SMITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 121, NUMBER 11
CfjarlesJ B. anb ifWarp "^aux OTialcottl^es^earcf) JfunbGEOLOGICAL BACKGROUND OF THElYATAYET ARCHEOLOGICAL SITE,CAPE DENBIGH, ALASKA(With Four Plates)
BYD. M. HOPKINSU. S. Geological SurveyAND
J. L. GIDDINGS, JR.University of Pennsylvania
(Publication 4110)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONJUNE 11, 1953

SMITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 121, NUMBER 11
Cfjarles; IB, anb JWarp "^aux fflSHalcott3&esJearct) jFunbGEOLOGICAL BACKGROUND OF THElYATAYET ARCHEOLOGICAL SITE,CAPE DENBIGH, ALASKA(With Four Plates)
BYD. M. HOPKINSU. S. Geological SurveyAND
J. L. GIDDINGS, JR.University of Pennsylvania
(Publication 4110)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONJUNE 11, 1953
Z^i £oti> (§aftimovi (preeeBALTIMOBE, MD.- U. 8. A.
Cfjarleg
©. anb iWarp ^aux 2!®alcott SAesfearcfj jFunbGEOLOGICAL BACKGROUND OF THElYATAYET ARCHEOLOGICAL SITE,CAPE DENBIGH, ALASKABy D. M. HOPKINSU. S. Geological SurveyAND
J. L. GIDDINGS, Jr.University of Pennsylvania(With Four Plates)INTRODUCTIONlyatayet, a stratified archeological site at Cape Denbigh, Alaska(fig. i), offers an exceptional opportunity to examine a sequence ofhuman occupations against a background of fluctuating climate. lya-tayet is the discovery site and the type locality of the Denbigh flintcomplex, the oldest large assemblage of cultural remains thus farfound in Alaska. The site was excavated by Giddings during 1948,1949, and 1950. Hopkins visited the site for three days in August1950-The discovery and excavation of the lyatayet site were importantsteps forward during the past 15 years in the study of early humanhistory in Alaska. Detailed geologic investigations during the sameperiod have clarified many aspects of the late Quaternary geomorphicand climatic history of the region. Until recently, however, little prog-ress has been made in relating archeological sequences to a geologicand climatic chronology. The younger archeological sites have beenoccupied too recently to record notable geologic changes, and most ofthe older sites have lacked distinctive features upon which to base achronology.Many problems in the archeology and Quaternary geology of Alaskaawait the accumulation of additional evidence. A better understandingof the time relationship between archeological and geological eventswill assist in their solution. The precise dating of these events, the
SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 121, NO. 11
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
nature of the Alaskan environment during early periods of humanhistory, and the nature and availability of routes for the spread ofpopulation at various times are among the problems w^hose solutionsdepend upon evidence to be gathered from both geological and archeo-logical investigations.Three culture layers at lyatayet separated by sterile zones or sharpphysical discontinuities record three distinct periods of occupationseparated by long periods during which the site was abandoned(table i). The deeper layers indicate by their character that they havebeen subject to soil movements and to soil-forming processes that areno longer active at lyatayet. These layers record a sequence of cli-matic changes similar to known sequences in other regions in Alaska.Converging lines of evidence indicate that the earliest dwellers atlyatayet lived during a warm period in Alaska more than 8,000 yearsago and perhaps more than 12,000 years ago. The later dwellers livedduring two periods in the past 2,000 years, and their deposits reflectminor climatic fluctuations in the Bering Sea region.GEOGRAPHIC SETTINGlyatayet is located on the west face of the Reindeer Hills, a fewmiles north of Cape Denbigh, Alaska (fig. 2). The Reindeer Hillsare isolated bedrock hills, ranging in altitude from 700 to 1,000 feet,which form the western extremity of a peninsula jutting from the eastcoast of Norton Bay near the village of Shaktolik. On the landwardside the hills slope gently eastward and merge into a poorly drainedcoastal plain which composes most of the peninsula. On the seawardside the hills terminate abruptly in rugged sea clififs, 100 to 300 feethigh, extending with only minor embayments from Cape Denbigh toPoint Dexter.The hills are drained by several straight, shallow valleys that enterNorton Bay between Cape Denbigh and Point Dexter. Most of thesevalleys have been truncated by the retreat of the sea clififs, and thestreams descend to the sea in cascades at the valley mouths. Thelargest stream, lyatayet Creek, has excavated its valley apace with theretreat of the cliffs, and its valley floor is graded to present sea level(pi. i). The lyatayet site is at the mouth of lyatayet Valley.Bedrock of the Reindeer Hills consists of coarsely crystalline,banded, reddish-brown and light-gray marble.^ The marble is tightly
1 The Reindeer Hills are shown erroneously as basalt on Smith and Eakin'sgeologic map (1911, pi. 6) of southeastern Seward Peninsula. Small bodies ofbasalt possibly may be present, but none were seen by the writers.
NO. 1 1 lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS
4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121folded and is cut locally by large dikes of dense, dark, igneous rocklargely altered to serpentine. Silt and peat lo to 20 feet thick underliethe coastal plain to the east of the Reindeer Hills. At least 100 feetof sand and gravel is believed present beneath the silt and peat.
