions and fin-ray counts that suggest a relationship to this group. These late Miocene sites are estimated to be 6 My old (Figure 14). 9. ^Siphateles bicolor (Girard) (tui chub). This polytypic species is widespread in the Lahontan Basin drainage as S. bi- color obesus, and it also extends into Oregon's lake basins as S. b. bicolor, and into the Harney Lake and lower Columbia River basins, where it has been referred to S. b. columbianus. Siphat- eles bicolor is diagnosed by a single row of pharyngeal teeth. Cyprinid bones found in indurated sediments of Pyramid and Winnemucca Lakes and other areas of the Lahontan Basin are probably Gila bicolor obesus. Fossils related to this species are NUMBER 33 209 known from the Pliocene of Honey Lake, Lassen County, Cali- fornia, and Mopung Hills, Churchill County, Nevada, and from numerous Holocene Indian middens reported in the archaeo- logical literature. The Blancan fossils occurred before Pleis- tocene tectonism began at the Honey Lake site, possibly 3 Ma. Samples from the Lahontan Basin, from the Owens River, Mono Lake, Mohave River, and Oregon Lakes basins, and from the Columbia River, analyzed by Dowling, differ from each other by 3.1% of the cytochrome b sequence (Table 2). Siphateles bicolor obesus is widespread throughout the streams and springs in the Lahontan Basin system, where it typically has 10-19 short gill rakers, a 5-4 pharyngeal tooth formula (Hubbs et al., 1974), and distinctive scales. Many un- named, geographically distinct forms, in addition to the named taxa summarized below, vary in gill raker number or morphol- ogy, scale counts, size, proportions, nuptial characters, and color patterns. For example, Sulfur Spring, in Diamond Valley, Eureka County, Nevada, contains a dwarf form with long gill rakers, complete lateral line, long pectoral fins, and distinctive colors; it is included with S. b. obesus because a downstream population has features that are intermediate between it and the Humboldt drainage forms (Hubbs et al., 1974). Hubbs and Miller (1948b) reported distinct subspecies in the valleys of Lake Dixie and Lake Toiyabe, in Big Smokey Val- ley, and in Railroad Valley in central Nevada, south of the Lahontan Basin drainage (Figure 4). Another distinct form lived on the east side of the valley of pluvial Lake Toiyabe in Lander and Nye Counties, Nevada, along with Rhinichthys os- culus lariversi (Lugaski, 1972). (Forms on the west side of the valley were apparently transplanted into the springs around the turn of the century; both populations of S. bicolor obesus are now extinct.) Just about every isolated population of S. bicolor in the Ore- gon Lakes has a distinctive combination of characters involv- ing size, body shape, fin shape, gill raker length and number, and color (Hubbs and Miller, 1948b; Hubbs et al., 1974). Siphateles bicolor eurysoma (Williams and Bond, 1981) was named from the Guano Basin, Nevada. The Catlow Valley chub in Malheur County and the Guano Valley chub in south- eastern Lake County, Oregon, are recognized as Siphateles bi- color eurysoma Williams and Bond (1981) (Hubbs and Miller, 1948b). Siphateles bicolor in the Oregon basins vary from forms with more than 16 gill rakers (Catlow Valley, Guano Lake, and Warner Lakes basins) to forms with fewer than 16 gill rakers (basins of the lower Columbia River, of Harney, Klamath, Fort Rock (Fossil), Alkali, and Summer Lakes, and of Lakes Abert and Chewaucan) (Minckley et al., 1986). 10. ^Siphateles sp. (chub). Pliocene(?) fossil Siphateles from Duck Valley, Washoe County, Nevada, is estimated herein to date from 2 Ma. Specimens like the modern Siphateles of the Lahontan Basin occur in Humboldt Canyon below the Lava Creek tephra, which was deposited 665 ? 10 Ka (Izet, 1981; Izet et al., 1992) and the Brunhes-Matuyama geomagnetic re- versal, 780 Ka (Baksi et al., 1992). Oregon and Nevada speci- mens differ by 3.1% sequence divergence (Table 2). G.I. Smith and colleagues of the United States Geological Survey (USGS) recovered a specimen from the Searles Lake core (R.R. Miller, pers. comm. to G.R.S., 1999). 11. jSiphateles bicolor subsp. (chub). Gobalet and Negrini (1992) reported a distinct but undescribed fossil form of S. bi- color from the Chewaucan Basin of southern Oregon; the form was stated to be more than 10 Ky old. Modern specimens of Siphateles from Oregon and Nevada differ from each other by 3.1% sequence divergence (Table 2). Modern representatives in the Chewaucan, Alkali, and Fort Rock Basins (Figure 4) are similar to each other (Hubbs and Miller, 1948b). 12. * Siphateles mohavensis Snyder (Mohave tui chub). This distinct chub was originally described as a spe- cies from the Mohave River in the valley of pluvial Lake Mo- have. It has eight anal-fin rays, uniserial teeth, and 18-29 gill rakers. Its numerous gill rakers characterize a plesiomorphic condition retained from its lacustrine ancestor (Hubbs and Miller, 1948b). Otherwise, it is similar to tui chubs (Miller, 1973). Fossils related to this form were reported from sedi- ments of China, Manix, and Searles Lakes by Buwalda (1914), Blackwelder and Ellsworth (1936), and Flint and Gale (1958). Numerous isolated bones collected by D.F. Hewett and D.W. Taylor from Lake Manix beds represent this species and have osteological characters like the Siphateles figured by Miller (1973). Teeth and vertebrae recovered by B. Leatham from "Lake Mohave" (also called Soda Lake) north of Baker, Cali- fornia, dated at 20 Ky B.P., are most likely from this species. Late Pliocene tui chub bones associated with Catostomus and Chasmistes from the White Hills near Airport Lake, Inyo County, California, may be the same as the Soda Lake form. 13. *Siphateles bicolor snyderi (Miller) (Owens tui chub). This chub inhabits springs and creeks of the Owens River drainage, Inyo and Mono Counties, California. Small lateral line scales, distinctive lateral radii on the scales, few dorsal- and anal-fin rays, a distinctively shaped pharyngeal arch, and usually 10-14 gill rakers diagnose this form. It is most similar to Siphateles bicolor obesus in the East Walker River, from which Miller (1973) presumed it was derived. Fragments as- sumed to represent the Owens River Siphateles were recovered throughout the Pleistocene Owens Lake core (Firby et al., 1997). Holocene Indian middens along the nearby Alabama Hills in Inyo County, California, contain abundant remains of this tui chub and Catostomus fumeiventris, but no other fishes. 14. ^Siphateles sp., (Mono Basin chub). Pliocene fossil bones from Mono Lake basin and Miocene to Pleistocene sedi- ments of Carson Valley are clearly plesiomorphic Siphateles. The dentaries from both sites are similar to each other and are sufficiently distinct from other Siphateles to be regarded as a single, different species (see number 15, below; to be described by K. Gobalet). The Mono Basin sample of Siphateles now in- cludes hundreds of identifiable specimens. They are similar to Gila crassicauda but are different from other Gila and other Siphateles in the distinct robustness of the anterior ramus of the 210 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES dentary and in the anterior position of the dorsal mental fora- men; the anterior edge of the coronoid process is angled more posteriorly in Gila. The masticatory platform of the basioccipi- tal is similar to Mylopharodon in the obtuse angle of the poste- rolateral processes at the origin of the caudal process; this part of the bone matches Siphateles rather than Mylopharodon. The pharyngeal arches bear a single row of teeth, which are robust, somewhat rounded, and reach large size (to 9 mm). These teeth indicate the fossil species had a larger maximum size than the modern record of 42 cm in total length for this species (Sigler and Sigler, 1987). We estimate the age of these fossils to be about 8 My old (based upon the morphological differences de- scribed here), which places them closer to the branch point with Mylopharodon than with Gila. A large, straight dentary and molariform pharyngeal teeth attributed to Mylopharodon by Miller and Smith (1981) might be large Siphateles with long jaws and rounded teeth. 15. ^Siphateles sp. (Carson Valley tui chub). Abundant Carson Valley materials are known from the early Blancan Buckeye Creek local fauna, the late Blancan Fish Spring Flat local fauna, and the early Irvingtonian Topaz Lake local fauna, Douglas County, Nevada. They are distinctive in their robust and posteriorly angular pharyngeals, somewhat like the Mono Basin forms. The dentary shows a more anterior position of the dorsolateral mental foramen than other forms, except those from Mono Basin. The masticatory plates of the basioccipitals are rounded ventroposteriorly, like fossil Siphateles from the Mono Basin and Humboldt River canyon and from Holocene forms from some Lahontan Basin localities and the Sacramento River drainage. The Carson Valley forms are estimated to have lived 5, 3, and 1 Ma. 16. ^Siphateles bicolor pectinifer (Snyder) (lake min- now). This lacustrine tui chub was described by Snyder (1917a) and placed in the distinct genus Leucidius. It can be dis- tinguished by its long, numerous gill rakers (29-36) and by its usually 5-5 pharyngeal-teeth pattern. It is sympatric with S. obesus (Snyder, 1917a; Hubbs et al., 1974:143-150). Extensive introgression has eliminated recognizable individuals of this form in Eagle Lake and has reduced its distinctiveness in Lake Tahoe, Walker Lake, and Pyramid Lake. The presence of hybrid pharyngeals in ancient Indian middens (Hubbs, 1961; Smith, 1985) is evidence that the parental forms hybridized but main- tained their differences for thousands of years (Hubbs, 1961). Remnant populations of S. b. pectinifer in different, uncon- nected habitats suggests that the ecologically significant differ- ences are genetically based (Hubbs, 1961; Hubbs et al., 1974). 17. Siphateles bicolor vaccaceps (Bills and Bond) (Cowhead Lake tui chub). This unique population was discovered by Hubbs and Miller (1948b) in Cowhead Lake, Modoc County, California. It is characterized by its small size (less than 12 cm SL), numerous, short gill rakers (19-25), and small scales. It is most similar to the unnamed chub in Warner Valley (to which Cowhead Lake was tributary in pluvial times), but that form has longer gill rakers (Bills and Bond, 1980). 18. Siphateles bicolor newarkensis (Hubbs and Miller) (Newark tui chub). This chub is the only native fish in the springs of the main basin of pluvial Lake Newark, in White Pine County, Nevada. It is characterized by having few, short, thick gill rakers, a usually complete supratemporal canal, long pectoral fins, and small size (less than 10 cm SL) (Hubbs and Miller, 1972; Hubbs etal., 1974). 19. Siphateles bicolor euchilus (Hubbs and Miller) (big- mouth tui chub). This is the sister form to S. b. newarkensis, with which it shares many features. It is found in springs of Fish Creek (Little Smokey) valley in the southwestern part of the valley of pluvial Lake Newark (southeast corner of Eureka County, Nevada). It is characterized by relatively large size (10-15 cm SL), especially anteriorly; large fins, especially the pectorals; and large scales (Hubbs and Miller, 1972; Hubbs et al., 1974). 20. Siphateles bicolor isolatus (Hubbs and Miller) (Clover tui chub). This represents one of two species found in Indepen- dence Valley, Elko County, Nevada, in the valley of pluvial Lake Clover. Siphateles b. isolata is small, with intermediate- length gill rakers reduced to about 11. It shares characteristics with S. b. newarkensis and S. obesus of nearby Bishop Creek of the Lahontan Basin drainage (Hubbs and Miller, 1972; Hubbs etal., 1974). 21. Siphateles alvordensis (Hubbs and Miller) (Alvord chub). This species inhabits creeks in the Alvord Basin of Malheur County, Oregon, and adjacent Humboldt County, Ne- vada. It is characterized by uniserial teeth, small, embedded scales with radii in all fields, usually seven dorsal and anal fin rays, and numerous, short gill rakers (16-22) (Hubbs and Miller, 1972). 22. Siphateles boraxobius (Williams and Bond) (Borax Lake chub). This small fish occurs in Borax Lake, Alvord Basin, Malheur County, Oregon. The diagnosis includes a large head, large eyes, small caudal peduncle, and a reduced lateral line. The species shares general characters with S. alvordensis, but it has fewer gill rakers, fewer lateral line scales, and fewer pecto- ral-fin rays, like many dwarf populations of Siphateles (Will- iams and Bond, 1980). Gila-Acrocheilus-Klamathella Subgroup 23. *Gila atraria (Girard) (Utah chub). This variable, abundant species is widespread in the Bonneville Basin and up- per Snake River above Shoshone Falls. It is distinctive in its 2,5-4,2 pharyngeal teeth with grinding surfaces, 9 dorsal-fin rays, 8 anal-fin rays, and 8-16 gill rakers. It is locally divergent and polymorphic in its mtDNA. The Bear River population di- verged from the Bonneville Basin populations more than 2 Ma, well before the complex Late Quaternary interactions between the Bear River and Lake Bonneville (Bouchard et al., 1998). The fossil record of Gila atraria is known back only to the late Pleistocene Black Rock Canyon, Smith Canyon, and Home- stead Cave faunas in the Lake Bonneville drainage basin, Utah NUMBER 33 211 and Nevada (Smith et al., 1968; Mead et al., 1982; Broughton, 2000a, 2000b). 24. Gila jordani Tanner (Pahranagat roundtail chub). This species lives in Ash Springs, Crystal Lake, and (formerly) Hiko Springs, all in Pahranagat Valley, pluvial White River system, Lincoln County, Nevada (Tanner, 1950). It has the unique com- bination of Gila robusta morphology and G. cypha and G. atraria mtDNA (Table 2); this unique combination is diagnos- tic for the species. Gila jordani is sympatric with Lepidomeda altivelis, Rhinichthys osculus, Catostomus clarki, and Crenich- thys baileyi grandis (LaRivers, 1962). It has been regarded as a subspecies of Gila robusta, with which it was presumably con- nected via the pluvial White River during the last 10 Ky (Hubbs and Miller, 1948b). The principle of individuality of distinct forms of hybrid origin, as applied to Gila seminuda (DeMarais et al., 1992), suggests that although Gila jordani has been separated for only 35 Ky, it should be recognized as a spe- cies with a unique history because of its unique combination of apomorphies. Gila (or Moapa) coriacea, the Moapa dace, is a localized rel- ative of Gila robusta (Figure 2) and G. intermedia that inhabits warm springs in the Moapa River, which is the former lower course of the pluvial White River channel in northern Clark County, Nevada. It is not a current member of the Great Basin fauna, but it is geographically peripheral and is a relative of Gila jordani. Moapa was named by Hubbs and Miller (1948b) who considered it to be related to Agosia, Rhinichthys, and Gila. According to the molecular cladogram (Figure 2), it is a mem- ber of the Gila clade. A Pliocene or Pleistocene fossil opercle of the Gila group was collected by G.R. Smith from sediments ex- posed in a roadcut near the town of Moapa, Nevada. 25. *Klamathella sp. (chub). Klamathella fossils are found in Pliocene sediments at Honey Lake, Madeline Plains, and Al- turus, California. The fossils are diagnosed by their Gz7a-like pharyngeal bones with two rows of pharyngeal teeth. Fossils from the Salt Lake group (~6 Ma) at Junction Hills, Box Elder County, Utah, were reported by McClellan (1977); they are similar to fossil Klamathella from the Snake River Plain. Kla- mathella was formerly a subgenus of Gila, which was erected for G. coerulea. We hereby raise it to generic rank on the basis of molecular evidence that Klamathella and Acrocheilus are sister groups (Figure 2). Separation of Klamathella from Gila makes Gila monophyletic. 26. *Acrocheilus alutaceus Agassiz and Pickering (chiselmouth). This is a Snake River and Columbia River species that also lives in the Harney-Malheur Basin. The basin was cut off from the Malheur River, a tributary to the Snake, by rather young volcanics (less than 1 Ma, Luedke and Smith, 1982). McClellan (1977) found Acrocheilus teeth in sediments of the Salt Lake group in the Junction Hills, Box Elder County, Utah. Fossils reported as Acrocheilus from the Madeline Plains in northeastern California (Wagner et al., 1997) may be highly variable Lavinia. In the molecular phylogeny (Figure 2), Acrocheilus is the sister of Klamathella coerulea, and together they are sister to the Gila clade. Acrocheilus and Klamathella differ by 6.4% in their cytochrome b sequence (Table 2); fossils show that the apomorphic jaw-shape of Acrocheilus was well established on the Snake River Plain 7 Ma (Smith et al., 1982), indicating di- vergence from Klamathella (plesiomorphic) occurred prior to that date. 27. "fGila turneri (Eastman) (chub). This fossil form is known from the Miocene Esmeralda Formation of south-cen- tral Nevada. Specimens are also known from slightly older (-12 Ma) lacustrine beds of Stewart Valley, Mineral County, Nevada. It is characterized by a slender caudal peduncle and 10 anal-fin rays. Gila turneri was assigned to Richardsonius by LaRivers (1962), but, based upon the large anal fin, it could be related to Colorado River Gila, which differ from other Gila by an average of about 6% sequence divergence. 28. f Gilail) esmeralda LaRivers (chub). This species from the Miocene (-12 Ma) Esmeralda Formation may be related to Siphateles. It has 8 anal-fin rays, the dorsal fin in advance of the pelvics, and a deep caudal peduncle (Lugaski, 1977, 1979; Smith and Miller, 1985). Gila and Siphateles differ by an aver- age of about 11% sequence divergence (Table 2). 29. f Gila{l) sp. (chub). A large chub, greater than 40 cm long, with 8 anal-fin rays, 9 pelvic-fin rays, long dentaries, and blunt, slightly hooked pharyngeal teeth with no grinding sur- face was collected by Michael Bell from the Truckee Forma- tion southeast of Ferndale, Nevada, in the Hemphillian Quarry (N81) along with numerous sticklebacks (~6 Ma). A similar fish was collected nearby by J. Platt Bradbury from the Eagle- Picher Mine in the Clarendonian section (~12 Ma) of the Truckee Formation (Axelrod, 1956). The Bell specimen has dentaries like those of Gila or Siphateles, but it has extremely short articular angulars, like those of Ptychocheilus or Mylo- pharodon. These forms could be related to Siphateles from Mono Lake and could be near the stem group for Gila, Pty- chocheilus, Siphateles, and Mylopharodon, judging from the shape of the angulars and the dentaries. Siphateles, Mylopharo- don, Ptychocheilus, and most Gila differ from each other by an average of about 10%?12% sequence divergence. Ptychocheilus?Relictus?Eremichthys Subgroup This group is not monophyletic in the analysis from which Figure 2 was drawn, because Ptychocheilus is polyphyletic in that tree. The possibility that Ptychocheilus is paraphyletic is supported by the existence of fossils from 15 Ma bearing Pty- chocheilus apomorphies (Smith et al., 2000), but the issue is in need of further study. 30. * Ptychocheilus oregonensis (Richardson) (northern pike- minnow). This Snake River-Columbia River species also lives in the Harney-Malheur Basin. Ptychocheilus occurred in the Pliocene Mono Basin, Alturus Basin, Madeline Plains, and Secret Valley, California (Wagner et al., 1997). It is known from Miocene (15 to ~6 Ma) and Pliocene (5.5 to 3 Ma) deposits in 212 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the Columbia River and Snake River basins (Smith et al., 1982; Smith et al., 2000). Numerous Miocene to Pliocene sites in Car- son Valley, Douglas County, Nevada (Kelly, 1994), have yielded Ptychocheilus dentaries, quadrates, vertebrae, a tooth, and a neurocranium. During this study a single vertebra was identified from Miocene deposits in Brady Park, Churchill County, Nevada. McClellan (1977) identified long, slender, ca- niniform pharyngeal teeth from the Miocene Salt Lake group in the Junction Hills, Box Elder County, Utah, as Ptychocheilus. Living Ptychocheilus oregonensis differ from their cladistic rel- atives by an average of 11.2% (s.d. = 2.1%) sequence diver- gence (Figure 14; Table 2). The oldest known fossil Pty- chocheilus bears a well-constrained date of 15 Ma from the Columbia Basin (Smith et al., 2000). 31. Relictus solitarius Hubbs and Miller (relict dace). This small fish is endemic to Ruby and Butte Valleys (pluvial lakes Franklin and Gale) and to Goshute and Steptoe Valleys (pluvial lakes Waring and Steptoe); each of the four valleys compose parts of both Elko and White Pine Counties, Nevada (Hubbs and Miller, 1972). Relictus displays the western spring mor- photype: small size (usually less than 10 cm SL), small rounded fins, no barbels, reduced lateral line, and scales not in orderly rows. It also has distinctive, cone-like nuptial tubercles on the head and pectoral fins (Hubbs and Miller, 1972). Ac- cording to mtDNA evidence, it may be cladistically, but dis- tantly related to Eremichthys (sequence divergence = ~15%), but its sequence is phenetically more similar to Gila robusta (sequence divergence = -11%) (Table 2). The data indicate that it is an ancient relict (Figure 2), as concluded by Hubbs and Miller (1972). 32. Eremichthys acros Hubbs and Miller (desert dace). This is a highly localized endemic of Soldier Meadows, in the north- western part of the Lahontan Basin. It is diagnosed by its small size (usually less than 6 cm SL), laterally broad pharyngeal arches, uniserial teeth, and specialized sheaths on the jaws (Hubbs and Miller, 1948b). According to parsimony cladistic analysis of mtDNA data, its closest relatives are Relictus (se- quence divergence = 15%) and Ptychocheilus (sequence divergence = 14%), but it also shares 13% similarity with sev- eral species of Gila. These data indicate that it is an ancient relict, as suggested by Hubbs and Miller (1948b). Snyderichthys-Lepidomeda Group 33. Lepidomeda albivallis Miller and Hubbs (White River spinedace). This species was found in springs and their efflu- ents in the upper White River valley of White Pine and Nye Counties, Nevada (Hubbs and Miller, 1960), but it is now ex- tinct. It differs from its sister species, Lepidomeda altivelis and L. mollispinis, from elsewhere in the Virgin-White drainage, by 0.5 % sequence divergence (Table 2). It lived with Rhinich- thys osculus, Catostomus clarki, and Crenichthys baileyi. 34. Lepidomeda altivelis Miller and Hubbs (Pahranagat spinedace). This localized species, now extinct, lived in the Ash Spring and Pahranagat Lake area of the White River val- ley, in Lincoln County, Nevada (Hubbs and Miller, 1960). It differed from its sister species, Lepidomeda albivallis, by hav- ing (on average) a more oblique mouth, a sharp snout, less dark pigment, and a larger and more pointed dorsal fin with a stron- ger spine. 35. Snyderichthys copei (Jordan and Gilbert) (leatherside chub). This apparently paraphyletic species is endemic to the Bonneville Basin, upper Snake River, and Little Wood River drainages in Utah, Wyoming, and Idaho. Its generic affinities have been ambiguous (Miller, 1945a), but Hubbs and Miller (1960) suggested a relationship to Lepidomeda. Snyderichthys shares mtDNA similarity at the level of 5% sequence diver- gence with the Virgin River Lepidomeda, which provides an important clue to its relationships and to western biogeography. Snyderichthys and the Virgin River and White River Lepi- domeda spp. differ from Lepidomeda vittata of the Little Colo- rado River and from more plesiomorphic members of the pla- gopterin clade (Simons and Mayden, 1997; Dowling et al., 2002) by 8%-9% sequence divergence (Figure 2; Table 2). This pattern of relative similarity in mtDNA indicates the rela- tionships among the northern and southern members of this species are complex (Dowling et al., 2002). The Bear River and Snake River populations are more similar to Virgin River drainage Lepidomeda (5% sequence divergence) than to Sevier River and other Bonneville Basin Snyderichthys (11%-12% se- quence divergence; Table 2), suggesting there has been inde- pendent loss of spines or introgressive transfer of mtDNA in Snyderichthys. This pattern of mtDNA similarity supports other evidence for a former drainage connection between the upper Snake River and the Virgin-White drainage of the Mio-Pliocene Bouse Embayment (see "Discussion"). Addi- tional evidence of this connection is the presence of the Muddy Creek Formation from the Lake Mead area to the southwestern Escalante Desert of the Bonneville Basin, prior to 4 Ma (Spen- cer Reber, pers. comm. to G.R.S., 1999). lotichthys-Richardsonius?Rhinichthys?Mylocheilus Group 36. Iotichthys phlegethontis (Cope) (least chub). This small species is endemic to the Bonneville Basin. It was formerly known from eastern tributaries to the Great Salt Lake south to Snake Valley and Juab County, Utah. It shares a unique mor- phological character, namely retarded closure of the posterior cranial fontanel by the parietal, with Gila atraria (Coburn and Cavender, 1992:346), but it shares its mtDNA haplotype with Richardsonius (sequence divergence 9% with R. balteatus; 18% with G. atraria) (Figure 2; Table 2). Incongruent morpho- logical and molecular characters suggest there has been intro- gressive transfer of either morphology or mtDNA from either Gila atraria or Richardsonius (respectively) to Iotichthys (Fig- ure 2 versus Figure 3). 37. *Richardsonius balteatus balteatus (Richardson) (Co- lumbia redside shiner). This subspecies is characterized by its long anal fin. It lives in the Columbia River drainage and in the Great Basin. In the latter, it inhabits the Silvies River of the NUMBER 33 213 Harney-Malheur Basin (Bisson and Bond, 1971). We infer that it dispersed north from the Columbia Basin postglacially be- cause it is not in northern refugia. Although Richardsonius is a solely western North American genus, it is closest to eastern North American Clinostomus (sequence divergence = 12%; Dowling, unpublished data). There is a fossil record of the R. balteatus lineage from the Pliocene (~3.5 Ma) of the Snake River Plain (Smith et al., 1982). 38. * Richardsonius balteatus hydrophlox (Cope) (Bonneville redside shiner). This subspecies differs from R. b. balteatus in having fewer anal-fin rays. It is found in the Bonneville Ba- sin, in isolated creeks of the Harney-Malheur Basin, and above barrier falls in the Snake River drainage and in the South Fork of the John Day River in Oregon (Bisson and Bond, 1971). The presence of relicts of this subspecies above various barrier falls in the Columbia River drainage suggests that the plesiomorphic morphology of this subspecies represents an earlier form than its sister subspecies, R. b. balteatus. It presumably dispersed after connection of the Snake River to the Columbia River about 2 Ma, but before formation of the falls and other causes of isolation in the late Pleistocene. A fossil right pharyngeal arch of R. balteatus, presumably R. b. hydrophlox, was recov- ered by G.R.S. from Wisconsinan deposits in Homestead Cave, northern Bonneville Basin (Broughton, 2000a). 39. Richardsonius egregius (Girard) (Lahontan red- side). This small shiner and its cohabitant in the Lahontan Basin, Catostomus tahoensis, are the only widespread endemic species of the Great Basin. Their absence from surrounding ba- sins is (weak) evidence against low-elevation drainage connec- tions between the Lahontan Basin and other basins in the last several million years. Richardsonius egregius is abundant in di- verse habitats, but it has not differentiated as much as Siphate- les or Rhinichthys. It differs from its sister species, R. baltea- tus, primarily in its smaller body size and smaller fins. 40. Rhinichthys cataractae (Valenciennes) (longnose dace). This one of the most widely distributed fishes in North America, especially in the eastern and northern parts of the continent. Within the Great Basin, it is found in the northeast- ern Bonneville Basin, specifically the Provo River, and north- ward into Utah, Idaho, Wyoming, and the Harney-Malheur Ba- sin of Oregon. Rhinichthys {Apocope) osculus (Girard) (speckled dace). This is a ubiquitous, polytypic species in the Great Ba- sin and surrounding waters; its subspecies are summarized in the entries to follow. Geographic and ecophenotypic variation is extreme in Great Basin representatives of this species (Hubbs et al., 1974). The general pattern of variation usually trends toward one of two ecological morphotypes. The lotic morphotype has a well-developed frenum and barbels, 2,4-4,2 pharyngeal teeth, small scales, medium-sized fins, and a fusi- form body shape that is adapted to rheophilic life. The lentic morphotype is a reductively evolved form; it has a reduced fre- num, barbels, and lateral line; 1,4-4,1 teeth; small, rounded fins; small body size; and inhabits quiet, isolated waters (Hubbs et al., 1974; Miller, 1984:15). The following subspecies represent only part of the variation in the group. 41. Rhinichthys osculus carringtoni (Cope) (Bonneville speckled dace). This subspecies name applies not only to Bonneville Basin, upper Snake River, and most Harney-Mal- heur Basin populations, but also to the lower Snake River and some Columbia Basin populations. This morphotype is prob- ably an early form of the genus. The features of the lotic mor- photype, listed above, characterize R. o. carringtoni; however, the Harney-Malheur Basin fish are smaller and have reduced barbels, especially in Warm Springs (Bisson and Bond, 1971). Hubbs and Miller (1948b) suggested that the Warner Basin contains three subspecies ofR. osculus. 42. Rhinichthys osculus adobe (Jordan and Evermann) (Sevier speckled dace). This dace of the southern Bonneville Basin is distinctive in its large head, large, horizontal mouth, flat snout, and many pigment differences (Hubbs et al., 1974). 43. Rhinichthys osculus velifer Gilbert (Pahranagat terete dace). This dace inhabits the swift water outflow of Ash Springs, in Pahranagat Valley, Lincoln County, Nevada. The fish is slender, with barbels, a frenum, and large, falcate fins (Miller, 1984:15,17-18). 44. Rhinichthys osculus robustus Rutter (western speckled dace). This name was applied to the form of Rhinichthys os- culus of the Sacramento River, Lahontan Basin, and some asso- ciated drainages by Snyder (1917a) and Hubbs et al. (1974). It is characterized by possession of barbels and usually 1,4-4,1 pharyngeal teeth. Prior to 3.2 Ma, the Sacramento and Lahon- tan drainages were connected by a pass at the head of San Joaquin Canyon in Mono and Madera Counties, California (G.I. Smith, pers. comm. to G.R.S.). The form in Soldier Meadows, Humboldt County, Nevada, shows possible evi- dence of introgression?it usually has the normal tooth formula of 1,4-4,1, but a unique specimen has 0,5-4,0, like sympatric Eremichthys acros (Hubbs and Miller, 1948a: 17). 45. Rhinichthys osculus nevadensis (Gilbert) (Amargosa speckled dace). This fish lives in Ash Meadows, Nye County, Nevada. It is small, with barbels, large scales, an incomplete lateral line, 1,4-4,1 teeth, and no frenum. Its closest relative was Rhinichthys osculus deaconi (Las Vegas dace), which was endemic to Las Vegas Creek and Las Vegas Springs, which were tributary to the Colorado River (Hubbs and Miller, 1948a; Miller, 1984). The Amargosa and Las Vegas drainages are con- nected by the Indian Springs Basin, a large section of which was captured by, and is now a tributary to, the Amargosa River (Gilluly, 1929:682; Miller, 1946a). The similarity of the sub- species R. o. deaconi and R. o. nevadensis indicates a conflu- ence of fish habitat and gene flow occurred between the Las Vegas and Indian Springs Valleys, probably in the pass occu- pied by the old railroad bed and Highway 95 east of Indian Springs. 46. Rhinichthys osculus deaconi Miller (Las Vegas speckled dace). This subspecies (now extinct) was endemic to Las Ve- gas Creek and Las Vegas Springs, which were tributary to the 214 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Colorado River in pluvial times. It was characterized by a uniquely rounded anal fin, large scales, small pectoral fins, 1,4-4,1 teeth, no barbel or frenum, and an incomplete lateral line (Miller, 1984). 47. Rhinichthys osculus reliquus Hubbs and Miller (Grass Valley speckled dace). This form lived in the valley of pluvial Lake Gilbert, Lander County, Nevada, and is probably extinct. It lacked barbels and had a much-reduced lateral line (Miller, 1984). 48. Rhinichthys osculus oligoporus Hubbs and Miller (Clover Valley speckled dace). Clover Valley is in the valley of plu- vial Lake Clover, Elko County, Nevada. Rhinichthys osculus oligoporus lacks barbels and has a reduced lateral line (Miller, 1984). 49. Rhinichthys osculus lethoporus Hubbs and Miller (Inde- pendence Valley speckled dace). This subspecies lives in In- dependence Valley, a tributary to the valley of pluvial Lake Clover, Elko County, Nevada. It was apparently derived from R. o. robustus of the Lahontan Basin drainage (Hubbs et al., 1974). It is small and compressed, lacks a frenum and barbels, and has a reduced lateral line. 50. Rhinichthys osculus lariversi Lugaski (Big Smokey speckled dace). Described as a new species by Lugaski (1972), this form was regarded as a subspecies by Hubbs et al. (1974) and Miller (1984). It lives in the valley of pluvial Lake Toiyabe in Lander and Nye Counties, Nevada. This fish has 62-70 lateral scale rows, 1,4-4,1 pharyngeal teeth, 8 dorsal-fin rays, 7 anal-fin rays, no frenum, and variable development of barbels and lateral line. The Lahontan form ofR. o. robustus is its closest relative; Hubbs and Miller (1948b) hypothesized that R. o. lariversi gained access to the valley of Lake Toiyabe through a stream connection with the valley of Lake Gilbert, on its northern border. The taxon is unusual in that it coinhabits the springs on the east side of the valley with a local form of Siphateles bicolor obesus. In most areas of the western Great Basin only a single form of Rhinichthys or Siphateles exists in a given spring, or Siphateles lives in the spring and Rhinichthys lives below the spring outflow. 51. Rhinichthys osculus moapae (Williams) (Moapa dace). This subspecies was described from the Moapa River, Clark Covmty, Nevada. It is a rheophile aivd has barbels, a fre- num, falcate fins, and silver coloration (Williams, 1978). 52. Rhinichthys osculus subsp. (Meadow Valley Wash dace). This inhabitant of springs in Lincoln County, Nevada, is a large dace with small scales, a complete lateral line, a fre- num, barbels, 1,4-4,1 pharyngeal teeth, and 37-39 vertebrae (Miller, 1984). 53. Rhinichthys osculus subsp. (Pahranagat speckled dace). The subspecies occurs in Lincoln and Clark Counties, Nevada. It has rather large scales, small fins, an incomplete lat- eral line, and reduced number and size of barbels, but it has 2,4-4,2 pharyngeal teeth (Miller, 1984). 54. Rhinichthys osculus subsp. (White River shortfin dace). The dace that lives in quiet waters of the White River valley in White Pine County, Nevada, is an undescribed sub- species with robust body, small rounded fins, no barbels, no frenum, and 2,4-4,2 pharyngeal teeth (Miller, 1984). 55. Rhinichthys sp. (Deep Creek dace). This endemic spe- cies has been collected in the western Bonneville Basin, from Fish Springs to Snake Valley. It differs from other robust Rhin- ichthys osculus in color, body shape, meristics, and 2,4-4,2 pharyngeal teeth (R.R. Miller and G.R. Smith, unpublished data). 56. ^Mylocheilus inflexus (Cope) (Chalk Hills chub). This fossil form of Mylocheilus from Idaho and Nevada is similar to Mylopharodon (Appendix, number 5) in the possession of mo- lariform teeth. The two lineages may have hybridized in the past, according to molecular and fossil evidence (Dowling and Smith, unpublished data). Mitochondrial DNA data indicate that Mylocheilus caurinus (number 58 in Table 2) belongs in the shiner clade with Richardsonius (sequence divergence = 33%, Table 2). Miocene(?) chub pharyngeals from Brady Pocket (Churchill County, Nevada, collected by Charles Dailey) are similar to specimens of M. inflexus from the late Miocene (~7 Ma) Snake River Plain (Smith et al., 1982, fig. 4A). This suggests a Miocene connection existed be- tween central Nevada and the Snake River Plain. 57. jMylocheilus sp. (chub). A pharyngeal arch and teeth of a Mylocheilus, intermediate between M. inflexus and M. ro- bustus, were collected from the Salt Lake group at Junction Hills, Box Elder County, Utah (McClellan, 1977). Correlation with other Mylocheilus suggests that the Junction Hills fauna is Hemphillian in age. The presence of this form and fossil Ameiurus in the fauna agrees with molluscan evidence for a Miocene connection and subsequent isolation between waters in northern Utah and the western Snake River Plain (Taylor, 1966). 58. ^Mylocheilus robustus (Cope) (baby-jaw chub). This species was described from late Miocene and Pliocene pharyn- geal arches from the Snake River Plain; the arches are large and have molar teeth. Similar arches with molariform teeth are in beds dated from the late Clarendonian to Hemphillian in the Virgin Valley Formation, Nevada. A single tooth fragment from the Barstovian to Hemphillian age in Esmeralda Forma- tion equivalent beds of Stewart Valley, Nevada, may represent this species (but see Mylopharodon from the Esmeralda Forma- tion). Mylocheilus robustus is replaced in the modern Snake River and Columbia River drainages by Mylocheilus caurinus. CATOSTOMIDAE 59. fDeltistes sp. (Fort Rock Lake sucker). Rare maxillae and dentaries from Pliocene localities at Fossil Lake, Fort Rock Basin, Oregon, have characters that separate these fish from the abundant Chasmistes in the basin. The maxilla has a posteriorly directed premaxillary process and a general shape similar to Deltistes luxatus of Klamath Lakes basin. The ventral keel of the maxilla is prominent anteriorly; the dorsal keel is weak, NUMBER 33 215 with a long convex slope anteriorly. The dentaries are shorter and more robust than those of the Chasmistes clade, with the gnathal ramus deflected mesially and lacking a ventral concav- ity. An almost complete Weberian apparatus, similar to that of the Deltistes clade of the Snake River Plain, was collected by Sue Ann Bilby from sediments of the Salt Lake group, near Georgetown, Idaho. 60. * Chasmistes cujus Cope (cui-ui). This is a large plank- tivore in Pyramid Lake that formerly occupied Lake Lahontan and its postpluvial remnants. It has the most specialized mor- phology among the Holocene species in the genus. Fossil den- taries, vomer, pterotic, and vertebrae from sediments of the Humboldt River canyon confirm the presence of Chasmistes in the pre-Lahontan system in the middle Pleistocene. Pleistocene fossils were collected from near Fallon, Nevada, by J.R. Al- corn. A similar form is found in Pliocene sediments at Secret Valley and Honey Lake, California (see below; Taylor and Smith, 1981). Chasmistes cujus is known from widespread Lahontan Basin Indian middens, such as those in Humboldt Cave, Lovelock Cave, Fishbone Cave, and Hidden Cave, Ne- vada, and in Karlo, California (Hubbs and Miller, 1948b:41; Follett, 1963, 1967, 1980; Miller and Smith, 1981; Smith, 1985). These specimens were probably transported by Ameri- can Indians from Pyramid and Walker Lakes. 61. ^Chasmistes liorus Jordan (June sucker). This species is now endemic to Utah Lake, but it was probably a Pleistocene inhabitant of Lake Bonneville. It is apparently not closely re- lated to Chasmistes muriei, the form that survived in the upper Snake River (Miller and Smith, 1981). Chasmistes that were not diagnostic to species were collected from the Thatcher Ba- sin near Preston in southeastern Idaho by R.C. Bright (1967). Chasmistes fossils were collected by McClellan (1977) from sediments of the Salt Lake group at Junction Hills, Box Elder County, Utah. Chasmistes muriei of Jackson Hole, Wyoming, was native to the upper Snake River drainage, where one specimen was col- lected in the 1930s, but it is now apparently extinct. The mor- phology of C. muriei is similar to that of C. batrachops (ex- tinct, Fossil Lake) or C. brevirostris (Klamath drainage). Pliocene fossils from the Teewinot Formation, upper Snake River drainage in Teton County, Wyoming, are similar to C. muriei in the Snake River and to C. brevirostris in the Klamath Basin (Miller and Smith, 1981). 62. ^Chasmistes batrachops Cope (Fossil Lake suck- er). This sucker was described from Pliocene or Pleistocene sediments of Fossil Lake, Fort Rock Basin, Oregon. Its dentary and maxillary osteology is generalized. The name may be ap- plicable to similarly generalized bones from the Pliocene Glenns Ferry Formation of Owyhee County, Idaho (not Chas- mistes spatulifer); Mono Basin, Mono County, California; China Lake and White Hills, Inyo County, California; and Duck Valley, Washoe County, Nevada (Miller and Smith, 1981). Chasmistes bones are variable in Fossil Lake, perhaps because Pliocene and Pleistocene remains might be mixed as lag clasts on deflation surfaces in the basin. 63. ^Chasmistes sp. indeterminant (Mono Basin lake sucker). This lake sucker is found in Pliocene sediments of the Mono Basin, California, with Ptychocheilus, Siphateles, Catostomus, and Pliocene mollusks (Taylor, 1966). Skulls from Mono Basin resemble skull bones of C. batrachops from Fossil Lake, Oregon. Dentaries and maxillae from Mono Basin are similar to those from Secret Valley, California. Dentaries are similar to those of C. batrachops from Fossil Lake (UMMP 74488) in the medially flattened ridge, in the depth of the groove between the gnathal ridges, and in the anterior recess caudal to the symphysis and gnathal ridges. The position of the median mental foramen and the end of the Meckelian groove are similar to those of C batrachops from the Snake River Plain (UMMP 72458). Maxillae are similar to those from Fos- sil Lake (UMMP 774490) in the narrow, pointed dorsal keel. 64. ^Chasmistes sp. (Carson Valley lake sucker). An early Blancan form from Carson Valley, Douglas County, Nevada, is known from dentaries and a maxilla; late Blancan fossils in- clude an opercle, dentaries, and a vomer; an Irvingtonian den- tary extends the record into the Pleistocene. The axial ridge of the vomer is broad and flat, as in Chasmistes and Xyrauchen, but lacks the postrostral constriction of Catostomus. The den- taries are more slender than those of Catostomus but are similar to those of Chasmistes. The medial ridge posterior to the sym- physis resembles that of C. batrachops, but it is less pro- nounced than those of the Mono and Lahontan Basins and those of C. spatulifer of the Snake River Plain. 65. ^Chasmistes sp. (Honey Lake sucker). This species was collected from Pliocene sediments of "The Island" of Honey Lake by Taylor and Smith (1981) and Charles Dailey. The Honey Lake sucker is intermediate between C. cujus of the Lahontan Basin system and C. spatulifer of the Snake River Plain (Miller and Smith, 1981). It is found with Oncorhynchus cf. "flacustris, Gila, Siphateles, and Catostomus. Similar bones are known from Duck Valley, Washoe County, Nevada, and from the Alturus Basin, the Madeline Plains, and Secret Valley of northeastern California. 66. * Catostomus ardens Jordan and Gilbert (Utah suck- er). This species is native to the Bonneville Basin drainage in Utah and to the upper Snake River drainage in Utah, Idaho, and Wyoming. It is an old lineage that is related to Catostomus macrocheilus of the lower Columbia drainage and to Catosto- mus latipinnis of the Colorado River drainage (Smith, 1992b; Figure 5). During the drought in the early 1930s, Catostomus ardens hybridized with Chasmistes liorus on spawning grounds in the Provo River, Utah, resulting in introgression that is still seen in Chasmistes liorus specimens (Miller and Smith, 1981). Late Pleistocene fossils were reported from the Hot Springs lo- cality, Salt Lake County, Utah (Smith et al., 1968); from Homestead Cave, Tooele County, Utah (Broughton, 2000a), and from the Old River Bed between northern Bonneville and Sevier Basins, Millard County, Utah (Oviatt, 1984). A small, 216 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES nearly complete Weberian apparatus, similar to that of Catosto- mus ardens, was collected from sediments of the Salt Lake group near Georgetown, Idaho, by Sue Ann Bilby and has been deposited in the Idaho State Museum of Natural History. 67. *Catostomus macrocheilus Girard (largescale suck- er). This Holocene species was a lower Columbia River (Rin- gold Formation) sucker in the Pliocene; fossil evidence indi- cates that it spread to the Snake River basin in the late Pliocene (~2.8 Ma), after capture of the Snake River at Hell's Canyon (Smith et al., 2000). It spread north of the Columbia as the gla- ciers receded. It persists in the Harney-Malheur Basin as a relict of the former Snake River outlet through Crane Gap or Malheur Gap. Fossil ages indicate that the colonization of the Harney-Malheur Basin could have occurred after the species immigrated south through Hell's Canyon and prior to (or even after, until about 9 Ka) occlusion of the outlet to the South Fork of the Malheur River by the Voltage lava flow (Piper et al., 1939), 0.04-1 Ma (Gehr and Newman, 1978; Luedke and Smith, 1982). 68. ^Catostomus cristatus Cope (sucker). This fossil spe- cies is best known from the Pliocene of the Snake River Plain. A complete maxilla was collected by Charles Dailey in de- formed Pliocene sediments on "The Island" of Honey Lake. The species is diagnosed by the slender anterior neck of the maxilla in the vicinity of the ventral maxillary muscle, anterior to the dorsal and ventral crests (Taylor and Smith, 1981, pi. 19: fig. 3) and the long, gently curved gnathal ramus of the dentary (Taylor and Smith, 1981, pi. 19: figs. 5, 6). 69. ^Catostomus sp. (Secret Valley sucker). This form dis- plays characters of C. cristatus (Pliocene) and C. macrocheilus (Holocene) of the Snake and Columbia Basins, respectively. As in C. macrocheilus, the dentaries are short, relatively straight, and wide anteriorly, with strong ridges for the labial cartilage (Smith, 1975, fig. 121). The maxillae have medium-high dorsal and ventral ridges and a general shape like those of the Wall Canyon Catostomus; however, the ridge for insertion of the pa- latine ligament (the anterior edge of the bone), although mostly vertical, has a small, horizontal, J-shaped curve that resembles the condition in C. shoshonensis (Smith, 1975, fig. 12A and points xpy on fig. 12G). McClellan (1977) collected similar sucker bones from the Salt Lake group at the Junction Hills, Box Elder County, Utah. Suckers from the Alturus Formation may be correlated with the Secret Valley form (Wagner et al., 1997). Age of the Alturus beds is correlated with the Pliocene Glenns Ferry Formation by a tephra at Crowder Flat Road near Alturus, California, which bears the chemical signature of a Pliocene ash of the Glenns Ferry Formation (Swirydczuk et al., 1982). 70. f*Catostomus shoshonensis Cope (Shoshone suck- er). This species (including its synonym, C. ellipticus) is from the Miocene and Pliocene Snake River Plain (Smith et al., 1982) and from Pliocene sediments at Honey Lake, California (Taylor and Smith, 1981). Catostomus shoshonensis is diag- nosed by an elliptical maxilla with the attachment of the pa- latine ligament high on the neck of the anterodorsal process (Smith, 1975, fig. 12A). This form is represented by the relict sucker of Wall Canyon, Surprise Valley, Nevada; this conclu- sion is based upon the shape of the maxilla. 71. Catostomus warnerensis Snyder (Warner sucker). This species lives in the Warner Valley in eastern Lake County, Ore- gon, and is similar to Catostomus shoshonensis of the Snake River Plain. It is also osteologically similar to the form in Wall Canyon, Surprise Valley, Nevada. 72. *Catostomus tahoensis Gill and Jordan (Tahoe suck- er). This Holocene sucker lives only in the Lahontan Basin. Fossils have been found in Pliocene sediments at Mopung Hills, Churchill County, Nevada (Taylor and Smith, 1981), in middle Pleistocene, pre-Lahontan sediments of the Humboldt River canyon, and in Holocene Indian midden material in the western part of the Lahontan Basin (Thomas, 1985:172, and references therein). The maxilla has a thick anterior segment in the vicinity of the ventral maxillary muscle attachment, with low dorsal and ventral keels that are similar to maxillae oiPan- tosteus (mountain suckers). The dentary has a mesially in- flected gnathal ramus suggestive of the plesiomorphic moun- tain suckers. Catostomus tahoensis was a basal member of the Pantosteus clade in the cladogram of Smith and Koehn (1971). This is supported by the maxillary shape of Holocene and fossil maxillae, which are Pantosteus-like (Taylor and Smith, 1981, pi. 19: 1, 2). Additional material representing C. tahoensis was collected by Marith Reheis (USGS) from sediments deposited ~1 Ma at Sunshine Amphitheater (northwest of Walker Lake) and Lone Tree Mine, Nevada, and by J.R. Alcorn from near Fallon, Nevada. 73. * Catostomus fumeiventris Miller (Owens sucker). This species is osteologically similar to C. tahoensis, except in de- tails described by Miller (1973). It is less derived than the Mopung Hills form of C. tahoensis. Fragments of Catostomus fumeiventris{l) were recovered from the Pleistocene Owens Lake core (Firby et al., 1997). Holocene Indian midden materi- als from along the Owens River near the Alabama Hills in Inyo County, California, are rich in remains of the Owens sucker and Siphateles bicolor snyderi. 74. ^Catostomus sp. (Mono Basin sucker). This Pliocene fish lived with Chasmistes sp., Ptychocheilus sp., and Siphate- les sp. Its dentary has an abrupt anterior edge of the coronoid process, which is similar to that of Catostomus fumeiventris of Owens Valley (Miller, 1973) and to C. tahoensis (see Reheis et al., 2002). The maxilla is similar to that of Catostomus crista- tus of the Miocene and Pliocene Snake River Plain, Idaho (Smith, 1975, fig. 12G), and Secret Valley, California. 75. ^Catostomus sp. (Carson Valley sucker). Pliocene sedi- ments in Carson Valley contain remains of a Catostomus asso- ciated with Siphateles, Ptychocheilus, Lavinia, Chasmistes, and Oncorhynchus. The dentary of this form is more similar to C. occidentalis than to C. tahoensis or C. fumeiventris in the weak peripheral labial ridge and the modest scalloping of the lateral surface of the precoronoid ramus. These characters sug- NUMBER 33 217 gest a pre-Lahontan connection existed that allowed popula- tions of C. occidentalis of the Sacramento River drainage to mix with the Carson Valley sucker of the pre-Lahontan drain- age. According to G.I. Smith (pers. comm. to G.R.S., 1999), such a connection existed at the head of the San Joaquin Can- yon prior to being blocked by volcanics 3.2 Ma. 76. *Catostomus columbianus (Eigenmann and Eigenmann) (bridgelip sucker). This species of mountain sucker (subge- nus Pantosteus) is native to the Columbia Basin and further north, but within the Great Basin it is restricted to the Har- ney-Malheur Basin. It normally has small scales, but in the Harney-Malheur Basin it has large scales and similarities to isolated headwater populations found in the Columbia-Snake Basin (Smith, 1966; Bisson and Bond, 1971). Catostomus co- lumbianus is represented in the Pliocene of the Snake River Plain. Miller and Smith (1981) described the Pliocene speci- mens as a new species, C. arenatus, but, because no evidence of lineage splitting has been found, there is no need for the fos- sil name. 77. Catostomus discobolus Cope (bluehead sucker). This species of mountain sucker (subgenus Pantosteus) arose in the upper Colorado River drainage, but it is also native to the We- ber River, Bear River, and upper Snake River drainages (Figure 7; Smith, 1966). It is the sister species to C. clarki of the lower Colorado River drainage; the two species are the most morpho- logically apomorphic in the genus. Catostomus discobolus is the most widespread mountain sucker in the upper Colorado River drainage, indicating this was the area of its origination and subsequent dispersion. Populations spread from the head- waters of the Green River to the headwaters of the Bear River via stream capture at the east end of the Uinta Mountains, per- haps before the upper Bear River was diverted from the Snake River to Lake Bonneville (Bright, 1967; Bouchard et al, 1998). It is reported from Pleistocene deposits in Homestead Cave near Great Salt Lake by Broughton (2000a). Stream capture in the headwaters of Chalk Creek, Summit County, Utah (Hansen, 1969) could have transferred C. discobolus from the Bear River to the Weber River. 78. Catostomus clarki (Baird and Girard) (desert suck- er). This sister species to C. discobolus occupies headwaters of the lower Colorado River drainage below the Grand Canyon. It inhabits Great Basin streams that were formerly connected to the Colorado River in the Pahranagat Valley-White River drainages in eastern Nevada and co-occurs with species of clades Lepidomeda, Rhinichthys, and Crenichthys. It is clinally variable in the Virgin River drainage. Northern populations possess high scale counts similar to those of C. platyrhynchus in the adjacent southern Bonneville Basin (see "Discussion"). 79. * Catostomus platyrhynchus (Cope) (mountain suck- er). This species of the subgenus Pantosteus lives in the Bonneville and Lahontan drainages, but probably early in the Pleistocene it spread north to the Snake and Columbia drain- ages, then east to the Colorado and Missouri drainages (Smith, 1966). Pleistocene fossils are unknown from the Great Basin, but specimens have been collected by J.D. Stewart from the Pleistocene of northwestern Nebraska. It is the sister to the clade of three species, C. santaanae (of the Los Angeles Ba- sin), C. clarki (lower Colorado River drainage), and C. disco- bolus (upper Colorado River drainage). Differences between the Bonneville Basin and Lahontan Basin forms are slight. The population of C. platyrhynchus in Shoal Creek, Washington County, Utah, shows introgressed characters from C. clarki of the Virgin River drainage, possibly as a result of genetic con- tact at the time of the outlet of Lake Bonneville to Meadow Valley Wash in the middle Pleistocene (Ives, 1948). ICTALURIDAE 80. ^Ameiurus hazenensis Baumgartner (Truckee bull- head). This is one of only two Late Cenozoic catfishes of the Great Basin. It is known only from late Miocene sediments of the Truckee Formation, near Hazen, Nevada (Baumgartner, 1982). Ameiurus hazenensis is related to A. peregrinus of the Chalk Hills Formation, Oregon, to A. vespertinus of the Glenns Ferry Formation, Idaho (Lundberg, 1992), to Ameiurus sp. from the Junction Hills, Box Elder County, Utah (see number 81, below), and to Ameiurus reticulatus from the Ringold For- mation of Washington (Smith et al., 2000). 81. ^Ameiurus sp. (Junction Hills bullhead). McClellan (1977) collected spines of a small bullhead, possibly of Hemp- hillian age, from the Junction Hills, Box Elder County, Utah. SALMON ID AE COREGONINAE 82. *Prosopium williamsoni (Girard) (mountain white- fish). This widespread western North American mountain species lives in the Bonneville, Harney, Lahontan, and sur- rounding basins (Figure 8). Fossil scales of Prosopium (not necessarily williamsoni) were taken from the Old River Bed units of Lake Bonneville (Oviatt, 1984) and the Pleistocene part of the Owens Lake core (Firby et al., 1997). The earliest known form of Prosopium, i.e., P. prolixus, occupied Lake Idaho on the Snake River Plain throughout much of the Pliocene (Smith, 1975). Cold adaptations, cladistic relation- ships, and paleogeography indicate that the genus Prosopium is northern in origin. 83. Prosopium abyssicola (Snyder) (Bear Lake white- fish). This whitefish is now endemic to Bear Lake, a part of the Bear River drainage, which is a tributary to Lake Bonne- ville. It is not represented among the abundant Lake Bonneville fish bones examined by Smith et al. (1968) or Broughton (2000a, 2000b). 84. *Prosopium spilonotus (Snyder) (Bonneville white- fish). This species is now a Bear Lake endemic. Prosopium spilonotus existed in late Lake Bonneville (Smith et al., 1968; Broughton, 2000a, 2000b). 218 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 85. *Prosopium gemmifer (Snyder) (Bonneville cis- co). Prosopium gemmifer is a specialized planktivore that is now endemic to Bear Lake, but a similar form also occurred in Lake Bonneville (Smith et al., 1968; Broughton, 2000a, 2000b). The two forms of P. gemmifer differ significantly in bones that reflect the shape of the head and jaws. The Bear River was a tributary to the Snake River prior to the late Pleis- tocene diversion of the Bear River into the Bonneville Basin at Soda Springs (Bright, 1967; Bouchard et al., 1998). THYMALLINAE 86. Thymallus arcticus (Pallas) (Arctic grayling). Grayling have not been previously acknowledged as native to the Great Basin fauna, but they were recognized among fishes collected in 1834 from the Bear River near Georgetown, Idaho, by John Kirk Townsend (1978:89). Townsend was a physician and zo- ologist working with Thomas Nuttall, a British botanist from Harvard University who was en route to the Columbia River. Nuttall was familiar with Great Lakes fauna, which includes grayling and char. Townsend was the first trained zoologist to cross the American continent, collecting specimens and record- ing observations of natural history. Townsend recorded gray- ling with trout and whitefish in the headwaters of the Green River (1978:84 (June 30)) and contrasted them in a way that implies that he and Nuttall knew the difference between gray- ling and whitefish: "The river, here, contains a great number of large trout, some grayling, and a small, narrow-mouthed white- fish, resembling a herring. They are all frequently taken with the hook." On the Bear River, about 12 miles south (upstream) of Soda Springs, Idaho, on 6 July 1834, Townsend (1978:89) wrote: "Trout, grayling, and a kind of char are very abundant here?the first very large." On the previous day, he recorded: "This is a fine stream of about 150 feet in width, with a move- able sandy bottom." (This was prior to the overgrazing, ero- sion, and destruction of these aquatic habitats that commenced three to four decades later.) Grayling are now native, nearby, to the upper Missouri River in Montana, in streams that were trib- utaries to the Hudson Bay during preglacial and early glacial times (Menely et al., 1957; Howard, 1960; Lemke et al., 1965:15). Grayling had access from the upper Missouri River to the North Platte River, the upper Green River basin (Sears, 1924; Bradley, 1936), the upper Snake River, the Bear River, and the Bonneville Basin through headwater-stream captures northeast of the Uinta Mountains (Hansen, 1969; see "Discus- sion"). There are no Holocene or fossil specimens from the Great Basin in the collections we examined. SALMONINAE 87. *Salvelinus confluentus (Suckley) (bull trout). Town- send (1978), in the account of his (1834) westward journeys, reported "charr" from the Bear River, as pointed out by Ros- tlund (1951). That the upper Bear River was sufficiently cold and stable in 1834 to support grayling, char, and large trout (Townsend, 1978:88; see Thymallus arcticus, above) has rarely been considered. In 1934, Steven D. Durrant, of the University of Utah, discovered a population of S. confluentus in the Jar- bidge drainage, a tributary to the Snake River drainage, north of the edge of the Great Basin (Miller and Morton, 1952; Cav- ender, 1978). Hubbs and Miller (1948b) captured specimens in the Lost River system on the eastern Snake River Plain, an area that shares most of the Bonneville Basin fauna. Bull trout could have colonized the Bonneville Basin from the Snake River while Lake Bonneville was a tributary to the Snake River (Bright, 1967). Salvelinus confluentus might have gained ac- cess to the Snake River via headwater transfer from the Salmon River (Behnke in Cavender, 1978) and then to the Bear River when it was a tributary to the Snake River. Char also could have occupied the Great Basin throughout the late Cenozoic. One premaxilla with diagnostic characteristics of S. confluen- tus was identified by Jack Broughton from abundant fossils in Homestead Cave (Broughton, 2000a, 2000b). We (GRS and KWG) have examined the maxilla and concur in the identifica- tion; we are searching for additional bones to confirm the re- cord. In the middle Miocene, about 14 Ma, Salvelinus was in central Nevada (Churchill County, collected by J.R. Alcorn and Ted Cavender; UMMP57107) and in the Columbia River drainage of Idaho (Smiley and Rember, 1985) as Hucho sp. (Smith and Miller, 1985). It lived in the Snake River Plain in the late Miocene (referred to as "\Hucho larsoni) (Kimmel, 1975; Smith et al., 1982). Native Salvelinus are now extinct in the Great Basin. 88. ^Oncorhynchus cynoclope (LaRivers) (Rabbit Hole trout). This trout was named for Rabbit Hole, Pershing County, Nevada, by LaRivers in 1964. "Salmo esmeralda" was named by LaRivers in 1966; the description was based upon fragmentary material from the late Miocene Alum Mine, Nye County, Nevada, in the upper portion of the Esmeralda Forma- tion. It is not clear how these trouts are related to each other or to a trout from the Miocene Truckee Formation of western Ne- vada. The oldest name is used until more material enables more detailed comparisons. Salmonid bones from Stewart Valley, Mineral County, Nevada, are slightly older than the specimens from the Esmeralda Formation (15-12 Ma, radiometric date); they were found with plants, insects, mammals, and other fishes. A late Hemphillian salmonid dentary was collected by Tom Kelley with mammals from the Coal Valley Formation, Mineral County, Nevada. A trout from Carson Valley, Nevada, is known from vertebrae of individuals about 45 and 54 cm in standard length. The perforated pattern on the centra of the ver- tebrae and the smooth connections of the arches to the centra indicate the vertebrae are different from recent salmon and trout. A dentary of a similar salmonid was collected by Lily Lugaski with cyprinid bones and fossil plants from the Buffalo Canyon Formation (-15 Ma), Churchill County, Nevada. 89. ^Oncorhynchus sp. (Truckee trout). The Truckee For- mation produced a small, nearly complete trout collected by NUMBER 33 219 Michael Bell from a diatomite mine near Hazen, Nevada. The late Miocene diatomite deposit also produced Fundulus, Gila, Ameiurus (Baumgartner, 1982), and a rich record of Gasteros- teus evolution described by Bell (1994) (see number 128, be- low). 90. *Oncorhynchus mykiss gairdneri Richardson (redband trout of the inland Columbia River and Oregon Lakes basins). The redband trout of the Columbia Basin between the Cascade Range and upstream barrier falls are closely re- lated to the redband trout of the Oregon Lakes division of the Great Basin and those of Goose Lake in Oregon and California (Figure 9; Behnke, 1992, map). Unusual variation exists within and among the many populations of the redband subspecies O. m. gairdneri. The fish typically have 130-170 large scales in the lateral line (mean = 135-160), 63-66 vertebrae, zero to many basibranchial teeth, 18-24 (mean = 20) gill rakers, 30-50 pyloric caeca, 58 chromosomes, intermediate-sized spots and parr marks, variable cutthroat marks, and light-colored tips on the dorsal, anal, and pelvic fins (Behnke, 1992). Some of this variation (especially in basibranchial teeth and cutthroat marks) is shared with cutthroat trout, O. clarki. Many popula- tions of redbands were originally regarded as O. clarki (Cope, 1889; Snyder, 1908; Hubbs and Miller, 1948b). In particular, the trout of the Fort Rock Basin frequently have a high basi- branchial tooth count (Behnke, 1992) and are more like O. clarki than O. mykiss, as were their Pliocene and Pleistocene ancestors in the basin (see below). The general trend of red- band characters is, however, less like O. clarki and more like O. mykiss. Cope (1879) and Behnke (1992) identified redband trout populations as O. mykiss. All authors have agreed that the populations in question show composites of cutthroat troat and rainbow trout charac- ters, but each has offered different ideas about whether the mixture was plesiomorphic or the product of ancient or post-in- troduction introgression. Gilbert and Evermann (1894) and Needham and Behnke (1962) regarded redband trout to be the product of introgression. Redband trout have been reported (Behnke, 1992) from the McCloud drainage (northern Califor- nia home of hatchery rainbows), from the Goose Lake drainage of the Sacramento system, and from five desert basins of the "Oregon Lakes system" (Snyder, 1908; Allison, 1940; Hubbs and Miller, 1948:67), namely, the Harney-Malheur, Fort Rock Valley, Catlow Valley, Lake Chewaucan, and Warner Lakes. These are west and north of the basins containing O. clarki, namely, Alvord and Whitehorse Basins (Legendre et al., 1972; Legendre, 1976; Smith, 1978; Behnke, 1992). Hubbs and Miller (1948b) described the close relationships of trouts, min- nows, and suckers to the pluvial lakes in these basins. The Harney-Malheur and Catlow Valley basins (central Har- ney County, Oregon) were occupied by Malheur and Catlow Lakes, respectively, in the late Pleistocene and possibly many times earlier. The Harney-Malheur Basin fishes are derived from two different colonizations, as documented by Bisson and Bond (1971): (1) the John Day River connection with the Co- lumbia River to the north; and (2) the Malheur River connec- tion east to the Snake River, dated to have occurred less than 1 Ma (Luedke and Smith, 1982). The Harney-Malheur Basin possibly received drainage from Catlow Valley in pluvial times, but this neighbor to the south is now isolated (Allison, 1982) and nearly fishless. Catlow Valley lost the heads of its eastern tributaries to capture by Donner and Blitzen Creeks of the Malheur Lake basin on the west side of Steens Mountain in southeast Harney County. The trouts in these two drainage ba- sins are similar, but those from Catlow Valley have larger scales, as expected for the more southern fish in a cline. Behnke (1992) regarded the native inhabitants of these basins to be redband trouts. This conclusion is supported by testimony from older inhabitants of Steens Mountain, indicating that na- tive redband trout were in Home, Three-Mile, and Skull Creeks in the Catlow Basin (west of Steens Mountain) as early as 1917. By contrast, these residents noted, cutthroat trout were native on the east side of Steens Mountain (letter from Stanley Jewett to Carl Hubbs, 13 Jan 1937). The Catlow Valley red- band trout in the UMMZ (collected by S. Jewett in 1936-1937) have 20-22 gill rakers, large, posteriorly concentrated spots, and no basibranchial teeth. Holocene redband trout examined by Behnke from the Fort Rock drainage were variable, with an intermediate frequency of basibranchial teeth that is expected in products of introgres- sion or plesiomorphic character retention. As a consequence of this mixture of characters, we conclude that the Fort Rock trout (ancient and modern) are not cladistically diagnosable as O. clarki or O. mykiss. They possess a mixture of the characters of O. mykiss and O. clarki from the various Oregon lakes. Appar- ently, as O. clarki and O. mykiss were diverging elsewhere in the Pliocene and Pleistocene, a mixture of ancestral characters were only partially sorted out in the Oregon Lakes system, pos- sibly under the influence of gene flow from the north and south. The population remained mixed in the Fort Rock Basin. A relative of f O. lacustris also apparently inhabited the Fort Rock Basin in the Pliocene. The reports of Pacific salmon in the Fort Rock Basin (Hubbs and Miller, 1948b) are based upon bones of redband trout (Allison and Bond, 1983). The redband trouts of the Warner Lakes (pluvial Lake Cole- man; Allison, 1982) and nearby Chewaucan Basin in southern Lake County, Oregon, tend to have slightly fewer vertebrae and more gill rakers, indicating closer relationships to the Goose Lake (Sacramento drainage) redband trout (Behnke, 1992). 91. Oncorhynchus mykiss aquilarum Snyder (Eagle Lake trout). These fish were recognized by Snyder (1917a) and Behnke (1972, 1992) as a population of rainbow trout, Onco- rhynchus mykiss, that possibly transferred from the Pit River drainage to Eagle Lake by a stream capture in Lasson County, California. Hubbs and Miller (1948b) identified O. m. aqui- larum as a hybrid between native O. clarki and introduced O. mykiss gairdneri, redband trout. Fossils resembling Pliocene Oncorhynchus lacustris of the Snake River Plain were collected by Taylor and Smith (1981) 220 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES and by Charles Dailey of Sierra College in deformed Pliocene sediments at "The Island" of Honey Lake, south of Eagle Lake in the northwestern part of what is now the Lahontan drainage. Although the Honey Lake fossils are not necessarily directly related to the Eagle Lake trout, both O. m. aquilarum and O. lacustris have mixtures of O. clarki and O. mykiss characters. The morphological and geographical circumstances are impor- tant because the diagnostic fossil maxillae (round in cross sec- tion) indicate the Pliocene presence of O. lacustris or O. mykiss characters in the area. The Honey Lake sediments also contain two species of Catostomus and one species each of Chasmistes, Mylopharodon, Klamathella (with two rows of pharyngeal teeth), and Siphateles (with a single row of teeth). The exact hydrographic relationship of the highest stages of Eagle Lake to the Lahontan Basin is hidden in several possible shorelines above and below the rim of Eagle Lake basin, which have been summarized by Hubbs and Miller (1948b:37-38). Observations on recent shorelines and connec- tions are not necessarily relevant to the history of these fishes, however. It is possible that the rainbow trouts called O. m. aquilarum are relicts of the ancient redband trout of the Snake River-Sacramento River drainage of the early Pliocene and are part of the series of redband trouts of the Oregon Lakes System, described above. Two Oncorhynchus maxillae from Secret Valley (Madeline Plains, Modoc County, California) sediments (Wagner et al., 1997) establish the presence of a trout similar to the O. clarki-O. mykiss series north of Eagle and Honey Lakes in the Pliocene. A left dentary from Savage Canyon in the Mineral Mountains, Mineral County, Nevada (MVP 58685), most closely resembles O. lacustris. 92. *Oncorhynchus clarki (Richardson) (cutthroat trout). These are inland trout of the Rocky Mountains (Fig- ure 9). They are the only native trout of the Bonneville and Lahontan Basins, as well as of the Alvord and Whitehorse Ba- sins located east of Steens Mountain, Oregon (Figure 9). But as noted above, redband populations west of Steens Mountain and in parts of the lower Columbia River basin possess O. clarki characters in varying amounts. Cutthroat trouts possess more primitive character states than other Oncorhynchus, and their morphological characters are predominant in the older (ancestral) fossil record. This assessment was based upon a cladistic analysis (Stearley and Smith, 1993), which included a trout specimen collected by Michael Bell from the late Mio- cene-Pliocene Truckee Formation (see number 89, above) and other trouts from this portion of western Nevada (LaRivers, 1964, 1966). Oncorhynchus clarki fossils occur in Pliocene strata in the Carson Sink, Churchill County, Nevada, portion of the Lake Lahontan area. Fossil bones of O. clarki were found in Pliocene sediments in the Mopung Hills, in the central part of the Lahontan Basin drainage (associated with Pliocene mol- lusks, Taylor and Smith, 1981); in Savage Canyon, Cedar Mountains, Mineral County, Nevada (UCMP 58685); and in pre-Lahontan sediments in Sunshine Amphitheater northwest of Walker Lake, Nevada, by Marith Reheis (USGS). A large sample of cutthroat trout bones from the Humboldt River can- yon, in sediments below the Lava Creek tephra (665 Ka), is herein diagnosed as O. clarki by the following characters: the anterodorsal strut of the posttemporal is tapered, not spatulate; the median shelf of the anterior portion of the dentary ap- proaches horizontal; and the ascending wing of the premaxilla is high anteriorly. Trout scales were recovered from the Owens Lake core at levels dated at 730 and 695 Ka (Firby et al., 1997). Molecular relationships of O. clarki are discussed be- low (numbers 93-99) and are shown in Figures 10 and 14; in- terpopulation genetic distances are shown in Tables 2 and 3. 93. ^Oncorhynchus clarki henshawi Gill and Jordan (Lahon- tan cutthroat trout). The Lahontan cutthroat trout inhabits Pyramid Lake and its tributaries, except the Humboldt River (inhabited by an unnamed subspecies) and Silver King Creek of the East Fork of the Carson River (O. c. seliniris, Paiute trout) (Behnke, 1992:111). These several populations form a group of three subspecies that are related to the (unnamed) Al- vord Basin and Whitehorse Basin subspecies (see numbers 96, 97, below), according to Behnke (1992) and our molecular data (Figure 10; Table 3). Oncorhynchus clarki henshawi differs from its closest rela- tives in the Lahontan Basin in having large spots over most of the body compared with the Paiute trout, O. c. seleniris, which lacks spots, and it has more gill rakers and smaller scales than the Humboldt River subspecies. The other characters of this trio are similar to those of the Whitehorse Basin and Alvord Basin trouts (Figure 10; Behnke, 1992). The five forms differ among themselves by only 0.1% sequence divergence, accord- ing to our analysis of their mtDNA restriction sites (Table 3). Originally, different spawning runs separated stocks in Pyra- mid Lake (Snyder, 1917a); an additional name, O. c. tahoensis (Jordan and Evermann (1896)), was applied to the "great trout" or "silver trout" from deep waters of Pyramid Lake, but Behnke (1981, 1992) found no morphological differences. Different size and color, if genetic and if characteristic of dif- ferent spawning runs, would be evidence of significant (prob- ably intralacustrine) differentiation. Oncorhynchus clarki henshawi is among Holocene fossils from shore-face sediments around Pyramid and Winnemucca Lakes and from various Indian midden sites reported in the ar- chaeological literature. These records provide documentary evidence and osteological material of the Holocene form for study; otherwise, museums lack adequate specimens from the period before the nineteenth century introductions, hybridiza- tion, and destruction. 94. Oncorhynchus clarki seleniris (Snyder) (Paiute cutthroat trout). This trout is limited to a few headwaters in Silver King Creek of the East Carson drainage in Alpine County, Cal- ifornia. It differs from O. c. henshawi only in the absence of spots. There is no evidence pertaining to the time of its isola- tion, except for the 0.1% sequence divergence (Table 3) from the other Lahontan Basin trouts. We assume that cutthroat trout inhabited the stable environments of the East Carson NUMBER 33 221 drainage and other mountain streams in the Lahontan Basin throughout the Miocene to Pleistocene because of the fossil re- cord of 0. clarki and its ancestors in the basin, extending back to the Miocene. 95. Oncorhynchus clarki subsp. (Humboldt cutthroat trout). Behnke (1992) pointed out the distinctiveness of cut- throat trout of the Humboldt River, relative to those of the Walker Lake, Truckee River, and Carson Valley drainages. The Humboldt cutthroat trout tends to have larger scales and fewer gill rakers, as expected for fish inhabiting warm, fluvial waters. It also has distinctive mtDNA (Williams, 1990, cited in Behnke, 1992), except in the Quinn River, where trout with Humboldt morphology have mtDNA characteristic of trout of the Lahontan Basin drainage. Behnke (1992:119) analyzed possible isolation dates in the context of the main dates of drainage separations, 13-10 Ka (Benson and Thompson, 1987), and he concluded that the divergence of the Humboldt form from O. c. henshawi dates from an earlier time, which is consistent with the 0.1% sequence divergence. The type speci- men of the Pliocene O. cynoclope (LaRivers) is from the range of this subspecies, indicating an ancient local ancestry for the clade, but it lacks sufficient characters to clarify the ancestry of this subspecies. 96. Oncorhynchus clarki subsp. (Alvord cutthroat trout). These trout were collected by Hubbs from Virgin Creek in 1933 and Trout Creek in 1934, in southeastern Har- ney County, Oregon. Hubbs' field notes indicate that hatchery- raised rainbow trout were introduced into Virgin Creek as early as fall, 1933. Behnke (1992) indicated that Alvord cut- throat trout were similar to Lahontan cutthroat trout (0. c. hen- shawi) but had fewer spots, rosier sides and opercles, larger scales, fewer pyloric caecae, and fewer basibranchial teeth (Behnke, 1981, 1992). The rosy color is reminiscent of red- band trout (0. m. gairdneri) (Tol and French, 1988). Mito- chondrial DNA restriction sites indicate that the Alvord trout is the sister subspecies to the Whitehorse trout (0. clarki subsp.) (Figure 10). Protein allele frequencies of Alvord trout (Tol and French, 1988) did not differ from frequencies of 0. c. henshawi at five diagnostic loci, but all of the specimens showed introgressive influence from Oncorhynchus mykiss. Tol and French (1988) suggested that Alvord trout were de- rived from the Lahontan (Quinn) drainage to the south, via Summit Lake, or were introduced. Behnke (1992:123) con- cluded that the Alvord trout was derived from Lahontan cut- throat trout in the late Pleistocene, assuming that a Wiscon- sinan connection was the first to provide access. This is consistent with mtDNA divergence of 0.1% (Table 3). The other fish lineage in the Alvord Basin, Siphateles alvordensis, may have been derived from the Lahontan Siphateles in the Pliocene. It is probable that both species would have used an existing connection; if one colonized earlier, it would probably have been the trout. Older Alvord "Trout Creek" sediments (now elevated) contain Miocene sunfish (Smith and Miller, 1985), but these were deposited in an earlier basin and are un- related to the trout problem. 97. Oncorhynchus clarki subsp. (Whitehorse cutthroat trout). This form lives in the Coyote Lake drainage in south- ern Harney County, Oregon. It bears polymorphic mtDNA haplotypes, namely the Lahontan and Quinn (not Humboldt) haplotype and a unique haplotype (Williams, 1990, in Behnke, 1992). Its mtDNA is similar to that of Alvord trout. White- horse trout share most morphological characters (except fewer pyloric caeca) with Humboldt cutthroat trout and (except for fewer gill rakers) Lahontan trout (Behnke, 1992). Behnke (1992) contrasted the Whitehorse trout with the morphologi- cally dissimilar Alvord trout and determined that the White- horse trout has fewer gill rakers, smaller scales, more pyloric caeca, more basibranchial teeth, and more spots than the Al- vord trout. The two are similar in their mtDNA, however. The distribution of morphological characters, as well as the iso- lated position of the Whitehorse Basin, suggests that the Whitehorse trout could represent the most plesiomorphic form in the Lahontan?Humboldt?Alvord series. The Whitehorse trout and Alvord trout characters require a connection to 0. clarki of the Humboldt River drainage. Ac- cording to sedimentary data (see Minckley et al., 1986:562; Reheis and Morrison, 1997), the lake beds of Alvord and Whitehorse Lakes (1231 m; 4040 ft) would have been inun- dated by one of two middle Pleistocene Eetza highstands of Lake Lahontan (with shoreline altitudes more than 1400 m (4600 ft)). The mtDNA molecules of the five subspecies of 0. clarki from the Lahontan Basin and Whitehorse/Willow Creek drainages are similar to each other at the 0.1% level. The esti- mated rate of divergence (0.5% per 1 My; Figure 14) suggests that these populations were connected 200 Ka. (The most re- cent Eetza highstand is dated as just prior to the 150 Ka Wads- worth tephra (Reheis and Morrison, 1997).) Subsequent char- acter sorting is responsible for the similarities and differences between Alvord trout and Whitehorse trout. Headwater capture has been important to the distribution of this form (Behnke, 1979). Lower Whitehorse Creek and its tributary, Willow Creek, flow north out of the Trout Creek Mountains to the dry lake floor 12 miles east of the Alvord desert. The drainage pat- tern (USGS 1:250,000 series, Adel Sheet, 1955) indicates two captures: (1) the headwaters of Whitehorse Creek were cap- tured by McDermitt Creek (T. 40 S., R. 39. E.), a tributary to the Quinn River (Humboldt River drainage); and (2) the head- waters of Trout Creek (T. 39 E., R. 37 W.) were apparently headwaters of Willow Creek (T. 39 S., R. 37 E.) prior to their capture at Flagstaff Butte by Trout Creek, a tributary to Alvord Lake. The Lahontan-Alvord-Whitehorse clade (Figures 9, 10) is the sister group to the Westslope cutthroat trout, Salmo clarki lewisi, from the Columbia Basin (e.g., Fish Lake, Wing Creek, and Cougar Creek, Idaho), from which they differ by 0.7%-0.9% sequence divergence (implying they were hydro- graphically connected 1.6 Ma). Oncorhynchus c. lewisi is 222 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES represented by relict populations throughout the Columbia River drainage, including as far south as the Boise River drainage near Atlanta, Idaho (UMMZ 234336 with cutthroat trout marks and O. clarki osteology). The distribution and mtDNA of Lahontan Basin and Columbia Basin trouts sup- port an early Pleistocene connection between the Lahontan Basin and Columbia River drainages, as hypothesized by Re- heis and Morrison (1997), but the other species in the two faunas are too different to support a direct spillover from Lake Lahontan to the Snake River drainage. Together, O. c. lewisi and O. c. henshawi and their relatives are the sister group to Oncorhynchus clarki clarki, from which they differ by 1.3%-1.8% sequence divergence, implying separation oc- curred about 3 Ma (Figure 10; Table 3). 98. * Oncorhynchus clarki Utah Suckley (Bonneville cut- throat trout). The Bonneville Basin contains 41 populations of cutthroat trout (Duff, 1988) differentiated into three kinds: Bonneville, Snake Valley, and Bear River (Loudenslager and Gall, 1980; Martin et al., 1985; Behnke, 1992). They were characterized by Behnke (1992) as subdued in coloration, 133-183 (160) lateral scale rows, 33^6 (38) scales above the lateral line, 16-24 (19) gill rakers, 25-54 (35) pyloric caeca, and 1-50 basibranchial teeth in at least half the population (Hickman, 1978; Duff, 1988). The counts are basically those of the Yellowstone cutthroat trout, O. c. bouvieri. Mitochon- drial DNA restriction-site analysis indicates that the Bonne- ville trouts, exclusive of the Bear River forms, are the plesio- morphic sister group to all other O. clarki (Figure 10), differing from them by 1 %-4% sequence divergence (Table 3). The Deaf Smith Creek (Salt Lake County, Utah) population differs from other Bonneville cutthroat trouts in its mtDNA (sequence divergence >2% from other Bonneville cutthroat trouts and 3%-4% from all other cutthroat trouts; Table 3). The Bonneville cutthroat form shares a protein allele with the Colorado River subspecies (Martin et al., 1985; Shiozawa and Evans, unpublished data) and has representatives in Santa Clara River headwaters in Pine Valley in southwestern Utah (see "Discussion"). Fossil cutthroat trout are known from the late Pleistocene Black Rock Canyon locality (shoreline dated at 11,000-13,000 years B.P.) in northeastern Bonneville Basin (Smith et al., 1968); from Homestead Cave, 10,160-11,170 years B.P. (Broughton, 2000a, 2000b); and from the late Pleis- tocene Old River Bed between the northern Bonneville and Sevier Basins (Oviatt, 1984). 99. * Oncorhynchus clarki subsp. (Snake Valley cutthroat trout). The Snake Valley cutthroat trout forms tend to have more spots than other Bonneville cutthroat trouts. Behnke (1992) suggested that trouts were possibly absent from the Bonneville Basin until the late Pleistocene, and Duff (1988) postulated they had an insufficient time of isolation to become a distinct species. We interpret the cladistic primitiveness of the diversity in the Bonneville Basin and the level of diver- gence of the Bonneville cutthroat trout forms to indicate an an- cient, possibly Miocene, presence of cutthroat trout in what became the Bonneville Basin. Fossils from 12 Ka or older have been reported from Smith Creek Canyon in Snake Valley (Meadetal., 1982). 100. Oncorhynchus clarki subsp. (Bear River cutthroat trout). The Bear River form shares proteins and mtDNA with the Yellowstone cutthroat trout, which is consistent with its former connections to the Snake River, between 140 Ka and 80 Ka (Bouchard et al., 1998). The Bear River and Thatcher Basin were most recently diverted to the Bonneville Basin sys- tem about 20 Ka (Bright, 1966, 1967; Bouchard et al., 1998). The Bear Lake relative of the Bear River form has more py- loric caecae, an azure blue color that is dense on the snout but diminishes posteriorly, and yellowish fins (Nielson and Lentsch, 1988). The Bear Lake form is distinct in its life his- tory, reproductive separation, and coloration (Nielson and Lentsch, 1988; Behnke, 1992). The Bear River form is the sis- ter group to the Yellowstone River and Snake River popula- tions (Figure 10), differing from them at the 0.3%-0.4% level in its mtDNA restriction sites (Table 3). These populations dif- fer from O. clarki Utah of the main Bonneville Basin system at the 1.4%?3.5% level, indicating their separate evolution through most of the Pleistocene. CYPRINODONTIDAE 101. Cyprinodon radiosus Miller (Owens pupfish). This species inhabits the Owens River valley, which is a tributary to Owens Lake, Mono County, California. It is the sister species to C. macularius of the lower Colorado River drainage (Ech- elle and Dowling, 1992; Echelle and Echelle, 1993; Figure 11); it differs from that species by 2.8% sequence divergence of mtDNA (Table 4). Cyprinodon macularius and C. radiosus are related to various species of pupfishes from the Death Val- ley system (Miller, 1946a, 1948) and northern Mexico (Figure 11). Cyprinodon radiosus is similar to C. macularius in having seven pelvic-fin rays, dorsal origin at the midbody, and dis- rupted vertical bars in females, but it shares an apomorphic scale structure with the Death Valley species (Miller, 1943b, 1948). It is additionally characterized by having large scales and a large, deep body, like C. nevadensis; its long dorsal and anal fins and long prehumeral length resemble those of C. di- abolis. Male coloration is unique, with amber-colored margins on the dorsal and anal fins and a pale or amber-colored (not black) terminal band on the caudal fin. Cyprinodon nevadensis Eigenmann and Eigenmann (Amargosa pupfish). This is a geographically variable and widespread pupfish in the Death Valley system. It is found in numerous springs from Ash Meadows in Nye County, Nevada, to Shoshone, Tecopa, and Saratoga Springs and other localities along the Amargosa River near the boundary of Inyo and San Bernardino Counties, California (Miller, 1948). The species is large and deep, with large scales, preorbital scales; broad, rounded, incised outer tooth-cusps; six pelvic-fin rays; and the dorsal-fin origin usually behind the midbody. Males are deep NUMBER 33 223 blue with a dark terminal border on the pale caudal fin; fe- males have continuous vertical bars. Protein and mtDNA evi- dence shows the C. nevadensis group (Figure 11) to be related to C. salinus, C. diabolis, and C. milleri, with evidence of pos- sible introgressive homoplasy (Echelle and Dowling, 1992; Echelle and Echelle, 1993). 102. Cyprinodon nevadensis nevadensis Eigenmann and Eig- enmann (pupfish). This subspecies, from Saratoga Springs in southern Death Valley, California, is characterized by its ro- bust body shape; high, narrow scales; numerous radii; and dense reticulations (Miller, 1948). 103. Cyprinodon nevadensis amargosae Miller (pup- fish). This subspecies is from the Amargosa River in San Bernardino County, California, including Death Valley. Small scales with few radii characterize it. Miller (1948) recognized several isolated populations as different races. 104. Cyprinodon nevadensis callidae Miller (pupfish). This subspecies, from Tecopa Hot Springs along the Amargosa River in Inyo County, California, is diagnosed by large scales, narrow interorbital distance, and a short caudal peduncle (Miller, 1948). 105. Cyprinodon nevadensis shoshone Miller (pupfish). This subspecies is from Shoshone Springs, Inyo County, California. It is characterized by a narrow, slender body and relatively straight ventral profile (Miller, 1948). 106. Cyprinodon nevadensis mionectes Miller (pup- fish). This subspecies is from Big Spring, Ash Meadows, Nye County, Nevada. Miller (1948) diagnosed it by its small size, few scales and fin rays, narrow body, posterior dorsal fin, long head, and arched predorsal profile. There are several dis- tinct races. 107. Cyprinodon nevadensis pectoralis Miller (pup- fish). This subspecies is from near Lovell's Spring, Ash Meadows, Nye County, Nevada. Small scales, large number of pectoral-fin rays, and robust body shape characterize it (Miller, 1948). 108. Cyprinodon salinus Miller (Salt Creek pupfish). This isolated species inhabits Salt Creek canyon, northern Death Valley, Inyo County, California (Miller, 1943a). A median ridge on the teeth, slender body, small scales, and scaleless preorbital region characterize it. It is similar to C. nevadensis and C. milleri, with which it shares small fins, especially the pelvics, posterior position of the dorsal fin, small head and eyes, and dark blue color. Miller (1946a) outlined geological evidence that the divergence of C. salinus and C. nevadensis occurred in the interval since isolation of the Salt River by the desiccation of Lake Manly, which was estimated at that time to be about 20 Ky (but see C. milleri, below). It is related to C. nevadensis and is the sister group to C. milleri, according to protein evidence (Echelle and Echelle, 1993). Molecular data indicate that C. salinus is the sister group to the other Death Valley and Ash Meadows species (Echelle and Dowling, 1992; Figure 11). Cyprinodon salinus does not differ from C. milleri in its mtDNA, but these two differ from diabolis + nevadensis by 0.13% sequence divergence, indicating more than 10 Ky of divergence. 109. Cyprinodon milleri LaBounty and Deacon (Cottonball Marsh pupfish). This distinctive species was discovered liv- ing under most severe conditions in a marsh near Salt Creek, several miles southeast of the type locality of C. salinus in Death Valley, Inyo County, California (LaBounty and Deacon, 1972). The population lives below sea level in waters that ex- perience a temperature range of 0?