tc e M nllnI t
Fig. 2.—Map of Norton Bay area showing location of lyatayet. After NortonBay quadrangle, Alaska Reconnaissance i : 250,000 series, U. S. Geological Sur-vey, 195 1.There is no evidence of glaciation in or near the Reindeer Hillsduring any part of the Quaternary.Perennially frozen ground is present throughout the coastal plaineast of the Reindeer Hills, except beneath lakes and streams. In theReindeer Hills, perennially frozen ground probably is present beneath
NO. 1 1 lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 5
areas that slope less than 5° and that are mantled by more than 2 feetof turf, peat, or silty soil. It is lacking, however, beneath slopesas steep as the walls of lyatayet Creek and beneath areas of exposedbedrock or rubble.No weather records are available for the Reindeer Hills or forShaktolik. Interpolation from weather records at Nome, St. Michael,and Nulato (U. S. Weather Bureau, 1943) indicates that the ReindeerHills have a mean annual temperature of about 25° F. The mean pre-cipitation is about 14 inches, of which about half falls during a well-defined rainy period from July through September. More than a thirdof the annual precipitation falls as snow.GENERAL FEATURES OF lYATAYET VALLEYlyatayet Creek is a small, perennial stream about a mile long. Thecreek flows on bedrock in a sharply cut valley, the walls of whichslope 10° to 15°. A narrow terrace 40 feet above the valley floor canbe traced 1,000 feet upstream from the mouth of the valley (fig. 3).Farther upstream on the northeast valley wall is an area of irregulartopography at approximately the same altitude as the terrace surface.The irregularities consist of semicircular benches 10 to 30 feet wide,separated from one another by steep fronts about 10 feet high. Thesefeatures, termed lobate soil terraces or soil lobes (Sigafoos and Hop-kins, 1952), formed at a time when the climate was colder. At presentthey are not active.A dense stand of alders, 6 to 10 feet tall, covers most of the lowerwalls of lyatayet Valley (pi. i). The alder thickets are interruptedlocally by small parklike areas in which there are dense stands of tallbunchgrass or open stands of dwarf birch, Alaska tea, blueberry, cran-berry, and spirea. Sedges, Equisetum, and scattered small alders, ar-ranged in stripes parallel to the slope, grow on poorly drained, gentlysloping areas on the higher parts of the valley walls (upper left, pi. i).Scattered spruce may grow elsewhere in the Reindeer Hills, but thereare none in lyatayet Valley.The steep slopes of lyatayet Valley are stable and free of solifluctionor other types of mass movement today and have been for manycenturies, as indicated by the lack of active soil lobes or other activefrost features, by the wide distribution of an undisturbed soil profileat the surface, and by the continuous cover of large, healthy, unde-formed alders. Solifluction probably is active, however, on higher,gentler slopes where drainage is poorer.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
O (J o,
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 7GEOLOGY AND SOILS OF lYATAYET VALLEYBedrock of marble lies a few feet beneath the hill slopes above theterraces. It is mantled by residual weathered material consisting of an-gular fragments of marble in a matrix of sandy silt. The fine-grainedmatrix consists chiefly of the least-soluble minerals in the marble.Feldspars and unidentified alteration minerals predominate in samplesexamined by Hopkins ; fine-grained muscovite is abundant ; and quartz,apatite, tremolite, and diopside are present in small quantities. Cal-cite constitutes more than 95 percent of the bedrock but is scarce orlacking in the fine-grained matrix of the weathered mantle.The 40-foot terraces on each side of the mouth of lyatayet Creekare composed of similar weathered marble debris. The terrace fill isvaguely stratified and poorly sorted ; individual fragments generallyare subangular. The deposit shows little evidence of water or wavehandling. It consists chiefly of material introduced into the valleybottom from the neighboring slopes by mudflow or solifluction duringa past cold period, and thus may be termed congeliturbate (Bryan,1946, p. 640). The apparent absence of turf or peat layers within theterrace fill suggests that accumulation was rapid and continuous.The top of the congeliturbate in the terrace represents approxi-mately the level of the bottom of lyatayet Valley after the accumulationof the fill and before the present era of stream cutting (fig. 4). Alayer of sandy silt up to 18 inches thick overlies the congeliturbate onthe surface of the terrace and at many places higher on the valleyslopes. An ancient soil profile is preserved at the top of the congeli-turbate, beneath the sandy silt, but has been removed or masked bylater soil-forming processes where the sandy silt is lacking (fig. 5).The top of the buried soil is marked by a i-inch layer of sticky,ashy-gray silt in which sand and rocks are lacking. The upper 2 or 3inches of congeliturbate beneath the silt are stained yellow-brown,grading downward into pale yellow or light olive-gray. Marble frag-ments increase in abundance and freshness with increasing depth inthe congeliturbate.The ashy-gray silt superficially resembles volcanic ash. A sampleexamined by Theodore Woodward, U. S. Geological Survey Petro-graphic Laboratory, consists chiefly of unidentified alteration minerals.Muscovite and orthoclase feldspar are common ; a few grains of quartzand hypersthene or enstatite were recognized. No volcanic glass orfragments of obvious volcanic origin were observed. All mineralspresent are present also in the underlying congeliturbate. The ashy-
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
^%} >
E IS
1 I fe
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 9gray silt, therefore, is believed to be derived by vv^eathering from theunderlying congeliturbate.The ashy-gray silt resembles the characteristic A 2 horizon of anancient podzol. The underlying 2 or 3 inches of stained congeliturbaterepresent the B horizon. Kellogg and Nygard (1951, pp. 49-58) re-port similar but much thicker and better-developed podzol profiles ata few localities elsewhere in Alaska. A thick turf layer forms the Aihorizon of modern podzols but has not been preserved in the buried
Subortt'le Brownf«r»»t Sou JBuriod podzol
Saborctlc BrownForett Soil
EXPLANATION
of . f , I
SOIL COLORS
I I Sonfly silt.Ashy-crpy silt.
Eed-brovm.Pnle red-brown,Yellov.'-hrovm.pale ji^llov orolive-pray.
E°€)aS?l Hooks in silt.Fig. 5.—Diagrammatic sketch showing distribution and arrangement of soillayers in lyatayet Valley. Ancient podzol is preserved in areas where it waslater covered with a layer of silt (left) but has been destroyed by later soil-forming processes elsewhere (right). Left side of section is approximately 4feet high.
soil at lyatayet. The buried soil profile records a period of relativelywarm climate and soil stability after the filling of the valley with theunderlying congeliturbate and preceding the deposition of the over-lying sandy silt during cold periods.The sandy silt that overlies the buried podzol is similar in size gradeand mineral composition to the matrix of the congeliturbate, but it isfree of large rock fragments. Locally it is laminated; the laminaeare contorted and range from one-eighth to one-half inch in thickness.A 2-inch layer of peat appears within the lower part of the silt in oneof the archeological excavations (pi. 2), and a Subarctic Brown Forestsoil profile appears beneath the peat.
10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121The sandy silt consists mostly of material washed from the frontsof active soil lobes higher on the slopes of the valley during a coldperiod. Small fans of similar material appear below soil lobes activeon Seward Peninsula at the present time. Upon thawing in spring theupper soil layers within the lobes become saturated with water. Un-derlying, still-frozen soil prevents excess moisture from percolatingdeeper, and a stout turf mat retards lateral drainage. Part of the ex-cess water eventually escapes through open, fibrous parts of the turf,carrying with it a load of silt and fine sand. Rocks are retained withinthe lobe by the turf. The sand and silt then are deposited on lower,gentler slopes (fig. 7).Wind-blown sand and silt, derived from beaches and coastal dunesat a time when sea level was lower and the strand line lay at leastseveral miles from the Reindeer Hills, may be included in the sandysilt layers, as well as in the underlying congeliturbate.A well-developed soil profile similar to the widespread SubarcticBrown Forest soil in interior Alaska described by Kellogg and Nygard(1951, pp. 49-58) appears beneath the surface turf throughout mostof lyatayet Valley. The profile is developed in sandy silt, congeli-turbate, or earthy midden, wherever these materials are near the sur-face. Similar soil profiles are found beneath ground levels buried bymidden and beneath the peat layer enclosed in the sandy silt (pi. 2).The soil profile generally consists of a layer of turf 4 to 8 inchesthick, underlain by 5 to 8 inches of rocky mineral soil deeply stainedreddish brown and containing a considerable admixture of organicmaterial. At greater depth the mineral soil grades in color throughyellow-brown into pale yellow or light olive-gray. Marble fragmentsare scarce and deeply weathered in the upper reddish-brown zone,where the profile is developed in congeliturbate, but rocks increase inabundance and freshness at greater depth (fig. 5, right side).Kellogg and Nygard's schematic soil association map of Alaska(1951) indicates that tundra soils predominate on the east coast ofNorton Bay. Tundra soils undoubtedly are present in less well-drainedsites, but the surface soils in lyatayet Valley resemble more closely theSubarctic Brown Forest soil.The Subarctic Brown Forest soil profile at the surface has developedsince active soil movements ceased on the slopes of lyatayet Valley.The similar buried soil profile near the base of the sandy silt in pitZ-4 (pi. 2) may record another period of warm climate and soil sta-bility that interrupted the accumulation of sandy silt after the buriedpodzol was developed.