-40? C. The fish can sur- vive a salinity range of 0-78 g/L. Large teeth, reduced or ab- sent pelvic fins, reduced lacrimal pores, short, slender caudal peduncle, small fins, and small scales diagnose the species. The large, tricuspid teeth have a broad, spatulate central cusp with no ridge on the outer face and a slim central shaft, like that of C. macularius Baird and Girard but with deeper inci- sions separating the outer cusps. It shares the absence of preor- bital scales with C. salinus and C. diabolis; it shares small scales and a slender body with C. salinus. Cyprinodon milleri and C. salinus differ at 4 of 31 protein loci (Rogers genetic dis- tance <0.06) according to Echelle and Echelle (1993), but the mtDNA of C. salinus and C. milleri do not differ. The docu- mentation of morphological differences by LaBounty and Deacon (1972) showed C. milleri to be the most phylogeneti- cally derived species in Death Valley. LaBounty and Deacon (1972) and Minckley et al. (1986) hypothesized an isolation interval of only a few hundred years, assuming no heterogene- ity at the time of the most recent hydrological connections documented by the geological studies of Hunt and Mabey (1966) and Hunt et al. (1966). This is consistent with the 0% sequence divergence and with geography, but it is in conflict with protein and morphological differences. Introgressive ho- moplasy of mtDNA is possible, in light of this inconsistency. The time of divergence is therefore ambiguous, but this is a possible example of rapid evolution. 110. Cyprinodon diabolis Wales (Devils Hole pup- fish). Devils Hole is a small cave fault that opens at an elevation of 2488 ft, in Ash Meadows, Nye County, Ne- vada. Its pool is the sole range of the 200-400 individuals of Cyprinodon diabolis. This small species has no pelvic fins, long dorsal and anal fins, a posterior dorsal fin, a convex caudal fin, and no preorbital scales. Mature males have a black terminal band on the caudal fin; mature fe- males lack vertical cross bars and show reversal of some sexually dimorphic characters (Miller, 1948). The popula- tion has lost individual territorial behavior (Liu in Turner, 1974). Mitochondrial DNA data (Figure 11) indicate that Cyprinodon diabolis is related to the C. nevadensis clade, and it shares most of its characters (e.g., large scales) with C. nevadensis; however, it shares two strikingly nonreduc- tive characters, increased dorsal- and anal-fin rays, with C. radiosus. Turner (1974) electrophoretically analyzed 31 presumptive loci and found C. diabolis to be similar to C. nevadensis and C. radiosus at the 0.968 and 0.936 lev- els of similarity, respectively. Morphological and mtDNA 224 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES homoplasies in combination suggest reticulate evolution in the history of C. diabolis (Echelle and Dowling, 1992). 111. ^Cyprinodon breviradius Miller (pupfish). This fossil form was found with Fundulus curryi in Titus Canyon on the east side of Death Valley, Inyo County, California (Miller, 1945b). The age was estimated by McAllister to be Miocene (Miller, 1981), but it could be Pliocene (see number 117, be- low). It is similar to C. macularius, but it has a slender body, like C. salinus, and 20 caudal vertebrae, which is several more than normal for the Death Valley group. On the basis of these apomorphies, the fossil is assumed to be ancestral to the ances- tor of the C. fontinalis?salinus?diabolis?nevadensis clade in Figure 11. This clade differs from its sister species by an aver- age of 2.9% sequence divergence. GOODEIDAE 112. *Empetrichthys merriami Gilbert (Ash Meadows pool- fish). This species, now extinct, was taken in five separated springs in Ash Meadows, Nevada. It was a small, robust fish with a deep body, large head, narrow mouth, and molariform pharyngeal teeth (Miller, 1948; Uyeno and Miller, 1962). Ap- parently always rare, it was possibly eliminated by scientific collecting (Miller, 1948:101). 113. *Empetrichthys latos Miller (Pahrump poolfish). This species lived in three springs in Pahrump Valley, Nye County, Nevada. Each spring had a distinct subspecies. Miller (1948) and Uyeno and Miller (1962) distinguished it from C. merri- ami by its slender body, broad and weak mouth, pointed teeth, and smaller scales. Springs that were inhabited by the three subspecies are geographically close to each other, and the taxa are morphologically similar. They differ from each other in body and fin shapes. A fossil relative, ^Empetrichthys erdisi (Jordan), from Pliocene sediments in the Ridge Basin, Los An- geles County, California, had conical teeth, smaller scales, and more vertebrae (Uyeno and Miller, 1962). Empetrichthys pha- ryngeals were identified by W.L. Minckley from the Glendale Formation (locality in Van Devender and Tessman, 1975) in southern Nevada (Minckley et al., 1986:582). 114. Crenichthys baileyi (Gilbert) (White River spring- fish). This species inhabits springs in the Pahranagat Valley and tributaries to the White River, in Clark, White Pine, Nye, and Lincoln Counties, Nevada. Geographic variation was rec- ognized, and four subspecies were named by Williams and Wilde (1981). 115. Crenichthys nevadae Hubbs (Railroad Valley spring- fish). This species differs from C. baileyi primarily in color- ation. It lives in Hot Springs in Railroad Valley and tributary Duckwater Valley, Nye County, Nevada. This drainage was probably connected to the Pahranagat Valley and the White River system through the Penoyer and Desert Valleys (Hubbs and Miller, 1948b). FUNDULIDAE 116. f Fundulus nevadensis (Eastman) (killifish). This is a late Miocene species (from at least 6 Ma) from the Truckee Formation in the Lahontan Basin near Hazen, Nevada. It is a long, slender fish with 11 or 12 dorsal-fin rays, the dorsal fin well in advance of the anal fin, 10-13 anal-fin rays, and small, circular scales in more than 65 lateral rows (Eastman, 1917; Miller, 1945b). LaRivers (1964) reported it in correlated beds in the Sahwave Mountains, Pershing County, Nevada. 117. ^Fundulus curryi Miller (killifish). This species, from the east side of Death Valley, was similar in some ways to Em- petrichthys (Miller, 1948:100). It is a short (44-66 mm SL), deep fish with 14 or 15 dorsal-fin rays, the dorsal fin much closer to the anal fin than to the pelvic fins, 15 or 16 anal-fin rays, a broadly rounded anal fin, and scales like those of F. eu- lepis (Miller, 1945b). Numerous specimens of Fundulus have been collected from Titus Canyon, Death Valley, California. Their ages are uncertain. The lower part of the Titus Canyon Formation has been dated as early Oligocene; the date was based upon the presence of Mesohippus, titanotheres, hydrac- odont rhinoceros, artiodactyls, and rodents, as well as on ra- diogenic dates (Saylor, 1994). But if F. curryi was collected near Cyprinodon breviradius Miller, it could be Miocene or Pliocene (Miller, 1945b). An old specimen of Fundulus was recently collected at the Bullfrog Mine, southeast of Beatty, Nevada. It was found in a dark gray to black shale beneath a late Oligocene to early Miocene volcanic unit dated at 23 Ma. 118. ^Fundulus eulepis Miller (killifish). This species is- from beds of Pliocene age in Death Valley near Furnace Creek. The specimens are relatively long (~9 cm SL), with 13 or 14 dorsal- and anal-fin rays, the dorsal fin in advance of the anal fin by about one-third of the distance from the anal fin to the pelvic-fin origins, and about 45-50 lateral scale rows (Miller, 1945b). 119. ^Fundulus spp. (killifishes). Undescribed samples of Fundulus have been collected from three localities on the Ne- vada Test Site by Richard V. Wyman (of the University of Ne- vada, Las Vegas) and others. Specimens studied so far are large (up to 9 cm SL), with conical teeth, 13 rays in the dorsal and anal fins, the dorsal fin slightly in advance of the anal fin, 6 pelvic-fin rays, moderately small scales on the body, scaled cheeks, and 20 caudal vertebrae. The caudal peduncle is mod- erately long, like that of F. eulepis. They lack evidence for modification of the anal fin, as seen in Goodeidae. Cyprinodon and Fundulus in the Miocene and Pliocene of the Great Basin are similar in morphology and diversity to modern cyprinod- ontoids (Empetrichthyines, Cyprinodon, Fundulus) in the southwestern United States. The fossil Fundulus add to the ev- idence of widespread, warm, low-gradient habitats during the Miocene. 120. ^Fundulus spp. (killifishes). Fundulus have been col- lected from two localities in the Esmeralda Formation, or beds reported to be Esmeralda Formation northwest of Death Val^ NUMBER 33 225 ley. These locations are in different and younger rocks than the Siebert tuff in which Fundulus lariversi was found. 121. ^Fundulus davidae Miller (killifish). A single specimen (Cal. Tech. 10276) of this species was obtained through Lore David from an oil driller. The specimen is from the Mohave Desert in northwestern San Bernardino County, California, but the formation of origin was not known. Miller (1945b) judged that the fossil was of Pliocene or early Pleistocene age. The specimen is short and moderately deep, with 11 or 12 dorsal- and anal-fin rays, 6 pelvic-fin rays, a large head, 40-45 lateral scale rows, and scale features like those of F. eulepis. Some of the characters are similar to those of Empetrichthys erdisi (Miller, 1945a). Specimens figured by Pierce (1959) from the Miocene Barstow Formation near Yermo in the Mohave Desert, might be Fundulus (Uyeno and Miller, 1962); the sup- posed fossil fish eggs reported from this locality were shown to be ostracods (Hubbs and Miller, 1962). 122. fFundulus lariversi Lugaski (killifish). This species was described from the Siebert tuff (which formed 17 Ma), from Tonopah, Nye County, Nevada. It is characterized by a slender body, conical teeth, dorsal fin over anal fin and half way between snout and caudal tip, 11 dorsal-fin rays, 13 anal- fin rays, 12 pectoral-fin rays, pelvic fin reduced to seven rays, 17 caudal-fin rays, and 31 vertebrae. This species is similar to Empetrichthys (Lugaski, 1978), but it possesses a pelvic fin. 123. ^Fundulus sp. (killifish). Fundulus was collected from the middle Miocene Humboldt Formation north of Wells, Ne- vada. The specimen is deposited at the Mackay School of Mines, Unversity of Nevada, Reno. COTTIDAE 124. *Cottus beldingi Eigenmann and Eigenmann (Paiute sculpin). This sculpin occupies the Lahontan, northern Bonneville, and Columbia-Snake drainages. Jordan (1924) re- ported fossils from Pleistocene cave deposits in the Lahontan Basin. Taylor and Smith (1981) reported the species from Pliocene rocks at Mopung Hills. A basioccipital diagnosed only as Cottus has been identified (by K.W.G.) from Pliocene sedi- ments in Secret Valley, California, and a cottid parasphenoid was collected by Julian Humphries from Virgin Valley, Nevada. An angular-articular from middle Pleistocene sediments of the Humboldt River gorge unquestionably represents Cottus beld- ingi. R.R. Miller collected specimens from Pliocene diatomites at Fish Cave, Churchill County, Nevada. This has apparently been a Lahontan Basin cool-water species since a time when the Snake River Plain had a strikingly different sculpin fauna (Smith, 1987) and the central Columbia Basin had a warm, lowland fauna lacking sculpins (G.R. Smith et al., 2000). 125. *Cottus bairdi Girard (mottled sculpin). This is the abundant, widespread sculpin of the Snake-Columbia and Bonneville drainages (as well as of eastern North America). Two different forms are found in the Harney-Malheur drainage (Bisson and Bond, 1971). The two morphs possibly colonized from the Columbia River drainage (by stream capture from the John Day River) and from the upper Snake River drainage (prior to separation of the Harney-Malheur drainage from the Snake River drainage). Pleistocene fossils were reported from the Black Rock and Hot Springs localities of Lake Bonneville (Smith et al., 1968). A Pliocene or Pleistocene Cottus was col- lected from Park Valley, Box Elder County, Utah (Felix, 1956), by Ted Cavender. 126. * Cottus extensus Bailey and Bond (Bear Lake scul- pin). The Bear Lake sculpin is now endemic to that lake, which is on the Utah-Idaho border, but it was common in Lake Bonneville in the late Pleistocene (Smith et al., 1968; Brough- ton, 2000a, 2000b). 127. Cottus echinatus Bailey and Bond (Utah Lake scul- pin). This extinct species is known only from Utah Lake, Utah. It has not been found among Lake Bonneville fossils, but it probably shared a common ancestor with Cottus extensus (Bailey and Bond, 1963). GASTEROSTEIDAE 128. ^Gasterosteus doryssus (Jordan) (Truckee stickle- back). The fossil three-spined stickleback (Gasterosteus do- ryssus) occurs in the Great Basin in the Miocene Truckee For- mation in west-central Nevada from Hazen (Bell, 1974) to the Sahwave Mountains (LaRivers, 1964). It represents a lowland fauna from the late Miocene (Bell et al., 1985) of Nevada when the drainage was west to the Pacific Ocean. Thousands of spec- imens have been collected and analyzed as a premier example of rates of evolution; this data has been summarized by Bell (1994, and references therein; Figure 12). 129. * Gasterosteus aculeatus Linnaeus (threespine stickle- back). Threespine sticklebacks are recorded from late Pleis- tocene sediments in the Mohave Basin in association with an- cient Lakes Manix and Thomson and Mohave River sediments (Roeder, 1985; Jefferson, 1991; Reynolds and Reynolds, 1991; Bell, 1994). Recently collected, living specimens of Gasteros- teus aculeatus williamsoni from high elevation in the Mohave Basin have some males with a unique melanistic nuptial pat- tern, suggesting that they are native (Bell, 1982). But trout (with stray sticklebacks) were transplanted by fisheries biolo- gists in that region in the 1930s or earlier, raising the possibility that the modern population is introduced (Leo Shapovilov, Cal- ifornia Department of Fish and Game, in litt. to R.R. Miller, University of Michigan, Museum of Zoology; Miller and Hubbs, 1969). CENTRARCHIDAE 130. ifArchoplites sp. (perch). Late Miocene sunfish occur in or near the Great Basin at several localities. These include Park Valley, Box Elder County, Utah (Felix, 1956); the Salt Lake group, (MVP 117073, collected by Patrick McClellan); Wind- ermere Hills near Elko, Nevada; the Snake River Plain in Idaho 226 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES and Oregon; and Trout Creek, Harney County, Oregon (Smith and Miller, 1985). Trout Creek specimens have been referred to the genus Plioplarchus Cope. Two partial dentaries of large sunfish were collected by J.A. Shotwell from the Juntura For- mation, outside the edge of the Great Basin in Malheur County, Oregon. Other specimens are from Miocene sediments in Kla- math Lake, Harney and Malheur Counties. These centrarchids are diverse and under study. Literature Cited Allendorf, F.W., and R.F. Leary 1988. Conservation and Distribution of Genetic Variation in a Polytypic Species, the Cutthroat Trout. Conservation Biology, 2:170-184. Allison, I.S. 1940. Study of Pleistocene Lakes of South Central Oregon. Carnegie In- stitute of Washington Year Book, 39:299-300. 1982. Geology of Pluvial Lake Chewaucan, Lake County, Oregon. Oregon State University Studies in Geology, 11:1-79. Allison, I.S., and C.E. Bond 1983. Identity and Probable Age of Salmonids from Surface Deposits at Fossil Lake, Oregon. Copeia, 1983(2):563-564. Anderson, R.E., and H.H. Mehnert 1979. Reinterpretation of the History of the Hurricane Fault in Utah. In G.W. Newman and H.D. Goode, editors, Basin and Range Sympo- sium and Great Basin Field Conference, pages 145-165. Denver: Rocky Mountain Association of Geologists and Utah Geological Association. Averitt, P. 1964. Table of Post-Cretaceous Geologic Events along the Hurricane Fault near Cedar City, Iron County, Utah. Bulletin of the Geological Soci- ety of America, 75:901-908. Axelrod, D.I. 1956. Mio-Pliocene Floras from West-Central Nevada. University of Cali- fornia Publications in Geological Sciences, 33:1-321. Bailey, R.M., and C.E. Bond 1963. Four New Species of Freshwater Sculpins, Genus Cottus, from Western North America. Occasional Papers of the University of Michigan Museum of Zoology, 634:1-27. Baksi, A.K., V. Hsu, M.O. Williams, and E. Farrar 1992. 40Ar/39Ar Dating of the Brunhes-Matuyama Geomagnetic Field Re- versal. Science, 256:356-357. Baumgartner, J.V. 1982. A New Fossil Ictalurid Catfish from the Miocene Middle Member oftheTruckee Formation, Nevada. Copeia, 1982(l):38-46. Behnke, R.J. 1972. The Systematics of Salmonid Fishes of Recently Glaciated Lakes. Journal of the Fisheries Research Board of Canada, 29(6):639-671. 1979. Monograph of the Native Trouts of the Genus Salmo of Western North America. Report to the United States Fish and Wildlife Ser- vice, Denver, Colorado, 215 pages. 1981. Systematic and Zoogeographical Interpretation of Great Basin Trouts. In R.J. Naiman and D.L. Soltz, editors, Fishes in North American Deserts, pages 95-124. New York: Wiley. 1992. Native Trout of Western North America. American Fisheries Society Monograph, 6: 275 pages. Bell, M.A. 1974. Reduction and Loss of the Pelvic Girdle in Gasterosteus (Pisces): A Case of Parallel Evolution. Contributions in Science, Los Angeles County Museum, 257:1?36. 1977. A Late Miocene Marine Threespine Stickleback, Gasterosteus ac- uleatus aculeatus, and Its Zoogeographic and Evolutionary Signifi- cance. Copeia, 1977(2):277-282. 1979. Persistence of Ancestral Sister-Species. Systematic Zoology, 28(1): Of onOJ?oo. 1982. Melanism in a High Elevation Population of Gasterosteus aculea- tus. Copeia, 1982:829-834. 1994. Palaeobiology and Evolution of Threespine Stickleback. In M.A. Bell and S.A. Foster, editors, The Evolutionary Biology of the Threespine Stickleback, pages 438-471. Oxford: Oxford University Press. Bell, M.A., J.V. Baumgartner, and E.C. Olson 1985. Patterns of Temporal Change in Single Morphological Characters of a Miocene Stickleback Fish. Paleobiology, 11(3): 25 8-271. Bell, M.A., and T.R. Haglund 1982. Fine Scale Temporal Variation of the Miocene Stickleback Gaster- osteus doryssus. Paleobiology, 8:282-292. Benson, L.V., D.R. Curry, R.I. Dorn, K.R. Lajoie, C.G. Oviatt, S.W. Robinson, G.I. Smith, and S. Stine 1990. Chronology of Expansion and Contraction of Four Great Lake Sys- tems during the Past 35,000 Years. Palaeogeography, Palaeoclima- tology, Palaeoecology, 78:241-286. Benson, L.V., and R.S. Thompson 1987. The Lake-Level Variation in the Lahontan Basin for the Past 50,000 Years. Quaternary Research, 28:69-85. Berg, W.J. 1987. Evolutionary Genetics of Rainbow Trout Parasalmo gairdneri (Ri- chardson). 192 pages. Doctoral dissertation, University of Califor- nia, Davis. Bills, F.T., and C.E. Bond 1980. A New Subspecies of Tui Chub (Pisces: Cyprinidae) from Cowhead Lake, California. Copeia, 1980(2):320-322. Bermingham, E., S.S. McCafferty, and A.P. Martin 1997. Fish Biogeography and Molecular Clocks. In T.D. Kocher and C.A. Stepien, editors, Molecular Systematics of Fishes, pages 113-128. New York: Academic Press. Bishop, P. 1995. Drainage Rearrangement by River Capture, Beheading and Diver- sion. Progress in Physical Geography, 19:449-473. Bisson, P.A., and C.E. Bond 1971. Origin and Distribution of the Fishes of Harney Basin, Oregon. Copeia, 1971(2):268-281. Blackwelder, E. 1933. Lake Manly: An Extinct Lake of Death Valley. Geography Reviews, 23:464-471. 1954. Pleistocene Lakes and Drainages in the Mohave Region, Southern California. In R.H. Jahns, editor, The Geology of Southern Califor- nia. California Division of Mines and Geology, Bulletin, 170:35-40. Blackwelder, E., and E.W. Ellsworth 1936. Pleistocene Lakes of the Afton Basin, California. American Journal of Science, 31:453-463. Bookstein, F.L. 1986. Random Walk and Existence of Evolutionary Rates. Paleobiology, 13:446^64. Bookstein, F.L., P.D. Gingerich, and A.G. Kluge 1978. Hierarchical Linear Modeling of the Tempo and Mode of Evolution. Paleobiology, 4(2): 120-134. NUMBER 33 227 Bouchard, D.P., D.S. Kaufman, A. Hochberg, and J. Quade 1998. Quaternary History of the Thatcher Basin, Idaho, Reconstructed from the 87Sr/86Sr and Amino Acid Composition of Lacustrine Fos- sils: Implications for the Diversion of the Bear River into the Bonneville Basin. Palaeogeography, Palaeoclimatology, Palaeo- ecology, 141:95-114. Bradley, W.H. 1936. Geomorphology of the North Flank of the Uinta Mountains, Utah. United States Geological Survey Professional Papers, 185-1:163? 199. Bright, R.C. 1966. Pollen and Seed Stratigraphy of Swan Lake, Southeastern Idaho: Its Relation to Regional Vegetation History and to Lake Bonneville History. Tebiwa, Journal of the Idaho State University Museum, 9(2): 1-47. 1967. Late Pleistocene Stratigraphy in Thatcher Basin, Southeastern Idaho. Tebiwa, Journal of the Idaho State University Museum, 10 (l):l-7. Broughton, J.M. 2000a. Terminal Pleistocene Fish Remains from Homestead Cave, Utah, and Implications for Fish Biogeography in the Bonneville Basin. Copeia, 2000(3):645-656. 2000b. The Homestead Cave Ichthyofauna, pages 103-121. In D.B. Mad- sen, editor, Late Quaternary Paleoecology in the Bonneville Basin. Utah Geological Survety Bulletin, 130: 190 pages. Broughton, J.M., D.B. Madsen, and J. Quade 2000. Fish Remains at Homestead Cave and Lake Levels of the Past 13,000 Years in the Bonneville Basin. Quaternary Research, 53: 392-401. Brown, J.H., and C.R. Feldmeth 1971. Evolution in Constant and Fluctuating Environments: Thermal Tol- erances of Desert Pupfish (Cyprinodon). Evolution, 25:390-398. Brown, W.J., and M.R. Rosen 1995. Was There a Pliocene-Pleistocene Fluvial-Lacustrine Connection between Death Valley and the Colorado River? Quaternary Re- search, 43:286-296. Buising, A.V. 1990. The Bouse Formation and Bracketing Units, Southeastern California and Western Arizona: Implications for the Evolution of the Proto- Gulf of California and the Lower Colorado River. Journal of Geo- physical Research, 95(B 12):20,111-20,132. Buwalda, J.P. 1914. Pleistocene Beds at Manix Lake in the Eastern Mohave Desert Re- gion. University of California Publications in Geological Sciences, 7:443-464. Cavender, T.M. 1978. Taxonomy and Distribution of the Bull Trout, Salvelinus confluen- tus (Suckley), from the American Northwest. California Fish and Game, 64:139-174. 1986. Review of the Fossil History of North American Freshwater Fishes. In C.H. Hocutt and E.O. Wiley, editors, The Zoogeography of North American Freshwater Fishes, pages 699-724. New York: Wiley-In- terscience Publication. Cavender, T.M., and M.M. Coburn 1992. Phylogenetic Relationships of North American Cyprinidae. In R.L. Mayden, editor, Systematics, Historical Ecology, and North Ameri- can Freshwater Fishes, pages 293-327. Stanford, California: Stan- ford University Press. Cavender, T.M., and R.R. Miller 1972. Smilodonichthys rastrosus, a New Pliocene Salmonid Fish from Western United States. Bulletin of the Oregon Museum of Natural History, 18:1-44. Chernoff, B. 1981. Character Variation among Populations and the Analysis of Bioge- ography. American Zoologist, 22(2):435-439. Coburn, M.M., and T.M. Cavender 1992. Interrelationships of North American Cyprinid Fishes. In R.L. May- den, editor, Systematics, Historical Ecology, and North American Freshwater Fishes, pages 328-373. Stanford, California: Stanford University Press. Cole, M.R., and J.M. Armentrout 1979. Neogene Paleogeography of the Western United States. In J.M Ar- mentrout and M.R. Cole, editors, Cenozoic Paleogeography of the Western United States. Pacific Coast Paleogeography Symposium, 3:297-323. Los Angeles: Pacific Section of the Society of Eco- nomic Petrologists and Mineralogists. Collins, T.M., K. Frazer, A.R. Palmer, G.J. Vermeij, and W.M. Brown 1996. Evolutionary History of Northern Hemisphere Nucella (Gastropoda, Muricidae): Molecular, Morphological, Ecological, and Paleonto- logical Evidence. Evolution, 50(6):2287-2304. Cope, E.D. 1868. On the Distribution of Fresh-Water Fishes in the Allegheny Region of Southwestern Virginia. Journal of the Academy of Natural Sci- ences of Philadelphia, 1868:207-247. 1872. On the Tertiary Coal and Fossils of Osino, Nevada. Proceedings of the American Philosophical Society, 12:475?481. 1879. The Fishes of Klamath Lake, Oregon. American Naturalist, 13: 784-785. 1883. On the Fishes of the Recent and Pliocene Lakes of the Western Part of the Great Basin, and the Idaho Pliocene Lake. Proceedings of the Academy of Natural Sciences of Philadelphia, 35:134?166. 1889. The Silver Lake of Oregon and Its Region. American Naturalist, 23: 970-982. Cracraft, J. 1983. Species Concepts and Speciation Analysis. Current Ornithology, 1: 159-187. Curry, D.R. 1990. Quaternary Paleolakes in the Evolution of Semidesert Basins, with Special Emphasis on Lake Bonneville and the Great Basin, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology, 76:241 ? 286. Dalrymple, G.B. 1964. Cenozoic Chronology of the Sierra Nevada, California. University of California Publications in Geological Sciences, 47:1?41. DeMarais, B.D., T.E. Dowling, M.E. Douglas, W.L. Minckley, and P.C. Marsh 1992. Origin of Gila seminuda (Teleostei: Cyprinidae) through Introgres- sive Hybridization: Implications for Evolution and Conservation. Proceedings of the National Academy of Science, 89:2747-2751. DiGuiseppi, W.H., and J.M. Bartley 1991. Stratigraphic Effects of Change from Internal to External Drainage in an Extending Basin, Southeastern Nevada. Bulletin of the Geo- logical Society of America, 103:48-55. Dowling, T.E., and G.J.P. Naylor 1997. Evolutionary Relationships of Minnows in the Genus Luxilus (Te- leostei: Cyprinidae) as Determined from Cytochrome b Sequences. Copeia, 1997(4):758-765. Dowling, T.E., L. Nyanti, and G.R. Smith In prep. Evidence for Endogenous Selection against Hybrids. Dowling, T.E., andC.L. Secor 1997. The Role of Hybridization and Introgression in the Diversification of Animals. Annual Review of Ecology and Systematics, 28:571- 593. Dowling, T.E., C.A. Tibbets, W.L. Minckley, and G.R. Smith 2002. Evolutionary Relationships of the Plagopterins (Telostei: Cyprin- idae) from Cytochrome b Sequences. Copeia, 2002:665-678. Duff, D.A. 1988. Bonneville Cutthroat Trout: Current Status and Management. In R.E. Greswell, editor, Status and Management of Interior Stocks of Cutthroat Trout. American Fisheries Society Symposium, 4:121? 127. 228 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Durham, J.W., and E.C. Allison 1960. The Geological History of Baja California and its Marine Faunas. Systematic Zoology, 9:47?91. Durrant, S.D. 1952. Mammals of Utah: Taxonomy and Distribution. University of Kan- sas Publications, Museum of Natural History, 6: 549 pages. Eastman, C.R. 1917. Fossil Fishes of the Collection of the United States National Mu- seum. Proceedings of the United States National Museum, 52: 235-304. Echelle, A.A., and T.E. Dowling 1992. Mitochondrial DNA Variation and Evolution of the Death Valley Pupfishes (Cyprinodon, Cyprinodontidae). Evolution, 46(1): 193-206. Echelle, A.A., and A.E. Echelle 1993. An Allozyme Perspective on mtDNA Variation and Evolution of Death Valley Pupfishes (Cyprinodon, Cyprinodontidae). Copeia, 1993(2):275-287. Feldmeth, C.R. 1981. The Evolution of Thermal Tolerance in Desert Pupfish, Cyprinodon macularius, in Mexico and Arizona. Copeia, 1981(2):478-483. Felix, C.E. 1956. Geology of the Eastern Part of the Raft River Range, Box Elder County, Utah. Guidebook of the Utah Geological Society, 11:76-97. Firby, J.R., S.E. Sharpe, J.F. Whelan, G.R. Smith, and W.G. Spaulding 1997. Paleobiotic and Isotopic Analysis of Mollusks, Fish and Plants from Core OL-92: Indicators of an Open or Closed Lake System. In G.I. Smith and J.L. Bischoff, editors, An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Spe- cial Papers of the Geological Society of America, 317:121?125. Fisher, R.A., and E.B. Ford 1947. The Spread of a Gene in Natural Conditions in a Colony of the Moth Panaxia dominula L. Heredity, 1:143-174. Flint, R.F., and W.A. Gale 1958. Stratigraphy and Radiocarbon Dates at Searles Lake, California. American Journal of Science, 256:689-714. Follett, W.I. 1963. Preliminary Report on Fish Remains from the Falcon Hill Site, Washoe County, Nevada. Nevada State Museum Anthropological Papers, 9:1-3. 1967. Fish Remains from Coprolites and Midden Deposits at Lovelock Cave, Churchill County, Nevada. Reports of the University of Cali- fornia Archeological Survey, 70:93-115. 1980. Fish Remains from Carlo Site (CA-LAS-7), Lassen County, Califor- nia. Journal of California and Great Basin Anthropology, 2(1): 114-122. Foote, M. 1996. On the Probability of Ancestors in the Fossil Record. Paleobiology, 22(2):141-151. Gale, H.S. 1915. Salines in the Owens, Searles, and Panamint Basins, Southeastern California. Bulletin of the United States Geological Survey, 580: 251-323. Gehr, K.D., and T.M. Newman 1978. Preliminary Note on the Late Pleistocene Geomorphology and Ar- chaeology of the Harney Basin, Oregon. The Ore Bin, 40(10): 165-170. Ghiselin, M.T. 1974. A Radical Solution to the Species Problem. Systematic Zoology, 23: 536-544. 1997. Metaphysics and the Origin of Species. 377 pages. Albany: State University of New York Press. Gilbert, C.H., and B.W. Evermann 1894. A Report upon Investigations in the Columbia River Basin with De- scriptions of Four New Species of Fish. Bulletin of the United States Bureau of Fisheries, 14:169-207. Gillespie, J.H. 1991. The Causes of Molecular Evolution. 336 pages. New York: Oxford University Press. Gilluly, J. 1929. Possible Desert-Basin Integration in Utah. Journal of Geology, 37: 672-682, figures 1-3. Gingerich, P.D. 1986. Temporal Scaling of Molecular Evolution in Primates and Other Mammals. Molecular Biology and Evolution, 3(3):205?221. Gobalet, K.W., and R.M. Negrini 1992. Evidence for Endemism in the Fossil Tui Chub, Gila bicolor, from Pleistocene Lake Chewaucan, Oregon. Copeia, 1992(2):539-544. Gould, S.J., andN. Eldredge 1977. Punctuated Equilibria: The Tempo and Mode of Evolution Recon- sidered. Paleobiology, 3:115-151. Grande, L. 1984. Paleontology of the Green River Formation, with a Review of the Fish Fauna. Bulletin of the Geological Survey of Wyoming, 63: 1-333. Grant, E.C, and B.R. Riddle 1995. Are the Endangered Springfish (Crenichthys Hubbs) and Poolfish (Empetrichthys Gilbert) Fundulines or Goodeids?: A Mitochondrial DNA Assessment. Copeia, 1995(l):209-212. Grinnell, J. 1924. Geography and Evolution. Ecology, 5:225-229. Hale, G.R. 1985. Mid-Pleistocene Overflow of Death Valley toward the Colorado River. In G.R. Hale, editor, Quaternary Lakes of the Eastern Mojave Desert California. Friends of the Pleistocene, Pacific Cell, Field Trip Guide, 1985:366-382. Hall, E.R. 1946. Mammals of Nevada. 710 pages. Berkeley: University of California Press. Hansen, W.R. 1969. The Geologic Story of the Uinta Mountains. Bulletin of the United States Geological Survey, 1291:i-xii, 1-144. 1985. Drainage Development of the Green River Basin in Southwestern Wyoming and Its Bearing on Fish Biogeography, Neotectonics, and Paleoclimates. The Mountain Geologist, 22:192-204. Haynes, C.V. 1967. Quaternary Geology of the Tule Springs Area, Clark County, Ne- vada. In H.M. Wormington and D. Ellis, editors, Pleistocene Studies in Southern Nevada. Anthropological Papers of the Nevada State Museum, 13:15-104. Hennig, W. 1966. Phylogenetic Systematics. 263 pages. Urbana: University of Illinois Press. Hershler, R., and W.L. Pratt 1990. A New Pyrgulopsis (Gastropoda: Hydrobiidae) from Southeastern California, with a Model for Historical Development of the Death Valley Hydrographic System. Proceedings of the Biological Society of Washington, 103:279-299. Hickman, T.J. 1978. Systematic Study of Native Trout of the Bonneville Basin. Master's thesis, Colorado State University, Fort Collins, Colorado. Howard, A.D. 1960. Cenozoic History of Northeastern Montana and Northwestern North Dakota with Emphasis on the Pleistocene. United States Geological Survey Professional Papers, 326:1?107. Howard, J.L. 1996. Paleocene to Holocene Paleodeltas of Ancestral Colorado River Offset by the San Andreas Fault System, Southern California. Geol- ogy, 24(9):783-786. NUMBER 33 229 Hubbs, C.L. 1940. Speciation of Fishes. American Naturalist, 74:198-211. 1941a. Fishes of the Desert. The Biologist, 22:61-69. 1941b. The Relation of Hydrological Conditions to Speciation in Fishes. In J.G. Needham, P.B. Sears, A. Leopold, and others, A Symposium on Hydrobiology, University of Wisconsin, 1940, pages 182-195. Mad- ison: University of Wisconsin Press. 1955. Hybridization between Fish Species in Nature. Systematic Zoology, 4:1-20. 1961. Isolating Mechanisms in the Speciation of Fishes. In W.F. Blair, edi- tor, Vertebrate Speciation, pages 5-23. Austin: University of Texas Press. Hubbs, C.L. and R.R. Miller 1943. Mass Hybridization between Two Genera of Cyprinid Fishes in the Mohave Desert, California. Papers of the Michigan Academy of Sci- ence, Arts, and Letters, 28:343-378. 1948a. Two New, Relict Genera of Cyprinid Fishes from Nevada. Occa- sional Papers of the University of Michigan Museum of Zoology, 507:1-30. 1948b. The Zoological Evidence: Correlation between Fish Distribution and Hydrographic History in the Desert Basins of Western United States. In The Great Basin, with Emphasis on Glacial and Postgla- cial Times. Bulletin of the University of Utah, 38(20): 17-166. 1960. The Spiny-Rayed Cyprinid Fishes (Plagopterini) of the Colorado River System. Miscellaneous Publications of the University of Michigan Museum of Zoology, 115: 39 pages. 1962. Supposed Miocene Fish Eggs from the Calico Mountains Nodules Identified as Ostracodes. Bulletin of the Southern California Acad- emy of Science, 60(3): 119-121. 1972. Diagnoses of New Cyprinid Fishes of Isolated Waters in the Great Basin of Western North America. Transactions of the San Diego So- ciety of Natural History, 17(8): 101 -106. Hubbs, C.L., R.R. Miller, and L.C. Hubbs 1974. Hydrographic History and Relict Fishes of the North-Central Great Basin. Memoirs of the California Academy of Sciences, 7:1-257. Huber, N.K. 1981. Amount and Timing of Late Cenozoic Uplift and Tilt of the Central Sierra Nevada, California?Evidence from the Upper San Joaquin River Basin. United States Geological Survey Professional Papers, 1197:1-28. Huber, N.K., and CD. Rinehart 1965. Geologic Map of the Devils Postpile Quadrangle, Sierra Nevada, California. United States Geological Survey Geologic Quadrangle Map, GQ-437, scale 1:62,500. 1967. Cenozoic Volcanic Rocks of the Devils Postpile Quadrangle, East- ern Sierra Nevada, California. United States Geological Survey Pro- fessional Papers, 5 54-D: 1 -21. Hunt, C.B., and D.R. Mabey 1966. General Geology of Death Valley, California, Stratigraphy and Structure. United States Geological Survey Professional Papers, 494-A: 1-165, map. Hunt, C.B., T.W. Robinson, W.A. Bowles, and A.L. Washburn 1966. Hydrologic Basin, Death Valley, California. United States Geologi- cal Survey Professional Papers, 494-B:iii-iv, B1-B138, figures 1- 73, plates 1, 2. Hutchinson, G.E. 1939. Ecological Observation on the Fishes of Kashmir and Indian Tibet. Ecological Monographs, 9:145-182. Huxley, J. 1943. Evolution, the Modern Synthesis. 645 pages. New York: Harper & Bros. Imbrie, J., and K.P. Imbrie 1979. Ice Ages. 224 pages, 49 figures. Cambridge, Massachusetts: Har- vard University Press. Ives, R.L. 1948. The Outlet of Lake Bonneville. The Scientific Monthly, 67:415-426. Izet, G.A. 1981. Volcanic Ash Beds: Recorders of Upper Cenozoic Silicic Pyroclas- tic Volcanism in the Western United States. Journal of Geophysical Research, 86:10,200-10,222. Izet, G.A., K.L. Pierce, and C. Jaworski 1992. Isotopic Dating of Lava Creek B Tephra in Terrace Deposits along the Wind River, Wyoming: Implications for Post 0.6 Ma Uplift of the Yellowstone Hotspot. Geological Society of America Abstracts with Programs, 24:A102. Jablonski, D. 1980. Apparent Versus Real Biotic Effects of Transgressions and Regres- sions. Paleobiology, 6:397-407. Jefferson, G.T. 1991. The Camp Cady Local Fauna: Stratigraphy and Paleontology of the Lake Manix Basin. San Bernardino County Museum Association Quarterly, 38:93-99. Jordan, D.S. 1878a. A Catalog of the Fishes of the Fresh Waters of North America. Bul- letin of the United States Geological and Geographical Survey, 4:407-442. 1878b. A Synopsis of the Family Catostomidae. Bulletin of the United States National Museum, 12:97-237. 1891. Report of Explorations in Colorado and Utah during the Summer of 1889, with an Account of the Fishes Found in Each of the Three River Basins Examined. Bulletin of the United States Fish Commis- sion, 9(1889):l-40. 1905. A Guide to the Study of Fishes. 2 volumes. New York: Henry Holt and Co. 1924. Description of a Recently Discovered Sculpin from Nevada Re- garded as Cottus beldingi. Proceedings of the United States Na- tional Museum, 65(6): 1?2. Jordan, D.S., and B.W. Evermann 1896. The Fishes of North and Middle America. Bulletin of the United States National Museum, 47(part 1): 1-618. Kelly, T.S. 1994. Two Pliocene (Blancan) Vertebrate Faunas from Douglas County, Nevada. PaleoBios, 16(1):1?23. Keyes, C.R. 1917. Orographic Origin of Ancient Lake Bonneville. Bulletin of the Geo- logical Society of America, 28:351-374. Kimmel, P.G. 1975. Fishes of the Miocene-Pliocene Deer Butte Formation, Southeast- ern Oregon. University of Michigan Museum of Paleontology, Pa- pers on Paleontology, 14:69-87. 1982. Stratigraphy, Age, and Tectonic Setting of the Miocene-Pliocene Lacustrine Sediments of the Western Snake River Plain, Oregon and Idaho. In Bill Bonnichsen and R.M. Breckenridge, editors, Cenozoic Geology of Idaho. Bulletin of the Idaho Bureau of Mines and Geol- ogy, 26:559-578. Kimura, M. 1980. A Simple Method for Estimating Evolutionary Rate of Base Substi- tutions through Comparative Studies of Nucleotide Sequences. Journal of Molecular Evolution, 16:111-120. Klicka, J., and R.M. Zink 1997. The Importance of Recent Ice Ages in Speciation: A Failed Para- digm. Science, 277:1666-1669. Kocher, T.D., and K.L. Carleton 1997. Base Substitutions in Fish Mitochondrial DNA. In T.D. Kocher and C.A. Stepien, editors, Molecular Systematics of Fishes, pages 559-578. New York: Academic Press. Kocher, T.D., and C.A. Stepien 1997. Molecular Systematics of Fishes. 314 pages. New York: Academic Press. 230 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Krishtalka, L., and R.K. Stucky 1985. Revision of the Wind River Faunas, Early Eocene of Central Wyo- ming, Part 7: Revision of Diacodexis (Mammalia, Artiodactyla). Annals of the Carnegie Museum of Natural History, 54:413?486. LaBounty, J.F., and J.E. Deacon 1972. Cyprinodon milleri, a New Species of Pupfish (Family Cyprinodon- tidae) from Death Valley, California. Copeia, 1972(4):769-780. LaRivers, I. 1962. Fishes and Fisheries of Nevada. 782 pages. Carson City: Nevada State Fish and Game Commission. 1964. A New Trout from the Barstovian (Miocene) of Western Nevada. Occasional Papers of the Biological Society of Nevada, 3:1-4. 1966. Paleontological Miscellanei. Occasional Papers of the Biological Society of Nevada, 11:1-8. Leary, R.F., F.W Allendorf, S.R. Phelps, and K.L. Knudsen 1984. Introgression between Westslope Cutthroat and Rainbow Trout in Clark Fork River Drainage, Montana. Proceedings of the Montana Academy of Science, 43:1-18. 1987. Genetic Divergence among Seven Subspecies of Cutthroat Trout and Rainbow Trout. Transactions of the American Fisheries Society, 116:580-587. Legendre, P. 1976. An Appropriate Space for Clustering Selected Groups of Western North American Salmo. Systematic Zoology, 25:13?195. Legendre, P., C.B. Schreck, and R.J. Behnke 1972. Taximetric Analysis of Selected Groups of Western North American Salmo with Respect to Phylogenetic Divergences. Systematic Zool- ogy, 21:292-307. Lemke, R.W., W.M. Laird, M.J. Tipton, and R.M. Lindvall 1965. Quaternary Geology of Northern Great Plains. In H.E. Wright, Jr., and D.G. Frye, editors, The Quaternary of the United States, pages 15-28. Princeton, New Jersey: Princeton University Press. Lewontin, R.C. 1974. The Genetic Basis of Evolutionary Change. 346 pages. New York: Columbia University Press. Lindsey, E.H., N.D. Opdyke, and N.M. Johnson 1984. Blancan-Hemphillian Land Mammal Ages and Late Cenozoic Mammal Dispersal Events. Annual Reviews of Earth and Planetary Science, 12:445-448. Loudenslager, E.J., and G.A.E. Gall 1980. Geographic Patterns of Protein Variation and Subspeciation in Cut- throat Trout, Salmo clarki. Systematic Zoology, 29:27-42. Luedke, R.G., and R.L. Smith 1982. Map Showing Distribution, Composition, and Age of Late Cenozoic Centers in Oregon and Washington. United States Geological Sur- vey Map, 1-1091-D. Lugaski, T. 1972. A New Species of Speckled Dace from Big Smokey Valley, Nevada. Occasional Papers of the Biological Society of Nevada, 30:1?8. 1977. Additional Notes and Discussion of the Relationships of Gila es- meralda LaRivers 1966 from the "Esmeralda" Formation. Occa- sional Papers of the Biological Society of Nevada, 43:1 -A. 1978. Fundulus lariversi, a New Miocene Fossil Cyprinodont Fish from Nevada. The Wasmann Journal of Biology, 35(2):203-211. 1979. Gila traini, a New Pliocene Cyprinid Fish from Jersey Valley, Ne- vada. Journal of Paleontology, 52(5): 1160-1164. Lundberg, J.G. 1992. The Phylogeny of Ictalurid Catfishes: A Synthesis of Recent Work. In R.L. Mayden, editor, Systematics, Historical Ecology, and North American Freshwater Fishes, pages 392-420. Stanford, California: Stanford University Press. 1993. African-South American Freshwater Fish Clades and Continental Drift: Problems with a Paradigm. In P. Goldblatt, editor, The Biotic Relationships between Africa and South America, pages 156-199. New Haven, Connecticut: Yale University Press. Lundelius, E.L., Jr., C.S. Churcher, T. Downs, C.R. Harrington, E.H. Lindsay, G.E. Schultz, H.A. Semken, S.D. Webb, and R.J. Zakrzewski 1987. The North American Quaternary Sequence. In M.O. Woodburne, editor, Cenozoic Mammals of North America, Geochronology and Biostratigraphy, pages 211-236. Berkeley: University of California Press. MacArthur, R.H., and E.O. Wilson 1967. The Theory of Island Biogeography. 203 pages. Princeton, New Jer- sey: Princeton University Press. Martin, A.P., and S.R. Palumbi 1993. Body Size, Metabolic Rate, Generation Time, and the Molecular Clock. Proceedings of the National Academy of Sciences, USA, 90: 4087-4091. Martin, M.A., D.K. Shiozawa, E.J. Loudenslager, and J.N. Jensen 1985. Electrophoretic Study of Cutthroat Trout Populations in Utah. Great Basin Naturalist, 45:677-687. Mayden, R.L., and R.M. Wood 1995. Systematics, Species Concepts, and the Evolutionary Significant Unit in Biodiversity and Conservation Biology. American Fisheries Society Symposium, 17:58-113. Mayr, E. 1963. Animal Species and Evolution. 797 pages. Cambridge, Massachu- setts: Belknap Press of Harvard University Press. 1984. Evolution of Fish Species Flocks: A Commentary. In A. Echelle and I. Kornfield, editors, Evolution of Fish Species Flocks, pages 3-11. Orono: University of Maine Press. McClellan, P.H. 1977. Paleontology and Paleoecology of Neogene Freshwater Fishes from the Salt Lake Beds, Northern Utah. 243 pages. Master's thesis, Uni- versity of California, Berkeley. McClellan, P.H., and G.R. Smith In press. Fishes of the Junction Hills Local Fauna, Salt Lake Group, North- ern Utah. Papers of the University of Michigan Museum of Paleon- tology. McCoy, W.D. 1987. Quaternary Aminostratigraphy of the Bonneville Basin, Western United States. Bulletin of the Geological Society of America, 98:99-112. McCune, A.R. 1990. Evolutionary Novelty and Atavism in the Semionotus Complex: Re- laxed Selection during Colonization of an Expanding Lake. Evolu- tion, 44(l):71-85. McCune, A.R., K.S. Thomson, and P.E. Olsen 1984. Semionotid Fishes from the Mesozoic Great Lakes of North Amer- ica. In A.A. Echelle and I. Kornfield, editors, Evolution of Fish Spe- cies Flocks, pages 27-44. Orono: University of Maine Press. Mead, J.I., R.S. Thompson, and T.R. Van Devender 1982. Late Wisconsinan and Holocene Fauna from Smith Creek Canyon, Snake Range, Nevada. Transactions of the San Diego Society of Natural History, 20( 1): 1 -26. Menely, W.A., E.A. Christiansen, and W.O. Kupsch 1957. Preglacial Missouri River in Saskatchewan. Journal of Geology, 65:441^147. Metzger, D.G. 1968. The Bouse Formation (Pliocene) of the Parker-Blythe-Cibola Area, Arizona and California. United States Geological Survey Profes- sional Papers, 600-D:D126-136. Miller, R.R. 1943a. Cyprinodon salinus, a New Species of Fish from Death Valley, Cali- fornia. Copeia, 1943(2):69-78. 1943b. The Status of Cyprinodon macularius and Cyprinodon nevadensis, Two Desert Fishes from Western North America. Occasional Pa- pers of the University of Michigan Museum of Zoology, 473:1-25. 1945a. Snyderichthys, a New Generic Name for the Leatherside Chub of the Bonneville and Upper Snake Drainages of Western United NUMBER 33 231 States. Journal of the Washington Academy of Sciences, 35:28. 1945b. Four New Species of Fossil Cyprinodont Fishes from Eastern Cali- fornia. Journal of the Washington Academy of Sciences, 35(10): 315-321. 1946a. Correlation between Fish Distribution and Pleistocene Hydrography in Eastern California and Southwestern Nevada, with a Map of the Pleistocene Waters. Journal of Geology, 54(l):43-54. 1946b. The Need for Ichthyological Surveys of the Major Rivers of West- ern North America. Science, 104:517-519. 1948. The Cyprinodont Fishes of the Death Valley System of Eastern Cal- ifornia and Southwestern Nevada. Miscellaneous Publications of the University of Michigan Museum of Zoology, 68:1-155. 1950. Speciation in Fishes of the Genera Cyprinodon and Empetrichthys, Inhabiting the Death Valley System. Evolution, 4:155-163. 1959. Origin and Affinities of the Freshwater Fish Fauna of Western North America. In C.L. Hubbs, editor, Zoogeography. Publications of the American Association for the Advancement of Science, 51:187?222. 1961a. Speciation Rates in Some Freshwater Fishes of Western North America. In W.F. Blair, editor, Vertebrate Speciation, pages 537-560. Austin: University of Texas Press. 1961b. Man and the Changing Fish Fauna of the American Southwest. Pa- pers of the Michigan Academy of Science, Arts, and Letters, 46: 365-404. 1965. Quaternary Freshwater Fishes of North America. In E.H. Wright and D.G. Frey, editors, The Quaternary of the United States, pages 569-582. Princeton, New Jersey: Princeton University Press. 1972. Threatened Freshwater Fishes of the United States. Transactions of the American Fisheries Society, 101:239-252. 1973. Two New Fishes, Gila bicolor snyderi and Catostomus fumeiventris, from the Owens River Basin, California. Occasional Papers of the University of Michigan Museum of Zoology, 667:1-19. 1981. Coevolution of Deserts and Fishes. In R.J. Naiman and D.L. Soltz, editors, Fishes in North American Deserts, pages 39-94. New York: John Wiley & Sons, Inc. 1984. Rhinichthys deaconi, a New Species of Dace (Pisces: Cyprinidae) from Southern Nevada. Occasional Papers of the University of Michigan Museum of Zoology, 707:1-21. Miller, R.R., and C.L. Hubbs 1969. Systematics of Gasterosteus aculeatus, with Particular Reference to Intergradation and Introgression along the Pacific Coast of North America: A Commentary on a Recent Contribution. Copeia, 1969 (1): 52-69. Miller, R.R., and W.M. Morton 1952. First Record of the Dolly Varden, Salvelinus malma, from Nevada. Copeia, 1952(3):207-208. Miller, R.R., and E.N. Shor 1997. Carl L. Hubbs (1894-1979): Collection Builder Extraordinaire. In T.W. Pietsch and W.D. Anderson, editors, Collection Building in Ichthyology and Herpetology. American Society of Ichthyologists and Herpetologists, Special Publication, 3:367-376. Miller, R.R., and G.R. Smith 1981. Distribution and Evolution of Chasmistes (Pisces: Catostomidae) in Western North America. Occasional Papers of the University of Michigan Museum of Zoology, 696:1-46. Miller, R.R., and M.L. Smith 1986. Origin and Geography of the Fishes of Central Mexico. In C.H. Hocutt and E.O. Wiley, editors, The Zoogeography of North Ameri- can Freshwater Fishes, pages 487-517. New York: John Wiley & Sons. Minckley, W.L., and J. Deacon 1991. Battle against Extinction: Native Fish Management in the American West. 517 pages. Tucson: University of Arizona Press. Minckley, W.L., D.A. Hendrickson, and C.E. Bond 1986. Geography of Western North American Freshwater Fishes: Descrip- tion and Relationships to Intercontinental Tectonism. In C.H. Hocutt and E.O. Wiley, editors, The Zoogeography of North American Freshwater Fishes, pages 519-614. New York: John Wiley and Sons. Mishler, B.D., and M.J. Donoghue 1982. Species Concepts: A Case for Pluralism. Systematic Zoology, 31(4):503-511. Moritz, C, S. Lavery, and R. Slade 1995. Using Allele Frequency and Phylogeny to Define Units of Conser- vation and Management. American Fisheries Society Symposium, 17:249-262. Morrison, R.B. 1991. Quaternary Stratigraphic, Hydrologic, and Climatic History of the Great Basin, with Emphasis on Lakes Lahontan, Bonneville, and Tecopa. In R.B. Morrison, editor, Quaternary Non-glacial Geology: Conterminous U.S. Boulder, Colorado: Geological Society of America: The Geology of North America, K2:283-320. 1965. Quaternary Geology of the Great Basin. In H.E. Wright, Jr., and D.G. Frye, editors, The Quaternary of the United States, pages 265-285. Princeton, New Jersey: Princeton University Press. Needham, P.R., and R.J. Behnke 1962. The Rationale of Preserving Genetic Diversity: Examples of the Uti- lization of Intraspecific Races of Salmonid Fish in Fisheries Man- agement. Proceedings of the Annual Conference Western Association of Fish and Wildlife Agencies, 52:559?561. Negrini, R.N., K..L. Verosub, and J.O. Davis 1987. Long-Term Non-geocentric Axial Dipole Directions and a Geomag- netic Excursion from the Middle Pleistocene Sediments of the Hum- boldt River Canyon, Pershing County, Nevada. Journal of Geophysical Research, 92(BIO): 10,617-10,627. Newberry, J.S. 1871. The Ancient Lakes of North America, Their Deposits and Drain- ages. American Naturalist, 4:641-660. Nielson, B.R., and L. Lentsch 1988. Bonneville Cutthroat Trout in Bear Lake: Status and Management. American Fisheries Society Symposium, 4:128-133. Orr, H.T. 1982. Tertiary and Quaternary Evolution of the Landscape in the Poca- tello, Idaho, Area. Northwest Geology, 11:31-36. Oviatt, C.G. 1984. Lake Bonneville Stratigraphy at the Old River Bed and Leamington, Utah. 122 pages. Doctoral dissertation, University of Utah, Salt Lake City, Utah. Oviatt, C.G., D.R. Curry, and D. Sack 1992. Radiocarbon Chronology of Lake Bonneville, Eastern Great Basin, USA. Palaeogeography, Palaeoclimatology, Palaeoecology, 99: 224-241. Parenti, L. 1981. Phylogenetic Study of Cyprinodontiform Fishes. Bulletin of the American Museum of Natural History, 168(4): 1?557. Parker, A., and I. Kornfield 1995. Molecular Perspective on Evolution and Zoogeography of Cyprin- odontid Killifishes (Teleostei: Atherinomorpha). Copeia, 1995(1): 8-21. Pierce, K.L., and L.A. Morgan 1992. The Track of the Yellowstone Hot Spot: Volcanism, Faulting, and Uplift. In P.K. Link, M.A. Kuntz, and L.B. Platt, editors, Regional Geology of Eastern Idaho and Western Wyoming. Memoirs of the Geological Society of America, 179: 1-53. Pierce, W.D. 1959. Silicified Eggs of Vertebrates from Calico Mts. Nodules. Bulletin of the Southern California Academy of Sciences, 58(2):79?83. Piper, A.M., T.W. Robinson, and C.F. Park 1939. Geology and Ground-Water Resources of the Harney Basin, Ore- gon. United States Geological Survey Water-Supply Paper, 841 :i?iv, 1-189. 232 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Reheis, M.C., and R.B. Morrison 1997. High, Old Pluvial Lakes of Western Nevada. Brigham Young Uni- versity Geological Studies, 42(l):459-492. Reheis, M.C., S. Stine, and A.M. Sarna-Wojcicki 2002. Drainage Reversals in Mono Basin during the Late Pliocene and Pleistocene. Geological Society of America Bulletin, 114(8):991- 1006. Repenning, C.A. 1987. Biochronology of the Microtine Rodents of the United States. In M.O. Woodburne, editor, Cenozoic Mammals of North America, Geochronology andBiostratigraphy, pages 236-265. Berkeley: University of California Press. Reynolds, R.E. 1994. Fishing for Paleogeographic Clues in the Holloran Hills, Eastern San Bernardino County, California. Geological Society of America Abstracts and Programs, 26(2):84. Reynolds, R.E., and R.L. Reynolds 1991. Late Pleistocene Faunas of Lake Thomson. San Bernardino County Museum Association Quarterly, 38:114-115. Ricklefs, R.E., and D. Schluter 1993. Species Diversity in Ecological Communities. 414 pages. Chicago: University of Chicago Press. Rieseberg, L., and J.F. Wendel 1993. Introgression and Its Consequences in Plants. In R.G. Harrison, edi- tor, Hybrid Zones and the Evolutionary Process, pages 70-109. New York: Oxford University Press. Robins, C.R., R.M. Bailey, C.E. Bond, J.R. Brooker, E.A. Lachner, R.N. Lea, and W.B. Scott 1991. Common and Scientific Names of Fishes from the United States and Canada. Fifth edition. Special Publication 20, 183 pages. Bethesda, Maryland: American Fisheries Society. Roeder, M.A. 1985. Late Wisconsin Records of Gasterosteus aculeatus (Threespine Stickleback) and Gila bicolor mohavensis (Mohave Tui Chub) from Unnamed Mohave River Sediments near Daggett, San Bernardino County, California. In R.E Reynolds, compiler, Geological Investi- gations along Interstate 15, Cajon Pass to Manix Lake, pages 171- 174. Riverside, California: San Bernardino County Museum. Rosenfeld, M. J. 1991. Systematic Studies of Members of the Genus Gila (Pisces: Cyprin- idae) from the Great Basin and Colorado River: Protein Electro- phoretic and Cytogenetic Variation. 183 pages. Doctoral dis- sertation, University of Utah, Salt Lake City, Utah. Rosenzweig, M.L. 1978. Competitive Speciation. Journal of the Linnaean Society, 10:275-289. Rostlund, E. 1951. Dolly Varden, Salvelinus malma (Walbaum). Copeia, 1951 (4):296. Sarich, V.M., and A.C. Wilson 1967. Rates of Albumin Evolution in Primates. Proceedings of the Na- tional Academy of Science, 58:142-148. Saylor, B.Z. 1994. 40Ar/39Ar Age Constraints on the Depositional History of the Oli- gocene Titus Canyon Formation, Death Valley, CA. Geological So- ciety of America Abstracts with Programs, 26(2):88. Sears, J.D. 1924. Relations of the Browns Park Formation and the Bishop Conglom- erate and Their Role in the Origins of the Green and Yampa Rivers. Bulletin of the Geological Society of America, 35:279-304. Shaffiqullah, M., RE. Damon, D.J. Lynch, S.J. Reynolds, W.A. Rehrig, and R.H. Haymond 1980. KJ-Ar Geochronology and Geologic History of Southwestern Ari- zona and Adjacent Areas. Arizona Geological Society Digest, 12: 201-260. Shiozawa, D.K., and R.P. Evans 1995. Relationships between Cutthroat Trout Populations from Eight Utah Streams in the Colorado River and Bonneville Drainages. Interim Report to the Utah Division of Wildlife Resources, Contract No. 94- 2377:1-12. Sigler, W.F., and J.W. Sigler 1987. Fishes of the Great Basin: A Natural History. 425 pages. Reno: Uni- versity of Nevada Press. 1996. Fishes of Utah. 203 pages. Salt Lake City: University of Utah Press. Simons, A.M., and R.L. Mayden 1997. Phylogenetic Relationships of the Creek Chubs and the Spinefins: An Enigmatic Group of North American Cyprinid Fishes (Actinop- terygii: Cyprinidae). Cladistics, 13(3): 187-205. 1998. Phylogenetic Relationships of the Western North American Phoxi- nins (Actinopterygii: Cyprinidae) as Inferred from Mitochondrial 12S and 16S Ribosomal RNA Sequences. Molecular Phylogenetics and Evolution, 9(2):308-329. Simpson, G.G. 1951. The Species Concept. Evolution, 5:285-298. Smiley, C.J., and W.C. Rember 1985. Physical Setting of the Miocene Clarkia Fossil Beds, Northern Idaho. In C.J. Smiley, editor, Late Cenozoic History of the Pacific Northwest, pages 11-31. San Francisco: Pacific Division of the American Association for the Advancement of Science. Smith, G.I. 1978. Field Guide to Examples of Quaternary Geology, Searles Valley, California. 31 pages. Menlo Park, California: United State Geologi- cal Survey. [Unpublished.] Smith, G.I., V.J. Barczak, G.F. Moulton, and J.C. Liddicoat 1983. Core KM-3, a Surface-to-Bedrock Record of Late Cenozoic Sedi- mentation in Searles Valley, California. United States Geological Survey Professional Papers, 1256:1-29. Smith, G.I., and J.L. Bischoff 1997. An 800,000-Year Paleoclimatic Record from Core OL-92, Owens Lake, Southeast California. Special Papers of the Geological Soci- ety of A merica, 317:1-160. Smith, G.R. 1966. Distribution and Evolution of North American Catostomid Fishes of the Subgenus Pantosteus, Genus Catostomus. Miscellaneous Publi- cations of the University of Michigan Museum of Zoology, 129: 1-132. 1975. Fishes of the Pliocene Glenns Ferry Formation, Southwest Idaho. University of Michigan Museum of Paleontology, Papers on Pale- ontology, 5(14): 1-68. 1978. Biogeography of Intermountain Fishes. In K.T. Harper and J.T. Re- veal, Intermountain Biogeography: A Symposium. Great Basin Nat- uralist Memoirs, 2:17-42. 1981a. Late Cenozoic Freshwater Fishes of North America. Annual Review of Ecology and Systematics, 12:163-193. 1981b. Effects of Habitat Size on Species Richness and Adult Body Sizes in Desert Fishes. In R.J. Naiman and D.L. Soltz, editors, Fishes in North American Deserts, pages 125-171. New York: John Wiley & Sons, Inc. 1985. Paleontology of Hidden Cave: Fish. In D.H. Thomas, The Archeol- ogy of Hidden Cave. Anthropological Papers of the American Mu- seum of Natural History, 61 (1): 171 -178. 1987. Fish Speciation in a Western North American Pliocene Rift Lake. Palaios, 2:436-445. 1992a. Introgression in Fishes: Significance for Paleontology, Cladistics, and Evolutionary Rates. Systematic Biology, 41(1):41?57. 1992b. Phytogeny and Biogeography of the Catostomidae, Freshwater Fishes of North America and Asia. In R.L. Mayden, editor, System- atics, Historical Ecology, and North American Freshwater Fishes, pages 778-826. Stanford, California: Stanford University Press. Smith, G.R., and R.K. Koehn 1971. Phenetic and Cladistic Studies of Biochemical and Morphological Characteristics of Catostomus. Systematic Zoology, 29(3):282-297. NUMBER 33 233 Smith, G.R., and R.R. Miller 1985. Taxonomy of Fishes from Miocene Clarkia Lake Beds, Idaho. In C.J. Smiley, editor, Late Cenozoic History of the Pacific Northwest, pages 75-83. San Francisco: American Association for the Ad- vancement of Science. Smith, G.R., N. Morgan, and E. Gustafson 2000. Fishes of the Mio-Pliocene Ringold Formation, Washington: Pliocene Capture of the Snake River by the Columbia River. Papers of the University of Michigan Museum of Paleontology, 32:1-47. Smith, G.R., J. Rosenfield, and J. Porterfield 1995. Processes of Origin and Criteria for Preservation of Fish Species. American Fisheries Society Symposium, 17:44?57. Smith, G.R., W.L. Stokes, and K.F. Horn 1968. Some Late Pleistocene Fishes of Lake Bonneville. Copeia, 1968(4):807-816. Smith, G.R., K. Swirydczuk, K. Kimmel, and B.H. Wilkinson 1982. Fish Biostratigraphy of Late Miocene to Pliocene Sediments of the Western Snake River Plain, Idaho. In B. Bonnichson and R.M. Breckenridge, editors, Cenozoic Geology of Idaho, pages 519-540. Moscow, Idaho: Idaho Bureau of Mines and Geology. Smith, G.R., and T.N. Todd 1984. Evolution of Species Flocks in Northern Temperate Lakes. In A. Echelle and I. Kornfield, editors, Evolution of Fish Species Flocks, pages 45-68. Orono: University of Maine Press. 1993. Morphological Cladistic Study of Coregonine Fishes. In T.N. Todd and M. Luczynski, editors, Biology and Management of Coregonid Fishes; Proceedings of the Fourth International Symposium, 1990. Polskie Archiwum Hydrobiologii, 39(3^t):479-490. Smith, M.L. 1980. The Evolutionary and Ecological History of the Fishes of the Rio Lerma Basin, Mexico. 191 pages. Doctoral dissertation, University of Michigan, Ann Arbor, Michigan. 1981. Late Cenozoic Fishes in the Warm Deserts of North America: A Re- interpretation of Desert Adaptations. In R.J. Naiman and D.L. Soltz, editors, Fishes in North American Deserts, pages 11-38. New York: John Wiley & Sons, Inc. Smith, P.B. 1960. Fossil Foraminifera from the Southeastern California Deserts. United States Geological Survey, Short Papers in the Geological Sciences, page B278. 1970. New Evidence for Pliocene Marine Embayment along the Lower Colorado River Area, California and Arizona. Bulletin of the Geo- logical Society of America, 81:1411-1420. Snyder, J.O. 1908. Relationships of the Fish Fauna of Southeastern Oregon. Bulletin of the United States Bureau of Fisheries, 27(636):69-102. 1912. A New Species of Trout from Lake Tahoe. Bulletin of the United States Bureau of Fisheries, 32:23?28. 1914. The Fishes of the Lahontan Drainage System of Nevada and Their Relationship to the Geology of the Region. Journal of the Washing- ton Academy of Science, 4:299-300. 1917a. The Fishes of the Lahontan System of Nevada and Northeastern California. Bulletin of the United States Bureau of Fisheries, 35(1915):31-86. 1917b. An Account of Some of the Fishes from Owens River California. Proceedings of the United States National Museum, 54:201-205. 1918. The Fishes of the Mohave River, California. Proceedings of the United States National Museum, 54:297-299. 1919. Three New Whitefishes from Bear Lake, Idaho and Utah. Bulletin of the United States Bureau of Fisheries, 36:1?9. 1924. Notes on Certain Catostomids of the Bonneville System, including the Types of Pantosteus virescens Cope. Proceedings of the United States National Museum, 64:1?6. Soltz, D.L., and R.J. Naiman 1978. The Natural History of Native Fishes in the Death Valley System. Nat- ural History Museum of Los Angeles County, Science Series, 30:1-76. Spencer, J.E., and J. Patchett 1997. Isotopic Evidence for a Lacustrine Origin for the Upper Miocene to Pliocene Bouse Formation, Lower Colorado River Trough, and Im- plications for Timing of Colorado Plateau Uplift. Bulletin of the Geological Society of America, 109(6):767-778. Stanley, S.M. 1975. A Theory of Evolution above the Species Level. Proceedings of the National Academy of Science of the United States of America, 72: 646-650. Stearley, R.F., and G.R. Smith 1993. Phylogeny of the Pacific Trouts and Salmons (Oncorhynchus) and Genera of the Family Salmonidae. Transactions of the American Fisheries Society, 122:1-30. Stirton, R.A. 1936. Succession of North American Continental Pliocene Mammalian Faunas. American Journal of Science, series 5, 32:161?206. Stokes, W.L. 1979. Paleohydrographic History of the Great Basin Region. In G.W. Newman and H.D. Goode, editors, Basin and Range Symposium and Great Basin Field Conference, pages 345-351. Denver: Rocky Mountain Association of Geologists. Swirydczuk, C, G.P. Larson, and G.R. Smith 1982. Volcanic Ash Beds as Stratigraphic Markers in the Glenns Ferry and Chalk Hills Formations from Adrian, Oregon, to Bruneau, Idaho. In B. Bonnichsen and R.M. Breckenridge, editors, Cenozoic Geology of Idaho. Idaho Bureau of Mines and Geology, Bulletin, 26:543- 558. Swoford, D.L. 1990. PAUP: Phylogenetic Analysis using Parsimony, Version 3.0. Cham- paign, Illinois: Illinois Natural History Survey. Tamura, K., and M. Nei 1993. Estimation of the Number of Nucleotide Substitutions in the Control Region of Mitochondrial DNA in Humans and Chimpanzees. Mo- lecular Biology and Evolution, 10:512-526. Tanner, V.M. 1936. A Study of the Fishes of Utah. Utah Academy of Sciences, Arts and Letters, 13:155-184. 1950. A New Species of Gila from Nevada (Cyprinidae). Great Basin Nat- uralist, 10:31-36. Taylor, D.W. 1960. Distribution of the Freshwater Clam Pisidium ultramontanum: A Zoogeographic Inquiry. American Journal of Science, 258-A: 325-334. 1966. Summary of North American Blancan Nonmarine Mollusks. Mala- cologia, 4:1?172. 1983. Late Tertiary Mollusks from the Lower Colorado River Valley. Con- tributions from the University of Michigan Museum of Paleontology, 26:289-298. 1985. Evolution of Freshwater Drainages and Molluscs in Western North America. In C.J. Smiley, Late Cenozoic History of the Pacific Northwest, pages 265-322. San Francisco: American Association for the Advancement of Science. Taylor, D.W, and R.C. Bright 1987. Drainage History of the Bonneville Basin. In R.S. Kopp and R.E. Cohenour, editors, Sites for Precious Metal and Hydrocarbon Accu- mulations. Utah Geological Association Publication, 16:239?256. Taylor, D.W. and G.R. Smith 1981. Pliocene Molluscs and Fishes from Northeastern California and Northwestern Nevada. Contributions from the University of Michi- gan Museum of Paleontology, 25(18):339?413. Tedford, R.H., M.F. Skinner, R.W. Fields, J.M. Rensberger, D.P. Whistler, T. Galusha, B.E. Taylor, J.R. Macdonald, and S.D. Webb 1987. Chapter 6: Faunal Succession and Biochronology of the Arikareean through Hemphillian Interval (Late Oligocene through Earliest 234 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Pliocene Epochs) in North America. In M.O. Woodburne, editor, Cenozoic Mammals of North America, Geochronology and Bios- tratigraphy, pages 153-210. Berkeley: University of California Press. Thomas, D.H. 1985. The Archaeology of Hidden Cave, Nevada. Anthropological Papers of the American Museum of Natural History, 61:1?430. Tol, D., and J. French 1988. Status of a Hybridized Population of Alvord Cutthroat Trout from Virgin Creek, Nevada. American Fisheries Society Symposium, 4:116-120. Townsend, J.K. 1978. Narrative of a Journey across the Rocky Mountains to the Columbia River. 259 pages. Lincoln: University of Nebraska Press. [Reprint; originally published in 1839.] Turner, B.J. 1974. Genetic Divergences of Death Valley Pupfish Species: Biochemical Versus Morphological Evidence. Evolution, 28:281?294. 1984. Evolutionary Genetics of Fishes. 639 pages. New York: Plenum Press, Monographs in Evolutionary Biology. Uyeno, T., and R.R. Miller 1962. Relationship of Empetrichthys erdisi, a Pliocene Cyprinodontid Fish from California, with Remarks on the Fundulinae and Cyprinodon- tinae. Copeia, 1962(3):520-532. Van Devender, T.R., and N.T. Tessman 1975. Late Pleistocene Snapping Turtles (Chelydra serpentina) from Southern Nevada. Copeia, 1975(2):249-253. Van Valen, L. 1976. Ecological Species, Multi-species, and Oaks. Taxon, 25:233-239. 1988. Species, Sets, and the Derivative Nature of Philosophy. Biology and Philosophy, 3:49-66. Vrba, E.S. 1993. Turnover-Pulses, the Red Queen, and Related Topics. American Journal of Science, 293-A:418-452. Wagner, H.M., C.B. Hanson, E.P. Gustafson, K.W. Gobalet, and D.P. Whistler 1997. Biogeography of the Pliocene and Pleistocene Vertebrate Faunas of Northeastern California and Their Temporal Significance to the De- velopment of the Modoc Plateau and Klamath Mountain Region. San Bernardino County Museum Association Quarterly, 44(1): 13-21. Wagner, W.H., Jr. 1983. Reticulistics. In N.I. Platnick and V.A. Funk, editors, Advances in Cladistics, pages 63-79. New York: Columbia University Press. Webb, S.A. 1998. Phylogenetic Studies of the Family Goodeidae (Teleostei, Cyprin- odontiformes). 280 pages. Doctoral dissertation, University of Michigan, Ann Arbor, Michigan. Weldon, R.J. 1982. Pleistocene Drainage and Displaced Shorelines around Manix Lake. In J.D. Cooper, editor, Geologic Excursions in the California Desert; Geological Society of America-Cordilleran Section? Guidebook, pages 77-81. Boulder, Colorado: Geological Society of America. Wheeler, H.E., and E.F. Cook 1954. Structural and Stratigraphic Significance of the Snake River Cap- ture, Idaho-Oregon. Journal of Geology, 62(6):525-536. Wiley, E.O. 1981. Phylogenetics: The Theory and Practice of Phylogenetic Systemat- ics. 439 pages. New York: Wiley. Williams, J.E. 1978. Taxonomic Status of Rhinichthys osculus (Cyprinidae) in the Moapa River, Nevada. Southwestern Naturalist, 23(3):511-518. Williams, J.E., and C.E. Bond 1980. Gila boraxobius, a New Species of Cyprinid Fish from Southeastern Oregon with a Comparison to G. alvordensis Hubbs and Miller. Pro- ceedings of the Biological Society of Washington, 93(2):291-298. 1981. A New Subspecies of Tui Chub (Osteichthyes: Cyprinidae) from Guano Basin, Nevada and Oregon. Southwestern Naturalist, 26: 223-230. Williams, J.E., and G.R. Wilde 1981. Taxonomic Status and Morphology of Isolated Populations of the White River Springfish, Crenichthys baileyi (Cyprinodontidae). Southwestern Naturalist, 25: 485-503. Wolfe, J.A., H.E. Schorn, C.E. Forest, and P. Molnar 1997. Evidence for High Altitudes in Nevada during the Miocene. Sci- ence, 276:1672-1675. Woodburne, M.O., editor 1987. Cenozoic Mammals of North America, Geochronology and Bios- tratigraphy. 336 pages. Berkeley: University of California Press. Wright, S. 1940a. Breeding Structure of Population in Relation to Speciation. Ameri- can Naturalist, 74:232-248. 1940b. The Statistical Consequences of Mendelian Heredity in Relation to Speciation. In J.S. Huxley, editor, The New Systematics, pages 161-183. Oxford: Oxford University Press. 1942. Statistical Genetics and Evolution. Bulletin of the American Mathe- matical Society, 28:223-246. 1948. On the Roles of Directed and Random Changes in Gene Frequency in the Genetics of Populations. Evolution, 2:279-295. Basins and Ranges: The Biogeography of Aquatic True Bugs (Insecta: Heteroptera) in the Great Basin Dan A. Polhemus and John T. Polhemus ABSTRACT The endorheic drainage catchments of the Great Basin form an obvious, topographically defined biogeographic region for the aquatic true bugs (Heteroptera) of North America. The Great Basin, however, does not have a homogenous regional fauna, because of its broad north-to-south extent and the consequent vari- ation of both present day climate zones and past limnological regimes. An east-west drainage divide, apparently linked to a transverse geological accommodation zone, separates the northern basins, including the Bonneville and Lahontan Basins and those of southeastern Oregon, from the southern basins lying in southeast- ern California and southern Nevada. The northern basins have depauperate water bug faunas characterized by a widespread suite of north temperate generalist species, and by the absence or near- absence of many families found at similar latitudes outside the Great Basin. We hypothesize that this pattern resulted from the presence of cold-water lakes in these northern basins at various times during the Pliocene and Pleistocene, which represented hos- tile habitats for many aquatic Heteroptera and which did not per- mit southerly escape along river corridors in response to periods of increased climatic stress. The southern basins have had a much different faunal history. One of the most remarkable concentra- tions of disjunct and endemic aquatic Heteroptera in North Amer- ica occurs in three of these isolated, endorheic drainages: Railroad Valley and the White River, in southern Nevada, and the Amar- gosa River drainage, straddling the adjacent California-Nevada border. As with other isolated aquatic elements, including spring Dan A. Polhemus, Department of Systematic Biology, MRC 105, Na- tional Museum of Natural History, Smithsonian Institution, Washing- ton, D.C. 20560-0105, USA. John T. Polhemus, Colorado Entomological Institute, 3115 South York Street, Englewood, Colo- rado 80110, USA. Reviewers: Richard W. Baumann, Department of Zoology, Brigham Young University, Provo, Utah, 84602-5255, USA; Thomas Henry, Systematic Entomology Lab, MRC 168, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0168, USA; Robert Her shier, Division of Mollusks, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0118, USA; W.L. Minckley, Arizona State University, Tempe, Arizona 85287, USA; Gerald R. Noonan, Invertebrate Zoology Section, Milwaukee Public Museum, 800 West Wells Street, Milwaukee, Wisconsin, 53233, USA. fishes, the nearest conspecific or congeneric populations of many of these isolates occur in northern Mexico or central Texas rather than in the more geographically proximal Colorado River basin. This indicates that these southern basins may have had an histori- cal connection to the Rio Grande system before the formation of the Colorado River in its current configuration. In addition, the southern basins also contain highly isolated populations of species adapted to cold-water regimes. Unlike the isolates associated with thermal waters, these cold-water species have their nearest associ- ated populations in sections of the upper Virgin and Colorado River systems immediately to the south. We suggest that this cold- water pattern is of more recent origin, and it has resulted from dis- persal through post-Pliocene cold-water drainages that once existed in the region. Introduction The endorheic drainage catchments of the Great Basin, as defined in the broad sense, form a topographically delimited biogeographic subregion for the aquatic true bugs (Het- eroptera) of North America. The basin as a whole, however, does not have a homogenous regional fauna, because of its broad north-to-south extent and the consequent variation of both current-day climate zones and past limnological regimes. Despite the obvious north-to-south expression of the surface topography, the Great Basin also contains a marked west-to- east drainage divide, cutting roughly across the middle of Ne- vada, then jogging southeastward along the Utah border (Fig- ure 1). This drainage divide is roughly correspondent to the Caliente-Enterprise accommodation zone, an east-west band of structural discontinuity and magmatic intrusion that has persisted transverse to regional extensional structures for nearly 30 My, from the Early Oligocene to the Holocene (Axen, 1998). This zone also appears to correspond to the Blue Ribbon-Warm Springs lineament, noted by various au- thors as a zone of volcanism and caldera formation from 30 to 20 million years ago (Ma) (Rowley et al., 1978; Mutschler et al., 1998; Rowley et al., 1998). As pointed out by Rowley et al. (1998), such transverse structures are analogous to the transform faults seen in oceanic basins running perpendicular 235 236 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ( Columbia ( River ? 30 FIGURE 1.?The Great Basin within the context of western North America. The cold, northern basins, as defined herein, lie within the dark hatched area, and the warm, southern basins lie within the area indicated by lighter hatch markings. to plate boundary spreading centers. In each case, they allow the crust to accommodate different amounts of spreading on either side of the structure, be this spreading because of mag- matism (as at mid-oceanic ridges) or extensional faulting (as in the Great Basin). Many of the basins north of the divide formed by the Cali- ente-Enterprise accommodation zone, which we term the northern basins, contained large cold-water lakes during the Pliocene and Pleistocene and had periodic drainage connec- tions to the Snake and Columbia Rivers. The basins south of the divide, which we term the southern basins, contained fewer and smaller lakes, had warmer Pliocene and Pleistocene climates, and had previous drainage connections that led southward toward the modern day Colorado River (Hubbs and Miller, 1948). Along the transverse divide itself are several basins of ambiguous affinity that appear to have been strictly endorheic throughout their histories; these include Dixie, Toiyabe, and Steptoe, plus the series of endorheic catchments devoid of native fishes that Hubbs and Miller (1948) referred to as the "area of sterile basins" (see Figure 2). For the pur- poses of the current report the first three basins have been grouped with the northern basins, whereas the latter group has been included in the southern basins (although it is treated as a separate faunal unit in Table 1). NUMBER 33 237 Flowing stream channel -'- Intermittent stream channel ? Dry lake (@) Major thermal spring ? ? State boundary FIGURE 2.?Map of the southern Great Basin, showing locations of thermal springs (hatch-filled spots) within the principal drainage basins discussed in the text. Localities: 1 = Duckwater, 2 = Lockes Ranch, 3 = upper White River, 4 = Pahranagat Valley, 5 = Moapa Warm Springs, 6 = Ash Meadows, 7 = middle Amargosa River basin, 8 = Saratoga Spring, 9 = Furnace Creek, 10 = Grapevine Springs. The divide separating the northern and southern basins has been a distinctive biological barrier as well as a topographic one, and this is reflected in the distributions of aquatic Het- eroptera and other aquatic insects. The northern basins, includ- ing the Bonneville Basin, the Lahontan Basin, and those of southeastern Oregon, have depauperate water bug faunas char- acterized by the absence or near-absence of many groups, such as Hebridae, Hydrometridae, most Nepidae and Veliidae, and the genus Buenoa in the Notonectidae, which are found at sim- ilar latitudes outside the northern Great Basin (Table 1). We hy- pothesize that this pattern resulted from the former presence of cold-water lakes in these northern basins at various times dur- ing the Pleistocene, which represented hostile habitats for many aquatic Heteroptera, precluding the possibility of escape along southward-leading river corridors in response to periods of increased climatic stress. Even among cold-tolerant families 238 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 1.?Geographic distribution of aquatic heteroptera within the Great Basin, by drainage basin: BV, Bonne- ville Basin; LH, Lahontan Basin; SO, southeastern Oregon closed basins; SB, area of sterile basins; RR, Railroad Valley and tributaries; WR, White River basin; AR, Amargosa River basin. X, species found in basin; -, species not found. Species, by family Belostomatidae Belostoma bakeri Montandon Belostoma flumineum (Say) Lethocerus americanus (Leidy) Lethocerus angustipes (Mayr) Corixidae Callicorixa audeni Hungerford Callicorixa scudderi Jansson Callicorixa vulnerata (Uhler) Cenocorixa bifida bifida Hungerford Cenocorixa utahensis (Hungerford) Cenocorixa wileyae (Hungerford) Corisella decolor (Uhler) Corisella edulis (Champion) Corisella inscripta (Uhler) Corisella tarsalis (Fieber) Graptocorixa abdominalis (Say) Graptocorixa serrulata (Uhler) Hesperocorixa laevigata (Uhler) Hesperocorixa vulgaris (Hungerford) Sigara alternata (Say) Sigara grossolineata Hungerford Sigara mckinstryi Hungerford Sigara nevadensis Walley Sigara omani (Hungerford) Sigara washingtonensis Hungerford Trichocorixa calva (Say) Trichocorixa reticulata (Guerin-Meneville) Trichocorixa uhleri Sailer Trichocorixa verticalis interiores Sailer Trichocorixa verticalis saltoni Sailer Gelastocoridae Gelastocoris oculatus oculatus (Fabricius) Nerthra martini Todd Gerridae Aquarius remigis (Say) Gerris buenoi Kirkaldy Gerris gillettei Lethierry & Severin Gerris incognitus Drake & Hottes Gerris incurvatus Drake & Hottes Limnoporus notabilis (Drake & Hottes) Hebridae Hebrus hubbardi Porter Merragata hebroides White Hydrometridae Hydrometra australis Say Mesoveliidae Mesovelia amoena Uhler Mesovelia mulsanti White Naucoridae Ambrysus amargosus La Rivers Ambrysus funebris La Rivers Ambrysus mormon Montandon Ambrysus relictus J. & D. Polhemus Ambrysus woodburyi Usinger Limnocoris moapensis (La Rivers) Pelocoris biimpressus biimpressus Montandon Pelocoris biimpressus shoshone La Rivers Northern basins BV X - X - X - X X X - X X X X - - X - - X - X - X - - - X - X _ X X X - - X - - - - - - - X - X - - - LH X - X - X - - - - X X X - - X - X - X - X X X X - - - - - X - X - X - X - - - - - X - - X - - - - - so -- X - X X - - - X X - X X - X X X - X - - X X - - - - - X - X X X X X - - - - - - - - - - - - - - SB X - - - - - - - - X - - - - - X - X - - - X - - X - - - - - - - X - - - - - - - - - - - - - - - - Southern basins RR X - - - - - - - - X X - - - - - X - - - - - X - - - - - X - X X - X - - - X X - - - - - - - X - - X WR X X - - - - - - - X X X X - - X X - - - - - - - X - X - X - X X - - - - - X X X - X - - X - X X - X AR X - - X - - - - - - - - - - - - - - - - - - - - - - - - - - X - - - - - - - X X X X X X X X - - X X Colorado River X X - - - X - - - X - X X X X X X - - - - X - X X X X X - X X X X X - - X X X X X X - - X - X - - - Snake River X - X - - X - X X X X - X X - - X X - X - - X X - X - - - X - X X X X X X - - - - - - - X - - - - - Mexico X X - X - - - - - - - X X X X X X - - - - - - - - X - - - X X X - - - - - - X X X X - - X - - - X - NUMBER 33 239 TABLE 1.?Continued. Species, by family Nepidae Ranatrafusca Palisot Beauvior Notonectidae Buenoa margaritacea Bueno Buenoa omani Truxal Buenoa scimitra Bare Notonecta kirbyi Hungerford Notonecta spinosa Hungerford Notonecta undulata Say Notonecta unifasciata Guerin-Meneville Ochteridae Ochterus barberi Schell Veliidae Microvelia beameri McKinstry Microvelia buenoi Drake Microvelia cerifera McKinstry Microvelia gerhardi Hussey Microvelia hinei Drake Microvelia signata Uhler Microvelia torquata Champion Rhagovelia becki Drake & Harris Rhagovelia distincta Champion Northern basins BV X - - - X X X X - - X - - - X - - X LH - - - - X X X X - - X X - X X - - - so X - - - X - - X - - X - - - - - - - SB - X - - X - - X - X - - - - - - - - Southern basins RR - - - - X - - X - X - - - X - - - - WR - - X - X - X X X X - X X X - X X X AR - - - X - - X X - X - - - X - - - X Colorado River - X - X X - X X X X - - X X - X - X Snake River X - - - X - X X - - - - - - - - - X Mexico - X X X - - X X X X - - X X X X X X such as the Gerridae and Corixidae, the present-day fauna of these basins is composed primarily of widespread generalist species also seen in the Snake and Columbia River systems (Table 1). The absence of interbasin or intrabasin endemism also indicates that the former lake systems of the northern ba- sins presented relatively homogenous habitats to those species capable of inhabiting them, and these systems were sufficiently interconnected to preclude allopatric speciation. The southern basins, particularly those lying in southern Ne- vada and the adjacent southeastern California-Nevada border area, have had a much different faunal history. One of the most remarkable concentrations of disjunct and endemic aquatic Heteroptera species in North America is found in the thermal springs of three of these catchments: Railroad Valley, White River, and Amargosa River (see Figure 2). The nearest conspe- cific or congeneric populations of many of these Heteroptera isolates are frequently found in central Texas or northern Mex- ico, rather than in the more geographically proximal drainages of the modern Colorado River basin and the Sierra Madre Oc- cidental of western Mexico, a pattern also seen in other simi- larly relictual aquatic elements, such as spring fishes. We hy- pothesize that these drainages were continuous and probably integrated in the Miocene, connecting to a drainage network that flowed southeast toward the modern Rio Grande and the Gulf of Mexico, as suggested by McKee et al. (1967). After the late Miocene to early Pliocene extension in the southern Great Basin (Stewart, 1998), these drainages then became endorheic and underwent desiccation, becoming reduced to sets of dis- continuous spring-fed oases. In addition to these thermal isolates, the southern basins also contain highly disjunct populations of certain species, such as creeping water bugs, adapted to cold-water regimes. These cold-water isolates have their nearest associated populations in sections of the upper Virgin and Colorado River systems lying immediately to the east. It is suggested that this pattern is of much more recent origin and has resulted from dispersal through Pleistocene or post-Pleistocene cold-water drainages that formed in the region, followed once again by subsequent desiccation and fragmentation. We review herein the evidence for the above patterns sup- plied by various families of aquatic Heteroptera. We detail the distributions of selected genera and species, address taxonomic problems relating to these, and discuss the biogeographic sig- nificance of the thermally associated fauna in the White River, Railroad Valley, and the Amargosa River basins. Distribution of Aquatic Heteroptera Species in the Great Basin The Great Basin distributions of aquatic Heteroptera species in the infraorders Gerromorpha and Nepomorpha are detailed in Table 1. The data for this table were obtained both from ex- amination of major museum collections and from existing tax- onomic literature (Drake and Harris, 1934; Hungerford, 1934, 1948; Harris and Shull, 1944; Truxal, 1953, 1979; Knowlton, 1956; Lauck, 1979; Menke, 1979a, 1979b, 1979c; Polhemus, J.T., 1979; Polhemus and Chapman, 1979a, 1979b, 1979c, 1979d, 1979e; Smith, 1980; Stonedahl andLattin, 1982, 1986; Henry and Froeschner, 1988; Sites and Polhemus, 1994). Three sets of collection data were of particular importance in regard to this analysis: (1) general collections from throughout the Great Basin by the authors from 1964 to the present; (2) 240 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES collections in southern Nevada and Death Valley, California, from 1992 to 1997, by the authors specifically in support of this study; and (3) results from a year-long entomological sur- vey (1996?1997) by Richard W. Baumann and colleagues in the Muddy River basin of the lower White River system, southern Nevada. In the following discussion, we use the terms "thermal en- demic" and "thermal disjunct" in reference to species confined to thermal spring waters and incapable of tolerating the ambi- ent water temperatures currently prevailing in the Great Basin. In cases where such taxa are endemic to single springs, we re- fer to them as thermal endemics. Alternately, when such taxa are distributed among several thermal springs in different drainage basins (Figure 3) or when they represent isolated pop- ulations having broader distributions independent of thermal springs in areas further to the south (Figures 4, 5), we refer to them as thermal disjuncts. Taxonomic Considerations Taxonomic interpretations given in Table 1 are valid under existing nomenclature and, for the most part, are stable. In a few groups, however, the taxonomy is undergoing renewed scrutiny, and species concepts are likely to be revised in the next several years. This section reviews several biogeographi- cally informative genera, in three families, that are under study by specialists. The problems involved are illustrative of the dif- ficulties encountered when dealing with the taxonomy of incip- iently speciating populations fragmented by recent tectonic and climatological events. FIGURE 3.?Distribution of the subspecies of Pelocoris biimpressus in the southwestern United States and northern Mexico. Hatch-filled spots = Pelo- coris biimpressus biimpressus. Solid black spots = Pelocoris biimpressus shos- hone. In addition to the localities shown here, P. biimpressus biimpressus is also recorded from eastern Texas, the Mexican states of Veracruz and Jalisco, and Guatemala; P. biimpressus shoshone is endemic to the southern Great Basin. Pelocoris is absent in the Colorado River basin?the locality shown in southern Arizona lies on a tributary of the Rio Conception flowing southward into Mexico. BELOSTOMATIDAE Belostoma saratogae was described by Menke (1958), who examined a series of specimens from Saratoga Spring, in south- ern Death Valley. Subsequent workers considered this a valid species endemic to the type locality. Data obtained from our re- cent collections of Belostoma from across the American South- west, however, reveal that the characters used to separate B. sa- ratogae from the more widespread and closely related B. bakeri (Figure 6) exhibit variation and that the two species in Ix??? "" " r~~???-^ t \. / i ?H/ / / 1110? r^T?~"? ...r[r~\ J7> -?-^ I / ^->L V nsr -4- J../J i ?4-(_ i i ?j < 0 < -i ? <* f1 / * K?jjg dwvV Llr ii L \ 30? 20? FIGURE 4.?Distribution of Rhagovelia becki (hatched areas). This species is an example of a thermal disjunct, as discussed in the text. Its primary distribu- tion is in southern Texas and northeastern Mexico, but a single, highly isolated population exists at Moapa Warm Springs in the White River basin of southern Nevada. NUMBER 33 241 FIGURE 5.?Distribution of Lethocerus angustipes (hatched areas). This spe- cies is another example of a thermal disjunct (discussed in the text), having its primary distribution in central Mexico with a single, highly isolated population in the Amargosa River basin of southern Nevada and California. FIGURE 6.?Belostoma bakeri Montandon. This giant water bug occurs throughout the Great Basin. In the southern basins it intergrades with Belos- toma saratogae Menke, a species found in the thermal springs of the Railroad Valley, White River, and Amargosa River basins. fact appear to intergrade. Character sets examined provide con- flicting diagnoses; these include male phallic features, airstrap measurements, and head structure. Two species, however, still may be involved, and they may segregate on the basis of thermal-spring versus cold-spring habitats. For instance, at Moorman Hot Spring (along the upper White River) we can distinguish a single male taken in the cold marsh from a long series of males taken in the warm waters of the outflow creek. If one bases the diagnosis solely on the ven- tral diverticulum of the phallus and ignores all other characters, then the "warm" specimens from Moorman Hot Spring are B. bakeri, and the "cold" specimen is B. saratogae. The situation at other spring-fed habitats around the region is less clear and provides no obvious geographic or ecological pattern. Populations inhabiting thermal springs at Duckwater and Lockes Ranch in Railroad Valley (see Figure 2) are B. bakeri, whereas specimens from thermal waters at Hot Creek along the upper White River are B. saratogae. Most of the specimens from thermal springs in Ash Meadows conform to a broad interpretation of B. saratogae except for a few indeter- minate specimens from a cool-water marsh. The population at the type locality conforms to the original description, but the airstrap measurements are variable even within this popula- tion, and this character is not as definitive as stated by Menke (1958). Given the above observations, it is reasonable to propose that B. saratogae is simply a synonym of B. bakeri. At the same time, the Belostoma populations in the Railroad Valley, White River, and Amargosa River drainages clearly represent a vari- able but related complex, which can in turn be separated con- sistently from populations of B. bakeri inhabiting surrounding regions. It may be necessary, therefore, to redefine the entire insular complex as B. saratogae and to interpret this species as more widespread and variable than was previously believed. All related populations outside the three basins noted above then would be considered B. bakeri. The only specimens fail- ing to conform to this arrangement are a series from San Diego, California, that falls within the new "broad" concept of B. sa- ratogae. This record is of interest biogeographically in light of hypotheses that the Mojave River once had a connection to the Pacific Ocean via the Los Angeles Basin (Hershler and Pratt, 1990). Because of this uncertainty, all records for this species complex are given as B. bakeri in Table 1. One additional Belostoma species, B. Jlumineum, is found at Ash Warm Springs on the middle White River and at Moapa Warm Springs and Meadow Valley Wash on the lower White River (Figure 2). This species, however, is a transconti- nental taxon widespread throughout the Colorado River sys- tem and the interior valleys of California and is quite distinct morphologically from the saratogae/ bakeri complex (Lauck, 1964); thus, its distribution has little bearing on the above analysis. 242 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 7.?Limnocoris moapensis (La Rivers). This creeping water bug is endemic to Moapa Warm Springs in the lower White River basin. This genus is absent in the Colorado River system, with the closest congeners occurring in Texas. NAUCORIDAE Limnocoris moapensis (Figure 7) was described by La Riv- ers (1950b) from material taken at Moapa Warm Springs, to which the species is endemic. It is one of only two taxa in the subfamily Limnocorinae found in the United States, the other being Limnocoris lutzi La Rivers from Texas. The primary character used by La Rivers (1950b) to separate Limnocoris from Usingerina was the unusual, hooked shape of the hemely- tral embolium in Usingerina. The suitability of this character as a basis for generic-level separation was subsequently ques- tioned by several authors (De Carlo, 1951; Sites and Willig, 1994), who synonymized Usingerina under Limnocoris. The presence of L. moapensis in southern Nevada provides evi- dence of former connections between the southern Great Basin and a previously more extensive Rio Grande system. It is also noteworthy that this species has not spread into the Colorado River system, despite the presence of a flowing water connec- tion via the Muddy River. La Rivers (1948) described Pelocoris shoshone (Figure 8) from material taken at Ash Warm Springs along the middle White River, and he subsequently reported additional speci- mens from Moapa Warm Springs along the lower White River (La Rivers, 1950a). Several years later an additional population of this species was found at Saratoga Spring in Death Valley, which La Rivers (1956) interpreted as a separate subspecies, P. shoshone amargosus. These discoveries were remarkable at the time because Pelocoris was considered to be a Neotropical ge- nus extending northward only to the southern and eastern United States. Further collecting during the past thirty years, FIGURE 8.?Pelocoris biimpressus shoshone La Rivers. This creeping water bug occurs in the Railroad Valley and White River basins of southern Nevada and in Grapevine Springs in northern Death Valley. Populations of the nomi- nate subspecies, P. biimpressus biimpressus, occur throughout the remainder of Death Valley and the lower Amargosa River basin. The species is absent from the Colorado River drainage but widespread in northern Mexico and southern Texas. however, revealed that Pelocoris is widespread as a series of disjunct populations throughout the southern Great Basin. By 1994, P. shoshone was known not only from the Ash Warm Springs, Moapa Warm Springs, and Saratoga Spring localities that La Rivers reported, but also from Sheep Creek and Grape- vine Springs (in Grapevine Canyon) in Death Valley, warm springs in the towns of Tecopa and Shoshone along the middle Amargosa River, many springs at Ash Meadows, and the warm springs at Duckwater in the northern Railroad Valley (see P. bi- impressus shoshone in Table 1). La Rivers (1956) suggested that Pelocoris shoshone shos- hone was confined to the White River drainage, whereas the subspecies P. shoshone amargosus occurred in the Amargosa River basin. Although intuitively appealing, this hypothesis of distribution has been disproven by this study of subsequent col- lections, which demonstrated that the characters used to sepa- rate the two subspecies?size and color?intergrade between the many populations now known. If La Rivers' (1956) diagno- sis for P. shoshone amargosus is strictly applied, the only popu- lations in the southern Great Basin that truly conform to his concept are those from Saratoga Spring, Tecopa Hot Springs, and Shoshone Spring, all lying in the Amargosa basin. This subspecies also occurs in the Atascosa Mountains of far south- ern Arizona (along southward flowing tributaries to the Rio Concepcion) and along the western foothills of the Sierra Ma- dre Occidental in Sonora, Mexico (see P. biimpressus biimpres- sus in Figure 3). All of the remaining Pelocoris populations in NUMBER 33 243 the southern Great Basin, including those in northern Death Valley, Ash Meadows, Railroad Valley, and the White River drainage, conform to La Rivers' (1948) concept of P. shoshone shoshone, a subspecies that does not have a wider distribution (see P. biimpressus shoshone in Figure 3). The taxonomic situation is further complicated by the recent determination that La Rivers' P. shoshone is a synonym of P. biimpressus, a poorly known species described by Montandon (1898) from Texas (Polhemus and Sites, 1995). As a result of this synonymy and a reanalysis of existing material, we treat the Pelocoris populations from the Amargosa River basin, with the exception of those at Grapevine Springs in northern Death Valley, as P. biimpressus biimpressus, whereas the populations occurring in the Railroad Valley and White River basins and at Grapevine Springs are considered P. biimpressus shoshone (Figure 3). The distribution of the latter taxon suggests that the northern and southern segments of Death Valley had differing biogeographic histories and that the Grapevine Canyon area may have had historical relationships with drainages east of the modern Amargosa River before the creation of the Death Val- ley trough. Montandon (1909) described Ambrysus mormon (Figure 9) on the basis of a type series collected at St. George, Utah, pre- sumably from the Virgin River or a nearby tributary. This spe- cies is widespread throughout the West, exhibiting extensive morphological variation. In particular, populations inhabiting thermal waters tend to be consistently dwarfed and display dis- tinctive morphological features. Two subspecies have been de- scribed from such habitats: A. mormon heidemanni, from the Firehole River in the Yellowstone caldera (Montandon, 1910), and A. mormon minor, from warm springs along the Owyhee River near Bruneau, Idaho (La Rivers, 1963). Similar trends to- ward morphological differentiation in thermal springs are also evident in the southern Great Basin, particularly with regard to the populations from Ash Warm Spring and Hot Creek in the White River basin. Decisions as to whether these populations merit recognition as distinct subspecies must await detailed analysis of morphological and molecular variation in ,4. mor- mon across its entire range. NOTONECTIDAE Notonecta unifasciata (Figure 10) was described by Guerin- Meneville (1857) from specimens taken in central Mexico. Hungerford (1934) listed the range as South America to British Columbia and recognized several subspecies. Notonecta uni- fasciata has been recorded in both hot springs and saline pools (including Francis Lake near Tecopa; Saline Valley in Inyo County) as well as in cool-water habitats. Usinger (1956) treated all populations in mainland California as the subspecies N. unifasciata andersoni, but this interpretation does not apply for the populations found in the southern Great Basin. Al- though N. unifasciata andersoni occurs in thermal waters at Hot Creek along the upper White River, other specimens taken in the cool marsh below Moorman Hot Spring, less than 30 km away, represent N. unifasciata unifasciata. Analysis of the available collection data indicates that N. unifasciata andersoni FIGURE 9.?Ambrysus mormon mormon Montandon. This is the most wide- spread creeping water bug in the western United States. Some populations in thermal springs of the Great Basin show morphological divergence indicative of incipient speciation. FIGURE 10.?Notonecta unifasciata andersoni Hungerford. This backswimmer is widely distributed in the Great Basin and occurs in thermal or saline waters. 244 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES inhabits thermal or saline waters in the southern Great Basin, whereas N. unifasciata unifasciata lives in cooler waters, such as marshes downstream from thermal springs. These subspe- cies may be clinal, however, and any firm conclusions regard- ing habitat partitioning must await resolution of the taxonomic problems involved. Throughout the southern Great Basin, N. unifasciata occurs primarily in thermal springs. It is found at Lockes Ranch and Twin Springs in southern Railroad Valley; Moorman Hot Spring and Hot Creek along the upper White River; Hiko Spring along the middle White River; and several springs in Ash Meadows on the middle Amargosa River. By contrast, cold springs of the upper White River at the towns of Lund and Emigrant (which are in close geographic proximity to the ther- mal outflows at Moorman Hot Spring and Hot Creek) harbor Notonecta kirbyi, a species with a distinct preference for cold waters. Also interesting is the presence of a third species, N. undulata, in thermal Saratoga Spring in southern Death Valley, a habitat where one might otherwise expect to find N. unifasci- ata. The recent discovery of another backswimmer, Buenoa omani, near Moapa in the lower White River basin is also of in- terest. This species is otherwise known only from the southern Coast Range and Los Angeles Basin of California southward into Mexico, where it is widespread?and not from the Colo- rado River system in the United States. Biogeography of the Northern Basins As noted previously, the aquatic Heteroptera biota of the northern basins consists of widespread generalist species and exhibits no local endemism. Instead, the species assemblage of the Bonneville Basin, Lahontan Basin, and southeastern Ore- gon basins is composed of north temperate taxa that are broadly distributed across much of the northwestern United States. Although thermal springs are widespread across the re- gion, none contain thermally disjunct or endemic species, pos- sibly because many of these waters were submerged beneath large pluvial lakes in the Pleistocene. A comparative assessment of distribution patterns in the northern basins also can be made by examining the distribu- tions of Ephmeroptera, or mayflies. Because of their short adult life spans, mayflies are poor dispersers and, given their toler- ance for cold waters, might be expected to develop localized endemism in the northern basins. Although 40 genera of may- flies are recorded from Utah, only 23 of these (i.e., 57%) occur in the Bonneville Basin, and only a single species is considered endemic to this basin (G.F. Edmunds, Jr., pers. comm. to DAP, 1998). In most cases, these genera have lower species diversi- ties than they exhibit in areas immediately to the south or east. Similarly, the mountain ranges within the Bonneville and Lahontan Basins contain a depauperate assemblage of genera and species compared with the Sierra Nevada and Wasatch ranges that form the basin borders. Thus, as in Heteroptera, the mayflies of the northern basins comprise a limited subset of a predominantly widespread north temperate biota. One aquatic Heteroptera record in the northern basins is wor- thy of note: the presence of the creeping water bug Ambrysus woodburyi in the Sevier River at Junction, Utah, in the far southern Bonneville Basin. This is the only known occurrence of A. woodburyi in the northern basins; all other populations occur in the Colorado River system or at isolated localities in several of the southern basins (i.e., Railroad Valley, White River). Given the otherwise southern distribution of this spe- cies, its presence in the Sevier drainage is anomalous and sug- gests that the Sevier River may have captured the headwaters of streams that previously drained southward to the basin of the modern Colorado River. Another notable aspect of the original aquatic Heteroptera fauna of the northern basins is the general paucity of Veliidae, particularly Microvelia and Rhagovelia (see Table 1), which are otherwise very widespread and ubiquitous throughout the northern United States. Rhagovelia distincta was recently in- troduced into the eastern Bonneville Basin via water diversions from the Uinta Basin associated with the Central Utah Project (Polhemus, D.A., 1997). This inadvertent introduction illus- trates that aquatic insects are often quick to utilize newly estab- lished hydrologic dispersal routes between basins, whether these corridors be natural or human-made. Biogeography of the Southern Basins Three of the southern basins lying in California and Nevada contain significant concentrations of thermally endemic or dis- junct aquatic Heteroptera: Railroad Valley, the White River ba- sin, and the Amargosa River basin. On the basis of extralimital faunal relationships, we hypothesize that these basins had pre- vious connections either to the Rio Grande system via a drain- age network similar to that proposed by McKee et al. (1967) or to other now disrupted drainages that once flowed southward toward Mexico. In some cases (i.e., lower White River) these basins also exhibit apparently recent connections to the modern Colorado River system. Railroad Valley and the White River basins lie in close proximity, being separated by a common drainage divide. The Amargosa River basin, by contrast, is more isolated, being separated from the previous two by the area of sterile basins (Hubbs and Miller, 1948), a region of nearly waterless endorheic basins devoid of native fishes, and by the Las Vegas Wash catchment, which drains to the present Colorado River system (Figure 2). RAILROAD VALLEY Railroad Valley is an extensive endorheic drainage basin covering approximately 15,540 km2 (~6000 square miles) in central Nevada. On the basis of biological and geomorphologi- cal evidence, Hubbs and Miller (1948) hypothesized that the valley had an historical connection from its southern end to the NUMBER 33 245 Pahranagat Valley (of the middle White River system) via Pe- noyer and Desert Valleys. This interpretation is supported by the distributions of aquatic Heteroptera. The former stream sys- tems that fed pluvial Lake Railroad are now almost entirely desiccated, as is the lake itself, which exists only in the form of a remnant playa. The two largest thermal springs in the valley are at Duckwater, in the far north, and at Lockes Ranch, along the western edge of the Lake Railroad playa below Black Rock Summit. Other thermal outflows along Hot Creek (not to be confused with a locality of the same name along the upper White River) and at Chimney Hot Springs, in the western and southern sections of Railroad Valley, respectively, have water temperatures too high to naturally support fish or aquatic Het- eroptera (although a fish refugium has been constructed below the outflow of the latter). However, complexes of cold springs, some with outflow creeks, occur near Currant along the eastern edge of the Lake Railroad playa and in Hot Creek Canyon above the major thermal outflow. A synopsis of the spring-fed habitats in Railroad Valley and their fish faunas was provided by Williams etal. (1985). The thermal springs at Duckwater are large and support dis- junct populations of a Belostoma species in the bakeril sarato- gae complex (Figure 6) and of Pelocoris biimpressus shoshone (Figure 8). Duckwater represents the most northern locality for the latter taxon. The thermal outflow at North Spring in the Lockes Ranch area supports a population of Nerthra martini (Figure 11); this locality also represents the most northern re- cord for N. martini in North America. Surveys of the cold springs east of the Lake Railroad playa have revealed no un- usual aquatic Heteroptera, but the cold spring outflows at mis- leadingly named Hot Creek contain a disjunct population of the FIGURE 11 .?Nerthra martini Todd. This littoral toad bug has a patchy distribu- tion from Mexico into the southern Great Basin. Highly isolated populations occur in thermal springs in the Railroad Valley, White River, and Amargosa River basins. FIGURE 12.?Rhagovelia distincta Champion. This small riffle bug is wide- spread in western North America but rare in the Great Basin. In the northern Great Basin it is known from only a few populations along the Provo and Weber Rivers in Utah, probably the result of recent trans-basin water diver- sions from the Green River catchment. In the southern Great Basin, by con- trast, this species occurs as naturally disjunct populations in thermal springs along the White and Amargosa Rivers. creeping water bug Ambrysus woodburyi, which is typical of the Colorado River drainage but is also known from cold- spring habitats in the White and Sevier River drainages. Am- brysus mormon (Figure 9), Rhagovelia distincta (Figure 12), and other species represented by disjunct populations in the thermal springs of the adjacent upper White River basin have not been collected in Railroad Valley, although additional sur- veys are needed, particularly of the Lockes Ranch area and of the cold springs near Currant. WHITE RIVER The White River is now a desiccated drainage that retains per- manent flow only in three separate spring zones. These occur along the upper segment of the drainage in the vicinity of Lund and Sunnyside, along the middle segment of the drainage in Pah- ranagat Valley, and along the lower reach of the river at Moapa Warm Springs. Each of these spring zones is separated from the others by long, dry sections of ancient river channel, and they contain individually distinctive assemblages of endemic or dis- junct aquatic Heteroptera species. In addition, the major eastern tributary of the White River, Meadow Valley Wash, contains several isolated reaches with cold, permanent, spring-fed flow, also separated by long stretches of dry channel. The upper White River valley contains a large concentration of both cold and thermal springs. The thermal springs emerge on the valley floor to the west of the dry river channel (Figure 2), and the cold springs lie along the base of the Egan Range. An overview of these habitats and their fish faunas was pro- vided by Williams et al. (1985). 246 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES The three major thermal outflows occur at Moorman Hot Spring, Hot Creek (not to be confused with the other Hot Creek in Railroad Valley), and Moon River (which lies on private land and has not been accessible for surveys). The spring at Hot Creek, now protected as a fish refuge, is very large and dis- charges into a deep creek and adjacent cool-water marsh. The outflow supports a disjunct population of Rhagovelia distincta (Figure 12), representing the most northern record of this spe- cies in Nevada, and a thermally dwarfed population of Am- brysus mormon. Moorman Hot Spring is smaller than the Hot Creek outflow but is also in relatively natural condition. A deep pool at the spring head drains into a small, formerly chan- nelized outflow creek that runs for a considerable distance be- fore entering a cool-water marsh. The most notable disjunct faunal element at this site is a Belostoma species in the bakeril saratogae complex (Figure 6), which is abundant in the out- flow channel. The cold springs of the upper White River drainage support an aquatic Heteroptera assemblage distinct from that seen in the thermal waters, including populations of certain species adapted to cold waters. Most notable among these is Ambrysus woodburyi, which is found at Sunnyside and Flagg Springs; this species also occurs along Meadow Valley Wash in the lower White River basin and, as previously noted, in the cold springs at Hot Creek Canyon in western Railroad Valley. These Nevada records aside, the next closest populations are far to the east, along the lower Virgin River in the Colorado River basin and along the Sevier River at Junction in the Bonneville Basin, suggesting some form of historic linkage between these areas. The next major spring complex, along the middle course of the White River, consists of three large thermal springs which emerge in Pahranagat Valley, near Hiko (Figure 2): Hiko Spring, Crystal Spring, and Ash Warm Springs. They contain a diverse array of thermally endemic or disjunct aquatic in- sect, snail, and fish taxa (Williams et al., 1985). Hiko Spring is on a private ranch and is the most disturbed and least ento- mologically interesting of the three springs in the valley. Be- cause of the highly altered nature of this habitat and the pres- ence of many introduced fish and aquatic plants, few aquatic insects now occur in the spring except for common generalist species. Crystal Spring lies in a grove of large cottonwoods and wil- lows just south of the junction of Nevada Highways 375 and 318. Two large spring pools discharge into a complex network of sloughs and flowing channels that finally coalesce into a sin- gle outflow, which exits to the east where there is a headgate. The area is in relatively natural condition, although the margins of the outflow channels are heavily overgrown with exotic aquatic plants, and the waters contain numerous introduced fishes, especially the convict cichlid Cichlasoma nigrofascia- tum (Gunther). The large spring pools support Belostoma of the bakerilsaratogae complex, and the outflow channels contain disjunct populations of Rhagovelia distincta (Figure 12) and Hydrometra australis (Figure 13), the latter being found else- where in Nevada only in the Muddy River Valley at the lower end of the White River drainage. Ash Warm Springs is the largest of the Pahranagat Valley springs. It has been the victim of numerous development at- tempts over the years, all of which have failed. Several spring outflows emerge from the base of a hill to the east of U.S. Highway 93 and flow into an extensive series of deep, tree- lined ponds. Species present here are Pelocoris biimpressus shoshone, which inhabits submerged vegetation along the margins of a small bayou at the northern end of the spring complex, and the thermally disjunct water beetle Stenelmis callida moapa La Rivers, found in the clear, flowing water of the spring heads in areas where a substrate of clean sand ex- ists. At the southern end of the ponds the combined waters from all the spring heads drain through a culvert under the highway and form a swift, shaded stream flowing down to the White River. This outflow stream harbors an extremely dense FIGURE 13.?Hydrometra australis Say. Although widespread in the southern half of the United States, this delicate, surface-dwelling insect is known from only three areas in the southern Great Basin: Saratoga Spring in southern Death Valley (Amargosa River basin); Crystal Spring along the middle course of the White River; and the Moapa Warm Springs valley along the Muddy River (a part of the lower White River system). NUMBER 33 247 population of Ambrysus mormon, the individuals of which are morphologically divergent from typical populations of this species (La Rivers, 1953) and which may deserve subspecies status. The terminal reach of the White River, known as the Muddy River, still connects to the Virgin River and thus to the Colo- rado River basin. As a result, the lower section of the White River system has had more potential for faunal interchange with the Colorado basin in recent times than any of the other areas under discussion. Nevertheless, Moapa Warm Springs (Figure 2) along this terminal reach retains perhaps the most distinctive aquatic fauna of any spring complex in the region. Many thermal endemics and disjuncts occur at this locality, no- tably the limnocorine naucorid Limnocoris moapensis (Figure 7), the riffle bug Rhagovelia becki (Polhemus, J.T., 1973) (Fig- ure 4), and the dytiscid beetle Neoclypeodytes discretus (Sharp), all of which are unknown elsewhere in the Virgin or Colorado River basins and which have their closest relatives or conspecific populations in eastern Mexico and central Texas. Other Heteroptera disjuncts found here are the backswimmer Buenoa omani, known elsewhere only from southwestern Cali- fornia and Mexico, and the toad bug Nerthra martini (Figure 11). In addition, several thermally endemic or disjunct water beetle taxa also occur at Moapa Warm Springs, including Mi- crocylloepus moapus moapus La Rivers (1949c) and Stenelmis callida moapa (La Rivers, 1949b), the latter of which is also found at Ash Warm Springs. Despite occasional collections by numerous specialists over the years and a recently completed year-long study conducted by R. Baumann (R. Baumann, pers. comm. to JTP, 1999), a comprehensive entomological inven- tory of this spring zone has not yet been published. In addition to the taxa listed above, the thermal outflows in this area also support mosquitoes in the Uranotaenia anhydor Dyar complex, the dragonfly Brechmorhoga mendax (Hagen), and the may- flies Isonychia intermedia (Eaton), Leptohyphes zalope Traver, and Camelobaetidus musseri (Traver and Edmunds), all of which can also be considered thermal disjuncts. The general setting of this spring oasis was described by Williams et al. (1985), and the habitat partitioning exhibited by its naucorid fauna was discussed by La Rivers (1950a). Near the town of Moapa, the White River is joined by Meadow Valley Wash, a drainage that occupies an extensive catchment along the Utah-Nevada border. Like the White River, this drainage makes a transition from low desert to up- land character as one progresses upstream; but unlike the White River, Meadow Valley Wash contains no thermal springs. Instead, several stretches of permanent flow are present from the headwaters to the lower reaches, all fed by cold-water springs. Surveys up the wash from Moapa to Cali- ente found that these permanent reaches support disjunct popu- lations of Ambrysus woodburyi and Belostoma Jlumineum, spe- cies typical of the Colorado River system. AMARGOSA RIVER The Amargosa River originates in the Bullfrog Hills of Ne- vada, flows southward for more than 150 km (crossing into California along the way), and hooks abruptly back to the north, terminating in Death Valley (Figure 2). Like the White River, the Amargosa is a desiccated drainage that retains per- manent flow in only a few sections. The major thermal springs in the Amargosa system occur along the upper middle section in a former tributary valley at Ash Meadows; on the lower middle section of the river in Tecopa and Shoshone and on a ranch at Resting Spring; and near the southern end of the drainage at Saratoga Spring in Death Valley. In addition, con- centrations of thermal springs occur at and near Furnace Creek in the middle section of Death Valley and at Grapevine Springs near the valley's northern end. Cool springs are also present along the upper course of the river near Beatty, but most of these have been extensively disturbed, and surveys have re- vealed no endemic aquatic Heteroptera in their waters. The Amargosa River basin is the most isolated of the three drainages under discussion, and it contains the richest concen- tration of endemic aquatic Heteroptera species. Situated in the lowest and hottest section of the North American continent, the basin undoubtedly retained a warm climate even during the Pleistocene, which, combined with the presence of constant- temperature thermal springs, allowed for the survival of a relict warm-water entomofauna. The largest concentration of thermal springs in the Amar- gosa system is found at Ash Meadows, where a complex of more than thirty seeps and springs supports a distinctive as- semblage of endemic plants, fish, and insects (Williams et al., 1985). The springs of Ash Meadows form three discrete groups defined by elevation and biota. The highest is Devils Hole Spring, which contains the endemic Devils Hole pupfish Cyprinodon diabolis Wales and the endemic riffle beetle Stenelmis callida callida Chandler (1949) but no aquatic Het- eroptera. At a slightly lower elevation, the second group com- prises several thermal springs?including North Scruggs, South Scruggs, North Indian, South Indian, Marsh, and School Springs?which support an endemic naucorid, Ambrysus relic- tus (Polhemus and Polhemus, 1994). Many of these springs also harbor thermally disjunct populations of a Belostoma spe- cies in the bakerilsaratogae complex and Pelocoris biimpres- sus biimpressus. The group at the lowest elevation consists of numerous springs with large thermal outflows, including Point of Rock Springs?the sole habitat for the endemic naucorid Ambrysus amargosus, whose habitat preferences were de- scribed in detail by La Rivers (1953). Many of the larger springs in Ash Meadows, including Point of Rock, were exten- sively modified in past decades to support planned agricultural and residential developments. As a result, Ambrysus amargo- sus was driven nearly to the point of extinction (Polhemus, D.A., 1993) and is now classified as a Federally threatened species; the disjunct populations of Rhagovelia distincta and 248 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Pelocoris biimpressus biimpressus that formerly occurred at Point of Rock Springs were extirpated. The thermal springs of the middle Amargosa River are con- centrated in the towns of Shoshone and Tecopa and at the nearby ranch at Resting Spring. All have suffered from munici- pal, recreational, or agricultural development. Access to Rest- ing Spring, which lies on private land, has been restricted by the ranch owner. The only aquatic Heteroptera species recorded from the site is the riffle bug Rhagovelia distincta, collected in 1956. Although widespread in the western United States, this species is found only as extremely localized populations in iso- lated thermal waters within the Great Basin (see previous dis- cussion concerning this species in the White River system). Be- cause all known localities for R. distincta in the southern Great Basin support additional thermal disjuncts or endemics, its presence at Resting Spring implies that this site may harbor ad- ditional Heteroptera of interest. The springs at Tecopa, a few miles west of Resting Spring, are the hottest in the Amargosa system, having an emergent temperature of 43? C; they previously supported an endemic pupfish subspecies that is now extinct. Pools near the spring sources also supported extremely large concentrations of Pelo- coris biimpressus biimpressus in the past, but its habitat there was completely eliminated in the course of development of rec- reational spas. Other less disturbed thermal outflows lie to the north of the Tecopa Hot Springs resort along the paved high- way between Shoshone and Tecopa, and these support species of both Pelocoris and Belostoma, although the occurrence of the latter appears to be seasonal (D. Threloff, pers. comm. to DAPandJTP, 1999). The thermal spring at Shoshone was similarly modified for municipal use: the spring head was boxed and the outflow creek is mostly piped to supply a local swimming pool. A small section of less disturbed habitat, consisting of a deep pond and a short stretch of flowing channel, remains between the spring head and the pipe collector system. The pond supports a popu- lation of Pelocoris biimpressus biimpressus, and the flowing channel contains Ambrysus mormon. Along the lower Amargosa River, the most notable thermal site is Saratoga Spring, which emerges at the base of the Ibex Hills; this spring habitat was described in detail by Belkin and McDonald (1956). The aquatic Heteroptera fauna of the spring is unique and distinctive, including a highly disjunct population of the giant water bug Lethocerus angustipes (Figures 5, 14), which is otherwise not known outside of central Mexico (Menke, 1963). Other disjunct species here are Pelocoris biim- pressus biimpressus, Belostoma saratogae (for which this is the type locality), Nerthra martini, and Hydrometra australis (see Table 1), plus a population of metallic-colored mosquitoes in the Uranotaenia anhydor complex (Belkin and McDonald, 1956). A second focus of endemism in the lower Amargosa River basin is found along the course of Furnace Creek. The thermal outflows concentrated in this area?most notably Travertine, Cow Creek, and Limestone springs?harbor the endemic nau- FlGURE 14.?Lethocerus amehcanus (Leidy). This giant water bug ranges across temperate North America, including the northern portion of the Great Basin. It is absent in the southern Great Basin, where the genus is represented by only a single highly disjunct population of L. angustipes (Mayr) in the Amargosa River basin; the closest known populations of this latter species are found in central Mexico. corid Ambrysus funebris and a disjunct population of Nerthra martini (Figure 11). The spring habitat at Cow Creek was de- scribed in detail by La Rivers (1949a, 1951). The most significant thermal springs in northern Death Val- ley are found at Grapevine Springs, which consist of many small thermal outflows perched on a bench above the valley floor northeast of Stovepipe Wells. The overall habitat here is in good condition, although many of the spring heads are cov- ered by introduced palms and other encroaching vegetation and, thus, do not present an original aspect. Similar outflows once occurred along the bed of Grapevine Creek in Grapevine Canyon to the east, but these were extensively altered and di- verted during the last hundred years to supply the small settle- ment at Scotty's Castle; as a result, they are not known to retain any of their former thermally dependant aquatic Heteroptera species. The remaining outflow channels in the Grapevine area are, for the most part, heavily overgrown and thus are difficult to sample. At a few spots the waters have pooled to form small ponds that support populations of Pelocoris biimpressus shos- hone. This is the most northerly occurrence of the genus in Cal- ifornia and the only place in the Amargosa River basin where P. biimpressus shoshone occurs. All other thermal outflows in this basin instead support P. biimpressus biimpressus (Figure 3), suggesting that the Amargosa basin may be a composite drainage, with its northwestern limb having had a separate his- tory more closely linked to the Railroad Valley and White River drainages. NUMBER 33 249 Biogeographic Implications The various internal drainages of the Basin and Range Prov- ince grouped under the term "Great Basin" form a definable unit in a topographic sense, but not biologically. As we have shown, the aquatic Heteroptera that are the subject of this paper exhibit an obvious faunal dichotomy between the cold basins to the north of the Caliente-Enterprise accommodation zone, many of which formerly drained to the Snake or Columbia Riv- ers, and the warm basins to the south, some of which have had ancient drainage relationships with the Colorado or Mojave Rivers, or even the Rio Grande (McKee et al., 1967; Lucchitta, 1990). The northern basins contain a broadly distributed cold- water biota devoid of notable endemic or disjunct species, whereas the thermal refugia of the southern basins harbor many such taxa. As a result, the northern basins provide little in the way of a compelling biogeographic story, whereas the southern basins contain intriguing faunal evidence of drainage basin de- velopment and capture during the Tertiary. Two competing hypotheses may be advanced to explain the current distribution of disjunct faunal elements in the southern basins: long-distance dispersal and vicariance. The first postu- lates intermittent chance colonizations of drainage systems over time by faunal passage across existing barriers, such as deserts or drainage divides. The latter implies an initial widespread dis- tribution of biotas, followed by subsequent fragmentation due to the imposition of barriers produced by tectonic or climatic events. A corollary axiom is that, in general, individual species will tend to disperse, at widely separate times, whereas entire communities will vicariate, more or less simultaneously. We favor the vicariance hypothesis to explain the distribu- tions of thermal relicts in the southern Great Basin, because this fauna represents entire disjunct communities of aquatic organ- isms with common geographic affinities?not the result one would expect from multiple episodes of chance long-distance dispersal. In addition to the evidence from aquatic Heteroptera presented in this paper, analyses of mitochondrial DNA varia- tion in Death Valley pupfishes also support the hypothesis of an eastern ancestor for some of these pupfish species, with the sis- ter group of the Amargosa River basin component of this radia- tion apparently occurring in the Laguna de Guzman area, east of the present continental divide in northern Mexico, rather than in the Colorado River basin (Echelle and Dowling, 1992). Simi- larly, Parenti (1981) presented a morphologically based phylo- genetic analysis showing that the cyprinodontiform fish genera Empetrichthys and Crenichthys, endemic to the southern Great Basin, had their sister group relationships with the Goodeidae of the Mexican Plateau. On the basis of morphological evidence, Hershler and Pratt (1990) also hypothesized an eastern deriva- tion for the ancestors of spring snail species now found at Moapa and Ash Meadows, via southeastward drainage connec- tions across Arizona in the Miocene. A vicariance scenario accommodating these relationships re- quires initial dispersal from the southeast, followed by subse- quent isolation. The geological evidence is quite clear that in the late Oligocene to early Miocene, before the onset of crustal ex- tension in the Great Basin, the terrain west of the Colorado Pla- teau was higher than the plateau itself, and sedimentation oc- curred in an easterly direction?a pattern that continued at least until the emplacement of the Peach Springs Tuff at 18 Ma (Young, 1970; Lucchitta, 1990). Certain authors have proposed that what is now the upper basin of the Colorado River drained eastward during that period, possibly to the Rio Grande via the Little Colorado River (McKee et al., 1967; Nilsen and McKee, 1979) (Figure 15), and that Oligocene drainage networks in southern Nevada could have been integrated with it. Such a sce- nario is consistent with the current distributions of Pelocoris and Limnocoris, which are absent in the warm streams of cen- tral Arizona (Figure 3), indicating that they dispersed from southeast to northwest along drainages lying north of the current Gila River. Any southeast-flowing connection to the Rio Grande, if it ex- isted, was probably severed in the late Oligocene to early Mio- cene by a region-wide episode of rifting, orogeny, and volcan- ism associated with the extension of the Basin and Range Province (Mutschler et al., 1998), in particular the eruption of the Mogollon-Datil volcanic field, which emplaced lava 3-5 km thick over southwestern New Mexico between 36 and 24 Ma (Schneider and Keller, 1994) (Figure 15). Subsequent ter- tiary volcanism in this region underwent a progressive north- ward migration out of the Arizona section of the Basin and Range Province and onto the southern margin of the Colorado Plateau from 15 to 2 Ma, further altering the course of the Little Colorado River (Moyer and Nealey, 1989; Mutschler et al., 1998). During this period, and before the establishment of a westward-flowing Colorado River, there is evidence that a for- merly eastward-flowing drainage may have pooled in the Hopi Buttes area from 12 to 4 Ma, forming the lacustrine facies of the Bidahochi Formation in northeastern Arizona and northwestern New Mexico (Hunt, 1969; Miller, 1981; Love, 1989). The pos- tulated Lake Bidahochi, depicted in Figure 15, thus could have continued to provide potential aquatic stepping stones for fauna moving northwestward from the Rio Grande Rift until 4 Ma, by which time the formation of the Grand Canyon and the gradual westward reorganization of regional drainage was well under- way (Lucchitta, 1990). After crustal extension, we hypothesize that vicariance oc- curred as Railroad Valley and the White River both became iso- lated endorheic basins, probably by the end of the Miocene, with the latter finally reestablishing an outlet to a reorganized, westward-flowing Colorado system via headwater extension of the Virgin River sometime in the Pliocene. The stratigraphic character of the Muddy Creek Formation in the Grand Wash Trough indicates interior drainage at the lower end of the White River basin throughout the Miocene, with through drainage to the Colorado River occurring no earlier than 5-6 Ma (Luc- chitta, 1972, 1990; Bohannon et al., 1993). The absence of any of the signature Railroad Valley and White River thermal en- demics or disjuncts in the current Colorado system (Figures 3, 250 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 15.?Late Miocene to Late Pliocene topographic features in southwestern North America. The present Great Basin is also depicted, with division into northern basins (area marked by right-leaning slashes) and south- ern basins (left-leaning slashes) as discussed in text. Black shading indicates Late Cenozoic basalt flows, emplaced 25-4 Ma (after Mutschler et al., 1998). Numbered features are as follows: 1 = course of Little Colo- rado River (gray line with arrowhead indicates hypothesized direction of southeasterly drainage toward the Rio Grande system (after McKee et al., 1967); 2 = hypothesized location of Lake Bidahochi (horizontal lines), 12-4 My (after Hunt, 1969; Love, 1989); 3 = location of Mogollon-Datil volcanic field (square dots), emplaced 36-24 Ma (after Schneider and Keller, 1994; Mutschler et al., 1998). 4), despite the modern-day presence of a flowing-water link via the Muddy River, argues that the reconnection of the lower White River may have come very late; the current outlet lies along the highly active Lake Mead fault zone, where crustal blocks such as Frenchman Mountain have been transported up to 80 km from their former location east northeast of their present position in the last 14 My (Duebendorfer et al., 1998), resulting in extensive rearrangement of local drainage net- works. The Amargosa River basin may have had a more isolated po- sition, as indicated by its faunal endemism, but it nevertheless exhibits faunal ties to the White River catchment and Railroad Valley. The distribution of Pelocoris subspecies (Figure 3) sug- gests that the current basin is a composite and that the northern section of Death Valley near Grapevine Canyon was previously linked in some fashion to the Railroad Valley and White River drainage system. Additional surveys of warm springs in the ba- sins northeast of the Amargosa drainage would prove useful in determining if Pelocoris populations are present in those areas, which might indicate a corridor that linked the Grapevine Springs populations to those of Railroad Valley and the White River. The presence of strikingly disjunct populations of otherwise Mexican taxa in the southern end of the Amargosa system (Fig- ure 5), coupled with the absence of strong relationships to the modern Colorado River basin, again indicates the relative NUMBER 33 251 youth of the Colorado River in its current, westward-flowing configuration and suggests a former Amargosa River outlet to the south or southeast. One clue regarding the location is pro- vided by the presence of another set of Mexican disjuncts along the upper Verde River system in central Arizona. This area con- tains populations of the backswimmer Martarega mexicana Truxal and the riffle bug Rhagovelia choreutes Hussey, both of which are typically Mexican taxa that do not occur elsewhere in Arizona or the Colorado River system (Menke and Truxal, 1966; Polhemus, J.T., 1966; Polhemus, D.A., 1997). Although neither of these species occurs in the Amargosa River basin, their presence as another cluster of disjunct Mexican taxa nev- ertheless suggests that a warm-water corridor to northern Mex- ico once ran through that region. Superimposed upon these presumably Oligocene to Pliocene warm-water vicariance patterns is another, apparently more re- cent pattern involving the isolation of cold-water taxa at vari- ous widely separated localities in the southern Great Basin, as exemplified by the distribution of Ambrysus woodburyi dis- cussed previously. This newer pattern may have resulted from the existence of cold-water drainage networks that formed dur- ing wet periods in the Pleistocene or late Pliocene. We infer that these conjectured drainages were not extensive or persis- tent enough to permit invasion into the southern basins by most of the aquatic insect biota of the modern Colorado River, and they were evidently too cold for utilization by species associ- ated with thermal waters. Our entomological evidence implies this second round of vicariance events occurred in the White River and Railroad Valley basins at some point after the frag- mentation of the warm-water biota. This hypothesis also im- plies that the Railroad Valley and White River drainage sys- tems were hydrologically integrated in post-Miocene times, and perhaps as late as the Pleistocene. Biological evidence for the sequence of events that eventu- ally isolated the southern basins from each other and from the previous river systems to the south may be difficult to deduce from phylogenetic analyses of the aquatic Heteroptera species involved. The thermal disjuncts do not form monophyletic clades, but instead they are either endemic isolates within gen- era with many other widespread species or locally disjunct populations which, though profoundly isolated, have not un- dergone extensive morphological differentiation. In some taxa, such as the Belostoma species, there is evidence of incipient speciation that has not developed into any geographic pattern. Molecular approaches offer the promise of additional resolu- tion of relationships between such potentially confusing and yet apparently conspecific populations. Until such data are available, however, our current biogeographic analysis of these basins, at least in terms of the evidence from aquatic Het- eroptera, must rest primarily on faunal similarity and sister group distributions, an approach which, though less rigorous than strict cladistic biogeography, can still discern the large- scale patterns of historical relationship within these systems. Many of the discoveries upon which the foregoing zoogeo- graphic hypotheses are based were made only during the last decade. Given the size and complexity of the Great Basin and the number of springs remaining to be sampled for aquatic in- sects, it is likely that important information will come to light in the future. As with all such hypotheses, our zoogeographic interpretations are open to further testing as additional collec- tions and analyses are made. We have little doubt that new evi- dence of the faunal signatures left by ancient lake and river sys- tems will continue to be uncovered amid the basins and ranges of this most remote segment of the American West. Biogeographic Summary In sum, the following biogeographic patterns are suggested by the distributions of aquatic Heteroptera in the Great Ba- sin: 1. The northern Great Basin, consisting of the Bonneville Basin, Lahontan Basin, and the basins of southeastern Oregon, contains a widespread assemblage of typical north-temperate species and lacks significant regional endemism. 2. The Sevier River, draining into the Bonneville Basin, contains species shared with the southern Great Basin and the Colorado River system and absent elsewhere in the northern Great Basin, indicating that this river may have had an histori- cal relationship with the upper Virgin River drainage. 3. Railroad Valley and the White River basin contain a unique set of disjunct taxa allied to species occurring in Texas and northern Mexico but absent in the Colorado River basin, indicating that these units had an historical relationship with the Rio Grande drainage and that modern integration of the White River and the Colorado River via the Virgin River is of very recent origin. 4. The Amargosa River basin contains a distinctive assem- blage of endemic and disjunct taxa absent in the Colorado River basin but in some cases found in northern and western Mexico, thus indicating that this basin was isolated before for- mation of the modern Colorado River and also that it had an historic outlet to the south or southeast. 5. Northern Death Valley contains taxa also found in Rail- road Valley and the White River basin but absent in the remain- der of the Amargosa River system, suggesting that modern Death Valley is a composite of previously distinct drainage ba- sins that had disparate histories. Acknowledgments The information upon which this report is based was col- lected by the authors during a period of thirty years. These sur- veys would not have been possible without the generous assis- tance of many persons, including Doug Threloff, National Park Service, Death Valley; Dave Ledig, U.S. Fish and Wildlife Ser- vice (USFWS), Ash Meadows; Ken Voget, USFWS, Las Ve- gas; and George F. Edmunds, Jr., University of Utah. We are 252 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES also indebted to R.W. Baumann, Brigham Young University, for sharing the results of a recent study of the insect fauna of the Muddy River basin, and to Warren Steiner, Smithsonian In- stitution, Washington, D.C., for providing us with additional re- cently collected material from this same region. We also ac- knowledge the contribution of the Smithsonian Institution's C.J. Drake fund, which funded certain later phases of our Great Basin survey work. Finally, we thank Robert Hershler of the Smithsonian Institu- tion for organizing the Great Basin Aquatic System History (GBASH) symposium in late 1997, at which the results of this analysis were initially presented. Literature Cited Axen, G.J. 1998. The Caliente-Enterprise Zone, Southeastern Nevada and South- western Utah. In J.E. Faulds and J.H. Stewart, editors, Accommoda- tion Zones and Transfer Zones: The Regional Segmentation of the Basin and Range Province. Special Papers of the Geological Society of America, 323:181-194. Belkin, J.N., and W.A. McDonald 1956. A Population of Uranotaenia anhydor from Death Valley, with De- scriptions of All Stages and Discussion of the Complex (Diptera, Culicidae). Annals of the Entomological Society of America, 49:105-131. Bohannon, R.G., J.A. Grow, J.J. Miller, and R.H. Blank, Jr. 1993. Seismic Stratigraphy and Tectonic Development of Virgin River De- pression and Associated Basins, Southeastern Nevada and North- western Arizona. Bulletin of the Geological Society of America, 105:501-520. Chandler, H.P. 1949. A New Species of Stenelmis from Nevada (Coleoptera, Elmidae). Pan-Pacific Entomologist, 25(3): 133-136. De Carlo, J.A. 1951. Genera Limnocoris Stal. Mision de Estudios Patalogia Regional Ar- gentina, 22:41-51. Drake, C.J., and H.M. Harris 1934. The Gerrinae of the Western Hemisphere (Hemiptera). Annals of the Carnegie Museum, 23:179?240. Duebendorfer, E.M., L.S. Beard, and E.I. Smith 1998. Restoration of Tertiary Deformation in the Lake Mead Region, Southern Nevada: The Role of Strike Slip Transfer Zones. In J.E. Faulds and J.H. Stewart, editors, Accommodation Zones and Trans- fer Zones: The Regional Segmentation of the Basin and Range Province. Special Papers of the Geological Society of America, 323: 127-148. Echelle, A.A., and T.E. Dowling 1992. Mitochondrial DNA Variation and Evolution of the Death Valley Pupfishes (Cyprinodon, Cyprinodontidae). Evolution, 46(1): 193-206. Guerin-Meneville, F.E. 1857. Ordre des Hemipteres, Latr. premiere section: Heteropteres, Latr. In M.R. Sagra, Historie physique, politique et naturelle de I'He Cuba, 7:359-424. Paris: Arthus Bertrand. Harris, H.M., and W.E. Shull 1944. A Preliminary List of Hemiptera of Idaho. Iowa State College Jour- nal of Science, 18:199-208. Henry, T.J., and R.C. Froeschner 1988. Catalog of the Heteroptera, or True Bugs, of Canada and the Conti- nental United States, i-xix + 1-958 pages. New York: E.J. Brill. Hershler, R., and W.L. Pratt 1990. A New Pyrgulopsis (Gastropoda: Hydrobiidae) from Southeastern California, with a Model for Historical Development of the Death Valley Hydrographic System. Proceedings of the Biological Society of Washington, 103(2):279-299. Hubbs, C.L., and R.R. Miller 1948. The Zoological Evidence: Correlation between Fish Distribution and Hydrographic History in the Desert Basins of Western United States. In The Great Basin, with Emphasis on Glacial and Postgla- cial Times. Bulletin of the University of Utah, 38(20): 18-166. Hungerford, H.B. 1934 (" 1933"). The Genus Notonecta of the World. Bulletin of the Univer- sity of Kansas, 34(5):5-195. [Date on title page is 1933; actual date of publication is 1934.] 1948. The Corixidae of the Western Hemisphere (Hemiptera). University of Kansas Science Bulletin, 32:1-827. Hunt, C.B. 1969. Geological History of the Colorado River. In The Colorado River Region and John Wesley Powell. United States Geological Survey Professional Paper, 669:59-130. Knowlton, G.F. 1956. Some Hemiptera and Homoptera of Utah?1956. Utah State Agri- cultural College Extension Service, Logan, Utah, Mimeograph Se- ries, 158:1-11. La Rivers, I. 1948. A New Species of Pelocoris from Nevada, with Notes on the Genus in the United States (Hemiptera: Naucoridae). Annals of the Ento- mological Society of Washington, 41(3):371?376. 1949a ("1948"). A New Species of Ambrysus from Death Valley, with Notes on the Genus in the United States (Hemiptera: Naucoridae). Bulletin of the Southern California Academy of Science, 47(3): 103-110. [Date on title page is 1948; actual date of publication is 1949.] 1949b. A New Subspecies of Stenelmis from Nevada (Coleoptera: Dryopi- dae). Proceedings of the Entomological Society of Washington, 51:218-224. 1949c. A New Species of Microcylloepus from Nevada (Coleoptera: Dry- opidae). Entomological News, 60:205-209. 1950a. The Meeting Point of Ambrysus and Pelocoris in Nevada (Hemi- ptera: Naucoridae). Pan-Pacific Entomologist, 26(1): 19-21. 1950b. A New Naucorid Genus and Species from Nevada (Hemiptera). An- nals of the Entomological Society of America, 43:368?373. 1951. A Revision of the Genus Ambrysus in the United States (Hemiptera: Naucoridae). University of California Publications in Entomology, 8(7):277-338. 1953. New Gelastocorid and Naucorid Records and Miscellaneous Notes, with a Description of the New Species, Ambrysus amargosus (Hemiptera: Naucoridae). Wasmann Journal of Biology, 11(1): 83-96. 1956. A New Subspecies of Pelocoris shoshone from the Death Valley Drainage (Naucoridae: Hemiptera). Wasmann Journal of Biology, 14(1):155-158. 1963. Two New Ambrysi (Hemiptera: Naucoridae). Occasional Papers of the Biological Society of Nevada, 1:1-7. Lauck, D.R. 1964. A Monograph of the Genus Belostoma, Part III: B. triangulum, NUMBER 33 253 bergi, minor, bifoveolatum, and flumineum Groups. Bulletin of the Chicago Academy of Science, 11:102-154. 1979. Family Corixidae: Water Boatmen. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of California (Heteroptera: Hemiptera). Bulletin of the California Insect Survey, 21:87-123. Love, D.W. 1989. Bidahochi Formation: An Interpretive Summary. New Mexico Geo- logical Society Guidebook, 40th Field Conference, Southeastern Colorado Plateau, 1989:273-280. Lucchitta, I. 1972. Early History of the Colorado River in the Basin and Range Prov- ince. Bulletin of the Geological Society of America, 83:1933-1948. 1990. History of the Grand Canyon and the Colorado River in Arizona. In S. S. Beus and M. Morales, editors, Grand Canyon Geology, pages 311-332. New York and Oxford: Oxford University Press. McKee, E.D., R.F. Wilson, W.J. Breed, and C.S. Breed 1967. Evolution of the Colorado River in Arizona. Bulletin of the Museum of Northern Arizona, 44:1-67. Menke, A.S. 1958. A Synopsis of the Genus Belostoma Latrielle of America North of Mexico, with the Description of a New Species. Bulletin of the Southern California Academy of Science, 57:154-171. 1963. A Review of the Genus Lethocerus in North and Central America, Including the West Indies (Hemiptera: Belostomatidae). Annals of the Entomological Society of America, 56:261-267. 1979a. Family Nepidae: Water Scorpions. In A.S. Menke, editor, The Semi- aquatic and Aquatic Hemiptera of California (Heteroptera: Hemi- ptera). Bulletin of the California Insect Survey, 21:70?75. 1979b. Family Belostomatidae: Giant Water Bugs, Electric Light Bugs, Toe Biters. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemi- ptera of California (Heteroptera: Hemiptera). Bulletin of the Califor- nia Insect Survey, 21:76-86. 1979c. Family Gelastocoridae: Toad Bugs. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of California (Heteroptera: Hemiptera). Bulletin of the California Insect Survey, 21:126-130. Menke, A.S., and F.S. Truxal 1966. New Distribution Data for Martarega, Buenoa, and Abedus, Includ- ing the First Record of the Genus Martarega in the United States (Hemiptera: Notonectidae, Belostomatidae). Los Angeles County Museum Contributions in Science, 106:1?6. Miller, R.R. 1981. Coevolution of Deserts and Pup fishes (Genus Cyprinodon) in the American Southwest. In R.J. Naiman and D.L. Soltz, editors, Fishes in North American Deserts, pages 39-94. New York: John Wiley and Sons. Montandon, A.L. 1898. Hemiptera Cryptocerata: notes et descriptions d'especes nouvelles. Bulletin de la Societe des Sciences de Bucarest-Roumanie, 7: 282-290. 1909. Naucoridae: descriptions d'especes nouvelles. Bulletin de la Societe des Sciences de Bucarest-Roumanie, 18:43-61. 1910. Hydrocorises de l'Amerique du Nord: notes et descriptions d'es- peces nouvelles. Bulletin de la Societe des Sciences de Bucarest- Roumanie, 18:180-191. Moyer, T.C., and L.D. Nealey 1989. Regional Compositional Variations of Late Tertiary Biomodal Rhyo- lite Lavas across the Basin Range/Colorado Plateau Boundary in Western Arizona. Journal of Geophysical Research, series B, 94:7799-7816. Mutschler, F.E., E.E. Larson, and D.L. Gaskill 1998. The Fate of the Colorado Plateau?A View from the Mantle. In J.D. Friedman and A.C. Huffman, Jr., editors, Laccolith Complexes of Southeastern Utah: Time of Emplacement and Tectonic Setting? Workshop Proceedings. Bulletin of the United States Geological Survey, 2158:203-222. Nilsen, T.H., and E.H. McKee 1979. Paleogene Paleogeography of the Western United States. In J.M. Armentrout, M.R. Cole, and H. TerBest, editors, Cenozoic Paleo- geography of the Western United States. Society of Economic Pale- ontologists and Mineralogists, Pacific Section, Pacific Coast Paleogeography Symposium, 3:257-276. Parenti, L.R. 1981. A Phylogenetic and Biogeographic Analysis of Cyprinodontiform Fishes (Teleostei, Atherinomorpha). Bulletin of the American Mu- seum of Natural History, 168:335-557. Polhemus, D.A. 1993. Conservation of Aquatic Insects: Worldwide Crisis or Localized Threats? American Zoologist, 33:588-598. 1997. Systematics of the Genus Rhagovelia Mayr (Heteroptera: Veliidae) in the Western Hemisphere (Exclusive of the angustipes Complex). 366 pages. Lanham, Maryland: Entomological Society of America. Polhemus, J.T. 1966. Some Hemiptera New to the United States (Notonectidae, Saldidae). Proceedings of the Entomological Society of Washington, 68:57. 1973. Notes on Aquatic and Semiaquatic Hemiptera from the Southwest- ern United States (Insecta: Hemiptera). Great Basin Naturalist, 33(2): 113-119. 1979. Family Naucoridae: Creeping Water Bugs, Saucer Bugs. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of Califor- nia (Heteroptera: Hemiptera). Bulletin of the California Insect Sur- vey, 21:131-138. Polhemus, J.T., and H.C. Chapman 1979a. Family Saldidae: Shore Bugs. In A.S. Menke, editor, The Semi- aquatic and Aquatic Hemiptera of California (Heteroptera: Hemi- ptera). Bulletin of the California Insect Survey, 21:16-33. 1979b. Family Mesoveliidae: Water Treaders. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of California (Heteroptera: Hemiptera). Bulletin of the California Insect Survey, 21:39-42. 1979c. Family Hydrometridae: Marsh Treaders. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of California (Heteroptera: Hemiptera). Bulletin of the California Insect Survey, 21:43?45. 1979d. Family Veliidae: Small Water Striders, Water Crickets, Riffle Bugs. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of California (Heteroptera: Hemiptera). Bulletin of the California In- sect Survey, 21:49-57. 1979e. Family Gerridae: Water Striders, Pond Skaters, Wherrymen. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of Califor- nia (Heteroptera: Hemiptera). Bulletin of the California Insect Sur- vey, 21:58-69. Polhemus, J.T., and D.A. Polhemus 1994. A New Species of Ambrysus Stal from Ash Meadows, Nevada (Het- eroptera: Naucoridae). Journal of the New York Entomological Soci- ety, 102:261-265. Polhemus, J.T., and R.W. Sites 1995. The Identity of Pelocoris biimpressus Montandon and Synonymy of Pelocoris Species in the Southwestern United States (Heteroptera: Naucoridae). Proceedings of the Entomological Society of Washing- ton, 97:654-658. Rowley, P.D., C.G. Cunningham, T.A. Steven, H.H. Mehnert, and C.W. Naeser 1998. Cenozoic Igneous and Tectonic Setting of the Marysvale Volcanic Field and Its Relation to Other Igneous Centers in Utah and Ne- vada. In J.D. Friedman and A.C. Huffman, Jr., editors, Laccolith Complexes of Southeastern Utah: Time of Emplacement and Tec- tonic Setting?Workshop Proceedings. Bulletin of the United States Geological Survey, 2158:167-201. Rowley, P.D., P.W. Lipman, H.H. Mehnert, D.A. Lindsey, and J.J. Andersen 1978. Blue Ribbon Lineament, an East-Trending Structural Zone within the Pioche Mineral Belt of Southwestern Utah and Eastern Nevada. United States Geological Survey Journal of Research, 6:175-192. 254 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Schneider, R.V., and G.R. Keller 1994. Crustal Structure of the Western Margin of the Rio Grande Rift and the Mogollon-Datil Volcanic Field, Southwestern New Mexico and Southeastern Arizona. In G.R. Keller and S.M. Cather, editors, Ba- sins of the Rio Grande Rift: Structure, Stratigraphy, and Tectonic Setting. Special Papers of the Geological Society of America, 291:207-226. Sites, R.W., and J.T. Polhemus 1994. Nepidae (Hemiptera) of the United States and Canada. Annals of the Entomological Society of America, 87(l):27-42. Sites, R.W., and M.R. Willig 1994. Efficacy of Mensural Characters in Discriminating among Species of Naucoridae (Insecta: Hemiptera): Multivariate Approaches and Ontogenetic Perspectives. Annals of the Entomological Society of America, 87(6):803-814. Smith, C.L. 1980. A Taxonomic Revision of the Genus Microvelia Westwood (Het- eroptera: Veliidae) of North America, Including Mexico, xv + 372 pages. Doctoral dissertation, University of Georgia, Athens, Geor- gia. Stewart, J.H. 1998. Regional Characteristics, Tilt Domains, and Extensional History of the Late Cenozoic Basin and Range Province, Western North Amer- ica. In J.E. Faulds and J.H. Stewart, editors, Accommodation Zones and Transfer Zones: The Regional Segmentation of the Basin and Range Province. Special Papers of the Geological Society of Amer- ica, 323:47-74. Stonedahl, G.M., and J.D. Lattin 1982. The Gerridae or Water Striders of Oregon and Washington (Hemi- ptera: Heteroptera). Technical Bulletin of the Oregon State Univer- sity Agricultural Experiment Station, 144:1-36. 1986. The Corixidae of Oregon and Washington (Hemiptera: Heteroptera). Technical Bulletin of the Oregon State University Agricultural Ex- periment Station, 150:ii-iv + 1-84. Truxal, F.S. 1953. A Revision of the Genus Buenoa (Hemiptera, Notonectidae). Uni- versity of Kansas Science Bulletin, 35:1351-1523. 1979. Family Notonectidae: Backswimmers. In A.S. Menke, editor, The Semiaquatic and Aquatic Hemiptera of California (Heteroptera: Hemiptera). Bulletin of the California Insect Survey, 21:139-147. Usinger, R.L. 1956. Aquatic Hemiptera. In R.L. Usinger, editor, Aquatic Insects of Cali- fornia, pages 182-228. Berkeley and Los Angeles: University of California Press. Williams, J.E., D.B. Bowman, J.E. Brooks, A.A. Echelle, R.J. Edwards, D.A. Hendrickson, and J.J. Landye 1985. Endangered Aquatic Ecosystems in North American Deserts with a List of Vanishing Fishes of the Region. Journal of the Arizona-Ne- vada Academy of Sciences, 20:1-62. Young, R.A. 1970. Geomorphological Implications of Pre-Colorado and Colorado Trib- utary Drainage in the Western Grand Canyon Region. Plateau, 42:107-117. Biogeography of Great Basin Aquatic Snails of the Genus Pyrgulopsis Robert Her shier and Donald W. Sada ABSTRACT Snails of the genus Pyrgulopsis (Caenogastropoda: Hydrobi- idae) are among the richest elements of aquatic biodiversity in the Great Basin (80 Holocene species), with most species endemic to single springs, spring systems, or drainage basins. Pyrgulopsis is an old Great Basin group, with a fossil record extending into the late Miocene (Truckee Formation). These tiny, gill-breathing gas- tropods (commonly known as springsnails) are obligately aquatic throughout their life cycle which, together with the above features, suggests that their biogeographic patterns are highly informative with respect to late Cenozoic hydrographic history. Distributions of Pyrgulopsis species define seven regions of endemism within the Great Basin. Five of these correspond to drainages harboring concentrations of other endemic aquatic biota (Death Valley system, Lahontan Basin, Bonneville Basin, Railroad Valley, upper White River basin), whereas snail endemism in Dixie and Steptoe basins appears to be unique. Each of the three largest regions (Death Valley system, Lahontan Basin, Bonneville Basin) contains two or three subregions of Pyrgulopsis endemism, most of which are not paralleled by other aquatic biota. Pyrgulopsis biogeography in part conforms to currently accepted interpretations of pluvial drainage in the Great Basin, but locally suggests different, presumably older relationships. Various historical relationships with adjacent external drainages are implied, although snail biogeography provides surprisingly little evidence of prior linkage with the Sacramento and upper Colorado River basins. In contrast to the fish-hook track, which is attributed to various mollusks and fishes and links the Snake River basin with the eastern Bonneville and western Lahontan Basins, Pyrgu- lopsis provides evidence of more continuous prior drainage inte- gration across the northern boundary of the Great Basin. Robert Hershler, Department of Systematic Biology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560- 0163, USA; e-mail: hershler.robert@nmnh.si.edu. Donald W. Sada, Division ofHydrologic Sciences, Desert Research Institute, 2215 Rag- gio Parkway, Reno, Nevada 89512-1095, USA. Reviewers: Marith C. Reheis, United States Geological Survey, MS- 980, Federal Center, Box 25046, Denver, Colorado 80225, USA; Rich- ard P. Vari, Department of Systematic Biology, National Museum of Natural History, Smithsonian Institution, Washington, D.C, 20560- 0159, USA; and two anonymous reviewers. Introduction Gastropods of the family Hydrobiidae (commonly known as springsnails) are a highly suitable group for biogeographic evaluation of historical relationships among drainages of the Great Basin. These gill-breathing snails are obligate to perma- nent waters and disperse only through their habitat (Taylor and Bright, 1987).1 Because of the antiquity (late Miocene) of these snails and their tight linkage with aquatic habitats, their pat- terns of phylogenetic relationships and distributions likely re- flect late Cenozoic regional hydrographic history. Native distri- butions of these snails in the Great Basin appear to have been little affected by introduced species in contrast to fishes (partic- ularly salmonids; Behnke, 1992). The diverse hydrobiid fauna of the Great Basin includes many locally endemic forms, sug- gesting a complex evolutionary history that can provide a rich source of biogeographic information. For this contribution we focus on the hydrobiid Pyrgulopsis, the most speciose genus of aquatic snails in North America. Our purposes are to describe the regions of endemism defined by Holocene Pyrgulopsis species in the Great Basin and to re- view the distributions and the inferred phylogenetic relation- ships of these snails as they relate to current interpretations of drainage history. Pyrgulopsis is a North American group of 131 described spe- cies that are considered valid, 61% of which live in the Great Basin. Members of the genus are found throughout much of western North America from the Rio Grande basin west to the California coast and from the Snake River-Columbia River ba- sin south to the Bolson de Mapimi depression in Chihuahua and Coahuila states, northern Mexico (Figure I).2 The scope of Pyrgulopsis remains incompletely known because many spe- cies are undescribed, including sizable faunas in northern Mex- ico and the northwestern United States. The Great Basin con- tains the greatest diversity of Pyrgulopsis (80 species), followed by a far smaller fauna in the Colorado River drainage (20 species). Allopatry of congeners is typical, although two or three species occasionally are sympatric, usually in large springs. 255 256 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 1.?The general distribution of Pyrgulopsis (shaded area) throughout the western United States and northern Mexico. The southern limit of the genus in Mexico is uncertain. Pyrgulopsis is composed of small snails (1-8 mm in maxi- mum shell dimension) having a smooth or weakly sculptured shell that ranges from low trochiform to narrow-conic in shape (Figure 2). Throughout the West Pyrgulopsis is a locally abun- dant member of benthic communities, living on rock or aquatic vegetation and grazing on periphyton (Mladenka, 1992). Fe- males are oviparous and deposit egg capsules on hard sub- strates. Larval development is direct, with "crawl-away" young emerging from the egg capsules. Pyrgulopsis typically lives in springs ranging from tiny seeps to large rheocrenes and limno- crenes. Snails generally are concentrated near spring sources, with their density declining downflow (e.g., Noel, 1954)?a pattern that surely contributes to their tendency for local differ- entiation. A few species live in lakes, rivers, and larger streams. Although some species live only in basin floor habi- tats, many others live in middle- and high-elevation springs and streams along montane slopes, suggesting that faunal FIGURE 2 (opposite).?Scanning electron micrographs of shells of selected Pyrgulopsis species. Top row (left to right): P. breviloba (USNM 873174, shell height 1.8 mm), P. bryantwalkeri (USNM 874305, 2.3 mm), P. dixensis (USNM 874391, 2.0 mm); second row: P. erythropoma (USNM 857864, 2.2 mm), P. hendersoni (USNM 874386, 5.1 mm), P. merriami (USNM 873395, 2.5 mm); third row: P. nevadensis (USNM 590364, 4.6 mm), P. notidicola (USNM 873215, 1.8 mm), P. orbiculata (USNM 873196, 1.4 mm); bottom row: P. owensensis (USNM 857955, 2.3 mm), P. papillata (USNM 873185, 2.0 mm), P. turbatrix (USNM 883978, 2.9 mm). NUMBER 33 257 258 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 3.?Pyrgulopsis truckeensis, from the Truckee Formation (late Mio- cene), Churchill County, Nevada (holotype, USNM 560438). Shell height 2.2 movement has been achieved by upland headwater transfers (e.g., stream capture) as well as by basinal integration. Whereas extant species typically live in small springs, fossil shells referable to the genus routinely are found in late Ceno- zoic lacustrine deposits (Taylor, 1966a). Paleoecological stud- ies that delineate habitat use by these snails are lacking. The known ecology of extant congeners allows us to conjecture that Pyrgulopsis was not widely dispersed in paleolakes but instead was restricted to well-oxygenated littoral zones or lived in springs or wetlands closely associated with these lakes. The fossil record of Pyrgulopsis is limited in utility because of the weak phylogenetic signal provided by the simple shells of these snails (Taylor, 1987:5). Nevertheless, the occurrence of P. truckeensis (Figure 3) in the Truckee Formation (stratigraphy discussed by Firby, 1993) in the northwest Great Basin sug- gests that the genus dates at least to the late Miocene (Yen, 1950; see Gregg and Taylor, 1965, for assignment of this spe- cies to Fontelicella (=Pyrgulopsis)).3 It has been suggested that Holocene (presumably cold-water) species in the genus gener- ally may be Pliocene in origin and that western hydrobiids liv- ing in thermal springs (which include many Pyrgulopsis spe- cies) are generally ancient, perhaps middle Tertiary in age (Taylor, 1987:5, as Fontelicella). The antiquity of this group suggests that, as in the case of Great Basin fishes (e.g., Smith, 1978; Minckley et al., 1986), ancestral biotic ranges were af- fected not only by extensive integration and subsequent frag- mentation of Late Quaternary pluvial drainage, but also by the spectrum of tectonic and volcanic events that shaped the mod- ern regional landscape during the past 17-18 My. Methods A rigorous biogeographical analysis (e.g., Wiley, 1988) of Great Basin Pyrgulopsis is precluded by the paucity of phylo- genetic information concerning these snails, and although we discuss morphological groupings of species (as summarized in Appendix), biogeographic conclusions derived from distribu- tions of these snails must be refined in the future. Most of the Great Basin Pyrgulopsis fauna was described during the past decade (Hershler and Sada, 1987; Hershler, 1989, 1995, 1998; Hershler and Pratt, 1990), and the study of these snails lags well behind that of regional aquatic vertebrates. Taxonomic re- analysis of relatively widespread and/or morphologically vari- able forms will be necessary, as evidenced by a recent allozy- mic study of P. wongi from the southwestern Great Basin; this study revealed fixed-allele differences among populations from different drainages (Hamlin, 1996), suggesting the presence of cryptic species. The only published phylogenetic analysis of Pyrgulopsis species provided little resolution (Hershler, 1994, fig. 31) and did not establish monophyly of the genus within the subfamily Nymphophilinae.4 Morphological diversity among species assigned to Pyrgulopsis is pronounced, particularly with respect to glandular penial ornament (Figure 4). Species groupings in use are based on very few characters (mostly from penial form and ornament) that often are noncongruent, and many snail species are strongly differentiated by autapomor- phies. We thus rely heavily on the technique of inferring past drain- age relationships on the basis of biotic distributions, although we acknowledge the inherent limitations of this method (Plat- nick and Nelson, 1978). A biotic distribution encompassing several now-isolated drainages may suggest previous integra- tion or connection of these waters (and/or relatively vagile or- ganisms), but this assumes that biotic components automati- cally expand or contract their ranges in concert with potential habitat. The extent to which this assumption is realistic de- pends greatly on ecological characteristics of species in ques- tion. Biotic distributions do not reveal relative timing of histor- ical relationships among inhabited areas, whereas fossils and/ or a well-resolved cladogram might provide this information. Relatively broadly distributed taxa are useful within this con- text, but locally endemic forms provide no information in the absence of phylogenetic resolution. (In contrast, broadly dis- tributed taxa have little utility for vicariance biogeography.) Despite these drawbacks, earlier efforts using this methodology often have been insightful for Great Basin aquatic biogeogra- phy, as exemplified by seminal works on fishes (Hubbs and Miller, 1948; Hubbs et al., 1974). We follow King's (1982) definition of the Great Basin (which, in turn, was based on that of Fremont, 1845) as all con- tiguous areas of currently internal drainage between the Sierra Nevada and Rocky Mountains. Thus, areas that may have had recent external drainage but now are endorheic, such as the Sal- ton Trough and disrupted drainage of the Colorado River in southeast Nevada, are treated as Great Basin. Drainage maps NUMBER 33 259 FIGURE 4.?Representative variation in penial glandular ornament among Pyr- gulopsis species groups (see Appendix for names of species exhibiting each form): a, simple penis lacking glands (group 1); b, ornament of a superficial ventral gland (group 2); c, ornament of a terminal gland (group 3); d, ornament of a terminal gland and elongate Dgl abutting filament (group 4); e, ornament of a terminal gland and small, basal penial gland (group 5);/ ornament of a ter- minal gland and weak, dot-like Dgl (group 6); g, ornament of a raised Dgl (group 8); h, ornament of a full complement of glands (group 9). The schematic drawings are not to scale. were constructed by using United States Geological Survey (USGS) 1:2,000,000 line graph data as base. Drainage basins and their boundaries were digitized from USGS 1:500,000 (state) hydrologic unit maps. The fit of hydrographic bound- aries onto the drainage base is not exact because these two sets of data were originally of different scales and map projections. Final maps were prepared using Adobe Illustrator (version 7.0). Regions of endemism refer to drainages having two or more endemic species and should not be confused with "areas of endemism," which imply congruence of distributional and phylogenetic pattern among taxa (Harold and Mooi, 1994). Acknowledgments We thank many colleagues (listed in Hershler, 1998) for pro- vision of specimens and data and for assistance with fieldwork. Mapping and geographic analyses were done by Dan Cole (Na- tional Museum of Natural History (NMNH), Smithsonian Insti- tution). Molly Ryan (NMNH) prepared final maps. Kary Dar- row (NMNH) inked anatomical drawings and assisted with preparation of plates. Fiscal support was provided by two awards from the Smithsonian Institution Scholarly Studies Pro- gram. This paper benefited from detailed and very useful criti- cisms from Marith Reheis, Richard Vari, and two anonymous reviewers. We also acknowledge an intellectual debt to Dwight W. Taylor, whose pioneering studies of western American freshwater molluscan biogeography (notably Taylor, 1960, 1966a, 1985; Taylor and Smith, 1981; Taylor and Bright, 1987) paved the way for the present work. Biogeography Pyrgulopsis is widespread in the Great Basin (Figure 5) and even occurs in some poorly watered, fishless regions such as the "area of sterile basins" of southern Nevada (Hubbs and Miller, 1948:45). We know of no major sampling gaps, al- though distributions of many species may be imprecisely known because of nonexhaustive local fieldwork and probable undocumented extinctions in historic time. Distributions of Great Basin species are given in Table 1. Most of the fauna is endemic to single springs, spring com- plexes, or drainage basins. Only 16 of the 80 species range across major drainage divides and, as mentioned above, a re- evaluation of the purported conspecificity of populations is necessary in these cases. Pyrgulopsis species describe seven major regions of endemism within the Great Basin (Figure 6), five of which correspond to drainages harboring other endemic biota (e.g., fishes; Smith, 1978; Minckley et al., 1986). On the other hand, snail endemism in Dixie and Steptoe Basins is not matched by any other aquatic organism. Species of Pyrgulopsis also define subregions of endemism within the largest of these drainages (Bonneville Basin, Lahontan Basin, Death Valley system); most of these subregions have no parallel among other aquatic biota. DEATH VALLEY SYSTEM.?Treatment of this fauna partly reiterates that of Hershler and Pratt (1990). The three compo- nents (Owens Valley, Armagosa River basin, and Mohave Desert) of the Death Valley hydrographic system contain a total of 14 species of Pyrgulopsis, all but four of which are endemic. Endemic species are concentrated in two subregions, upper Owens Valley and the Amargosa River basin (Figure 7), both of which harbor other endemic biota (Sada et al., 1995). The upper Owens Valley subregion contains an endemic flock of three species?P. aardahli, P. owensensis, P. pertur- bata?that are most similar to the widespread Lahontan species P. gibba. The distribution of another species from this subre- gion, P. wongi (Figure 8a), suggests prior integration among upper Owens Valley?the southeastern Lahontan region (Car- son and East Walker River basins); Mono Valley, which com- prised the uppermost segment of the pluvial Owens River drainage (Putnam, 1949); and several valleys to the east (Deep Springs Valley, Huntoon Valley, Fish Lake Valley, Teels Salt Marsh). Reheis et al. (2002) present shoreline evidence that 260 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 0 40 80 120 160 200 I I I I I I km FIGURE 5.?Distribution of Pyrgulopsis in the Great Basin of the western United States. Dots often represent more than one locality. The heavy line shows the boundary of the Great Basin hydrographic system. NUMBER 33 261 FIGURE 6.?Regions of endemism represented by Pyrgulopsis in the Great Basin. 1, Death Valley system; 2, Lahontan Basin; 3, Dixie Basin; 4, Railroad Valley; 5, upper White River basin; 6, Steptoe Basin; 7, Bonneville Basin. 262 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE l.?Pyrgulopsis species recorded (X) from the Great Basin, by drainage: DV, Death Valley system; LA, Lahontan Basin; RB, Railroad Basin; ST, Steptoe Basin; DX, Dixie Basin; WR, upper White River basin; IB, other isolated basins within the Great Basin; BO, Bonneville Basin; and OL, "Oregon Lakes" region. Extralimital drain- ages report the distributions of these species outside the Great Basin: SA, Sacramento River basin; CA, California coastal drainage; SN, Snake River basin; and CO, Colorado River basin. Y = species extinct. Species Drainage Great Basin Extralimital DV LA RB ST DX WR IB BO OL SA CA SN CO aardahli aloba amargosae anatina anguina augustae aurata basiglans bifurcata breviloba bryantwalkeri carinata chamberlini cruciglans crystalis deaconi dixensis eremica erythropoma fairbanksensis fusca gibba giulianii gracilis hamlinensis hendersoni hovinghi hubbsi humboldtensis imperialis inopinata intermedia isolata kolobensis landyei lata lentiglans leporina limaria lockensis longae longiglans longinqua marcida merriami micrococcus militaris millenaria montana nanus neritella nevadensis nonaria notidicola orbiculata owensensis X1 - X - - - - - - - - - - - X - - - X X - - X - - - - - - - - - X1 - - - - - - - - - - - - X - - - X - - - - - X X1 X Y1 XXX X1 X X X1 X X X X X X NUMBER 33 263 Species papillata peculiaris pellita perturbata pictilis pilsbryana pisteri planulata plicata ruinosa sadai sathos saxatilis serrata sterilis sublata sulcata transversa turbatrix umbilicata variegata villacampae vinyardi wongi DV - - - X - - X - - - - - - - - - - - X - - - - X LA RB X - X1 - X1 - - - - - - X - - - - - - - - X - X X X TABLE 1.?Continued. Drainage Great Basin ST DX - - - - - - - X1 - - X - - X - - X - - - - - - - WR IB - - - - - - _ - - Y1 - Y" - - X X1 - - X - - - - X Extralimital BO OL SA CA SN CO _ _ Y - - - _ _ - - - X - - _ _ - - - X1 - - - _ _ - - - - X - _ _ X1 - - - - - _ _ _ _ _ _ _ _ Y Y _ _ _ X - - - X - _ _ - - - - - - 'Species is endemic to a single spring. M 0 40 80 120 160 1 I I I I km v, Mono Valley's pluvial Lake Russell may have had an interval of northern drainage to the East Walker River basin, and Rus- sell (1889:299-301) implied earlier integration of Mono Valley and the Lahontan Basin prior to disruption by Quaternary vol- canics. Perhaps early integration between Mono and Huntoon Valleys also occurred, because their divide is only slightly above the highest shoreline (~2257 m; Reheis et al., 2002) doc- umented for Lake Russell. Distribution of/! wongi in isolated valleys to the east of the Owens Valley trough correlates, in part, with an emerging concept that Lake Lahontan may have extended sufficiently south of Walker Lake to integrate with Fish Lake Valley (Mifflin and Wheat, 1979; Reheis and Morri- son, 1997; Reheis et al., 2002). However, integration of this conjectured drainage with Deep Springs Valley would have been unlikely, because pluvial Deep Spring Lake was isolated by very high divides during the late Pleistocene (Miller, 1928). A fossil Pyrgulopsis (Fontelicella sp.; Taylor, 1966a:67) was recorded from the Waucoba Lake beds (fide Walcott, 1897), which contain tephra layers from 2.22-2.17 Ma (A. Sarna- Wojcicki, in litt., August, 1998) that are now uplifted above FIGURE 7 (left).?Regions of endemism by Pyrgulopsis within the Death Val- ley system. A, upper Owens Valley (north of Tinemaha Reservoir); B, Amar- gosa River basin (including its terminus, the Death Valley trough). The heavy line shows the boundary of the Death Valley pluvial drainage system. SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES V \J ? wongi ? giulianii * turbatrix 0 40 80 120 160 I I I km \ FIGURE 8.?Distributions of selected Pyrgulopsis species of the Death Valley System: a, P. giulianii, P. turbatrix, P. wongi; b, P. micrococcus. Shading indicates the regions of endemism (from Figure 7). eastern Owens Valley. This record suggests that P. wongi may have dwelled early in the proto-Sierra Nevada/White Moun- tain block and attained current distribution as a result of later tectonic events that created Owens, Deep Springs, and Fish Lake Valleys during the past 2.3 My (Bachman, 1978; Reheis and Sawyer, 1997; also see Lueddecke et al., 1998). This hy- pothesis is supported by allozyme evidence that P. wongi popu- lations of Deep Springs Valley are more similar to snails of Owens Valley than to samples from upper segments of the plu- vial Owens River drainage (Hamlin, 1996). Pyrgulopsis giulianii (Figure 8a), which is morphologically very similar to the upper Owens Valley species flock described above, lives in Sierra Nevada streams that drain to Indian Wells Valley, into which pluvial Lake Owens spilled. Records of this species along the south fork of the Kern River west of the Si- erra Nevada suggest prior transmontane drainage across the Si- erra Nevada divide (also indicated by Dibblee, 1967, figs. 79, 80), which presumably was disrupted by uplift and westward tilting of the southern portion of this range since the late Neo- gene (Loomis and Burbank, 1988; Unruh, 1991; Christiansen and Yeats, 1992).5 Of the seven species endemic to the Amargosa River basin, six are restricted to the Ash Meadows spring oasis, several of which (P. crystalis, P. erythropoma, P. pisteri) form a putative clade closely similar to P. avernalis from Moapa Valley in southeast Nevada (Colorado River basin).6 The seventh species endemic to the Amargosa River basin, P. amargosae, lives in springs along the lower Amargosa River and most closely re- sembles a species from the Snake River basin (P. bruneauen- sis). This putative relationship is suggestive of the fish-hook distributional pattern described for various freshwater mollusks and fishes (Taylor 1960, 1966a), although this possibility would be bolstered by presence of linking populations or spe- cies in the large intervening area.7 The distribution of another snail, P. turbatrix (Figure 8a), also implies that the Amargosa River basin had a previous drainage relationship with the Colorado River basin as well as with several isolated valleys to the east. Pyrgulopsis turbatrix lives along slopes of the Spring Mountains in Amargosa Flat (which drains to Ash Meadows), isolated Indian Springs and Pahrump Valleys, and Las Vegas Valley (Colorado River ba- sin). (Note that another congener, P. deaconi, also lives along the western and eastern flanks of the Spring Mountains in Pah- rump and Las Vegas Valleys, respectively.) Pyrgulopsis turba- trix also lives in a single spring in Frenchman Flat, a small, iso- lated basin west of Indian Springs Valley. Many have speculated that the Death Valley system had a past connection with the Colorado River basin (see Miller, 1946); although an NUMBER 33 265 oft-postulated connection via spilling of the integrated pluvial lake system to the south (through Bristol, Cadiz, and Danby Lake basins) is unlikely (Brown and Rosen, 1995; Hooke, 1998), other forms of integration may have occurred. Distribu- tion of P. turbatrix conforms in part to previous integration of the Amargosa River basin and Las Vegas Valley around the north flank of the Spring Mountains, as proposed by Hubbs and Miller (1948:101-102; also see Nolan in Gilluly, 1929:682). This integrated system presumably was disrupted as a result of stream captures across the low, largely alluvial divides in the region. Hubbs and Miller (1948:85) argued that "Lake Pah- rump" spilled to the Amargosa River basin, but Mifflin and Wheat (1979) denied existence of a late Pleistocene pluvial lake in this valley. Integration of the drainage postulated above with that of Frenchman Flat may have occurred via the low di- vides separating the Amargosa River basin from Rock and Mercury Valleys to the west (which in turn are weakly sepa- rated from Indian Springs Valley). Alternatively, drainage inte- gration may have been disrupted by the several kilometers of movement along local, northeast-trending strike-slip faults, at least some of which occurred after the Miocene (Carr, 1984; Stewart, 1988). Distribution of P. turbatrix along flanks of the Spring Mountains also could have arisen via transmontane stream capture, perhaps associated with late Cenozoic tilting of this range (Burchfiel et al., 1983). Pyrgulopsis micrococcus (Figure 86) lives in all three com- ponents of the Death Valley hydrographic system, although it is found in only a lower segment of the pluvial Owens River drainage, Panamint Valley. The distribution of this snail pro- vides a possible exception to the assertion by Taylor (1985: 317-318) that "spillover from Lake Panamint into Death Val- ley, and into Lake Manly in Death Valley, had no recognizable effect on mollusc distribution." However, this species lives in high-elevation springs in the Panamint Mountains, and its dis- tribution across this divide may reflect headwater transfer as opposed to pluvial dispersal. Distribution of this snail also im- plies early (before the late Pleistocene) drainage integration of the Death Valley system with both Saline Valley (northwest of Panamint Valley) and the California coastal drainage south of the Transverse Ranges. Saline Valley, a small, profoundly closed basin that lacked a pluvial lake, perhaps integrated with Panamint Valley before the faulting that was responsible for basinal development during the past 3.0 My (Burchfiel et al., 1987). The distribution of P. micrococcus is paralleled by an- other prosobranch snail, Assiminea infima, which lives in Sa- line (Hershler, unpublished) and Panamint Valleys (as Ho- locene fossil; Taylor, 1966b:208), as well as in the Amargosa River basin (Hershler, 1987). The transmontane distribution of P. micrococcus in the San Bernardino Mountains implies early presence in the region?before the late Miocene to early Pliocene uplift, which truncated and reversed the southward Mojave drainage (Meisling and Weldon, 1989).8 Pyrgulopsis biogeography (1) implies various prior drainage connections between the Death Valley system and other, mostly adjacent, areas and (2) suggests that the upper Owens Valley and Amargosa River basin subregions of endemism do not share a close historical relationship. Interestingly, early drainage integration between the northwest portion of the Death Valley system and the Sacramento Valley before the late Cenozoic Sierra Nevada uplift (e.g., Huber, 1981) is not well reflected in Pyrgulopsis biogeography. The complete absence of species overlap and the entirely different sets of historical drainage relationships implied by the Pyrgulopsis faunas of Owens Valley and the Amargosa River basin are striking, given the pluvial integration of these drainages, yet they agree with the distribution patterns of other mollusks (Taylor, 1985) and of fishes (Minckley et al., 1986; also see Echelle and Dowling, 1992). Taylor (1985:317) suggested that the disparity of fresh- water mollusks from west and east of Death Valley may reflect a dispersal barrier provided by extension of the Miocene proto- Gulf of California into the area. Although fossils may imply presence of a late Cenozoic estuary in portions of the Mojave Desert (Durham and Allison, 1960; Smith, 1970; disputed by Spencer and Patchett, 1997), no physical record of such condi- tions has been found in the Death Valley region (Buising, 1990, fig. 1), and the local fossil record is equivocal in this regard.9 Physical evidence of middle Pleistocene integration of the Owens and Amargosa River basins is clear (Smith, 1976; Jan- nik et al., 1991; Fitzpatrick and Bischoff, 1993), yet hydro- graphic conditions nonetheless may have provided a barrier to aquatic dispersal (Miller, 1981) that perhaps was positioned be- tween Indian Wells and Panamint Valleys. LAHONTAN BASIN.?This large region harbors 22 species of Pyrgulopsis. Four endemic species of fishes also live in this ba- sin (Hubbs et al., 1974; Deacon and Williams, 1984). The 16 en- demic species of Pyrgulopsis are concentrated in four subregions (Figure 9), only one of which, the Black Rock Desert, harbors other endemic aquatic biota. Pyrgulopsis biogeography implies considerable prior drainage integration within the Lahontan Ba- sin and between this and several adjacent drainages. A southwest extension of Reese River Valley (Antelope Val- ley) contains two endemic species (P. augustae, P. pictilis) that are morphologically very similar to other Lahontan snails. Ba- sin floor springs in the Carico Lake basin (part of the middle segment of the Humboldt River basin) harbor two endemic spe- cies (P. basiglans, P. bifurcata) that form a putative clade simi- lar to the group of snails from the Amargosa and White River basins discussed above. Two snails (P. bryantwalkeri, P. hum- boldtensis) belonging to large groups of similar species are en- demic to the upper Humboldt River basin. The distribution of another snail, P. leporina (Figure 10a), implies that drainage of the upper Humboldt River basin (Lamoille Valley) was previ- ously integrated (before the late Pleistocene) with that of north- ern Ruby Valley, perhaps reflecting spill of pluvial Lake Frank- lin into Lake Lahontan via Lake Clover as suggested by Snyder et al. (1964; but disputed by Mifflin and Wheat, 1979:57). Sol- dier Meadows, a prominent zone of thermal springs that drains south to the floor of Black Rock Desert via Mud Meadows 266 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 9 (left).?Regions of endemism by Pyrgulopsis within the Lahontan system. A, Black Rock Desert; B, Reese River basin; C, middle Humbolt <(_ River basin (includes the section of river from Palisade to Golconda Summit; f& i also Carico Lake, Crescent, and Pumpernickel Valleys); D, upper Humboldt River basin (includes Marys River drainage). The heavy line shows the boundary of the Lahontan pluvial drainage system. nj -| \ FIGURE 10 (below).?Distributions of selected Pyrgulopsis species of the Lahontan system: a, P. eremica, P. leporina, P. longiglans, P. militaris, P. sadai; b, P. gibba. Shading indicates the regions of endemism (from Figure 9). 