NO. 1 1 lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS IITHE lYATAYET ARCHEOLOGICAL SITElyatayet Valley has been occupied by human beings during threedistinct periods (table i). Evidence of occupation is found on thesurface and slopes of the terrace remnants on each side of the valley.Cultural remains of the earliest period have been termed the Denbighflint complex (Giddings, 1951) and include a few artifacts similar tothose found at several sites of seemingly great age in the interior ofAlaska (Giddings, 1950; Irving, 1951 ; Solecki, 1951). The Denbighculture layer is overlain by a sterile layer of sandy silt, and then byTable i.—Summary of archcological and geological features of lyatayet Valley
IArcheologicalsequence
12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121Gaps in the layer caused by soil movements are found near the inneredge of the terrace. Filled excavations and house sites of the Paleo-Eskimos interrupt the continuity of the layer elsewhere on the terracesurface. The flint layer ends abruptly near the front edge of the ter-race and is replaced on the steep foreslope by a diffuse zone of mixedDenbigh flint and Paleo-Eskimo material (fig. 4). The Denbigh flintmaterial in the mixed zone probably consists of material encounteredin house excavations and thrown over the terrace scarp with otherrefuse by Paleo-Eskimos.Cultural material of the Denbigh flint complex consists of sparselydistributed flint flakes and artifacts (chert, obsidian, and chalcedony,in order of their abundance) pressed flush against the buried podzol.Neither these nor the small beach pebbles and angular marble frag-ments found in some parts of the layer are intruded into or belowthe underlying ashy-gray silt. Recognizable organic remains are ex-tremely scarce and consist of tiny fragments of charcoal and decom-posed bone. No house floors or tent rings have been recognized, buta series of five hearths, roughly in line, have been uncovered at theupslope edge of the terrace. Four of the hearths are shown on figure 6.The distribution of cultural material in the Denbigh culture layer,pressed flush against the surface of the buried soil, is open to twointerpretations. The site may have been occupied only during winter,when the ground was frozen to the surface and was too hard to allowstones or wood to penetrate under the pressure of human feet. Ifthis were the case, however, it is difficult to understand why the stonesof hearths and the ground around these obviously fire-exposed areasalso should be limited by the underlying dense soil. A more appealinginterpretation would be that the campers at lyatayet left their hearthsand flinty materials on top of a sod cover, the organic parts of whichhave long since decomposed and disappeared. Occupation would thenbe limited to no particular season, but probably would be limited torather brief visits. Under continuous occupation the protecting turflayer should have been broken, permitting penetration of artifacts intothe soil, as occurred during the later Paleo-Eskimo occupation.Nearly 1,500 artifacts have been found in the Denbigh culture layer.The collection is characterized by tools and techniques that, with fewexceptions, bear little resemblance to cultural materials generally rec-ognized as Eskimo (Giddings, 1951).Burins, or "gravers" in the European sense, and their spalls dis-carded in sharpening comprise about a quarter of the Denbigh flintcollection (Giddings, 1951, figs. 59a, 6oa). Burins apparently had notbeen recognized elsewhere in America before the excavation of lyata-
EXPLANATION
vertical f horizontal scalee I t t * s reefMAPs \ \ s Area In whith Denbigh\ ^ ^ ^ flint layer li o/u/jiqAtit cf fcH in Denbighj-s^ fttnt laytr, Hachuras^*"m, hdicott JMeper croi/arturiit4 liab
'i^^'-^^Charrtd grea--
''"nri> probably a hearth'^c* Larqe stoneiSECTIONS
NeO'EskimomiddenPalst-CskimomiddenSterile sandytiltDenbiah flint fST tlayer r^~\Roeks, eand,(. silt EI3
/\\\\ N\\\ \\
^^^.|\ "^ \^ ^ \ \ \ v"
West Wall, pes
^^^s:^SP?p
v\ \ Rocks, .iand.am silt Longitudinal Section olonq Nortli WallFig. 6.—Map and cross sections showing folds in the Denbigh culture layer in pits PE-ithrough PE-8. Note large gaps where Denbigh culture layer is missing upslope from folds(lower left part of map). A well-defined turf layer separates the Paleo-Kskimo and Neo-Eskimo middens in many parts of the site but is lacking in the area shown here.
14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
yet. A variety of forms are present, several of which fall into typesknown from the Aurignacian and later periods of European pre-his-tory. Referable to the same periods and places are a few otherspecial forms found at lyatayet, including keeled scrapers, flake knives,and gravette-like blades (op. cit., fig. 59b).Lamelles and cores—microlithic blade developments known asearly as the late Paleolithic, and with a geographic spread that in-cludes large parts of Eurasia and Africa—comprise more than halfof the collection (Giddings, 1951, fig. 6ib). These have been reportedpreviously from a few other sites in America. Many of the smallblades at lyatayet have been carefully pressure-flaked diagonally toproduce the most delicate and elaborate microlithic forms known tothe authors.Several points or blades (Giddings, 195 1, fig. 60), a form of "chan-neled" scraper (op. cit., fig. 62), and a single graver (op. cit., fig, 59b)resemble forms found in "Yuma" ^ and Folsom sites of southwesternUnited States so closely that there can be little doubt of technologicalrelationship.A few other blades and scrapers are unique at lyatayet, and stillothers are of a generalized nature that renders them presently unclassi-fiable. Short, wide, and thin blades with concave bases (Giddings,195 1, fig. 64) closely resemble specimens found hafted as sealingharpoon blades at Ipiutak (Larsen and Rainey, 1948). A bit of birchbark that appears to have been part of a lashed or sewn vessel con-stitutes the only recognizable organic artifact recovered in the Denbighflint layer.Assemblages of artifacts in the Denbigh flint complex tradition havebeen found in the Kugurorok River Valley (fig. i) in the eastern partof the Brooks Range (Solecki, 1951), in Anaktuvuk Pass in the cen-tral part of the range (Irving, 1951), and near the Anaktuvuk Riverat the north front of the range (Solecki and Hackman, 1951). TheKugurorok River site occupies glaciated bedrock knolls, and the Anak-tuvuk River site occupies unconsolidated deposits of glacial origin.The Anaktuvuk Pass site undoubtedly also lies within a glaciated area.These sites were not necessarily covered with ice, however, duringthe latest glacial advance in the Brooks Range. At least four distinctPleistocene ice advances, each less extensive than its predecessor, are
2 This term is used here in the sense generally understood by archeologiststo include several forms and techniques of early-man flint work. Although"Yuma" is in current disfavor, it has not been replaced by a similarly compre-hensive term.