0 40 80 120 160 * I I I I I km ? sadai longivlans " eremica CALIFORNIA; 7 0 40 80 120 160 I I I I i km ( 0 40 80 120 160 I I I I I km J NUMBER 33 267 Creek, contains a presumed flock of three endemic species (P. limaria, P. notidicola, P. umbilicata), which closely resemble an eastern Lahontan (Squaw Valley) species. Soldier Meadows also contains an endemic species of desert dace (Nyquist, 1963). Distribution of a fourth snail (P. militaris) belonging to this local group implies earlier drainage integration between Soldier Meadows and Alvord Basin to the north (Figure 10a), which in turn had (at least temporarily) a pluvial connection with the Snake River basin during the late Pleistocene (Lind- berg and Hemphill-Haley, 1988). Snail transfer also could have been mediated through intervening Summit Lake basin, which Mifflin and Wheat (1979:27, 30) suggested had a complex Qua- ternary history that included intervals of drainage to Soldier Meadows and Alvord Basin. Note that Soldier Meadows springs are associated with a lineament extending through the Alvord Basin, which Garside and Schilling (1979:38-39, fig. 28) interpreted as a fault trace. In addition to the endemic faunas described above, the Lahontan Basin harbors scattered locally endemic species. Of particular note is P. nevadensis, which lived in several western Lahontan lakes but is now probably extinct (Hershler, 1994). This snail is closely similar to P. archimedis from Klamath Lake basin, and the implied historical drainage relationship between these areas is part of the fish-hook track described by Taylor (1960, fig. 1; 1985, fig. 17) for these and other species having similar shell sculpture. We are not confident, however, that other components of this track, consisting of fossil taxa from the Snake River region and southern California, are closely re- lated to these living species because development of carinate shell sculpture has been subject to homoplasy in this group (Hershler, 1994, 1998). Several widespread snails also reflect considerable past con- nectivity of drainage in the Lahontan region. Pyrgulopsis wongi, discussed above, lives in both the Carson River and East Walker River basins (Figure 8a). Pyrgulopsis longiglans (Fig- ure 10a) is distributed in a north-south manner in the western portion of the Lahontan Basin (overlapping the Black Rock Desert) whereas P. sadai (Figure 10a) is distributed in a similar fashion in the central Lahontan area (overlapping the Reese River and the middle Humboldt subregions). Distribution of this species also implies prior drainage integration between the Lahontan Basin and the Snake River basin in northeast Nevada, although the single locality in the Snake River drainage (at the base of the Santa Rosa Range) is suggestive of a headwater transfer. The distribution of P. eremica (Figure 10a) in the west- ern Lahontan Basin conforms to interpretations that Eagle Lake and Horse Lake basins (north of Honey Lake Valley), both of which contained pluvial lakes, overflowed to Lake Lahontan (Gester, 1962; Snyder et al., 1964; Morrison, 1991). In the Ea- gle Lake basin, this snail lives along Willow Creek, which may have drained pluvial Lake Acaps'kati to the south before the damming caused by Quaternary basalt flows (Gester, 1962). Pyrgulopsis gibba lives throughout much of the western Lahon- tan Basin, including all of the subregions of endemism (Figure 106). Distribution of this snail implies past connections be- J? MY \ FIGURE 11.?Regions of endemism by Pyrgulopsis within the Bonneville Basin. A, Thousand Springs basin (includes Pilot Creek, Tecoma, and Thou- sand Springs Valleys); B, Snake Valley (includes Hamlin and Snake Valleys); C, upper Sevier River basin (extending to Sevier Bridge Reservoir above Gun- nison; exclusive of the West Fork Sevier River below Piute Reservoir). The heavy line shows the boundary of the Bonneville pluvial drainage system. tween the Lahontan Basin, several closely proximate Nevadan valleys (Big Smokey, Dixie, Granite Springs, and Grass Val- leys), and the "Oregon Lakes" region, although the northern component of this distribution also may be attributed to drain- age disruption caused by tens of kilometers of right-lateral slip (Miocene to Holocene) along the Eugene-Denio fault zone (Lawrence, 1976). BONNEVILLE BASIN.?Taylor and Bright (1987) previously discussed the aquatic molluscan biogeography of this region (although not focusing on Pyrgulopsis), which constitutes the largest drainage unit in the Great Basin. The Bonneville Basin harbors 17 species of Pyrgulopsis, 14 of which are endemic. This basin also contains eight endemic species of fishes (Smith, 1978). The Pyrgulopsis fauna is very distinct from that of the Lahontan Basin (paralleling that of fishes), and only a single species ranges across both regions. Endemic species are con- centrated in three subregions (Figure 11), of which only the Snake Valley (along the Nevada-Utah border, east of the Snake 268 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Y ir1 IN ? peculiaris A transversa * variegata [, ^ , , ? FIGURE 12.?Distributions of selected Pyrgulopsis species of the Bonneville Basin: a, P. peculiaris, P. trans- versa, P. variegata; b, P. kolobensis, P. pilsbryana. Shading indicates regions of endemism (from Figure 11). Range; not to be confused with the Snake River basin to the north) contains other endemic biota. The Thousand Springs subregion harbors five species, in- cluding three endemics (P. hovinghi, P. lentiglans, P. millena- ria) distributed along segments of the Thousand Springs Creek drainage. Although relationships of these endemic snails are uncertain, the distribution of another snail, P. variegata (Figure 12o), suggests a prior drainage relationship existed with an ad- jacent portion of the northwest Bonneville Basin and Goose Creek, which is in the Snake River basin to the north. Physical evidence also suggests that Goose Creek was part of the Bonneville drainage basin before the late Miocene drainage re- versal that was associated with passage of the Yellowstone Hot Spot (Hildebrand and Newman, 1985; Rodgers et al., 1991). Three species are locally endemic in the Snake Valley subre- gion (see Figure 11). One of these, P. hamlinensis, is closely similar to P. montana, which lives in the uppermost portion of Meadow Valley Wash (a disrupted segment of Colorado River drainage) just over the White Rock Mountain divide to the west. Distributions of these species may reflect past historic upland stream capture between these drainages. Pyrgulopsis saxatilis is most similar to a species from the White River val- ley to the west (P. lata), whereas the third endemic, P. anguina, is most similar to P. chamberlini from the Sevier River basin to the east. Prior drainage integration with areas to the west and east is also implied by the distribution of P. peculiaris (Figure 12a), which lives in the Snake Valley as well as in the lower segment of the Sevier River basin and the isolated Spring Val- ley in eastern Nevada. Five species are endemic in the upper segment of the Sevier River basin in southern Utah, confirming the distinctiveness of this fauna as noted by Taylor and Bright (1987). Affinities of these species imply prior historical rela- tionships with other segments of the Bonneville Basin, as does the distribution of P. transversa (Figure \2a). The distribution of P. kolobensis suggests extensive prior connectiveness of Bonneville Basin drainage (this snail over- laps all three subregions of endemism) as well as integration with the eastern Lahontan Basin, isolated valleys situated be- tween these two regions, and the Colorado River basin (Figure 126). (This snail exhibits considerable morphological differen- tiation across its broad range and almost certainly represents a species complex.) Past integration with the Colorado River ba- NUMBER 33 269 0 40 80 120 160 I I I I km 0 40 80 120 160 1 I I I I km FIGURE 13.?Distributions of selected Pyrgulopsis species in Steptoe Basin (a, P. cruciglans) and in the upper White River basin (b, P. marcida, P. breviloba). The heavy lines show the boundaries of these drainage basins. sin (in southern Utah) probably represents headwater transfers, for which there is also ichthyological evidence (Hubbs and Miller, 1948:29-30) as well as physical evidence (Averitt, 1962; Anderson and Mehnert, 1979). Undescribed material (USNM collection) from the Snake River basin of southeast Idaho is closely similar to or possibly conspecific with this snail, conforming with the physical record of early southerly drainage of southeast Idaho basins (Ore, 1982; Bobo, 1991; Osier, 1995).10 The distribution of P. pilsbryana reflects inte- gration of the Bear Lake and Bear River basins (Figure 126) and is noteworthy in that this snail does not range below the Oneida Narrows, which comprised the divide between pluvial Lake Thatcher and the Bonneville Basin (Bright, 1963). STEPTOE BASIN.?This large Nevadan basin (which includes Antelope, Goshute, Spring, and Steptoe Valleys) contains eight species, six of which are endemic. Endemic species are con- centrated in the northern and southern segments of this drain- age, and only the widespread P. kolobensis lives in both areas. Five of the endemic species are localized in a large basin floor spring province along the east flank of the Egan Range north of Ely. These snails are united by several unique features and probably represent a species flock, the affinities of which are not clear. The distribution of P. cruciglans in the northern seg- ment of Steptoe Basin (Figure 13a) and in the adjacent western Bonneville Basin is enigmatic because shorelines indicate that although pluvial Lake Waring was relatively deep, its high stand was well below basinal sills, and it did not integrate with Lake Bonneville during the late Pleistocene (Hubbs et al., 1974; Currey et al., 1984). Note that the sole native fish living in the Steptoe Basin, Relictis solitarius, does not cross the Bonneville divide but instead ranges to the west and south (Butte and Ruby Valleys; Hubbs et al., 1974; Deacon and Will- iams, 1984). UPPER WHITE RIVER BASIN.?This large area includes Pah- ranagat and White River Valleys (above Pahranagat Wash) but excludes Moapa Valley (the lowermost segment of the White River basin), which maintains external drainage to the Colo- rado River basin. DiGuiseppi and Bartley (1991) suggested that the White River drainage was internally drained before capture by the Colorado River during the Quaternary. This region con- tains eight species (five endemic); this fauna neither overlaps nor shows affinities with the two species living in Moapa Val- ley, which instead have affinities with snails in the Amargosa River basin (P. avernalis) and Colorado River basin of northern Arizona (P. carinifera). The upper White River basin also con- tains an endemic species of cyprinid fish (Miller and Hubbs, 1960). Of the endemic snails for which affinities are clear, P. merriami is most similar to fauna of the Rio Grande and upper 270 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Gila River basins, whereas P. lata is most similar to P. saxatilis from the Bonneville Basin. Distributions of two nonendemic species (Figure 136) sug- gest prior relationships between the White River region and two fishless valleys to the east (Cave and Dry Lake Valleys) despite physical evidence that the pluvial lakes of these small valleys did not overflow during the late Pleistocene (Mifflin and Wheat, 1979). The distribution of/! marcida supports the contention of Hubbs et al. (1974:65) that, although Cave Valley appears to be part of a structural trough including Steptoe Val- ley to the north, Cave Valley instead probably drained early (i.e., prior to the late Quaternary pluvial period) to the White River valley. We were unable to confirm the assertion of Hubbs and Miller (1948:98) that there may be a low, alluvial saddle separating the southern portion of Cave Valley from the White River valley; maps indicate that these valleys instead are sepa- rated in this area by middle Tertiary andesite flows. The distri- bution of P. breviloba (Figure 136) conforms to an interpreta- tion (Hubbs and Miller, 1948:98) that Dry Lake valley (therein referred to as Bristol Valley) may have drained early, south to the White River valley via Delamar Valley. Drainage integra- tion between these valleys may have been disrupted by signifi- cant late Cenozoic lateral slip along the northeast-trending Pah- ranagat Fault Zone (Rowan and Wetlaufer, 1981; Jayko, 1990). RAILROAD BASIN.?Railroad Valley in south-central Ne- vada is closed to the east, north, and west principally by the Grant, White Pine, and Pancake Ranges, respectively, but it opens to the southwest into a large basin comprising Reveille, Hot Creek, and Little Fish Lake valleys (Figure 6). Mifflin and Wheat (1979:30) suggested that Big Sand Springs Valley also may have opened southeast to Railroad Valley before the Late Quaternary emplacement of the Lunar Crater Volcanic Field (see Scott and Trask, 1971; Luedke and Smith, 1991). The northern segment of Railroad Valley harbors six endemic spe- cies of Pyrgulopsis, and widespread P. kolobensis also occurs locally, both in Railroad and Reveille Valleys. Four of the local endemics?P. aloba, P. carinata, P. lockensis, P. papillata?be- long to a group represented in western California and the Lahontan Basin, northern Bonneville Basin, Snake River basin, and Colorado River basin. Three of these snails compose a mor- phologically distinct subgroup confined to thermal limnocrenes. DIXIE BASIN.?This large, Nevada basin includes Buffalo, Buena Vista, Dixie, Edwards Creek, Fairview, and Pleasant Valleys. It contains two dissimilar endemic species?P. aurata, P. dixensis?that closely resemble other Lahontan forms. The only endemic fish in this drainage is an undescribed subspecies of Gila bicolor, which Hubbs and Miller (1948) suggested was closely related to Lahontan taxa. In support of these connec- tions, Reheis et al. (2002) infer that a high stand of Lake Lahontan flooded into Dixie Basin during the middle Pleis- tocene. OREGON LAKES.?Pyrgulopsis is poorly represented in this large region, which consists of a series of valleys and pluvial lakes in northeast California and southeast Oregon. Only four intermedia hendersoni 7 I j;0 40 80 120 160 ^ I I I I I km FIGURE 14.?Distributions of Pyrgulopsis intermedia and P. hendersoni of the "Oregon Lakes" region. The heavy line shows the northwest boundary of the Great Basin. species are present, none of which are endemic, and relatively large areas (e.g., Goose Lake and Fort Rock basins) apparently lack Pyrgulopsis. Biogeographic patterns of these snails, never- theless, are informative, suggesting past integration of compo- nent valleys as well as relationships with both the Lahontan Basin (as discussed above) and the Snake River basin. The dis- tribution of P. gibba (Figure 106) suggests prior integration of Long and Surprise Basins, whose pluvial lakes apparently did not overflow during the late Pleistocene (Mifflin and Wheat, 1979; Allison, 1982, respectively). The snail also lives in Cole- man Valley, a small endorheic basin that Hubbs and Miller (1948:65) suggested was flooded by an arm of pluvial Lake Warner. However, Weide's (1974, fig. 19) map of the lake indi- cates that this did not occur in the late Pleistocene. The distri- bution of P. hendersoni (Figure 14) suggests prior integration of the widely separated Lake Abert and Malheur Lake basins, which is also suggested by the distribution of the fish Gila bi- color (Minckley et al., 1986). This snail also lives in springs along the Malheur River, conforming to the oft-postulated prior drainage relationship between the Oregon Lakes region and the Snake River basin (e.g., Wheeler and Cook, 1954; Taylor and Smith, 1981; disputed by Repenning et al., 1995). Shoreline NUMBER 33 271 evidence suggests that pluvial Lake Malheur probably spilled east to the Malheur River (Gehr and Newman, 1978), and a similar eastward drainage through Malheur or Crane Gaps may have occurred earlier, before late-Pleistocene emplacement of lava flows (Piper et al., 1939; Bisson and Bond, 1971; Bald- win, 1976). The distribution of P. intermedia also suggests a prior relationship between regional drainage and the Snake River basin. This snail lives in Crooked Creek, a tributary to the Owyhee River (Snake River basin), as well as in Barren Valley, a small basin often depicted on maps as draining to the Snake River basin (e.g., Snyder et al., 1964; Williams and Be- dinger, 1984) but which, in fact, is endorheic and seemingly a northern extension of the Alvord Basin trough. Snail popula- tions from the Pit River and Klamath Lake basins are similar to P. intermedia?although probably not conspecific, as sug- gested by Taylor (1985:309). Conclusions Pyrgulopsis biogeography within the Great Basin conforms in part to current interpretations of pluvial drainage configura- tions, but in many cases it defines additional or more compli- cated patterns that presumably reflect older relationships. Vari- ous drainage relationships between the Great Basin and adjacent regions also are implied by these data, although there is surprisingly little evidence of past integration between the Great Basin and the Sacramento River basin to the west and the upper Colorado River basin to the northeast. This may partly reflect inadequate sampling of these extra-limital areas. In con- trast to previous suggestions that the Great Basin had prior con- nections along its northeast and northwest boundaries with the Snake River basin (Taylor, 1966a, 1985; Minckley et al., 1986; Taylor and Bright, 1987), Pyrgulopsis biogeography suggests a more continuous pattern of northern integration. As mentioned above, a more rigorous biogeographic analysis of Pyrgulopsis must await preparation of a well-corroborated phylogeny for its species, which will prove difficult using mor- phological criteria alone and thus will probably require the use of DNA sequence data. Determination of cladistic structure among Holocene species of this group can provide a hierarchi- cal depiction of historical relationships among drainages, but the fossil record also must be better utilized to constrain the chronology more tightly. Given that hydrobiids are difficult to identify at the species and, often, generic ranks based on tradi- tional shell characters, new methods will be needed to extract a phylogenetic signal from fossils, perhaps from details of shell structure or composition of preserved organic shell compo- nents. Ecological characteristics of fossil Pyrgulopsis also must be better understood if we are to determine, for instance, whether interbasinal patterns reflect vicariance of pluvial lake biota or transmontane headwater transfers. This information may perhaps be obtained from a combination of environmen- tally sensitive isotopes preserved in shell fabric and more de- tailed study of sedimentary facies. Appendix Species Groups of Pyrgulopsis The informal groups described below are defined mostly by characters of penial form and ornament. We follow the termi- nology of Hershler, 1994. 1. A small number of species, mostly from the Great Basin, share a simple penis bereft of distal lobe and glands (Figure 4a). These snails otherwise closely resemble typical Pyrgu- lopsis, and a close relationship is inferred, although separate generic status for some or all of these may be suggested upon cladistic analysis. Subgroups are discernable on the basis of shape of shell and penial filament. Species included are aloba, augustae, carinata, coloradensis (Virgin River basin), dixensis, eremica, greggi (Kern River basin), locken- sis, millenaria, and papillata. 2. Snails of this group have a glandular ornament of a sin- gle, superficial ventral unit (Figure 4b). This gland differs in appearance from the narrow, raised ventral gland typical of Pyrgulopsis and resembles that of eastern American Bir- gella (Thompson, 1984, figs. 81, 82) and Rhapinema (Thompson, 1970, fig. 3D,E). With the exception of P. ari- zonae, the snails of this group also lack a penial lobe, per- haps indicating a close relationship with the group described above. Subgroups are recognizable on the basis of shell shape and gland position and size. Species included are ari- zonae (Gila River basin), avernalis (Colorado River basin), basiglans, bernardina (Rio Yaqui basin), bifurcata, crysta- lis, erythropoma, and pisteri. 3. A large number of species share a relatively narrow, dis- tally bifurcate penis ornamented with a terminal gland along the distal edge of the rather small lobe (Figure 4c). Other dorsal glands are absent in these snails, although some spe- cies have a single ventral gland. The terminal gland is often oriented longitudinally rather than transversely, as in groups described below. In some species the terminal gland is weakly developed (e.g., stearnsiana) or often absent {bry- antwalkeri). This group conforms in large part to the "Fon- telicella stearnsiana series" described by Taylor (1987:10). Subgroups are recognizable on the basis of shell shape, shape of penial lobe and filament, and development of the terminal gland. Species included are amargosae, aurata, bacchus (Colorado River basin), brandi (Rio Casas Grandes basin), breviloba, bruneauensis (Snake River basin), bryant- walkeri, carinifera (Colorado River basin), conica (Colo- 272 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES rado River basin), fairbanksensis, hamlinensis, hubbsi, imperialis, isolata, lentiglans, manantiali (Cuatro Cienegas basin), micrococcus, montana, morrisoni (Verde River basin), nanus, nonaria, sathos, simplex (Verde River basin), sola (Verde River basin), stearnsiana (California coastal drainage), thompsoni (Gila River basin), and transversa. 4. A medium-sized group comprises species in which the outer edge of the dorsal penis has an elongate gland (Dgl) abutting or slightly overlapping the filament (Figure Ad). Penial ornament in this group is otherwise variable, although most species have a terminal gland, ventral gland, and sometimes additional dorsal glands. The penial lobe of these snails often is enlarged relative to the filament. This group includes the large snails allocated to the subgenus Natricola Gregg and Taylor, 1965, except for intermedia. Species included are glandulosa (Verde River basin), hend- ersoni, hovinghi, idahoensis (Snake River basin), inopinata, montezumensis (Verde River basin), pecosensis (Rio Grande basin), pellita, robusta (Snake River basin), and sublata. 5. Another large group consists of snails in which the penis has a short penial gland at the base of the filament and gen- erally no other dorsal glands other than a terminal gland and occasionally a ventral gland (Figure 4e). Subgroups are rec- ognizable on the basis of shell sculpture, shape of penial lobe and filament, development of terminal gland, presence of additional dorsal glands, and position of ventral gland. Species included are anatina, archimedis (Klamath Lake basin), diablensis (Sacramento River basin), fusca, hum- boldtensis, kolobensis, lata, leporina, limaria, longae, mar- cida, militaris, nevadensis, notidicola, pictilis, pilsbryana, saxatilis, serrata, sterilis, taylori (California coastal drain- age), trivialis (Gila River basin), turbatrix, umbilicata, var- iegata, and vinyardi. 6. Six species have a penis with a well-developed terminal gland, a ventral gland, and a weakly developed additional dorsal gland corresponding to Dgl (Figure Af). Species included are aardahli, gibba, giulianii, owensensis, pertur- bata, and ruinosa. 1. The five species locally endemic in southwestern Steptoe Valley are united by strong protoconch microsculpture, a broadly conical or subglobose shell, and unique features of the central radular teeth (fused lateral cusps, basal cusps absent or very weakly developed, elongate basal tongue). Although penial form and ornament varies considerably among these species, we nevertheless interpret the group as a strongly differentiated clade based on the synapomorphic radular features. Species included are landyei, neritella, orbiculata, planulata, and sulcata. 8. Two Great Basin species share a distinctive pattern of penial ornament consisting of a large, raised, proximal dorsal unit (Figure 4g). These species are cruciglans and pi'icata. 9. A large number of species share a massive penis having a long gland on the dorsal surface of the penial filament (Pg); typically they also share a terminal gland (typically trans- verse, and often U-shaped), one or two ventral glands, and one or more additional dorsal glands (Figure Ah). Members of this group usually have an enlarged, pyriform bursa copu- latrix and a narrow bursal duct. The seminal receptacle is often very small relative to the bursa copulatrix and usually is well posterior in position, broadly overlapping the bursa copulatrix. This group corresponds, at least in part, to the " californiensis series," as defined by Taylor (1987). Sub- groups are recognizable from the presence of secondary dor- sal glands and from the shape and multiplication of penial glands. Species included are anguina, californiensis (Cali- fornia coastal drainage), chamberlini, chihuahua (Rio Car- men basin), chupaderae (Rio Grande basin), davisi (Rio Grande basin), deaconi, deserta (Virgin River basin), fausta (Colorado River basin), gracilis, intermedia, longiglans, longinqua, metcalfi (Rio Grande basin), neomexicana (Rio Grande basin), peculiaris, roswellensis (Rio Grande basin), sadai, thermalis (Gila River basin), ventricosa (Sacramento River basin), villacampae, and wongi. Notes 1. Ponder et al. (1994) suggested that passive dispersal on birds may be a sig- nificant factor in biogeography of some Australian hydrobiids. 2. Twelve eastern American species currently are allocated to the genus (Hershlerand Thompson, 1987; Hershler, 1994; Thompson, 1995); recent mor- phological studies indicate that this fauna is a distinct clade that is not closely related to Pyrgulopsis (Thompson and Hershler, In press). 3. Tall-spired western Paleogene hydrobiids, usually allocated to Hydrobia or Micropyrgus (e.g., from Flagstaff Formation; La Rocque, 1960), approach the range of shell variation of modern Pyrgulopsis, but they probably represent other clades (Taylor, 1975). 4. Other genera in the subfamily are Birgella, Cincinnatia, Notogillia, Nym- phophilus, Rhapinema, Spilochlamys, and Stiobia (Thompson, 1979). 5. These and other early concepts of extent and timing of late Cenozoic uplift of the Sierra Nevada (e.g., Axelrod, 1962; Huber, 1981) have recently been dis- puted (Small and Anderson, 1995; Wernicke et al., 1996; Wolfe et al., 1997). 6. An earlier conjecture (Hershler and Pratt, 1990:292) that all of the endemic Pyrgulopsis in Ash Meadows were derived from taxa in the White River Valley (whose lowermost segment is Moapa Valley) to the east was premature. 7. This refers to a distributional track encompassing the Bonneville Basin, the middle Snake River drainage, the Lahontan Basin, and the Death Valley system. 8. Nagy and Murray (1996) suggested that this drainage reversal may not have occurred before 0.5 Ma. 9. Late Cenozoic fossil Foraminifera from Panamint Valley may have lived in a saline lake instead of in an estuary (Smith, 1960). 10. Whether or not the late Pleistocene Bonneville Flood (Malde, 1968; O'Connor, 1993) also contributed to snail exchange with the Snake River basin is conjectural. NUMBER 33 273 Literature Cited Allison, I.S. 1982. Geology of Pluvial Lake Chewaucan, Lake County, Oregon. Ore- gon State Monographs, Studies in Geology, 11: 79 pages Anderson, R.E., and H.H. Mehnert 1979. Reinterpretation of the History of the Hurricane Fault in Utah. In G.W. Newman and H.D. Goode, editors, Basin and Range Sympo- sium and Great Basin Field Conference, pages 145-165. Denver, Colorado: Rocky Mountan Association of Geologists; and Salt Lake City, Utah: Utah Geological Association. Averitt, P. 1962. Geology and Coal Resources of the Cedar Mountain Quadrangle, Iron County Utah. United States Geological Survey Professional Papers, 389: 72 pages, plates 1-3. Axelrod, D.I. 1962. Post-Pliocene Uplift of the Sierra Nevada, California. Bulletin of the Geological Society of America, 73:183-198. Bachman, S.B. 1978. Pliocene-Pleistocene Break-Up of the Sierra Nevada?White-Inyo Mountains Block and Formation of Owens Valley. Geology, 6: 461-^63. Baldwin, E.M. 1976. Geology of Oregon. Revised edition, 147 pages. Dubuque, Iowa: Kendall/Hunt. Behnke, R.J. 1992. Native Trout of Western North America. American Fisheries Soci- ety Monograph, 6: 275 pages. Bisson, P.A., and C.E. Bond 1971. Origin and Distribution of the Fishes of Harney Basin, Oregon. Copeia, 1971(2):268-281. Bobo, R. 1991. Basin Evolution of Arbon Valley and Its Relation to Snake River Plain Volcanism and Basin and Range Tectonics. Geological Soci- ety of America Abstracts with Programs, 23(3):6. [Abstract.] Bright, R.C. 1963. Pleistocene Lakes Thatcher and Bonneville, Southeastern Idaho. 292 pages. Doctoral dissertation, University of Minnesota, Minne- apolis, Minnesota. Brown, W.J., and M.R. Rosen 1995. Was There a Pliocene-Pleistocene Fluvial-Lacustrine Connection between Death Valley and the Colorado River? Quaternary Re- search, 43:286-296. Buising, A.V. 1990. The Bouse Formation and Bracketing Units, Southeastern Califor- nia and Western Arizona: Implications for the Evolution of the Proto-Gulf of California and the Lower Colorado River. Journal of Geophysical Research, 95(B12):20,l 11-20,132. Burchfiel, B.C., G.S. Hamill IV, and D.E. Williams 1983. Structural Geology of the Montgomery Mountains and the North- ern Half of the Nopah and Resting Spring Ranges, Nevada and Cal- ifornia. Bulletin of the Geological Society of America, 94: 1359-1376. Burchfiel, B.C., K.V. Hodges, and L.H. Royden 1987. Geology of Panamint Valley-Saline Valley Pull-Apart System, Cal- ifornia: Palinspastic Evidence for Low-Angle Geometry of a Neo- gene Range-Bounding Fault. Journal of Geophysical Research, 92(B10): 10,422-10,426. Carr, W.J. 1984. Regional Structural Setting of Yucca Mountain, Southwestern Ne- vada, and Late Cenozoic Rates of Tectonic Activity in Part of the Southwestern Great Basin, Nevada and California. United States Geological Survey Open-File Report, 84-854: 109 pages. Christiansen, R.L., and R.S. Yeats 1992. Post-Laramide Geology of the U.S. Cordilleran Region. In B.C. Burchfiel, P.W. Lipman, and M.L. Zoback, editors, The Geology of North America, volume G-3: The Cordilleran Orogen, Contermi- nous U.S., pages 261-406. Boulder, Colorado: Geological Society of America. Currey, D.R., C.G. Oviatt, and J.E. Czarnomski 1984. Late Quaternary Geology of Lake Bonneville and Lake Waring. In G.J. Kerns and R.L. Kerns, Jr., editors, Geology of Northwest Utah, Southern Idaho, and Northeast Nevada. Utah Geological Associa- tion Publication, 13:227-237. Deacon, J.E., and J.E. Williams 1984. Annotated List of the Fishes of Nevada. Proceedings of the Biolog- ical Society of Washington, 97(l):103-118. Dibblee, Jr., T.W. 1967. Areal Geology of the Western Mojave Desert. United States Geo- logical Survey Professional Papers, 522: 153 pages. DiGuiseppi, W.H., and J.M. Bartley 1991. Stratigraphic Effects of Change from Internal to External Drainage in an Extending Basin, Southeastern Nevada. Bulletin of the Geo- logical Society of America, 103:48-55. Durham, J.W., and E.C. Allison 1960. The Geologic History of Baja California and its Marine Faunas. Systematic Zoology, 9(2):47-91. Echelle, A.A., and T.E. Dowling 1992. Mitochondrial DNA Variation and Evolution of the Death Valley Pupfishes (Cyprinodon, Cyprinodontidae). Evolution, 46(1): 193-206. Firby, J.R. 1993. A New Species of Vorticifex (Gastropoda: Planorbidae) from Late Cenozoic Lake Deposits, Nevada. Journal of Paleontology, 67(3): 370-374. Fitzpatrick, J.A., and J.L. Bischoff 1993. Uranium-Series Dates on Sediments of the High Shoreline of Pana- mint Valley, California. United States Geological Survey Open-File Report, 93-232: 15 pages. Fremont, J.C. 1845. Report of the Exploring Expedition to the Rocky Mountains in the Year 1842, and to Oregon and North California in the Years 1843-'44. 693 pages, maps, tables. U.S. Congress, 28th Congress, 2nd session, 1844-1845, Senate Document, volume 11, number 174; Serial 461. Washington, D.C.: Gales and Seaton. Garside, L.J., and J.H. Schilling 1979. Thermal Waters of Nevada. Bulletin of the Nevada Bureau of Mines and Geology, 91:1-163, plate 1. Gehr, K.D., and T.M. Newman 1978. Preliminary Note on the Late Pleistocene Geomorphology and Ar- chaeology of the Harney Basin, Oregon. Ore Bin, 40(10): 165-170. Gester, G.C. 1962. The Geological History of Eagle Lake, Lassen County, California. Occasional Papers of the California Academy of Sciences, 34: 1-29. Gilluly, J. 1929. Possible Desert-Basin Integration in Utah. Journal of Geology, 37: 672-682. Gregg, W.O., and D.W. Taylor 1965. Fontelicella (Prosobranchia: Hydrobiidae), a New Genus of West American Freshwater Snails. Malacologia, 3(1): 103-110. Hamlin, R. 1996. Conservation Genetics of Remnant Springsnail, Pyrgulopsis wongi, Populations in Desert Valleys of California and Nevada. 53 pages. 274 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Master's thesis, University of Nevada, Reno, Nevada. Harold, A.S., and R.D. Mooi 1994. Areas of Endemism: Definition and Recognition Criteria. System- atic Biology, 43(2):261-266. Hershler, R. 1987. Redescription of Assiminea infima Berry, 1947, from Death Valley, California. Veliger, 29(3):274-288. 1989. Springsnails (Gastropoda: Hydrobiidae) of Owens and Amargosa River (Exclusive of Ash Meadows) Drainages, Death Valley Sys- tem, California-Nevada. Proceedings of the Biological Society of Washington, 102( 1): 176-248. 1994. A Review of the North American Freshwater Snail Genus Pyrgu- lopsis (Hydrobiidae). Smithsonian Contributions to Zoology, 554: 115 pages. 1995. New Freshwater Snails of the Genus Pyrgulopsis (Rissooidea: Hy- drobiidae) from California. Veliger, 38(4):343-373. 1998. A Systematic Review of the Hydrobiid Snails (Gastropoda: Ris- sooidea) of the Great Basin, Western United States, Part I: Genus Pyrgulopsis. Veliger, 41(1):1?132. Hershler, R., and W.L. Pratt 1990. A New Pyrgulopsis (Gastropoda: Hydrobiidae) from Southeastern California, with a Model for Historical Development of the Death Valley Hydrographic System. Proceedings of the Biological Society of Washington, 103(2):279-299. Hershler, R., and D.W. Sada 1987. Springsnails (Gastropoda: Hydrobiidae) of Ash Meadows, Amar- gosa Basin, California-Nevada. Proceedings of the Biological So- ciety of Washington, 100(4):776-843. Hershler, R., and F.G. Thompson 1987. North American Hydrobiidae (Gastropoda: Rissoacea): Redescrip- tion and Systematic Relationships of Tryonia Stimpson, 1865 and Pyrgulopsis Call and Pilsbry, 1886. Nautilus, 101(l):25-32. Hildebrand, R.T., and K.R. Newman 1985. Miocene Sedimentation in the Goose Creek Basin, South-Central Idaho, Northeastern Nevada, and Northwestern Utah. In R.M. Flores and S.S. Kaplan, editors, Cenozoic Paleogeography of the Western United States. Rocky Mountain Paleogeography Sympo- sium, 3:55-70. Denver, Colorado: Society of Economic Paleontol- ogists and Mineralogists (Rocky Mountain Section). Hooke, R. Le B. 1998. Did Lake Manly Overflow at Ash Hill? Earth Surface Processes andLandforms, 23:377-384. Hubbs, C.L., and R.R. Miller 1948. The Zoological Evidence: Correlation between Fish Distribution and Hydrographic History in the Desert Basins of Western United States. In The Great Basin, with Emphasis on Glacial and Postgla- cial Times. Bulletin of the University of Utah, 38(20): 17-167, 1 plate. [Biological Series, 10(7).] Hubbs, C.L., R.R. Miller, and L.C. Hubbs 1974. Hydrographic History and Relict Fishes of the North-Central Great Basin. Memoirs of the California Academy of Sciences, 7: 259 pages. Huber, N.K. 1981. Amount and Timing of Uplift and Tilt of the Central Sierra Ne- vada?Evidence from the Upper San Joaquin Basin. United States Geological Survey Professional Papers, 1197: 28 pages. Jannik, N.O., F.M. Phillips, G.I. Smith, and D. Elmore 1991. A 36C1 Chronology of Lacustrine Sedimentation in the Pleistocene Owens River System. Bulletin of the Geological Society of Amer- ica, 103:1146-1159. Jayko, A.S. 1990. Shallow Crustal Deformation in the Pahranagat Area, Southern Ne- vada. In B.P. Wernicke, editor, Basin and Range Extensional Tec- tonics near the Latitude of Las Vegas, Nevada. Memoirs of the Geological Society of America, 176:213-236. King, G.Q. 1982. Morphometry of Great Basin Playas. 137 pages. Doctoral disserta- tion, University of Utah, Salt Lake City, Utah. La Rocque, A. 1960. Molluscan Faunas of the Flagstaff Formation of Central Utah. Memoirs of the Geological Society of America, 78: 100 pages. Lawrence, R.D. 1976. Strike-Slip Faulting Terminates the Basin and Range Province in Oregon. Bulletin of the Geological Society of America, 87: 846-850. Lindberg, D.N., and M.A. Hemphill-Haley 1988. Late-Pleistocene Pluvial History of the Alvord Basin, Harney Co., Oregon. Northwest Science, 62(2):81. [Abstract.] Loomis, D.P., and D.W. Burbank 1988. The Stratigraphic Evolution of the El Paso Basin, Southern Califor- nia: Implications for the Miocene Development of the Garlock Fault and Uplift of the Sierra Nevada. Bulletin of the Geological Society of America, 100:12-28. Lueddecke, S.B., N. Pinter, and P. Gans 1998. Plio-Pleistocene Ash Falls, Sedimentation, and Range-Front Fault- ing along the White-Inyo Mountains Front, California. Journal of Geology, 106:511-522. Luedke, R.G., and R.L. Smith 1991. Quaternary Volcanism in the Western Conterminous United States. In R.B. Morrison, editor, The Geology of North America, volume K- 2: Quaternary Nonglacial Geology, Conterminous U.S., pages 75- 92. Boulder, Colorado: Geological Society of America. Malde, H.E. 1968. The Catastrophic Late Pleistocene Bonneville Flood in the Snake River Plain, Idaho. United States Geological Survey Professional Papers, 596: 52 pages. Meisling, K.E., and R.J. Weldon 1989. Late Cenozoic Tectonics of the Northwestern San Bernardino Mountains, Southern California. Bulletin of the Geological Society of America, 101:106-128. Mifflin, M.D., and M.M. Wheat 1979. Pluvial Lakes and Estimated Pluvial Climates of Nevada. Bulletin of the Nevada Bureau of Mines and Geology, 94:1-57, plate 1. Miller, R.R. 1946. Correlation between Fish Distribution and Pleistocene Hydrogra- phy in Eastern California and Southwestern Nevada, with a Map of the Pleistocene Waters. Journal of Geology, 54(1):43?53. 1981. Coevolution of Deserts and Pupfishes (Genus Cyprinodon) in the American Southwest. In R.J. Naiman and D.L. Soltz, editors, Fishes in North American Deserts, pages 39-94. New York: John Wiley & Sons. Miller, R.R., and C.L. Hubbs 1960. The Spiny-Rayed Cyprinid Fishes (Plagopterini) of the Colorado River System. Miscellaneous Publications of the Museum of Pale- ontology, University of Michigan, 115:1-39, plates 1-3. Miller, W.J. 1928. Geology of Deep Springs Valley, California. Journal of Geology, 36:510-525. Minckley, W.L., D.A. Hendrickson, and C.E. Bond 1986. Geography of Western North American Freshwater Fishes: De- scription and Relationships to Intracontinental Tectonism. In C.H. Hocutt and E.O. Wiley, editors, Zoogeography of North American Freshwater Fishes, pages 519-613 (+ bibliography). New York: John Wiley & Sons. Mladenka, G.C. 1992. The Ecological Life History of the Bruneau Hot Springs Snail (Pyr- gulopsis bruneauensis). 116 pages. Unpublished (final) report pre- pared for the United States Fish and Wildlife Service. NUMBER 33 275 Morrison, R.B. 1991. Quaternary Stratigraphic, Hydrologic, and Climatic History of the Great Basin, with Emphasis on Lakes Lahontan, Bonneville, and Tecopa. In R.B. Morrison, editor, The Geology of North America, volume K-2: Quaternary Nonglacial Geology, Conterminous U.S., pages 283-320. Boulder, Colorado: Geological Society of America. Nagy, E.A., and B. Murray 1996. Plio-Pleistocene Deposits Adjacent to the Manix Fault: Implica- tions for the History of the Mojave River and Transverse Ranges Uplift. Sedimentary Geology, 103:9-21. Noel, M.S. 1954. Animal Ecology of a New Mexico Springbrook. Hydrobiologia, 6: 120-135. Nyquist, D. 1963. The Ecology of Eremichthys acros, an Endemic Thermal Species of Cyprinid Fish from Northwestern Nevada. 247 pages, 27 plates, 2 maps. Master's thesis, University of Nevada, Reno, Nevada. O'Connor, J.E. 1993. Hydrology, Hydraulics, and Geomorphology of the Bonneville Flood. United States Geological Survey Special Papers, 21 A: 83 pages. Ore, H.T. 1982. Tertiary and Quaternary Evolution of the Landscape in the Poca- tello, Idaho, Area. Northwest Geology, 11:31-36. Osier, T.A. 1995. Neogene-Quaternary Geology and Geomorphology of the Poca- tello Basin, Southeastern Idaho: Effects of the Migrating Yellow- stone Hotspot. Geological Society of America Abstracts with Programs, 27(6): A-55. [Abstract.] Piper, A.M., T.W. Robinson, and C.F. Park, Jr. 1939. Geology and Ground-Water Resources of the Harney Basin, Ore- gon. United States Geological Survey Water-Supply Paper, 841: 189 pages, plates 1-20. Platnick, N.I., and G. Nelson 1978. A Method of Analysis for Historical Biogeography. Systematic Zo- ology, 21 (\):\-\6. Ponder, W.F., D.J. Colgan, G.A. Clark, A.C. Miller, and T. Terzis 1994. Microgeographic, Genetic, and Morphological Differentiation of Freshwater Snails?the Hydrobiidae of Wilsons Promonotory, Vic- toria, South-Eastern Australia. Australian Journal of Zoology, 42: 557-678. Putnam, W.C. 1949. Quaternary Geology of the June Lake District, California. Bulletin of the Geological Society of America, 60:1281-1302, plates 1-7. Reheis, M., and R. Morrison 1997. High, Old Pluvial Lakes of Western Nevada. In P.K. Link and B.J. Kowallis, editors, Proterozoic to Recent Stratigraphy, Tectonics, and Volcanology, Utah, Nevada, Southern Idaho and Central Mex- ico. Brigham Young University Geology Studies, 42(l):459-492. Reheis, M.C., A.M. Sarna-Wojcicki, R.L. Reynolds, C.A. Repenning, and M.D. Mifflin 2002. Pliocene to Middle Pleistocene Lakes in the Western Great Basin: Ages and Connections. In R. Hershler, D.B. Madsen, and D.R. Cur- rey, editors, Great Basin Aquatic Systems History. Smithsonian Contributions to the Earth Sciences, 33:53-108. Reheis, M.C., and T.L. Sawyer 1997. Late Cenozoic History and Slip Rates of the Fish Lake Valley, Emi- grant Peak, and Deep Springs Fault Zones, Nevada and California. Bulletin of the Geological Society of America 109(3): 280-299. Repenning, C.A., T.R. Weasma, and G.R. Scott 1995. The Early Pleistocene (Latest Blancan-Earliest Irvington) Froman Ferry Fauna and History of the Glenns Ferry Formation, South- western Idaho. Bulletin of the United States Geological Survey, 2105: 86 pages. Rodgers, D.W., H.T. Ore, R. Bobo, E.P. Henderson, and A.D. Huerta 1991. Drainage Reversal in Three Basin and Range Grabens, Southeast- ern Idaho: Evidence for Miocene Passage of the Yellowstone Hot Spot. Geological Society of America Abstracts with Programs, North-Central Section, 23(3):88. [Abstract.] Rowan, L.C., and P.H. Wetlaufer 1981. Relation between Regional Lineament Systems and Structural Zones in Nevada. Bulletin of the American Association of Petro- leum Geologists, 65(8): 1414-1432. Russell, I.C. 1889. Quaternary History of Mono Valley, California. Eighth Annual Re- port of the United States Geological Survey to the Secretary to the Interior, 1886-'87, part 1:261-394, plates 17-44. Sada, D.W., H.B. Britten, and P.F. Brussard 1995. Desert Aquatic Ecosystems and the Genetic and Morphological Di- versity of Death Valley System Speckled Dace. American Fisheries Society Symposium, 17:350-359. Scott, D.H., and N.J. Trask 1971. Geology of the Lunar Crater Volcanic Field, Nye County, Nevada. United States Geological Survey Professional Papers, Contribu- tions to Astrogeology, 599-1: 22 pages, plate 1. Small, E.E., and R.S. Anderson 1995. Geomorphically Driven Late Cenozoic Rock Uplift in the Sierra Nevada, California. Science, 270:277-280. Smith, G.R. 1978. Biogeography of Intermountain Fishes. In K.T. Harper and J.L. Re- veal, symposium organizers, Intermountain Biogeography: A Sym- posium. Great Basin Naturalist Memoirs, 2:17?42. Smith, P.B. 1960. Fossil Foraminifera from the Southeastern California Deserts. United States Geological Survey Professional Papers, 400-B: 278-279. 1970. New Evidence for a Pliocene Marine Embayment along the Lower Colorado River Area, California and Arizona. Bulletin of the Geo- logical Society of America, 81:1411-1420. Smith, R.S.U. 1976. Late-Quaternary Pluvial and Tectonic History of Panamint Valley, Inyo and San Bernardino Counties, California. 295 pages. Doctoral dissertation, California Institute of Technology, Pasadena, Califor- nia. Snyder, C.T., G. Hardman, and F.F. Zdenek 1964. Pleistocene Lakes in the Great Basin. United States Geological Survey Miscellaneous Geologic Investigations, map 1-416. Spencer, J.E., and P.J. Patchett 1997. Sr Isotope Evidence for a Lacustrine Origin for the Upper Miocene to Pliocene Bouse Formation, Lower Colorado Trough, and Impli- cations for Timing of Colorado Plateau Uplift. Bulletin of the Geo- logical Society of America, 109(6):767-778. Stewart, J.H. 1988. Tectonics of the Walker Lane Belt, Western Great Basin: Mesozoic and Cenozoic Deformation in a Zone of Shear. In W.G. Ernst, edi- tor, Metamorphism and Crustal Evolution of the Western United States, Rubey volume, 7, pages 683-713. Englewood Cliffs, New Jersey: Prentice Hall. Taylor, D.W. 1960. Distribution of the Freshwater Clam Pisidium ultramontanum: A Zoogeographic Inquiry. American Journal of Science, Bradley vol- ume, 258-A:325-334. 1966a. Summary of North American Blancan Nonmarine Mollusks. Mala- cologia, 4(1): 1-172. 1966b. A Remarkable Snail Fauna from Coahuila, Mexico. Veliger, 9(2): 152-228. 1975. Early Tertiary Mollusks from the Powder River Basin, Wyo- 276 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ming-Montana, and Adjacent Regions. United States Geological Survey, Open-File Report, 75-331: 515 pages. 1985. Evolution of Freshwater Drainages and Molluscs in Western North America. In C.J. Smiley, editor, Late Cenozoic History of the Pa- cific Northwest; Interdisciplinary Studies on the Clarkia Fossil Beds of Northern Idaho, pages 265-321. San Francisco, California: American Association for the Advancement of Sciences. 1987. Fresh-Water Molluscs from New Mexico and Vicinity. Bulletin of the New Mexico Bureau of Mines & Mineral Resources, 116: 50 pages. Taylor, D.W., and R.C. Bright 1987. Drainage History of the Bonneville Basin. In R.S. Knoop and R.E. Cohenour, editors, Cenozoic Geology of Western Utah, Sites for Precious Metal and Hydrocarbon Accumulations. Utah Geological Association Publication, 16:239-256. Taylor, D.W., and G.R. Smith 1981. Pliocene Molluscs and Fishes from Northeastern California and Northwestern Nevada. Contributions from the Museum of Paleon- tology, University of Michigan, 25(18):339-413. Thompson, F.G. 1970 ("1969"). Some Hydrobiid Snails from Georgia and Florida. Qua- ternary Journal of the Florida Academy of Sciences, 32:241-265. [Date on title page is 1969; actual date of publication is 1970.] 1979. The Systematic Relationships of the Hydrobiid Snail Genus Nym- phophilus Taylor 1966 and the Status of the Subfamily Nympho- philinae. Malacological Review, 12:41-49. 1984. North American Freshwater Snail Genera of the Hydrobiid Sub- family Lithoglyphinae. Malacologia, 25:109-141. 1995. A New Freshwater Snail from the Coosa River, Alabama (Gas- tropoda: Prosobranchia: Hydrobiidae). Proceedings of the Biologi- cal Society of Washington, 108(3):502-507. Thompson, F.G., and R. Hershler In press. Two Genera of North American Freshwater Snails: Marstonia Baker, 1926, Resurrected to Generic Status, and Floridobia, New Genus (Prosobranchia: Hydrobiidae: Nymphophilinae). The Ve- liger, 45(3):278-280. Unruh, J.R. 1991. The Uplift of the Sierra Nevada and Implications for Late Cenozoic Epeirogeny in the Western Cordillera. Bulletin of the Geological Society of America, 103:1395-1404. Walcott, CD. 1897. The Post-Pliocene Elevation of the Inyo Range, and the Lake Beds of Waucobi Embayment, Inyo County, California. Journal of Geol- ogy, 5:340-348. Weide, D.L. 1974. Postglacial Geomorphology and Environments of the Warner Val- ley-Hart Mountain Area, Oregon. 293 pages. Doctoral dissertation, University of California, Los Angeles, California. Wernicke, B., and 18 coauthors 1996. Origin of High Mountains in the Continents: The Southern Sierra Nevada. Science, 271:190-193. Wheeler, H.E., and E.F. Cook 1954. Structural and Stratigraphic Significance of the Snake River Cap- ture, Idaho-Oregon. Journal of Geology, 62(6):525-536. Wiley, E.O. 1988. Vicariance Biogeography. Annual Review of Ecology and Systemat- ics, 19:513-542. Williams, T.R., and M.S. Bedinger 1984. Selected Geologic and Hydrologic Characteristics of the Basin and Range Province, Western United States: Pleistocene Lakes and Marshes. United States Geological Survey Miscellaneous Investi- gations Series, map 1-1522-D. Wolfe, J.A., H.E. Schorn, C.E. Forest, and P. Molnar 1997. Paleobotanical Evidence for High Altitudes in Nevada during the Miocene. Science, 276:1672-1675. Yen, T.-C. 1950. A Molluscan Fauna from the Type Section of the Truckee Forma- tion. American Journal of Science, 248(3): 180-193. Anthropogenic Changes in Biogeography of Great Basin Aquatic Biota Donald W. Sada and Gary L. Vinyard ABSTRACT The Great Basin, the most arid region in North America, has small, widely spaced wetlands compared with those in more mesic regions. These wetland habitats generally support aquatic taxa that are widespread throughout North America. Recent work has com- plemented the studies of endemic fishes by Carl Hubbs and Robert Miller and has served to identify a diverse fauna of insects and mollusks endemic to the Great Basin. The presence of this fauna in comparatively small wetlands shows that habitat size is not a good indicator of the importance of these wetlands to unique biological communities. Historical and current records show that anthropogenic activi- ties during the last 120 years have modified the structure of Great Basin aquatic communities by altering habitats and by the translo- cation of species. Fifty nonnative fish taxa and several invertebrate taxa have been introduced into the region by the public and/or by fishery management agencies. Twenty-four fish taxa endemic to the Great Basin also have been translocated into habitats both within and outside of their native range. Most of these transloca- tions were undertaken to create refuge populations and, thus, to reduce the possibility of extinction. Within the Great Basin, introductions of nonnative species and habitat modification have caused the extinction of 16 endemic spe- cies, subspecies, or other distinctive populations (12 fishes, three mollusks, and one aquatic insect) since the late 1800s. Declines in abundance or distribution were attributable (in order of decreasing importance) to water flow diversions, competitive or predatory interactions with nonnative species, livestock grazing, introduc- tions for sport fisheries management, groundwater pumping, spe- cies hybridization, timber harvest, pollution, recreation, and habitat urbanization. Most affected taxa were influenced by more than one of these factors. Donald W. Sada, Division of Hydrologic Sciences, Desert Research Institute, 2215 Raggio Parkway, Reno, Nevada 89512-1095, USA. Gary L. Vinyard (deceased), formerly at the Department of Biology, University of Nevada, Reno, Nevada 89557, USA. Reviewers: Robert Hershler, Division of Mollusks, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560- 0118, USA; W. Ronald Heyer, Division of Herpetology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0162, USA; and two anonymous reviewers. The temporal pattern of decline in endemic taxa was examined by determining the calendar decades of first population loss, of major decline (loss of one-half of either a taxon's distribution or abundance), and of extinction for 199 endemic taxa (102 fishes, 85 mollusks, nine aquatic insects, two amphibians, and one fairy shrimp). Population loss has affected approximately 50 percent of taxa for which information was available (135 distinct taxa: 99 fishes, 24 mollusks, and all taxa in the three other aquatic animal groups), and 58 percent of these taxa have suffered major declines. Differences among rates of population loss, major decline, and extinction were not significant (ANCOVA,/? > 0.05). Declines and extinctions were first recorded in the late 1800s. Rates peaked first after World War I and again in the 1970s after a long increase that began after World War II. Status was comparatively static during both World Wars. This pattern indicates that declines can be attrib- uted to regional economic conditions and increased immigration. Declines slowed in the 1980s and 1990s because most taxa had previously declined, not because threats had diminished. Declines during the last 120 years have been greatest in the most narrowly distributed and vulnerable populations. All extinct taxa and most taxa suffering major declines (68 percent) had fewer than five populations. If past trends continue into the future, additional extinctions will occur (primarily in narrowly distributed taxa), and extinctions also may begin to affect widespread taxa. These changes will accompany environmental change that characterizes consumption patterns of increasing human populations. Avoiding future changes in Great Basin biogeography that result from declines in taxon status will require new, innovative programs that protect wetland habitats from environmental degradation and the deleterious effects of nonnative species while allowing appropriate human uses of wetland resources. Introduction The Great Basin, the driest physiographic province in North America, contains fewer small rivers, lakes, streams, springs, and marshes than do more mesic regions, and all of these waterbodies are more widely dispersed in the Great Basin than elsewhere. Aquatic communities in the region include species common to other North American wetlands as well as a diverse fauna of endemic fishes, mollusks, and aquatic insects predom- inantly associated with springs. Although these unique charac- teristic have long been recognized (Gilbert, 1893; Brues, 277 278 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1932), recent surveys have revealed a surprisingly high diver- sity of additional endemic macroinvertebrates (e.g., Hershler and Sada 1987; Shepard, 1990; Hershler, 1998). We examine herein the major causes of historical change in biodiversity in an effort to understand the predominant ways in which anthro- pogenic uses of wetlands have affected Great Basin biogeogra- phy. Humans have used Great Basin wetlands since the early Ho- locene (Janetski and Madsen, 1990). Intermittent settlement near wetlands and daily activities of hunting and gathering nat- urally tended to focus upon the biotic resources concentrated at these habitats. For a short period before the region was settled by European immigrants, a few indigenous tribes also diverted water to irrigate land and increase production of commonly eaten native plants. The Fremont Tribe diverted water from large rivers near the Wasatch Range (Madsen, 1989), and the Owens River basin Paiutes diverted streams from the east flank of the Sierra Nevada (Steward, 1933). Great Basin wetlands also produced food organisms, such as Cyprinodon radiosus (Owens pupfish), Catostomus tahoensis (Tahoe sucker), Chas- mistes cujus (cui-ui), Oncorhynchus clarki henshawi (Lahontan cutthroat trout), Gila bicolor (tui chub), and Ephydra hians (brine fly), which were eaten by Paiutes along the eastern Si- erra Nevada (Steward, 1933; Knack and Stewart, 1984). Exca- vation of sites known to have been inhabited by Fremont Tribe members revealed that their diet included Catostomus ardens (Utah sucker), Gila atraria (Utah chub), and Oncorhynchus clarki utah (a cutthroat trout subspecies) (Janetski, 1990). Greenspan (1990) also documented fish remains from archaeo- logical sites in the Great Basin of Oregon, and Drews (1990) reviewed the dietary role of Great Basin shellfishes. Although Native Americans used and altered Great Basin wetlands, the major anthropogenically induced changes to aquatic biogeography accompanied habitat alterations and in- troductions of nonnative taxa during the last 120 years. Since the beginning of the twentieth century, most rivers in the region have been dammed and diverted for irrigation, flood control, or power generation (La Rivers, 1962; Sigler and Sigler, 1987). Many springs and streams have been modified by livestock grazing, water flow diversion, and groundwater use (see Miller, 1961; Dudley and Larson, 1976; Fleischner, 1994). Native aquatic biota also have been affected by the introduction of 50 nonnative fish species for sport, pest management (Sigler and Sigler, 1987), and unspecified recreational or commercial inter- ests (Moyle, 1984). Fisheries management agencies have en- hanced nonnative sport fish populations by poisoning thou- sands of miles of streams, causing the consequent reduction or elimination of native fish and macro invertebrate populations (Moyle et al., 1986; Rinne and Turner, 1991; Andersen and Deacon, 1996). These habitat modifications and introductions have caused the decline of many populations of native taxa and have driven some species to extinction (Minckley and Deacon, 1968; Williams et al., 1985; Miller et al., 1989; Minckley and Douglas, 1991). The changing status of fishes and their habitats have been re- ported as discrete events affecting localized habitats or individ- ual taxa (e.g., Williams et al., 1985; Williams et al., 1989; Moyle et al., 1995). The status of endemic invertebrates and the effects of human activities on smaller aquatic habitats (e.g., springs not occupied by native fishes) have not been examined in detail. We assess herein how habitat modifications and spe- cies translocations have affected Great Basin aquatic biogeog- raphy by altering the historical distribution of endemic species. This information was then used to examine the temporal pat- tern of past changes as a means of providing insight into poten- tial future changes in the biogeography of aquatic animals en- demic to the Great Basin. Habitat and Biotic Diversity Taxonomic studies during the last 140 years have consis- tently increased recognition of endemic plants and animals in Great Basin wetlands (see Hubbs and Miller, 1948a; Hershler, 1998). Although many taxa have been described, a number of additional populations may qualify for recognition as either en- demic species or endemic subspecies (see Hubbs et al., 1974; Deacon and Williams, 1984; Sada et al., 1995; Hamlin, 1996). Recent descriptions of mollusks and insects (e.g., Hershler and Sada, 1987; Shepard, 1990; Hershler, 1998) indicate that addi- tional undiscovered aquatic macroinvertebrates in Great Basin springs may exist. Currently, 118 species (two amphibians, 23 fishes, 84 mollusks, eight insects, and one fairy shrimp) and 45 subspecies (one aquatic insect and 44 fishes) are endemic to the Great Basin. We also studied 36 additional endemic forms that have been identified in published and unpublished reports as morphologically or genetically distinct. All of these described and undescribed forms will be collectively referred to herein as "distinctive taxa." Distinctive taxa occupy a wide diversity of habitats including rivers, lakes, and streams, as well as cold and thermal springs (La Rivers, 1962; Sigler and Sigler, 1987; Hershler, 1998). Most of these taxa occupy habitats below 2200 m in elevation, where wetlands are most common and the historical use of wa- ter by humans has been greatest. Most distinctive taxa occupy specific habitats. Several taxa are primarily lentic but require lotic habitats for spawning (e.g., Catostomus warnerensis, Chasmistes cujus, Chasmistes liorus), and some continuously inhabit both lentic and lotic habitats (e.g., Catostomus ardens, Richardsonius egregius). Others only inhabit thermal springs (e.g., Cyprinodon diabolis, Eremichthys acros, Gila boraxo- bius, Pyrgulopsis militaris) or cold springs (e.g., P. ruinosa, P. wongi). Cyprinodon radiosus and Iotichthys phlegethontis oc- cupy both lentic and spring habitats. Most endemic macroin- vertebrates are restricted to springs, but several are lentic (e.g., P. nevadensis, Capnia lucustra, Artemia monicd). The primary habitats of most distinctive taxa are springs (153 taxa), fol- lowed by lotic (18 taxa) and lentic (10 taxa) habitats; 10 taxa are lentic species that require lotic spawning habitat; and eight NUMBER 33 279 taxa are equally abundant in both springs and streams (see Ap- pendix). Many habitats are occupied by a single endemic, and many habitats support several endemics. Communities with the highest diversity of endemics are concentrated in thermal springs of southern Nevada (e.g., Ash Meadows, Pahranagat Valley), which support endemic fishes, mollusks, and aquatic insects (Hershler and Sada, 1987; U.S. Fish and Wildlife Ser- vice, 1990; Polhemus and Polhemus, 1994; Hershler, 1998). The unique aspects of these wetland communities are not limited to the aquatic animals inhabiting them. Great Basin wetlands of Nevada and southeastern Oregon also support 19 endemic plant species and five endemic plant varieties (Nevada Heritage Program Data Base, unpublished; Oregon Heritage Program Data Base, unpublished); four endemic vole subspe- cies?namely, Microtus californicus vallicola (Owens vole), M. c. scirpensis (Amargosa vole), M. montanus fucosus (Pahra- nagat vole), and M. m. nevadensis (Ash Meadows vole)?in- habit southern Nevada and southern California wetlands (Bailey, 1898; Hall, 1946). No endemic plants or mammals, however, are known from wetlands in the Utah portion of the Great Basin. Introductions and Translocations Many nonnative species have been introduced into North American waters, and these introductions have changed the structure of native vertebrate and invertebrate communities and caused extinctions of native biota through predation, hy- bridization, and competition (see Courtenay and Stauffer, 1984; Mooney and Drake, 1986; Moyle et al., 1986; Johnson and Padilla, 1996). Nonnative aquatic species in the Great Ba- sin were introduced from other parts of North America and from Europe, Asia, Africa, and South America (Deacon and Williams, 1984; Sigler and Sigler, 1987). Many of these intro- duced species are now widespread, and most Great Basin fish assemblages are dominated by nonnative taxa. Deacon and Williams (1984) and Sigler and Sigler (1987) identified 50 nonnative fish taxa in the region, which exceeds the number of native Great Basin fish taxa (43 species). Most intentional in- troductions were made to accommodate the recreational sport- fishing industry, but some fishes were introduced as biological control agents. Some fish species were released from home aquaria, and others escaped from commercial aquaculture fa- cilities (Courtenay et al., 1985; U.S. Fish and Wildlife Service, 1997a). Little is known about nonnative invertebrate introductions; several introduced species, however, have become widespread in the region, including Procambarus sp. and Pacifastacus le- nusculus (crayfishes), Corbicula manilensis (Asian clam), and Thiara (=Melanoides) tuberculata (red-rimmed thiara, a snail). Crayfish introductions may have reduced refuge populations of Cyprinodon radiosus (U.S. Fish and Wildlife Service, 1998a); T. tuberculata may have contributed to declines in springsnail populations in southern Nevada (Hershler and Sada, 1987); and the disappearance of Anodonta californica (California floater, a bivalve mollusk) from the Owens River basin may have been caused by Corbicula manilensis. Disappearance of Daphnia sp. (ostracode) from Lake Tahoe is attributed to the introduction of Mysis relicta (oppossum shrimp) (Richards et al., 1975; Mor- gan et al., 1978). No private or state programs exist to translocate endemic Great Basin macroinvertebrates or amphibians, and no records indicate that these species have been translocated. However, several state and Federal programs have actively translocated 24 endemic Great Basin fishes into refuges both within and outside of their native ranges (Appendix). Most fishes were translocated to expand their distributions and minimize threats, thus reducing the possibility of extinction. Other translocations were conducted to enhance sportfishing opportunities (e.g., Oncorhynchus clarki henshawi). Some species were estab- lished in single refuges whereas others have been broadly in- troduced (e.g., O. c. henshawi into 33 refuges). Some of these translocations probably prevented imminent extinction (e.g., Cyprinodon radiosus, Empetrichthys latos latos (Pahrump poolfish), and Crenichthys baileyi grandis (Hiko White River springfish). Refuge populations may influence biogeography in several ways. They may expand distributions and confound future bio- geographic interpretations. Also, many refuges have been es- tablished with very few individuals, which presumably repre- sents a comparatively small portion of the genome of the species. These factors suggest that caution should be used when interpreting biogeographic information accumulated from refuge populations. In addition, refuge construction may adversely affect other native taxa. Hershler (1989) reported that Pyrgulopsis perturbata (Fish Slough springsnail) disappeared from a spring that had been impounded to create a refuge for Cyprinodon radiosus. Effects of refuge development probably have been greatest when construction modified a previously unaltered habitat (e.g., channelized or impounded spring brooks), thus causing functional changes in the aquatic ecosys- tem's community structure. Status Changes We sought to determine the extent of decline for all aquatic taxa endemic to the Great Basin. Status information was gath- ered from personal communications and from approximately 100 published and unpublished survey reports (undertaken within the private sector and by federal and state agencies) writ- ten during the last 120 years. Early records were first prepared before widespread settlement of the region (Fremont, 1845; Beckwith, 1855; Merriam, 1893; Davidson, 1976), and they frequently have been used to compare historical and current distributions of most taxa. Distributional information was also gathered from early taxonomic surveys (e.g., Gilbert, 1893; Hubbs and Miller, 1948a). Most information provided by early records is less comprehensive than that from contemporary 280 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES records, which frequently describe demography and distribu- tion of extant populations. Early records, however, usually pro- vide baseline distributional information that can be effectively contrasted with current distributions. For each distinctive taxon, we identified the year in which taxonomic distinction was initially recognized. We then sought to determine the calendar decade(s) in which three possible levels of decline had occurred (if any): the first population loss (loss of first population, i.e., reduced distribution and absolute abundance), a decrease in historical distribution or absolute abundance by at least one-half (a "major decline"), and extinc- tion. Factors that affected the status of each taxon (i.e., threats, when known) were also identified. All known distinctive aquatic taxa in the Great Basin were considered: 102 fishes, 85 mollusks, nine aquatic insects, two amphibians, and one fairy shrimp. Status surveys, however, had not been conducted for all distinctive taxa (surveys have assessed the status of few recently described taxa), thus status could be reviewed for only 135 taxa?99 fishes, 24 mollusks, and all taxa in the three other aquatic animal groups. By deter- mining the year that each distinctive taxon was identified, we could assess the relationships between changes in status and taxonomic recognition. These dates often predated formal tax- onomic descriptions, and they were selected for analysis be- cause Federal and state agencies frequently have initiated sta- tus surveys soon after distinctive populations were identified. Relationships between status changes and recognition of dis- tinction may facilitate our understanding of temporal trends in status change. Positive correlation between these factors may indicate that status changes occur quickly after recognition of distinctive taxa. Either a negative correlation or the absence of correlation may suggest that decline rates occur sporadically, regardless of when distinctivenes is recognized. Data for each distinctive taxon are summarized in the Appendix. The information provided by these records has limitations. Surveys usually have been qualitative, and status changes are rarely quantified. It is also difficult to precisely identify when populations disappeared or declined because surveys have been sporadic as a result of budget limitations and changing agency priorities. Taxonomic studies have also revealed many endemic taxa (approximately 90 mollusks and aquatic insects since 1990) that were uncollected and undescribed until re- cently (e.g., Shepherd, 1992; Hamlin, 1996; Hershler, 1998). Many of these populations occupy springs that are in poor con- dition, and traditional human uses of these wetlands have caused several taxon extinctions and population losses in the last five years (Sada and Nachlinger, 1996; Hershler, 1998). Although it is not possible to determine how many extinctions of uncollected taxa have occurred, recent extinctions and popu- lation losses suggest that current status information may under- estimate the total number of historical Great Basin extinctions. Despite of these limitations, adequate status information was obtained for a sufficient number of taxa such that a meaningful analysis could be conducted. The status reports identified 10 factors (threats) that influ- enced abundance and distribution of distinctive endemic aquatic taxa: water flow diversions (affected 90 taxa), nonna- tive species (78 taxa, which include introductions for fisheries management purposes as well as other other nonnative taxa; e.g., Asian clam and crayfishes, mentioned above in "Introduc- tions and Translocations"), livestock grazing in riparian zones (54 taxa), introductions for sportfisheries management (33 taxa), groundwater pumping (17 taxa), hybridization of species (eight taxa), timber harvest (four taxa), pollution and recre- ation (three taxa each), and urbanization of habitat (two taxa). Most affected taxa were influenced by more than one factor: two affected 37 taxa, three affected 26 taxa, four affected 12 taxa, and five affected five taxa, whereas 67 taxa were affected by only one factor. This suggests synergistic effects may affect status changes?e.g., the combined effects of degraded habi- tats and nonnative species on endemic taxa may be greater than the summed effect of individual threats (Moyle and Light, 1996). Nonnative species, water flow diversions, and ground- water use caused most extinctions, whereas livestock grazing and pollution each caused a single extinction. All extinct taxa and most (68%) of the taxa experiencing major decline were narrowly distributed within the Great Basin (<5 populations each); this indicates that taxa with limited distribution are acutely vulnerable to catastrophic changes in status. Sixty-eight distinct taxa (50% of those reviewed) lost at least one population during the last 140 years, 78 experienced a ma- jor decline (58%), and 16 became extinct. Only 28 taxa main- tained what is believed to be their approximate historical abun- dance and distribution. Incidences of first population loss, major decline, and extinction occurred irregularly throughout the period of record (Figure 1). Rates of decline increased and decreased three times since the late 1800s; however, differ- ences among the rates of population loss, major decline, and extinction from the late 1800s through the late 1990s were not significant (analysis of covariance (ANCOVA): slopes, p> 0.05; regression, p < 0.05), indicating that changes in these rates followed similar temporal trends. The first rate increase began before the turn of the twentieth century, another began during the early decades after the turn of the century, and the third began during the 1950s. Decline rates peaked sharply and then leveled off somewhat for several decades. Both the fre- quencies of increasing decline rates and amplitudes of the three rates increased with time, indicating that periods of rapid de- clines have become more frequent. Several factors may explain general trends in the changes of status. Increases in the rates of decline appear to coincide with periods of human population expansion and economic growth. Status declines were most rapid during the 1890s, 1930s, and 1970s?periods that are associated (respectively) with the greatest revenues from mining and completion of the transcon- tinental railroad in the late 1800s (Hulse, 1991), the population expansion before the Great Depression, and the growing econ- omy prior to the 1970s recession. Conversely, rates of decline NUMBER 33 281 | | MAJOR DECLINE FIRST POPULATION LOSS DISTINCTIONS O - 100 1850 1870 1890 1910 1930 1950 1970 DECADES 1990 FIGURE I.?Cumulative percent decline (by calendar decade) of aquatic animals endemic to the Great Basin since the 1850s, as shown by percent of taxa affected by first population loss, major decline (a decrease in histor- ical distribution or abundace by at least one-half), and extinction. Differences among rates of first population loss (y = 0.516.x - 967.0, r2 = 0.88), major decline (y = 0.587.x - 1101.8, r2 = 0.87), and extinction (y = 0.143* - 268.4, r2 = 0.84) are not significant. "Cumulative distinctions" illustrate the cumulative increases (by decade) in the identifications of 199 endemic taxa determined in published and unpublished taxonomic studies to be distinct within the Great Basin. See "Status Changes" in the text for a description of distinctive taxa. - 200 - 150 -^ m o o were slowest during the period before the region was actively settled (pre-1850), during the second decade of the twentieth century (during World War I), and in the 1940s (during World War II). During the last century, status was most stable during both World Wars, when the national economy focused on ar- mament production and when immigration was constrained by fuel shortages and rationing. After World War II, status de- clines increased through the 1970s to an unprecedented level, and most taxa known in 1980 had either lost one population (68 taxa) or suffered major decline (78 taxa). Rates of decline slowed again during the 1980s and 1990s. This decrease in the decline of taxa coincides with enactment of environmental leg- islation (i.e., Endangered Species Act, Clean Water Act, Na- tional Environmental Policy Act, and many similar state stat- utes) and with initiation of many state, Federal, and private conservation programs, which seems to suggest these pro- grams are successful. It is problematic, however, to attribute these decreases solely to conservation programs because most distinctive taxa had already declined before these decades, leaving a comparatively smaller portion of endemic taxa to suffer their first population loss and/or major decline. In the future, major declines and the first population loss of a taxon will be recorded mostly for newly discovered taxa. Nonethe- less, the high number of extant taxa (199 distinct forms) com- pared with the numbers of distinct taxa having lost a popula- tion (68) and having undergone major decline (78) suggests that these conservation programs may have lowered the extinc- tion rate. Discussion Composition and distribution of Great Basin aquatic commu- nities, and the patterns of human resource use, both reflect the influence of regional aridity. Although Great Basin wetlands are small and isolated, they provide most of the water in a dry region and they contain a more diverse flora and fauna than other habitats (Thomas et al., 1979; Brode and Bury, 1984). Ri- parian vegetation provides nesting, roosting, and migratory habitat for resident and migratory birds and provides food and cover for mammals. Importance of these wetlands to terrestrial wildlife and to the endemic aquatic fauna shows that even small aquatic habitats are important to Great Basin biodiver- sity. Water has also been a focus for human activities in the Great Basin. Hunter-gatherers frequently placed their settle- ments near rivers, marshes, and springs (Fowler and Fowler, 1990), and all of the largest contemporary cities are proximal to large water resources (e.g., the Humboldt, Carson, and Truckee Rivers, Utah Lake, etc.). 282 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Anthropogenic use has affected terrestrial and aquatic eco- systems in many ways. Hypotheses relating the extinction of mammals and birds to the arrival of humans approximately 10,000 years ago suggest that some declines were attributed to human predation or climate change (Martin, 1990; Gray son, 1991). A thorough assessment of these impacts is beyond the scope of this paper, and additional evidence is needed to bet- ter understand how humans effected Holocene biotic changes in the Great Basin. However, information compiled during the last 120 years shows that historical status declines and extinc- tions began soon after the region was first settled by European immigrants. Evidence that the impacts of recent human activi- ties on biogeography greatly exceeded those of hunter-gather- ers is indicated by archaeological and fossil records and by taxonomic studies. Wetlands have continuously provided important resources for humans in the Great Basin. Hunter-gatherers did not live in permanent settlements because food resources were fre- quently scarce and movement was necessary to maximize for- aging opportunities that changed seasonally. In wet years, they utilized higher elevations to harvest pinon (Pinus mono- phylla and P. edulis) nuts. When food was scarce (e.g., during droughts) they temporarily settled near wetlands (Yellen, 1977; Fowler and Fowler, 1990). Springs and marshes border- ing pluvial lakes supported small animals, tubers, and seeds that were foods unavailable in other places (Janetski and Madsen, 1990). Halford (1998) identified a number of edible plant species (e.g., currant, Ribes sp.; elderberry, Sambucus sp.; and wild rose, Rosa woodsii) from pack-rat middens that were located near small springs and an ephemeral lake in the west central Great Basin. Hunter-gatherers affected endemic aquatic taxa by manipu- lating habitats and by harvesting fish and shellfish. It is diffi- cult to assess the effects that these activities had on Great Ba- sin biogeography because little information exists that indicates the response of aquatic taxa to human activities. Mehringer and Warren (1976) found that fire was used period- ically to clear mesquite (Prosopis spp.) bosques and to in- crease food availability in Ash Meadows, a spring province in southern Nevada. Although fire may dramatically affect ripar- ian vegetation, and fish populations have been extirpated by fire in the southwest (Rinne, 1996), little information suggests that it has been a major cause of extinction. Scoppettone et al. (1998) observed declines in native fish abundances after a fire along spring brooks in southern Nevada, but no extinctions were documented and populations soon recovered to preburn levels. The minimal impact of fire on Great Basin macroin- vertebrates was also indicated by the presence of large popu- lations of the springsnail Pyrgulopsis gibba in a small, re- cently burned spring in northern Nevada (Sada, field notes, 1996). Owens Basin Paiutes and the Fremont Tribe manipu- lated aquatic habitats by diverting streams for irrigation to in- crease the abundance of edible native vegetation (Steward, 1933; Madsen, 1989). These activities may have affected aquatic taxa by drying streams and by creating barriers that interrupted spawning migrations. The amount of habitat af- fected by these activities and the abundance and wide distri- bution of taxa affected by these diversions during early taxo- nomic surveys indicate that these activities did not cause extinction. A number of Native American tribes also harvested shell- fishes and endemic fishes (Steward, 1933; Knack and Stew- art, 1984; Drews, 1990; Greenspan, 1990; Janetski, 1990). The abundance of shellfishes and endemic fishes among mid- den fossils (Dansie, 1990; Greenspan, 1990; Janetski, 1990; Schmitt and Sharp, 1990) suggests that hunter-gatherers ac- tively harvested aquatic taxa. Harvesting did not cause extinc- tion because the midden fossils are all extant taxa that also were collected during initial taxonomic surveys. The occur- rence of extinct taxa in middens would suggest that harvesting or other Holocene events caused extinction. There can be little doubt that harvest and habitat manipulation by hunter-gather- ers affected the abundance of aquatic taxa and that similar ma- nipulations since the beginning of the twentieth century have extirpated many populations of aquatic taxa within the Great Basin (Minckley and Deacon, 1968; Williams et al., 1985; Miller et al., 1989; Hershler, 1998); however, no evidence documents extirpations that may have been caused by hunter- gatherer harvesting and habitat manipulations. Early Ho- locene extinctions of aquatic species are not likely attributable to hunter-gatherers because of their transitory settlement be- havior, their limited ability to manipulate large amounts of aquatic habitat, and the large size of habitats that supported harvested populations. Although hunter-gatherers affected the abundance of some endemic aquatic species, it is unlikely that their activities extensively affected aquatic biogeography. Biogeography could have been affected by hunter-gatherers who transported aquatic species outside of their native range, thus establishing new populations. Translocations may have been inadvertent (e.g., accidental captures in water containers when settlements were moved) or intentional. Effects of these translocations on biogeography are also difficult to determine, but their possible influence is indicated by several factors. If movement was accidental, a wide diversity of taxa (e.g., fishes, mollusks, and aquatic insects) would have been in- volved because endemic aquatic species occupy many aquatic habitats that would have been encountered because of the transitory lifestyle of hunter-gatherers. These taxa could have been accidentally captured in water jugs or carried in moist clothing and moved to new locations. Intentional transloca- tions would have served a purpose such as broadening the dis- tribution of food fishes, which suggests that a small number of taxa would have been selected. If intentional translocations were common, it is doubtful that they were limited to intraba- sin movements. Interbasin translocations from the diverse fish assemblages in surrounding basins (e.g., Colorado River, Kla- math River, Sacramento River, and Columbia River basins) would have been likely, and representatives of these faunas NUMBER 33 283 would have been recorded in the Great Basin during the early biological surveys. Evidence that hunter-gatherers translocated aquatic taxa is equivocal. Unintentional capture and movement was highly probable for some taxa, but establishment of viable popula- tions would have been less likely because of differences among aquatic habitats (e.g., water chemistry, water tempera- ture, current velocity, etc.) and competitive interactions that discourage colonization. Translocation of springsnails would have been unlikely; they can persist for several hours in moist, cool conditions but do not appear to survive by rapidly accli- mating to waters with chemistries different from their founder habitat (Sada, unpublished data). Unintentional fish transloca- tions likely would have been infrequent because of fish re- quirements for adequate water and tolerable temperatures and dissolved oxygen concentrations. Water quality differences also may have limited aquatic insect translocation. If uninten- tional translocation had been common, interbasin genetic dif- ferences between populations could not be easily detected and there would be little pattern that could be described by plu- vial-period interbasin connectivity. Evidence for intentional translocations is also weak. Most aquatic taxa of the Great Basin were either too small or too scarce to be used by hunter-gatherers; this suggests that movement of larger fish and invertebrates (e.g., California floater Anodonta californica) was more likely. The paucity of large Great Basin taxa also indicates that the efficacy of trans- location would have been enhanced if species had been brought in from surrounding basins. There is little evidence, however, that intentional movement was common; the distri- bution of endemic taxa within the Great Basin can usually be explained by models describing the fluvial connectivity of ba- sins during the Pleistocene (Echelle and Dowling, 1992; Smith, 1992; Hamlin, 1996), and the Great Basin fish fauna has few taxa in common with surrounding basins. Many aquatic species in the Great Basin instead are derived from ancestral taxa occupying surrounding basins (Smith, 1978; Minckley et al., 1991). Unassisted entrance of aquatic taxa into the Great Basin is shown by the fossil record. Pliocene and Miocene fish and mollusk fossils indicate past connectivity between the Great Basin and its surrounding basins (e.g., Miller and Smith, 1981; Smith, 1981; Taylor and Smith, 1981; Taylor, 1985). Firby et al. (1997) recorded bones of Oncorhynchus (trout) and Catostomus (subgenus Pantosteus) (sucker) from 695,000-725,000 year-old sediments in Owens Lake, Califor- nia. Neither of these taxa inhabit the Owens Basin today, but closely related taxa are found in the Lahontan Basin to the north and in coastal drainages of southern California (Smith, 1966). Past occurrence and extirpation of these fishes before the arrival of humans indicates that these fish moved into the Great Basin during pluvial periods and became extinct with- out the influences of hunter-gatherer activity. European immigrants had an immediate effect on Great Ba- sin aquatic biogeography when they first settled the region during the nineteenth century. Expansion of the livestock in- dustry, which began in the middle 1800s, resulted in some of the first disturbances of many aquatic habitats and riparian communities (Davis, 1977; Mack, 1981). Introductions of nonnative fishes also began in this period, when interest in sportfishing accompanied the expanding population that pur- sued the silver and gold mining bonanzas then underway (Hulse, 1991; Dill and Cordone, 1997). The resulting effects on biogeography were greater than those caused by hunter- gatherers, because the settlers used technologies that caused large-scale habitat modifications (e.g., pipes, dredging, dams, impoundments) and they introduced species from throughout the world. As a result, many existing wetlands probably bear little resemblance to their condition prior to settlement by the European Americans and immigrants. Water flow diversion and introduced species have degraded most large wetlands, whereas smaller habitats?streams and springs?have been perturbed by water diversion and riparian zone trampling that accompanies pastoral activities. These changes have reduced the distribution and abundance of most endemic aquatic taxa and the aquatic biodiversity within the Great Basin. Historical declines in Great Basin endemic aquatic taxa can be attributed to a comparatively small number of factors (e.g., water diversion, livestock use, and introduction of nonnative species). The scarcity of proposals for ecologically beneficial changes in future land use, the increasing demands for water, and the changes in abundance and distribution of most en- demic species during the last 100 years all indicate that these factors can be expected to cause additional extinctions (pri- marily in narrowly distributed taxa) and changes in biogeog- raphy. Continued declines in abundance and distribution may begin to affect widespread taxa as well, which will limit our ability to interpret pluvial climates and interbasin connectiv- ity; and a natural laboratory will be eliminated wherein the ecology, diversity, and speciation of the aquatic fauna within the Great Basin?almost 20 percent of the area of the United States?can be examined. The small size of most Great Basin wetlands and the small number of activities causing status declines indicate that stop- ping the losses of aquatic taxa does not require substantial funding or a large commitment of natural resources. Conserv- ing these habitats requires innovative programs to allow rea- sonable human uses of wetland areas while protecting Great Basin aquatic ecosystems from further degradation. To meet this challenge, land and resource management strategies must be developed that will allow use of resources while conserving Great Basin biota. This can be accomplished only by increasing ecological knowledge of endemic species and by using this in- formation to create innovative techniques that allow wetland use by humans while providing for conservation of fauna and flora unique to the Great Basin. Experimental work that exam- ines interactions between native and exotic macroinvertebrates 284 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES is also needed to determine mechanistic causes for interactions between native aquatics and normative invertebrates so that ef- fective conservation strategies can be developed. Acknowledgments J.E. Williams and P. Holden generously assisted in deter- mining status and causes for decline of southeastern Oregon and Utah fishes, respectively. J. Stein, R. Haskins, and G. Scoppettone assisted in accumulating information about Ne- vada fishes, and S. Vrilakas and J. Moorefield identified plants that are endemic to Great Basin wetlands. The manu- script was greatly improved by comments from R. Hershler, J. Deacon, R. Heyer, K. Halford, P. Pister, and two anonymous reviewers. Appendix Primary Habitat and Status of Known Distinctive Aquatic Biota in Great Basin Wetlands Distinct biota are either described taxa or potential, undescribed taxa whose morphological or genetic distinctiveness has been recognized in published or unpublished reports. Potential taxa are identified by their geographic location in quotes after the taxon name. In parentheses are numbers of translocated populations (if known). The year distinctiveness was first recognized (Dis- tinct); the calendar year or decades of first population loss (First Loss), major decline, and extinction; and the anthropogenic activities causing status change (Threats) are shown for each form. A dash (-) indicates no population loss, major decline, extinction, or threats are known. Threats: 1 = livestock grazing, 2 = nonnative species introduction, 3 = sport fisheries management, 4 = water diversion, 5 = groundwater pumping, 6 = urbanization, 7 = hybridization, 8 = timber harvest, 9 = pollution, 10 = recreation. A question mark (?) denotes in- stances in which population translocation, status change, or threats are proba- ble but existing information is insufficent for an accurate assessment. Distinct biota Amphibians Bufo exsul Bufo nelsoni Fishes Catostomus ardens Catostomus clarki intermedius Catostomus clarki 'Meadow Valley' Catostomus fumeiventris (4) Catostomus occidentalis lacusanserinus Catostomus platyrhynchus lahontan (?) Catostomus platyrhynchus platyrhynchus Catostomus tahoensis (?) Catostomus warnerensis (1) Catostomus 'Wall Canyon' Chasmistes cujus Chasmistes liorus liorus Chasmistes liorus mictus (3) Cottus bairdi 'Mahleur' Cottus echinatus Cottus extensus (1) Cottus pitensis Crenichthys baileyi albivallis Crenichthys baileyi baileyi (1) Crenichthys baileyi grandis (1) Crenichthys baileyi moapae Crenichthys baileyi thermophilus Crenichthys nevadae (4) Cyprinodon diabolis (2) Cyprinodon nevadensis amargosae (3) Cyprinodon nevadensis calidae Cyprinodon nevadensis mionectes (2) Cyprinodon nevadensis nevadensis Habitat Spring Spring Lotic/Lentic Lotic Lotic Lotic Lotic/Lentic Lotic Lotic Lotic/Lentic Lotic/Lentic Lotic Lotic/Lentic Lotic/Lentic Lotic/Lentic Lotic Lentic Lentic Lotic Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Distinct 1942 1893 1881 1942 1948 1948 1948 1903 1874 1878 1908 1948 1883 1981 1981 1948 1963 1963 1963 1981 1981 1981 1981 1981 1932 1930 1948 1948 1948 1948 First loss - 1980 1930 1960 1950 ? 