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS I5
recognized farther east by R. F. Detterman, of the U. S. GeologicalSurvey (personal communication).Hopkins examined air photos of the Anaktuvuk River site. Themorainal ridges upon w^hich the site is found appear similar in degreeof modification and dissection to morainal ridges deposited during thesecond of the three glacial advances recognized by Detterman. Theother two sites at which Denbigh cultural material is found may alsobe in areas that remained ice-free during the latest glacial advance inthe Brooks Range. Solecki (1951) states that "the fact that theseearly manifestations were found in glaciated areas conclusively pointsout that these sites were occupied within post-glacial times." On thebasis of present knowledge, however, it can only be said that the sitesare younger than an early glacial advance ; they are not necessarilyyounger than the latest Pleistocene glacial advance.Cultural material comparable with, but believed by the writers tobe younger than, the Denbigh flint complex characterizes the lowerlevels of stratified cave deposits excavated by Helge Larsen at TrailCreek, Seward Peninsula (fig. i). Mesolithic and early Americantraits, including core and microblade, diagonal flaking, and "Yuma"-like points, are common to the lower levels at Trail Creek and to theDenbigh flint complex at lyatayet. The Paleolithic traits of the Den-bigh flint complex, including burins and spalls, are completely lacking,however, at Trail Creek. These relationships suggest that no part ofthe Trail Creek sequence known at present is as old as the Denbighflint complex at lyatayet.A carbon-14 analysis of willow stems collected in one of these lowerlevels at Trail Creek, but not definitely associated with cultural ma-terial, gives an age of 5,993 ±280 years (Johnson, 1951, p. 16).^ Be-cause of the abundant sources of sample error in carbon-14 analyses,data obtained from single samples must be treated with extreme cau-tion (see discussion of sample error in Flint and Deevey, 195 1, pp.259-260; and Bartlett, 1951). If, however, future analyses confirmthe age of the lower part of the Trail Creek sequence, one may assumethat the Denbigh flint complex at lyatayet is older than 6,000 years.The so-called "Siberian Neolithic," a widespread cultural horizonknown from the Ural Mountains to the Lena River, contains flint
3 A sample of charcoal from the Denbigh flint layer was submitted for ageanalysis at the end of the 1951 field season. When this sample was found toosmall for reliable use, a special trip to lyatayet was scheduled for the 1952 fieldseason. Aided by Alex Ricciardelli, Giddings obtained charcoal enough formore than one analysis from the Denbigh flint layer, from which carbon-14 datesare anticipated in the near future.
l6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
traditions strongly suggestive of connectives both with the Denbighflint complex and with later Paleo-Eskimo flint developments inAlaska. Briefly, the Siberian Neolithic includes microblade and coreindustry, diagonal flaking, pottery in abundance and variety, and well-preserved bone and other organic material. Microburins are reportedat the Kullata site near Yakutsk (Okladnikov, 1950) but are lackingat other sites. According to Grigoriev (1950, p. 177), Okladnikovbelieves that the Kullata site was occupied during a warm, dry period4,000 to 6,000 years ago ; other Siberian Neolithic sites are assigneddates by Okladnikov and others of 6,000 to 3,000 years ago.Full discussion of the relationships between the Siberian Neolithicand cultural horizons represented at lyatayet is reserved for a laterpaper by Giddings. For the present it may be said that the Siberianmaterial appears to the authors transitional between the Denbigh flintcomplex and the Paleo-Eskimo culture and that the Siberian Neo-lithic is believed to represent a later culture than the Denbigh flintcomplex.*The several lines of archeological evidence described above indicatea considerable antiquity for the Denbigh flint complex. With the ex-ception of the probable harpoon blades, none of the artifacts can beconsidered characteristically Eskimo, and thus a profound separationfrom later, Eskimo cultures is indicated. Many of the tools and tech-niques of the Denbigh flint complex are found typically in UnitedStates and Old World sites occupied prior to the "postglacial" thermalmaximum ; some of these distant sites were occupied still earlier inthe last (Wisconsin or Wiirm) glacial stage. The traits common tothe Denbigh flint complex and to ancient sites in the Old World orthe United States are so distinctive and so unlikely to have been in-vented twice that one must assume that the Denbigh flint complexis in a direct line of heritage with these far-distant sites. The archeo-logical evidence argues strongly for an age greater than that of thethermal maximum as recognized in the United States, believed to havebegun about 6,000 years ago (Flint and Deevey, 1951, p. 258), andmay permit an age greater than that of the Mankato substage ofglaciation, believed to have reached its climax between 8,500 and11,000 years ago (op. cit., pp. 261-267; Schultz, Lueninghoener, andFrankforter, 1951, pp. 34-36). Geological evidence, summarized in
* The authors acknowledge with full respect the views expressed in writingand in conversations by Profs. H. L. Movius, Jr., and Lauriston Ward thatmore resemblances to the Denbigh flint complex appear in the Siberian Neolithichorizons than in known sites of the east Asiatic upper Paleolithic, and recognizein these views a difficulty of correlation with the dating herein proposed.
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 1/
a later section, indicates that the Denbigh culture layer at lyatayet wasdeposited during a warm period at least 8,000 years ago and possiblymore than 12,000 years ago.FOLDS IN THE DENBIGH CULTURE LAYERComplex folds with axes approximately parallel to the surface con-tours disturb the Denbigh flint layer in several parts of the site (pis. 2and 3). The folds are best developed near the base of the steep bed-rock slope at the inner edge of the terrace, but gentler folds are notedelsewhere (fig. 4). Involved in the folds are the upper part of thecongeliturbate, the ashy-gray silt, the Denbigh culture layer, and atleast part of the sterile sandy silt. The folds are not reflected at thesurface or in any part of the Paleo-Eskimo layer.Axes of the folds are broadly arcuate and convex downslope. In-dividual folds decrease in ampHtude laterally and fade out a few tensof feet from the point where they are best developed. A younger setof folds in one part of the site appears superimposed upon an olderfold (pit PE-2 in figure 6).The folds are asymmetric or overturned, with the steeper or over-turned limb on the downslope side. Thin beds retain their continuityunbroken within the folds, but upslope from each set of folds arefound gaps several feet wide in which both the Denbigh culture layerand the ashy-gray silt are missing. Flint chips and stones pressedflush against the lower surface of the Denbigh layer faithfully reflectthe attitude of the layer, being horizontal and above the ashy-graysilt on the upslope limbs, vertical at the crests of the folds, and tiltedbeneath the ashy-gray silt on the overturned limbs.The folds in the Denbigh culture layer formed in the course ofdevelopment of soil lobes in the overlying sandy silt (fig. 7) duringa cold period long after the early occupation ended. Active soil lobesstudied by Sigafoos and Hopkins (1952) creep downslope during latespring, when the thawed soil near the surface is saturated with melt-water. At deeper levels the soil remains frozen in position and cannotpartake of the surface movements. Differential heaving during thefall freezing season results in similar differential movements (op. cit.,fig. i). Shear resulting from frictional drag between mobile, satu-rated, sandy silt above and immobile, frozen, rocky debris below pro-duced the intricate folds in the Denbigh culture layer. The remarkablecontinuity of distinctive zones throughout the folds indicates that themovements were slow, perhaps totaling a few inches each year. Dur-ing the folding, the total downslope length of the area occupied by
i8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
STAGE I
STAGE or
STAGE nr
THAWED SOILPERSISTENT FROZEN SOIL[I^ DENBIGH FLINT LAYER
Fig. 7.—Development of folds in the Denbigh culture layer by creep and solifluction.Stage i : Ground thaws during late spring but frozen ground is still present below thesurface. Denbigh culture layer is thawed in most places but locally is anchored firmly infrozen ground (A). Stage 2: Saturated surface layer creeps downslope, forming soil lobe.Thawed part of Denbigh culture layer overrides part of layer still anchored in frozen ground(A). Gap forms upslope from fold, downslope from another area where Denbigh culturelayer is frozen (B). Stage 3' Front of soil lobe advances, overriding surface turf, duringspring in a later year. Additional folds are formed in Denbigh culture layer and gapwidens where layer is missing. Note tilted shrub with roots trailing upslope.