1940 - 1900 1900 1940 ? - 1900 - 1960 1890 - 1940 1970 - 1960 1960 1960 - 1970 - 1950 1960 - Major decline Extinct - 1980 - 1960 - 1940 1940 - - - 1940 ? _ 1930 1930 1930 1950 1940 1900 1930 - 1890 - 1960 1960 - - 1980 1970 - 1950 1970 1970 - Threats 1 1,2,4 4 2,3,4 2,4 1,2,3,4 1,2,4,8 2,3,4 3,4 1,2,3,4 1,2,3,4,8 2 2,3,4,6 2,3,4 2,3,4 1,2,4,8 2,4 - 1,2,8 1,2,3 1,2,4 1,2,3 2,4,5 2,3,4 2,3,4 5 2,4 4,7 2,4,5 - References Szewczak, 1997 Altig, 1981; Maciolek, 1983; Hoff, 1993; Heinrich, 1996 Kershner, 1995 Courtenay et al., 1985 Hubbs and Miller, 1960; Stein, 1997 Moyle et al., 1995; USFWS, 1998a J.E. Williams, pers. comm., 1996 Decker, 1989 Kershner, 1995 La Rivers, 1962 Andraesen, 1975; Kittredge, 1987; Williams et al., 1990 Hubbs and Miller, 1948a; Heinrich, 1993 Scoppettone and Vinyard, 1991; USFWS, 1992 Sharp, 1905; Miller and Smith, 1981 Sharp, 1905; Miller and Smith, 1981; Kershner, 1995 Bond, 1974; C.E. Bond, pers. comm. with J.E. Williams, 1996 Heckman et al., 1981; Miller et al., 1989 Sharp, 1905; Holden et al., 1996 Oregon Department of Wildlife files, 1992 Williams and Wilde, 1981; Courtenay et al., 1985 Williams and Wilde, 1981; Courtenay et al., 1985; USFWS 1998b Williams and Wilde, 1981; Courtenay et al., 1985; USFWS 1998b Williams and Wilde, 1981; Courtenay et al., 1985 Williams and Wilde, 1981; Courtenay et al., 1985; Heinrich, 1993 Williams and Williams, 1981; USFWS, 1997a Deacon, 1979; Heinrich, 1993 Moyle et al., 1989; Sada et al., 1997 Miller et al., 1989; Moyle et al., 1995 Williams and Sada, 1985; Scoppettone et al., 1995 Deacon and Deacon, 1979; Sada et al., 1997 NUMBER 33 285 APPENDIX.?Continued. Distinct biota Fishes (continued) Cyprinodon nevadensis pectoralis Cyprinodon nevadensis shoshone Cyprinodon radiosus (4) Cyprinodon salinus milleri Cyprinodon salinus salinus Empetrichthys latos concavus Empetrichthys latos latos (3) Empetrichthys latos pahrump Empetrichthys merriami Eremichthys acros Gila alvordensis Gila bicolor euchila Gila bicolor eurysoma Gila bicolor isolata Gila bicolor mohavensis Gila bicolor newarkensis Gila bicolor obesa Gila bicolor oregonensis Gila bicolor pectinifer Gila bicolor snyderi (3) Gila bicolor vaccaceps Gila bicolor 'Big Smokey Valley' Gila bicolor 'Blue Eagle Spring' Gila bicolor 'Bull Creek' Gila bicolor 'Butterfield Spring' Gila bicolor 'Catlow Valley' Gila bicolor 'Diamond Valley' Gila bicolor 'Dixie Valley' (2) Gila bicolor 'Eagle Lake' Gila bicolor 'Fish Lake Valley' Gila bicolor 'Green Springs' Gila bicolor 'High Rock' Gila bicolor 'Hot Creek Valley' Gila bicolor 'Hutton Spring' Gila bicolor 'Kate Springs' Gila bicolor 'Little Fish Lake Valley' Gila bicolor 'Pleasant Valley' Gila bicolor 'Railroad Valley' Gila bicolor 'Summer Valley' Gila boraxobius (1) Gila robusta jordani (1) lotichthys phlegethontis (1) Lampetra tridentata 'Goose Lake' Lepidomeda albivallis Lepidomeda altivelis Lepidomeda mollispinis pratensis (1) Oncorhynchus clarki henshawi (33) Oncorhynchus clarki seleniris (4) Oncorhynchus clarki Utah (23) Oncorhynchus clarki 'Alvord' Oncorhynchus clarki 'Humboldt' (?) Prosopium abyss icola Prosopium gemmifer (1) Prosopium spilonotus Relictus solitarius (4) Rhinichthys osculus lariversi Rhinichthys osculus lethoporus Rhinichthys osculus nevadensis Habitat Spring Spring Spring/Lotic Spring Spring Spring Spring Spring Spring Spring Spring/Lotic Spring Spring Spring Lotic Spring Lotic Spring/Lotic Lentic Lotic Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring/Lotic Lentic Spring/Lotic Spring Spring/Lotic Lotic Lotic Lotic/Lentic Lotic Lotic/Lentic Lentic Lentic Lentic Spring Spring Spring Spring/Lotic Distinct 1948 1948 1948 1972 1943 1948 1948 1948 1893 1948 1972 1948 1948 1948 1938 1948 1856 1977 1917 1948 1948 1948 1948 1948 1948 1948 1948 1948 1943 1948 1948 1948 1948 1948 1948 1948 1948 1948 1948 1980 1950 1874 1948 1960 1960 1960 1845 1933 1874 1934 1978 1919 1919 1919 1948 1972 1948 1893 First loss 1970 1940 1930 - - 1940 1970 1940 1930 - 1960 - - - 1930 1950 1890 - - 1940 - 1960 1970 - - - 9 - - 1960 - 1980 1970 - - - 1970 - 1930 - 1960 1900 - 1960 1930 1950 1890 1920 1870 1890 1890 - - - 1930 - - 1971 Major decline Extinct - 1960 1940 - - 1950 1970 1950 1930 - - 1970 1890 1970 1940 - 1940 1890 - 1940 - 1980 1970 - - 1890 - 1910 1980 - 1980 - - - - 1980 - 1960 - 1940 1900 1940 1970 1930 1950 1930 1920 1910 1890 1930 - - - 1930 - 1970 1970 - 1960 - - - 1950 - 1950 1950 - - - - - - - - - - - - - - - - - - - - - - 1980 - - - - - - - - - - - - 1940 - - - - 1940 - - - - - - - - Threats 2 2,4 2,4 - - 5 2,5 5 2,5 1,4 1,2,4 1,2,3,4 1 1,2,4 2,4,7 1 1,2,3,4 1,4 3,4 2,3,4,7 1,4 1,2,5 4 - - 1,4 ? 2 2,3 2,3 - 2 1 - - - 2,3 1,2,3 1,2,3,4,7 5 2,4 2,3,4,6 2,4 2,4,9 2,3,4 2,4 1,2,3,4,7 1,2,7 1,2,3,4 1,2,3,4,7 1,2,3,4,7 - - - 2,3 1 2,4 2,4,5 References USFWS, 1976; Soltz and Naiman, 1978; Scoppettone et al., 1995 Taylor et al., 1988; Moyle et al., 1995 Miller, 1948; 1961; Miller and Pister, 1971; USFWS, 1998a LaBounty and Deacon, 1972; Sada and Deacon, 1995 Miller, 1943; Sada and Deacon, 1995 Miller, 1948; Sokol, 1954; Miller, 1961 Miller, 1948; Minckley and Deacon, 1968 Miller, 1948; Sokol, 1954; Miller, 1961 Miller, 1948; Sokol, 1954; Miller, 1961; Miller, 1969 Hubbs and Miller, 1948b; Vinyard, 1996; USFWS, 1997b Williams and Bond, 1983 Hubbs et al., 1974; Baugh et al., 1986 Williams et al., 1980; Williams and Bond, 1981 Hubbs et al., 1974; Vinyard, 1984; Heinrich, 1993 Hubbs and Miller, 1943; USFWS, 1984 Hubbs etal., 1974; Hardy, 1980; Haskins, 1996 La Rivers, 1962; Snyder, 1917a Bills, 1977; J.E. Williams, pers. comm., 1997 La Rivers, 1962 Snyder, 1917b; Miller, 1973; USFWS, 1998a Bills and Bond, 1980; Moyle et al., 1995 Deacon and Pedretti, 1984; Pedretti et al., 1987 J.R. Stein, pers. comm., 1997 D.W. Sada, field notes, 1995; J.R. Stein, pers. comm., 1997 J.R. Stein, pers. comm., 1997 Bills, 1977; J.E. Williams, pers. comm., 1997 Hubbs etal., 1974 Rissler et al., 1991; Vinyard, 1994 Kimsey, 1954; Moyle et al., 1991 D.W. Sada, field notes, 1993 D.W. Sada, field notes, 1995; J.R. Stein, pers. comm., 1997 Moyle, 1993 J.R. Stein, pers. comm., 1997 Bills, 1977; J.E. Williams, pers. comm., 1997 J.R. Stein, pers. comm., 1997 J.R. Stein, pers. comm., 1997 Nevada Division of Wildlife files, 1995 Williams and Williams, 1981 Bills, 1977; J.E. Williams, pers. comm., 1997 Williams and Bond, 1980; 1983; USFWS, 1987 Courtenay et al., 1985; USFWS, 1986, 1998b Kershner, 1995; USFWS, 1995a J.E. Williams, pers. comm., 1997 Courtenay et al., 1985; Scoppettone et al., 1992; USFWS, 1994 Hubbs and Miller, 1960; Miller et al., 1989 Hubbs and Miller, 1960; USFWS, 1993 Sumner, 1940; Behnke, 1992; USFWS, 1995b USFWS, 1985; Behnke, 1992 Kershner, 1995; Duff, 1996; Holden et al., 1996 Williams and Bond, 1983; Miller et al., 1989; Behnke, 1992 Behnke, 1992; USFWS, 1995b Sigler and Sigler, 1987 Sigler and Sigler, 1987 Sigler and Sigler, 1987 Hubbs et al., 1974; Stein and Salisbury, 1994; Haskins, 1995 Pedretti et al., 1987 Hubbs etal., 1974; Vinyard, 1984 Williams and Sada, 1985; USFWS, 1990; Scoppettone et al., 1995 286 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES APPENDIX.?Continued. Distinct biota Fishes (continued) Rhinichthys osculus oligoporus Rhinichthys osculus reliquus Rhinichthys osculus robustus (?) Rhinichthys osculus velifer Rhinichthys osculus 'Amargosa' Rhinichthys osculus 'Benton Valley' Rhinichthys osculus 'Diamond Valley' Rhinichthys osculus 'Foskett Spring' (1) Rhinichthys osculus 'Little Lake' Rhinichthys osculus 'Long Valley' Rhinichthys osculus 'Meadow Valley' Rhinichthys osculus 'Monitor Valley' Rhinichthys osculus 'Oasis Valley' Rhinichthys osculus 'Pahranagat Valley' Rhinichthys osculus 'Upper White River' Richardsonius egregius (?) Mollusks Assiminea infima Fluminicola dalli Fluminicola modoci Fluminicola turbiniformis Pyrgulopsis aardahli Pyrgulopsis aloba Pyrgulopsis amargosae Pyrgulopsis anatina Pyrgulopsis anguina Pyrgulopsis augustae Pyrgulopsis aurata Pyrgulopsis basiglans Pyrgulopsis bifurcata Pyrgulopsis breviloba Pyrgulopsis bryantwalkeri Pyrgulopsis cariniata Pyrgulopsis chamberlini Pyrgulopsis cruciglans Pyrgulopsis crystalis Pyrgulopsis dixiensis Pyrgulopsis eremica Pyrgulopsis erythropoma Pyrgulopsis fairbanksensis Pyrgulopsis fusca Pyrgulopsis gibba Pyrgulopsis gracilis Pyrgulopsis hamlinensis Pyrgulopsis hendersoni Pyrgulopsis hovinghi Pyrgulopsis hubbsi Pyrgulopsis humboldtensis Pyrgulopsis imperalis Pyrgulopsis inopinata Pyrgulopsis isolata Pyrgulopsis landeyi Pyrgulopsis lata Pyrgulopsis lentiglans Habitat Spring Spring Lotic Spring Lotic Spring Spring/Lotic Spring Spring Spring Lotic Spring Spring Spring Spring Lotic/Lentic Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Distinct 1948 1948 1903 1893 1995 1995 1948 1948 1995 1995 1948 1948 1995 1984 1948 1859 1947 1884 1912 1865 1989 1998 1987 1998 1998 1998 1998 1998 1998 1998 1916 1998 1998 1998 1987 1998 1995 1893 1987 1998 1995 1998 1998 1933 1998 1998 1998 1998 1998 1987 1998 1998 1998 First loss 1970 1940 1900 1969 - 1940 ? - 1940 1960 - - - - 1970 1900 1930 - - - ? ? - ? ? ? 7 ? ? ? ? ? ? 7 - ? 9 - - ? ? 7 9 ? 7 ? ? ? 7 - 7 ? ? Major decline Extinct 1970 1950 1969 - 1969 - 1940 ? - 1940 1940 1960 - - 1980 1960 1970 - - - - - 7 _ ? - ? _ 7 7 7 _ 7 _ ? 7 7 _ ? 7 _ 7 _ 1970 ? - 1970 1970 7 _ ? 7 _ ? ? 7 _ ? _ ? ? 7 _ - ? 7 _ ? Threats 1,2,5 2 1,3,4 2,3,4 2 2,4 7 1 2 2 - 1 2,4 1,2 2,3,4,9 1,2,3,4 10 - - - 4 7 - ? 1 4 1 1 1 4 ? 4 10 ? 2,4 1 1 2,4 2,4 7 1,4 ? 1 7 7 4 ? 7 4 - ? 4 7 References Vinyard, 1984 Hubbs et al., 1974; Miller et al., 1989 La Rivers, 1962 Miller, 1984; Tuttle et al., 1990 Williams et al., 1982; Moyle et al., 1995; Sada et al., 1995 Sada, 1989; Sada etal., 1995; USFWS 1998a Hubbs etal., 1974 Andreasen, 1975; Williams et al., 1990 Sada, 1989; Sada etal., 1995 Sada, 1989; Sada et al., 1995; USFWS 1998a Hubbs and Miller, 1960; Stein, 1997 Heinrich, 1991; D.W. Sada, field notes, 1996 Heinrich, 1993; Sada etal., 1995 Tuttle etal., 1990 Courtenay et al., 1985; Scoppettone et al., 1992 La Rivers, 1962 Pistrang and Kunkel, 1958; Hershler 1987; Sada, 2001 Hershler and Frest, 1996 Hershler and Frest, 1996 Hershler and Frest, 1996 Hershler, 1989 Hershler, 1998 Hershler, 1989 Hershler, 1998 Hershler, 1998 Hershler, 1998 D.W. Sada, field notes, 1996; Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1998 D.W. Sada, field notes, 1991; Hershler, 1994 Hershler, 1998 Hershler, 1998 Hershler, 1998 Sada, 1985; Hershler and Sada, 1987; D.W. Sada, field notes, 1997 D.W. Sada, field notes, 1991; Hershler, 1998 Hershler, 1995 Sada, 1985; Hershler and Sada, 1987 Hershler and Sada, 1987 Hershler, 1998 Hershler, 1995; Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1994; Hershler, 1998 Hershler, 1998 Courtenay et al., 1985; Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1998 Sada, 1985; Hershler and Sada, 1987 Hershler, 1998 Hershler, 1998 Hershler, 1998 NUMBER 33 287 APPENDIX.?Continued. Distinct biota Mollusks (continued) Pyrgulopsis leporina Pyrgulopsis limaria Pyrgulopsis lockensis Pyrgulopsis longae Pyrgulopsis longiglans Pyrgulopsis marcida Pyrgulopsis merriami Pyrgulopsis micrococcus Pyrgulopsis militaris Pyrgulopsis millenaria Pyrgulopsis montana Pyrgulopsis nanus Pyrgulopsis neritella Pyrgulopsis nevadensis Pyrgulopsis nonaria Pyrgulopsis notidicola Pyrgulopsis orbiculata Pyrgulopsis owensensis Pyrgulopsis papillata Pyrgulopsis peculiaris Pyrgulopsis pellita Pyrgulopsis perturbata Pyrgulopsis pictillis Pyrgulopsis pilcata Pyrgulopsis pilsbryana Pyrgulopsis pisteri Pyrgulopsis planulata Pyrgulopsis ruinosa Pyrgulopsis sathos Pyrgulopsis saxatalis Pyrgulopsis serrata Pyrgulopsis steralis Pyrgulopsis sulcata Pyrgulopsis transversa Pyrgulopsis umbilicata Pyrgulopsis variegata Pyrgulopsis villacampae Pyrgulopsis vinyardi Pyrgulopsis wongi Pyrgulopsis 'Longstreet Spring' Tryonia angulata Tryonia elata Tryonia ericae Tryonia margae Tryonia robusta Tryonia rowlandsi Tryonia salina Tryonia variegata Aquatic insects Ambrysus amargosus Ambrysus funebris Ambrysus relictus Belostoma saratogae Capnia lucustra Microcylloepus formicoideus Microcylloepus moapus fraxinus Stenelmis calida Stenelmis lariversi Fairy shrimp Artemia monica Habitat Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Lentic Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Spring Lentic Spring Spring Spring Spring Lentic Distinct 1998 1998 1998 1995 1998 1998 1998 1893 1998 1998 1998 1987 1998 1883 1998 1998 1998 1989 1998 1998 1998 1989 1998 1998 1998 1987 1998 1998 1998 1998 1998 1998 1998 1998 1998 1998 1998 1998 1989 1981 1987 1987 1987 1989 1989 1989 1989 1987 1953 1948 1994 1958 1965 1990 1949 1949 1992 1964 First loss ? 9 ? ? 9 ? 7 - 9 ? ? - ? 1890 ? 9 9 9 ? ? ? - 9 ? ? ? 9 1990 9 ? ? ? ? ? 9 ? 9 9 9 1970 ? 7 7 7 1930 9 - 9 1970 1930 - - 1960 1930 9 - - - Major decline Extinct 9 ? _ ? 9 9 _ ? ? - ? ? 9 _ - 9 _ 1890 1890 ? ? ? - ? ? ? _ ? _ 7 _ ? _ 7 _ ? ? 1990 1990 7 ._ ? ? ? - ? _ ? _ 7 _ ? 7 _ ? - 1970 1970 1970 1970 1970 7 _ - 7 _ - 1970 1970 1930 1970 - 1960 1960 1930 1960 1970 ? - Threats 1 1 4 4 1 1 ? 5 1 ? 9 2,4 ? 4 7 1 9 4 4 10 ? 4 4 ? 7 2,4,5 7 4 4 7 ? 1 7 4 1 1 4 ? 5 5 2,4 2,4 2,4 9 4 7 - 2,4 4,5 4 4 - 2,9 4 2,4 5 ? 4 References Hershler, 1998 D.W. Sada, field notes, 1996; Hershler, 1998 Hershler, 1998 Hershler, 1995 D.W. Sada, field notes, 1996; Hershler, 1998 Hershler, 1998 Hershler, 1998 Sada, 1985; Hershler and Sada, 1987; Hershler, 1989 D.W. Sada, field notes, 1996; Hershler, 1998 Hershler, 1998 Hershler, 1998 Sada, 1985; Hershler and Sada, 1987 Hershler, 1998 Galatetal., 1981; Hershler, 1994 Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1989; Hershler and Pratt, 1990 Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1989 Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler and Sada, 1987 Hershler, 1998 D.W. Sada, field notes 1991; Hershler, 1998 Courtenay et al., 1985; Hershler, 1998 Hershler, 1998 Hershler, 1998 D.W. Sada, field notes, 1992; Hershler, 1998 Hershler, 1998 Hershler, 1998 D.W. Sada, field notes, 1996; Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1998 Hershler, 1989; Hershler and Pratt, 1990 Taylor, 1980; Sada, 1985 Sada, 1985; Hershler and Sada, 1987 Sada, 1985; Hershler and Sada, 1987 Hershler and Sada, 1987 Hershler, 1989; Pratt and Hoff, 1992 Hershler, 1989; Pistrang and Kunkel, 1958 Hershler, 1989 Hershler, 1989; Sada and Deacon, 1995 Sada, 1985; Hershler and Sada, 1987; Hershler, 1989 La Rivers, 1953; Scoppettone et al., 1995 La Rivers, 1948; Pratt and Hoff, 1992; Polhemus and Polhemus, 1995 Polhemus and Polhumus, 1994 Pratt and Hoff, 1992; Polhemus and Polhemus, 1995 California Natural History Database (1996 unpublished data) Shepard, 1990; Pratt and Hoff, 1992 La Rivers, 1949 Chandler, 1949; Deacon and Deacon, 1979; Schmude, 1992 Schmude, 1992 Jellison et al., 1993; R.S. Jellison, pers. comm., 1997 288 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Literature Cited Altig, R. 1981. Status Report on the Amargosa Toad {Bufo nelsoni). Report to the U.S. Fish and Wildlife Service, Boise, Idaho. [Unpublished.] Andersen, M.E., and J.E. Deacon 1996. Status of Endemic Non-Salmonid Fishes in Eastern Nevada. Jour- nal of the Arizona-Nevada Academy of Science, 29:124-133. Andreasen, J.K.. 1975. Systematics and Status of the Family Catostomidae in Southern Or- egon. 76 pages. Doctoral dissertation, Oregon State University, Cor- vallis, Oregon. Bailey, V. 1898. Descriptions of Eleven New Species and Subspecies of Voles. Pro- ceedings of the Biological Society of Washington, 12:85-90. Baugh, T.M., J.W. Pedretti, and J.E. Deacon 1986. Status and Distribution of the Fish Creek Springs Tui Chub, Gila bi- color euchila. Great Basin Naturalist, 46:441-444. Beckwith, E.G. 1855. Report of Explorations for a Route for the Pacific Railroad of the Line of the Forty-First Parallel of North Latitude. In Reports of Ex- plorations and Surveys to Ascertain the Most Practicable and Eco- nomical Route for a Railroad from the Mississippi River to the Pacific Ocean, Made under the Direction of the Secretary of War, in 1853-54, According to Acts of Congress of March 3, 1853, May 31, 1854, and August 5, 1854, volume 2. 33rd Congress, 2nd Session, Senate Executive Document No. 78:1-132. Behnke, R.J. 1992. Native Trout of Western North America. American Fisheries Society Monograph, 6: 275 pages. Bills, F. 1977. Taxonomic Status of the Isolated Populations of Tui Chub Referred to Gila bicolor oregonensis (Snyder). 165 pages. Master's thesis, Oregon State University, Corvallis, Oregon. Bills, F., and C.E. Bond 1980. A New Subspecies of Tui Chub (Pisces: Cyprinidae) from Cowhead Lake, California. Copeia, 1980:320-322. Bond, C.E. 1974. Endangered Plants and Animals of Oregon, I: Fishes. 8 pages. Cor- vallis, Oregon: Agricultural Experiment Station, Oregon State Uni- versity. Brode, J.M., and R.B. Bury 1984. The Importance of Riparian Systems to Amphibians and Reptiles. In R.E. Warner and K.E. Hendrix, editors, Proceedings of the Confer- ence on California Riparian Systems, pages 30-36. Davis, Califor- nia: University of California. Brues, C.T. 1932. Further Studies on the Fauna of North American Hot Springs. Pro- ceedings of the American Academy of Arts and Science, 67:185- 303. Chandler, H.P. 1949. A New Species of Stenelmis from Nevada. Pan-Pacific Entomolo- gist, 25:133-136. Courtenay, W.R., Jr., J.E. Deacon, D.W. Sada, R.C. Allan, and G.L. Vinyard 1985. Comparative Status of Fishes along the Course of the Pluvial White River, Nevada. The Southwestern Naturalist, 30:503-524. Courtenay, W.R., and J.R. Stauffer, Jr., editors 1984. Distribution, Biology, and Management of Exotic Fishes. 430 pages. Baltimore, Maryland: Johns Hopkins University Press. Dansie, A.J. 1990. Prehistoric Carnivore Usage in the Wetland Habitats of Western Ne- vada. In J.C. Janetski and D.B. Madsen, editors, Wetland Adapta- tions in the Great Basin. Occasional Papers of the Museum of Peoples and Cultures, Brigham Young University, 1:159-172. Davidson, J.W. 1976. Report of the Results of an Expedition to Owens Lake and River, with the Topographical Features of the Country, Climate, Soil, Tim- ber, Water, and also Habitats, Arms, and Means of Subsistence, of the Indian Tribe Seen from the March; July and August, 1859. In P.J. Wilke and H.W. Lawton, editors, The Expedition of Capt. J.W. Davidson from Fort Tejon to Owens Valley in 1859. Publications in Archaeology, Ethnology, and History, 8:14-31. Socorro, New Mex- ico: Ballena Press. Davis, G.A. 1977. Management Alternatives for the Riparian Habitat in the Southwest. In R.R. Johnson and D.A. Jones, technical coordinators, Importance, Perspective and Management of Riparian Habitat: A Symposium. United States Department of Agriculture, U.S. Forest Service, Gen- eral Technical Report RM-43, pages 59-76. Fort Collins, Colorado: Rocky Mountain Range Experiment Station. Deacon, J.E. 1979. Endangered and Threatened Fishes of the West. Great Basin Natu- ralist Memoirs, 3:41-64. Deacon, J.E., and M.S. Deacon 1979. Research on Endangered Fishes in the National Parks with Special Emphasis on the Devil's Hole Pupfish. In R.M. Linn, editor, Pro- ceedings of the First Conference on Scientific Research in the Na- tional Parks. United States Department of the Interior, National Park Service Transactions and Proceedings Series, 5:9?19. Deacon, J.E., and J.W. Pedretti 1984. Population Status and Distribution of Gila Complex in Big Smoky Valley, Nevada. Report to the Nevada Department of Wildlife, Reno, Nevada. [Unpublished.] Deacon, J.E., and J.E. Williams 1984. Annotated List of the Fishes of Nevada. Proceedings of the Biologi- cal Society of Washington, 97:103-118. Decker, L.M. 1989. Coexistence of Two Species of Sucker, Catostomus, in Sagehen Creek, California, and Notes on Their Status in the Western Lahon- tan Basin. Great Basin Naturalist, 49:540-551. Dill, W.A., and A.J. Cordone 1997. History and Status of Introduced Fishes in California, 1871-1996. California Department of Fish and Game, Fish Bulletin, 178:1-414. Drews, M.P. 1990. The Dietary Role of Freshwater Shellfish from Stillwater Marsh. In J.C. Janetski and D.B. Madsen, editors, Wetland Adaptations in the Great Basin. Occasional Papers of the Museum of Peoples and Cul- tures, Brigham Young University, 1:63?74. Dudley, W.W., Jr., and J.D. Larson 1976. Effect of Irrigation Pumping on Desert Pupfish Habitats in Ash Meadows, Nye County, Nevada. United States Geological Survey Professional Paper, 927:1-52. Duff, D. 1996. Bonneville Cutthroat Trout, Oncorhynchus clarki Utah. In D.A. Duff, technical editor, Conservation Assessment for Inland Cut- throat Trout: Distribution, Status and Habitat Management Implica- tions, pages 35-73. Ogden, Utah: United States Department of Agriculture, U.S. Forest Service, Intermountain Region. Echelle, T.A., and T.E. Dowling 1992. Mitochondrial DNA Variation and Evolution of the Death Valley Pupfishes (Cyprinodon, Cyprinodontidae). Evolution, 46:193-206. Firby, J.R., S.E. Sharpe, J.F. Whelan, G.R. Smith, and W.G. Spaulding 1997. Paleobiotic and Isotopic Analysis of Mollusks, Fish, and Plants from Core OL-92: Indicators for an Open or Closed Lake System. In G.S. Smith and J.L. Bischoff, editors, An 800,000-Year Paleocli- matic Record from Core OL-92, Owens Lake, Southeast California. NUMBER 33 289 Geological Society of America Special Paper, 317:121-125. Fleishner, T.L. 1994. Ecological Cost of Livestock Grazing in Western North America. Conservation Biology, 8:629-644. Fowler, C.S., and D.D. Fowler 1990. A History of Wetlands Anthropology in the Great Basin. In J.C. Jan- etski and D.B. Madsen, editors, Wetland Adaptations in the Great Basin. Occasional Papers of the Museum of Peoples and Cultures, Brigham Young University, 1:5-16. Fremont, J.C. 1845. Report of the Exploring Expeditions to the Rocky Mountains in the Year 1842, and to Oregon and North California in the Years 1843-'44. 583 pages. Washington, D.C.: Printed by Order of the U.S. Senate. Galat, D.L., E.L. Lider, and S.R. Robertson 1981. Limnology of a Large, Deep, North American Terminal Lake, Pyra- mid Lake, Nevada, U.S.A. Hydrobiologia, 82:281-317. Gilbert, C.H. 1893. Report on the Fishes of the Death Valley Expedition, Collection in Southern California and Nevada in 1891, with Descriptions of New Species. North American Fauna, 7:229-234. Grayson, D.K. 1991. Late Pleistocene Extinctions in North America: Taxonomy, Chro- nology, and Explanations. Journal of World Prehistory, 5:193-232. Greenspan, R.L. 1990. Prehistoric Fishing in the Northern Great Basin. In J.C. Janetski and D.B. Madsen, editors, Wetland Adaptations in the Great Basin. Oc- casional Papers of the Museum of Peoples and Cultures, Brigham Young University, 1:207-232. Halford, F.K. 1998. Archaeology and Environment on the Dry Lakes Plateau, Bodie Hills, California: Hunter-Gatherer Coping Strategies for Holocene Environmental Variability. 188 pages. Master's thesis, University of Nevada, Reno, Nevada. Hall, E.R. 1946. Mammals of Nevada. 710 pages. Berkeley, California: University of California Press. Hamlin, R.A. 1996. Conservation Genetics of Remnant Springsnail, Pyrgulopsis wongi, Populations in Desert Valleys of California and Nevada. 54 pages. Master's thesis, University of Nevada, Reno, Nevada. Hardy, T. 1980. The Inter-Basin Area Report. Proceedings of the Desert Fishes Council, 11:5-28. Haskins, R.L. 1995. Status and Distribution of Selected Populations of Relict Dace (Relictus solitarius) in White Pine County. 12 pages. Elko, Nevada: Nevada Division of Wildlife. [Unpublished report.] 1996. Status and Distribution of Newark Valley Tui Chub (Gila bicolor newarkensis). 20 pages. Elko, Nevada: Nevada Division of Wildlife. [Unpublished report.] Heckmann, R.A., C.W. Thompson, and D.A. White 1981. Fishes of Utah Lake. Great Basin Naturalist Memoirs, 5:107-127. Heinrich, J. 1991. Native Nongame Fish Program Progress Report, January 1, 1991 through December 31, 1991. 23 pages. Las Vegas, Nevada: Nevada Division of Wildlife. [Unpublished report.] 1993. Native Nongame Fish Program Progress Report, January 1, 1992 through December 31, 1992. 23 pages. Las Vegas, Nevada: Nevada Division of Wildlife. [Unpublished report.] 1996. August 1995 Oasis Valley Surveys for the Amargosa Toad, Bufo nelsoni. 13 pages + 4 appendices. Las Vegas, Nevada: Nevada Divi- sion of Wildlife. [Unpublished report.] Hershler, R. 1987. Redescription of Assiminea infima Berry, 1947, from Death Valley, California. Veliger, 29:274-288. 1989. Springsnails (Gastropoda: Hydrobiidae) of Owens and Amargosa River (Exclusive of Ash Meadows) Drainages, Death Valley Sys- tem, California-Nevada. Proceedings of the Biological Society of Washington, 102:176-248. 1994. A Review of the North American Freshwater Snail Genus Pyrgulop- sis (Hydrobiidae). Smithsonian Contributions to Zoology, 554: 1-115. 1995. New Freshwater Snails of the Genus Pyrgulopsis (Rissooidea: Hy- drobiidae) from California. Veliger, 38:343-373. 1998. A Systematic Review of the Hydrobiid Snails (Gastropoda: Ris- sooidea) of the Great Basin, Western United States, Part 1: Genus Pyrgulopsis. Veliger, 41:1-132. Hershler, R., and T.J. Frest 1996. A Review of the North American Freshwater Snail Genus Flumini- cola (Hydrobiidae). Smithsonian Contributions to Zoology, 583: 41 pages. Hershler, R., and W.L. Pratt 1990. A New Pyrgulopsis (Gastropoda: Hydrobiidae) from Southeastern California, with a Model for Historical Development of the Death Valley Hydrographic System. Proceedings of the Biological Society of Washington, 103:279-299. Hershler, R., and D.W. Sada 1987. Springsnails (Gastropoda: Hydrobiidae) of Ash Meadows, Amar- gosa Basin, California-Nevada. Proceedings of the Biological Soci- ety of Washington, 100:776-843. Hoff, K.S. 1993. Status of the Amargosa Toad, 1992. Report to the U.S. Fish and Wildlife Service, Reno, Nevada. [Unpublished.] Holden, P.B., S.J. Zucker, and P.D. Abate 1996. Assessment of the Effects of Fish Stocking in the State of Utah: Past, Present, and Future. Logan, Utah: Report to the Utah Division of Wildlife Resources, BIO/WEST PR-565-1. [Unpublished.] Hubbs, C.L., and R.R. Miller 1943. Mass Hybridization between Two Genera of Cyprinid Fshes in the Mojave Desert, California. Papers of the Michigan Academy of Sci- ence, Arts, and Letters, 28:343-378. 1948a. The Zoological Evidence: Correlation between Fish Distribution and Hydrographic History in the Desert Basins of Western United States. In The Great Basin, with Emphasis on Glacial and Postgla- cial Times. Bulletin of the University of Utah, 38:17-166. 1948b. Two New, Relict Genera of Cyprinid Fish from Nevada. Occasional Papers of the Museum of Zoology, University of Michigan, 507: 30 pages. 1960. The Spiny-Rayed Cyprinid Fishes (Plagopterini) of the Colorado River System. Miscellaneous Publications of the Museum of Zool- ogy, University of Michigan, 115: 39 pages. Hubbs, C.L., R.R. Miller, and L.C. Hubbs 1974. Hydrographic History and Relict Fishes of the North-Central Great Basin. Memoirs of the California Academy of Sciences, 7: 259 pages. Hulse, J.W. 1991. The Silver State, Nevada: History Reinterpreted. 371 pages. Reno, Nevada: University of Nevada Press. Janetski, J.C. 1990. Wetlands in Utah Valley Prehistory. In J.C. Janetski and D.B. Mad- sen, editors, Wetland Adaptations in the Great Basin, Occasional Papers of the Museum of Peoples and Cultures, Brigham Young University, 1:233-258. Janetski, J.C, and D.B. Madsen, editors 1990. Wetland Adaptations in the Great Basin. Occasional Papers of the Museum of Peoples and Cultures, Brigham Young University, 1: 283 pages. Jellison, R.S., J.M. Melack, and G.L. Dana 1993. A Modeling Analysis of Artemia Dynamics in Mono Lake. In Mono Basin Environmental Impact Report, Auxiliary Report Number 13, 97 pages. Report to the California State Water Resources Board, 290 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Sacramento, California. [Unpublished.] Johnson, L., and D.K. Padilla 1996. Geographic Spread of Exotic Species: Ecological Lessons and Op- portunities from the Invasion of the Zebra Mussel, Dreissena poly- morpha. Biological Conservation, 78:23-34. Kershner, J.L. 1995. Bonneville Cutthroat Trout. In M.K. Young, technical editor, Con- servation Assessment for Inland Cutthroat Trout. United States De- partment of Agriculture, U.S. Forest Service, General Technical Report, RM-GTR-256, pages 28-35. Kimsey, J.B. 1954. The Life History of the Tui Chub, Siphatales bicolor (Girard) from Eagle Lake, California. California Fish and Game, 40:395^10. Kittredge, W. 1987. Owning It All. 182 pages. St. Paul, Minnesota: Graywolf Press. Knack, M.C., and O.C. Stewart 1984. As Long as the River Shall Run: An Ethnohistory of Pyramid Lake Indian Reservation. 433 pages. Berkeley, California: University of California Press. La Bounty, J.F., and J.E. Deacon 1972. Cyprinodon milleri, a New Species of Pupfish from Death Valley, California. Copeia, 1972(4):769-780. La Rivers, I. 1948. A New Species of Ambrysus from Death Valley, with Notes on the Genus in the United States (Hemiptera: Naucoridae). Bulletin of the Southern California Academy of Science, 47:103-110. 1949. A New Species of MicrocyIloepus from Nevada (Coleoptera: Dry- opidae). Entomological News, 60:205-209. 1953. New Gelastocorid and Naucorid Records and Miscellaneous Notes, with a Description of a New Species, Ambrysus amargosus (Hemi- ptera: Naucoridae). Wasmann Journal of Biology, 11:83-96. 1962. Fishes and Fisheries of Nevada. 782 pages. Carson City, Nevada: Nevada State Fish and Game Commission. Maciolek, J.A. 1983. Status Report: Amargosa Toad. 21 pages. Report to the U.S. Fish and Wildlife Service, Reno, Nevada. [Unpublished.] Mack, R.N. 1981. Invasion of Bromus tectorum L. into Western North America: An Ecological Chronicle. Agroecosystems, 7:145-165. Madsen, D.B. 1989. Exploring the Fremont. University of Utah Occasional Publication, 8: 70 pages. Martin, P.S. 1990. Who or What Destroyed our Mammoths? In D. Agenbraod, J.I. Mead, and L.W. Nelson, editors, Meagfauna and Man: Discovery of America's Heartland, pages 109-117. The Mammoth Site of Hot Springs, South Dakota, Inc., Scientific Papers, 1. Mehringer, P.J., Jr., and C.N. Warren 1976. Marsh, Dune and Archaeological Chronology, Ash Meadows, Am- argosa Desert, Nevada. In R. Elson, editor, Holocene Environmental Change in the Great Basin, pages 120-150. Nevada Archaeological Survey Research Paper, 6. Merriam, C.H. 1893. The Death Valley Expedition: A Biological Survey of Parts of Cali- fornia, Nevada, Arizona, and Utah, Part II. North American Fauna, 7:1-394. Miller, R.R. 1943. Cyprinodon salinus, a New Species of Fish from Death Valley, Cali- fornia. Copeia, 1943(2):69-78. 1948. The Cyprinodont Fishes of the Death Valley System of Eastern Cal- ifornia and Southwestern Nevada. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 68:1-155. 1961. Man and the Changing Fish Fauna of the American Southwest. Pa- pers of the Michigan Academy of Science, Arts, and Letters, 46: 365^04. 1969. Conservation of Fishes of the Death Valley System in California and Nevada. Transactions of the California-Nevada Section, The Wild- life Society, 1969:107-122. 1973. Two New Fishes, Gila bicolor snyderi and Catostomus fumeiventris, from the Owens River Basin, California. Occasional Papers of the Museum of Zoology, University of Michigan, 667:1-19. 1984. Rhinichthys deaconi, a New Species of Dace (Pisces: Cyprinidae) from Southern Nevada. Occasional Papers of the Museum of Zool- ogy, University of Michigan, 707:1-21. Miller, R.R., and E.P. Pister 1971. Management of the Owens Pupfish, Cyprinodon radiosus, in Mono County, California. Transactions of the American Fisheries Society, 100:502-509. Miller, R.R., and G.R. Smith 1981. Distribution and Evolution of Chasmistes (Pisces: Catostomidae) in Western North America. Occasional Papers of the Museum of Zool- ogy, University of Michigan, 696:1-46. Miller, R.R., J.D. Williams, and J.E. Williams 1989. Extinctions of Northern American Fishes during the Past Century. Fisheries, 14(6):22-38. Minckley, W.L., and J.E. Deacon 1968. Southwestern Fishes and the 'Enigma' of Endangered Species Man- agement. Science, 159:1424-1432. Minckley, W.L., and M.E. Douglas 1991. Discovery and Extinction of Western Fishes: A Blink of the Eye in Geologic Time. In W.L. Minckley and J.E. Deacon, editors, Battle against Extinction. Native Fish Management in the American West, pages 7-18. Tucson, Arizona: University of Arizona Press. Minckley, W.L., D.A. Hendrickson, and C.E. Bond 1991. Geography of Western North American Freshwater Fishes: Descrip- tion and Relationships to Intracontinental Tectonism. In C.H. Hocutt and E.O. Wiley, editors, The Zoogeography of North American Freshwater Fishes, pages 519-614. New York: John Wiley and Sons. Mooney, H.A., and J.A. Drake, editors 1986. Ecology of Biological Invasions of North America and Hawaii. 321 pages. New York: Springer-Verlag, Ecological Series, 58. Morgan, M.D., S.T. Threlkeld, and C.R. Goldman 1978. Impact of the Introduction of Kokanee Salmon {Oncorhynchus nerkd) on the Opossum Shrimp (Mysis relicta) on a Subalpine Lake. Journal of the Fisheries Research Board of Canada, 35:1572-1579. Moyle, P.B. 1984. Fish Introductions into North America: Patterns and Ecological Im- pact. In H.A. Mooney and J.A. Drake, editors, Ecology of Biological Invasions of North America and Hawaii, pages 27-43. New York: Springer-Verlag, Ecological Series, 58. 1993. Fish: An Enthusiast's Guide. 272 pages. Berkeley, California: Uni- versity of California Press. Moyle, P.B., H.W. Li, and B.A. Barton 1986. The Frankenstein Effect: Impact of Introduced Fishes on Native Fishes in North America. In R.H. Stroud, editor, Fish Culture in Fisheries Management, pages 415-426. Bethesda, Maryland: Amer- ican Fisheries Society. Moyle, P.B., and T. Light 1996. Fish Invasions in California: Do Abiotic Factors Determine Suc- cess? Ecology, 77:1666-1670. Moyle, P.B., T. Kennedy, D. Kuda, L. Martin, and G. Grant 1991. Fishes of Bly Tunnel, Lassen County, California. Great Basin Natu- ralist, 51:267-270. Moyle, P.B., J.E. Williams, and E.D. Wikramanayake 1989. Fish Species of Special Concern of California. 222 pages. Report to the California Department of Fish and Game, Inland Fisheries Divi- sion, Rancho Cordova, California. [Unpublished.] Moyle, P.B., R.M. Yoshiyama, J.E. Williams, and E.D. Wikramanayake 1995. Fish Species of Special Concern of California, Second edition, 272 NUMBER 33 291 pages. Final Report to the California Department of Fish and Game, Inland Fisheries Division, Rancho Cordova, California. [Unpub- lished.] Pedretti, J.W., T.M. Baugh, and J.E. Deacon 1987. Status of Native Fishes in Big Smoky Valley, Nevada. Journal of the Arizona-Nevada Academy of Science, 22:117-121. Pistrang, M.A., and F. Kunkel 1958. A Brief Geologic and Hydrologic Reconnaissance of the Furnace Creek Wash Area, Death Valley National Monument, California. United States Geological Survey Open-File Report, 58-12: 63 pages. Polhemus, D.A., and J.T. Polhemus 1995. A Preliminary Survey of the Aquatic and Semiaquatic Hemiptera Occupying the Springs of the Railroad Valley, White River Valley, and Amargosa River Drainage Systems, Nevada and California, with Special Reference to Thermal Endemics. 22 pages. [Unpub- lished manuscript.] Polhemus, J.T., and D.A. Polhemus 1994. A New Species of Ambrysus Stal from Ash Meadows, Nevada (Het- eroptera: Naucoridae). Journal of the New York Entomological Soci- ety, 102:261-265. Pratt, W.L., and K. Hoff 1992. Aquatic Invertebrate and Amphibian Survey of Saratoga Springs, Travertine Springs, and Amargosa River in Death Valley National Monument. 39 pages. Report to the U.S. National Park Service, Lake Mead National Recreation Area, Boulder City, Nevada. [Un- published.] Richards, R.C., C.R. Goldman, T.C. Frantz, and R. Wickwire 1975. Where Have all the Daphnia Gone: The Decline of a Major Cla- doceran in Lake Tahoe, California-Nevada. International Vereini- gungfiir Theoretische undAngewandte Limnologie, Verhandlungen, 19:835-842. Rinne, J.N. 1996. Short-Term Effects of Wildfire on Fishes and Aquatic Invertebrates in the Southwestern United States. North American Journal of Fish- eries Management, 16:653-658. Rinne, J.N., and PR. Turner 1991. Reclamation and Alteration of Management Techniques, and a Re- view of Methodology in Stream Renovation. In W.L. Minckley and J.E. Deacon, editors, Battle against Extinction: Native Fish Manage- ment in the American West, pages 219-244. Tucson, Arizona: Uni- versity of Arizona Press. Rissler, PH., S. Byers, G.G. Scoppettone, and D. Withers 1991. Status of Tui Chub and Other Fishes on Navy Lands in Dixie Valley. 16 pages. Report to the U.S. Naval Air Station, Fallon, Nevada. [Unpublished.] Sada, D.W. 1985. Collection of Hydrobiid Snails in Ash Meadows, Nevada, Novem- ber 6-11, 1985. 16 pages. United States Fish and Wildlife Service Report. Reno, Nevada. [Unpublished.] 1989. Status and Distribution of Speckled Dace (Rhinichthys osculus) in the Owens River System, Inyo and Mono Counties, California. 33 pages. Report to the California Department of Fish and Game, Ran- cho Cordova, California. [Unpublished.] 2001. Demography and Habitat Use of the Badwater Snail (Assiminea in- fima), with Observations on Its Conservation Status, Death Valley National Park, California, U.S.A. Hydrobiologia, 466:255-265. Sada, D.W., H.B. Britten, and P.B. Brussard 1995. Desert Aquatic Ecosystems and the Genetic and Morphological Di- versity of Death Valley System Speckled Dace. In J. Nielsen, editor, Evolution and the Aquatic Ecosystem: Defining Unique Units in Population Conservation. American Fisheries Society Symposium, 17:350-359. Sada, D.W., and J.E. Deacon 1995. Spatial and Temporal Variability of Pupfish (Genus Cyprinodon) Habitat and Populations at Salt Creek and Cottonball Marsh, Death Valley National Park, California. 76 pages. Report to the U.S. Na- tional Park Service, Death Valley National Park. [Unpublished.] Sada, D.W., and J.L. Nachlinger 1996. Spring Mountains Ecosystem: Vulnerability of Spring-fed Aquatic and Riparian Systems to Biodiversity Loss. 46 pages + 5 appendi- ces. Report to the U.S. Fish and Wildlife Service, Reno, Nevada. [Unpublished.] Sada, D.W., K. Pindal, D. Threloff, and J.E. Deacon 1997. Spatial and Temporal Variability of Pupfish (Genus Cyprinodon) Habitat and Populations at Saratoga Springs and the Lower Amar- gosa River, Death Valley National Park, California. 101 pages. Re- port to the U.S. National Park Service, Death Valley National Park. [Unpublished.] Schmitt, D.N., and N.D. Sharp 1990. Mammals in the Marsh: Zooarchaeological Analysis of Six Sites in the Stillwater National Wildlife Refuge, Western Nevada. In J.C. Ja- netski and D.B. Madsen, editors, Wetland Adaptations in the Great Basin. Occasional Papers of the Museum of Peoples and Cultures, Brigham Young University, 1:75-96. Schmude, K.L. 1992. A Revision of the Riffle Beetle Genus Stenelmis (Coleoptera: Elmi- dae) in North America. 388 pages. Unpublished doctoral disserta- tion. University of Wisconsin, Madison, Wisconsin. Scoppettone, G.G., and G.L. Vinyard 1991. Life History and Management of Four Endangered Lacustrine Suck- ers. In W.L. Minckley and J.E. Deacon, editors, Battle against Ex- tinction: Native Fish Management in the American West, pages 359-377. Tucson, Arizona: University of Arizona Press. Scoppettone, G.G., J.E. Harvey, S.P. Shea, and J. Heinrich 1992. Relative Abundance and Distribution of Fishes in the White River Valley, Nevada, with Special Emphasis on the White River Spine- dace {Lepidomeda albivallis). Reno, Nevada: Nevada Department of Wildlife and U.S. Fish and Wildlife Service, Seattle National Fisheries Research Center. [Unpublished report.] Scoppettone, G.G., P.H. Rissler, S. Byers, S. Shea, B. Nielsen, and J. Sjoberg 1995. Information on the Status and Ecology of Ash Meadows Fishes and Ambrysus. 111 pages. Reno, Nevada: National Biological Service. Scoppettone, G.G., P.H. Rissler, B. Nielsen, and J.E. Harvey 1998. The Status of Moapa coriaceae and Gila seminuda and Status Infor- mation in other Fishes of the Muddy River, Clark County, Nevada. The Southwestern Naturalist, 43:115-122. Sharp, J. 1905. Report of the State Fish and Game Commission for the Years 1903 and 1904. 203 pages. Salt Lake City, Utah: Star Publishing Com- pany. Shepard, W.D. 1990. Microcylloepus formicoideus (Coloeoptera: Elmidae), a New Riffle Beetle from Death Valley National Monument, California. Entomo- logical News, 101:147-153. 1992. Riffle Beetles (Coleoptera: Elmidae) of Death Valley National Mon- ument, California. Great Basin Naturalist, 52:378-381. Sigler, W.F., and J.W. Sigler 1987. Fishes of the Great Basin: A Natural History. 425 pages. Reno, Ne- vada: University of Nevada Press. Smith, G.R. 1966. Distribution and Evolution of North American Catostomid Fishes of the Subgenus Pantosteus, Genus Catostomus. Miscellaneous Publi- cations of Museum of Zoology, University of Michigan, 126:1-132. 1978. Biogeography of Intermountain Fishes. Great Basin Naturalist Memoirs, 2:17-42. 1981. Late Cenozoic Freshwater Fishes of North America. Annual Review of Ecology and Systematics, 12:163?193. 1992. Phylogeny and Biogeography of the Catostomidae, Freshwater Fishes of North America and Asia. In R.L. Mayden, editor, Sys- 292 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES tematics, Historical Ecology, and North American Freshwater Fishes, pages 778-826. Stanford, California: Stanford University Press. Snyder, J.O. 1917a. The Fishes of the Lahontan System of Nevada and Northeastern California. Bulletin of the United States Bureau of Fisheries, 35 (for 1915-1916):33-86. 1917b. An Account of Fishes from Owens River, California. Proceedings of the United States National Museum, 54:201-205. Sokol, O. 1954. To the Desert for Fishes. Aquarium Journal, 25:178-182. Soltz, D.L., and R.J. Naiman 1978. The Natural History of Native Fishes of the Death Valley System. Natural History Museum of Los Angeles County, Science Series, 30: 1-76. Stein, J.R. 1997. 1996 Survey Report: Meadow Valley Wash Desert Sucker (Catosto- mus clarki ssp.) and Meadow Valley Wash Speckled Dace (Rhinich- thys osculus ssp.). 4 pages. Las Vegas, Nevada: Nevada Division of Wildlife. [Unpublished report.] Stein, J.R., and J. Salisbury 1994. Status and Distribution of Relict Dace {Relictus solitarius). 36 pages, 2 appendices. Reno, Nevada: Nevada Department of Conser- vation and Natural Resources, Division of Wildlife. [Unpublished report.] Steward, J.H. 1933. Ethnography of the Owens Valley Paiute. University of California Publications in American Archaeology and Ethnology, 33:233- 350. Sumner, F.B. 1940. The Decline of the Pyramid Lake Fishery. Transactions of the Amer- ican Fisheries Society, 69:216-224. Szewczak, S.M. 1997. Summary Report on Bufo exsul. 5 pages, 10 appendices. Report to Deep Springs College, Big Pine, California. [Unpublished.] Taylor, D.W. 1980. Endangered and Threatened Mollusks of Amargosa Drainage, Cali- fornia-Nevada. 41 pages. Proposal to Office of Endangered Species, U.S. Fish and Wildlife Service, Washington, D.C. [Unpublished.] 1985. Evolution of Freshwater Drainages and Molluscs in Western North America. In C.J. Hocutt and A.B. Leviton, editors, Late Cenozoic History of the Pacific Northwest, pages 265-321. San Francisco: American Association for the Advancement of Science and Califor- nia Academy of Science. Taylor, D.W., and G.R. Smith 1981. Pliocene Molluscs and Fishes from Northeastern California and Northwestern Nevada. Contributions from the Museum of Paleon- tology, University of Michigan, 25:339-413. Taylor, F.R., R.R. Miller, J.W. Pedretti, and J.E. Deacon 1988. Rediscovery of the Shoshone Pupfish, Cyprinodon nevadensis shos- hone (Cyprinodontidae), at Shoshone Springs, Inyo County, Califor- nia. Bulletin of the Southern California Academy of Science, 87: 67-73. Thomas, J.W, C. Maser, and J.E. Rodiek 1979. Wildlife Habitats in Managed Rangelands?The Great Basin of Southeastern Oregon: Riparian Zone. United States Bureau of Land Management General Technical Report, PNW-80: 18 pages. Tuttle, P., G. Scoppettone, and D. Withers 1990. Status and Life History of Pahranagat River Fishes: Completion Re- port (1990). 51 pages. Seattle, Washington: U.S. National Fishery Research Center, and Reno, Nevada: Nevada Department of Wild- life. [Unpublished report.] United States Fish and Wildlife Service (USFWS) 1976. Warm Springs Pupfish Recovery Plan. 13 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1984. Mojave Tui Chub Recovery Plan. 56 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1985. Paiute Cutthroat Trout Recovery Plan. 63 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1986. Pahranagat Roundtail Chub Recovery Plan. 71 pages. Portland, Or- egon: U.S. Fish and Wildlife Service. 1987. Borax Lake Chub Recovery Plan. 81 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1990. Recovery Plan for the Endangered and Threatened Species of Ash Meadows, Nevada. 86 pages. Portland, Oregon: U.S. Fish and Wild- life Service. 1992. Cui-ui (Chasmistes cujus) Recovery Plan. Second revision, 47 pages, 5 appendicies. Portland, Oregon: U.S. Fish and Wildlife Ser- vice. 1993. Big Spring Spinedace, Lepidomeda mollispinis pratensis, Recovery Plan. 43 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1994. White River Spinedace, Lepidomeda albivallis, Recovery Plan. 45 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1995a. Endangered and Threatened Wildlife and Plants: Proposal to Deter- mine the Least Chub {Iotichthys phlegethontis) an Endangered Spe- cies with Critical Habitat. Federal Register, 60:50518-50530. 1995b. Recovery Plan for the Lahontan Cutthroat Trout. 108 pages. Port- land, Oregon: U.S. Fish and Wildlife Service. 1997a. Railroad Valley Springfish Recovery Plan. 57 pages. Portland, Ore- gon: U.S. Fish and Wildlife Service. 1997b. Recovery Plan for the Rare Species of Soldier Meadows. 50 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1998a. Owens Basin Wetland and Aquatic Species Recovery Plan, Inyo and Mono Counties, California. 127 pages. Portland, Oregon: U.S. Fish and Wildlife Service. 1998b. Recovery Plan for the Aquatic and Riparian Species of Pahranagat Valley. 82 pages. Portland, Oregon: U.S. Fish and Wildlife Service. Vinyard, G.L. 1984. A Status Report about the Independence Valley Speckled Dace (Rhinichthys osculus lethoporus), Independence Valley Tui Chub (Gila bicolor isolata), and Clover Valley Speckled Dace (Rhinich- thys osculus oligoporus); Three Fishes of the Northeastern Portion of Nevada. 21 pages. Report to the U.S. Fish and Wildlife Service, Reno, Nevada. [Unpublished.] 1994. Report of Population Stocking and Management Options for Dixie Valley Tui Chub (Gila bicolor ssp.) and an Assessment of Horse Creek Brook Trout (Salvelinus fontinalis). \6 pages. Report to U.S. Naval Air Station, Fallon, Nevada. [Unpublished.] 1996. Distribution of a Thermal Endemic Minnow, the Desert Dace, (Ere- michthys acros), and Observations of Impacts of Water Diversion on Its Population. Great Basin Naturalist, 56:360-368. Williams, CD., T.P. Hardy, and J.E. Deacon 1982. Distribution and Status of Fishes of the Amargosa River Canyon, California. 116 pages. Report to the U.S. Fish and Wildlife Service, Sacramento, California. [Unpublished.] Williams, CD., and J.E. Williams 1981. Distribution and Status of Native Fishes of the Railroad Valley Sys- tem, Nevada. Cal-Neva Wildlife Transactions, 1981:48-51. Williams, J.E., and C.E. Bond 1980. Gila boraxiobus, a New Species of Cyprinid Fish from Southeastern Oregon with a Comparison to Gila alvordensis. Proceedings of the Biological Society of Washington, 93:293-298. 1981. A New Subspecies of Tui Chib (Osteichthyes: Cyprinidae) from Guano Basin, Nevada and Oregon. The Southwestern Naturalist, 26:223-230. 1983. Status and Life History Notes on the Native Fishes of the Alvord Basin, Oregon and Nevada. Great Basin Naturalist, 43:409-420. Williams, J.E., D.B. Bowman, J.E. Brooks, A.A. Echelle, R.J. Edwards, D.A. Hendrickson, and J.J. Landye 1985. Endangered Aquatic Ecosystems of North American Deserts with a NUMBER 33 293 List ofVanishing Fishes of the Region. Arizona-Nevada Academy of Science, 20:1-62. Williams, J.E., J.E. Johnson, D.A. Hendrickson, S. Contreras-Balderas, J.D. Williams, M. Navarro-Mendoza, D.E. McAllister, and J.E. Deacon 1989. Fishes of North America Endangered, Threatened, or of Special Concern: 1989. Fisheries, 14(6):2-20. Williams, J.E., and D.W. Sada 1985. Status of Two Endangered Fishes, Cyprinodon nevadensis mion- ectes and Rhinichthys osculus nevadensis, from Two Springs in Ash Meadows, Nevada. The Southwestern Naturalist, 30:475-484. Williams, J.E., M.A. Stern, A.V. Munhall, and G.A. Anderson 1990. Conservation Status of the Threatened Fishes in Warner Basin, Ore- gon. Great Basin Naturalist, 50:243-248. Williams, J.E., and G.R. Wilde 1981. Taxonomic Status and Morphology of Isolated Populations of the White River Springfish, Crenichthys baileyi (Cyprinodontidae). The Southwestern Naturalist, 25:485-503. Williams, J.E., CD. Williams, and C.E. Bond 1980. Fishes of the Sheldon National Wildlife Refuge. In R. A. Tubb, prin- ciple investigator, Survey of Fishes, Amphibians and Reptiles on the Sheldon National Wildlife Refuge, Nevada, 58 pages. Unpublished report to Sheldon National Wildlife Refuge, United States Fish and Wildlife Service, Lakeview, Oregon. Yellen, J.E. 1977. Long Term Hunter Gather Adaptation to Desert Environments: A Biogeographical Perspective. World Archaeology, 8:262-274. Late Neogene Environmental History of the Northern Bonneville Basin: A Review of Palynological Studies Owen K. Davis ABSTRACT Pollen analysis of sediments from Great Salt Lake in the north- ern Bonneville Basin indicate six vegetational and pluvial stages since the deposition of Bishop volcanic ash 0.76 Ma (million years ago) and Lava Creek B volcanic ash 0.6 Ma. These analyses and other pollen diagrams from the region generally record elevated percentages of forest conifers (pine, spruce, fir, and Douglas fir) during the last two pluvials: 24-14.5 14C Ky B.P. (radiocarbon- dated, thousand years before present) and -150-130 Ka (thousand years ago). The regional vegetation was a mosaic of cold-dry pine woodland and sagebrush steppe. This mosaic was replaced by steppe vegetation 10.8 14C Ky B.P., with xeric indicators reaching maximum percentages 7-5 14C Ky B.P. Similarly, pluvial Lake Bonneville was at about the Gilbert Lake level from 87 to 24 Ka, except for a dry interval -70 Ka, and it was above the Gilbert level from 24 to 14.5 14C Ky B.P. It then fell below the Gilbert level (i.e., it did not occupy the Wendover Basin) except for brief periods after 11,300 ? 300 14C years B.P. Pollen analysis of sediments provided by the Amoco Production Company indicated that the upland vegetation (nonriparian, nonlit- toral) has been desert from the late Miocene to the present. Esti- mated temperature was highest and moisture least during the Pliocene, with temperatures cooling and moisture increasing there- after. The late Miocene (<5 Ma) was dominated by pollen of xero- phytic Sarcobatus and Ephedra, which was replaced by Chenopodiaceae-Amaranthus and Artemisia pollen during the Pliocene and early Pleistocene (5-0.75 Ma). After 0.75 Ma, per- centages of Artemisia and Pinus abruptly increase. Owen K. Davis, Department of Geosciences, University of Arizona, Tucson, Arizona 85721-0077, USA. Reviewers: W.A. Dimichele, Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0121, USA; David B. Madsen, Environmental Sciences, Utah Geologial Survey, Salt Lake City, Utah 84114, USA; Cathy Whit- lock, Department of Geological Sciences, University of Oregon, Eugene, Oregon 97403, USA; Peter E. Wigand, Quaternary Sciences Center, Desert Research Institute, 2215 Raggio Parkway, Reno, Nevada 89512,USA. Three noteworthy events punctuate the gradual evolution of the desert flora: a warm-cold oscillation 3.7 Ma, a cold interval 2.6 Ma, and abrupt cooling after 0.75 Ma. Pluvial-interpluvial cycles begin after the 0.75 Ma event, each with an average length of 0.1 million years. Diversion of the Bear River into the Great Salt Lake basin is indicated by increased pollen concentration after 0.31 Ma, near the oxygen isotope stage 8-9 transition. Introduction The Great Salt Lake region within the northern Bonneville Basin has been the subject of various geological and biological studies, including many vegetation histories. Early investiga- tions (e.g., Gilbert, 1885) shaped our understanding of Ceno- zoic climate as much as did studies of the glacial deposits of the midwestern United States. However, understanding of con- tinental glaciation has steadily developed through investigation of ocean sediments, whereas the pluvial record has relied on scattered outcrops woven together by radiocarbon (and other) dates. My goals in this paper are to synthesize previously pub- lished palynological studies of the northern Bonneville Basin region, to publish for the first time pollen analyses by Harry Yokum and Paul S. Martin (University of Arizona), and to compare the shorter records of Yokum and Martin with pollen diagrams created with data from the Amoco Production Com- pany's wells, which rival ocean cores in the data they provide. The Great Salt Lake contains an uninterrupted accumulation of pollen and other sediments spanning millions of years. The shorter, more detailed records of Yokum and Martin confirm trends shown in the uppermost layers of the Great Salt Lake wells; and the Amoco wells unite and extend the findings of these briefer investigations. The interpretation of pollen diagrams is based on the rela- tionship between vegetation and pollen. For the Great Salt Lake region, the modern vegetation and pollen rain is discussed in detail by Bright (1967) and Davis (1995). Briefly, valley bot- toms are covered by halophytic, shadscale, and greasewood 295 296 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES steppes dominated by Chenopodiaceae-Amaranthus (hereafter Cheno-^ms) and Sarcobatus pollen. Further upslope is sage- brush steppe dominated by Artemisia (sagebrush) pollen. After that is woodland characterized by Pinus (pine), Juniperus (ju- niper), and Quercus pollen; pine forest with Pinus and Pseudotsuga pollen; and subalpine forest with Pinus, Abies (fir), and Picea (spruce) pollen. The palynological study of Great Salt Lake environmental history began with Bright's (1967) study of Swan Lake, in Red Rock Pass (elevation, 1453 m), the overflow for pluvial Lake Bonneville. The lowermost date, 12,090 ? 300 14C years B.P. (radiocarbon-dated years before present; United States Geolog- ical Survey, National Center Laboratory number W-1338, -795-805 cm), on the 955 cm core provides a limiting age for the spillover, now dated at -14.5 Ka (thousand years ago) (O'Connor, 1993). Bright subsequently obtained additional ra- diocarbon dates for the Swan Lake core and for additional cores from a nearby spillover basin that he named the "Downey Core," but these analyses were never published. The basal pollen zones (12,090 ? 300 to 10,190 ? 250 14C years B.P.) of the Swan Lake pollen diagram are dominated by 55%?75% Pinus pollen, with moderate percentages (l%-5%) of Picea and Abies. From ~7.5 to ~2.8 14C Ky (thousand years) B.P. the diagram is dominated by Gramineae (grass), Artemisia, and Cheno-^m5 pollen. From -2.8 to -1.7 14C Ky B.P., Pinus increased to 30%, and thereafter Artemisia pollen dominates the Swan Lake diagram. The sequence in Bright's unpublished Downey Core record is very similar (Robert C. Bright, pers. comm., 1975). A third "Bonneville spillover" core, collected and analyzed by Stephen T. Jackson (pers. comm., 1985) near McCammon, Idaho, has a similar pollen stratigraphy, but its sediments also contain the Mount Mazama volcanic ash, whereas Bright's cores do not. In many respects, David B. Madsen's pollen diagram (Figure 1) of Great Salt Lake sediments near Carrington Island (Core C; GREAT SALT LAKE Core "C" 29,000 yr B.P. - 20 40 60 80 20 20 40 60 20 40 20 2000 FIGURE 1.?Percentage pollen diagram for Core C, from the southern (Tooele) basin of the Great Salt Lake, selected types. Pollen of wetland trees and herbs (curves at right) are outside the pollen sum. Age estimates are based on linear interpolation between the dated horizons. (Pollen analysis by D.B. Madsen in Spencer et al., 1984.) NUMBER 33 297 Spencer et ah, 1984, fig. 1) is similar to Bright's Swan Lake di- agram (not shown). Conifer (Pinus, Abies, Picea) pollen is abundant from 550 to 450 cm (-24-10.8 14C Ky B.P.). During the middle Holocene (near the position of Mt. Mazama volca- nic ash, 6.7 14C Ky B.P.), the abundances of Artemisia, Ambro- sia (ragweed), and Cheno-^ms pollen reach maxima, and Juni- perus pollen increases after 7.5 14C Ky B.P. (Figure 1). Two pollen diagrams from the Great Salt Lake were pub- lished by Peter J. Mehringer, Jr. (1985). A 1.7 m (~7 Ky) core taken near Bird Island (see Figure 2) records steady percent- ages of Pinus (<15%), Artemisia (30%), and Gramineae (5%). As in Bright's and Madsen's (Figure 1) pollen diagrams, Am- brosia, Sarcobatus, and Cheno-Ams pollen are most abundant during the middle Holocene ~7-5 14C Ky B.P. Juniperus pollen increases to a maximum of ~25% near the top of the diagram (Mehringer, 1985). Mehringer (1985) also analyzed pollen from 4.26 to 2.66 m deep in a core from Crescent Springs, west of Great Salt Lake. Just above 3.65 m (i.e., after 26,700 ? 900 14C years B.P.), the pollen of Picea, Abies, and Pinus increase, as they do in Mad- sen's Core C diagram (Figure 1). Pollen of wetground plants? Cyperaceae (sedge) + Triglochin (pondweed)?reach 18% be- low 3.65 m, indicating an extensive marsh near Crescent Springs. Decreased percentages of these wetground types above 3.65 m probably indicate that the Crescent Springs site (elevation, 1300 m) was flooded by Lake Bonneville, with the result that no marsh vegetation existed near the core site. Three pollen diagrams from the Wendover Basin, west of the Great Salt Lake basin, are published herein for the first time in their entirety. The pollen analyses were completed by Yokum and Martin in the early 1960s. Pinus curves for the Wendover and Knolls cores were published in Martin and Mehringer (1965). The "Carrozi Pit" pollen diagram (Figure 3) is based on samples collected by Paul Martin from a borrow pit near the railroad crossing of Carrozi, near Knolls, Utah. The samples were taken beneath a tufa layer dated at 11,300 ? 300 14C years B.P. by Meyer Rubin (Eardley et al., 1957:1171). In Figure 3, the sedimentation rate is assumed to be constant, and is esti- mated from the radiocarbon date and an age of 24 Ky at 110 cm for the Artemisia decveasQ-Pinus increase, based on the age of this event at Crescent Springs. The Wendover and Knolls diagrams are from "Eardley Cores" (Williams, 1993) taken in 1960. The Knolls core is from the eastern Wendover Basin (elevation, 1289 m; 113?20'W, 40?45'N) and is 152 m in length. The Wendover core is from the western Wendover Basin (elevation, 1285 m; 115?5'W, 40?45'N) and is 171 m in length. Sediments of the upper 29 m were analyzed for the Knolls core (Figure 4), and those from the upper 30 m were analyzed for the Wendover core (Figure 5). The chronologies for the Wendover and Knolls diagrams (Figures 4, 5) are based on the sedimentation rates derived from dated volcanic ashes by Williams (1993), including the Pahoa Ash from 160 Ka in the Wendover core (Figure 5). Ap- proximate ages of events discussed in the text are indicated as "est." on the left in Figures 4 and 5. For example, the beginning of the last interglacial, -130 Ka (oxygen isotope stage (OIS) 5, Martinson et al., 1987), occurs near 2200-2400 cm in the Knolls core (Figure 4) and at about 2000 cm in the Wendover core (Figure 5). The pollen preservation of the Wendover Basin cores (Fig- ures 3-5) is poor in many samples, probably because of inter- mittent desiccation of the sediments and subsequent pollen de- struction. However, similarities among these three diagrams and to the other pollen diagrams of this region permit some general statements of Late Quaternary vegetation change. The increase in Pinus pollen and the decrease in Artemisia pollen that are seen near the top of each diagram (Figures 3-5), in the Crescent Springs diagram of Mehringer (1985), and in the Core C diagram (Figure 1; Spencer et al., 1984) occurred ca. 24 Ka. The higher Cheno-^ms percentages below 110 cm (estimated at 24 Ka) in the Carrozi Pit pollen diagram (Figure 3) are con- sistent with growth of A triplex and other halophytes of the Chenopodiaceae family near the core site. This suggests expo- sure of the basin margin and relatively low lake depth, near the Gilbert Lake level (elevation, -1300 m) during OIS 3. The in- crease of Botryococcus (a colonial green algae) above 110 cm in the core may indicate lower water salinity and increased wa- ter depth during OIS 2; however, the expansion of Botryococ- cus does not occur in the Knolls core during OIS 2 (above 300 cm). In the Carrozi Pit and Knolls diagrams (Figures 3, 4), the percentages of Picea, Abies, and Pseudotsuga (hereafter PAP) indicate greater abundances of montane conifers (i.e., lower vegetation zones) during OIS 2. Increases in PAP during OIS 2 do not occur in the Wendover diagram (Figure 5), possibly be- cause of poor pollen preservation; however, both the Knolls and Wendover diagrams indicate increased PAP percentages near the estimated depth of OIS 6?the last glacial (Knolls, 2600 cm; Wendover, 2800 cm). The low values (l%-5%) of montane conifers and high values of Pinus (50%-60%) are consistent with low values of Picea and Abies (l%-6%) and moderate values of Pinus (20%?40%) in the Crescent Springs diagram (Mehringer, 1985). They suggest that glacial (i.e., OIS 2, OIS 6) vegetation on the western and northern margins of Lake Bonneville was xeric pine woodland and sagebrush steppe. In general, Juniperus percentages are low during inter- vals with elevated montane conifer percentages, so a pinon-ju- niper woodland (such as may be found at low elevation today) is not indicated. The Knolls diagram (Figure 4) contains a peak in Cyper- aceae (20%-40%) and Typha (cattail) pollen at 1500-1100 cm (estimated at 87-64 Ka?i.e., OIS 4). The Wendover diagram (Figure 5) shows modest increases in Cyperaceae and Typha pollen from 1300 to 600 cm at about the same time as in the Knolls diagram, based on sediment deposition rates (Williams, 1993). A zone of no pollen preservation from 1100 to 950 cm (estimated at -70 Ka) in the Wendover diagram suggests an in- terval of low water depth during which the Wendover core site was above the lake surface elevation and the Knolls site was 298 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 114 ?T"113 112 Swan Lake VBear River NEVADA ? UTAH Wendover 41 40 39 FIGURE 2.?Map of the Bonneville Basin showing location of Amoco Production Company wells and other sites of pollen analysis (bullets). The blackened area is approximately the current extent of Great Salt Lake. The shaded area is the approximate extent of the Gilbert Lake level, the solid line represents the Provo Lake level, and the dashed line represents the Lake Bonneville level. NUMBER 33 299 CARROZI PIT, UTAH 11,300 ?200 est. 24,000 yr 60 70 -| 80 90 100- 110- 120 130H 140 150 -p 160-5 170- ^ 180? 190 : 03 200- Q 210 220- 230: 240: 250 260- 270- 280- 290 300 310 s? :& :<&>4v ss // 100 20 40 20 20 40 60 20 20 40 FIGURE 3.?Percentage pollen diagram for Carrozi Pit, near Knolls, Utah. Pollen of Cyperaceae and Botryococ- cus (curves at right) are outside the pollen sum, selected types. Picea, Abies, and Pseudotsuga {PAP) were not recorded individually in the original counts and are shown as a combined curve. Shading for Juniperus and PAP is a tenfold exaggeration. Black square at left is position of radiocarbon date (stratigraphically above the sedi- ments analyzed). (Pollen analysis by H. Yokum and P. Martin, unpublished.) below it. These elevations are near the post-pluvial Gilbert shoreline level. Isostatic depression during the Gilbert stage was about 10 m in the basin center (e.g., near the Knolls site) but less than 1 m at the basin margin (near Wendover; Currey, 1990), so the relative elevations of the Wendover and Knolls sites would have been reversed during the 70 Ka drying event. In summary, lowering of vegetation zones during OIS 2 (-24-15 Ka) and OIS 6 (-150-130 Ka) is indicated by higher percentages of Pinus pollen and moderate increases of PAP pollen (Figures 4, 5). Modern analogs for these assemblages can be found 500-1500 m higher in elevation (Davis, 1981), so cooling by 3?-9? C relative to present is indicated. Modern pol- len samples from closed-canopy montane forest, however, have much higher percentages of montane conifer pollen (e.g., 10%-20% PAP). Therefore, the full-glacial vegetation from 1550 to 1600 m above sea level was a mosaic of cold-dry pine woodland and sagebrush steppe rather than mesic closed-can- opy forest. Steppe vegetation expanded and pine woodland contracted beginning about 10.8 14C Ky B.P., with maximum percentages of xeric indicators from ~7 to 5 14C Ky B.P. The pluvial lake level in the Wendover Basin was near the Gilbert Lake level during OIS 4 and OIS 3 (87-24 Ka, esti- mated) except for an interval of drying ~70 Ka (estimated; Fig- ures 4, 5). Lake depth was well above the Gilbert level during OIS 2 (24-14 Ka) and presumably during OIS 6. Lake Bonne- ville did not permanently occupy the Wendover Basin after 11,300 ? 300 14C years B.P. Pollen of littoral vegetation (e.g., Cyperaceae and Typha) is most abundant in the Knolls and Wendover cores (Figures 4, 5) during low-water (Gilbert level) phases when the lake shore is near the core sites. Methods Pollen analysis of well cuttings, made available by Amoco Production Company, extend the vegetation and climatic re- cord of the Great Salt Lake Basin into the Miocene. THE AMOCO WELLS.?From June 1978 through December 1980, Amoco Production Company oversaw the drilling of 15 wells into the sediments of Great Salt Lake from floating plat- forms. Several of these wells exceeded 3 km in depth, and the cumulative thickness of sediment drilled exceeded 26 km. Sed- iment cuttings were collected from the wells at selected inter- vals of 3 m or 10 m. Several kinds of geophysical and sedimen- tological data were gathered for each well, and approximately 2700 line-kilometers of seismic reflection data also were col- lected. These data were provided to the University of Arizona, Department of Geosciences, as part of a cooperative research agreement with Amoco Production Company. 300 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 0 oO ?8888888888888888888888 (wo) Hld3Q oooooooooooooo n*IO en3 O ht)bO trieg 8 C4 ? 3 kOJH XI 'en_>> "3c ca c_u "oQ- B.P. ?e years ea *520 pm fractions before being mounted on separate slides. For counting purposes, the relative pollen concentration on each slide was assumed to be equals so equal numbers of rows were counted on each slide until a full count (300 grains) was reached. This procedure produced results very similar to undivided samples. The Amoco samples were supplemented with samples pre- pared at 3 m intervals from the Indian Cove Well (Figure 2). The site is about 5 km offshore (west) from Little Valley of the Promontory Range. Well cuttings were available at 3 m (10 ft) intervals from 15 to 220 m deep (15-606 Ka) and at 9 m (30 ft) intervals from 232 to 628 m (643-1484 Ka). Given the sedi- mentation rate for the Indian Cove well, this sample density gives 8 Ky resolution to 600 Ka and 23 Ky resolution from 0.6 to 1.5 Ma (million years ago). POLLEN COUNTING.?Three hundred pollen grains of upland species (nonriparian, nonlittoral) were counted whenever pos- sible. This sum?used as a divisor for calculating percentages of all types?did not included algae, fungal spores, or the pol- len of wet-ground plants. Samples devoid of pollen were not included in the pollen diagrams. Pollen extraction of 1 cm3 samples of Indian Cove cuttings followed routine methods, including acid digestion of minerals with HC1 and HF, removal of organic material with acetolysis, and removal of organic acids with 10% KOH. Lycopodium spore tablets were added to permit calculation of pollen con- centration. VOLCANIC ASH IDENTIFICATIONS.?Chronological control for the Amoco wells is based on the identification of volcanic ashes on known age. The methods and identifications are as published by Moutoux (1995), Moutoux and Davis (1995), Davis (1998), Davis and Moutoux (1998), and Moutoux et al. (in prep.). CLIMATE RECONSTRUCTIONS.?Quantitative climatic recon- structions were made by comparing fossil pollen samples with 1367 modern pollen samples from the western United States (Davis, 1995). A PASCAL computer program calculated the 302 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES squared chord distance between each fossil sample and the 1367 modern analogs by using 28 common pollen types. Wet- land plants (e.g., Cyperaceae and Typha) were not used. Mod- ern samples were considered analogs if the squared chord dis- tance was 0.15 or less. The climate estimates are the averages of annual temperature and precipitation for the analogs. Values were not calculated if modern analogs were not found. To provide a basinwide climate reconstruction, the climate estimates for three wells?Gunnison, Indian Cove, and Bridge?were combined. Values for two other wells?South Rozell and Carrington Island?were not included because pres- ervation was poor in some intervals. The climate estimates for the three wells were averaged after climate had been interpo- lated at 50 Ky intervals. The average curve was then smoothed by an 11 sample moving average. The calculations were then made for tectonic subsidence of the eastern Great Basin during the last 5 My (million years). During the Sevier orogeny of the Cretaceous, compressional tectonics resulted in the uplift of the Great Salt Lake region to approximately 1000-2000 m above the current average eleva- tion (DeCelles, 1994). This uplift relaxed during two intervals of Cenozoic extension: 50 to 40 Ma and -30 Ma to the present. Wolfe et al. (1997) estimated equivalent lowering for the west- ern Great Basin?but at a later date, ~ 13 Ma. Approximately one-half of the sediment in the Great Salt Lake Basin accumulated during the last 5-10 Ma (Mohapatra, 1996). The thickness of sediment accumulation in the central Great Salt Lake basin is more than twice the lowering of sur- face elevation for the region, but I assumed that because both resulted from Cenozoic extension, their timing is similar; i.e., about one-half of the lowering occurred during the last 5-10 Ma. I used temperature rates intermediate to Miocene (2.0? C/ km; Wolfe et al., 1997) and contemporary (6.0? C/km; Davis, 1981) values to estimate ~4? C warming and 225 mm lowering of precipitation during the last 5 Ma. These values were used to correct the analog-based climate estimates for tectonic lower- ing. Results and Interpretation Pollen analysis of the Amoco well cuttings provided data that are readily comparable to those of routine analysis of cores and contemporary pollen samples. Modern analogs for the fos- sil samples were abundant from 5.5 Ma onward (Figure 6). Per- haps this was the result of "time averaging" of pollen in routine pollen samples (each packet of sediment containing pollen from several seasons of plant growth). Also, "habitat averag- ing" of the pollen from different vegetation zones in mountain- ous topography may produce assemblages similar to the "depth-averaged" cuttings. Furthermore, each cutting sample integrated sediment from a range of depths, so the procedure (followed by Amoco personnel) of combining three samples only widened the mixing range. The regional vegetation recorded in the Amoco pollen dia- grams was desert from the late Miocene onward. The pollen spectra were dominated by xerophytic Sarcobatus and Ephedra pollen during the late Miocene (>5 Ma). They were replaced by Cheno-Ams and Artemisia during the Pliocene and early Pleistocene (5-0.75 Ma) as the Chenopodiaceae and Composi- tae (Asteraceae) diversified and spread through the flora (Le- opold and Denton, 1987; Cerling et al., 1997). Artemisia and Pinus percentages abruptly increased in the late Pleistocene (0.75 Ma). The evident antiquity of desert vegetation in the eastern Great Basin conflicts with the hypothesis that deserts expanded during the last few million years in response to uplift of the Sierra Nevada (Axelrod, 1957; Axelrod and Bailey, 1969). These reconstructions were based on leaves and other plant macrofossils, but in Great Salt Lake sediment the pollen of mesic elements?"Tertiary exotics" such as Ulmus (elm), Carya (hickory), Ostrya-Carpinus (hop-hornbeam)?was rare, although it persisted into the middle Pleistocene. These plants probably existed in riparian settings, as do birch, maple, ash, and other "mesic" trees today. The differences between environmental reconstructions based on pollen and those based on macrofossils must result, in part, from the proximity of these mesic trees to the rivers, del- tas, and other sedimentary environments in which the macro- fossil floras are preserved (Axelrod, 1957; Axelrod and Bailey, 1969). In contrast, the recruitment area for airborne pollen en- compasses a much larger area, including the xeric uplands. MAJOR VEGETATION EVENTS.?Three noteworthy events at 3.7, 2.6, and 0.75 Ma punctuated the gradual evolution of the desert flora. The 3.7 Ma event was best represented in the Bridge and South Rozel wells, and the 2.6 Ma event was best represented in the East Gunnison and Indian Cove wells. The 3.7 Ma event was characterized in the upland (nonwetland) pollen by an abrupt increase in Pinus and Artemisia and a de- cline in Cheno-^W5 (Figure 6); it was similar to the 0.75 Ma event. In the Bridge well (Figure 6), poor preservation and high Cheno-^m5 percentages from 3980 to 3682 Ka (1698-1607 m) were followed by high percentages of Artemisia and Pinus (also Abies and Cupressaceae) from 3580 to 3320 Ka (1576- 1454 m). In the South Rozel well, the same transition occurred between 3790 and 3580 Ka (2200-2100 m). This event matched the marine record for cooling and increased ice vol- ume dated 3.8-3.6 Ma (Tiedmann et al., 1994). During the 2.6 Ma event, the pollen diagrams from the East Gunnison and Indian Cove wells show reduced percentages of Cheno-Ams and increased Pinus, Artemisia, Abies, and Cupres- saceae. However, the percentages of upland taxa changed less during this event than during the 3.7 or 0.75 Ka events. In all Great Salt Lake wells investigated (Moutoux, 1995; Moutoux and Davis, 1995; Moutoux et al., in prep.), the 2.6 Ma event co- incided with the beginning of high percentages of wetland taxa (Cyperaceae, Typha, and Azolla) and algae (Pediastrum and Botryococcus). In the marine record, this event marked the be- ginning of the Pleistocene, with increasing 5 I8O values and a NUMBER 33 303 S ? 2 E 60 J>it ? 43? to ? .22 2^ V .? o >> ?2 S in W'?^ s cs-^3 ja SSI . os" ? 3 5 ? ? ? 3 C ?2 - E I ?2 I c 3 if O 60 ? .5 o 5 E ? s ? 304 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES greater amplitude of glacial-interglacial fluctuations (Tied- mannetal., 1994). The 0.75 Ma event was the most distinctive transition of the Great Salt Lake pollen record (Moutoux, 1995; Moutoux and Davis, 1995; Moutoux et al., in prep.). Before 750 Ka, Cheno- Ams pollen exceeded 50% of the pollen assemblage; thereafter, Artemisia and Pinus predominated, Cupressaceae and other Compositae increased, and Gramineae decreased (Figure 6). This event appears to be a permanent downward displacement of vegetation zones by ~500 m in elevation, in which grease- wood steppe (with high percentage of Cheno-Ams) was re- placed by sagebrush steppe in the intermountain basins. Terres- trial (versus marine) pollen records indicate significant changes occurred during the late Pleistocene, such as the general in- crease of steppe vegetation {Artemsia and Gramineae) after 620 Ka in northern California (Adam et al., 1989) and the ex- pansion of Quercus after 900 Ka in Colombia (Hoogheimstra, 1984); however, no other terrestrial record indicated a 750 Ka event that was of greater magnitude than the 2.6 Ma one. Ma- rine records showed increased amplitude of 100 Ky periodicity after 750 Ka (Raymo et al., 1987; Ruddiman et al., 1987) but indicated no change in average conditions. CLIMATE RECORD.?The climate reconstruction for Great Salt Lake wells indicated maximum temperature and minimum moisture during the Pliocene, with cooling and increasing moisture thereafter. This trend differs from that of earlier Great Salt Lake publications (Moutoux, 1995; Moutoux and Davis, 1995) because of the incorporation of tectonic subsidence in the climate estimates. A pronounced temperature peak (5? C greater than modern values) at 3.8 Ma coincided with the Cheno-^4m5 maximum preceding the 3.7 Ma event (Figure 7). The peak itself should be considered tentative because the cli- mate reconstruction for that time included only the Bridge well data. The peak was followed by oscillations leading to a decline in temperature ~3.0 Ma that was reflected in a decline of polar temperatures (Funder et al., 1985; Webb and Harwood, 1991) and a lowering of eustatic sea levels by 35 m (Krantz, 1991). The 2.6 Ma event (2? C cooling; Figure 7) appears in the Bridge well pollen diagram (Figure 6) as decreased Cheno-,4ww and increased conifer pollen. Thereafter, temperatures gradu- ally declined and precipitation increased until 0.75 Ka, when the overall (average of glacial + interglacial) temperature dropped 3? C and annual precipitation increased by 100 mm. The uppermost samples provided estimates of temperature at 10.8? C and precipitation at 310 mm, which are equivalent to present-day values (Figure 7). PLUVIAL CYCLES.?The relationships between glacial indi- cators, interglacial indicators, wetland types, and algae were used to produce the water depth reconstruction in Figure 8. Six pluvial-interpluvial cycles followed the deposition of Lava Creek B ash, with an average cycle length of-100 Ky (Figure 8). Water depth minima coincided with low values of wetland plants and high values of glacial indicators; water depth max- ima also coincided with low values of wetland plants but high 2.5 n 3.0 - 3.5 - 4.0 - 4.5 - 18 n 16 - 14 - 12 - "~ 10 - ^600 -i 37 2.6 ODP 846 o Q. E CD 400 H 200 H Combined Temperature Combined Precipitation i ' 1??? 1 5000 4000 3000 2000 Age (Ky) 1000 0 FIGURE 7.?Comparison of Great Salt Lake climate estimates with global records of climate change (bold lines). Climate reconstruction for Great Salt Lake wells was based on modern analogs (squared chord distance <0.14) in a database of 1367 contemporary samples from western North America. Average annual temperature (Temp.) and precipitation (Precip.) curves are smoothed 11-point averages of estimates for the Gunnison, Indian Cove, and Bridge wells. Dashed lines are the temperature and precipitation estimates that have not been corrected for tectonic subsidence of the region. The 8I8O line is for Ocean Drilling Program (ODP) site 846 (Pisias et al., 1992). values of interglacial indicators. The amplitude of the oscilla- tions increased through time, with minima becoming lower and maxima becoming greater. Evidence from core samples indicates that before 2.6 Ma pollen preservation was poor, fungal spores were abundant, and the pollen of wetland plants was scarce. Apparently, the lakes and marshes of the Great Salt Lake basin were shallow and in- termittent. From 2.6 until 0.75 Ma, pollen preservation was good, and percentages of wetland plants (Cyperaceae, Typha, and Myriophyllum) and planktonic algae {Pediastrum and Bot- ryococcus) were generally high (Figure 6). During this interval (early Pleistocene), shallow lakes and marshes apparently per- sisted in the northern basin of Great Salt Lake. Beginning 0.75 Ma, spanning the sedimentary interval con- taining Bishop (0.76 Ma) and Lava Creek B (0.6 Ma) ashes (Sarna-Wojcicki et al., 1991), and contemporaneous with the cooling and moistening recorded by the upland pollen data (Figure 7), the percentages of wetland plants and algae de- clined (Figures 1, 4). The event was marked in the aquatic en- vironment by a peak of Botryococcus or Pediastrum soon after the deposition of Lava Creek B ash. After the transition, the close-interval samples from the Indian Cove well (Figure 8) show a regular pattern of montane confers {Abies + Picea + Pseudotsuga) during glaciations and of basin floor taxa (Cu- pressaceae + Ambrosia + Sarcobatus) during interglacials (Davis and Moutoux, 1998). Figure 8 shows that high percent- ages of the interglacial indicators coincided with the odd stages NUMBER 33 305 INDIAN COVE WELL Great Salt Lake Pollen Percent 500 1000 1500 20 20 40 60 20 20 40 20 20 40 60 100 200 300 FIGURE 8.?Indian Cove pollen diagram showing selected types. The chronology is based on the presence of Lava Creek B and Bishop volcanic ashes, adjusted by matching the marine oxygen isotope chronology. Curves to the right of Cupressaceae + Ambrosia + Sarcobatus are not included in the pollen concentration sum. Inferred water depth is a smoothed curve of pluvial indicators {Abies + Picea + Pseudotsuga, weighted xlO) plus shallow water indicators (wetland types and Botryococcus + Pediastrum) minus interpluvial indicators (Cupressaceae + Ambrosia + Sarcobatus) expressed as percent of the pollen sum. The values are smoothed by a 3-point moving average. Correlation with the marine oxygen isotope chronology is indicated to the right, where the ages of the even-number stages are marked. of the marine oxygen isotope chronology, and that high per- centages of the glacial indicators matched the even stages. Algae and the pollen of wetland plants was most abundant at the beginning of the even-numbered stages, shown in Figure 8; i.e. rare during interpluvials when the lake was saline and rare during pluvials when littoral vegetation was far from the core site, but common when the lake was fresh and shallow. This pattern implies global climatic control of lake cycles during the last 0.75 My. DIVERSION OF BEAR RIVER.?The Indian Cove diagram (Figure 8) shows a dramatic increase in total pollen concentra- tion (from an average of 6000 to 30,000 grains/cm3) after 310 Ka (marine OIS 9). The concentration remains high to the sur- face and represents a fundamental change in the pollen sedi- mentation of the lake. This transformation may have resulted from the diversion of the Bear River (the largest tributary) into the Great Salt Lake basin. Before its diversion, the river had flowed northward into the Snake River drainage via Lake Thatcher. Spillover of Lake Thatcher into the Great Salt Lake may have lead to the catastrophic overflow of Lake Bonneville into the Snake River plain 14.5 Ky B.P. Bright (1967) dated the diversion from the Snake River drainage at 35 14C Ky B.P., but unpublished K/Ar dates re- corded for the damming basalt are highly variable (R.C. Bright, pers. comm., 1994). Strontium isotope data of Bouchard et al. (1998) suggest that the history of Lake Thatcher and the Bear River may involve more than one diversion, with the earliest evidence for Bear River in the Thatcher Basin dating to marine OIS 6. Regardless of whether the proposed diversion dates to OIS 9 or OIS 6, it appears not to have permanently increased 306 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the lake's depth. Figure 8 suggests that the amplitude of the pluvial-interpluvial oscillations steadily increased through the late Pleistocene, culminating in the catastrophic overflow through Red Rock Pass. Conclusions Although the time resolution of the Amoco well pollen sam- ples is comparatively low, they permit, for the first time, the re- construction of Late Cenozoic vegetation change for the Great Basin at a single locality. Furthermore, the pollen diagrams from the Amoco 30 m amalgam samples are comparable to the higher resolution 10 m unmixed samples (Figure 8), which are in turn consistent with the data from core sediments (Figure 1). Long cores from the Great Salt Lake are needed to produce a detailed reconstruction of Late Cenozoic vegetation and cli- mate, but the Amoco cuttings have produced an unparalleled record of long-term change of terrestrial vegetation. The pollen diagrams trace the development of the Great Ba- sin desert during the Late Cenozoic, produce a climatic recon- struction comparable to that obtained from marine sediments, and afford an indication of Lake Bonneville pluvial cycles of increasing amplitude, beginning 620 Ka. The cooling and transformation of the vegetation and pluvial cycles 0.75 Ma are a unique feature of this study. It is a robust conclusion, re- corded in all wells and in both terrestrial and aquatic indicators. The transformation is different from the increased amplitude of oscillations of the marine record and should be thoroughly in- vestigated in Great Salt Lake cores and in other long terrestrial records. ACKNOWLEDGMENTS.?The sediments used in this study were provided by Amoco Production Company. Analyses were supported by National Science Foundation grants EAR- 9219505 and EAR-9205065. Data for pollen analysis of Core C near Bird Island were provided by David P. Madsen. Data for the Carrozi Pit, Knolls, and Wendover cores were provided by Paul S. Martin. Nicholas Pisias generously provided data for Ocean Drilling Program (ODP) site 846 (Pisias et al., 1992), and Ralf Tiedman provided the ODP site 659 data (Tiedmann et al., 1994). Tom Moutoux counted the pollen for the Bridge well and identified the volcanic ashes with the generous assis- tance of Michael Perkins and William P. Nash. Literature Cited Adam, D.P., A.M. Sarna-Wojcicki, H.J. Rieck, J.P. Bradbury, W.E. Dean, and R.M. Forester 1989. Tule Lake, California, the Last 3 Million Years. Palaeogeography, Palaeoclimatology, and Palaeoecology, 73:89-103. Axelrod, D.I. 1957. Late Tertiary Floras and the Sierra Nevada Uplift. Bulletin of the Geological Society of America, 68:19-46. Axelrod, D.I., and H.P. Bailey 1969. Paleotemperature Analysis of Tertiary Floras. Palaeogeography, Palaeoclimatology, and Palaeoecology, 6:163?195. Bouchard, D.P., D.S. Kaufman, A. Hochberg, and J. Quade 1998. Quaternary History of the Thatcher Basin, Idaho, Reconstructed from the 87Sr/86Sr and Amino Acid Composition of Lacustrine Fos- sils: Implications for the Diversion of the Bear River into the Bon- neville Basin. Palaeogeography, Palaeoclimatology, Palaeoeco- logy, 141(l-2):95-114. Bright, R.C. 1967. Late-Pleistocene Stratigraphy in Thatcher Basin, Southeastern Idaho. Tebiwa, 10:1-7. Cerling, T.E., J.M. Harris, B.J. MacFadden, M.G. Leakey, J. Quade, V. Eisen- mann, and J.R. Ehleringer 1997. Global Vegetation Change through the Miocene/Pliocene Boundary. Nature, 389:153-158. Currey, D.R. 1990. Quaternary Paleolakes in the Evolution of Semidesert Basins, with Special Emphasis on Lake Bonneville and the Great Basin, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology, 76:189-214. Davis, O.K. 1981. Vegetation Migration in Southern Idaho during the Late-Quaternary and Holocene. 252 pages. Doctoral dissertation, University of Min- nesota, St. Paul. 1995. Climate and Vegetation Patterns in Surface Samples from Arid Western U.S.A.: Application to Holocene Climatic Reconstructions. Palynology, 19:97-120. 1996. The Impact of Climatic Change on Available Moisture in Arid Lands: Examples from the American Southwest. In A.N. Angelakis and A.S. Issar, editors, Diachronic Climatic Impacts on Water Re- sources, pages 283-299. NATO ASI Series I, volume 36. Berlin; Heidelberg: Springer-Verlag. 1998. Palynological Evidence for Vegetation Cycles in a 1.5 Million Year Pollen Record from the Great Salt Lake, Utah, U.S.A. Palaeogeog- raphy, Palaeoclimatology, Palaeoecology, 138:175?185. 1999. Preliminary Pollen Analysis of Cenozoic Sediments of the Great Salt Lake, Utah, U.S.A. In J.H. Wrenn, J.-P. Sue, and S.A.G. Leroy, edi- tors, The Pliocene: Time of Change, pages 227-240. Dallas, Texas: American Association of Stratigraphic Palynologists Foundation. Davis, O.K., and T.E. Moutoux 1998. Tertiary and Quaternary Vegetation History of the Great Salt Lake. Journal of Paleolimnology, 19:417-427. DeCelles, P.G. 1994. Late Cretaceous-Paleocene Synorogenic Sedimentation and Kine- matic History of the Sevier Thrust Belt, Northeast Utah and South- east Wyoming. Bulletin of the Geological Society America, 106: 32-56. Eardley, A.J., V. Gvodetsky, and R.E. Marsell 1957. Hydrology of Lake Bonneville and Sediments and Soils of Its Basin. Bulletin of the Geological Society America, 66:1141-1202. Funder, S., N. Abrahamsen, O. Bennike, and R.W. Feyling-Hanssen 1985. Forested Arctic: Evidence from North Greenland. Geology, 13: 542-546. Gilbert, G.K. 1885. The Topographic Features of Lake Shores. United States Geological Survey, Annual Report, 5-B:69-123. Hoogheimstra, H. 1984. Vegetational and Climatic History of the High Plain of Bogota, Co- lombia: A Continuous Record of the Last 3.5 Million Years. Disser- NUMBER 33 307 tationes Botanicae, 79: 368 pages. Krantz, D.E. 1991. A Chronology of Pliocene Sea-Level Fluctuations, U.S. Atlantic Coastal Plain. Quaternary Science Reviews, 10:163-174. Leopold, E.B., and M.F. Denton 1987. Comparative Age of Grassland and Steppe East and West of the Northern Rocky Mountains. Annals of the Missouri Botanical Gar- den, 74:841-867. Martin, P.S., and P.J. Mehringer, Jr. 1965. Pleistocene Pollen Analysis and Biogeography of the Southwest. In H.E. Wright, Jr., and D.G. Frey, editors, The Quaternary of the United States, pages 433-451. Princeton, New Jersey: Princeton University Press. Martinson, D.G., N.G. Pisias, J.D. Hays, J. Imbrie, T.C. Moore, Jr., and N.J. Shackleton 1987. Age Dating and the Orbital Theory of the Ice Ages: Development of a High-Resolution 0-300,000-Year Chronostratigraphy. Quater- nary Research, 27:1-29. Mehringer, P.J., Jr. 1985. Late-Quaternary Pollen Records from the Interior Pacific Northwest and Northern Great Basin of the United States. In V.M. Bryant and R.M Holloway, Pollen Records of Late-Quaternary North American Sediments, pages 168-189. Dallas, Texas: Foundation of the Ameri- can Association of Stratigraphic Palynologists. Mohapatra, G.K. 1996. Faulting and Basin Geometry Beneath the Great Salt Lake: Implica- tions for Basin Evolution and Cenozoic Extension. 159 pages. Doc- toral dissertation, University of Arizona, Tucson, Arizona. Moutoux, T.E. 1995. Palynological and Tephra Correlations among Deep Wells in the Modern Great Salt Lake, Utah, U.S.A. 30 pages. Master's thesis, University of Arizona, Tucson. Moutoux, T.E., and O.K. Davis 1995. Neogene through Pleistocene Paleoclimate of the Great Salt Lake Region, Northeastern Great Basin, U.S.A. California Department of Water Resources Ecological Studies Program, Technical Report, 40: 127-135. Moutoux, T.E., O.K. Davis, M. Perkins, and W.P. Nash In prep. Palynological and Tephra Correlations among Deep Wells in the Modern Great Salt Lake, Utah, U.S.A. Bulletin of the Geological Society America. O'Connor, J.E. 1993. Hydrology, Hydraulics, and Geomorphology of the Bonneville Flood. Geological Society of America Special Paper, 274: 83 pages. Pisias, N.G., L.A. Mayer, T.R. Janecek, A. Palmer-Julson, and T.H. van Andel 1992. Site 846. In Initial Reports of the Deep Sea Drilling Project, 138: 265-333. United States Government Printing Office. Raymo, M.E., W.F. Ruddiman, and B.M. Clement 1987. Pliocene-Pleistocene Paleoceanography of the North Atlantic at Deep Sea Drilling Project Site 609. In Initial Reports of the Deep Sea Drilling Project, 94(2):895-901. United States Government Printing Office. Ruddiman, W.F., A. Mclntyre, and M. Raymo 1987. Paleoenvironmental Results from North Atlantic Sites 607 and 609. In Initial Reports of the Deep Sea Drilling Project, 94(2):855-878. United States Government Printing Office. Sarna-Wojcicki, A.M., K.R. Lajoie, C.E. Meyer, D.P. Adam, and J.J. Rieck 1991. Tephrochronologic Correlation of Upper Neogene Sediments along the Pacific Margin, Conterminous United States. In R.B. Morrison, Editor, Quaternary Nonglacial Geology: Conterminous U.S. The Geology of North America, K-2:117-140. Boulder, Colorado: Geo- logical Society of America. Spencer, R.J., M.J. Baedecker, J.P. Eugster, R.M. Forester, M.B. Goldhaber, B.F. Jones, K. Kelts, J. Mckenzie, D.B. Madsen, S.L. Rettig, M. Rubin, and C.J. Bowser 1984. Great Salt Lake and Precursors, Utah: The Last 30,000 Years. Con- tributions to Mineralogy and Petrology, 86:321-334. Tiedmann, R., M. Sarnthein, and N.J. Shackleton 1994. Astronomic Timescale for the Pliocene Atlantic A18O and Dust Flux Records of Ocean Drilling Program Site 659. Paleoceanography, 9: 619-638. Webb, P.-N., and D.M. Harwood 1991. Late Cenozoic Glacial History of the Ross Embayment, Antarctica. Quaternary Science Reviews, 10:215-223. Williams, S.K. 1993. Tephrochronology and Basinal Correlation of Ash Deposits in the Bonneville Basin, Northwest Utah. 104 pages. Mater's thesis, Uni- versity of Utah, Salt Lake City. Wolfe, J.A., H.E. Schorn, C.E. Forest, and P. Molnar 1997. Paleobotanical Evidence for High Altitudes in Nevada during the Miocene. Science, 276:1672-1675. Great Basin Vegetation History and Aquatic Systems: The Last 150,000 Years Peter E. Wigand and David Rhode ABSTRACT The Great Basin was seen by early settlers as a terrifying barrier to be crossed on the way to the golden promise that was California. Although the region was visited by fur trappers and mountain men in the 1820s and 1830s, it was not until the 1840s that the Euro- pean pathfinders and empire builders tarried to explore the wood- land-covered mountains and playa-bottomed valleys of the Great Basin. Mormon farmers and spirited miners settled this wild land, uprooting the native people who had lived there for millennia. As the region opened to European settlement, evidence of a long and complex environmental history was discovered and explored. But as recently as 50 years ago our knowledge of Late Quaternary lake histories and climates was minimal, and our knowledge of the dynamics of Pleistocene and Holocene vegetation was virtually nonexistent. Our understanding has advanced dramatically in the last half century, with advances in analyses of pollen, woodrat middens, and tree rings. A substantial pollen and plant macrofossil database has been assembled, spanning tens of thousands to hun- dreds of thousands of years, such that our knowledge of Late Qua- ternary vegetation history in the Great Basin rivals that of any other region in the world. These data, coupled with the long-term lake histories deduced from the basins of the Lahontan, Bonne- ville, Chewaucan, and Owens-to-Death Valley lake systems, per- mit detailed examination of late Pleistocene vegetation dynamics in relation to aquatic system history. By examining proxy records from the various subregions of the Great Basin and the northern Mojave Desert, we have compiled a compelling picture of past cli- mate and the vegetation and hydrologic responses during the last 50,000-150,000 years. Peter E. Wigand, Quaternary Sciences Center, Desert Research Insti- tute, 2215 Raggio Parkway, Reno, Nevada 89512, USA. David Rhode, Earth and Ecosystem Sciences, Desert Research Institute, 2215 Rag- gio Parkway, Reno, Nevada 89512, USA. Reviewers: Owen K. Davis, Department of Geosciences, University of Arizona, Tucson, Arizona 85721-0077, USA; DavidB. Madsen, Envi- ronmental Sciences, Utah Geologial Survey, Salt Lake City, Utah 84114, USA; P. Peterson, Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560- 0166 USA; G. Spaulding, Dames and Moore, 7180 Pollock Drive, Suite 200, Las Vegas, Nevada 89119, USA. Introduction Our understanding of Late Quaternary biological history in the Great Basin, particularly its vegetation history, has ad- vanced substantially in the half century since the publication of works by Antevs (1948) and by Hubbs and Miller (1948). By that time evidence derived from the interpretation of modern biogeographic indicators such as disjunct taxa and isolated hy- brid populations had led to the recognition that Pleistocene gla- ciopluvial episodes and mid-Holocene warming wrought sub- stantial shifts in vegetation zones, but even the barest outlines of these shifts remained uncertain. In 1948 the number of Late Quaternary fossil plant localities studied in the Great Basin was negligible (Deevey, 1949). Laudermilk and Munz (1934, 1938) had analyzed late Pleistocene pollen and plant fragments from dung balls of the Shasta ground sloth {Nothrotheriops shaten- sis), found in Gypsum and Rampart Caves along the lower Col- orado River just outside the southern margins of the Great Ba- sin. Plant remains had been recovered from other cave sites (e.g., Smith Creek Cave; Thompson, 1984) but these were as yet unstudied. Only two Late Quaternary pollen sequences from within the Great Basin?Chewaucan and Warner marshes in south-central Oregon?had been analyzed by Ernst Antevs' colleague Henry Hansen (Hansen, 1947). These pollen records, in combination with knowledge of the northern European cli- matic sequence (Blytt-Sernander) (Iversen, 1973:13-14), were instrumental in Antevs' formulation of his tripartite Neother- mal climatic scheme (Antevs, 1948). Another early pollen lo- cality, Tsegi Canyon in northeastern Arizona (Sears, 1937), lay well beyond the Great Basin's southeastern edge. Use of tree ring analysis to provide clues of the Holocene climates driving vegetation changes was still in its descriptive stage, and chro- nologies were still relatively limited in length (Hardman and Reil, 1936; Antevs, 1938; Schulman, 1956). The revolution in reconstruction of vegetation and climate history of arid areas, brought about by the recognition of the potential of packrat middens as a fossil record (Betancourt et al., 1990), was still nearly a decade and a half in the future. Indeed, a half century ago, we had more direct information about Tertiary floras in the 309 310 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Great Basin and adjacent regions than we did about Late Qua- ternary vegetation history (La Motte, 1936; Axelrod, 1939, 1944, 1948; Brown, 1949). Today our much-improved understanding of late Pleistocene terrestrial vegetation history can be usefully compared with the history of Great Basin aquatic and landscape systems, with de- tailed human and faunal records, and, most fundamentally, with global and regional climate reconstructions. The advances in our understanding of Great Basin vegetation history come from three main sources: (1) millennia-long tree ring records, (2) pollen and plant macrofossils recovered from lake, marsh, and dry cave settings, and (3) plant materials obtained from ancient woodrat nests (Figure 1). In addition, studies of plant macro- fossils and tree ring series provide information regarding plant physiology through study of anatomical features (e.g., Van de Water et al., 1994) and stable isotopic signatures (Siegal, 1983; Feng and Epstein, 1994; Van de Water et al., 1994; Wigand, 1995; Jahren, 1996). The potential of future molecular and ge- netic studies of these remains (e.g., Hamrick et al., 1994; Poi- nar et al., 1998) will offer additional information regarding phylogeography and the response of plant populations to changes in past climate. FIGURE 1.?Map of the western United States showing the Great Basin and localities, as discussed in the text, from which tree ring records, pollen and plant macrofossils, and plant materials from ancient woodrat nests were obtained. 1, Wildcat Lake; 2, Craddock Meadow; 3, Summer Lake (Ana River, Bed and Breakfast); 4, Alkali Lake Valley; 5, Chewaucan marsh; 6, Bicycle Pond; 7, Diamond Pond; 8, Warner Lake; 9, Steens Mountain (Fish Lake, Wildhorse Lake); 10, Lava Beds; 11, McCoy Flat; 12, Smoke Creek Desert; 13, Jackson Range; 14, Pyramid Lake; 15, Painted Hills; 16, Lake Tahoe; 17, Little Valley; 18, Lovelock Cave; 19, Carson Sink (Lead Lake); 20, Hidden Cave; 21, Walker Lake; 22, White Mountains (Methusaleh Walk); 23, Alabama Hills; 24, Owens Lake; 25, Little Lake; 26, Warm Sulphur Springs; 27, Searles Lake; 28, Eureka Valley; 29, Cofer Spring; 30, Eleana Range; 31, Lower Pahranagat Lake/Pahranagat Range; 32, Meadow Valley Wash; 33, Specter Range/Spotted Range/Little Skull Mountain; 34, Pintwater Range (Pintwater Cave); 35, Sheep Range; 36, Ash Meadows/Owl Canyon; 3-7, Tule Springs; 38, Gypsum Cave; 39, Sandy Valley; 40, McCullough Range; 41, Owyhee River Valley; 42, Jarbidge Mountains (Mission Cross Bog); 43, Grouse Creek Range (Swallow Shelter); 44, Raft River Mountains (Curelom Cirque); 45, Grays Lake; 46, Swan Lake; 47, Bear River (Cutler Dam formation); 48, Crescent Spring/ Hogup Cave; 49, Lakeside Range (Homestead Cave, Gillespie Hills); 50, Great Salt Lake; 51, Snowbird Bog; 52, Onaqui Range (Devils Gate); 53, Knolls; 54, Silver Island Canyon; 55, Wendover (Danger Cave, Leppy Overhang, Triple Barrel); 56, Pilot Range (Raven Cave); 57, Pequop Range (Icicle Cave); 58, Potato Canyon Bog; 59, Gatecliff Shelter; 60, Ruby Marsh; 61, Cherry Creek Range; 62, Goshute/Toano Range (Top of the Terrace, Marblehead Mine); 63, West Bonneville Basin (Bonneville Estates, Twin Peaks, Pristine Shelter); 64, Indian Farm Canyon; 65, Granite Canyon; 66, Northern Snake Range (Council Hall Cave, Smith Creek Cave, Arch Cave, Ladder Cave, Old Man Cave); 67, Confusion Range; 68, Cricket Mountains; 69, Wah Wah Range; 70, Garrison; 71, Southern Snake Range. NUMBER 33 311 Our intention in this paper is not to summarize everything that we currently know about vegetation history in the region, but rather to highlight major vegetation patterns and changes that reveal underlying climate trends and variability throughout the Great Basin. ENVIRONMENTAL SETTING The Great Basin of western North America is a region of dra- matic topographic relief and striking vegetation transitions over short distances. Lying between the Sierra Nevada/Cas- cade mountain chain on the west and the Wasatch Front on the east, the Great Basin comprises a great mosaic of diverse habi- tats. Great Basin climate reflects the effects of both the hetero- geneous topography and the varying impact of three major air mass systems that intersect over it (Houghton, 1969; Houghton etal., 1975). Pacific: a regime dominated by maritime polar air masses. These moist cool air masses from the Pacific Ocean produce cool, wet winters, and their absence in summer makes the growing sea- son hot and dry. Plant species such as Sierran white fir {Abies concolor van lowiana (Gordon) Lemmon) and whitebark pine (Pinus albicaulis Engelm.) depend upon the reliable winter rainfall that this air mass provides. These species extend into the Great Basin from the west and northwest, dotting mountain tops that lie in the path of the Pacific storm tracks. Gulf: a regime dominated by maritime and continental tropical air masses. These warm, moist air masses entering the southwest- ern United States from either the Gulf of Mexico or the Gulf of California produce hot, moist summers with midsummer tor- rential rainfall (originating during convective storms) and warm, dry winters. Plants, such as creosote bush (Larrea tri- dentata (Sesse & Moc. ex DC.) Cov.), and perhaps singleleaf piiion pine (Pinus monophylla Torn & Fremont), that favor warmer winter temperatures and summer rainfall extend into this area from the southern desert, except in places where win- ter or summer temperatures are too harsh. Polar: a regime dominated by continental polar and arctic air masses. These cold, dry continental air masses from the North American interior extend into the Great Basin from the north- east during the winter. Such air mass system "outbreaks" re- strict the plant species able to grow within the area of greatest influence to those with physiological limits tolerant of cold, dry winters coupled with periodic drought (e.g., limber pine, Pinus flexilis James). Movements of these air systems, and their effects on local and regional climate, are reflected in changes in hydrology, erosion/deposition processes, and vegetation. Displacement of winter and summer storm tracks and the variations in penetra- tion of the summer monsoon are all affected by the realignment of these pressure systems through time. Because movements of these three large air masses reflect changes in global circula- tion and are potentially revealed in the climatic proxy data of the Great Basin, changes in a large portion of the earth's cli- mate system may be monitored through time. In addition to the effects of these air masses, Great Basin cli- mates and vegetation are further complicated by topography. The effect of topographic diversity in the Great Basin on the distribution of precipitation has resulted in the formation of a multiplicity of habitats within close proximity, thus fostering rich biotic diversity (Billings, 1951). Elevational changes in plant distributions can be dramatic in response to the oro- graphic effect of north-south-trending mountain chains that lie across the paths of Pacific storms or, conversely, that funnel summer monsoonal rainstorms northward. Finally, the overrid- ing rain shadow effect of the Sierra Nevada mountains to the west exerts a massive impact upon rainfall distribution in the Great Basin. The rising of these mountains that began during the Pliocene and accelerated during the course of the Pleis- tocene accentuated this effect (Axelrod and Raven, 1985; Axel- rod, 1990; Thompson, 1991). \^GETATION RESPONSE TO CLIMATE CHANGE During the Late Quaternary most changes in vegetation com- munities occurred in response to climate change. Vegetation changes observed in the Great Basin have included large-scale latitudinal shifts of plant taxa that typically occurred in the south and east as well as the primarily altitudinal shifts that characterized the central and northwestern Great Basin. Smaller scale shifts in abundance of major taxa within plant as- sociations have characterized the entire region during both low- and high-magnitude climate changes. Differences in the response of vegetation to variation in climate result from sev- eral factors. Primary among these factors are the magnitude and duration of the new climatic regime to which individual plant species are adjusting. Secondary are the variable effects of topographic relief in the formation of microhabitats, each with its own microclimate. Third in importance is the distance to source areas from which plant species would have to expand during climatic transitions. The climatic sensitivity of vegetation in the Great Basin lies in the physiological characteristics of its species. As a result of millions of years of adaptation to arid and semiarid condi- tions?even before the Great Basin was formed (Tidwell et al., 1972)?many of the plants that grow in this region are opportu- nistic. They have evolved to quickly take advantage of in- creased precipitation. Many species respond to wetter condi- tions through increased pollen and seed production and rapid vegetative growth. Because the annual production of biomass is linearly related to effective annual precipitation (Walter, 1954), biomass production in arid and semiarid environments is a sen- sitive indicator of changing climate. Such changes can be mea- sured directly today through satellite imagery and historically through repeat photography (Hastings and Turner, 1965). In- creased productivity is evidenced in the fossil record by wider tree and shrub growth rings, greater needle and leaf size, and copious production of pollen. Therefore, even relatively brief episodes of greater rainfall that resulted in increased pollen 312 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES production may be evidenced in high-frequency pollen records as sudden rises in pollen influx. Plant adaptations favoring rapid migration are aided by Great Basin topography. Many plants disperse their pollen and seeds with the wind, favoring rapid dispersal of propagules. Seeds of some plants are ingested or held in specialized organs by cer- tain animals, then carried great distances to favorable locations where they may germinate and grow (e.g., Clark's nutcracker (Nucifraga columbiana) and whitebark pine). Because of the steep relief of Great Basin mountain ranges, species can move the equivalent of several degrees of latitude by moving verti- cally only a few hundred meters. The north-south-aligned mountain ranges of the Great Basin also serve as corridors (at various elevations) that facilitate the north-south movement of plant and animal species in response to the vagaries of climate. (Intervening valleys may impede east-west movement between mountain ranges, however.) As a consequence, the potential lag time between climatic input and plant response to reach a new equilibrium may be much less in the Great Basin than it is in areas of reduced topographic relief, where plants must move hundreds of kilometers latitudinally to reach a new environ- mental equilibrium. VEGETATION CHANGE, THE HYDROLOGIC RECORD, AND CYCLES OF EROSION AND DEPOSITION In the Great Basin, rates of erosion and deposition are usu- ally closely tied to climate. The low rainfall and high evapora- tion rates that characterize much of the Great Basin today sup- port a relatively sparse vegetation cover that leaves much of the ground surface exposed to the erosive effects of rainfall. Areas with greater rainfall have relatively greater ground cover offer- ing more protection against erosive processes. Great Basin plant communities respond to increased rainfall through greater foliage production by extant plants and through establishment of extra local plant species. Eventually, increased biomass pro- vides a better buffer to increases in rainfall and reduces the amount of water reaching streams and lakes. Sudden increases in rainfall (either during a single event or during the course of a year) can overwhelm the ability of relatively sparse vegetation cover to assimilate the excess. This is almost immediately re- flected in increased runoff and accumulation of water in the playa lakes, the sinks of Great Basin fluvial systems. If in- creased effective precipitation continues, the increase in the biomass of existing plants and the establishment of new plants providing cover (or protection) for previously exposed ground will eventually reduce the rate of sediment movement. A return to sparser vegetation cover will mark the return to lower effec- tive precipitation and greater erosion potential. Thus, monitor- ing changes in sediment deposition rates in streams and lakes, directly through historical records and prehistorically through paleoecological studies, can provide excellent proxy data of climate-induced vegetation change. THE NATURE OF THE DATA The natural distribution of paleobotanical data sets is con- strained to a large degree by moisture differences within the Great Basin. Low amounts of effective precipitation have fa- vored the preservation of certain kinds of proxy data. Plant macrofossils preserved in dry cave deposits and woodrat mid- dens provide intermittent records of local vegetation spanning tens of thousands of years. Wood from long-dead trees, pre- served by the dry conditions that often characterize upper tree lines in the Great Basin, provides continuous evidence of cli- matic variation spanning the Holocene. A few lakes and desert springs provide long and often complex sedimentary records that can span thousands to hundreds of thousands of years. High organic production in some of these areas has resulted in rapid rates of deposition, and the records are particularly valu- able for the examination of high-frequency pollen and the mac- rofossils of local aquatic plants. These records can be com- pared with tree ring data to generate longer, more detailed records of regional climate and vegetation response. In general, the number of well-preserved palynological records decreases to the south (reflecting the greater rarity of lakes and marshes), and the number of woodrat midden locali- ties decreases to the north (reflecting increasingly wetter condi- tions that favor dissolution of the protective urine coating of the nests). Tree ring series are taken mainly from near the tree line in the higher mountains of the margins of the Great Basin, most commonly in the southern half of the Basin. Fortunately, recent field studies are alleviating some of these problems of data scarcity. Integration of currently available paleobotanical proxy data reveals Great Basin vegetation dynamics in great detail for the last 4 Ka (thousand years ago) and more coarsely for the last 35 Ka. Earlier than 40 Ka, only a few pollen or midden records are available (e.g., from the Summer Lake/Chewaucan Lake basin in the northwest, Owens Lake in the west, Searles Lake in the south, and the Bonneville Basin in the east). For the period af- ter 40 Ka, data obtained from pollen records are supplemented by evidence obtained from ancient radiocarbon-dated woodrat middens. Several high-resolution pollen records in the north- western Great Basin and the northern Mojave Desert provide excellent coverage of both terrestrial and aquatic vegetation history from 6 Ka to the present. Ostracodes and diatoms from the northern Mojave Desert record are being used to recon- struct regional groundwater histories as well. In the Lahontan and Bonneville Basins, vegetation history inferred from mid- dens and pollen records can be correlated with well-dated Late Quaternary lake histories. In general, evidence of the relocation of semiarid and subal- pine woodlands is most obvious for the late Pleistocene/early Holocene transition, especially in the woodrat midden record of the Great Basin and northern Mojave Desert mountain ranges (Van Devender et al., 1987). These records show that during the Pleistocene, woodland species often grew as much as 1000 m (and in a few instances >1000 m) below their NUMBER 33 313 present-day elevation. Substantial shifts in woodland species, however, have occurred during the Holocene as well. During the middle Holocene, some tree species were displaced by per- haps as much as 500 m above where they are found today, es- pecially in the western and northwestern Great Basin. Down- ward displacement of semiarid woodland by as much as 150 m appears to have occurred several times late Holocene. Changes in the distribution of individual shrub species and changes in the composition of shrub communities have also been dramatic during the late Holocene. This is most obvious in the southern Great Basin and the northern Mojave Desert but also has occurred in the northern Great Basin. Changes are most obvious in woodrat midden records of lowland and inter- mediate elevations, but some less obvious changes in shrub communities are also apparent in the few pollen records of the Great Basin. Because the plant communities of the northern and northwestern Great Basin are situated well away from ma- jor vegetation formation boundaries, they do not appear to be as dynamic as do those of southern Nevada that currently lie on the boundary between the Great Basin and the Mojave Desert. However, as the paleobotanical evidence described below will show, this perception is more apparent than real. These various factors have differentially shaped the vegeta- tion history of three broad regions of the Great Basin: (1) the northwest (Lahontan Basin and adjacent mountain ranges), (2) the south (southern Great Basin and northern Mojave Desert), and (3) the northeast (Bonneville Basin and adjacent mountain ranges). Late Quaternary vegetation histories of these three re- gions are discussed separately below. In these discussions, chronological estimates are based on uncalibrated radiocarbon years, unless otherwise noted. Chronological divisions (e.g., the Wisconsin/Holocene break) follow general vegetation and geomorphic patterns within each separate region. Northern and Western Great Basin LATE PLEISTOCENE (-250 TO 12.5 KA) Mifflin and Wheat (1979) indicated that the Great Basin ri- valed the Minnesota of today as a land of lakes during the Pleistocene. The Holocene remnants of these Pleistocene lakes in the northern and western Great Basin were visited by Euro- pean fur trappers, including Peter Skene Ogden, as early as the 1820s (Davies et al., 1961). Several lakes, including Pyramid Lake, were described by J.C. Fremont during his explorations of the West during the 1840s (Fremont, 1845). The first study of pluvial Lake Lahontan was made by I.C. Russell (1885), who described two lake phases indicated by lower and upper lacustral clays. Morrison (1964, 1965, 1991) and Morrison and Frye (1965) further defined the lake's history using such pictur- esque formation names as the Eezta (early Wisconsin or ear- lier), the Sehoo (late Wisconsin), and the Fallon (Holocene). They correlated these lakes with stades of the Wisconsin glaci- ation or other Late Quaternary episodes of cooler, wetter cli- mate. Since then, Davis (1978) and Benson et al. (1987, 1990, 1992, 1997) have further narrowed the chronology and details of lake history. These understandings serve as a backdrop for reconstruction of the vegetation history of the region. PRE-LATE WISCONSIN (>35 KA).?A growing body of re- search in the northern and western Great Basin is clarifying the relationships among past climate history, lake and marsh his- tory, and vegetation change (Mehringer, 1985, 1986). Two long pollen records that include more than just the Holocene?Sum- mer and Owens Lakes (respectively: Cohen et al., 2000; Smith and Bischoff, 1997)?are available from the region, not includ- ing a series of pollen cores from the southern Cascades near Klamath Lake (Adam et al., 1995). Although the proxy record of past climate is far from complete, the general pattern sug- gests that high lakes may be more closely linked to the onsets and declines of glacial episodes in the region than to the glacial maxima. The Summer Lake basin of south-central Oregon currently lies on the transition between the pine woodlands of the eastern slope of the Cascades and the sagebrush (Artemisia spp.) steppe of the northern Great Basin. Three pollen localities from the Summer Lake basin span much of the last 250 Ka (Cohen et al., 2000). The Summer Lake records include the lower two- thirds of oxygen isotope stage (OIS) 6, all of OIS 4 and OIS 3, and the lower portion of OIS 2 as defined by Pisias et al. (1984) and Shackleton and Opdyke (1973). There also seem to be a few samples from the top dating to OIS 1. Shifts in the domi- nance of mixed montane conifers, cold steppe shrubs, and desert scrub species record the varying effects of precipitation and temperature and their changing seasonal influence. Greater abundance of fir (Abies spp.) and spruce (Picea spp.) pollen in a sediment column taken from the Ana River Canyon indicates that climate in the northern Great Basin was more mesic during the OIS 6 section (-170-220 Ka) than during OIS 4 and OIS 2 (Cohen et al., 2000). Unfortunately, an erosional episode (or episodes) removed much of the sediment that accumulated dur- ing the end of OIS 6 and much of OIS 5. Therefore, we do not have a good indication of the nature of vegetation and lake his- tory in the Summer Lake basin during the latter portion of OIS 6 or at the beginning of OIS 5. Indices of drought (based upon a ratio of the pollen of dry- to that of wet-loving plant species) and temperature (based upon spore abundance) reflect much greater climatic variability during OIS 6 than in later stages (Figure 2). After the relatively brief cool, wet episode of OIS 4 (about 70 Ka), OIS 3 climates were relatively cool and dry and apparently comparably stable. The dominance of sagebrush and saltbush (Atriplex spp.) shrub communities during this stage suggests that plant communities on the valley floors were not much different than they are today. The Summer Lake pale- oenvironmental records (Ana River, Wetlands Levee, and Bed and Breakfast localities; see Negrini et al., 2000; Negrini, 2002 (this volume)) are all truncated at about 22 Ka just as condi- tions were becoming cooler and wetter (the beginning of OIS 2 conditions). The nature of effectively wetter climatic episodes differs throughout the record. 314 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 2 Ana River locality (1264 m) J' 0 26 52 78 126.8 0 26 52 78 100 2 240 0 26 52 78 126.8 0 26 52 78 100 2 l< abundance ratios >l l<- o -1 6 18O (o/oo) -2 E2 B2 Bi FIGURE 2.?Comparison of moisture and temperature indices from the Ana River in Summer Lake basin, south- central Oregon (elevation, 1264 m), with the oceanic 18O curve (Martinson et al., 1987). The cool-moist index (or rather, "drought" index) was generated by setting relative abundances of conifer pollen against those of salt- bush and greasewood pollen; values represent a ratio of pollen from the dry-loving species to that of the wet-lov- ing species (moister is to the right on the x-axis). The temperature index is a measurement of the relative abun- dance of spike moss (Selaginella densa) spores (to the right is colder). A combination of radiocarbon and paleomagnetic dating techniques was used to obtain ages (Ka = thousands of years ago). Glacial maxima appear to have been characterized by ex- tremely cold temperatures and relatively little rainfall. The max- ima of the interstades seem to have been typified by warmer temperatures and, in some cases, by substantial drought. The transitions generally were typified by moister conditions and varied only in whether they were warmer or cooler. Records for both Summer Lake and Owens Lake indicate that cool, wet pe- riods resulted in the regional expansion of conifers, but these NUMBER 33 315 expansions occurred during the initial and terminal (transitional) stages of glacial stades. Once the glacial stage approached its climax, temperature (extremely cold conditions) became more important than moisture as a control for vegetation. Therefore, the transitional periods between warm, dry climates and cool, dry climates were the most amenable for the expansion of for- ests and woodlands. Marsh expansion and algae productivity also was greatest during these periods. When lake levels were high, the extent of marsh area in the Summer Lake basin was restricted by the steep gradients of the mountains on the north, west, and south margins. Marsh expansion was most extensive at the shallow north end of the basin during periods of inter- mediate-depth lakes, when the water table would have been close to the surface. Today the northern fringing marsh is artifi- cially maintained by levees constructed by the U.S. Fish and Wildlife Service. The levees result in high water table condi- tions in the northern part of the Summer Lake basin?condi- tions that mimic those of the intermediate-depth lakes found lo- cally during the Pleistocene and the moist periods of the Holocene. Marsh areas during shallow-lake periods would have been severely restricted to areas immediately surrounding stream entries at lake edges or around spring seeps. This climatic pattern also appears to be reflected in the Owens Lake pollen record (Woolfenden, 1993, 1997). A ratio of juniper {Juniperus sp.) pollen to Ambrosia-type pollen re- flects the effect of Late Quaternary climatic conditions on veg- etation; this ration clearly indicates that periods of greater juni- per pollen were warmer and periods of Ambrosia-type pollen were cooler (Woolfenden, 1997, fig. 17). Although wetter than those of today, Owens Lake climates >35 Ka were relatively dry. More closely spaced samples at Owens Lake revealed greater detail in the record, but major transitions to a warmer, drier climate at 120 Ka, and to a cooler, wetter climate at 70 Ka are clear in the records from Owens Lake and the Summer Lake basin. Dugas (1998) recorded a highstand of pluvial Lake Malheur northeast of Summer Lake between 80 and 70 Ka based on the occurrence of water-deposited Pumice Castle tephra. Gehr (1980) dated mollusk shells from a beach ridge in- dicating a highstand of pluvial Lake Malheur at 32 Ka, and Du- gas (1998) presented additional dates that Gehr obtained on shells that confirm this age from other beach ridge positions. In addition, the occurrences of indicator species reflect the warm conditions of the interglacial and even, to some extent, the in- terstades for the last 150 Ka. At Owens Lake, oak (Quercus sp.) and walnut (Juglans sp.) pollen were more common during the last interglacial and the interstades (Woolfenden, 1997, fig. 8). In the record from the Ana River, western hemlock (Tsuga heterophylla (Raf.) Sarg.) is the indicator. PRE-FULL GLACIAL (35 TO 26 KA).?Terrestrial conifer pollen evidence from the Bed and Breakfast and Wetlands Levee localities at Summer Lake (Figure 3) and from woodrat midden strata west of Pyramid Lake indicates that cooler, wet- ter conditions prevailed from 34 to 30 Ka. Haploxylon pines in this area primarily prefer a cooler and/or wetter climate pines whereas diploxylon pines generally prefer a warmer and/or drier climate. This shift was heralded by an expansion of juni- per (Figure 2) and sagebrush. Increased regional abundance of grass together with retreat of saltbush and greasewood (Sarco- batus sp.) are indicated (Figure 3). Slight differences between the Wetlands Levee and the Bed and Breakfast records may re- flect differences in both sample spacing and distance from shore within the basin. Dramatically reduced productivity of the eutrophically adapted algae Botryococcus at the Bed and Breakfast locality, indicates colder water temperatures during this period (Figure 3). A return to drier conditions at ca. 29.6 Ka is indicated by a decrease in fir and spruce and an increase in saltbush (Cohen et al., 2000). A major drought from 28 to 26 Ka is flanked by regionally correlatable grass expansions in the woodrat midden pollen record at Pyramid Lake as well as in the sediments of Summer Lake (Figures 3, 4). This drought is evidenced by saltbush expansion and is reflected in lowered lake levels throughout the northwestern Great Basin (Negrini and Davis, 1992). In the north-central Great Basin, subalpine and/or tundra species of juniper (prostrate or creeping juniper, Juniperus hor- izontalis Moench, and common juniper, J. communis L.) ap- pear in a woodrat midden from the Owyhee River Valley, dat- ing from about 27 Ka (Wells, 1983). Their appearance in a woodrat midden dating to this period of drought (Benson et al., 1990; Negrini and Davis, 1992) suggests that the period may have been not only dry but cold as well. The pollen records from Summer Lake and McCoy Flat (a valley floor site along Pine Creek, west of Eagle Lake on the eastern fringe of northern California's Modoc Plateau), and the woodrat midden records from around the pluvial Lake Lahon- tan basin, indicate that juniper occasionally was locally abun- dant between 50 and 12 Ka in the northwestern Great Basin (Figures 5, 6). This occasional juniper pollen, which differs somewhat in appearance from juniper pollen typically found in the sediments of Summer Lake and McCoy Flat during most of the Late Quaternary, might be that of common and/or prostrate juniper. This might indicate colder, more tundra-like or alpine- like conditions. Woodlands dominated by Utah juniper (/. osteosperma (Torr.) Little), as evidenced by the presence of Utah juniper twiglets dating to ~34 Ka in woodrat middens, were growing at least as far north as the shores of pluvial Lake Lahontan (Wi- gand and Nowak, 1992; Nowak et al., 1994a, 1994b). Compar- ison of juniper with other species in the pollen and macrofossil records from the region indicates that juniper expansions oc- curred during cool, moist periods of climate before and after the glacial maxima. LAST GLACIAL CYCLE (26 TO 12.5 KA).?Effectively moister climate after 24.5 Ka, coincident with renewed growth of pluvial Lake Lahontan, encouraged the spread of Utah juni- per woodlands and allowed whitebark pine to intrude into scat- tered localities as low as 1380 m in elevation (Wigand and No- wak, 1992). Slopes currently dominated by saltbush scrub were 316 Wetlands Levee locality west-central Summer Lake basin 100 38.7 0 40 80 111.4 0 Relative abundance ratios 20.2 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Bed and Breakfast locality southwestern Summer Lake basin i> / S\ l FIGURE 3.?Ratios of the abundances (x-axis) of selected major pollen types from the Wetlands Levee and Bed and Breakfast localities in the Summer Lake basin, south-central Oregon, by date (years before present, B.P.). Increases in the ratio of (fir + spruce + haploxylon (Hap) pine)/diploxylon (Dip) pine indicates cooler, moister conditions. Increases in the ratio of (juniper + sagebrush)/Dip pine herald the expansion of juniper woodland. Increased regional moisture is indicated with increases in the ratio of grass/(sal tbush + greasewood). Dramati- cally reduced productivity of the algae Botryococcus is revealed by plotting the inverse of its abundance, 1 IBot- ryococcus. All ratios indicate moister conditions to the right and drier conditions to the left. characterized then by sagebrush steppe and by occasional pe- rennial seeps lush with fewflower spikerush (Eleocharis quin- queflora (F.X. Hartmann) Schwarz) (Wigand and Nowak, 1992). Utah juniper macrofossils that were found in a woodrat midden dating to 20,460 ? 990 years B.P. (before present; labo- ratory no. Beta-41919), located just above the highest shoreline of pluvial Lake Lahontan on the western margin of the Carson Sink, reflect the peak of pre-glacial maximum expansion of ju- niper well beyond its current limits (Wigand and Nowak, 1992). This period corresponds to the rapid rise of pluvial Lake Lahon- tan at the beginning of the last glacial cycle (Benson et al., 1997, fig. 4). By 20 Ka the onset of much cooler, drier glacial maximum conditions led to a decline in Utah juniper, disap- pearance of whitebark pine from the midden record, and expan- sion of mixed sagebrush and shadscale (Atriplex confertifolia Torr. & Fremont) desert scrub communities (Nowak et al., 1994a). Approximately 23 and 12 Ka, during the wetter periods leading up to and following but not during the glacial maxi- mum, whitebark pine occurred along the western margin of the Lahontan Basin, more than 1000 m lower than its current eleva- NUMBER 33 317 Little Skull Mountain Painted Hills Bed and Breakfast Core 6.6 0 5 10 20 23.2 10 20 30 40 50 % of Total Terrestrial Pollen minus Pine FIGURE 4.?Comparison of Late Quaternary relative abundances (as percentages) of grasses (Poaceae) in lacus- trine core sediment from the Bed and Breakfast locality in Summer Lake basin, Oregon, with the relative grass percentages from the woodrat midden pollen records obtained from Painted Hills west of Pyramid Lake, west- central Nevada, and from Little Skull Mountain in the northern Mojave Desert, southern Nevada, by date (years before present, B.P.). The pollen record from Pyramid Lake was obtained from a single stratified woodrat mid- den, whereas the Little Skull Mountain record was from several adjacent woodrat middens. tion (Wigand and Nowak, 1992; Nowak et al., 1994a, 1994b). This provides an indication of just how cold the glacial maxi- mum may have been in northwestern Nevada. Estimates of the temperature difference for the onset and decline of the glacial maximum, based upon the current temperature limits of white- bark pine in the area today, indicate that the mean annual tem- perature (MAT) was at least 7? C colder during the onset and decline of the glacial maximum. Considering that woodland in this region disappeared during the glacial maximum because conditions were so cold, the MAT may have been l?-2? C colder still. The harsh (much colder and drier) conditions during the gla- cial maximum around 18 Ka also may have severely restricted semiarid woodland. Juniper even may have been locally or re- gionally extinct in the northwestern Great Basin. Utah juniper appears to have been present before and after the glacial maxi- mum along the western shore of pluvial Lake Lahontan (Wi- gand and Nowak, 1992; Nowak et al., 1994a, 1994b). After the glacial maximum, woodrat midden data indicate that western juniper (Juniperus occidentalis Hook.) was present in places along the shores downwind of pluvial Lake Lahontan, where it may benefitted from "lake effect" (Thompson et al., 1986). Absence of woodrat midden strata and the scarcity of pollen records dating to the glacial maximum preclude making defi- nite inferences regarding the distribution of juniper on the landscape. Greater shrub dominance, however, in both the 318 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES tIS O JS1) O g a 3 1 fi .8 2 * ?III a^ vi 6 gl e S I g 5.(/> 3 1 5 tive . U0^ D Lak a me o ? liptio 1 * E .2 < 96V NUMBER 33 319 8- -t*?? i ? -??? J - ' - 1^: L_l__f* ? i ??i???^ ?*??! m m m ?? 4 O S1 a ??? (LJUO) 320 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES pollen and macrofossil samples from woodrat midden strata immediately before and after the glacial maximum suggests that juniper may have been relatively scarce, probably occupy- ing protected locations. On the other hand, shrub communities dominated by sagebrush, and at times saltbush, characterized extensive portions of the northwestern Great Basin (Wigand and Nowak, 1992). These data suggest that it was either too dry or too cold (or both) for juniper woodland to survive in the northern Great Basin during the last glacial maximum. Relatively warmer, wetter conditions after the glacial maxi- mum led to dramatic increases in sagebrush-dominated commu- nities, and limited reexpansion of whitebark pine as well (Wi- gand and Nowak, 1992; Nowak et al., 1994a). Findings of Utah juniper in woodrat middens dated to 12,260 ? 140 years B.P. (Beta-47203) and 11,980 ?110 years B.P. (Beta-47204) at the highest strandline on the eastern shore of the Carson Sink reflect reexpansion of semiarid woodland after the glacial maximum. By 12.5 Ka (concurrent with the final late-Pleistocene high- stands of pluvial Lake Lahontan (Benson et al., 1990, 1992); see Negrini, 2002) mesic plant communities, reflecting lake effect and orographic forcing of rainfall along the eastern shore of plu- vial Lake Lahontan, contrasted sharply with more xeric plant communities to the west (Wigand and Nowak, 1992). The pollen record from a large woodrat midden west of Pyra- mid Lake that spans the last 34 Ky (thousand years) suggests substantial differences between climates at the onset (~24 Ka) and at the end (~12 Ka) of the last glacial cycle (Wigand and Nowak, 1992; Figure 7). Despite the expansion of whitebark pine during both periods (evidenced by the presence of macro- fossils in the midden), the substantially greater abundance of sagebrush pollen and the appearance of mountain mahogany (Cercocarpus sp.) in the macrofossil record of the latter period suggest cooler temperatures and less precipitation 12.5 Ka than -24 Ka. In part this could also reflect greater variability in cli- mate, as suggested by the deep-water tufa record from Pyramid Lake (Pelican Point) (Benson et al., 1995; Benson et al., 1996). Painted Hills woodrat midden (1380 m) 0 40 89.4 0 40 97.1 0 55.5 4.5 0 32.1 22.9 0 20 65.1 24.4 6.6 0 41.4 8.1 35000 \ I i n i r i i i i 0 40 89.4 0 40 97.1 0 i I 55.5 4.5 0 32.1 i i n 22.9 0 20 65.1 24.4 6.6 0 41.4 8.1 % terrestrial pollen sum ? = or < 2 % % terrestrial pollen sum minus pine & juniper FIGURE 7.?Relative percentage diagram of major pollen types from a single stratified woodrat midden near Pyramid Lake, western Nevada (elevation, 1380 m). Dates are radiocarbon-based. A bullet (?) indicates a value of<2%. NUMBER 33 321 LATEST PLEISTOCENE/EARLY HOLOCENE (12.5 TO 8 KA) This period is characterized by often incredibly diverse plant communities that contained species typical of the cooler condi- tions of the glacial period along with pioneering plant species that heralded the Holocene. Mehringer (1985) noted several characteristics that tied many of these plant communities to- gether during this time, especially at the higher elevations and in the more xeric areas of the northern Great Basin. This in- cludes (1) a treeless sagebrush steppe, often with abundant grass; (2) common occurrence of russet buffaloberry (Shep- herdia canadensis (L.) Nutt), occasionally juniper (probably common juniper), and more rarely spruce and fir; (3) combina- tions of pollen types such as Rumex-Oxyria, Bistorta, Polem- onium, Eriogonum, and Koenegia that indicate subalpine or al- pine plant communities; and (4) abundance of the spores of cold-climate plants such as Selaginella densa, S. selaginoides, Botrychium, and Lycopodium. For example, the pollen record from 2250 m at Fish Lake near Steens Mountain, Oregon, re- flects the substantially colder conditions of the transition from the late Pleistocene to the early Holocene. The pollen record is dominated by sagebrush {Artemisia) and grass (Poaceae), which are representative of a cool, moist steppe (Mehringer, 1985). It also contains juniper pollen, which may signal the movement of common juniper from its lower elevation distri- bution during the Pleistocene to its new Holocene position high atop the mountain (Mehringer, 1985). By 9.7 Ka the last glacial ice remnants disappeared from the high, southeast-facing cir- que basin near the top of Steens Mountain that became Wild- horse Lake (2560 m). Common juniper was probably estab- lished in the surrounding subalpine grassland that had replaced the snow and ice of the Pleistocene (Mehringer, 1985, 1986). The departure of ice from this basin and initiation of the lake may record, in part, the impact of the early Holocene thermal maximum. The 9000-year-long pollen record from Bicycle Pond in the Warner Valley, south-central Oregon, indicates that the current lower limit of western juniper on the west-facing slope of Hart Mountain was probably dominated by a grassy, sagebrush steppe 9 Ka (Figure 8). Juniper, though present then, was less abundant than today. Although scattered pine trees may have occurred in nearby canyons, pine woodland probably lay where it does today: around the tops of the mountains lying west of the Warner Valley and in the west-facing canyons and on the east slope of Hart Mountain. The aquatic pollen record sug- gests that throughout the Pleistocene, Bicycle Pond was fed by a seasonal stream and springs that flowed from the layered ba- salts that composed the mountain above it. Bicycle Pond is cra- dled by a steep-sided, Pleistocene landslide basin that restricts littoral plant growth to a narrow margin. The well-drained slopes behind the littoral margin support a grassy, sagebrush steppe, which extends a couple of hundred meters to the floor of the Warner Valley. Grass pollen values that were higher in the early Holocene than today reflect a wetter environment then; and abundant sedge pollen (Cyperaceae) from that period indicates that the marsh around Bicycle Pond was much more extensive then than it is today and reflects substantial early Ho- locene spring activity. At lower elevations, the early Holocene expansion of sage- brush steppe vegetation is matched by expansion of saltbush- dominated desert scrub vegetation before the fall of Mazama ash (see Chenopodiineae values in Figure 8). The climax of saltbush scrub vegetation before and after the fall of Mazama ash is clearly evident in the pollen records from the Warner and Chewaucan valleys analyzed by Hansen (1947). Analysis of pollen from sediments in the Alkali Lake Valley in south-cen- tral Oregon, undertaken by one of the authors (Wigand), indi- cates that marshes (abundant sedge pollen) were present in the valley floor until ~9.5 Ka but disappeared soon after to be re- placed by saltbush-dominated shrub communities. Gehr (1980) and Dugas (1998) recorded a series of radiocarbon dates be- tween 9.6 and 7.4 Ka on mollusks from beach ridges and char- coal that indicated multiple pluvial highstands of Lake Malheur including some within five meters of its overflow into the Snake River drainage at Malheur Gap. The latter of these lakes may have been considerably lower. Mehringer and Cannon (1994) recorded a wet episode in the Fort Rock Basin, dating to around the fall of Mazama ash. In any case, generally drier conditions resulted in the establishment of saltbush communi- ties by the time that Mazama volcanic ash fell in the region. The pollen record from Fish Lake indicates that sagebrush steppe had expanded, becoming denser during the climatic warming of the early Holocene on Steens Mountain (Meh- ringer, 1985, 1986). This climaxed shortly before the fall of Mazama volcanic ash. Further south in northwestern Nevada, the persistence of Utah juniper during the early Holocene at lower elevations (be- low where it occurs today) ended ~9.5 Ka. Radiocarbon (14C) dates for Utah juniper of 9640 ?110 years B.P. (Beta-46256) and 9700 ? 90 years B.P. (Beta-46255) record the last time that it appeared during the early Holocene beyond its current distri- bution west of the Lahontan Basin. These dates correspond di- rectly with the last dates on lithoid tufa that Benson et al. (1995) obtained from Pyramid Lake (Pelican Point), suggesting that retreat of juniper and dramatically lowered lake levels are contemporaneous and coincide with high northern hemisphere solar radiation during the early Holocene (Kutzbach and Guetter, 1986; Kutzbach et al., 1993). In areas more distant from the current margin of the semiarid woodland, juniper may have disappeared as much as a thousand years earlier. Radiocarbon dates for Utah juniper from woodrat middens on the southwest margin of the Smoke Creek Desert, north- western Lahontan Basin, suggest that it disappeared from that area by 10.3 Ka. A cluster of these dates ranging from 10,830 ? 100 years B.P. (Beta-47200) to 10,340 ? 210 years B.P. (Beta-35814) are the last from that area until a late-Ho- locene reexpansion of juniper. As juniper was disappearing from areas it occupied during the Pleistocene, it was expanding northward into new areas. The 322 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES '#%*"*-&&?> .a s 2 2 8. a -?asIS- l-c O ?CB u U |JS1 * i -S o ?^ u H ill .2 O is111 2 ? ? U. OH O NUMBER 33 323 presence of semiarid juniper woodland in the northwestern Great Basin by ~8.5 Ka is evidenced by juniper pollen values in pollen records from McCoy Flat in northern California and from Bicycle Pond on the west slope of Hart Mountain in the Warner Valley, south-central Oregon (Figures 6, 8). This is fur- ther corroborated by directly dated Utah juniper twigs from woodrat middens in the Jackson Range of northern Nevada with a date of 8490 ? 90 14C years B.P. (Beta-64360). Although we know from the pollen records mentioned above that juniper oc- curred in south-central Oregon, no macrofossils have been re- covered that indicate the species. This is the earliest species identification for the area. The pollen record from Hidden Cave in the Carson Sink also reflects the early Holocene decline of woodland and expansion of desert shrub vegetation at the expense of sagebrush steppe (Wigand and Mehringer, 1985). Although the high lakes had disappeared from the Lahontan Basin by 11.0 Ka, cattail (Typha) pollen in Hidden Cave sediments indicates that exten- sive marshes remained until about 9.6 Ka (Wigand and Meh- ringer, 1985, fig. 36). The pollen record indicates that dramatic regional decline in pine and sagebrush and equally dramatic in- creases in greasewood characterized this period. The ongoing effects of regional drought at low elevations are reflected in the fourfold increase in Chenopodiineae (primarily saltbush) pollen in Hidden Cave sediments by ~9.0-8.0 Ka, levels that would characterize the remainder of the Holocene. Additional evi- dence of desiccation of the lake basin is indicated in analysis of the sediments from the cave. Statistical parameters of grain size analysis of Hidden Cave sediments conducted by P.J. Mehringer, Jr. (see Wigand and Mehringer, 1985), in the late 1970s indicate that between 7.5 and 5.5 Ka the energy of colluvial processes was declining as the energy of early Holocene storms was waning (Figure 9). How- ever, eolian processes (indicated by more-positive skewness of sediments) were increasingly more active during this period of transition in the Lahontan Basin (Figure 9). Although the collu- vial contribution to transportable sediment was declining during this transitional period, the regional reduction in plant cover Hidden Cave locality (1244 m) -i?i?i?i?'?i?i?i?i?i?i?i?i?i?|?i?i?i?i?i?i?i?i?i?|?i?i?i?i?i?i?i?i?i?:?i?i?i?i i i i r 4000 6000 8000 100002000 Radiocarbon age (yr B.P.) FIGURE 9.?Wet-to-dry index, (sagebrush + grass pollen)/(saltbush + greasewood pollen), generated with sedi- ment pollen data from Hidden Cave in the Carson Sink, Nevada, compared with the skewness and sorting of Hid- den Cave sediments; the latter two are statistical parameters derived from grain-size analysis and can be related to understandings of erosional and depositional processes. Skewness is used as a measure of eolian activity, and sorting is used as a measure of the amount of energy in the medium that moved sediment. "Up" on the drought index indicates wetter climate, up on the skewness index indicates greater eolian activity, and up on the sorting index indicates greater energy in the medium that was moving sediment. Sample dates were based on deposition rates extrapolated from radiocarbon dates (Davis, 1985). 324 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES density in response to drought was exposing large areas to eolian erosion (Figure 9). The pollen record from Lovelock Cave (By- rne et al., 1979) corroborates these trends. The general corre- spondence between the Hidden Cave and Lovelock Cave pollen records is close. Preliminary pollen analyses of human fecal material, re- moved from the 9400-year-old Spirit Cave burial site in the Carson Sink of west-central Nevada, corroborate the environ- mental reconstruction (Wigand, 1997a). (The cave contained the partially mummified body of a man, the Spirit Cave Man, 5'2" tall and about 40-50 years old.) The fecal material was composed almost entirely of the remains of small fish (Eiselt, 1997). However, a very low pollen background in the samples may have mirrored the regional pollen rain. The pollen may have found its way into the fecal material through residues present on plants eaten by the early Nevadan or may have en- tered through the stomachs of the fingerlings that were in- gested. Fish may have eaten the pollen as it floated on the wa- ter surface of the marsh during flowering time (suggested by PJ. Mehringer, Jr.). This pollen record is statistically identical to pollen of the same age recovered from nearby Hidden Cave (Wigand, 1997a). It reflects a saltbush-dominated environment with evidence for persistent marshes and eolian habitats. Com- parison of the drought index from Hidden Cave with one pro- duced from the Hogup Cave pollen record analyzed by Kelso (1970) indicates that this was a regional pattern at low eleva- tions across the northern Great Basin (Figure 10). Today the McCoy Flat area comprises a mosaic of mixed-co- nifer-forested, Late-Tertiary/early Pleistocene volcanic cones that rise above a lattice of wide, low-grade stream valleys dom- inated by sagebrush steppe. The pollen record from McCoy Flat indicates that by the early Holocene a regional decline in pine and juniper woodland corresponded to an increase in steppe dominated by sagebrush and rabbitbrush (Chrysotham- nus-type pollen) (Figure 6). An increase in grass values at this time, corresponding to the increase seen at Bicycle Pond and in - 6 - 4 | CO CO (AE O (A -QCD O) CO CO - 0 I I I I I I | I I I I I I I I I | I I I I I I 2000 4000 i i I i i i ii i i i 6000 8000 Radiocarbon age (yr B.R) i i 10000 FIGURE 10.?A comparison of two drought indices generated from pollen analyses of sediments obtained from Hidden Cave in the Lahontan Basin, Nevada (Wigand and Mehringer, 1985), and Hogup Cave in the Bonneville Basin, Utah (Kelso, 1970). Up indicates wetter climate. Differences between the two records primarily reflect sample spacing and error introduced through extrapolation of sample ages from radiocarbon-based dates. NUMBER 33 325 the midden pollen record from Pyramid Lake, may reflect a re- gional response to increased rainfall in late spring or early sum- mer, an increase also hinted at in vegetation records further south and east in the Great Basin. Increased cattail pollen in the record may simply reflect decreased competence of the local stream, which was now filling the deep channel cut that had characterized Pine Creek during the Pleistocene. MIDDLE HOLOCENE (8 TO 5.5 KA) The McCoy Flat and Bicycle Pond records indicate that by the beginning of the middle Holocene the retreat of juniper woodland and the expansion of sagebrush steppe had reached a climax (Figures 6, 8). Both of these records also indicate that saltbushes (Chenopodiineae pollen) were regionally more im- portant than they had been before. The decline of grass in the McCoy Flat record reflects the middle-Holocene drought. At Steens Mountain, the Fish Lake pollen record indicates that sagebrush steppe became the dominant vegetation at eleva- tions lying above the juniper woodland zone (Mehringer, 1985). Sagebrush pollen to grass pollen ratios from Fish Lake indicate that the climate that typified the period between 8.8 and 5.6 Ka was the driest to occur there during the Holocene (Mehringer, 1986). The Wildhorse Lake record evidences the arrival of sagebrush steppe by ~8 Ka (Mehringer, 1986). It was to remain dominant at these elevations until about 3.8 Ka. In southern Oregon, dry pluvial lake basins are indicated by dune deposits dated to this period, including ones in the Catlow Val- ley at Skull Creek (Mehringer and Wigand, 1986), Fort Rock Valley (Mehringer and Cannon, 1994), and the Malheur Valley (Dugas, 1998). The all-pervasive drought of the middle Holocene is re- flected in the dearth of paleobotanical records throughout the northern Great Basin at low elevations and especially around the Lahontan Basin. The almost continuous midden record that exists in the Painted Hills on the west shore of Pyramid Lake is interrupted for almost 4000 years, from ~8 to 4 Ka (Wigand and Nowak, 1992). Juniper macrofossils disappear from wood- rat middens lying 200 m above the early and late-Holocene lower elevational limits of juniper evidenced in the woodrat midden record. This suggests that juniper retreated upward in elevation by at least 200 m or more above its current distribu- tion during the middle Holocene. The Hidden Cave pollen record indicates dominance of salt- bush shrub communities in the margins of the Lahontan Basin (Wigand and Mehringer, 1985). Statistical parameters gener- ated from the grain size analysis of Hidden Cave sediments in- dicate that transport of sediment through eolian (less positive skewness) and colluvial (better sorting) processes in the Carson Sink had declined substantially (Figure 9). The near absence of evidence of high-energy colluvial transport may have resulted from the rarity of high-energy storms in the Lahontan Basin. Reduced colluvial transport despite the reduced vegetation cover indicates that rainfall was at a minimum, especially dur- ing the summer. The dramatic reduction in eolian activity may be attributable to the depletion of transportable materials in the basin. Because colluvial and fluvial activity was at a minimum, there was a reduction in the material that would have been available for eolian transport. A comparison of the Hidden Cave drought index with one generated from the pollen record of Kelso (1970) from Hogup Cave in northwestern Utah sug- gests that widespread drought characterized much of the north- ern Great Basin (Figure 10). Pyramid Lake may have reached its lowest level at this time (Born, 1972). Woodrat midden evidence from the White Mountains on the California-Nevada border indicates that middle-Holocene drought displaced semiarid woodlands upward in elevation by as much as 300-500 m (Jennings and Elliot-Fisk, 1993). Taken together with the evidence from the Lahontan Basin, it appears that this displacement may have characterized large portions of the western Great Basin (Wigand and Nowak, 1992). Upward woodland displacement in this area coincides with the age of drowned trees found 10-15 m below the modern sur- face of Lake Tahoe. These trees became established in the Tahoe Basin during periods of intense, long-duration drought -6.5 Ka (Lindstrom, 1990). Throughout the northern Great Ba- sin lakes and marshes desiccated and pollen of drought-tolerant salt desert species increased substantially (Wigand, 1987; Meh- ringer and Wigand, 1990). At Diamond Pond south of Lake Malheur in south-central Oregon, greasewood pollen reached levels of 80% of the terrestrial pollen, whereas currently it makes up less than ~20% of the modern pollen record. These changes coincided with dramatically reduced evidence of the activities of Native American populations (Grayson, 1993). This is consistent with a decline in foraging resources and cor- roborates the severity of this drought (Aikens, 1986). EARLY LATE HOLOCENE (5.5 TO 2 KA) Beginning ~5.4 Ka, the extreme drought of the middle Ho- locene came to an end. Temperatures remained warm, but pol- len and macrofossil data indicate gradually increasing annual precipitation punctuated by periodic increases in rainfall abun- dance (Davis, 1982; Mehringer, 1986; Wigand, 1987). A brief, extremely dramatic climatic event that can be correlated from the Columbia Plateau of eastern Washington to the spring de- posits of Ash Meadows in the northern Mojave Desert effec- tively signals the end of the middle Holocene in the intermoun- tain West. Evidence for this panregional event begins in the north at Wildcat Lake in eastern Washington. At 5.4 Ka, the deposition of reworked Mazama ash in the lake sediments suddenly ended (Peter J. Mehringer, Jr., pers. comm. to Wigand, 1978). This in- dicates that vegetation density on the landscape became great enough to prevent the resuspension of the volcanic ash either by wind or water. This increase in vegetation density probably directly reflects a dramatic shift toward greater rainfall. 326 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES At Diamond Pond in south-central Oregon there was a shift from dry-climate pollen assemblages to much more mesic as- semblages (Wigand, 1987). Just after 5.4 Ka, pollen spectra that previously had been dominated by as much as 80% grease- wood pollen were instead characterized within a few decades by as much as 40% sagebrush pollen. Although a shift back to- ward drier climate occurred during the next 1400 years, the conditions that characterized the period before 5.4 Ka never re- occur. This event is also noted in the Tahoe Basin on the west-cen- tral margin of the Great Basin, where montane forests were so rapidly submerged beneath the cold waters of Lake Tahoe at 5.5 Ka that they had no opportunity to decay (Martin Rose, pers. comm. to Wigand, 1995). An 8000-year-long tree ring re- cord of a bristlecone pine (Pinus longaeva D.K. Bailey) from Methuselah Walk in the White Mountains reveals this same rapid shift in climate that ended a period of drought that may have lasted as long as 1500 years (Graybill et al., 1994). The paleolimnological record from Walker Lake, the southernmost subbasin of ancient pluvial Lake Lahontan, indicates that the Walker Lake basin, which had been a dry playa, was refilled by -4.7 Ka (Bradbury et al., 1989). This probably resulted from wetter conditions at this time, but it also may have been caused by a shift in the course of the Walker River from the Carson Basin to the Walker Basin. This shift, however, also may have resulted from changes in the Walker River alluvial process pre- cipitated by a change in climate. The shift to wetter conditions at ~5.5 Ka had other ramifica- tions as well. It initiated the reexpansion of woodlands and montane forests. The earliest evidence of western juniper re- ported in its historic range of northeastern California and east- ern Oregon are twigs from ancient woodrat middens at Lava Beds, California (5.4 Ka), and Diamond Craters, Oregon (4 Ka) (Mehringer and Wigand, 1987, 1990). Just east of the Tahoe Basin, a 3.5 meter core (unpublished) recovered from a bog north of Little Valley, Nevada, revealed responses of both local montane forest and regional semiarid woodland and sagebrush steppe to climate change since ~6.3 Ka, when the pond formed (Figure 11). Little Valley is on the east slope of the Sierra Nevada near the lower edge of the mon- tane forest. Across the valley to the east lie the pinon- and juni- per-dominated ranges of the Great Basin. Cluster analysis of the Little Valley core pollen reflects the shift from dry middle- Holocene conditions to more mesic late-Holocene conditions. Local sagebrush predominance between 6.3 and 5.6 Ka (Zone I) indicates relatively more xeric conditions than any of the succeeding periods. A retreat of sagebrush and an increase in pine and juniper initiated a period of more mesic climate be- tween 5.6 and 4 Ka (Zone II) (Figure 11). Sedges probably dominated the margins of the pond during both periods. Two more mesic periods occurred between 5.6 and 4 Ka, when fir may have first expanded into the valley from above. This pe- riod corresponds with the drowning of low-lying trees in the Tahoe Basin by rising lake levels. Between 4 and 1.9 Ka (Zone III), major downward expansions of fir (both red fir, Abies magnified A. Murr., and white fir, A. concolor (Gordon & Glend.) Lindl. ex Hildebr.) into Little Valley occurred (Figure 11). Abundant sedge pollen at this time records filling of the pond and generation of sedge peat. In addition, alder (Alnus), birch (Betula), and willow (Salix) pollen evidences the estab- lishment of these taxa in mesic communities in Little Valley. This period was clearly cooler and more mesic than those that preceded or followed it. Local fire severity seems to have in- creased substantially between 5.6 and 4 Ka, perhaps in re- sponse to the greater availability of fuel. After 3 Ka, fire sever- ity appears to have declined substantially. Expansions of juniper woodland into areas at elevations lower than where it is currently found in the northern Great Ba- sin are recorded in the Diamond Pond juniper pollen record and by western juniper macrofossils from woodrat middens in Dia- mond Craters east of Diamond Pond (Wigand, 1987; Meh- ringer and Wigand, 1990). Expansion of western juniper wood- land into the lower elevation sagebrush steppe began about 4.5 Ka (Wigand, 1987). The rate of juniper woodland expansion increased dramatically -3.7 Ka and again 2.7 and 2.3 Ka. Ex- cept for two lapses, western juniper woodland remained near this lower elevational limit (~150 m below its current extent) until about 1.9 Ka (Wigand, 1987; Mehringer and Wigand, 1990). Broad Neopluvial expansion of pine in the woodlands on the northernmost boundary of the Great Basin is recorded in the Craddock Meadow, Oregon, pollen record (Wigand, 1989). Further south a pollen record from Potato Canyon Bog in Grass Valley, central Nevada, also shows evidence of expansion of semiarid woodland during this period (Madsen, 1985). Three episodes (~3.7, 2.7, and ~2.3 Ka) of fir expansion at Little Val- ley (Figure 11) mirror the episodes of moister climate revealed at Diamond Pond as well. Together these data reflect a pattern of forest and woodland expansion that occurred throughout the entire northern and western Great Basin. These episodes are characterized not only by the expansion of both forests and woodlands, but also by a regional rise in the water table and a reexpansion of lakes and marshes (Wigand, 1987; Grayson, 1993). Even the record from Walker Lake indi- cates deep lake levels between 4 and 2.9 Ka (Bradbury et al., 1989, fig. 11). Although lake levels had fallen by 2.2 Ka, the absence of dated samples between 2.9 and 2.2 Ka leaves lake level elevation during that period in question. In general, cli- mates in the northern Great Basin during the Neoglacial or Neopluvial from 4 to 2 Ka were cooler and substantially wetter, with winter precipitation dramatically increased relative to summer precipitation (Davis, 1982; Wigand, 1987). Although macrofossils indicate the presence of western juni- per at lower elevations than its current distribution, juniper pol- len values from Diamond Pond indicate that juniper woodlands at the lower elevations during this Neopluvial expansion were less dense than those currently found in the area (Wigand, 1987). Recent analyses of 813C in western juniper twigs and Utah juniper twigs from ancient woodrat middens indicated that water stress?either as a result of drought or frequent freezing NUMBER 33 327 w.(*. CO CO o _CD3 1.8 14.4 3 " -i00 i 1 ? 9.5 35.7 ( ?-!? *H- .? ?, ? ? 1 1 1 1 ??i *-?i 1 *!?* A. ? k ??^??-?i?*?i? ?i 1 1 r ?, , , r F--. r , 1 1 1 1 ?! P* f ""1 a. m| ?, r ? ? f ?^ s? v O .Sj CM go 35 KA) Profile I at Tule Springs apparently contains pollen spectra dating to before 35 Ka (Mehringer, 1967, fig. 31). The profile documents a cool, wet event dating to more than 37 Ka. This may actually be one of the two periods of wetter climate indi- cated in the Browns Room record dating to either ~40 or ~45 Ka, or it may be an earlier event (Szabo et al., 1994, fig. 4). The episode is characterized by both pine and fir expansion. Neither the age nor the duration of this event can be determined, but it seems to document the same kind of vegetation pattern that is seen in later pluvial episodes. The pollen record indicates that desert shrub communities dominated by saltbushes character- ized low elevations. Juniper and sagebrush pollen suggests that juniper woodland, possibly mixed with pine, was also com- mon. As cooling progressed, the contribution of pine pollen to the record became overwhelming. This resulted, in part, from the increased area that was available for pine to expand into as its habitat was displaced down the mountainsides (see Meh- ringer's (1967) discussion of the impact upon pollen abun- dance by the displacement of species along the area elevation curve). Fir pollen became more abundant than that of juniper. This may have resulted from poor preservation of juniper pol- len, but values of about seven percent suggest a substantial in- crease in fir regionally. Pollen from juniper and sagebrush indi- cates that juniper woodland was still widespread. The decrease in pine and fir values and the disappearance of juniper near the top of the profile seem to signal a return to a drier climate. This pattern is repeated throughout the late Pleistocene in the north- ern Mojave Desert. Unfortunately, most woodrat middens dated to before 35 Ka have multiple dates and reflect substantial mixing of materials of many ages. Typical of this problem are middens from the Eleana Range (Spaulding, 1985). This mixing may be an indi- cation that climatic conditions since the initial formation of these woodrat middens have been wet enough to cause re-solu- tion of the crystallized urine cementing them. LATE PLEISTOCENE (-35 TO 12.5 KA) Woodrat midden studies during the last 30 years have en- abled identification to the species level of the likely contribu- tors of both pine and fir pollen from the Tule Springs pollen re- cord. It is clear that in the southern Great Basin and the northern Mojave Desert, expansions and contractions of limber pine and white fir (possibly Rocky Mountain white fir, Abies concolor var. concolor) record periods of wetter climate during the cooler, drier spans of the late Pleistocene (Figures 15, 16). Originally recorded in the pollen record obtained from Tule Springs (Mehringer, 1967), they have been corroborated and refined by the substantial woodrat midden record that has been accumulated (Mehringer and Ferguson, 1969; Spaulding, 1977, 1981, 1985, 1994; Forester et al., 1996). The woodrat midden record indicates that during much of the last glacial cycle between 33 and 11 Ka, limber pine formed a NUMBER 33 333 2800 q 2500- 2200- 1900- 1600- 1300 800^ w o 5 0 800 8P = Dated middens containing limber pine New data Data from previous projects 35 30 25 20 15 10 Radiocarbon age (ky B.P.) FIGURE 15.?Elevational distribution of limber pine, as ascertained with macrofossils from ancient woodrat mid- dens in the northern Mohave Desert, by radiocarbon age. (Top) Our own and previously published data from middens containing limber pine (tree symbols) in a 200 km radius around Yucca Mountain; the dashed line is the approximate lower elevational distribution limit of limber pine in the Sheep Range today. (Middle, Bottom) Radiocarbon dates on limber pine for this study (middle) and for other studies (bottom) plotted as a normal distri- bution at three standard deviations around the mean. For these kurtosis plots, the normal distributions are a mea- sure of radiocarbon date precision?i.e., the more pointed the distribution curve, the more precise the date is; and the flatter the curve, the less precise the date is. subalpine woodland that occasionally extended down to at least 1600 m in elevation, ~1000 m below where it grows today in the Sheep and Spring Ranges of southern Nevada (Table 1; Figure 15). This is roughly equivalent to the period of highest water table in Browns Room at Devils Hole (Szabo et al., 1994). Limber pine macrofossils appear in the midden record of-33-30 Ka, 26.5-23 Ka, 21-14 Ka, and 13-11 Ka. White fir occasionally appeared in this woodland also. At those times it certainly grew along the bases of cliffs and in canyons (where woodrat middens are preserved), but whether conditions were moist enough for it to grow on exposed slopes is uncertain. Pe- riods of downward translocation of white fir are centered at -35-32 Ka, 26.5-21 Ka, 16-14 Ka, and 13-12 Ka (Table 2; Figure 16). Limber pine disappears from the midden record be- tween 35 and 33 Ka and between 23 and 21 Ka, leaving white fir as the only tree species present. Late Pleistocene climates can be extrapolated by using mod- ern analogue temperature and precipitation requirements for limber pine and white fir. Limber pine favors cool, dry conti- nental conditions with rainfall distributed relatively evenly dur- ing the year; it is drought tolerant. White fir favors cool, moist conditions with a winter peak in rainfall, but it often survives in canyons where shallow groundwater is available to it. In either case, white fir expansion indicates moister conditions. Rocky Mountain white fir appears to survive under cooler and drier conditions than does Sierran white fir. The seasonal distribu- tion of rainfall appears to vary considerably as well where Rocky Mountain white fir grows. The modern analogue cli- mates of these two species suggest the following climatic sce- nario for the late Pleistocene. The wettest periods were those when limber pine disap- peared from the woodland record. These appear to coincide 334 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 2800 ? 2500 c ? 2200 1600H 1300 800 = Dated middens containing white fir ?jLJt ??2 8go New data 800 s8 (O O So Data from previous projects \ i | i i i i | i i r 35 30 I ' ' ' ' l25 20n i i r~15 I ' ' ' ' I10 5 " I Radiocarbon age (ky B.P.) FIGURE 16.?Elevational distribution of white fir, as ascertained with macrofossils from ancient woodrat mid- dens in the northern Mohave Desert, by radiocarbon age. (Top) Our own and previously published data from middens containing white fir (tree symbols) in a 200 km radius around Yucca Mountain; the dashed line is the approximate lower elevational distribution limit of white fir in the Sheep Range today. (Middle, Bottom) Radio- carbon dates on white fir for this study (middle) and for other studies (bottom) plotted as a normal distribution at three standard deviations around the mean. See Figure 15 caption for explanation of these kurtosis plots. with the onset (23-21 Ka) and demise (16-14 Ka and 13-12 Ka), but not the climax, of the late glacial maximum. An earlier cold/wet period centered between -35 and 33 Ka apparently corresponds to one noted in the pollen records from the Sum- mer Lake basin and from the Pyramid Lake midden (see mid- dle Holocene above in "Northern and Western Great Basin"). The pollen record from Profile II at Tule Springs records in- creased pine and fir corresponding with the pre-glacial maxi- mum episodes (Mehringer, 1967, fig. 32). Such episodes sug- gest that Pacific air masses played a greater role in the climates of southern Nevada at that time. Conversely, periods of cold, dry continental climates (continental polar and Arctic air mass predominance) occurred from 16 to 21 Ka and from 29 to 32 Ka based on the absence of white fir in communities with abundant limber pine. Our knowledge of the modern mean annual precipitation (MAP) requirements of key plant species together with changes in their distribution during the Pleistocene?as re- vealed in the plant macrofossil data from woodrat middens in southern Nevada (and northern Nevada)?indicate rapid shifts toward wetter winter precipitation patterns lasting from about one century to as long as one to two millennia. The MAP ranges for modern analogue plant species suggest a Pleistocene MAP range between 380 and 635 mm at elevations between 1500 and 2000 m (Forester et al., 1996). This is roughly an in- crease of 50%-90% above the current range of 200-250 mm. NUMBER 33 335 TABLE 1 .?Lower elevational occurrences of limber pine (Pinus flexilis) within a 200 km radius of Yucca Moun- tain, Nevada, as revealed in the current study (in bold), from lowest to highest elevation. Midden data analyzed by other investigators are Fortymile Canyon (Spaulding, 1994), Sheep Range (Spaulding, 1981), Eleana Range (Spaulding, 1985), Potosi Mountain (Mead et al., 1978), and Clark Mountain (Mehringer and Ferguson, 1969). All sites are in Nevada unless noted otherwise (NTS = Nevada Test Site). Locality; midden no. Fortymile Canyon, NTS; FMC-11A (1) Sheep Range; Penthouse 2(1) Sheep Range; Willow Wash 4C(1) Sheep Range; Willow Wash 4C(2) Sheep Range; Basin Canyon 1 Pahranagat Range; PR23O594ARM1(1) Pahranagat Range; PR23O594ARM1(1)SSIDE Pahranagat Range; PR110688PEW3(2) Pahranagat Range; PR230594ARMl(2) Pahranagat Range; PR110688PEW3(3) Double Cave, Pahroc Range; DC18O595JPB1(1) Double Cave, Pahroc Range; DC18O595JPB1(2) Sheep Range; Flaherty Mesa 1 Sheep Range; Wagon Canyon 2 Eleana Range, NTS; ER-3(14) Eleana Range, NTS; ER-2(4) Eleana Range, NTS; ER-2(4) Eleana Range, NTS; ER-2(5) Eleana Range, NTS; ER-2(6) Eleana Range, NTS; ER-2(10) Ribbon Cliffs, Nellis Air Force Base; RC250588LCP11 Ribbon Cliffs, Nellis AFB; RC060795PEW41A Ribbon Cliffs, Nellis AFB; RC060795PEW41B Ribbon Cliffs, Nellis AFB; RC060795PEW41C Potosi Mountain, Spring Range; Potosi Mtn 2C2 Sheep Range; Eyrie 3(1) Sheep Range; Eyrie 1(3)1 Sheep Range; Eyrie 1(3)2 Sheep Range; Eyrie 5(2) Sheep Range; Eyrie 5(3) Clark Mountain, California; CM-1 Clark Mountain, California; CM-2 Stockade Wash, NTS; STW220594ARM113 Stockade Wash, NTS; STW220594ARM114 Stockade Wash, NTS; STW220594ARM111 Stockade Wash, NTS; STW220594ARM112U Stockade Wash, NTS; STW220594ARM112 Sheep Range; Deadman 1(1) Sheep Range; Deadman 1(2) Sheep Range; Deadman 1 (4) Sheep Range; South Crest 1(4)2 Sheep Range; South Crest 1(3) Sheep Range; Spires 2(1) Sheep Range; Spires 2(2) Cliff Springs, Belted Range, NTS; CS0909900GHH11 Clark Mountain, California; CM-2 Sheep Range; Hidden Forest 3A Sheep Range; Hidden Forest 3B Sheep Range; Hidden Forest 3C(1) Sheep Range; Hidden Forest 3C(2) Sheep Range; Hidden Forest 3C(3) Sheep Range; Hidden Forest 1 Radiocarbon age (years) 15,870 ?70 28,960 ? 2,000 24,400 ? 760 19,020 + 750 15,610 ?260 16,010 ?80 18,160 ?80 18,765 ? 180 23,890 ? 160 25,250 ? 140 32,070 ? 230 30,000 ? 240 20,390 ? 340 25,000 ?710 14,200 ?100 14,090 ?320 15,300?130 14,260?110 14,500?110 17,100?90 18,075 ? 180 24,330 ? 110 26,210 ? 130 25,850 ? 140 14,430 ?150 16,490 ?220 23,380 ?490 9,370 ?120 30,470 ? 740 19,750 + 450 23,600 ? 980 28,720 ?1,800 11,400 ? 60 12,040 ? 60 12,290 ? 70 12,340 ?60 12,550 ?70 17,420 ?250 16,800 ?250 18,680 ?280 21,700 ?500 25,140 ?900 18,800 ?130 16,200 ?350 30,270 ?360 19,900 ?1,500 1,600 ?120 820 ?100 10,060 ?130 11,940 ?140 11,860 ?160 11,570 ?120 Laboratory no. QL-4223 A-1773 A-1751 UCR-729 WSU-1856 BETA-74775 BETA-86066 BETA-32405 BETA-74772 BETA-86063 BETA-86031 BETA-86105 WSU-1864 A-1656 USGS-1266 A-2927 USGS-875 USGS-874 USGS-885 USGS-1069 BETA-32407 BETA-86080 BETA-86082 BETA-86084 LJ-4005 WSU-1853 WSU-1864 WSU-1863 WK-168 WK-167 1-3557 1-3648 BETA-86123 BETA-86088 BETA-74777 BETA-86121 BETA-74779 LJ-3707 WSU-1860 WSU-1857 LJ-2840 A-1669 USGS-199 A-1875 BETA-64355 GAK-1987 A-1742 WSU-2046 LJ-3729 LJ-3730 LJ-3728 LJ-4046 Material dated Fecal pellets Utah juniper Utah juniper Utah juniper Utah juniper Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Utah juniper Utah juniper Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine Utah juniper Utah juniper Utah juniper Utah juniper Fecal pellets Twigs Twigs Limber pine Limber pine Limber pine Limber pine Limber pine Limber pine/bristlecone pine Utah juniper Utah juniper Limber pine/bristlecone pine Utah juniper Fecal pellets Limber pine/bristlecone pine Limber pine Limber pine/bristlecone pine Utah juniper Rocky Mountain juniper Limber pine/bristlecone pine Limber pine/bristlecone pine Limber pine/bristlecone pine Limber pine/bristlecone pine Elevation (m) 1310 1580 1585 1585 1635 1695 1695 1695 1695 1695 1743 1743 1770 1790 1800 1810 1810 1810 1810 1810 1829 1829 1829 1829 1830 1855 1860 1860 1860 1860 1910 1910 1950 1950 1950 1950 1950 1970 1970 1970 1990 1990 2040 2040 2133 2140 2380 2380 2380 2380 2380 2400 336 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 2.?Lower elevational occurrences of white fir (Abies concolor) within a 200 km radius of Yucca Moun- tain, Nevada, as revealed in the current study (in bold), from lowest to highest elevation. Midden data analyzed by other investigators are from Fortymile Canyon (Spaulding, 1994), Sheep Range (Spaulding, 1981), Potosi Mountain (Mead et al., 1978), and Clark Mountain (Mehringer and Ferguson, 1969). All sites are in Nevada un- less noted otherwise. Locality; midden no. Fortymile Canyon, Nevada Test Site; FMC-11A (1) Sheep Range; Willow Wash 4A Sheep Range; Willow Wash 4C(1) Pahranagat Range; PR23O594ARM1(1) Pahranagat Range; PR110688PEW3(2) Pahranagat Range; PR230594ARMl(2) Pahranagat Range; PR110688PEW3(3) Double Cave, Pahroc Range; DC18O595JPB1,1 Double Cave, Pahroc Range; DC18O595JPB1(2) Potosi Mountain, Spring Range; Potosi 2C L Potosi Mountain, Spring Range; Potosi 2A (2) Clark Mountain, California; CM-3 Clark Mountain, California; CM-1 Clark Mountain, California; CM-2 Stockade Wash, Nevada Test Site; STW220594ARM111 Sheep Range; South Crest 1(4)2 Sheep Range; South Crest 1(3) Sheep Range; Hidden Forest 3B Sheep Range; Hidden Forest 3A Sheep Range; Hidden Forest 3C(1) Radiocarbon age (years) 15,870 ?70 21,3501420 24,400 ? 760 12,280 ? 60 23,150 ?270 23,820 ? 170 25,200 ? 140 32,790 ? 260 34,260 ?410 14,450 ?250 14,900 ?180 12,460 ?190 23,600 ? 980 28,720 ?1,800 11,890 ? 70 21,700 ?500 25,140 ?900 820 ?100 1,600 ?120 10,060 ?130 Laboratory no. QL-4223 WSU-1858 A-1751 BETA-74776 BETA-32404 BETA-74773 BETA-86064 BETA-86032 BETA-86106 A-1778 LJ-4004 1-3690 1-3557 1-3648 BETA-74778 LJ-2840 A-1669 WSU-2046 A-1742 LJ-3729 Material dated Fecal pellets Utah juniper Utah juniper White fir White fir White fir White fir White fir White fir White fir Miscellaneous Piiion pine Twigs Twigs White fir Limber pine/ bristlecone pine Utah juniper Rocky Mountain juniper Utah juniper Limber pine/ bristlecone pine Elevation (m) 1310 1585 1585 1695 1695 1695 1695 1743 1743 1830 1880 1910 1910 1910 1950 1990 1990 2380 2380 2380 Shifts from the dominance of xeric climate shrub species to Utah juniper during the Pleistocene at elevations from ~700 to 1500 m record precipitation increases ranging between 33% and 66%, a gain of 50-100 mm above current values (Forester et al., 1996). Whereas today's MAP is only 150 mm, the Pleis- tocene MAP was at least 200 mm or as high as 255 mm at ele- vations between ~700 and 1500 m. Other evidence for these wetter episodes is provided by the accumulating database of radiocarbon dates on spring deposits in the northern Mojave Desert. Dates from the Las Vegas Val- ley, Corn Creek Dunes, Three Lakes Valley, and Indian Springs Valley indicate several episodes of late Pleistocene and early Holocene spring activity (Quade et al., 1995, 1998). Quade et al. (1995) suggest that based upon the radiocarbon dates at Corn Creek Dunes there may be three distinct phases of dis- charge: an early set centered around or prior to 31 Ka, a second set from ca. 25 to 14 Ka, and a third set from 14 to 8 Ka. As more dates accumulate, these episodes may become more con- strained and may match the vegetation data more closely. To the southeast in the Lake Mojave basin, a series of lakes? some lasting for as many as 2000 years and others lasting only a few hundred?span the period between -22 and ?11.5 Ka (Brown et al., 1990, fig. 8). These may, in part, correspond to the precipitation events seen in the record from southern Ne- vada; however, these lakes reflect climatic conditions in trans- verse ranges, which may have experienced a quite different se- quence of rainfall history because of their proximity to the Pacific Ocean. These occasional wet episodes are set against the major por- tion of the Pleistocene between 35 and 12 Ka, which was dom- inated by more continental conditions characterized by cold- dry climate about 1.33 times wetter than today (Forester et al., 1996). Based upon a comparison of the temperature require- ments of Rocky Mountain and Sierran white firs versus limber pine, the onset and decline of the late glacial maximum may have been 2?