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS I9
the Denbigh flint layer was shortened. This shortening was compen-sated by the formation of gaps in the layer upslope from the folds.The peat layer found within the sandy silt in one pit probably is aremnant of the strong turf cover that retained the soil lobes.THE PALEO-ESKIMO LAYERRemains of the Paleo-Eskimo culture are found in a layer i8 to36 inches thick on the top and steep foreslope of the terrace on thenortheast side of lyatayet Creek. The Paleo-Eskimo layer fades im-perceptibly at its lower boundary into the sterile sandy silt that coversthe Denbigh flint layer. A layer of grass sod locally separates thePaleo-Eskimo zone from the overlying Neo-Eskimo layer, but in someplaces the two layers are in direct contact. They are distinguishedeasily, however, by a sharp break in the degree of preservation oforganic matter.The Paleo-Eskimo layer generally consists of a rich, red-brown orlight yellow-brown mixture of humic material and silty soil, but insome places it consists of soft, friable, black, peaty material. In manyparts of the site, one or more former ground levels can be distin-guished. Paleo-Eskimo artifacts are found in the sandy silt beneaththe buried ground levels in attitudes suggesting that they have beenthrust or tramped into the ground. Cross-bedded middens are foundabove the buried ground levels. One house and several charred areashave been recognized. Many filled excavations are present (pi. 2) ;some of these extend below the Denbigh culture layer.The Paleo-Eskimo cultural material consists of stone tools, pottery,beach pebbles, angular marble fragments, charcoal, and abundant butpoorly preserved organic material. Hard, firm bone, antler, and ivoryare rare ; generally these materials are preserved as small, friable frag-ments of about the same consistency as the enclosing sandy silt. Shellsare abundant, but only the papery, chitinous cover of the limy valvescan be recovered. Wood generally is represented by a few limp fibersor by a humic stain in the enclosing silt and sand. Rarely a recoverablelog is found.In striking contrast with the Denbigh flint complex, the Paleo-Eskimo layer contains the sorts of household equipment, tools, andweapons commonly associated with Eskimo developments locally andNeolithic developments in other parts of the world. Stone lamps ofa triangular, "sadiron" shape, as well as of irregular forms, have beenpecked out of igneous or metamorphic rock in relatively large num-bers. Smaller stone vessels of teacup and soap-dish size and shape may
20 SMITHSONIAN MISCEXLANEOUS COLLECTIONS VOL. 121have been mortars and paint dishes rather than lamps. Thick potteryvessels were occasionally made in cnide imitation of stone lamps, butthe great bulk of potsherds collected were those of thin, well-firedcooking pots. The pots had been either round or conical-based orflat-based, and had been imprinted over all the outer surface withstriations or small check-stamps. Cord markings appear on a fewsherds, and two or three vessels represented in the collection weretreated on the outer surface only by smoothing. All in all, the methodsof treating this pottery are in keeping with upper Neolithic practicesin Asia and with Woodland and other traditions of earliest potterymaking in the eastern United States.The grinding and polishing of slate to be seen on blade fragments arein no case as refined as are these techniques on Neo-Eskimo artifacts.Hea\'y scratches characterize the semilunar and other knife blades ofthe Paleo-Eskimos. A finer polishing distinguishes the tongiie-shaped,partly chipped adz blades and the long, medial labrets of stone andjet that also are found in the Paleo-Eskimo deposits. A polished formof small, hard stone implement appears to be a cutter or groover tobe used in the manner of a burin ; it is fairly numerous, and is charac-terized by the meeting of two planes at nearly right angles at one nar-row edge. Adz blades, labrets, and burinlike instruments are made ofhighly silicified slate or other materials that could be prepared byconchoidal flaking. Whetstones of several shapes and grain sizes aswell as fragments of larger grinding stones appear in both Paleo-Eskimo and Xeo-Eskimo deposits, but are lacking in the Denbigh flintcomplex.The flinty materials chosen by the Paleo-Eskimo were mainlybasalts and silicified slates—materials almost totally lacking in theDenbigh flint complex. They were skillfully handled, however, andfurnished a base for most of the weapon points and other blades.Some use was made of chert and obsidian, but it is often diflicult todetermine which of the objects made of these materials belong to themiddle levels and which were displaced from underlying disturbedsections of the Denbigh flint layer.The few objects of organic materials obtained from these levelsare, in almost all cases, relatable to pieces from the Near Ipiutak ^houses and middens at Point Hope. These include harpoon heads oftwo types, arrowheads slotted for side blades and for end blades, bone
5 Larsen and Rainey (1948) recognized the contents of a few isolated houseand midden deposits in Point Hope as distinct from Ipiutak cuUure materials,though closely related. They named this aspect "Near Ipiutak" but were unableto say with certainty whether it represented an earlier or a later stage of culture.
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 21
and ivory awls or marlinespikes, ice picks, flaking hammers, and anengraving tool. Engraved decoration is lacking. A single crude, arm-less "doll" of ivory has a long trunk and an oval head, but lacks facialfeatures.All in all, the materials obtained from the Paleo-Eskimo levels atlyatayet show close identity with those from Near Ipiutak at PointHope, and much less similarity to those from Ipiutak proper.No part of the Paleo-Eskimo site is perennially frozen, and the poorpreservation of organic material indicates that it has not been peren-nially frozen during most of the time elapsed since the period of Paleo-Eskimo occupation. Small-scale plications in charcoal lenses charac-terize the deeper parts of the Paleo-Eskimo layer, below old groundlevels. Some of these were produced by the formation and destructionof clear ice lenses during annual cycles of freezing and thawing ; otherswere formed by local slumping as wood and other organic material inthe soil rotted away. Still others probably resulted from the pressureof tramping feet when the ground was wet and soft each spring.Large-scale folds, comparable with those involving the Denbigh cul-ture layer, are lacking. A Subarctic Brown Eorest soil appears tohave developed beneath Paleo-Eskimo ground levels, indicating thatthe soil was relatively stable and immobile. Additional quantities ofsandy silt, however, were washed into some parts of the site. Some ofthe silt may have been introduced during brief cool periods when soillobes formed higher on the slope, but most of the added sedimentprobably was derived from paths, dog holes, and other artificial bareareas. Intrusion of Paleo-Eskimo artifacts below the surface on whichthey lived indicates that occupation continued for long periods, throughsummers as well as winters. Carbon- 14 analysis of a charcoal sampleindicates that an older part of the Paleo-Eskimo zone at lyatayet is2,016 ±250 years old ; a younger phase is 1,460 ±200 years old (sam-ples 563 and 506—Johnson, 195 1, pp. 15-16).THE NEC-ESKIMO LAYERDeposits of the Neo-Eskimo culture are found on both sides oflyatayet Creek, on the steep slopes as well as the surfaces of theterraces. The culture layer is found immediately below the surfacesod. A well-defined sod layer underlies much of the Neo-Eskimomidden and separates it from the deeper Paleo-Eskimo layer. Theculture layer ranges from a few inches to 6 feet in thickness and con-sists of peaty, dark-brown to black midden material. Parts of the mid-den are perennially frozen, although underlying Paleo-Eskimo midden
22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
is thawed. The midden shows no evidence, however, of disturbanceby frost action.Organic materials, including bone, antler, ivory, and bivalve shells,are abundant and well preserved. Wood is abundant but is generallysoft and friable. Some logs are fresh enough, however, to permitcutting sections for tree-ring study.All the later deposits at lyatayet contain sparse cultural materialsdisplaced from older levels. This is to be expected in any village orcamp site that lies above another, not only because of the Eskimo prac-tice of digging house and cache pits, but also because of a curiosity andarcheological bent of many individuals such as may be seen amongEskimos living in the area today. No doubt the Paleo-Eskimos wereas keenly interested in the microlithic work of the earliest inhabitantsof the site as are the Neo-Eskimos of the present day in all the curiousearly workmanship. The earlier artifacts are readily distinguished,however, in the deposits of the Neo-Eskimos. This is partly becauseof a difference in tradition of style and workmanship between the twoperiods. A more definitive aid in separating the later deposits at lya-tayet, however, is Giddings's excavation of a pure Neo-Eskimo site,Nukleet, on the south side of Cape Denbigh (Giddings, 1949). Sev-eral thousand artifacts from the Nukleet site delineate a sequence ofstylistic changes that affords a rather complete picture of the Neo-Eskimo in the Cape Denbigh region during their period of occupation.Dendrochronological studies indicate that occupation extended fromabout A.D. iioo to 1600 (op. cit., p. 86).The Neo-Eskimos were a fishing, sealing, caribou-hunting peoplewhose dependence on the inland, forested regions nearby is attestedby their extensive use of beaver teeth as knives, their use of birchbark as containers, and their extensive use of antler. An enumerationof their other artifacts is in preparation ; we may note, however, thattheir cultural balance seems much like that of present-day Eskimosliving at points along the northeastern Bering Sea,LATE QUATERNARY HISTORY OF lYATAYET VALLEYCLIMATIC SIGNIFICANCE OF FEATURES OF lYATAYET VALLEYA series of changes in climate and in the position of sea level maybe inferred from the geologic and archeologic deposits in lyatayetValley (table i). The recognized history begins with the carving ofthe bedrock valley. Because the bedrock valley is graded approxi-mately to present sea level, one must assume that sea level stood nearits present position at that time. The climate probably was not colderthan at present.
NO. 1 1 lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 23A subsequent period of valley filling during an interval w^ith coldclimate is recorded by the 40-foot terrace. Lack of rounding and theunsorted character of the debris indicate, in this region, that the fillconsists of congeliturbate introduced from neighboring slopes by soli-fluction during a period of intense frost action. The lack of wave-handled material in the terrace fill and the absence of elevated wave-cut rock terraces elsewhere along the rugged coast between CapeDenbigh and Point Riley indicate that sea level stood no higher duringthe deposition of the congeliturbate than it does at present. Instead,sea level is believed to have been lower. The shore line probablyshifted several miles—perhaps several hundred miles—across the flatfloor of Bering Sea away from the Reindeer Hills.lyatayet Creek was unable to transport all the load furnished bysolifluction on the valley walls, and the valley was filled to a depth of40 feet. The fill probably extended as an alluvial fan beyond themouth of the valley across the dry floor of lyatayet Cove.Valleys draining the hills between the Koyuk and Kwik Rivers, 40miles north of Cape Denbigh (fig. 2), resemble lyatayet Creek afterits valley was filled with congeliturbate. Minor streams within thehills in the Koyuk-Kwik area meander across flat, debris-choked val-ley bottoms (pi. 4). Longitudinal profiles of the valleys are steep.The streams debouch onto adjoining coastal lowlands across broaddebris fans similar to that inferred at the mouth of lyatayet Creek.A small rise in sea level in the Koyuk-Kwik area would subject theunconsolidated fans to erosion by waves and longshore currents, andthe fans would be removed rapidly to the edges of the bedrock hills.Stream gradients at the mouths of the valleys would be steepened,and the streams then would re-excavate the valleys, leaving remnantsof the valley fill as terraces like those in lyatayet Valley.The podzol soil profile, developed in the congeliturbate during awarm interval while lyatayet Valley was occupied intermittently bypeople of the Denbigh flint culture, records a brief but importantinterruption in the filling of the valley. The profile indicates a periodof soil stability during which soil movements ceased on the adjoiningslopes, to be resumed when the podzol with its included artifacts wasburied and folded beneath sandy silt washed in from the walls. Theprecise climatic significance of the podzol is uncertain. Podzols aremost widely distributed in Alaska in areas with a warmer summertemperature than that in the Norton Bay region, but Subarctic BrownForest soils are abundant in the same areas (Kellogg and Nygard,1951, pp. 51-83). It seems certain that the climate of the intervalduring which podzol formed at lyatayet was at least as warm as, and