-6? C warmer than its climax. Precipitation gradients increased substantially during the Pleistocene. The MAP was approximately 133% of what it is today at ~750 m in elevation, based upon the displacement of Utah juniper (Wells and Jorgensen, 1964; Wells and Berger, 1967; Forester et al., 1996). Limber pine displacement indi- cates that MAP was approximately 190% of what it is today at ~1500 m (Forester et al., 1996). If the increase in regional MAP had been regular, the percent of increase should be the same at all elevations. However, the difference in MAP be- tween these two elevations during the Pleistocene reveals that precipitation was ~50% greater at 1500 m than it was at 750 m. Lack of woodrat midden data above 2000 m for the Pleistocene prevents extrapolation of precipitation values to higher eleva- tions. This reconstructed rainfall gradient indicates a dramati- cally enhanced orographic precipitation effect during the late Pleistocene in southern Nevada. NUMBER 33 337 TABLE 3.?A general sequence of climate and mean annual precipitation (MAP) for southern Nevada from ca. 35 to 12 Ka. Radiocarbon date (Ky B.P.) -35-33 33-26.5 26.5-23 23-21 21-16 16-14 14-13 13-12 Relative climate Very wet, cool Dry, cooler Wet, cool Very wet, cool Dry, cold Wet, cool Dry, cooler Wet, cool Estimated MAP (mm) 560-635 406-457 457-560 560-635 406-457 457-560 406-457 457-560 By using the MAPs derived from the vegetation data summa- rized above, a general sequence of climate is generated for the period from ~35 to 12 Ka for southern Nevada (Table 3). This applies only for elevations around 1500 m. At elevations around 750 m, estimates should be about 50% of those at 1500 m. Above 2000 m, absence of data prevents estimation. Cli- mate designations in Table 3 are only relative. At lower elevations, desert shrub communities dominated by saltbushes, rabbitbrush (Chrysothamnus spp.), mormon tea {Ephedra spp.), and other shrubs appear to have prevailed through much of the late Pleistocene (Spaulding, 1985, 1990, 1994; Wells and Woodcock, 1985). Investigations in the Pint- water Cave area, north of the Spring Range in southern Ne- vada, revealed shrub community dominance with intermittent incursions of Utah juniper for most of the last 34 Ky. A brief appearance of Rocky Mountain juniper (Juniperus scopulorum Sarg.) in the area of the cave after ~34 Ka (numerous seeds in the cave sediments) suggests cooler, wetter climates at that time than later in the Pleistocene. Similar climatic conditions seem to have allowed Rocky Mountain juniper to exist on the east shore of pluvial Owens Lake between 23 and 17.5 Ka (Koehler and Anderson, 1994). The cool temperatures gener- ated in the Owens Valley by the nearby Tioga glacial advance may have provided conditions suitable for the survival of Rocky Mountain juniper far from its current upper-elevation distributions. EARLY HOLOCENE (12.5 TO 8 KA) In southern Nevada, replacement of subalpine woodland dominated by limber pine with juniper-dominated semiarid woodland at intermediate elevations preceded the disappear- ance of low-elevation juniper woodland by desert scrub (Spaulding, 1977, 1981; Wigand et al., 1995). Although woodrat midden records indicate that pinon pine appeared in several areas around southern Nevada including the Pahrana- gat Range well before 8.5 Ka, it seems to have been a much less important component of early Holocene semiarid wood- lands than was juniper (Spaulding, 1985, 1990; Thompson, 1990; Wigand et al., 1995). However, records from the north- central and eastern Great Basin (Thompson and Hattori, 1983; Thompson and Kautz, 1983; Wells, 1983; Madsen and Rhode, 1990; Thompson, 1990) confirm pinon pine's rapid northward expansion. Despite evidence of warming climates in the southern Great Basin, the persistence of juniper woodland outliers after 9.5 Ka may reflect occasional wetter and/or cooler periods of climate that aided their persistence. At Little Skull Mountain on the southern edge of the Nevada Test Site, juniper persisted until 10.5 Ka but disappeared from the record by 10 Ka (Wigand, unpublished data). By 9.5 Ka, however, it reappeared in the woodrat midden macrofossil record at Little Skull Mountain. Organic mats (residues) dating primarily from 10.2 to 8.6 Ka also reflect these episodes of moister climate (Quade et al., 1998, especially fig. 5). Based upon the vegetation during this span of time, the rainfall from two different climatic regimes may have been responsible for formation of the black mats. Climates may have been considerably cooler from 11.2 to 10.2 Ka than from 12 to 11.5 Ka, resulting in reduced evaporation rates and greater effective precipitation. Climates were warmer from 10.2 to 8.6 Ka, and from 9.5 to 8.6 Ka the summer rainfall may have increased considerably. Evidence of a shift to summer seasonal rainfall during the warm portion of the latter early Holocene between 9.5 and 9 Ka is provided by the appearance of netleaf hackbeiry {Celtis laevigata var. reticulata (Torn) L. Benson) seeds in the paleo- botanical record of southern Nevada. Hackberry seeds (en- docarps) have been recovered from woodrat middens located on the xeric southeast slope of Little Skull Mountain (Wigand et al., 1994), from middens in the Sheep Range north of Las Vegas (Spaulding, 1990), and from the Sandy Valley (Spauld- ing, 1994), and several hundred seeds were recovered from the sediments of Pintwater Cave (Jahren et al., 2001). Today it oc- curs in canyons, seeps, and washes at elevations of 500-1700 m. Hackberry ranges from the southern Sierra Nevada foot- hills and the San Bernardino Mountains in the southwestern United States and northern Mexico into the southern plains to Kansas, with outliers in Washington (Benson and Darrow, 1981; Krajicek and Williams, 1990). In southern Nevada it is found associated with a spring near Pahroc Spring in the Pah- roc Range, in one locality in Fortymile Canyon on the Nevada Test Site, and in canyons at the southern end of the Spring Range. The appearance of hackberry remains in middens sug- gests it grew at the bases of cliff faces or in arroyo bottoms, areas that are currently too dry to support it, and reveals a pat- tern of greater regional abundance of the tree between 9.5 and 9Ka. These reconstructed climatic conditions compare well with evidence of increased spring discharge found by Quade et al. (1998) that apparently may correspond to incursions of mon- soon rains during the thermal maximum (Spaulding, 1985; Spaulding and Graumlich, 1986). In particular, well-developed black mats, which Quade et al. (1998) describe from the val- leys surrounding the Spring Range, dating to between 9.5 and 8.6 Ka probably reflect the period of increased summer mon- soon activity. Additional evidence of episodes of the incursion 338 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES of moisture during this period are the intermittent lakes that ap- pear in the Lake Mojave basin (Silver and Soda Lake playas) at this time (Brown et al., 1990). The evidence indicates that as many as six lakes (with intervening drying) occupied the Lake Mojave basin between 11.5 and ~8.5 Ka (Brown et al., 1990, fig. 8). Therefore, during a period when other climate proxies indicate that conditions were much warmer and that mean an- nual rainfall was decreasing, a shift of rainfall toward the sum- mer season coinciding with the early Holocene thermal maxi- mum would explain the presence of perennial seeps, as evidenced by the latter period of black mat formation, the ma- jority of which Quade et al. (1998) have dated to between 10.2 and 8.6 Ka. In the Alabama Hills north of the Owens Lake basin, the ter- mination of Pleistocene woodland occurred 9.5 Ka (Koehler and Anderson, 1995). This coincides with the disappearance of juniper woodland from outlier positions in the Lahontan Basin, as noted above, and also with the termination of lithoid tufa formation around the Lake Lahontan shoreline. This is also, as just noted, the beginning of the period of greater penetration of monsoon rainfall into southern Nevada. Together these data suggest a major northward displacement of regional climatic regimes at this time. MIDDLE HOLOCENE (8 TO 5.5 KA) As Spaulding (1990) indicated, the early Holocene vegetation assemblages of the northern Mojave Desert were, in most cases, unique and did not survive the middle Holocene. Pollen sam- ples from Profile V in the Tule Springs record indicate a dra- matic increase in low-spine (Ambrosia-type) pollen shortly after 8.5 Ka (Mehringer, 1967). This probably corresponds to the ap- pearance of white burrobush (Ambrosia dumosa (Gray) Payne), one of the two codominants of the typical Mojave Desert creo- sote bush-white burrobush vegetation assemblage (Spaulding, 1990, fig. 9.13). According to the woodrat midden macrofossil record, creosote bush appears slightly later in the area (but see Grayson's (1993:200) discussion). Further west the appearance of these two species is even later (Spaulding, 1980). Unfortu- nately, pollen can not be used to document the presence or ap- pearance of creosote bush because of its rarity in the pollen re- cord. Both these species arrived as migrants from further south, but the climatic impetus for their arrival is unclear. Truncation of the pollen record in Profile IV at Tule Springs sometime after 7.5 Ka corresponds with the beginning of a gen- eral absence of paleobotanic data spanning the next 2 Ka in the southern Great Basin and northern Mojave Desert (Mehringer, 1967). The dearth of vegetation data is probably a direct reflec- tion of the impact of severe drought in the region. Springs and marshes dried and woodrats may have retreated into more fa- vorable areas. Severe eolian erosion may have characterized parts of this period, because late-Holocene spring and marsh deposits in the northern Mojave Desert generally lie upon ero- sion surfaces that were formed during the middle Holocene (Mehringer and Warren, 1976; Mehringer and Sheppard, 1978; Wigand, 1997b). This suggests that vegetation cover may have been very sparse. Only a few middens are dated to this period. One collection of middens from the McCullough Range southwest of Las Vegas is dominated by thermophilic (warm-temperature) plant species (Spaulding, 1991). These reflect the warm, dry conditions of the middle Holocene. Spaulding also suggested that no evidence exists for summer seasonality of rainfall between 6.8 and 5 Ka. LaMarche (1974) suggested that upper tree line temperatures were considerably warmer during this period and that upper tree lines extended upward. Graybill et al. (1994) indicated ex- tended periods of drought as well. These data suggest that the middle Holocene in the Great Basin was characterized by con- siderable warmer and drier conditions than currently exist in the Great Basin. EARLY LATE HOLOCENE (5.5 TO 2 KA) In the northern Mojave Desert there is a dramatic change in climate at this time. It is most clearly manifested in the sudden formation of peats that overlie erosional unconformities. Peats occurred at the base of the Little Lake record in the Owens River valley (Mehringer and Sheppard, 1978), in Ash Mead- ows south of the Nevada Test Site (Mehringer and Warren, 1976), and at the base of the Lower Pahranagat Lake cores in the White River valley northeast of Las Vegas (Wigand, 1997b). These peats attest to the sudden rejuvenation of springs in the region. Mesophilic plant abundance in woodrat middens of the McCullough Range began in strata dating to 5.5 Ka (Spaulding, 1991). A woodrat midden stratum in the southern Pahranagat Range dated to 5.6 Ka reveals a brief lowering in elevation of Utah juniper below its current elevational distribu- tion (Wigand et al., 1995). In southern Nevada, increased abundance of woodrat mid- dens containing Utah juniper and pifion pine macrofossils evi- dence a major Neoglacial reexpansion and/or increase in den- sity of pinon-juniper woodland beginning ~3.8 Ka (Wigand et al., 1995, fig. 4). This early phase of semiarid woodland expan- sion culminated about 2.3 Ka. At Lower Pahranagat Lake, a pollen record analyzed back to 3.8 Ka with a resolution of a sample every 14 or so years pro- vides a clear record of drought cycles for the northern Mojave Desert and the southern Great Basin (Figure 17; Wigand, 1997b). This record is correctable with the White Mountain tree ring record of Graybill et al. (1994) (Figure 18). It provides an indication not only of relative increases in precipitation but also of shifts in seasonal distribution as well. Wetter climate events of the Neoglacial are reflected in the record from Lower Pahranagat Lake, but two differences from the Neoglacial wet events in the Diamond Pond record of the Great Basin are evident. First, only two of the three major wet NUMBER 33 339 oo w ooo oo ooo CM oo IO CM OO OCO os CO oo o 3 e5 3? 60 Si c 2 O E I S B O to ^24,500 -29,000 est 0% Increasing Conifers 100% FIGURE 20.?Pollen ratios of xeric desert scrub/conifers for the last 30,000 years from Great Salt Lake Core C samples (Madsen and Kay, 1982; Spencer et al., 1984), with temporal controls in the right margin provided by Thompson et al. (1990). The curve shows relative percentages of Chenopodiaceae- Amaranthus+Sarcobatus+Ambrosia-type Asteraceae (scale at top) versus Abies+Picea+Pinus+Pseudotsuga (scale at bottom) and is smoothed by a weighted three-level moving average. (Modified from Madsen, 2000, fig. 4.) dilation and development of the Stansbury Shoreline (Oviatt et al., 1990). According to Oviatt et al. (1990, 1992), the Stans- bury oscillation represented a major change in hydrologic bud- NUMBER 33 345 get, though the climatic implications are still uncertain. Oviatt (1997) correlated this event with the termination of an ice-raft- ing (Heinrich) event in the northern Atlantic Ocean, suggesting a global-scale climatic cause. The vegetation record dating to the millennia after 28 Ka shows continued dominance of bristlecone pine (without spruce or Utah juniper) in the southern Bonneville Basin, based on Council Hall Cave pollen (Thompson, 1984) and a couple of packrat midden samples (Wells, 1983). However, the inter- val between 28 and 18 Ka is poorly represented in the existing packrat midden record in the southern Bonneville Basin. Northward, at Top of the Terrace (Figure 19), several samples document that spruce increased in abundance while limber pine declined to low amounts (Rhode, 2000a; Madsen et al., 2001). Upper montane meadow plants such as currant and cinquefoil reappeared and expanded in abundance at the expense of fern- bush; sagebrush maintained fairly consistent representation throughout. These vegetation changes suggest increasingly moist and cool conditions by ca. 22 Ka; however, the midden record for the period is more gap than record. Fortunately, pollen and macrofossils in sediments and sedi- ment cores provide more continuous, long-term records of veg- etation change in the Bonneville Basin. The record from Cres- cent Spring (Mehringer, 1977) shows relatively low to moderate values of pine and spruce prior to ca. 24 Ka and rela- tively large values for sagebrush and Chenopodiineae, presum- ably shadscale and greasewood, reflecting widespread sage- brush-shadscale associations in valley bottoms, with pine and some spruce in uplands. A transition to deep, freshwater lake conditions occurred at ca. 24 Ka (Mehringer, 1985), with large increases in pine pollen, the appearance of spruce, and the de- cline of chenopods. Davis (2002) shows a very similar record for pollen taken from Wendover and Knolls. Pines may not have increased much in overall abundance on the landscape, but their pollen may have been concentrated and overrepre- sented by large lake effects. Chenopods, on the other hand, probably lost substantial habitat as rising levels of Lake Bonne- ville flooded valley bottoms and margins. The Great Salt Lake Core C (Spencer et al., 1984; Thompson et al., 1990) indicates a similar transition to a fresh, deepwater lake by ca. 22.5 Ka, with greatly increased conifer pollen at the expense of sagebrush and xerophytic chenopods (Figure 20). A similar increase in lake level after 23 Ka is indicated in the lacustrine record from Ruby Marsh (Thompson, 1992), though pollen values suggest that sagebrush steppe continued to domi- nate the local landscape. Pollen from Grays Lake, in the up- lands adjacent to the Snake River Plain (Beiswenger, 1991) shows that a pine/spruce woodland increased in abundance by ca. 30 to 26 Ka, intruding into what had been sagebrush steppe, and suggesting increasing moisture as well as continued cold temperatures. Finally, scattered occurrences of spruce and fir wood dating between 26 and 20 Ka are found in lake deposits on the east side of the Bonneville Basin (Scott et al., 1983), presumably indicating spruce/fir forests in the Wasatch Front. Wetter and cold conditions are indicated. MIDDLE BONNEVILLE CYCLE, FULL GLACIAL (20 TO 14 KA) During this interval Lake Bonneville rose abruptly from the Stansbury level?rapidly for the first 2000 years, then more slowly after 18 Ka?reaching the Bonneville highstand by ca. 15 Ka (Oviatt et al., 1992). Age estimates of glaciation in the Wasatch Range suggest the maximum extent of glaciation was between 20 and 18 Ka (Richmond, 1964, 1986; Madsen and Currey, 1979; Scott, 1988), at about the time Lake Bonneville was rising rapidly but a couple thousand years before the lake reached its highstand. The difference in timing between the age of the maximum extent of glaciation and the highstand of Lake Bonneville has been accounted for by the southward movement of the polar jet stream (Antevs, 1948; Benson and Thompson, 1987; COHMAP Members, 1988; Oviatt et al., 1990; Thompson et al., 1993). The rise of the lake was inter- rupted several times by substantial millennial-scale lake level declines (on the order of 50-100 m in depth). Oviatt (1997:155) has identified six "falling-lake events" ranging in age from 21 to 10 Ka that appear to be correlated with termina- tions of iceberg-rafting events in the northern Atlantic Ocean (Heinrich, 1988; Bond and Lotti, 1995). These events also ap- pear to be correlated with glacial extent and climate changes in western North America (Allen and Anderson, 1993; Phillips et al., 1994; Clark and Bartlein, 1995). Oviatt (1997) suggested that these events may be related to latitudinal shifts in the polar jet stream that occurred several times during the late-Wiscon- sin interval in response to the topography of the continental ice sheet. Vegetation records dating from the full glacial are somewhat more abundant than from the previous period. In the southern Bonneville Basin, packrat middens reveal that bristlecone pine continued to dominate, with additional representation of subal- pine conifers including spruce, common juniper, Rocky Moun- tain juniper, and (rarely) limber pine; and shrubs including snowberry, spiny greasebush (Glossopetalon spinescens Gray), sagebrush, and (rarely) shadscale (Wells, 1983; Thompson, 1984). Pollen from Council Hall Cave contains large propor- tions of pine pollen with increasing spruce and decreasing abundance of juniper-type pollens and relatively stable amounts of sagebrush pollen. To the north, a midden sample dating to ca. 17 Ka at Top of the Terrace (Figure 19) shows an impoverished local flora, suggesting a tundra steppe domi- nated by sagebrush, some grass, and sparse cinquefoil, with rare occurrences of limber pine and spruce. This sample also contains several wetland plants, including bulrush (Scirpus sp.) and pondweed {Potamogeton sp.), as well as fish bones. It is highly unlikely that pondweed grew within the foraging radius of woodrats living at the midden locality. More likely is the possibility that seeds of aquatic plants were ingested by fish or waterfowl, which then became part of the diet of raptors, 346 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES regurgitated in the shelter, and subsequently collected by woodrats. The seeds imply a nearby permanent body of water, probably Lake Waring in the valley below, but possibly also Lake Bonneville, east of the Goshute Mountains. Slightly younger samples from this midden locality document the reap- pearance of currant and an absence of limber pine. Pollen from the Great Salt Lake Core C (Spencer et al., 1984) shows a continued dominance of spruce and pine during this period; the relative abundance of sagebrush and especially chenopod shrubs dropped substantially after ca. 22 Ka (Figure 20). In Ruby Valley, the large, freshwater Lake Franklin deep- ened to ~30 m from ca. 18 Ka until after 15 Ka, drowning what had been marshland (Thompson, 1992). Pollen from these sed- iments, however, shows relatively little change in the composi- tion or abundance of sagebrush steppe vegetation surrounding the lake. It is likely that the lake resulted from reduced evapo- ration rate in a cold but relatively dry continental climate; in- creased precipitation likely would have resulted in more meso- philic shrubs and conifers in the region, which is not indicated in the pollen record. In all, increasingly cold conditions are indicated after ca. 20 Ka; whether precipitation was substantially greater is less cer- tain. Thompson and Mead (1982) noted that a steep vegetation gradient existed between the northern and southern Bonneville Basin (above the tree line in the north, below the tree line to the south), probably reflecting the distribution of growing-season temperatures and possibly also growing-season precipitation (dry in the north, more mesic in the south). The vegetation re- cord is not yet sufficiently detailed or well dated to distinguish vegetation changes during the several falling-lake events re- ported by Oviatt (1997). Rapid fluctuations of pollen values in the Great Salt Lake Core C samples (Spencer et al., 1984) sug- gest that certain taxa (conifers and sagebrush) varied in abun- dance or that tree lines may have shifted in elevation in re- sponse to global-scale climatic changes during the full glacial (Figure 20). LATE BONNEVILLE CYCLE, TERMINAL WISCONSIN (14 TO IOKA) Dating evidence summarized by Oviatt et al. (1992) and Ben- son et al. (1992) suggests that, after Lake Bonneville stabilized at the Provo level after the Bonneville Flood, the lake receded precipitously after 14 Ka, falling possibly to below modern lev- els by ca. 12 Ka, and marking the end of the Bonneville cycle. Subsequently, a large but much shallower lake rose to the Gil- bert shoreline by around 10.9 Ka (possibly as early as 12 Ka), receding again by ca. 10 Ka. As Oviatt et al. (1992) noted, how- ever, the timing of the recession from the Provo level is one of the least-known parts of the Bonneville lake cycle; dating of the Gilbert lake phase is little better (but see Oviatt et al., 2001). Recent dating efforts (Light, 1996; D.B. Madsen and C.G. Oviatt, pers. comm. to Rhode, 1996), as well as climate recon- structions from terrestrial vegetation data (Rhode and Madsen, 1995; see below), suggest that details of both the timing and magnitude of the post-Provo recession may need to be revised. Timing of the recession of Lake Bonneville also contrasts with the pluvial Lake Lahontan record, which indicates a deepening lake between 14.5 and 12.5 Ka (Benson et al., 1992), and with the evidence from the vegetation records in the northwestern Great Basin and the northern Mojave Desert, which suggests that this period was wetter than the glacial maximum (see "Last Glacial Cycle ..." above in "Northern and Western Great Basin" and see "Late Pleistocene ..." above in "Southern Great Basin and Northern Mohave Desert"). In contrast to the poorly dated recessional lake record, the terminal Wisconsin period is one of the best represented in the entire paleovegetation record, with many packrat middens and a couple of pollen sequences available (Madsen and Currey, 1979; Wells, 1983; Thompson, 1984; Rhode and Madsen, 1995). Several substantial vegetation changes occurred after 14 Ka. In the southern Bonneville Basin (Thompson, 1990), meso- philic shrubs tended to decline in abundance, and xerophytes or thermophilous shrubs took their place. Bristlecone pine re- mained common until shortly before 10 Ka. Spruce is no longer found in the record of this period, and several montane shrubs also drop out at particular localities. Limber pine in- creased substantially in frequency, especially after 11 Ka. Utah juniper and douglas fir (Pseudotsuga menziesii (Mirbel) Franco) also began to appear in the south near the end of this period. In the northern Bonneville Basin, numerous middens (Rhode and Madsen, 1995; Rhode, 2000a) document the presence of montane meadow vegetation comprising snowberry, currant, cinquefoil, sagebrush, and grass dating from ca. 14 to 13 Ka (Figure 21). From 13 to ca. 11 Ka, middens from relatively low elevations in the Bonneville Basin are dominated by limber pine, prostrate juniper, and sagebrush, and often contain re- mains of other montane shrubs. This evidence suggests that an extensive limber pine woodland (or mosaic of woodland and sagebrush steppe) dominated upper valley margins, while lim- ber pine forest mixed with small numbers of spruce or fir ap- parently grew at higher elevations. Pollen data from Swan Lake, in the old Lake Bonneville overflow channel in southern Idaho, also show high values of limber pine and probably lodgepole pine (Pinus contorta Dougl. ex Loud.), along with spruce and fir, between 12 and 11 Ka (Bright, 1966). Pollen data from the Great Salt Lake Core C (Spencer et al., 1984; Thompson et al., 1990) show high relative abundance of pine and spruce until after 12.4 Ka, when sagebrush, grease wood, and Chenopodiineae rapidly increase. Pollen from Grays Lake (Beiswenger, 1991) documents a shift by ca. 12 Ka from a pine-sagebrush parkland reflecting cold, dry glacial conditions to a woodland/steppe mosaic, including spruce, pine, and sage- brush, suggesting cool but moister conditions. Subsequent postglacial warming resulted in the upslope retreat of spruce and pine and the expansion of sagebrush and xerophytic cheno- pods in the Grays Lake basin. NUMBER 33 347 Terminal Wisconsin-Age Packrat Middens, Northeast Great Basin a 03 14 I I - I 1 - - 1 1 1 1 1 1 1 1 1 1 1 ll ll 1 1 1 1 1 ,ll 1 1.1 III Illl 1 III 1 Illl ll Illl 1 1 1 1 1 1 1 1 1 1 1 1 Gooseberry Snowberry _ Cinquefoil Sagebrush Common Juniper Limber Pine Shadscale 13 12 11 Radiocarbon Ky B. P. 10 FIGURE 21.?Abundances of major plant constituents from selected radiocar- bon-dated strata in packrat middens dating between 14 and 10 Ky B.P. in the northern Bonneville Basin (Rhode and Madsen, 1995). Radiocarbon dates are corrected for I3C/12C but are not calibrated. Abundance values follow an ordi- nal scale ranging from 0 (none) to 5 (abundant), as described by Rhode and Madsen (1995). Together these records indicate warmer and possibly moister conditions ca. 14-12 Ka than during the full glacial, supporting montane meadows and woodlands in the north and more ther- mophilic woodland plants in the southern part of the Bonne- ville Basin. However, the climatic amelioration that occurred between 14 and 12 Ka was not so great as to prevent presently subalpine conifers from colonizing large areas at low eleva- tions along the Bonneville highstand shoreline. Rhode and Madsen (1995) used the evidence that limber pine woodlands expanded substantially to the margins of the Bonneville Basin to suggest that growing season temperatures were -6? C cooler than today. Insect assemblages from some of these middens also indicate substantially increased effective moisture relative to today (Scott Elias, pers. comm. to Rhode, 1998). In addition, the presence offish bones in middens adjacent to the Great Salt Lake Desert and in Homestead Cave, in the Lakeside Range near Great Salt Lake, suggests relatively high lake levels per- haps as late as 11.5 Ka. Together, these biological data conflict with models of the decline of Lake Bonneville to below-mod- ern levels at ca. 13-12 Ka (Benson et al., 1992; Oviatt et al., 1992). By 12 Ka, the trend to drier conditions apparently strength- ened. Pollen from Swan Lake indicates that spruce/pine wood- lands were replaced by sagebrush steppe vegetation sometime between 11 and 10 Ka (Bright, 1966), possibly signaling cool but drier conditions during the Younger Dryas. The paleovege- tation record from 11 to 10 Ka in the Bonneville Basin, how- ever, is quite sparse at present. Middens dating to this interval are extremely scarce (Figure 21), and the Great Salt Lake C core (Spencer et al., 1984) is marked by a substantial temporal gap dating between ca. 12 and 9.6 Ka (Thompson et al., 1990) (Figure 20). What little data there are suggest a substantial re- duction in pine woodlands and an increase in sagebrush-grass and shadscale associations, probably by 11 Ka and certainly by 10 Ka, reflecting the general trend of warming and drying. There is presently not enough evidence in the Bonneville Basin to show that vegetation changes occurring during the Gilbert lake cycle reflected enhanced moisture conditions, through ei- ther postulated Younger Dryas cooling (Benson et al., 1992) or increased summer monsoon precipitation (Currey, 1990; Oviatt etal., 1990). EARLY HOLOCENE (10 TO 7 KA) Early Holocene midden records are scarce by comparison with the terminal Wisconsin. The reason for the relative dearth of middens dating to the early Holocene remains unknown (Thompson, 1990), but it is possible that a continued warming and drying trend restricted Neotoma cinerea (bushy-tailed woodrat) habitat and population size. Fortunately, several pol- len records from lacustrine settings exist for the Bonneville Pa- sin and vicinity. Comparison of vegetation patterns with the Holocene aquatic history of the Bonneville Basin also is hampered by lack of a well-known Holocene lake record for this Basin (Madsen et al., 2001). With the waning of Lake Bonneville after recession from the Provo level, lake level fluctuations became much smaller in scale. These smaller fluctuations are not as well dated as the lake events of the late Pleistocene (but see Currey and Madsen, 1974; Oviatt and McCoy, 1986), and some of the crucial evidence (for very low lake levels) is now under water and mostly inaccessible. It is known that lakes substantially larger than the present Great Salt Lake existed at several times 348 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES during the Holocene (Currey et al., 1984). Evidence for an un- dated but probably early Holocene lake with a level somewhat below the Gilbert shoreline suggests a mesic interval (relative to today) some time between 10 and 7 Ka. Marsh deposits dated 10-9 Ka were identified recently in the vicinity of Wen- dover and near the Old River Bed channel (D.B. Madsen, pers. comm. to Rhode, 1998). Both marshes indicate greater surface water and a higher regional water table than what prevails to- day. Woodrat midden records in the southern Bonneville Basin suggest that woodlands were dominated by newly immigrant Utah juniper and Rocky Mountain juniper, at least toward the end of the early Holocene; bristlecone pine, limber pine, and common juniper persisted in protected montane settings such as north-facing slopes (Thompson, 1990). The Council Hall Cave pollen record suggests that conifers were relatively scarce, whereas trees such as quaking aspen (Populus tremu- loides Michx.), mountain mahogany, and chokecherry (Prunus virginiana L.) increased in abundance, and sagebrush and grass were common (Thompson, 1984). Cooler temperatures in the early Holocene apparently allowed mesophilic plants to grow at lower elevations than today (Thompson, 1990). In the northern Bonneville Basin, the Great Salt Lake Core C pollen record (Spencer et al., 1984) shows that the abundance of pollen from xerophytic shrub communities relative to coni- fer woodlands was about the same 9.5 Ka as it was about 12 Ka, bridging the 2500-year-long unconformity in the core (Fig- ure 20). Soon after 9 Ka, however, xeric desert scrub increased strongly at the expense of conifer woodlands. Packrat middens (Rhode, 2000b; Madsen et al., 2001) also indicate that limber pine woodlands were replaced by more xe- rophytic sagebrush and shadscale scrub in lowlands by at least 9.5 Ka and probably a millennium earlier, as shadscale steppe continued to expand into the valley bottoms vacated by Lake Bonneville and Lake Gilbert (Figure 22). At the Leppy Over- hang locality, a midden sample dating to 9.3 Ka contains re- mains of shadscale, horsebrush (Tetradymia sp.), and other arid desert shrubs, with no conifers represented. A sample dating to ca. 7.5 Ka at the same site is dominated by shadscale, with sagebrush, greasebush, rabbitbrush, Stansbury cliffrose (Pur- shia stansburiana (Torr.) Henrickson), desert snowberry (Sym- phoricarpos longiflorus Gray), and thornberry (Lycium sp.) also present. This locality had harbored limber pine woodlands at approximately 11.8-10.8 Ka (see "Late Bonneville Cycle ..." section above in "Bonneville Basin and Eastern Great Basin"). At the Twin Peaks locality, a midden sample that dates to 9.3 Ka is dominated by sagebrush, with shadscale as a relatively minor component. Limber pine, in association with common juniper, sagebrush, and buffaloberry, had occurred at the same locality some 3000 years before. The same contrast can be found in the modern vegetation at the two localities: sagebrush dominates at the Twin Peak locality, whereas the Leppy Over- hang site is dominated by shadscale. Holocene Packrat Middens, Northeastern Great Basin Limber Pine 8 COo 4 o ?o Rocky Mt "Juniper l I I 1 : Utah Juniper 5 i? SagebrushJ II I I I ill ill IShadscalel 10 6 4 Radiocarbon Ky B. P. i .hill I i FIGURE 22.?Major plant constituents from selected radiocarbon-dated strata in packrat middens dating younger than 10 Ky B.P. in the northern Bonneville Basin (Rhode and Madsen, 1995; Rhode, 1998, and unpublished data). Radio- carbon dates are corrected for 13C/12C but are not calibrated. Abundance val- ues follow an ordinal scale ranging from 0 (none) to 5 (abundant), as described by Rhode and Madsen (1995). Several records suggest that climatic conditions during the early Holocene were cooler and/or moister than at present, at least before ca. 8.5 Ka. Thompson's (1984) midden data indi- cating a cooler, mesic early Holocene have been noted previ- ously. Further north, a midden from the Cherry Creek Range dating to ca. 7.7 Ka shows that limber pine survived in sheltered upland settings below its present altitude. Limber pine nut hulls in Danger Cave, directly dating between 8 and 7 Ka, also sug- gest that populations of this now subalpine conifer grew at rela- tively lower elevations nearby (Rhode and Madsen, 1998). Lim- NUMBER 33 349 ber pine retreated upslope to its scattered subalpine stations during the middle Holocene, as climates continued to warm and as thermophilous competitors migrated into the region. To the west, woodrat middens from near Gatecliff Shelter in the To- quima Range, central Nevada, are dominated by mesophilic trees and shrubs such as quaking aspen, serviceberry (Amelan- chier utahensis Koehne), willow (Salix spp.), and Woods' rose (Rosa woodsii Lindl.), with sagebrush and rabbitbrush also present; conifers are rare or absent (Thompson and Hattori, 1983). At Homestead Cave and the Lakeside Range in Utah, small mammal remains also indicate that the early Holocene was rel- atively cool and moist compared with conditions today (Gray- son, 1998). Hackberry apparently was common in the outcrops around the cave and at lower elevations in the Lakeside Range from about 10 to 7 Ka (Hunt et al., 2000; Rhode, 2000b; Mad- sen et al., 2001). The climate today is too dry to support hack- berry around Homestead Cave although it still occurs in the Lakeside Range (Albee et al., 1988) and is commonly found in the more massive and better-watered Oquirrh and Stansbury Mountains not far away. Hackberry prefers rocky substrates where it can absorb water trapped in cracks in outcrops (De- Bolt and McCune, 1995). The presence of hackberry in the Lakeside Range during the early Holocene could have resulted from the more mesic conditions (perhaps summer precipita- tion) of the time, but it also could have resulted from a higher water table and more-saturated rock outcrops remaining from the Bonneville pluvial episode. The early Holocene hackberry in the Lakeside Range coincides with a similar increase in re- gional abundance of hackberry in southern Nevada, as noted above in the early Holocene section in "Southern Great Basin and Northern Mojave Desert." Other regional vegetation records indicate substantial warm- ing during the early Holocene, perhaps to temperatures warmer than today. The pollen record from Swan Lake (Bright, 1966) reveals a transition from sagebrush steppe to sagebrush/shad- scale steppe ca. 9-8.5 Ka, reflecting increasing xericity through the early Holocene. A similar sequence occurred at Grays Lake (Beiswenger, 1991), where the transition is dated to ca. 8.5 Ka, and Ruby Marsh (Thompson, 1992), where the transition dates from 8.5 to 8 Ka. A decrease in the ratio of conifer pollen to all other pollen from Snowbird Bog (Madsen and Currey, 1979) also indicates an early Holocene warming prior to ca. 8 Ka (Figure 23). These records are consistent with general circula- tion models, suggesting increased solar insolation and warming of continental interiors during the early Holocene (COHMAP Members, 1988). It is notable that early Holocene records indicating relatively cooler, moister conditions than today are typically (1) derived from packrat middens, (2) montane in distribution, and (3) more southerly in distribution; whereas those records empha- sizing the early Holocene warming trend tend to be (1) pollen based, (2) from lowland or valley bottom settings, and (3) more northerly in distribution. The differing interpretations might have resulted from the different kinds of data sets being inter- 1870 ?110 5220 ?150 5660 ?190 Radiocarbon Date(Yr BP) 7920 ?210 12,300 ?330 Cooler Warmer FIGURE 23.?Estimate of Holocene temperature change in the Wasatch Range, illustrated by curve showing changes in the relative abundance ratio of conifer pollen (Abies + Picea + Pinus + Pseudotsuga) to all other pollen in Snowbird Bog, Little Cottonwood Canyon (Madsen and Currey, 1979), with the average ratio (dashed line) and selected temporal controls indicated. Cooler tempera- tures are indicated by relatively greater abundances of conifer pollen. The curve is smoothed by a weighted three-level moving average. (Modified from Madsen, 2000, fig. 5.) preted, or geographical climate gradients, or the differential re- sponse of plants to early Holocene climate in lowland versus montane habitats. Enhanced seasonality resulting from orbital forcing during the early Holocene (Thompson et al., 1993) may have affected montane woodlands in a different way than low- land shrub communities. Vegetational inertia (Cole, 1985; Davis, 1986) may also play a role here. It seems likely that, be- cause of a warmer and drier early Holocene climatic regime, mesophilic plants had already been eliminated from valley floors; and although they had not yet disappeared from uplands they may have been growing there under stress. New tech- niques for assessing plant physiological stress from macrofos- sil remains may address this issue. 350 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Beginning in the latest Wisconsin period and continuing through the early Holocene and into the middle Holocene, a va- riety of plant taxa underwent substantial range shifts, apart from simple changes in local abundance or upslope retreat. These range shifts are best attested by the evidence from pack- rat middens. One example is Rocky Mountain juniper, a tem- perate mesophilic conifer. First known in the region during the terminal Wisconsin, Rocky Mountain juniper became common in upland settings in the early Holocene and through the middle Holocene as well. Singleleaf pifion pine also began to expand northward during the early Holocene. Its distribution apparently remained south of Meadow Valley Wash (37?30'N) until at least 8.5 Ka (Mad- sen, 1973), but by 7 Ka, pinon pine had reached the vicinity of Danger Cave (41?N). Expansion of pinon pine woodlands is typically facilitated by corvid birds, which can carry the heavy pine nuts for distances of up to 22 km (Lanner, 1983). The rapid initial northward expansion is faster, however, than current esti- mates allow for pinon pine migration via corvid birds (Lanner, 1983; Wells, 1983). The initial spread of pinon pine northward may also have been aided by people who brought pine nuts with them on their journeys (Mehringer, 1986; Madsen and Rhode, 1990; Rhode and Madsen, 1998). The chronology of pinon pine migration is complicated by the Danger Cave remains, which are older than all other pifion pine records north of ~37?N in the central or eastern Great Basin. How it happened to arrive in the vicinity of Danger Cave by 7 Ka is difficult to trace in the ab- sence of any intervening records of earlier age. Utah juniper was common in the southern Bonneville Basin by the beginning of the Holocene, and it reached the Onaqui and Stansbury Mountains by about 6.6 Ka (Rhode and Madsen, 1995), but it was apparently a rather late arrival elsewhere in the northern Bonneville Basin. Because Utah juniper and pinon pine are rather closely associated today, why do they have dis- tinctly different Holocene migration histories? West et al. (1978) noted that the two taxa have somewhat different cli- matic requirements and tolerances, and perhaps these climatic factors account for the differences in timing and range. MIDDLE HOLOCENE (7 TO 4 KA) Vegetation records dating from the middle Holocene are rela- tively sparse but come from a variety of sources, including woodrat middens, archaeological sites, and pollen cores. These provide limited, but nevertheless useful, information on the timing of migration of certain taxa into the region, on the na- ture of vegetation community composition, and on climatic conditions during this interval. The general size and level of Great Salt Lake during the period is not known, but it is likely to have been low (Currey, 1980). As the middle Holocene began in the southern Bonneville Basin, conifers such as limber pine, bristlecone pine, and Rocky Mountain juniper, as well as mesophilic shrubs, had de- clined substantially in abundance or were restricted to higher elevations (Thompson, 1984). Pinon pine, Utah juniper, ponde- rosa pine (Pinus ponderosa P. Lawson & C. Lawson) and white fir became increasingly abundant. Thompson (1984, 1990) sug- gested that increased abundance of some of these taxa may re- flect increased amounts of summer precipitation from subtropi- cal sources, as well as warmer growing-season temperatures. Further west, at Gatecliff Shelter, singleleaf pinon pine ap- pears in the region by about 5.3 Ka (Rhode and Thomas, 1983) in a conifer woodland containing both Rocky Mountain juniper and Utah juniper (Thompson and Hattori, 1983). Using sedi- mentary evidence from Gatecliff Shelter and nearby Triple T Shelter, and closely following the arguments of Van Devender and Spaulding (1979), Davis (1983) proposed that the interval prior to 6300 years B.P. was relatively cool and winter-wet with relatively abundant snowfall but deficient in summer moisture. Davis thought that after this time until about 5000 years B.P., annual temperatures were somewhat warmer and summer pre- cipitation increased with often intense storms, leading him to support the "moist Altithermal" of Martin (1963). It is not cer- tain how far northward from central Nevada the enhanced stormy summer precipitation regime may have occurred. It is notable that pollen samples from Gatecliff and Triple T Shel- ters show a marked increase in representation of juniper and pine between 6 and 5 Ka (Thompson and Kautz, 1983). After 5000 years ago, according to Davis (1983), temperatures ap- parently cooled slightly and summer rainfall decreased in amount. Several middens dating to the middle Holocene are known from the northern Bonneville Basin and vicinity (Rhode, 2000b; see Figure 22). Upland areas in this region appear to have contained woodlands dominated by Rocky Mountain ju- niper, with a secondary component of pinon pine. Rocky Mountain juniper grew with pinon pine in the Toano Range, in the northern Pequop Range, and in the Cherry Creek Range; it is the only conifer known from the Silver Island Range. Sage- brush and rabbitbrush were the most common understory shrubs. In a few protected settings, such as the north-facing Ici- cle Cave in the Pequop Range, limber pine and common juni- per were also minor components of the vegetation, growing at relatively low elevations compared with today. Utah juniper was quite scarce in the northern Bonneville Basin. One record from the Toano Range contains Utah juniper remains dating from the middle Holocene, and it is also known from a midden in the Onaqui Range, east of the Bonneville Basin, dating to 6.6 Ka. The dominance of Rocky Mountain juniper, rather than the more thermophilic Utah juniper, may indicate that middle- Holocene climate in the uplands was not so warm and dry as is sometimes thought. Alternatively, it may suggest that growing- season precipitation derived from summer convective storms was more common in the Bonneville Basin than elsewhere in the Great Basin. Lowland vegetation patterns are better shown in pollen records. At Great Salt Lake, high ratios of shadscale and other chenopod shrubs relative to sagebrush and conifers prevailed NUMBER 33 351 from 7 Ka until about 5.5 Ka, suggesting that the early part of the middle Holocene was warmer and drier than today (Meh- ringer, 1977, 1985). Between 5.5 and 4 Ka, relative pollen ra- tios approached modern values, suggesting that climatic condi- tions were generally cooler and/or wetter in the latter part of the middle Holocene than in the first part. At Snowbird Bog in the Wasatch Range, Madsen and Currey (1979) found pollen evidence indicating that the period from 8 to 6 Ka was rela- tively warm and dry, whereas the following period until 5.2 Ka was warm but wetter (Figure 23). The Ruby Marsh sediment record documents lower-than-modern water levels and an ex- pansion of shadscale at the expense of sagebrush steppe vege- tation from ca. 7.4 to 5 Ka. Mission Cross Bog, in the Jarbidge Mountains, shows evidence of an increase in juniper woodland and shadscale steppe and a decline in subalpine woodlands from about 8 to 5 Ka (Thompson, 1984). LATE HOLOCENE (<4 KA) Great Salt Lake rose in elevation and expanded substantially in area at least twice during the late Holocene (Currey et al., 1984). The lake rose to an elevation of 1284 m, then receded, then advanced again to an elevation of 1287 m, flooding most of the Great Salt Lake Desert and forming a broad shallow lake that extended to the Utah-Nevada border. The midden record representing the late Holocene is more abundant than that representing the previous several thousand years (Figure 22), possibly because of better preservation of the younger middens, but also perhaps because woodrats may have been more abundant during the late Holocene than earlier. Mid- den records indicate that the flora of the region was largely modern by this time and that plant taxa were for the most part close to their modern distribution. Many of the midden samples have contents that are quite similar to the vegetation that grows around them today. Thompson (1984) found only rather modest vegetation changes in the Snake Range and southern Bonneville region, suggesting that late-Holocene climates were relatively stable, at least on the scale recorded in midden plant records (Thomp- son, 1984). The midden record from the northern Bonneville Basin (Rhode, 2000b) indicates that although modern vegeta- tion associations were largely in place by the late Holocene, the abundance of certain plant taxa within those vegetation associ- ations varied substantially through time (Figure 22). Some taxa only achieved their present importance in plant communities within the last 2000 years. The distribution of some taxa shifted in elevation in response to climatic variation. A prominent late-Holocene vegetation shift concerns the dis- tribution of Utah juniper and Rocky Mountain juniper. Utah ju- niper apparently arrived in northeastern Nevada, as noted pre- viously, during the middle to late Holocene. In some areas Utah juniper was a minor component in woodlands dominated by Rocky Mountain juniper (e.g., Cherry Creeks, Silver Island Range); but in other nearby areas, midden contents show it was the dominant woodland conifer (e.g., Goshute Range). These differences are probably related to differences in elevation or topographic position of the midden sites, as well as to other factors. Utah juniper is the most common juniper today in all of these areas, and Rocky Mountain juniper has become much less common in the last 2000 years. A similar pattern is shown in middens from Gatecliff Shelter, central Nevada, where Rocky Mountain juniper disappears from the midden record between 2.3 and 1.7 Ka (Thompson and Hattori, 1983). Several midden records from Antelope Island and the Prom- ontory Mountains, near Great Salt Lake, indicate that Utah ju- niper grew 50-100 m lower during the Neoglacial interval be- tween ca. 3.5 and 2.9 Ka than it occurs today (Rhode, 2000b). Lower-than-modern elevations of Utah juniper are also re- corded in middens dating to 2.3 and 1.0 Ka from the Cricket Mountains in the southernmost Bonneville Basin (Madsen, 1984). It is likely that Utah juniper retreated upslope from these relatively low elevations toward elevations where we see it today as the late-Holocene climate warmed and dried after the Neoglacial. Presumably the lower tree line of Utah juniper has migrated up- and downslope several times in response to climatic variation during the past few thousand years. Another example of late-Holocene range changes may be green mormon tea {Ephedra viridis Cov.). Green mormon tea is a common shrub in uplands throughout the region, including the vegetation surrounding midden localities on the western slope of the Goshute Range today, and it is found in modern midden samples in the region; but midden samples dating to 3.8-3.1 Ka lack green mormon tea. Other paleoecological records also indicate that green mormon tea was a late entry into the vegetation of the northern and central Great Basin, within the past 3000 years (e.g., Thompson and Kautz, 1983). It should be noted, however, that mormon tea is generally un- common in late-Holocene midden samples, suggesting that the lack of mormon tea in middens may be partly related to collec- tion habits of woodrats rather than abundance in the landscape. Additional late-Holocene vegetation changes of importance are shown in pollen records from the region. A pollen record from Crescent Spring (Mehringer, 1977, 1985) shows that sagebrush and conifer pollen became increasingly important after 4 Ka, suggesting overall greater effective moisture than during the middle Holocene, albeit with several warmer and/or drier intervals during the period since then. These warmer/ drier excursions occurred approximately 3.2, 2.5, and 1.8 Ka; substantial cool/wet episodes centered on 3.6, 2.7, 2.2, and 1.5 Ka (Mehringer, 1985). As noted in previous sections, these pe- riods correspond to drier and wetter periods indicated in the pollen record of the northern Great Basin (Wigand, 1987) and the northern Mojave Desert (Wigand, 1997b). The pollen record at Ruby Marsh indicates that, after ca. 4.5 Ka, "shadscale declined in importance in the valley bottom and marsh waters deepened, in turn reflecting cooler temperatures and/or greater precipitation" (Thompson, 1992:13). In addi- tion, an increase in sagebrush pollen and higher water levels at 352 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Ruby Marsh during the last 500 years indicate that "the coolest and/or moistest conditions since the early Holocene" presum- ably correlated with the Little Ice Age (Thompson, 1992:13). The late-Holocene record from Mission Cross Bog in the Jar- bidge Mountains shows an increase in pollen from spruce, fir, and pine after 5.0^4.3 Ka, with a peak around 3.6 Ka, suggest- ing this was the most mesic episode in the last 8000 years (Thompson, 1984). At Swan Lake (Bright, 1966), warm, dry middle-Holocene conditions became cooler by 3.1 Ka, with a lowering of the forest tree line and a decrease in sagebrush steppe. Wanner conditions apparently returned by 1.7 Ka, with increased sagebrush and less limber pine. Cooler late-Holocene climate from ca. 5.0 Ka onward is also reflected in the pollen from Snowbird Bog in the Wasatch Range (Figure 23; Madsen andCurrey, 1979). At Gatecliff Shelter, central Nevada, packrat midden and sedimentary evidence (Davis, 1983; Thompson and Hattori, 1983) indicate that a cool, winter-wet interval began rather sud- denly about 3400 years ago, during which time mormon tea be- came established in the region. The period between 3400 and 2800 years B.P. may have been cooler and moister than any time in the last 6000 years. These records also suggest that summer-dominant rainfall may have appeared between 1500 and 800 years ago. Pollen and plant macrofossil evidence from archaeological sites (Aikens, 1970; Harper and Alder, 1970, 1972; Kelso, 1970; Dalley, 1976) also document late-Holocene vegetation changes (see Currey and James, 1982). At Swallow Shelter and Hogup Cave, grass pollen began to increase ca. 1.5 Ka, again suggesting an increase in effective growing-season moisture. This increase in summer precipitation may have helped to sup- port the spread and success of the Fremont peoples' maize- based horticultural systems in the Bonneville Basin from ca. 1.5 to 0.9 Ka. Harper and Alder (1970, 1972) used the record of plant remains from Hogup Cave and Danger Cave to suggest that the last 1000-600 years have been among the driest and warmest of the entire Holocene. DISCUSSION: VEGETATION AND AQUATIC HISTORY IN THE BONNEVILLE BASIN In general outline, vegetation history, as represented by widespread pollen and macrofossil records, and aquatic history, as represented by the lake chronology and other records, appear to be mostly but not entirely concordant. Shallow saline lakes during the middle Wisconsin are matched with cold but dry limber pine woodlands. As Lake Bonneville began to rise from 28 to 20 Ka, evidence for more mesophilic montane plants be- came increasingly common. By the time of the Bonneville highstand ca. 16-15 Ka, very cold conditions are reflected in now subalpine conifers in the southern Bonneville Basin and in cold sagebrush steppe tundra in the north. The recession of Lake Bonneville from the Provo shoreline coincides with con- siderable evidence of expansion of plants requiring warmer and drier conditions than prevailed during the full glacial. How- ever, as noted above (see discussion of late Bonneville cycle in "Bonneville Basin and Eastern Great Basin"), the terrestrial vegetation record from the northern Bonneville Basin is not consistent with the timing and magnitude of certain models of Lake Bonneville recession (e.g., Oviatt et al., 1992). The postu- lated Younger Dryas rise of Lake Gilbert is also not strongly re- flected in the vegetation records, although available vegetation records are quite rare. Holocene fluctuations in vegetation ap- pear to coincide with lacustrine history of Great Salt Lake, al- though comparison is limited by the state of knowledge of Ho- locene lake history and by lags in response of vegetation to climate change. A number of reasons may account for discordance between lake history and vegetation history: 1. Differential sensitivity or rapidity of response of different systems to the same climatic factors (e.g., inertia of vegeta- tion to climate change). 2. Responses of different systems to different parameters within "climate space" (e.g., temperature versus precipita- tion; growing-season precipitation versus winter precipita- tion). 3. The presence of local or regional factors that affect one sys- tem but not the other (e.g., downcutting that lowered Lake Bonneville in the absence of climatic change; tectonics or isostatic rebound that altered the configuration of the lake basin; or altered growth tolerances of certain plants in rela- tion to atmospheric chemistry or competitive relationships with other plants). 4. Data gaps in one or both proxy records causing difficulties of correlation. 5. Uncertainties in how one or both proxy records are related to underlying climatic factors. Causes for some of the discordance between vegetation and aquatic history in the Bonneville Basin (e.g., vegetation change during the post-Provo recession and Gilbert Lake interval) re- main to be worked out in detail. The packrat midden record suggests a substantial north-to- south gradient in both floristic and vegetation representation existed through the most of the record. For one example, bris- tlecone pine was abundant in the south and nearly absent in the north, replaced by limber pine; for another, during the full gla- cial the southern area bore conifer forests while a cold sage- brush steppe prevailed in the north. How and why this gradient might have varied through time, and the location of substantial biotic discontinuities, are subjects for future biogeographic re- search. Summary The foregoing regional discussions highlight the wealth of empirical evidence now available to reconstruct the history of vegetation in the Great Basin. The combined and complemen- tary evidence found in pollen profiles, packrat middens, ar- NUMBER 33 353 chaeological sites, and other localities present a detailed record of broad regional trends and local variability of vegetation through time. The existing paleoenvironmental record for the last 35 Ka re- veals periods when vegetation changes occurred simulta- neously and in the same relative direction throughout the Great Basin. At other times, each of the subregions seems to have re- sponded quite differently. These differences directly reflect the relative influence of monsoonal versus Pacific storm systems during the Holocene and earlier interglacials, and the relative strength of polar air masses during glacials. Although the vegetation history of the Great Basin is marked by numerous high-frequency changes in climate, there are eas- ily recognizable crosscutting events that can be tracked in each subregion by tracing the response of indicator species. These species are characterized by wide spatial distribution and rela- tively rapid response to changes in climatic input. In addition, these species are easily identified and can be tracked in one or more of the most commonly used paleoenvironmental data sets mentioned above. To summarize some of the trends from a basinwide perspec- tive and to get a broader view of Great Basin vegetation history during the last 40,000 years, we compare what was happening in the three regions considered above during a dozen separate snapshot "moments" in the past. 40 KA The limited empirical record available from this time sug- gests that vegetation in much of the Great Basin and on the Snake River Plain was dominated by sagebrush or saltbush steppe, with a mosaic of sagebrush, montane brush, and patches of conifers in upland areas. Pollen records from the northeastern Great Basin suggest that conifers were not very abundant in the region (Beiswenger, 1991; Thompson, 1992), although they may have been more common further south (Thompson, 1984). The conifers consisted of limber pine and bristlecone pine at relatively low elevations in the eastern ba- sin, with mesophilic shrubs as understory or shrub layer asso- ciates. A cold and somewhat mesic interval is indicated. In the southern Great Basin, woodlands dominated by Utah juniper and mountain mahogany were present in upland areas, where present vegetation appears broadly similar, as well as in lower elevation areas now dominated by desert scrub species (Spaul- ding, 1985). According to Spaulding, the presence of certain desert plant species in these middens indicate that "winter minimum temperatures were within 2? to 3?C of today's val- ues" (Spaulding, 1985:30). The pollen record from the Bed and Breakfast locality in the Summer Lake basin indicates that sagebrush steppe was widespread. Juniper woodland was present as well but not as abundant as it was to be during the succeeding period. Lower grass pollen values than before 40 Ka and after 34 Ka indicate that conditions were drier during this period than either before or after. Sedge-dominated marshes together with abundant eutrophic algae {Botryococ- cus sp.) indicate that pluvial Summer Lake was low. The same may have been true of pluvial Lake Lahontan during this pe- riod. 32 KA In the northwestern Great Basin the onset of a series of cool, moist cycles from ~34 to 30 Ka resulted in the extension of patches of mixed-conifer woodland?comprising pines (prob- ably whitebark pine, western white pine {Pinus monticola Dou- glas ex D. Don), and lodgepole pine), fir, and occasional spruce?into relatively low elevations in mountainous terrain. More-continuous woodlands dominated by Utah juniper with an understory of sagebrush grew over much of the area, with a lower elevational range greater than at present. This expansion appears to have climaxed around 32-30 Ka and was followed by a dramatic return to drier conditions. In the southern Great Basin and northern Mojave Desert, subalpine conifers also grew at relatively low elevations in the mountains. Whereas white fir had been a substantial part of this woodland between 34 and 32 Ka, it declined in abundance after 32 Ka, leaving limber pine alone at these lower elevations, which suggests that climate was cool but becoming increasingly arid. Utah juniper was abundant or common from as low as ~ 750 m in elevation, growing in association with shadscale and rabbitbrush, to ele- vations as high as 1860 m (Spaulding, 1985). In the northeast- ern Great Basin, limber pine dominated the patchy mountain woodlands, accompanied by montane shrubs that indicate drier conditions than had occurred at 40 Ka. Sagebrush steppe con- tinued to prevail in valley floors and on the Snake River Plain. In the southern Bonneville Basin, bristlecone pine woodlands dominated the mountains, while sagebrush and shadscale occu- pied valley bottoms. 28 KA In the northwestern Great Basin, this time marks the begin- ning of a major drought stretching from ~28 to 26 Ka. It is im- mediately preceded and followed by regionally correlatable grass expansions in the woodrat midden pollen record at Pyra- mid Lake as well as in the sediments of Summer Lake. The drought is characterized by saltbush expansion and is reflected in lowered lake levels throughout the northwestern Great Ba- sin. In the southern Great Basin, Utah juniper continued to grow fairly abundantly at lower elevations, below about 1300 m; piiion pine is virtually unreported the region (Spaulding, 1985). Limber pine and white fir disappeared from the record of the area but may have retreated to widely scattered settings above about 2000 m in elevation. In the eastern Great Basin, sagebrush and shadscale steppe continued to dominate valley settings, but Lake Bonneville was beginning to rise, flooding portions of the valley habitat in that large basin. In montane settings, limber pine was declining in abundance and spruce 354 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES increased; montane mesophilic shrubs also increased in abun- dance. The climate appears to have become wetter and colder than previously. 22 KA Increasingly moister conditions beginning ~24 Ka were reaching a pre-glacial maximum climax by 22 Ka. A transition toward much colder temperatures, however, was continuing and affecting vegetation in all parts of the Great Basin and the northern Mojave Desert. In the northwest Great Basin, the ex- pansion of patchy woodlands, including subalpine whitebark pine, to elevations below 1380 m reflected moister conditions with a reliable wet season. Lake Lahontan was apparently ris- ing rapidly and was probably at the level of the Astor Pass sill (1222 m), but it had not yet reached the Darwin Pass sill (1265 m) to spill into the Carson Basin (Benson et al., 1995). Accord- ing to Benson et al. (1995), this rapid rise probably coincided with passage of the jet stream over the Lahontan Basin. In the southern Great Basin and northern Mojave Desert, cooler, wet- ter conditions are also indicated for this period, as subalpine conifer woodlands dominated by limber pine with some bristle- cone pine, and also containing white fir, descended to eleva- tions below 1600 m (Spaulding, 1985; Forester et al., 1996). In the eastern Great Basin, Lake Bonneville had also risen rapidly, and patchy, subalpine conifer woodlands dominated by spruce (in the north) and bristlecone pine (in the south) were present in upland environments. Cold but moister conditions are indicated in the Bonneville Basin and also in the Snake River Plain dur- ing that time (Beiswenger, 1991). 18KA At the height of the full glacial, vegetation throughout the Great Basin records cold and fairly dry climatic conditions. In the northwest Great Basin, these cold, dry glacial conditions fostered a mixed-shrub steppe comprising primarily sagebrush, saltbushes, and rabbitbrush with rare clumps of juniper in scat- tered sheltered areas. Whitebark pine, which had appeared be- tween 24 and 22 Ka, disappeared. To the south, cold and fairly dry conditions are also indicated by the widespread presence of limber pine and bristlecone pine, without white fir, in the re- gion's mountains, to elevations as low as 1600 m. Juniper woodlands retreated upward almost 450 m to elevations around 1200 m. Pinon pine was rare to absent in most of these wood- lands. In the northeastern Great Basin, a sagebrush steppe-tun- dra apparently existed in montane settings, with very limited patches of conifers including spruce and occasional limber pine. Southward, bristlecone woodlands prevailed in uplands. Sagebrush steppe dominated in lowland settings, while deepen- ing lakes existed in many of the valley basins. The vegetation is consistent with expectations drawn from global circulation models (COHMAP Members, 1988; Thompson et al., 1993) that suggest atmospheric circulation in the region at the time was dominated by cold, dry air moving southwestward from the continental glacial sheet. 12.5 KA This period marks the transition from the Pleistocene to the Holocene in the northwestern Great Basin. Whitebark pine re- appeared briefly at elevations around 1380 m on the west shore of pluvial Lake Lahontan around 12.5 Ka but disap- peared shortly thereafter. Although Utah juniper woodland re- expanded after the glacial maximum, it was sagebrush that ex- perienced its most dramatic expansion of the last 35 Ky during this period. By 11.5 Ka, however, severe drought had not only driven whitebark pine from the area but also had brought an end to the sagebrush expansion and encouraged the regional expansion of saltbushes as well. Pluvial Lake Lahon- tan dropped dramatically from its late-Pleistocene high at this time. The so-called "Younger Dryas," centered around 10.6 Ka, is evidenced by a renewed abundance of juniper wood- land and by an apparent still-stand of pluvial Lake Lahontan. The same pattern of change was occurring in the southern Great Basin and northern Mojave Desert. Limber pine re- treated up the mountainsides while juniper retreated from the valley floors. By 10 Ka, limber pine no longer grew below about 2300 m in elevation (Spaulding, 1985), and juniper was approaching its modern elevational distribution. In the north- eastern Great Basin, limber pine had expanded widely into lower montane settings. It also occurred at fairly low eleva- tions around the northwest side of the Bonneville Basin in a mosaic with sagebrush, prostrate juniper, and mesophilic shrubs. To the south, woodlands were dominated by bristle- cone pine. Thermophilous shrubs such as Atriplex spp. were just beginning to appear in the region, and these would expand after ca. 11 Ka as climate grew warmer and drier. It is at this time that plant communities were undergoing major restruc- turing as they moved into new areas on the landscape. What probably had been a mosaic of highly variable plant commu- nity patches occupying discrete microhabitats regulated by slope, aspect, and soils, as well as effective moisture, began the transformation into the more familiar vegetation zones of the Holocene. 9KA In the northwestern Great Basin, the early Holocene thermal maximum was characterized by major expansion of desert scrub vegetation around the margins of still-drying pluvial lake basins. In the Carson Sink, a subbasin of pluvial Lake Lahontan, extensive marshes were exploited by Native Ameri- cans for food and raw materials to make basketry, matting, and other items. Expansion and density of sagebrush steppe in the uplands reached its Holocene climax at this time. Low-eleva- tion outliers of Utah juniper woodland that had survived early Holocene droughts or had reexpanded during the Younger Dr- yas finally disappeared. Evidence of monsoonal penetration, which is weak in the northern Great Basin, is strong in the southern Great Basin and northern Mojave Desert during this period. The expansion of grasses without a corresponding ex- NUMBER 33 355 pansion of winter-rainfall-dependent plants suggests that sum- mer rainfall had increased. The appearance of hackberry in lo- cations indicating a proliferation of seeps along cliff-faces suggests the presence then of year-round water where it is not found today. Summer monsoons provided the additional rain- fall nurturing these springs. Black mat formation, whether it was related to increased spring discharge or soil formation (or perhaps to the increase in grasses), certainly reflects greater or- ganic production, which may have been stimulated by in- creased summer precipitation in combination with warm tem- peratures. In the eastern Great Basin, the early Holocene was marked by warmer and drier conditions than the terminal Wis- consin, but the degree of warming is not certain. Some records indicate substantial warming by ca. 9 Ka to levels as warm or warmer than today, whereas other records indicate that early Holocene climate was cooler and wetter than today, though warmer than the terminal Wisconsin. The limited vegetation record indicates that shadscale and xerophytic shrub associa- tions were well established in valley bottoms and around the margins of the Bonneville Basin, alongside sagebrush steppe. In montane settings, limber pine (in the north) and bristlecone pine (in the south) declined in abundance, and more-temperate conifers, such as Rocky Mountain juniper, began to dominate. Mesophilic mountain shrubs were declining in abundance and were being replaced by more thermophilous plants. 7KA Although the thermal maximum had passed, climatic condi- tions were still harsh, and severe regional drought character- ized the northwestern Great Basin. Desert scrub communities dominated the slopes surrounding dry playas. Eolian activity generated extensive, and very active, dune fields downwind of valley bottom playas (Mehringer and Wigand, 1986). Absence of vegetation cover prevented stabilization of these dunes for the next 3 Ka. In the southern Great Basin, little information exists from this period. However, the few woodrat middens and pollen records that do exist indicate periodic, severe drought. True Mojave Desert plant species, such as creosote bush and white burrobush, became dominant during this period. Their appearance reflects not only warmer temperatures but also es- tablishment of the current pattern of convective summer storms. On the western fringe of the Great Basin, pinon-juni- per woodland was forced to retreat into areas occupied by lim- ber pine and bristlecone pine near the top of the White Moun- tains. In the eastern Great Basin, warm and relatively dry conditions also prevailed, but a possible increase in growing- season precipitation from summer monsoons may have en- hanced the northward migration of pinon pine, ponderosa pine, douglas fir, and other temperate conifers. Rocky Mountain ju- niper was a major element of woodland forests in the northeast- ern Great Basin, while limber pine still persisted in protected localities at relatively low elevations. Utah juniper was com- mon in the southern Bonneville Basin. 5.5 KA This marks the termination of the long middle-Holocene drought in the northern Great Basin as a result of a dramatic in- crease in winter rainfall. Areas that had been dominated by greasewood just decades earlier were replaced by sagebrush steppe. Ephemeral ponds began to retain water year-round. Pe- riodic droughts still punctuated the next 1.5 Ka, but there was never a return to the conditions that had characterized the mid- dle Holocene. In the southern Great Basin and northern Mojave Desert, renewed growth of peat in the White River Valley, at Ash Meadows, and at Little Lake indicates a rejuvenation of springs in response to the sudden increase in rainfall. Juniper woodland seems also to have begun moving downslope again. Pollen records indicate that the eastern Great Basin at about this time underwent a similar change to relatively cooler and perhaps more winter-wet conditions, as sagebrush and conifers increased slightly at the expense of shadscale and saltbush. A number of middens dating to this moment suggest that montane woodlands in the northeastern Great Basin were dominated by Rocky Mountain juniper and pinon pine, with a small amount of newly immigrant Utah juniper present as well. A similar sit- uation apparently prevailed in the southern Bonneville Basin and into central Nevada, although Utah juniper was more com- mon in these latter areas. 3.5 KA The climax of the so-called "Neoglacial" or "Neopluvial" in the northwestern Great Basin occurred at this time. Juniper woodland expansion in south-central Oregon, which began dra- matically 500 years earlier, reached the first of its three greatest late-Holocene limits. Fostered by major increases in winter precipitation, regional water tables probably reached their late- Holocene maximum ~3.7 Ka. (Two additional expansions of woodland ~2.7 and 2.1 Ka occurred during what remained of the Neoglacial.) In the west-central Great Basin on the east slope of the Sierra Nevada, fir lowered its elevational distribu- tion and previously grassy meadows became sedge-dominated bogs that were soon invaded by birch {Betula sp.), and willow. The increase in winter precipitation was not as marked in the southern Great Basin or the northern Mojave Desert. Although each of the three major wet episodes of the Neoglacial are evi- denced in the south, they are much less pronounced there than in the north. This period is marked by the increased abundance of pinon pine on the landscape. It clearly becomes a codomi- nant in the semiarid woodland of the southern Great Basin and northern Mojave Desert. In the eastern Great Basin, the Neo- pluvial episode resulted in increased lake levels; greater abun- dance of sagebrush, grass, and conifers relative to chenopod shrubs; and Utah juniper occurring at elevations more than 100 m below its current distribution. Climatic conditions substan- tially cooler and more winter-wet than today are suggested. Rocky Mountain juniper still persisted in some montane set- tings, but Utah juniper and pinon pine were increasingly the dominant lower-montane conifers represented. 356 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1.5 KA This period marks the climax of late-Holocene warm, wet conditions. Beginning about 1.8 Ka, annual rainfall decreased substantially. At the same time, however, the peak of its annual distribution slipped later into the year, into summer. Grass abundance in sagebrush steppe increased greatly in the north- ern Great Basin at the same time that juniper woodland con- tracted. Ponds indicate very shallow eutrophic conditions, though they do not seem to dry seasonally. Although the cli- matic variability of this period was great, there was an overall increase in precipitation that peaked around 1.5 Ka. Bison fol- lowed the expansion of grasses into the sagebrush steppe. Fur- ther south in the west-central Great Basin, pifion pine expanded dramatically northward and downward in elevation, finally be- coming a major component of the semiarid woodland. In south- ern Nevada, pifion pine increased in abundance in the wood- land as well. Summer-shifted rainfall enabled shallow lakes to survive year-round despite a general decline in annual rainfall. In the eastern Great Basin, an increase in growing-season pre- cipitation at this moment is suggested by increases in grass pol- len in some records. Modern plant communities are fairly well established by now, including the presence of mormon tea through most of the region. The increase in summer precipita- tion may have also facilitated the success and spread of agricul- tural societies, known as the Fremont cultures, through much of the eastern Great Basin. 0.5 KA Drought beginning around 1 Ka continued with brief inter- ruptions until about 0.35 Ka. Retrenchment of juniper wood- land in the northern Great Basin and of pinon-juniper wood- land in the west-central Great Basin characterized this period. Increased fire activity marked the retreat of the forests. Ponds that had been wet for almost 5000 years occasionally dried dur- ing this period. This drought ended about 0.35 Ka with the Lit- tle Ice Age. Renewed expansion of juniper woodland in the northern and western Great Basin marked this period. It was only after the end of the Little Ice Age that pinon pine began its reexpansion in the northern and western Great Basin. Drought also characterized this period in the southern Great Basin and northern Mojave Desert. Saltbush expansion characterized many areas. Lakes became marshes until the wetter conditions of the Little Ice Age led to groundwater recharge and the return of shallow lakes. Expansion of woodland also characterized the Little Ice Age in southern Nevada with pifion pine expansion first occurring at the end of the period. In the eastern Great Ba- sin, the record from 500 years ago is somewhat mixed, possibly reflecting natural variability and the finer resolution that comes with recency. In some areas such as Ruby Marsh, this moment marked the beginning of the coolest and moistest period since the early Holocene, with increases in marsh level and expan- sion of sagebrush steppe at the expense of saltbush and shad- scale. On the other hand, some researchers (e.g. Harper and Al- der, 1972) find the last 1000 years to have been among the warmest and driest of the Holocene. Clearly, the climatic vari- ability inherent in the late Holocene should not be underesti- mated. Retrospective and Future Directions One of the landmark findings in plant ecology of the last 50 years is the widespread documentation that plants respond in highly individualistic fashion to both long- and short-term changes in climate and other ecological forcing mechanisms. This has led to the understanding that vegetation associations are rather loosely knit assemblages of plant taxa brought to- gether by historical contingency and individual response to en- vironmental factors, instead of strongly coevolved communi- ties held together in a dynamic equilibrium. In the first part of the twentieth century, such a position had been favored by plant ecologists such as Shreve and Hinckley (1937) and Gleason (1939), but it was countered by the prevailing climax commu- nity viewpoint of Clements (1916, 1936) and others (e.g., Weaver and Clements, 1938). However, improved understand- ing of the Late Quaternary paleovegetation record within the last 50 years has provided strong evidence supporting the indi- vidualistic, nonequilibrium model (Davis, 1986; Huntley, 1991). In response to climatic amelioration since the last glacial pe- riod, plant taxa did not respond as communities or vegetation formations, but singly and uniquely, with different tempos of migration. The rapid rates and directions of the shifts taken by various species may be explained in part by the concepts of thresholds, steady states, and multiple successional pathways described by Tausch et al. (1993). Some taxa responded by large-scale range shifts, others by more spatially restricted shifts in elevation (Thompson, 1988). Such migrations and range expansions did not commence with the Holocene. In the northern Bonneville Basin, for example, substantial expansion of shadscale began about 12 Ka, limber pine began at 13 Ka, and expansion of montane shrubs such as currant occurred be- fore that. Nor did substantial range shifts end in the early or middle Holocene, as the examples of western juniper, Utah ju- niper, and mormon tea demonstrate. In fact, some major range changes were not completed until the last few hundred years: e.g., singleleaf pifion pine in northwestern Nevada. An addi- tional finding is that plants that now do not occur together very frequently were often associated in the past, when climatic con- ditions were quite different than anything found today. In west- ern North America, the evidence of this was revealed by nu- merous packrat midden studies conducted in the 1970s and 1980s (e.g., Van Devender and Spaulding, 1979; Spaulding, 1981, 1985; Thompson, 1984; Cole, 1985; Betancourt, 1990; Van Devender, 1990a, 1990b). These insights set the modern baseline of understanding for a variety of approaches that seek to explain why changing plant distributions took the form that they did. As the above sum- mary indicates, changing climate has played the major role dur- ing the Late Quaternary (Whitlock and Bartlein, 1997) but not the only role. Paleoecologists quickly realized that it was not NUMBER 33 357 simply different tolerances to climate that affected how rapidly plant taxa changed their distributions through time. Available mechanisms for dispersal of disseminules and how well the plant is able to compete with existing vegetation are also im- portant components that figure into the concept of vegetation "inertia," or the occasional apparent lag of biogeographic range response after a climatic change (Cole, 1985; Markgraf, 1986; Thompson, 1990:232-233). In addition, some plants are rela- tively tolerant to climatic variation and, therefore, did not ex- hibit substantial range changes through widely variable condi- tions, whereas other taxa showed extensive range alterations (Barnosky, 1987; Thompson, 1988). Furthermore, the biogeo- graphic patterns of different taxa are also related to the high to- pographic variability in parts of western North America, espe- cially in the Great Basin, which can create a mosaic of microhabitats that may serve as suitable refugia for dispersed relict populations throughout a region (Whitlock and Bartlein, 1993). Chance may play an important role here too, for not all small, relict populations may survive in all the same microhab- itats, simply through stochastic processes. Nonclimatic factors such as atmospheric CO2 concentration and fire frequency (though these may correlate with climatic parameters) also play major roles in the fitness of plant taxa and their ability to sur- vive and spread. The population genetics of certain taxa in these settings may be critical to understanding biogeographic patterns (Hamrick et al., 1994). Nowak et al. (1994b) asserted that in the northwest- ern Great Basin, Utah juniper was present through much of the late Wisconsin and that it probably remained in the region be- cause its interbreeding populations were sufficiently variable genetically to enable the species to adapt via genetic heteroge- neity. Following example of Barnosky (1987), these research- ers called species with this conservative mode of geographic range movement "orthoselective," and they contrasted Utah ju- niper with pinon pine, which they considered to be a migratory species, one that expanded its range when climatic conditions are right. It should be noted, however, that the interpretation of Nowak et al. (1994b) regarding the sensitivity of juniper to cli- mate change was based upon a single 35,000-year-long wood- rat midden record of juniper from northwestern Nevada. It is clear even from that record that juniper may have been absent from the locality during periods of the most extreme climate conditions, including the glacial maximum and the middle-Ho- locene drought (Wigand and Nowak, 1992; see discussion of middle Holocene above in "Northern and Western Great Ba- sin"). The pollen record from this midden also suggests sub- stantial variation in the abundance of juniper during the last 35,000 years. For these reasons we disagree with the interpreta- tion of Nowak et al. (1994b) that Utah juniper was as compla- cent or unchanging as they suggest. We find the distinction drawn by Barnosky (1987), however, to be potentially useful in accounting for differential responses of taxa to climate change (see also Thompson, 1988). Recall, however, that Utah juniper had a very different his- tory in the northeastern Great Basin, where it was absent since before 50,000 years B.P. until approximately 6600 years B.P. In this region, Utah juniper must be considered migratory rather than orthoselective. Why should this be? Does it relate to the severe continentality of climate in the eastern Great Basin dur- ing the late Wisconsin? Is it a function of a possible lack of ge- netic variability in the eastern Great Basin compared with the western Great Basin (perhaps a lack of introgression with west- ern juniper)? Determining why different plant taxa respond dif- ferently to climatic changes is an important challenge for ecol- ogy in the twenty-first century, particularly as the global climate and atmospheric composition are undergoing rapid hu- man-induced change, with dramatic ecophysiological response. Fortunately, new genetic techniques are rapidly being devel- oped to allow studies of these issues on fossil materials, pro- viding a longer-term perspective of similar changes in the past. The view that Great Basin plant communities are, in large measure, products of historical circumstance (a corollary to the recognition that plant taxa have individualistic histories) is partly supported by the fossil record but is also tempered by it. Certain combinations of plants often appear to recur and persist in dynamic equilibria through long periods of time, as appro- priate environmental conditions allow. The composition of sagebrush steppe, for example, appears largely unchanged since the terminal Wisconsin and possibly since the middle Wisconsin. Indeed, some researchers see marked long-term sta- bility in many steppe and woodland communities in the eastern Great Basin (Thompson, 1988, 1992; Thompson et al., 1994; Moutoux and Davis, 1995; Davis, 2002). Again, this is not to say that the sagebrush steppe formation did not respond to past climate change; it clearly has, primarily through major shifts in its geographical distribution. In any event, the interplay of indi- vidual histories and tolerances that led to long-term associa- tions or recurrence of associated taxa remains to be explored more fully. Certain plants may have altered their climatic tolerances dur- ing the last 30,000 years, affecting their utility as proxy records for climatic change. Limber pine may have substantially in- creased its water use efficiency after the late Wisconsin in re- sponse to rising concentrations of atmospheric CO2 (Van de Water et al., 1994). If so, it suggests that late-Pleistocene popu- lations were more susceptible to water stress from warm, dry (or cold, dry) growing seasons than are modern populations. If not taken into account, this factor could result in substantial un- derestimation of the true rainfall and temperature values for the southern Great Basin and the northern Mojave Desert during the late Pleistocene, as projected by Forester et al. (1996). Vari- ous species of saltbush also may have altered their tolerances, as the genus apparently went through substantial genetic reas- sortment during the latest Pleistocene and Holocene when it re- colonized valley bottoms vacated by Lakes Bonneville and Lahontan (Stutz, 1978). The effects of genetic variability upon 358 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the climatic tolerances of Atriplex species are currently un- known. In the last few years, some studies of past environments and climates have begun to move beyond documentation of the his- torical sequence of change in vegetation and climate toward the relationships among climatic input, vegetation response, land- scape change, and feedbacks into vegetation change. Indeed, the requirements of current paleoenvironmental studies are leading the field in establishing the climatic limits of many plant species (e.g., Bartlein et al., 1998; Thompson et al., 2000). One recent model that relates climate and plant distributions (Thompson et al., 2000; cf. also Bartlein et al., 1998) uses a geographic grid of current climatic parameters and overlays it onto the current distribution of some of the more common plant species. The model then derives the climatic tolerances of individual plant species through the correspondence of plant species with the modern analogue climate. Once these parameters have been de- rived they are used to estimate past climates from dated assem- blages of plant macrofossils at specific points on the landscape. Contemporaneous, spatially distributed assemblages of macro- fossils (e.g., dated woodrat midden strata) can then be used to generate isobars of past rainfall and temperature or other cli- matic parameters as modeled. Care should be taken in interpreting the results of such mod- els, however. Bartlein et al. (1998) point out several possible discrepancies between model predictions and actual plant dis- tributions. One potential problem is that plant distributions may be constrained not by climatic forces but by competition with associated or neighboring taxa. For example, Thompson (1990) suggested that the lower limit of limber pine in the mountains of the eastern Great Basin often may be constrained more by competition from pinon pine than by climatic factors. This would not have been the case before the arrival of pinon pine toward the end of the early Holocene. Consequently the forces constraining a plant species' distribution may vary through time?and, indeed, through the range of the plant?in ways that may not be related directly to climate. Another potential problem is the use of modern plant distri- butions to establish climatic limits: climatic limits of what? Seedling establishment? Survival of mature individuals? Re- generating plant associations? A modern plant community is not unlike a medieval manuscript that has been reused numer- ous times; each successive use of the palimpsest manuscript is visible. Likewise, modern species distributions combine cur- rently reproducing populations, relict stands established during previous climatic regimes, and effects of disturbance factors (e.g., insect infestations, fire, pollutants). For example, the old- growth forests of the Sierra Nevada contain large white fir and red fir that were established during cooler, wetter conditions 200-300 years ago. Beneath this canopy lie ponderosa pine and/or Jeffrey pine (Pinus Jeffreyi Grev. & Balf.) established during warmer, drier conditions 100 years ago. Finally, there are saplings of red fir and white fir that were established during the wetter conditions of the late 1970s and early 1980s. A simi- lar pattern can be seen in the semiarid woodlands of the north- ern Great Basin where stands of Utah juniper, established as far back as 700-800 years ago, now stand side by side with pinon pines less than 150 years old (Tausch et al., 1981). Testing of the model by predicting the geographic distribution of a partic- ular species based upon current climates would not reveal the problem. Because the data used to generate the model include modern plant distributions affected by both modern and past climates, the model will predict the current distribution of both regenerating and relict plant populations on the current land- scape. Therefore the model does not predict the distribution of currently active, regenerating forests. As noted above, a third problem that may affect use of such models for reconstruction of Pleistocene and early Holocene climates is the uniformitarian assumption that plant taxa have the same response to present-day climate as they did in the past. Anatomical evidence has shown that plants may not have had a stable response to climatic factors?e.g., the change in stomatal density of limber pine from the Pleistocene to Ho- locene and its relationship to atmospheric CO2 and water use efficiency (Van de Water et al., 1994). Decreased water use ef- ficiency during the late Wisconsin could have resulted in sub- stantially greater plant sensitivity to climatic conditions of that time. Therefore, estimates of Pleistocene climates obtained by using modern analogue climates that were derived from mod- ern limber pine distribution might be in error (cf. Rhode and Madsen, 1995). In the future, the use of other corroborating proxy data sets and refinement of our understanding of the physiology of the more important indicator plant species will allow us to resolve some of these problems. Discussion of Late Quaternary vegetation history would not be complete without reference to the effects of human popula- tions upon the distributions of individual plant species and the composition and structure of modern and past plant communi- ties. Two areas of potential effect are especially germane: (1) the role of human land management practices in fire histories, and (2) humans as agents of substantial range expansion of cer- tain taxa. During the last 50 years, Federal land managers have continually discussed the influences of grazings by sheep and cattle and of fire suppression in forming today's rangeland and woodland plant communities (Miller and Wigand, 1994). In- deed, some recognition has even been made of the role of Na- tive Americans' use of fire in shaping the environment as well (Barrett and Arno, 1982). The roles played by Native American populations (if any) in shaping Great Basin vegetation history, however, remain to be much more fully explored. With respect to fire history, the paleoenvironmental record indicates that climate is still a prime mover (perhaps the prime mover) in effecting the frequency and magnitude of fire (e.g., Wigand, 1987; Swetnam, 1990; Swetnam and Betancourt, 1990). A strong deterministic link is clear between climate, fuel build-up, cycles of drought, and the resulting occurrence of fire. That human land use practices are important in this dy- namic is also obvious, as amply demonstrated by the historic NUMBER 33 359 record, but the importance of fire in prehistoric Native Ameri- can "range management" in the Great Basin is, as yet, rather poorly understood (Stewart, 1954, 1955a, 1955b, 1956, 1963, 1982; Downs, 1966; Mellars, 1976; Lewis, 1977, 1982, 1985; Barrett and Arno, 1982; Hemphill, 1983; Fowler, 1996). Pre- historic peoples may have substantially affected fire histories on a local scale. Effects by humans may have been important under certain threshold conditions of population, fuel buildup, and drought, but at present we do not know what thresholds (if any) may have existed or whether the effects were substantial. Interregional correlation among the histories of climate, vege- tation, and fire may serve to show that climate is the driving force in regional fire histories in the prehistoric and possibly historic periods and that prehistoric human activities were a relatively minor element in regional fire histories (e.g., Swet- nam and Baisan, 1996). A great deal of additional research is needed to determine how the dramatic changes in land use practices (including fire suppression) that occurred during the last 1000 years may have affected plant community distribu- tion and structure in the context of changing climates. The role of people as disseminators of plant species during the historic period is also clear with regard to cheatgrass (Bro- mus tectorum L.) and other range weeds as well as exotic tree and shrub species. The role of Native Americans in the spread of native plants is less clear. The ethnographic record shows that economically valuable plants were often managed and pruned, collecting areas were cleared or burned or irrigated, and in some cases plants may have been artificially trans- planted or sown to increase their range (Steward, 1930, 1938; Downs, 1966; Fowler, 1986, 1996). The magnitude and antiq- uity of these practices is not known, however (Winter, 1976; Winter and Hogan, 1986). As noted above, people may have had a role in the rapid early Holocene expansion of pifion pine (Mehringer, 1986; Madsen and Rhode, 1990; Rhode and Mad- sen, 1998). This speculation is based in part upon the rapid rate of pifion pine expansion as indicated by the paleobotanical re- cord compared with the presumed speed of its expansion based on an avian dispersal model, which seems too slow to account for the rate of postglacial reexpansion into the Great Basin. The available archaeological record indicates that early Holocene people consumed and transported pifion pine nuts (Rhode and Madsen, 1998), and mobile human groups could be a much faster agent of seed dispersal than corvids. But it is not certain whether the avian dispersal model is accurate, nor is the early Holocene pifion pine distribution or its rate of expansion known with enough certainty to completely rule out the avian model. In addition, for any attempt at dissemination of such species to be successful, climatic conditions would have to be favorable. The paleoenvironmental record indicates that pinon pine expansion can be quite rapid, and it is not necessary for human intervention to be involved (Tausch et al, 1981; and see above). The rapid spread of pifion pine during the last 150 years presumably did not have, nor did it need, the benefit of human intervention. It is likewise probable that pifion pine's expansion in the northwestern Great Basin 1.6 Ka was a re- sponse to natural processes (i.e., climate) because many other plant species including grasses, juniper, saltbush, and sage- brush responded as well (Wigand, 1997b). The evidence pre- sented in the discussion above suggests that pifion pine has re- sponded as rapidly to stressful climates as it apparently did to favorable climates. Retreat of pifion pine woodland after 0.9 Ka was as rapid as its expansion around 1.6 Ka in the northern Great Basin. The historic expansion of western juniper in the northern Great Basin via animal vectors of seed dispersal has been as dramatic as any prehistoric expansion of plant species. These cases confirm the premise that if conditions are favor- able, and appropriate agents of dispersal are available, the plant species will spread. Native Americans may have been appro- priate agents of dispersal for some plant species, substantially affecting their ranges, but the role of native peoples in the dis- tribution of pifion pine or other plants in the Great Basin re- mains uncertain and, indeed, may have been quite limited. In any case, and as Whitlock and Bartlein (1997) and Pren- tice et al. (1991) have previously observed in other areas, most of the dramatic changes in plant distributions and community structure observable in the Great Basin during the Late Quater- nary were ultimately constrained by climatic causes?although mediated by a host of more local, proximate factors. As our un- derstanding of plant responses to past climates, landscape change, and the role of disturbance phenomena (including fire, plant diseases, insect infestations, and human manipulation) is refined, the lines between paleoecology, modern plant ecology, plant physiology, and genetics will become ever more blurred. These are directions for the next half century of study. Literature Cited Adam, D.P., J.P. Bradbury, W.E. Dean, J.V. Gardner, and A.M. Sarna-Wojcicki 1995. Report of 1994 Workshop on the Correlation of Marine and Terres- trial Records of Climate Changes in the Western United States. United States Geological Survey Open-File Report, 95-34: 92 pages. United States Department of the Interior, United States Geological Survey. Aikens, CM. 1970. Hogup Cave. University of Utah Anthropological Papers, 93: 286 pages. 