24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121possibly slightly warmer than, the present climate. The shallow depthof profile development suggests that the warm period was brief com-pared with later intervals during which Subarctic Brown Forest soilswere formed.The sandy silt that covers the Denbigh culture layer marks thereturn of a cold climate that terminated the development of the podzol.Renewed intense frost action reactivated soil lobes higher on the valleyslopes. Silt washed from the advancing soil lobe fronts was depositedlower on the slopes and on the terrace surface, which at that time stillformed the valley bottom. The buried podzol, the Denbigh culturelayer, and part of the sandy silt were tightly folded as soil movementprogressed.The thin peat and underlying Subarctic Brown Forest soil exposedin pit Z-4 probably record a warm interval that interrupted the pre-vailingly cold period during which the sandy silt was deposited. Be-cause this buried profile appears in only one of the many good ex-posures of the full thickness of the sandy silt layer, it could be con-sidered to have formed in a local area of stable soil at a time whensolifluction was active in adjoining areas. Poorly drained silt, however,is particularly subject to frost heaving. Stirring due to frost heavingprevents the formation of a stratified soil profile in silt underlain ata depth of 2 or 3 feet by perennially frozen ground on Seward Penin-sula (Hopkins and Sigafoos, 195 1, pp. 61, 63). The sandy silt inwhich the buried Subarctic Brown Forest soil appears lies at the baseof a slope in a position where, if perennially frozen ground were pres-ent, it would maintain a high water content throughout most of theyear. It is extremely unlikely that any part of the sandy silt on theold valley floor of lyatayet Creek would have remained sufficientlystable to permit formation of a well-developed Subarctic Brown Forestsoil profile during a period when frost action was intense a few feetaway. Consequently, the buried soil exposed in pit Z-4 is believedto have formed during a warm interval, when soil movements ceasedthroughout lyatayet Valley. The profile probably developed in thesandy silt throughout the valley but was obscured by frost stirringduring the ensuing cold period except in areas where it was laterburied by an exceptionally large thickness of sandy silt.Deposition of the sandy silt was terminated finally by the periodof warmer climate that has extended with only minor fluctuations tothe present. Sea level rose and the shore line approached its presentposition. The alluvial fan assumed to have covered the floor of lya-tayet Cove was removed by waves and longshore currents. lyatayetCreek re-excavated its valley to bedrock, leaving remnants of the valley
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 25
fill as terraces. Well-developed soil profiles beneath Paleo-Eskimoground levels indicate that soil movements had ceased and that awarm climate had prevailed for a long time prior to Paleo-Eskimooccupation.A minor climatic fluctuation during the past 2,000 years is suggestedby the difference in preservation of organic material in Paleo-Eskimoand Neo-Eskimo middens. The Paleo-Eskimo midden is earthy, butfriable fragments of bone and antler and soft fibers of w^ood testifyto a former abundance of organic material now largely rotted away.The Neo-Eskimo midden, on the other hand, is peaty in character
;
organic material is well preserved and makes up much of the bulk ofthe deposit. The contrast in preservation of organic material indicatesa climate slightly warmer than at present during and after Paleo-Eskimo occupation and a return to slightly cooler conditions duringand after Neo-Eskimo occupation.LATE QUATERNARY CLIMATIC FLUCTUATIONS ON SEWARD PENINSULAAND IN THE FAIRBANKS DISTRICTThe climatic fluctuations recorded at lyatayet undoubtedly werecommon to other parts of Alaska. Glacial sediments mapped by Hop-k'ns in the Kigluaik Mountains of Seward Peninsula (field notes,1949, 1950) may have been deposited during one or more of the coldperiods recorded at lyatayet. Unfortunately, the glacial deposits ofSeward Peninsula at present can be neither dated nor correlated withcertainty with glacial deposits of known age elsewhere in Alaska.Sequences of silt and muck (silt rich in organic material) exposedin placer mine excavations record climatic fluctuations in unglaciatedareas of Seward Peninsula and central Alaska. Hopkins has studiedsilt and muck stratigraphy in some unglaciated valleys in central andnorthern Seward Peninsula. Muck layers present in many valleyshave been dated by carbon-14 analyses of wood specimens from threelocalities. Pewe (1952) has studied more thoroughly the stratigraphyof unconsolidated sediments in upland valleys in the Fairbanks dis-trict. Carbon-14 analyses of wood from several different muck layersprovide data by which several climatic fluctuations during the past20,000 years can be dated.Silt layers in upland valleys in the Fairbanks district consist ofloess deposited on valley slopes and ridge tops and carried into valleybottoms during glacial episodes in the nearby Alaska Range (Pewe,1950) . Muck layers consist of stream and mudflow deposits and resid-ual peat and forest beds that accumulated in stream valleys during
26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121periods when glaciers in the Alaska Range were less extensive andwhen the climate was at least as warm as at present. Pewe recognizesthree muck layers separated by two silt layers which record threewarm intervals and two glacial intervals.The oldest muck layer rests on auriferous gravel in valley bottomsand is at least 20,000 years old. A thick layer of slightly reworkedloess overlying the oldest muck is more than 16,000 years old. Over-lying the loess is a second muck layer from which samples of organicmaterial 12,000 to 16,000 years old have been collected. A secondloess layer less than 12,000 and more than 4,200 years old overliesthe second muck layer and is overlain in turn by a third muck depositfrom which samples of organic material 3,500 to 4,200 years old havebeen collected. The cold period represented by the second, upper loessdeposit probably is equivalent in age to the Mankato substage of glaci-ation in north-central United States considered by Flint and Deevey(1951, pp. 261-267) to have occurred. about 11,000 years ago.® Atleast part of the middle muck layer of the Fairbanks district was de-posited during the interstadial interval between the Gary and Mankatosubstages as recognized in north-central United States.Two muck units of different ages are tentatively distinguished byHopkins in unglaciated creek valleys of central and northern SewardPeninsula. Specimens collected at Claim No. 4 above Discovery onCofifee Creek, a tributary of the Kougarok River 60 miles north-north-east of Nome, at the head of Black Gulch, a tributary of the NoxapagaRiver 80 miles north-northeast of Nome, and in the valley of MudCreek near the southeast shore of Kotzebue Sound (fig. i) have beendated by carbon- 14 analysis by J. L. Kulp, Lament Geological Ob-servatory, Columbia University. Wood collected at Coffee Creek inthe older muck unit is 8,350 ±200 years old ; wood collected at BlackGulch in the older muck is 8,800± 200 years old ; and wood collectedat Mud Creek in the younger muck is 3,600 ±200 years old.The dated specimen from Coffee Creek was collected from frozenmuck free of large ice masses that overlies blue-gray silt containingice wedges arranged in a polygonal pattern. The blue-gray silt con-tains little or no organic material and is thought to consist either ofwind-blown silt or of silt derived from frost-rived bedrock farther
6 A different opinion is offered by Schultz, Lueninghoener, and Frankforter(i95i> PP- 34-36), who argue that the "Mankato climax may have been as re-cent as eight or even seven thousand years ago." Pollen sequences indicate,however, that Mankato ice had retreated in Minnesota and that spruce and firgrew in the glaciated area as early as 8,000 years ago (Flint and Deevey, 1951,pp. 272-273).