1984. Archaeology of Oregon. 134 pages. Portland, Oregon: United States Department of the Interior, Bureau of Land Management. Albee, B.J., L.M. Shultz, and S. Goodrich 1988. Atlas of the Vascular Plants of Utah. Occasional Publication of the Utah Museum of Natural History, 7: 670 pages. Allen, B.D., and R.Y. Anderson 1993. Evidence from Western North America for Rapid Shifts in Climate 360 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES during the Last Glacial Maximum. Science, 260:1920-1923. Antevs, E. 1938. Rainfall and Tree Growth in the Great Basin. Special Publication of the Carnegie Institution of Washington, 469: 97 pages. 1948. Climatic Changes and Pre-White Man. In The Great Basin, with Emphasis on Postglacial Times. Bulletin of the University of Utah, 38(20):167-191. Axelrod, D.I. 1939. A Miocene Flora from the Western Border of the Mohave Desert. Publications of the Carnegie Institution of Washington, 516:1-129. 1944. The Alvord Creek Flora (Oregon). Publications of the Carnegie In- stitution of Washington, 553:263-284. 1948. Climate and Evolution in Western North America during Middle Pliocene Time. Evolution, 2:127-144. 1990. Age and Origin of Subalpine Forest Zone. Paleobiology, 16: 360-369. Axelrod, D.I., and P.H. Raven 1985. Origins of the Cordilleran Flora. Journal of Biogeography, 12: 21-47. Barnosky, C.W. 1987. Response of Vegetation to Climate Changes of Different Duration in the Late Neogene. Trends in Ecology and Evolution, 2:247-250. Barrett, S.W., and S.F. Arno 1982. Indian Fires as an Ecological Influence in the Northern Rockies. Journal of Forestry, 80:647-651. Bartlein, P.J., K.H. Anderson, P.M. Anderson, M.E. Edwards, C.J. Mock, R.S. Thompson, R.S. Webb, T. Webb III, and C. Whitlock 1998. Paleoclimate Simulations for North America over the Past 21,000 Years: Features of the Simulated Climate and Comparisons with Pa- leoenvironmental Data. Quaternary Science Reviews, 17:549-585. Beiswenger, J.M. 1991. Late Quaternary Vegetational History of Grays Lake, Idaho. Ecolog- ical Monographs, 61:165-182. Benson, L., and R.A. Darrow 1981. Trees and Shrubs of the Southwestern Deserts. 416 pages. Tucson: University of Arizona Press. Benson, L.V., D.R. Currey, R.I. Dorn, K.R. Lajoie, C.G. Oviatt, S.W. Robin- son, G.I. Smith, and S. Stine 1990. Chronology of Expansion and Contraction of Four Great Basin Lake Systems during the Past 35,000 Years. Palaeogeography, Palaeocli- matology, Palaeoecology, 78:241-286. Benson, L.V., D. Currey, Y. Lao, and S. Hostetler 1992. Lake-size Variations in the Lahontan and Bonneville Basins be- tween 13,000 and 9000 14C yr B.P. Palaeogeography, Palaeoclima- tology, Palaeoecology, 95:19-32. Benson, L.V., M. Kashgarian, and M. Rubin 1995. Carbonate Deposition, Pyramid Lake Subbasin, Nevada, 2: Lake Levels and Polar Jet Stream Positions Reconstructed from Radiocar- bon Ages and Elevations of Carbonates (Tufas) Deposited in the Lahontan Basin. Palaeogeography, Palaeoclimatology, Palaeo- ecology, 117:1-30. Benson, L.V., J.P. Smoot, M. Kashgarian, A. Sarna-Wojcicki, and J.W. Burdett 1997. Radiocarbon Ages and Environments of Deposition of the Wono and Trego Hot Springs Tephra Layers in the Pyramid Lake Subba- sin, Nevada. Quaternary Research, 47:251-260. Benson, L.V., and R.S. Thompson 1987. The Physical Record of Lakes in the Great Basin. In W.F. Ruddiman and H.E. Wright, Jr, editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, K-3: 241-260. Boulder, Colorado: Geological Society of America. Benson, L.V., L.D. White, and R.O. Rye 1996. Carbonate Deposition, Pyramid Lake Subbasin, Nevada, 4: Compar- ison of the Stable Isotope Values of Carbonate Deposits (Tufas) and the Lahontan Lake-Level Record. Palaeogeography, Palaeoclima- tology, Palaeoecology, 122:45-76. Betancourt, J.L. 1990. Late Quaternary Biogeography of the Colorado Plateau. In J.L. Bet- ancourt, T.R. Van Devender, and P.S. Martin, edittors, Packrat Mid- dens: The Last 40,000 Years of Biotic Change, pages 259-292. Tucson: University of Arizona Press. Betancourt, J.L., T.R. Van Devender, and P.S. Martin, editors 1990. Packrat Middens: The Last 40,000 Years of Biotic Change. 467 pages. Tucson: University of Arizona Press. Billings, W.G. 1951. Vegetational Zonation in the Great Basin of Western North America. In Les Bases ecologiques de la regeneration de la vegetation des zones arides. International Union de la Societe Biologique, Series 5,9:101-122. Bond, G., and R. Lotti 1995. Iceberg Discharges into the North Atlantic on Millennial Time Scales during the Last Glaciation. Science, 267:1005-1010. Born, S.M. 1972. Late Quaternary History, Deltaic Sedimentation, and Mudlump Formation at Pyramid Lake, Nevada, 97 + xi pages. Reno, Nevada: Center for Water Resources Research, Desert Research Institute. Bradbury, J.P., R.M. Forester, and R.S. Thompson 1989. Late Quaternary Paleolimnology of Walker Lake, Nevada. Journal of Paleolimnology, 1:249-267. Bright, R.C. 1966. Pollen and Seed Stratigraphy of Swan Lake, Southeastern Idaho: Its Relation to Regional Vegetational History and to Lake Bonneville History. Tebiwa, 9(2): 1-47. Bright, R.C, and O.K. Davis 1982. Quaternary Paleoecology of the Idaho National Engineering Labo- ratory, Snake River Plain, Idaho. American Midland Naturalist, 108(l):21-33. Brown, R.W. 1949. Pliocene Plants from Cache Valley, Utah. Journal of the Washington Academy of Science, 39:224-229. Brown, W.J., S.G. Wells, Y. Enzel, R.Y. Anderson, and L.D. McFadden 1990. The Late Quaternary History of Pluvial Lake Mojave-Silver Lake and Soda Lake Basins, California. In At the End of the Mojave: Quaternary Studies in the Eastern Mojave Desert, pages 55-72. 1990 Mojave Desert Quaternary Research Center Symposium, May 18-21, 1990. Special Publication of the San Bernardino County Mu- seum Association. Byrne, R., C. Busby, and R.F. Heizer 1979. The Altithermal Revisited: Pollen Evidence from the Leonard Rock- shelter. Journal of California and Great Basin Anthropology, 1(2): 280-294. Clark, P.U., and P.J. Bartlein 1995. Correlation of Late Pleistocene Glaciation in the Western United States with North Atlantic Heinrich Events. Geology, 23(6):483- 486. Clements, F.E. 1916. Plant Succession: Analysis of the Development of Vegetation. Pub- lications of the Carnegie Institute of Washington, 242:1-512. 1936. Nature and Structure of the Climax. Journal of Ecology, 24:252- 284. Cohen, A.S., M. Palacios, R.M. Negrini, P.E. Wigand, and D.B. Erbes 2000. A Paleoclimate Record for the Past 250,000 Years from Summer Lake, Oregon, U.S.A., II: Sedimentology, Paleontology, and Geo- chemistry. Journal of Paleolimnology, 24(2): 151-182. COHMAP Members 1988. Climatic Changes of the Last 18,000 Years: Observations and Model Simulations. Science, 241:1043-1052. Cole, K.L. 1985. Past Rates of Change, Species Richness, and a Model of Vegeta- tional Inertia in the Grand Canyon, Arizona. Quaternary Research, 25:392^100. NUMBER 33 361 Currey, D.R. 1980. Coastal Geomorphology of Great Salt Lake and Vicinity. In J.W. Gwynn, editor, The Great Salt Lake: A Scientific, Historical and Economic Overview. Bulletin of the Utah Geological and Mineral- ogical Survey, 116:69-82. Salt Lake City. 1990. Quaternary Paleolakes in the Evolution of Semidesert Basins, with Special Emphasis on Lake Bonneville and the Great Basin, USA. Palaeogeography, Palaeoclimatology, Palaeoecology, 76:189-214. Currey, D.R., G. Atwood, and D.R. Mabey 1984. Major Levels of Great Salt Lake and Lake Bonneville. Utah Geo- logical and Mineralogical Survey Map, 73. Salt Lake City. Currey, D.R., and S.R. James 1982. Paleoenvironments of the Northeastern Great Basin and Northeast- ern Basin Rim Region: A Review of Geological and Biological Evi- dence. In D.B. Madsen and J.F. O'Connell, editors, Man and Environment in the Great Basin. Papers of the Society for American Archaeology, 2:27-52. Washington, D.C. Currey, D.R., and D.B. Madsen 1974. Holocene Fluctuations of Great Salt Lake. American Quaternary Association Programs with Abstracts, 3:74. Currey, D.R., and C.G. Oviatt 1985. Durations, Average Rates, and Probable Causes of Lake Bonneville Expansions, Still-Stands, and Contractions during the Last Deep- Lake Cycle, 32,000 to 10,000 Years Ago. In P.A. Kay and H.F. Diaz, editors, Problems of and Prospects for Predicting Great Salt Lake Levels: Papers from a Conference Held in Salt Lake City, March 26-28, 1985, pages 9-24. Salt Lake City: University of Utah Center for Public Affairs and Administration. Dalley, G.F. 1976. Appendix II: Palynology of the Swallow Shelter Deposits. In G.F. Dalley, Swallow Shelter and Associated Sites. University of Utah Anthropological Papers, 96:171-174. Salt Lake City. Davies, K.G., A.M. Johnson, and D.O. Johansen, editors 1961. Peter Skene Ogden's Snake Country Journal 1826-27. 255 pages. London: The Hudson's Bay Record Society. Davis, J.O. 1978. Quaternary Tephrochronology of the Lake Lahontan Area, Nevada and California. Research Paper of the Nevada Archaeological Sur- vey, 7: 137 pages. 1982. Bits and Pieces: The Last 35,000 Years in the Lahontan Area. In D.B. Madsen and J.F. O'Connell, editors, Man and Environment in the Great Basin. Papers of the Society for American Archaeology, 2:53-75. Washington, D.C. 1983. Geology of Gatecliff Shelter: Sedimentary Facies and Holocene Cli- mate. In D.H. Thomas, The Archaeology of Monitor Valley, 2: Gate- cliff Shelter. Anthropological Papers of the American Museum of Natural History, 59(l):64-87. New York. 1985. Sediments and Geological Setting of Hidden Cave. In D.H. Thomas, The Archaeology of Hidden Cave, Nevada Anthropological Papers of the American Museum of Natural History, 61 (l):80?103. New York. Davis, M.B. 1986. Climatic Instability, Time Lags, and Community Disequilibrium. In J. Diamond and T.J. Case, editors, Community Ecology, pages 269- 284. New York: Harper and Row. Davis, O.K. 1981. Vegetation Migration in Southern Idaho during the Late-Quaternary and Holocene. 252 pages. Doctoral dissertation, University of Min- nesota, St. Paul. 2002. Late Neogene Environmental History of the Northern Bonneville Basin: A Review of Palynological Studies. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems His- tory. Smithsonian Contributions to the Earth Sciences, 33:295-307. Davis, O.K., J.C. Sheppard, and S. Robertson 1986. Contrasting Climate Histories for the Snake River Plain, Idaho, Re- sulting from Multiple Thermal Maxima. Quaternary Research, 26: 321-339. DeBolt, A.M., and B. McCune 1995. Ecology of Celtis reticulata in Idaho. Great Basin Naturalist, 55: 237-248. Deevey, E.S., Jr. 1949. Biogeography of the Pleistocene, Part I: Europe and North America. Bulletin of the Geological Society of America, 60:1315-1416. Downs, J.F. 1966. The Significance of Environmental Manipulation in Great Basin Cultural Development. In W.L. d'Azevedo, W.A. Davis, D.D. Fowler, and W. Suttles, editors, The Current Status of Anthropologi- cal Research in the Great Basin: 1964. Desert Research Institute Technical Report, Series S-H, Social Sciences and Humanities Pub- lications, 1:39-56. Reno, Nevada. Dugas, D.R 1998. Late Quaternary Variations in the Level of Paleo-Lake Malheur, Eastern Oregon. Quaternary Research, 50(3):276-282. Eardley, A.J., and V. Gvosdetsky 1960. Analysis of Pleistocene Core from Great Salt Lake, Utah. Geologi- cal Society of America Bulletin, 71:1323-1344. Eardley, A.J., R.T. Shuey, V. Gvosdetsky, W.P. Nash, M. Dane Picard, D.C. Grey, and G.J. Kukla 1973. Lake Cycles in the Bonneville Basin, Utah. Geological Society of America Bulletin, 84(1 ):211-216. Eiselt, B.S. 1997. Fish Remains from the Spirit Cave Paleofecal Material: 9,400 Year Old Evidence for Great Basin Utilization of Small Fishes. Quarterly of the Nevada Historical Society, 40(1): 117-139. Ely, L.L., Y Enzel, V.R. Baker, and D.R. Cayan 1993. A 5000-Year Record of Extreme Floods and Climate Change in the Southwestern United States. Science, 262:410-412. Enzel, Y, W.J. Brown, R.Y Anderson, L.D. McFadden, and S.G. Wells 1992. Short-Duration Holocene Lakes in the Mojave River Drainage Ba- sin, Southern California. Quaternary Research, 38(l):60?73. Enzel, Y, D.R. Cayan, R.Y. Anderson, and S.G. Wells 1989. Atmospheric Circulation during Holocene Lake Stands in the Mojave Desert: Evidence of Regional Climate Change. Nature, 341:44?48. Feng, X., and S. Epstein 1994. Climatic Implications of an 8000-Year Hydrogen Isotope Time Se- ries from Bristlecone Pine Trees. Science, 265:1079-1081. Forester, R.G., J.P. Bradbury, C. Carter, A.B. Elvidge, M.L. Hemphill, S.C. Lundstrom, S.A. Mahan, B.D. Marshall, L.A. Neymark, J.B. Paces, S.E. Sharpe, J.F. Whelan, and RE. Wigand 1996. Synthesis of Quaternary Response of the Yucca Mountain Unsatur- ated and Saturated Zone Hydrology to Climate Change. 1996 Mile- stone Report, 3GCA102M: 110 pages. United States Geological Survey, Yucca Mountain Project Branch. Fowler, C.S. 1986. Subsistence. In W.L. d'Azevedo, volume editor, Great Basin. Hand- book of North American Indians, vol. 11:64-97. W.C. Sturtevant, general editor.Washington, D.C: Smithsonian Institution. 1996. Historical Perspectives on Timbisha Shoshone Land Management Practices, Death Valley, California. In E.J. Reitz, L.A. Newsom, and S.J. Scudder, editors, Case Studies in Environmental Archaeology, pages 87-101. New York: Plenum Press. Fremont, J.C. 1845. Report of the Exploring Expedition to the Rocky Mountains in the Year 1842, and to Oregon and California in the Years 1843?1844. 427 pages. Washington, D.C: Goles and Seaton. Fritts, H.C. and W. Xiangdig 1986. A Comparison between Response-Function Analysis and Other Re- gression Techniques. Tree-ring Bulletin, 46:31-46. Gehr, K.D. 1980. Late Pleistocene and Recent Archaeology and Geomorphology of 362 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the South Shore of Harney Lake, Oregon. Master's thesis, Portland State University, Portland, Oregon. [Unpublished.] Gleason, H.A. 1939. The Individualistic Concept of the Plant Association. American Midland Naturalist, 21:92-l 10. Graumlich, L.J. 1993. A 1000-Year Record of Temperature and Precipitation in the Sierra Nevada. Quaternary Research, 39:249?255. Graybill, D.A., M.R. Rose, and F.L. Nials 1994. Tree-Rings and Climate: Implications for Great Basin Paleoenviron- mental Studies. High-Level Radioactive Waste Management Confer- ence: Proceedings of the Fifth Annual International, Sponsored by the American Society of Civil Engineers and the American Nuclear Society; Co-sponsored by the American Chemical Society... fet al.J in Cooperation with U.S. Department of Energy ... fet al.J, volume 4:2569-2573. La Grange Park, Illinois: American Nuclear Society; New York: American Society of Civil Engineers. Grayson, D.K. 1993. The Desert's Past: A Natural Prehistory of the Great Basin. 356 pages. Washington, D.C.: Smithsonian Institution Press. 1998. Moisture History and Small Mammal Community Richness during the Latest Pleistocene and Holocene, Northern Bonneville Basin, Utah. Quaternary Research, 49:330-334. Hamrick, J.L., A.F. Schnabel, and P.V. Wells 1994. Distribution of Genetic Diversity within and among Populations of Great Basin Conifers. In K.T. Harper, L.L. St. Clair, K.H. Thorne, and W.M. Hess, editors, Natural History of the Colorado Plateau and Great Basin, pages 147-162. Niwot, Colorado: University Press of Colorado. Hansen, H.P. 1947. Postglacial Vegetation of the Northern Great Basin. American Jour- nal of Botany, 34(3): 164-171. Hardman, G., and O.E. Reil 1936. The Relationship between Tree-Growth and Stream Runoff in the Truckee River Basin, California-Nevada. University of Nevada Ag- ricultural Experiment Station Bulletin, 141: 38 pages. Harper, K.T., and G.M. Alder 1970. Appendix I: The Macroscopic Plant Remains of the Deposits of Hogup Cave, Utah, and Their Paleoclimatic Implications. In CM. Aikens, Hogup Cave. University of Utah Anthropological Papers, 93:215-240. Salt Lake City. 1972. Paleoclimatic Inferences concerning the Last 10,000 Years from a Resampling of Danger Cave, Utah. In D.D. Fowler, editor. Great Basin Cultural Ecology. Desert Research Institute Publications in the Social Sciences, 8:13-23. Reno. Hastings, J.R., and R.M. Turner 1965. The Changing Mile: An Ecological Study of Vegetation Change with Time in the Lower Mile of an Arid and Semiarid Region. 317 pages. Tucson: University of Arizona Press. Harvey, A.M., RE. Wigand, and S.G. Wells. 1999. Response of Alluvial Fan Systems to the Lake Pleistocene to Ho- locene Climatic Transition: Contrasts between the Margins of Plu- vial Lakes Lahontan and Mojave, Nevada and California, U.S.A. Catena, 36:255-281. Heinrich, H. 1988. Origin and Consequences of Cyclic Ice Rafting in the Northeast At- lantic Ocean during the Past 130,000 Years. Quaternary Research, 29:189-199. Hemphill, M.L. 1983. Fire, Vegetation, and People?Charcoal and Pollen Analysis of Sheep Mountain Bog, Montana: The Last 2800 Years. 70 pages. Master's thesis, Department of Anthropology, Washington State University. [70 pages.] Holmes, R.L., R.K. Adams, and H.C. Fritts 1986. Tree-Ring Chronologies of Western North America: California, Eastern Oregon and Northern Great Basin with Procedures Used in the Chronology Development Work Including Users Manuals for Computer Programs COFECHA and ARSTAN. Chronology Series VI: 182 pages. Tucson, Arizona. Houghton, J.G. 1969. Characteristics of Rainfall in the Great Basin. 205 pages. Reno: Desert Research Institute, University of Nevada. Houghton, J.G., CM. Sakamoto, and R.O. Gifford 1975. Nevada's Weather and Climate. Special Publication of the Nevada Bureau of Mines and Geology, 2: 78 pages. Reno: Mackay School of Mines, University of Nevada. Hubbs, C.L., and R.R. Miller 1948. II: The Zoological Evidence. In The Great Basin, with Emphasis on Glacial and Post-glacial Times. Bulletin of the University of Utah, Biological Series, 10(7): 18-166. Hunt, J., D. Rhode, and D.B. Madsen 2000. Homestead Cave Flora and non-Vertebrate Fauna. In D.B. Madsen, Late Quaternary Paleoecology in the Bonneville Basin. Utah Geo- logical Survey Bulletin, 130:47-58. Huntley, B. 1991. How Plants Respond to Climate Change: Migration Rates, Individu- alism and the Consequences for Plant Communities. Annals of Bot- any, 67:15-22. Iversen, J. 1973. The Development of Denmark's Nature since the Last Glacial; Translation from Danmarks Natur, vol. 1, pp. 345^445, 1967, by Michael Robson. Danmarks Geologiske Undersagelse, series 5, Raekke, number 7-c. Copenhagen: C.A. Reitzels Forlag. Jahren, A.H. 1996. The Stable Isotope Composition of the Hackberry (Celtis) and Its Use as a Paleoclimate Indicator. 136 pages. Doctoral dissertation, University of California, Berkeley. [Unpublished.] Jahren, A.H., R. Amundson, C. Kendall, and RE. Wigand 2001. Paleoclimatic Reconstruction Using the Correlation in Delta I8O of Hackberry Carbonate and Environmental Water, North America. Quaternary Research, 56(2):252-263. Jennings, S.A., and D.L. Elliot-Fisk 1993. Packrat Midden Evidence of Late Quaternary Vegetation Change in the White Mountains, California-Nevada. Quaternary Research, 39:214-221. Kaufman, D.S., S.L. Forman, and J. Bright 2001. Age of the Cutler Dam Alloformation (Late Pleistocene), Bonne- ville Basin, Utah. Quaternary Research, 56:322-334. Kelso, G. 1970. Hogup Cave, Utah: Comparative Pollen Analysis of Human Copro- lites and Cave Fill. In CM. Aikens, Hogup Cave. Anthropological Papers of the University of Utah, 93:251-262. Koehler, P.A., and R.S. Anderson 1994. Full-Glacial Shoreline Vegetation during the Maximum Highstand at Owens Lake, California. Great Basin Naturalist, 54(2): 142-149. 1995. Thirty Thousand Years of Vegetation Changes in the Alabama Hills, Owens Valley, California. Quaternary Research, 43:238-248. Krajicek, J.E., and R.D. Williams 1990. Celtis occidentalis L. Hackberry. In R.M. Burns and B.H. Honkala, coordinators, Silvics of North America. United States Department of Agriculture, Agriculture Handbook, 654:262?265. Ku, T.L., T.K. Lowenstein, J. Li, and R.J. Spencer 1998. U-Series Chronology of Lacustrine Deposits in Death Valley, Cali- fornia. Quaternary Research, 50(3):261?275. Kutzbach, J.E., and P.J. Guetter 1986. The Influence of Changing Orbital Parameters and Surface Bound- ary Conditions on Climate Simulations for the Past 18,000 Years. Journal of the Atmospheric Sciences, 43( 16): 1726-1759. Kutzbach, J.E., P.J. Guetter, P.J. Behling, and R. Selin 1993. Simulated Climatic Changes: Results of the COHMAP Climate- NUMBER 33 363 Model Experiments. In H.E. Wright, Jr., J.E. Kutzbach, T. Webb III, W.F. Ruddiman, F.A. Street-Perrott, and P.J. Bartlein, editors, Glo- bal Climates since the Last Glacial Maximum, pages 24-93. Minne- apolis: University of Minnesota Press. LaMarche, V.C., Jr. 1974. Paleoclimatic Inferences from Long Tree-Ring Records. Science, 183(4129):1043-1048. La Motte, R.S. 1936. The Upper Cedarville Flora of Northwestern California and Adja- cent California. Publications of the Carnegie Institution of Washing- ton, 455:57-142. Lanner, R.M. 1983. The Expansion of Singleleaf Pinon in the Great Basin. In D.H. Tho- mas, The Archaeology of Monitor Valley, 2: Gatecliff Shelter. An- thropological Papers of the American Museum of Natural History, 59(1):169-171. Laudermilk, J.D., and P.A. Munz 1934. Plants in the Dung of Nothrotherium from Gypsum Cave, Nevada. Publications of the Carnegie Institution ofWashington,453:29-37. 1938. Plants in the Dung of Nothrotherium from Rampart and Muav Caves, Arizona. Publications of the Carnegie Institution of Wash- ington, 487:271-281. Leonard, S., R. Miles, and H. Summerfield 1987. Soils of the Pinon-Juniper Woodlands. In R. Everett, editor, Pro- ceedings?Pinon-Juniper Conference, Reno, Nevada, January 13- 16, 1986. United States Forest Service, General Technical Report, INT-215:227-230. Lewis, H.T. 1977. Maskut: The Ecology of Indian Fires in Northern Alberta. Western Canadian Journal of Anthropology, 7(1): 15-52. 1982. Fire Technology and Resource Management in Aboriginal North America and Australia. In N.M. Williams and E.S. Hunn, editors, Resource Managers: North American and Australian Hunter-Gath- erers. American Association for the Advancement of Science, Se- lected Symposium, 67:45-67. Washington, D.C. 1985. Why Indians Burned: Specific Versus General Reasons. In J.E. Lo- tan, B.M. Kilgore, W.C. Gischer and R.W. Mutch, coordinators, Proceedings, Symposium and Workshop on Wilderness Fire, Mis- soula, Montana, November 15-18, 1983. United States Department of Agriculture, Forest Service, Intermountain Forest and Range Ex- periment Station, General Technical Report, INT-182:75-80. Ogden, Utah. Li, J., T.K. Lowenstein, C.B. Brown, T. Ku, and S. Luo 1996. A 100 ka Record of Water Tables and Paleoclimates from Salt Cores, Death Valley, California. Palaeogeography, Palaeoclima- tology, Paleoecology, 123:161-178. Light, A. 1996. Amino Acid Paleotemperature Reconstruction and Radiocarbon Shoreline Chronology of the Lake Bonneville Basin, USA. 142 pages. Master's thesis, Department of Geography, University of Colorado, Boulder, Colorado. Lindstrom, S. 1990. Submerged Tree Stumps as Indicators of Mid-Holocene Aridity in the Lake Tahoe Basin. Journal of California and Great Basin An- thropology, 12(2):146-157. Madsen, D.B. 1973. Late Quaternary Paleoecology in the Southeastern Great Basin. 117 pages. Doctoral dissertation, Department of Anthropology, Univer- sity of Missouri, Columbia. 1984. Pollen, Packrats, and Pine along the I.P.P. In S. Simms and M.C. Is- green, editors, Archaeological Excavations in the Sevier and Escal- ante Deserts, Western Utah: The Intermountain Power Project Transmission Line, Utah Section IPP Generating Plant to Deadhorse Junction. University of Utah, Department of Anthropology, Reports of Investigations, 83-12:353-363. Salt Lake City. 1985. Two Holocene Pollen Records for the Central Great Basin. In B.F. Jacobs, PL. Fall, and O.K. Davis, editors, Late Quaternary Vegeta- tion and Climates of the American Southwest. American Associa- tion of Stratigraphic Palynologists Contributions, 16:113-126. 2000. Late Quaternary Paleoecology in the Bonneville Basin. Utah Geo- logical Survey Bulletin, 130: 190 pages. Salt Lake City, Utah. Madsen, D.B., and D.R. Currey 1979. Late Quaternary Glacial and Vegetation Changes, Little Cottonwood Canyon Area, Wasatch Mountains, Utah. Quaternary Research, 12: 254-270. Madsen, D.B., and P.A. Kay 1982. Late Quaternary Pollen Analysis in the Bonneville Basin. American Quaternary Association Conference Abstracts, page 128. Seattle, Washington. Madsen, D.B., and D. Rhode 1990. Early Holocene Pinyon (Pinus monophylla) in the Northeastern Great Basin. Quaternary Research, 33:94?101. Madsen, D.B., D. Rhode, D.K. Grayson, J.M. Broughton, S.D. Livingston, J. Hunt, J. Quade, D.N. Schmitt, and M.W. Shaver III 2001. Late Quaternary Environmental Change in the Bonneville Basin, Western USA. Palaeoecology, Palaeogeography, Palaeoclima- tology, 167(3/4):243-271. Markgraf, V. 1986. Plant Inertia Reassessed. American Naturalist, 127':725?726. Martin, P.S. 1963. The Last 10,000 Years: A Fossil Pollen Record of the American Southwest. 87 pages. Tucson: University of Arizona Press. Martin, P.S., and P.J. Mehringer, Jr. 1965. Pleistocene Pollen Analysis and Biogeography of the Southwest. In H.E. Wright, Jr., and D.G. Frey, editors, The Quaternary of the United States, pages 433-451. Princeton, New Jersey: Princeton University Press. Martinson, D.G., N.G. Pisias, J.D. Hays, J. Imbrie, T.C. Moore, Jr., and N.J. Shackleton 1987. Age Dating and the Orbital Theory of the Ice Ages: Development of a High-Resolution 0 to 300,000-year Chronostratigraphy. Quater- nary Research, 27(1): 1-29. Marwitt, J.P. 1973. Median Village and Fremont Culture Regional Variation. Anthropo- logical Papers of the University of Utah, 95: 193 pages. McCoy, W.D. 1981. Quaternary Aminostratigraphy of the Bonneville and Lahontan Ba- sins, Western U.S., with Paleoclimatic Implications. 603 pages. Doctoral dissertation, University of Colorado, Boulder. Mead, J.I., R.S. Thompson, and A. Long 1978. Arizona Radiocarbon Dates IX: Carbon Isotope Dating of Packrat Middens. Radiocarbon, 20:171-191. Mehringer, P.J., Jr. 1967. Pollen Analysis of the Tule Springs Area, Nevada. In H.M. Worm- ington and D. Ellis, editors, Pleistocene Studies in Southern Nevada. Anthropological Papers of the Nevada State Museum, 13:129-200. 1977. Great Basin Late Quaternary Environments and Chronology. In D.D. Fowler, editor, Models and Great Basin Prehistory: A Sympo- sium. Desert Research Institute Publications in the Social Sciences, 12:113-167. 1985. Late-Quaternary Pollen Records from the Interior Pacific Northwest and Northern Great Basin of the United States. In V.M. Bryant, Jr., and R.G. Holloway, editors, Pollen Records of Late-Quaternary North American Sediments, pages 167-189. Dallas, Texas: Ameri- can Association of Stratigraphic Palynologists. 1986. Prehistoric Environments. In W.L. d'Azevedo, editor, Great Basin. Handbook of North American Indians, vol. 11:31-50. Washington, D.C: Smithsonian Institution. 1987. Late Holocene Environments on the Northern Periphery of the Great Basin, Final Report. 71 pages. Portland, Oregon: United 364 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES States Bureau of Land Management. Mehringer, RJ., and W.J. Cannon 1994. Volcaniclastic Dunes of the Fort Rock Valley, Oregon: Stratigraphy, Chronology, and Archaeology. In CM. Aikens and D.L. Jenkins, editors, Archaeological Research in the Northern Great Basin: Fort Rock Archaeology since Cressman. Anthropological Papers of the University of Oregon, 50:283-327. Mehringer, RJ., Jr., and C.W. Ferguson 1969. Pluvial Occurrence of Bristlecone Pine (Pinus aristatd) in a Mojave Desert Mountain Range. Journal of the Arizona Academy of Sci- ence, 5(4):284-292. Mehringer, P.J., Jr., and J.C. Sheppard 1978. Holocene History of Little Lake, Mojave Desert, California. In E.L. Davis, editor, The Ancient Californians: Rancholabrean Hunters of the Mojave Lakes Country. Science Series, 29:153-185. Mehringer, P.J., Jr., and C.N. Warren 1976. Marsh, Dune and Archaeological Chronologies, Ash Meadows, Amargosa Desert, Nevada. Research Report of the Nevada Archaeo- logical Survey, 6:120-150. Mehringer, P.J., and RE. Wigand 1986. Holocene History of Skull Creek Dunes, Catlow Valley, Oregon, U.S.A. Journal of Arid Environments, 11:117-138. 1987. Western Juniper in the Holocene. In R.L. Everett, editor, Proceed- ings of the Pinon-Juniper Conference. United States Forest Service, General Technical Report, INT-215:13-16. 1990. Comparison of Late Holocene Environments from Woodrat Mid- dens and Pollen. In J.L. Betancourt, T.R. Van Devender, and P.S. Martin, editors, Packrat Middens: The Last 40,000 Years ofBiotic Change, pages 294-325. Tucson: University of Arizona Press. Mellars, P. 1976. Fire Ecology, Animal Populations and Man: A Study of Some Eco- logical Relationships in Prehistory. Proceedings of the Prehistoric Society, 42:15-45. Mifflin, M.D., and M.M. Wheat 1979. Pluvial Lakes and Estimated Pluvial Climates of Nevada. Bulletin of the Nevada Bureau of Mines and Geology, 94: 57 pages. Miller, R.F., and RE. Wigand 1994. Holocene Changes in Semiarid Pinon-Juniper Woodlands: Re- sponse to Climate, Fire, and Human Activities in the US Great Ba- sin. BioScience, 44(7):465^74. Morrison, R.B. 1964. Lake Lahontan: Geology of Southern Carson Desert, Nevada. United States Geological Survey Professional Paper, 401: 156 pages. 1965. Quaternary Geology of the Great Basin. In H.E. Wright, Jr., and D.G. Frey, editors, The Quaternary of the United States, pages 265-285. Princeton, New Jersey: Princeton University Press. 1991. Quaternary Stratigraphic, Hydrologic, and Climatic History of the Great Basin, with Emphasis on Lakes Lahontan, Bonneville, and Tecopa. In R.B. Morrison, editor, Quaternary Nonglacial Geology: Conterminous U.S. The Geology of North America, K-2:283-320. Boulder, Colorado: Geological Society of America. Morrison, R.B., and J.C. Frye 1965. Correlation of the Middle and Late Quaternary Successions of the Lake Lahontan, Lake Bonneville, Rocky Mountain (Wasatch Range), Southern Great Plains, and Eastern Midwest Areas. Report of the Nevada Bureau of Mines, 9: 45 pages. Moutoux, T.E., and O.K. Davis 1995. Neogene through Pleistocene Paleoclimate of the Great Salt Lake Region, Northeastern Great Basin, USA. In CM. Isaacs and V.L. Tharp, editors, Proceedings of the Eleventh Annual Pacific Climate (PACLIM) Workshop, April 19-22, 1994. Interagency Ecological Program, Technical Report, 40:127-135. Sacramento: California Department of Water Resources. Mozingo, H.N. 1987. Shrubs of the Great Basin: A Natural History. 342 pages. Reno: University of Nevada Press. Negrini, R.M. 2002. Pluvial Lake Sizes in the Northwestern Great Basin throughout the Quaternary Period. In R. Hershler, D.B. Madsen, and D.R. Currey, editors, Great Basin Aquatic Systems History. Smithsonian Contri- butions to the Earth Sciences, 33:11-52. Negrini, R.M., and J.O. Davis 1992. Dating Late Pleistocene Pluvial Events and Tephras by Correlating Paleomagnetic Secular Variation Records from the Western Great Basin. Quaternary Research, 38:46-59. Negrini, R.M., D.B. Erbes, K. Faber, A.M. Herrera, A.P. Roberts, A.S. Cohen, RE. Wigand, and F.F. Foit, Jr. 2000. A Paleoclimate Record for the Past 250,000 Years from Summer Lake, Oregon, U.S.A., I: Chronology and Magnetic Proxies for Lake Level. Journal of Paleolimnology, 24(2):125?149. Nowak, C.L., R.S. Nowak, R.J. Tausch, and P.E. Wigand 1994a. A 30,000 Year Record of Vegetation Dynamics at a Semi-Arid Lo- cale in the Great Basin. Journal of Vegetation Science, 5:579-590. 1994b. Tree and Shrub Dynamics in Northwestern Great Basin Woodland and Shrub Steppe during the Late-Pleistocene and Holocene. Ameri- can Journal of Botany, 81(3):265-277. Oviatt, C.G. 1997. Lake Bonneville Fluctuations and Global Climate Change. Geology, 25:155-158. Oviatt, C.G., D.R. Currey, and D.M. Miller 1990. Age and Paleoclimatic Significance of the Stansbury Shoreline of Lake Bonneville, Northeastern Great Basin. Quaternary Research, 33:291-305. Oviatt, C.G., D.R. Currey, and D. Sack 1992. Radiocarbon Chronology of Lake Bonneville, Eastern Great Basin, USA. Palaeogeography, Palaeoclimatology, Palaeoecology, 99: 225-241. Oviatt, C.G., and W.D. McCoy 1986. New Radiocarbon and Amino Acid Constraints on Holocene Expan- sions of Great Salt Lake, Utah. American Quaternary Association, Programs with Abstracts, Ninth Biennial Meeting, page 103. Oviatt, C.G., W.D. McCoy, and R.G. Reider 1987. Evidence for a Shallow Early or Middle Wisconsin Lake in the Bonneville Basin, Utah. Quaternary Research, 27:248-262. Oviatt, C.G., D.M. Miller, C. Zachary, and J.P. McGeehin 2001. Refining the Age of the Lake Transgression to the Gilbert Shoreline in the Bonneville Basin. Eos, Transactions of the American Geo- physical Union, 82(47), fall meeting supplement, abstract PP22A- 0496. Phillips, F.M., A.R. Campbell, G.I. Smith, and J. Bischoff 1993. Interstadial Climate Cycles: A Link between Western North Amer- ica and Greenland? Geology, 22:1115-1118. Pisias, N.G., D.G. Martinson, T.C Moore, Jr., N.J. Shackleton, W. Prell, J. Hays, and G. Boden 1984. High Resolution Stratigraphic Correlation of Benthic Oxygen Isoto- pic Records Spanning the Last 300,000 Years. Marine Geology, 46: 217-233. Poinar, H.N., M. Hofreiter, W.G. Spaulding, P.S. Martin, B.A. Stankiewicz, H. Bland, R.P. Evershed, G. Possnert, and S. Paabo 1998. Molecular Coproscopy: Dung and Diet of the Extinct Ground Sloth Nothrotheriops shastensis. Science, 281(5375):402?406. Prentice, I.C, RJ. Bartlein, and T. Webb III 1991. Vegetation and Climate Change in Eastern North America since the Last Glacial Maximum. Ecology, 72:2038-2056. Quade, J., R.M. Forester, W.L. Pratt, and C. Carter 1998. Black Mats, Spring-Fed Streams, and Late-Glacial-Age Recharge in the Southern Great Basin. Quaternary Research, 49(2): 129-148. NUMBER 33 365 Quade, J., M.D. Mifflin, W.L. Pratt, W. McCoy, and L. Burckle 1995. Fossil Spring Deposits in the Southern Great Basin and Their Impli- cations for Changes in Water-Table Levels near Yucca Mountain, Nevada, during Quaternary Time. Geological Society of America Bulletin, 107(2):213-230. Rhode, D. 1998. 50,000 Years of Vegetation Change in the Northeastern Great Basin: An Analysis of Packrat Middens Located on Federal Lands Admin- istered by the Bureau of Land Management, Elko District. 57 pages. Report prepared for the United States Department of Interior, Bu- reau of Land Management, Elko, Nevada. [Unpublished.] 2000a. Middle and Late Wisconsin Vegetation History in the Bonneville Basin. In D.B. Madsen, Late Quaternary Paleoecology in the Bonneville Basin. Utah Geological Survey Bulletin, 130:137-148. 2000b. Holocene Vegetation History in the Bonneville Basin. In D.B. Mad- sen, Late Quaternary Paleoecology in the Bonneville Basin. Utah Geological Survey Bulletin, 130:149-164. Rhode, D., and D.B. Madsen 1995. Early Holocene Vegetation in the Bonneville Basin. Quaternary Re- search, 44(2):246-256. 1998. Pine Nut Use in the Early Holocene and Beyond: The Danger Cave Archaeobotanical Record. Journal of Archaeological Science, 25: 1199-1210. Rhode, D., and D.H. Thomas 1983. Flotation Analysis of Selected Hearths. In D.H. Thomas, The Ar- chaeology of Monitor Valley, 2: Gatecliff Shelter. American Mu- seum of Natural History Anthropological Papers, 59( 1): 151-156. Richmond, G.M. 1964. Glaciation of Little Cottonwood and Bells Canyon, Wasatch Moun- tains, Utah. United States Geological Survey Professional Paper, 454-D: 41 pages. 1986. Stratigraphy and Correlation of Glacial Deposits of the Rocky Mountains, Colorado Plateau, and the Ranges of the Great Basin. Quaternary Science Reviews, 5:99-127. Russell, I.C. 1885. Geological History of Lake Lahontan, A Quaternary Lake of North- western Nevada. United States Department of the Interior, Geologi- cal Survey, Monograph, 11. Washington, D.C.: Government Printing Office. Schroedl, G.F. 1973. The Archaeological Occurrence of Bison in the Southern Plateau. Reports of Investigations, 51:77 pages. Pullman, Washington: Lab- oratory of Anthropology, Washington State University. Schulman, E. 1956. Dendroclimatic Changes in Semiarid America. 142 pages. Tucson: University of Arizona Press. Scott, WE. 1988. Temporal Relations of Lacustrine and Glacial Events at Little Cot- tonwood and Bells Canyons, Utah. In M.N. Machette, editor, In the Footsteps of G.K. Gilbert?Lake Bonneville and Neotectonics of the Eastern Basin and Range Province. Miscellaneous Publication of the Utah Geological and Mineralogical Survey, 88-1:78-81. Scott, W.E., W.D. McCoy, R.R. Shroba, and M. Rubin 1983. Reinterpretation of the Exposed Record of the Last Two Cycles of Lake Bonneville, Western United States. Quaternary Research, 20: 261-285. Sears, P.B. 1937. Pollen Analysis as an Aid in Cultural Deposits in the United States. In G.G. MacCurdy, editor, Early Man, pages 61-66 . London: Lip- pincott & Company. Shackleton, N.J., and N.D. Opdyke 1973. Oxygen Isotope and Paleomagnetic Stratigraphy of Equatorial Pa- cific Core V28-238: Oxygen Isotope Temperature and Ice Volumes on a 10,000 and 100,000 Year Time Scale. Quaternary Research, 32:39-55. Shreve, F., and A.L. Hinckley 1937. Thirty Years of Change in Desert Vegetation. Ecology, 18:463-478. Siegal, R.D. 1983. Paleoclimatic Significance of D/H and 13C/I2C Ratios in Pleis- tocene and Holocene Wood. 105 pages. Master's thesis, University of Arizona, Tucson. Smiley, T.L., and P.J. Mehringer, Jr. n.d. Post-Pluvial History of the Spring-Fed Salt Marshes of the Amar- gosa Drainage. Report to National Science Foundation for Research Supported by NSF Grant GB-8646, 35 pages. Smith, G.I., and J.L. Bischoff, editors 1997. An 800,000-Year Geologic and Climatic Record from Owens Lake, California: Core OL-92. Special Paper of the Geological Society of America, 317: 165 pages. Spaulding, W.G. 1977. Late Quaternary Vegetation Change in the Sheep Range, Southern Nevada. Journal of the Arizona Academy of Sciences, 22(2): 183-188. 1980. Presettlement Vegetation of the California Desert. 97 pages. Unpub- lished report submitted to the Desert Planning Staff, California Bu- reau of Land Management, Riverside. 1981. The Late Quaternary Vegetation of a Southern Nevada Mountain Range. 271 pages. Doctoral dissertation, University of Arizona, Tucson. [Unpublished.] 1985. Vegetation and Climates of the Last 45,000 Years in the Vicinity of the Nevada Test Site, South-Central Nevada. United States Geologi- cal Survey Professional Paper, 1329: 83 pages. 1990. Vegetational and Climatic Development of the Mojave Desert: The Last Glacial Maximum to Present. In J.L. Betancourt, T.R. Van Dev- ender, and P.S. Martin, editors, Packrat Middens?The Last 40,000 Years ofBiotic Change, pages 166-199. Tucson: University of Ari- zona Press. 1991. A Middle Holocene Vegetation Record from the Mojave Desert of North America and Its Paleoclimatic Significance. Quaternary Re- search, 35:427^137. 1994. Paleohydrologic Investigations in the Vicinity of Yucca Mountain: Late Quaternary Paleobotanical and Paleontological Records. 80 pages. Las Vegas, Nevada: Mifflin & Associates, Inc. Spaulding, W.G., and L. Graumlich 1986. The Last Pluvial Climatic Episodes in the Deserts of Southwestern North America. Nature, 320:441-444. Spencer, R.J., M. Baedecker, H.P. Eugster, R.M. Forester, M.B. Goldhaber, B.F. Jones, K. Kelts, J. McKenzie, D.B. Madsen, S.L. Rettig, M. Rubin, and C.J. Bowser 1984. Great Salt Lake and Precursors, Utah: The Last 30,000 Years. Con- tributions to Mineralogy and Petrology, 86:321-334. Steward, J.H. 1930. Irrigation Without Agriculture. Papers of the Michigan Academy of Science, Arts, and Letters, 12:149-156. 1938. Basin-Plateau Aboriginal Sociopolitical Groups. Bulletin of the Bu- reau of American Ethnology, 120: xii + 346 pages. Stewart, O.C. 1954. Forest Fires with a Purpose. Southwestern Lore, 20(3):42-47. 1955a. Why Were the Prairies Treeless? Southwestern Lore, 20(4):59-64. 1955b. Forest and Grass Burning in the Mountain West. Southwestern Lore, 1956. Fire as the First Great Force Employed by Man. In W.L. Thomas, editor, Man s Role in Changing the Face of the Earth, pages 115- 133. Chicago: The University of Chicago Press. 1963. Barriers to Understanding the Influence of Use of Fire by Aborigi- nes on Vegetation. Proceedings, Second Annual Tall Timbers Fire Ecology Conference, March 14-15, Tallahassee, Florida, pages 117-126. 1982. Origin of Control of Fire. Southwestern Lore, 48(4):14-21. 366 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Stine, S. 1990. Late Holocene Fluctuations of Mono Lake, Eastern California. Palaeogeography, Palaeoclimatology, Palaeoecology, 78:333-381. Stutz, H.C. 1978. Explosive Evolution of Perennial Atriplex in Western America. Memoirs of the Great Basin Naturalist, 2:161-168. Swetnam, T.W. 1990. Fire History and Climate in the Southwestern United States. In J.S. Krammes, editor, Effects of Fire Management of Southwestern Nat- ural Resources: Proceedings of the Symposium. United States For- est Service, General Technical Report, RM-191:6-17. Swetnam, T.W., and C.H. Baisan 1996. Historical Fire Regime Patterns in the Southwestern United States since AD 1700. In C. Allen, editor, Fire Effects in Southwestern Forests: Proceedings of the Second La Mesa Fire Symposium. United States Forest Service, General Technical Report, RM-286: 11-32. Swetnam, T.W., and J.L. Betancourt 1990. Fire-Southern Oscillation Relations in the Southwestern United States. Science, 249:1017-1020. Szabo, B J., P.T. Kolesar, A.C. Riggs, I.J. Winograd, and K.R. Ludwig 1994. Paleoclimatic Inferences from a 120,000-Yr Calcite Record of Wa- ter-Table Fluctuations in Browns Room of Devils Hole, Nevada. Quaternary Research, 41:59?69. Tausch, R.J., N.E. West, and A.A. Nabi 1981. Tree Age and Dominance Patterns in Great Basin Pinon-Juniper Woodlands. Journal of Range Management, 34:259-264. Tausch, R.J., RE. Wigand, and J.W. Burkhardt 1993. Viewpoint: Plant Community Thresholds, Multiple Steady States, and Multiple Successional Pathways: Legacy of the Quaternary? Journal of Range Management, 46:439-447. Thompson, R.S. 1984. Late Pleistocene and Holocene Environments in the Great Basin. 256 pages. Doctoral dissertation, Department of Geosciences, Uni- versity of Arizona, Tucson. 1988. Western North America. In B. Huntley and T. Webb III, editors, Vegetation History. Handbook of Vegetation Science, 7:415-459. Dordrecht, The Netherlands: Kluwer Academic Publishers. 1990. Late Quaternary Vegetation and Climate in the Great Basin. In J.L. Betancourt, T.R. Van Devender, and P.S. Martin, editors, Packrat Middens: The Last 40,000 Years ofBiotic Change, pages 200-239. Tucson: University of Arizona Press. 1991. Pliocene Environments and Climates in the Western United States. Quaternary Science Reviews, 10:115-132. 1992. Late Quaternary Environments in Ruby Valley, Nevada. Quaternary Research, 37:1-15. Thompson, R.S., K.H. Anderson, and P.J. Bartlein 2000. Atlas of Relations between Climatic Parameters and the Distribu- tions of Important Trees and Shrubs in North America. United States Geological Survey Professional Paper, 1650A-B: 423 pages. Denver, Colorado. Thompson, R.S., L.V. Benson, and E.M. Hattori 1986. A Revised Chronology for the Last Pleistocene Lake Cycle in the Central Lahontan Basin. Quaternary Research, 25:1-10. Thompson, R.S., and E. Hattori 1983. Paleobotany of Gatecliff Shelter: Packrat {Neotoma) Middens and from Gatecliff Shelter and Holocene Migrations of Woodland Plants. In D.H. Thomas, The Archaeology of Monitor Valley, 2: Gatecliff Shelter. Anthropological Papers of the American Museum of Natural History, 59(1): 157-1 67. Thompson, R.S., and R.R. Kautz 1983. Paleobotany of Gatecliff Shelter: Pollen Analysis. In D.H. Thomas, The Archaeology of Monitor Valley, 2: Gatecliff Shelter. Anthropo- logical Papers of the American Museum of Natural History, 59(1): 136-151. Thompson, R.S., and J.I. Mead 1982. Late Quaternary Environments and Biogeography in the Great Ba- sin. Quaternary Research, 17:39-55. Thompson, R.S., C.G. Oviatt, A.P. Roberts, J. Buchner, R. Kelsey, C. Bracht, R.M. Forester, and J.P. Bradbury 1994. Stratigraphy, Sedimentology, Paleontology, and Paleomagnetism of Pliocene-Early Pleistocene Lacustrine Deposits from Western Utah. United States Geological Survey Open-file Report, 95-1: 91 pages. Thompson, R.S., L.J. Toolin, R.M. Forester, and R.J. Spencer 1990. Accelerator-Mass Spectrometer (AMS) Radiocarbon Dating of Pleistocene Lake Sediments in the Great Basin. Palaeogeography, Palaeoclimatology, Palaeoecology, 78:301-313. Thompson, R.S., C. Whitlock, P.J. Bartlein, S.P. Harrison, and W.G. Spaulding 1993. Climate Changes in the Western United States since 18,000 yr B.P. In H.E. Wright, Jr., J.E. Kutzbach, T. Webb III, W.F. Ruddiman, F.A. Street-Perrott, and P.J. Bartlein, editors, Global Climates since the Last Glacial Maximum, pages 468-513. Minneapolis: University of Minnesota Press. Tidwell, W.D., S.R. Rushforth, and D. Simper 1972. Evolution of Floras in the Intermountain Region. In A. Cronquist, A.H. Holmgren, N.H. Holmgren, and J.L. Reveal, Intermountain Flora: Vascular Plants of the Intermountain West, U.S.A., volume 1, pages 19-39. New York: Hafner Publishing Co. Van Devender, T.R. 1990a. Late Quaternary Vegetation and Climate of the Sonoran Desert, United States and Mexico. In J.L. Betancourt, T.R. Van Devender and P.S. Martin, editors, Packrat Middens: The Last 40,000 Years of Biotic Change, pages 134-163. Tucson: University of Arizona Press. 1990b. Late Quaternary Vegetation and Climate of the Chihuahuan Desert, United States and Mexico. In J.L. Betancourt, T.R. Van Devender, and P.S. Martin, editors, Packrat Middens: The Last 40,000 Years of Biotic Change, pages 104-133. Tucson: University of Arizona Press. Van Devender, T.R., and W.G. Spaulding 1979. Development of Vegetation and Climate in the Southwestern United States. Science, 204:701-710. Van Devender, T.R., R.S. Thompson, and J.L. Betancourt 1987. Vegetation History of the Deserts of Southwestern North America: The Nature and Timing of the Late Wisconsin-Holocene Transition. In W.F. Ruddiman and H.E. Wright, Jr., editors, North America and Adjacent Oceans during the Last Deglaciation. The Geology of North America, K-3:323-352. Boulder, Colorado: Geological Soci- ety of America. Van de Water, P.K., S.W. Leavitt, and J.L. Betancourt 1994. Trends in Stomatal Density and 13C/12C Ratios of Pinus flexilis Needles during Last Glacial-Interglacial Cycle. Science, 264:239- 243. Walter, H. 1954. Le facteur eau dans les regions arides et sa signification pour l'or- ganisation de la vegetation dans les contrees sub-tropicales. In Les Divisions Ecologiques du Monde, pages 27-39. Paris: Centre Natio- nale de la Recherche Scientifique. Weaver, J.E., and F.E. Clements 1938. Plant Ecology. Second edition, 601 pages. New York: McGraw-Hill Book Co. Wells, P.V. 1983. Paleobiogeography of Montane Islands in the Great Basin since the Last Glaciopluvial. Ecological Monographs, 53:341-382. Wells, P.V, and R. Berger 1967. Late Pleistocene History of Coniferous Woodland in the Mohave Desert. Science, 155(3770): 1640-1647. Wells, P.V, and CD. Jorgensen 1964. Pleistocene Wood Rat Middens and Climatic Change in the Mohave Desert: A Record of Juniper Woodland. Science, 143(3611): 1171- NUMBER 33 367 1174. Wells, P.V., and D. Woodcock 1985. Full-Glacial Vegetation of Death Valley, California: Juniper Wood- land Opening to Yucca Semidesert. Madrono, 32:11-32. West, N.E., R.J. Tausch, K.H. Rea, and P.T. Tueller 1978. Phytogeographical Variation within Juniper-Pinon Woodlands of the Great Basin. In K.T. Harper and J.L. Reveal, editors, Intermoun- tain Biogeography: A Symposium. Great Basin Naturalist Memoirs, 2:119-136. Whitlock, C, and P.J. Bartlein 1993. Spatial Variations of Holocene Climatic Change in the Yellowstone Region. Quaternary Research, 39:231-238. 1997. Vegetation and Climate Change in Northwest America during the Past 125 kyr. Nature, 388(6637):57-61. Wigand, P.E. 1987. Diamond Pond, Harney County, Oregon: Vegetation History and Water Table in the Eastern Oregon Desert. Great Basin Naturalist, 47:427^158. 1989. Vegetation, Vegetation History, and Stratigraphy. In K.C. Reid, J.A. Draper, and P.E. Wigand, editors, Prehistory and Paleoenvironments of the Silvies Plateau, Harney Basin, Southeastern Oregon. Project Report, 8:37-85. Pullman, Washington: Center for Northwest An- thropology, Washington State University. 1995. Stable Isotopic Analyses in Paleoclimatic Reconstruction. In W.J. Waugh, editor, Proceedings of the Workshop?Climate Change in the Four Corners and Adjacent Regions: Implications for Environ- mental Restoration and Land-Use Planning, CONF-9409325, pages 27-38. Grand Junction, Colorado: United States Department of En- ergy. 1997a. Native American Diet and Environmental Contexts of the Holocene Revealed in the Pollen of Human Fecal Material. Nevada Historical Society Quarterly, 40(l):105-116. 1997b. A Late-Holocene Pollen Record from Lower Pahranagat Lake, Southern Nevada, USA: High Resolution Paleoclimatic Records and Analysis of Environmental Responses to Climate Change. In CM. Isaacs and V.L. Tharp, editors, Proceedings of the Thirteenth Annual Pacific Climate (PACLIM) Workshop, April 15-18, 1996. Inter- agency Ecological Program, Technical Report 53:63?77. Sacra- mento: California Department of Water Resources. Wigand, P.E., M.L. Hemphill, and S. (Manna) Patra 1994. Late Holocene Climate Derived from Vegetation History and Plant Cellulose Stable Isotope Records from the Great Basin of Western North America. High-Level Radioactive Waste Management Con- ference: Proceedings of the Fifth Annual International, Sponsored by the American Society of Civil Engineers and the American Nu- clear Society; Co-sponsored by the American Chemical Society ... [et al.J in Cooperation with U.S. Department of Energy ... [et al.J, volume 4:2574-2583. La Grange Park, Illinois: American Nuclear Society; New York: American Society of Civil Engineers. Wigand, P.E., M.L. Hemphill, S.E. Sharpe, and S. Patra 1995. Great Basin Woodland Dynamics during the Holocene. In W.J. Waugh, editor, Proceedings of the Workshop-Climate Change in the Four Corners and Adjacent Regions: Implications for Environmen- tal Restoration and Land-Use Planning, CONF-9409325, pages 51? 69. Grand Junction, Colorado: United States Department of Energy. Wigand, P.E., and P.J. Mehringer, Jr. 1985. Pollen and Seed Analyses. In D.H. Thomas, The Archaeology of Hidden Cave, Nevada. Anthropological Papers of the American Mu- seum of Natural History, 61(1): 108-124. Wigand, P.E., and C.L. Nowak 1992. Dynamics of Northwest Nevada Plant Communities during the Last 30,000 Years. In C.A. Hall, V. Doyle-Jones, and B. Widawski, edi- tors, The History of Water: Eastern Sierra Nevada, Owens Valley, White-Inyo Mountains. White Mountain Research Station Sympo- sium, 4:40-62. Wigand, P.E., and M.R. Rose 1990. Calibration of High Frequency Pollen Sequences and Tree-Ring Records. High-Level Radioactive Waste Management: Proceedings of the International Conference Topical Meeting, Sponsored by the American Society of Civil Engineers and the American Nuclear So- ciety; Co-sponsored by the American Chemical Society... [etal.], in Cooperation with the U.S. Department of Energy ... [et al.], volume 2:1240-1250. La Grange Park, Illinois: American Nuclear Society; New York: American Society of Civil Engineers. Winograd, I.J., T.B. Coplen, J.M. Landwehr, A.C. Riggs, K.R. Ludwig, B.J. Szabo, P.T. Kolesar, and K.M. Revesz 1992. Continuous 500,000-Year Climate Record from Vein Calcite in Devils Hole, Nevada. Science, 258:255-260. Winograd, I.J., J.M. Landwehr, K.R. Ludwig, T.B. Coplen, and A.C. Riggs 1997. Duration and Structure of the Past Four Interglacials. Quaternary Research, 48:141-154. Winograd, I.J., B.J. Szabo, T.B. Coplen, and A.C. Riggs 1988. A 250,000-Year Climatic Record from Great Basin Vein Calcite: Implications for Milankovitch Theory. Science, 242(4883): 1275- 1280. Winter, J.C. 1976. The Processes of Farming Diffusion in the Southwest and Great Ba- sin. American Antiquity, 41(4):421-429. Winter, J.C, and P.F. Hogan 1986. Plant Husbandry in the Great Basin and Adjacent Northern Colo- rado Plateau. In C.J. Condie and D.D. Fowler, editors, Anthropology of the Desert West: Essays in Honor of Jesse D. Jennings, pages 117-144. Salt Lake City: University of Utah Press. Woolfenden, W.B. 1993. Pollen Present in Cores OL-92-1, -2, and -3. In Core OL-92 from Owens Lake, Southeast California. G.I. Smith and J.L. Bischoff, ed- itors, United States Geological Survey Open File Report, 93-683: 314-332. 1997. Late-Quaternary Vegetation History of the Southern Owens Valley Region, Inyo County, California. 347 pages, 22 figures, 5 tables. Doctoral dissertation, Department of Geosciences, University of Ar- izona, Tucson. [Unpublished.] Great Basin Mammals and Late Quaternary Climate History DonaldK. Gray son ABSTRACT Our exacting knowledge of the history of small mammals in the Great Basin has had a substantial effect on our understanding of mammalian biogeography in this region but has yet to be ade- quately incorporated into more general syntheses of Great Basin climate history, particularly those produced by climate modelers. I review herein the rich late-Pleistocene and Holocene small-mam- mal record recently obtained from Homestead Cave in the north- ern Bonneville Basin, Utah. Together with a diverse variety of other small-mammal sequences, this record indicates that the early Holocene (ca. 10,000-7500 years B.P. (before present)) climate of the northern Great Basin was cool and moist, not warm and moist as has been argued on the basis of computer modeling. This sequence also demonstrates the substantial effects of the generally hot and dry climate of the middle Holocene (ca. 7500-4500 years B.P.) on small mammals in this area as well as the responses of those mammals to the end of this period. Perhaps most surpris- ingly, Homestead Cave data document the local extinction and recent recolonization of a species that was at one time felt to be isolated on Great Basin mountains. This discovery joins newly available neoecological data in showing that the history of "mon- tane" mammals in the Great Basin is far more complex than once had been thought, and it simplies that low-elevation dispersal cor- ridors must be provided in conjunction with the creation of moun- tain-based wildlife reserves in the Great Basin. Introduction The volume whose legacy we are now celebrating (Black- welder et al., 1948) was remarkable in many ways. Barely visi- ble to Quaternary scientists who are not archaeologists is the Donald K. Grayson, Department of Anthropology, University of Wash- ington, Seattle, Washington 98915, USA. Reviewers: J. Broughton, Department of Anthropology, University of Utah, Salt Lake City, Utah 84112, USA; R. Hoffman, Division of Mam- mals, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560-0108, USA; T.E. Lawlor, Department of Bio- logical Sciences, Humboldt State University, Arcata, California 95521-8299, USA; David B. Madsen, Environmental Sciences, Utah Geological Survey, Salt Lake City, Utah 84114, USA. fact that Ernst Antevs' contribution to this volume was of fun- damental importance in structuring both the chronological as- signments of archaeological sites and the interpretation of the record that was dated accordingly. Antevs' work, together with that by archaeologist Luther Cressman, played a major role in ushering in the tradition of conjoint paleoenvironmental and ar- chaeological research in the Great Basin, a tradition that contin- ues to thrive today. Indeed, Antevs (1948) is still cited by Great Basin prehistorians (e.g., Aikens, 1993), and for good reason: on a very general level, he appears to have gotten it right in terms of both his general climatic sequence and the effect of cli- matic change on human population densities in the Great Basin. Antevs' Neothermal classification was based essentially on temperature; he argued that it was temperature and not mois- ture that supported the long-distance correlations that allowed him to date his sequence on the basis of Scandinavian varve chronologies. Because of this, it is interesting to observe that although Antevs routinely characterized the early Holocene, or Anathermal, of the arid west as moister than anything that came later, his view of early Holocene temperatures in this re- gion changed through the years. In the 1940s and early 1950s, Antevs routinely described the Anathermal as having temperatures that at first were like those of today but that grew warmer, eventually culminating in the famed Altithermal (e.g., Antevs, 1948, 1952). Influenced by the work of Magnus Fries, however, Antevs soon began to characterize the Anathermal as not only the wettest but also the coolest subdivision of his Neothermal sequence (e.g., Antevs, 1953, 1955a, 1955b; see Figure 1). No longer can there be any question that Antevs was correct in his characterization of the early Holocene in the Great Basin. Because it was clearly moist, one of his descriptions must be right. It remains to be determined, however, whether Antevs I, with a warm and moist early Holocene, or Antevs II, with a cool and moist early Holocene, was closer to the truth. A wide range of empirical data converges to suggest that much of the Great Basin was cool during the early Holocene. Certain cli- matic models, however, suggest that this period of time was quite warm. 369 370 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 0- -4' -8- 600- 0- DEGLACIAL NEOTHERMAL ANA- ALTI- MEDI- BONNEVILLE BASIN WATER LEVELS \ PLUVIAL \ \ POSTPLUVIAL I I I I I i I I I I I I i i i i i I i i r 20,000 15.000 10,000 5,000 0 FIGURE 1.?Antevs' later reconstruction of Late Quaternary temperatures in the arid west and water levels in the Bonneville Basin (after Antevs, 1955a) (B.P. = before present, i.e., before 1950; temperature scale indicates devi- ation from modern conditions). These conflicting views lead to different interpretations of biotic and moisture history in the Great Basin. I examine herein small-mammal sequences from the latest Pleistocene and Ho- locene in the Great Basin that are relevant to this issue. Mam- mal data have been cited by some climatic modelers when those data match the predictions of the models but have been consistently ignored when the data are at odds with model-de- rived predictions. The greatest mismatch occurs during the early Holocene, the period of primary focus for this paper. Vertebrates other than small mammals are also relevant to these issues. Fishes are discussed elsewhere in this volume; our paleontological knowledge of reptile and amphibian history from the latest Pleistocene and Holocene has been reviewed by Mead and Bell (1994); and the timing and nature of the extinc- tions of birds and large mammals within the Great Basin to- ward the end of the Pleistocene also has been explored (Gray- son, 1993, 1994). Given these reviews, I restrict my attention herein to small-mammal history, focusing on the most detailed history available for latest Pleistocene and early Holocene mammals of the Great Basin. This history is provided by Homestead Cave in the northern Bonneville Basin of north- central Utah. I discuss the paleoclimatic implications of the Homestead Cave mammal sequence and examine the implica- tions of this fauna for our understanding of modern small- mammal distributions within the Great Basin. Homestead Cave Homestead Cave is a few kilometers to the southwest of Great Salt Lake on a northwestern spur of the Lakeside Moun- tains (Figures 2, 3). This spur, Homestead Knoll, is a low (max- imum elevation, 1615 m), rocky promontory that is devoid of active springs and permanent streams. The barren playa of Pleistocene Lake Bonneville is immediately west and north- west of this rocky spur, but the knoll itself supports vegetation dominated by shrubs and grasses in addition to a few scattered junipers (Juniperus osteosperma). Shadscale (Atriplex conferti- folid) and horsebrushes (Tetradymia spp.) are the most abun- dant shrubs, but bud sagebrush (Artemisia spinescens), rabbit- brush (Chrysothamnus sp.), and greasewood (Sarcobatus vermiculatus) are also present, as is big sagebrush {Artemisia tridentata) along seasonally moist drainages. Greasewood be- comes increasingly common as the valley bottom is ap- proached, whereas the invasive Eurasian cheatgrass (Bromus sp.) is abundant on the flats beneath the knoll. Homestead Cave sits on the northwestern edge of Homestead Knoll at an elevation of 1406 m. Initial test excavations con- NUMBER 33 371 KM FIGURE 2.?Location of Homestead Cave, northwestern Lakeside Mountains, in the Bonneville Basin, north- central Utah, USA. ducted by D.B. Madsen of the Utah Geological Survey re- vealed that the sediments of Homestead Cave are remarkably rich in vertebrate remains. This accumulation was primarily caused by owls using the cave; the deposits contained numer- ous owl pellets, and many isolated mammal bones and teeth, regardless of their age, still had remains of those pellets adher- ing to them. In 1993 and 1994, a 1 x 1 m column sample was stratigraph- ically excavated from this site, faunal material in that column was collected by sieving the deposits through a nested series of screens with mesh openings that measured 1/4 inch (6.4 mm), 1/8 inch (3.2 mm), and 1/16 inch (1.6 mm). Remains of fishes from the resultant sample were analyzed by J.M. Broughton (University of Utah), those of birds by S.D. Livingston (Desert Research Institute), and the mammals by me. Detailed analyses of all of this material, including the mammals, are presented by Madsen (2000). The details of the depositional history of the sediments in the cave are not relevant to the arguments developed herein, but I note that we now have available 21 radiocarbon (14C) dates for the 18 Homestead Cave strata from the column itself (Table 1). These dates document that the earliest stratum within the cave was deposited between about 11,300 and 10,200 14C yr B.P. (years before present?i.e., before 1950), thus encompassing the period of time that includes the Gilbert phase of Lake Bonneville history. The available radiocarbon chronology indi- cates that superincumbent stratum II at Homestead Cave accu- mulated between 8900 and 8500 14C yr B.P. and, thus, also sug- gests that there may have been a depositional gap of about 1000 14C years between the accumulation of Stratum I and Stratum II. No other similar gaps are suggested by the radiocar- bon chronology, but because some parts of the sequence remain weakly dated, such gaps may exist although they are undetec- ted. The latest radiocarbon date for the sequence is for Stratum XVII, and it falls at about 1000 14C yr B.P. The Homestead Cave faunal sequence thus covers the latest Pleistocene and nearly all of the Holocene. 372 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 3.?Location of Homestead Cave on Homestead Knoll (white arrow). Prominent terraces were produced by the waters of Pleistocene Lake Bonneville. The faunal sample that we excavated from Homestead Cave is of almost overwhelming size. Some 184,000 mammal bones and teeth from the 1/4-inch and 1/8-inch sample fractions have been identified to at least the genus level (Table 2). These spec- imens represent 40 mammalian taxa from 15 of the 18 strata; the faunal assemblages from Strata XIII, XIV, and XV have not been identified, whereas only the kangaroo rats {Dipodomys spp.) from Stratum X have been analyzed. The Homestead Cave sample is so large, so well stratified, and so well dated that it provides an unparalleled view of Late Quaternary small- mammal history in the Great Basin. Although my focus herein will be the paleoclimatic implications of this fauna, I will also discuss briefly a series of relevant biogeographic issues that are related to those paleoclimatic matters. Selected Faunal Histories The paleoclimatic implications of the Homestead Cave mammalian fauna are best understood from analysis of the his- tories of a group of environmentally sensitive taxa. Discussion of their broader implications follows the individual histories. Authorities for the mammal taxa discussed herein are provided by Wilson and Reeder, 1993. KANGAROO RATS The aptly named chisel-toothed kangaroo rat {Dipodomys microps) is characterized by flattened, or chisel-shaped, ante- rior faces and tips of the lower incisors. They are distinctly dif- ferent from the rounded and pointed, or awl-shaped, lower inci- sors of Ord's kangaroo rat {D. ordii). Kenagy (1972, 1973; see also Garrison and Best, 1990) demonstrated that D. microps uses these incisors to shave off the hypersaline external sur- faces of Atriplex leaves, then consumes the palatable inner por- tions of those leaves. In contrast, D. ordii is primarily a grani- vore and is often found associated with habitats in which sagebrush is common (Hayssen, 1991). In an analysis of the spatial relationships within the rodent community of Warm Springs Valley, west-central Nevada, O'Farrell (1980) found little spatial overlap between these two species, with D. mi- crops restricted to habitats dominated by shadscale. This is not to say that D. microps cannot exist in habitats that lack Atriplex NUMBER 33 373 TABLE 1.?Radiocarbon (14C) dates (years before present, B.P.) from Home- stead Cave strata and their calibrated ages, in calendar years, as calculated with CALIB 3.0.3c (Stuiver and Reimer, 1993). TABLE 2.?Number of identified specimens (NISP) of mammals by stratum at Homestead Cave. Qfri*otiirvi kjll CHL4111 XVII XVI XIV XIII XII X VII VI V IV II I (upper 5 cm) I (general) I (lower 5 cm) 14C date (years B.P.) 1,020 ?40 1,200 ?50 2,850 ?50 3,480 ?40 3,400 ? 60 5,330 ?65 6,160 ?85 6,185 ?105 7,120 ?70 8,230 ? 70 8,195 ?85 8,520 ?80 8,790 ? 90 8,830 ? 240 10,160 ?85 10,350 ?80 10,910 ?60 11,065 ?105 11,180 + 85 11,265 ?85 11,270 ?135 Laboratory number Beta 101877 Beta 66940 Beta 103692 Beta 101878 Beta 63179 AA 14822 AA 14824 AA 14825 AA 14826 AA 16810 AA 14823 AA 14821 AA 14820 Beta 63438 AA 14819 AA 14818 Beta 72205 AA 14817 AA 16808 AA 16809 AA 14816 Calendar age (years) 933 1,078 2,948 3,707 3,632 6,105-6,171 7,017 7,028-7,148 7,912 9,210 9,047-9,193 9,484 9,694-9,852 9,873 11,865 12,228 12,832 12,978 13,087 13,172 13,177 or that D. ordii cannot exist in habitats in which Atriplex is common. When both taxa are present, however, D. microps will tend strongly to be found in Atriplex habitats, and D. ordii will be found in habitats marked by other plant associations, in- cluding those dominated by Artemisia. Kangaroo rats are by far the most abundant mammal in the Homestead Cave fauna: 59.1% of the identified specimens (ex- Stratum XVIII XVII XVI XII XI X IX VIII VII VI V IV III II I TOTAL Mammals (NISP) 1,047 15,548 6,296 22,860 10,096 6,60 la 22,088 8,289 13,905 24,330 5,109 26,615 2,884 7,855 10,275 183,798 a Only the kangaroo rats from this stratum have been identified. eluding Stratum X, for which only kangaroo rats were identi- fied) pertain to this genus; of these, some 8300 specimens have been identified to the species level. Figure 4 provides the ratio of D. microps to both D. microps and D. ordii through time at Homestead Cave (including Stratum X) and shows that the late-Pleistocene and early Holocene faunas of this area were marked by abundances of D. ordii that have not been seen since. Figure 5 presents the same data in terms of the ratio of D. microps to D. ordii for each stratum and shows that the middle Holocene, particularly between about 7000 and 5000 14C yr B.P., had elevated proportions of D. microps compared with 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 CO I*" CO CM m CM o in i? co T- CM T- h- o co,_ in 2 8 FIGURE 4.?Proportion of the kangaroo rat fauna contributed by D. microps at Homestead Cave, by stratum. The total number of Dipodomys specimens identified as either D. microps or D. ordii is above the bar for each stratum. Corresponding ages indi- cated below the strata are in radiocarbon years. o oo o CO CM * x = > = E X X ^ > STRATUM 374 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 125 100 H 75 H 50 H 25 H FIGURE 5.?Ratio of D. microps to ?>. on/? at Home- stead Cave, by stratum. Above the bar for each stra- tum is the total number of Dipodomys specimens identified as either D. microps or D. ordii. Corre- sponding radiocarbon ages are indicated below the strata. what came before and after. This history strongly suggests the progressive replacement of sagebrush by shadscale vegetation during the early Holocene in the Homestead Knoll area, as well as a middle-Holocene peak in xericity here. It is also possible to take a broader view of the history of Di- podomys in the Homestead Knoll area. Although D. ordii and D. microps may have different environmental implications, they are both kangaroo rats, which in general are extremely well adapted to arid habitats (French, 1993). As a result, chang- ing relative abundances of kangaroo rats through time in the Homestead Cave fauna should provide a clear indication of lo- cal desertification. As Figure 6 shows, kangaroo rats provided a relatively small proportion of the late-Pleistocene fauna at Homestead Cave. During the early Holocene, kangaroo rats be- came increasingly abundant. Beginning soon after 8300 14C yr B.P. (Stratum IV), more than half of the identified mammalian fauna comprised these animals. By about 6200 14C yr B.P. (Stratum VII), they began to contribute more than two-thirds of w o HI o LLJ COfs. 00 T- CO Ifl 00r- a> ^> oM CO CD COUi CO CO 00^? o 75 52 60 H 45 H 30 H 15 H FIGURE 6.?Relative abundance of Dipodomys as a percentage of the total number of identified speci- mens (NISP) of mammals at Homestead Cave by stratum, excluding Stratum X. The number of identi- fied Dipodomys specimens is above the bar for each stratum. Corresponding radiocarbon ages are indi- cated below the strata. NUMBER 33 375 Q_O DC O W CLO DC O 1.00 0.75 - 0.50 - 0.25 0.00 0.50 0.75 1.00 1.25 1.50 LOG % DIPODOMYS FIGURE 7.?Relationship between the relative abun- dance of Dipodomys (expressed as a logarithm of the percentage of Dipodomys specimens in the mammal fauna of each stratum) and the proportion of the kanga- roo rat fauna contributed by D. microps at Homestead Cave. 1.75 2.00 the specimens in each stratum and (at least in the analyzed units) were never less than this fraction in subsequent strata. Given that the increasing abundance of D. microps appears to track decreasing abundances of big sagebrush, and that in- creasing abundances of kangaroo rats in general indicate in- creasing aridity, I found that the percentage of each stratum's mammal fauna contributed by kangaroo rats is tightly corre- lated with the proportion of that stratum's kangaroo rat fauna that is contributed by D. microps (Figure 7). PYGMY RABBITS It might be argued that environmental changes other than a replacement of sagebrush by shadscale vegetation might have driven the changing abundances of D. ordii documented by the Homestead Cave fauna. After all, D. ordii does not require Atriplex in its diet. However, the history of pygmy rabbits (Brachylagus idahoensis) at Homestead Cave would seem to make sense in no other way. Pygmy rabbits are heavily dependent on dense stands of big sagebrush, both for food (pygmy rabbit winter diets may be composed of almost nothing else) and as locations for their bur- rows (Green and Flinders, 1980; Weiss and Verts, 1984; Dobler and Dixon, 1990; Katzner and Parker, 1997). Today they are found, albeit in declining numbers, in the sagebrush valleys of the northern two-thirds of the Great Basin and in the immedi- ately adjacent Columbia Plateau. There is also an isolated, and endangered, set of populations in eastern Washington. The latest Quaternary history of pygmy rabbits is reasonably well known. In the late Pleistocene, they were found as far south as central New Mexico (Harris, 1985) as well as outside their modern distribution in eastern Washington (Lyman, 1991). Within the Great Basin and adjacent Columbia Plateau, they are known to have undergone two separate declines in abundance, one at about 10,000 14C yr B.P., and a second either at the onset of or during the middle Holocene (Grayson, 1993). For instance, at Danger Cave (approximately 100 km south- west of Homestead Cave in the Silver Island Range, western Utah, at an elevation of 1314 m), 35 of 37 pygmy rabbit speci- mens were deposited between about 10,500 and 10,000 14C yr B.P., even though the subsequent sequence spans much of the Holocene (Grayson, 1988; Madsen and Rhode, 1990). The first indication that pygmy rabbits had undergone significant prehis- toric declines in abundance came from work at the Wasden Site (Snake River plain, southern Idaho), where Butler (1972) showed that they declined dramatically in abundance at about 7000 14C yr B.P. The record from the Connley Caves (Fort Rock Basin, south-central Oregon, at 1355 m) is in strong ac- cord with this observation. This site contains deposits that span the intervals from 11,200 to 7200 and from 4400 to 3000 14C yr B.P., but whereas pygmy rabbits were common before 7000 14C yr B.P., they were rare after that time (Grayson, 1977). Other sites?for instance, Gatecliff Shelter (Toquima Range, central Nevada; see Grayson, 1983)?show similar histories. Both the latest-Pleistocene and middle-Holocene declines in pygmy rab- bit abundance appear to be associated with declines in abun- dance of Artemisia tridentata (see the review in Grayson, 1993). The record for pygmy rabbits at Homestead Cave is clear (Figure 8). These animals were common in the Homestead Knoll area during the late Pleistocene and were clearly present during the early Holocene, but they declined rapidly and had become locally extinct, or nearly so, by 8300 14C yr B.P. There is no sure evidence to suggest that pygmy rabbits have been in the Homestead Knoll area during the last 7000 years or so. Given the tight association between pygmy rabbits and dense stands of big sagebrush, it again is extremely likely that A. tridentata began a substantial decline in this area after about 376 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 2.50 FIGURE 8.?Relative abundance of pygmy rabbits as a percentage of the total number of identified speci- mens (NISP) of mammals at Homestead Cave by stratum, excluding Stratum X. The number of identi- fied Brachylagus idahoensis specimens is above the bar for each stratum. Corresponding radiocarbon ages are indicated below the strata. 10,000 14C yr B.P.; and that soon after 8300 14C yr B.P., there were insufficient stands of big sagebrush to support popula- tions of these animals. The record is fully consistent with other paleontological pygmy rabbit data, and it also fully consistent with the local history for Dipodomys ordii provided by Home- stead Cave. VOLES Most species of Microtus occupy grassy habitats and live in cool environments. Indeed, these animals show poor ability to thermoregulate at high temperatures, and no species of Micro- tus is restricted in distribution to arid areas. Of the species that currently occur in the Great Basin even Microtus longicaudus, which is least tied to moist habitats, is more common in cooler, moister habitats here than in warmer, drier ones (Getz, 1985; Rose and Birney, 1985; Wunder, 1985). As Figure 9 shows, Microtus was common in the late-Pleis- tocene and early Holocene deposits of Homestead Cave. Its numbers dwindled dramatically after 8300 14C yr B.P., and it has clearly been uncommon in this area for the last 7000 14C years. Much the same is true for Lemmiscus curtatus (Figure 10). Sage voles are, as the name suggests, often associated with habitats dominated by stands of tall sagebrush (usually A. tri- 3.00 > E = D. > >m g STRATUM o X x FIGURE 9.?Relative abundance of Microtus sp. as a percentage of the total number of identified speci- mens (NISP) of mammals at Homestead Cave by stratum, excluding Stratum X. The number of speci- mens identified as Microtus sp. is above the bar for each stratum. Corresponding radiocarbon ages are indicated below the strata. NUMBER 33 377 FIGURE 10.?Relative abundance of Lemmiscus cur- tatus as a percentage of the total number of identified specimens (NISP) of mammals at Homestead Cave by stratum, excluding Stratum X. The number of specimens identified as Lemmiscus curtatus is above the bar for each stratum. Corresponding radiocarbon ages are indicated below the strata. dentata) and grasses, although they can also be found in com- munities in which the dominant shrub is rabbitbrush as well as in habitats dominated by grasses alone. Within these habitat pa- rameters, there is also some suggestion that sage vole numbers are highest in settings marked by relatively warm winters cou- pled with cool and moist summers (Hall, 1946; Maser, 1974; Carroll and Genoways, 1980). These animals seem to be able to meet their moisture requirements through the ingestion of green vegetation, and surface moisture is clearly less critical for them than it is for most species of Microtus. In short, L. curtatus is most abundant today in habitats do- mininated by A. tridentata and grasses in contexts marked by relatively cool summers. These modern habitat preferences likely account for the strong similarity between the histories of this animal and that of Brachylagus idahoensis at Homestead Cave (compare Figures 8, 10). Both were most abundant dur- ing the latest Pleistocene in the Homestead Knoll area, both continued to be present during the early Holocene, and both saw their numbers fall dramatically as the early Holocene came to an end in this area at about 8300 14C yr B.P. POCKET GOPHERS Several other taxonomic pairs show similar, though by no means identical, behavior. There are, for example, two species of pocket gopher in Utah today: the northern pocket gopher, Thomomys talpoides, and Botta's pocket gopher, T. bottae. The northern pocket gopher is mostly confined to the mountains and high valleys of the state (although it is has been reported from the Promontory Peninsula: see Durrant, 1952), whereas T. bottae resides in Utah's lower valleys. To my knowledge, only T. bottae is known from the region immediately west of the Great Salt Lake. Pocket gophers are also abundant in the Homestead Cave fauna. Approximately 15,000 specimens have been identified from the Homestead sample, of which 969 have been identified to the species level. Although Stratum I at Homestead provided only two gopher specimens that could be identified to species, both of these are T. talpoides. All identified gopher specimens from later strata represent T. bottae. POCKET MICE Figure 11 shows the conjoined histories of two of the three species of pocket mouse represented in the Homestead Cave fauna: the little pocket mouse, Perognathus longimembris, and the Great Basin pocket mouse, P. parvus (the long-tailed pocket mouse, Chaetodipus formosus, is the third species). Perognathus longimembris is common in the valleys and on the lower mountain slopes of the Great Basin and often co-occurs with P. parvus. This latter, however, ranges into altitudes far higher than those tolerated by the little pocket mouse. Hall (1946), for instance, noted that the highest elevation in Nevada from which P. longimembris had then been found was 6500 ft (1981 m) but that P. parvus routinely occupied higher elevation habitats?for instance, at 10,000 ft (3048 m) on Mt. Grant in central Nevada. Perognathus parvus is also found well into British Columbia, Canada, but the little pocket mouse exists no further north than southeastern Oregon (Hall, 1981). Given what I have reported about the faunal sequence at Homestead Cave, it is perhaps no surprise that P. parvus was the abundant pocket mouse in deposits from the late Pleis- tocene in Homestead Cave and that P. longimembris was not present in the fauna during that time. The little pocket mouse first appeared in Stratum II, sometime after 10,000 14C yr B.P. In comparison with P. longimembris, P. parvus remains fairly 378 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 121 1.00 o to DC O_ W CE Q. FIGURE 11.?Ratio of Perognathus parvus to P. par- vus and P. longimembris at Homestead Cave by stra- tum, excluding Stratum X. The number of specimens identified as P. parvus is above the bar for each stra- tum. Corresponding radiocarbon ages are indicated below the strata. m X oo o common through the early Holocene. It becomes rare, how- ever, during the early centuries of the middle Holocene, ap- pears to have become regionally extinct between about 6000 and 2000 14C yr B.P., and does not reappear until late prehis- toric times. of the bushy-tailed woodrat specimens in the entire faunal as- semblage were deposited between about 10,500 and 7000 14C yr B.P. The remaining six specimens found in Danger Cave de- posits were scattered across the subsequent 7000 years of his- tory registered in this site. WOODRATS Two species of woodrat are widespread in the modern Great Basin: Neotoma lepida, the desert woodrat, and N. cinerea, the bushy-tailed woodrat. Neotoma lepida is found in the lower el- evations of much of the Great Basin, whereas N. cinerea tends to occupy the mountains and the more northerly high valleys of the region. In addition, N. cinerea ranges as far north as the Northwest Territories, Canada, whereas N. lepida extends no further north than central Oregon (Hall, 1981). Both species are found on Homestead Knoll today (Grayson et al., 1996). Woodrats are abundant in the Homestead Cave deposits: only Stratum XVIII has fewer than 650 specimens identified to the species level. As Figure 12 shows, Neotoma cinerea was the dominant woodrat in the Homestead Knoll area during the late Pleistocene and remained common through much of the early Holocene. After about 8300 14C yr B.P., it underwent a dramatic decline; Neotoma lepida became, and has remained, the numerically dominant woodrat in the Homestead Knoll area. Indeed, bushy-tailed woodrats seem to have become lo- cally extinct around 5000 14C yr B.P. and may have recolonized the area shortly before 1000 14C yr B.P. (see "Recolonization and Small-Mammal Biogeography," below). The history of bushy-tailed woodrats at Homestead Cave is quite similar to that provided by the much smaller sample of vertebrate remains provided by Danger Cave, where 95% (125) MARMOTS Yellow-bellied marmots (Marmota flaviventris), which also tend strongly to be found only on Great Basin mountains (al- though descending to much lower elevations to the north), are likewise present in the deeper strata of Homestead Cave (Fig- ure 13). Because the Homestead Cave fauna owes its accumu- lation largely to owls, and because adult marmots exceed the prey size routinely taken by these birds, it is not surprising that marmots were never abundant in the Homestead Cave deposits. As a result, precisely when these animals became locally ex- tinct cannot be determined, but they are reasonably common in late-Pleistocene Stratum I and appear to have been present in the area until roughly 8000 14C yr B.P. That 66 specimens were present in the late-Pleistocene and early Holocene Homestead Cave samples clearly documents the presence of local popula- tions of these mammals during these periods. Just as clearly, the complete absence of marmot specimens in these deposits during the last 5000 14C years shows that they had been extir- pated from the area by the end of the middle Holocene?and perhaps long before the end of this period. Only one other well-dated stratified fauna in the northern Bonneville Basin sheds light on the history of marmots in this area. Danger Cave contains a small sample of marmot specimens that were deposited between about 10,500 and 10,000 14C yr B.P.; later deposits do not contain the remains of these animals. NUMBER 33 379 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 242 1 9 0 306 FIGURE 12.?Proportion of the woodrat fauna con- tributed by Neotoma lepida at Homestead Cave by stratum, excluding Stratum X. The number of speci- mens identified as N. cinerea is above the bar for each stratum. Corresponding radiocarbon ages are indicated below the strata. ~~ Q. CQ 300 > fl- ea 000 > VII /III STRATUM CO 000 X CQ X oo o HARVEST MICE The western harvest mouse (Reithrodontomys megalotis) is found in a wide variety of habitats but is most abundant in set- tings marked by dense grasses or other herbaceous vegetation (Webster and Jones, 1982). These animals can be found in arid contexts, but in Nevada they tend to be most abundant in places that are both grassy and near water (Hall, 1946), and they may undergo significant population reductions in response to drought (Whitford, 1976). The history of Reithrodontomys at Homestead Cave is sum- marized in Figure 14 (the numbers plotted in this figure com- bine specimens identified as Reithrodontomys sp. with those identified as R. megalotis, the only species demonstrably present in the fauna). These animals were most abundant here during the early Holocene, declined dramatically during the middle Holocene, and rebounded late in prehistoric times. Paleoclimatic Implications THE EARLY HOLOCENE Recent climatic models suggest that the early Holocene in the Great Basin was both wetter and warmer than today, with in- creased summer insolation causing increased summer mon- 0.50 <2 0.40 - 0.30 - 0.20 - 0.10 - 0.00 o oo o CO CM 0Q ,300 Q. CQ ,000 STRATUM Q_ CQ 000 FIGURE 13.?Relative abundance of marmots as a percentage of the total number of identified speci- mens (NISP) of mammals at Homestead Cave by stratum, excluding Stratum X. The number of speci- mens identified as Marmota flaviventris is above the bar for each stratum. Corresponding radiocarbon ages are indicated below the strata. 380 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1.25 w 1.00 -