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 27
upslope, at a time when the cUmate was colder than at present. Theoverlying muck contains abundant willow wood, a few sticks of poplar(R. S. Sigafoos, botanist, U. S. Geological Survey, oral communica-tion), and a log chewed by a beaver whose incisors were no larger thanthe incisors of Castor canadensis, the modern beaver in Alaska (H. W.Setzer, zoologist, United States National Museum, oral communi-cation).The dated specimen from Black Gulch was collected in muck gener-ally similar to, but less well exposed than, the muck at Coffee Creek.Ice wedges are present in the muck at Black Gulch. Wood in themuck includes a few logs as large as 6 inches in diameter. The datedspecimen was identified by Sigafoos as poplar wood.The dated specimen from Mud Creek was collected near the inneredge of a rolling coastal plain at the southeast corner of KotzebueSound, 2 miles west of Candle, Alaska. The specimen consists ofwood from a buried beaver dam at the base of a deposit about 30feet thick of interbedded muck and fibrous peat. The organic sedi-ments are underlain by Quaternary gravel many tens of feet thick.Mining operations in 1949 exposed the surface of an ancient flood-plain upon which the beaver dam was built. Sedge turf, willow oralder thickets, and a birch stump 6 inches in diameter were seen rootedin place in the vicinity of the beaver dam. Sticks in the beaver damhad been carved by beavers whose incisors were comparable in sizewith those of Castor canadensis.Sparse willow shrubs grow today in the valleys of Coffee Creek,Black Gulch, and Mud Creek, but poplar and birch trees do not. Wil-lows are much less abundant on the modern surface of the valley ofMud Creek than on the exhumed flood plain on which the dated woodwas collected. The western limit of beaver during the past 50 yearshas been 100 miles to the east of Coffee Creek and Black Gulch andmore than 50 miles to the east of Mud Creek. The older muck, repre-sented by the muck at Coffee Creek and Black Gulch, and the youngermuck, represented by the muck at Mud Creek, must have accumulatedat times when the climate was warmer than at present, and when treesand large shrubs grew in areas that now support tundra vegetation,with only a few dwarf shrubs.The dated muck of Seward Peninsula and the youngest muck inthe Fairbanks area may have accumulated during a single, long, post-Mankato warm interval that extended from about 9,000 to about 3,500years ago. Alternately, the muck at Coffee Creek and Black Gulchmay have accumulated during a brief, early post-Mankato warm inter-val 8,000 to 9,000 years ago, and the muck at Mud Creek and the
28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121youngest muck in the Fairbanks district may have accumulated duringa separate, much later warm interval 3,500 to 4,500 years ago.CORRELATION BETWEEN LATE QUATERNARY EVENTS AT lYATAYET ANDIN NEARBY PARTS OF ALASKAForty feet of congeliturbate accumulated in lyatayet Valley duringa period when the climate was cold and frost action and solifluctionwere intense. Sea level stood much lower than at present, suggestingthat a major, world-wide, glacial event is represented. The lack of arecognized break in deposition suggests that accumulation was notinterrupted by a period of stability and thus that the congeliturbateaccumulated during a single, late Wisconsin substage of glaciation.Exposures are poor, however, and it is entirely possible that morethan one substage is represented by the congeliturbate.The podzol beneath the Denbigh culture layer formed during a briefwarm interval that interrupted the prevailing cold climate and markedthe end of deposition of the rocky congeliturbate. The distributionof the buried soil and culture layer suggests that the surface of the40-foot terrace still formed the floor of lyatayet Valley during thewarm interval. This lack of dissection of the terrace suggests that sealevel had not yet risen to its present position. The later cold intervalduring which the Denbigh culture layer was covered and folded ap-pears to have been briefer and less intense than the cold interval dur-ing which the rocky congeliturbate was deposited, because the layerof sandy silt that was added to the surface is thin and is different incharacter from the rocky congeliturbate. Accumulation of the sandysilt probably was interrupted briefly by another warm period duringwhich the peat and Subarctic Brown Forest soil exposed in pit Z-4were formed.Comparison of the Denbigh flint complex at lyatayet with flints inthe oldest culture layer recognized in the Trail Creek caves suggeststhat the Denbigh culture layer is more than 6,000 years old. Evidencethat lyatayet Valley had not yet been dissected, and thus that sea levelhad not yet risen to its present position, also suggests that the Denbighculture layer and the associated podzol are older than 6,000 years.Sea level is believed to have been a few feet higher than at present inmost unglaciated areas at the beginning of the post-]\Iankato thermalmaximum as recognized in the United States and Europe. Flintand Deevey (1951, p. 258) estimate that the thermal maximum in theUnited States began about 6,000 years ago.Two possible correlations must be considered for the deposits ofthe pre-Christian era in lyatayet Valley (table 2). The rocky congeli-
NO. II lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 29
1
>.
>.
"0uu3s(I.
30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121
turbate may have been deposited more than 16,000 years ago duringthe cold interval represented by the lower, pre-Mankato loess in theFairbanks area (correlation B, table 2). The buried podzol may haveformed and the Denbigh culture layer may have been deposited about12,000 years ago, during part of the warm interval recorded by themiddle muck in the Fairbanks area. The culture layer then would havebeen covered and folded during the Mankato substage at the sametime that part of the upper loess of the Fairbanks area was deposited.The buried peat layer and Subarctic Brown Forest soil layer exposedin pit Z-4 would have formed during a single, long, warm interval9,000 to 3,500 years ago represented by the dated mucks of SewardPeninsula and by the youngest muck in the Fairbanks area.It is alternately possible that part or all of the rocky congeliturbatewas deposited during the Mankato substage while the upper loess wasbeing deposited in the Fairbanks area (correlation A, table 2). Thepodzol and the Denbigh flint layer may have formed during a briefwarm interval about 8,000 to 9,000 years ago that is represented bythe dated muck at Coffee Creek and Black Gulch, Seward Peninsula.The culture layer then would have been covered and folded during aminor cold interval between 8,000 and 4,200 years ago, and the peatand Subarctic Brown Forest soil exposed in pit Z-4 would have formedduring a later warm interval 4,200 to 3,500 years ago, represented bythe dated muck at Mud Creek and the youngest muck of the Fairbanksarea.The writers favor correlation of the podzol and the Denbigh culturelayer with the older muck of Seward Peninsula, deposited 8,000 to9,000 years ago, because the thinness of the podzol at lyatayet sug-gests that the warm period during which it formed was very brief.The middle muck layer in the Fairbanks area, with which the podzoland culture layer instead may be correlative, represents a warm periodthat lasted at least 4,000 years. Moreover, the sandy silt which overliesthe Denbigh culture layer appears to represent a relatively minor coldinterval compared with the cold interval during which the congeli-turbate was deposited. The thickness of the upper loess of the Fair-banks area suggests that the Mankato substage was a major cold inter-val in Alaska. The archeological relationships are more readily ex-plained if the Denbigh flint complex at lyatayet is assumed to be about8,500 years old, but they do not rule out the alternate assumption thatthe culture layer is more than 12,000 years old. The possibility cannotbe eliminated that lyatayet was occupied by people of the Denbighflint complex during an interstadial interval earlier than the Mankatosubstage.
1
NO. II lYATAYET SITE, ALASKA
—
HOPKINS AND GIDDINGS 3IEvidence that the cHmate at lyatayet was warmer than at presentduring and after Paleo-Eskimo occupation from A.D. i to 500 ismatched by evidence for a warm or dry climate near this time inGreenland (Brooks, 1949, pp. 342-343), southwestern United States(op. cit., pp. 356-357), Ohio, and the Valley of Mexico (Sears, 1951).The cooler period during and since the Neo-Eskimo occupation fromabout A.D. iioo to 1600 corresponds broadly to the "little ice age"(Matthes, 1942), recognized in several places in Alaska (Lawrence,1950; Pewe, 195 1 ; Sharp, 195 1) and in nearly every glaciated moun-tain range in the Northern Hemisphere (Manley, 1951).ACKNOWLEDGMENTSArcheological investigations at Cape Denbigh were supported in1948 and 1949 by grants from the Arctic Institute of North Americain cooperation with the Office of Naval Research, U. S, Departmentof the Navy. The field work was continued in 1950 as a part of theBering Strait Expedition—a joint enterprise financed by the Wenner-Gren Foundation for Anthropological Research, the University ofAlaska, the University of Pennsylvania, and the Danish National Mu-seum. Giddings was aided in these studies by a contract between theOffice of Naval Research and the University of Pennsylvania (NR160-903). S. J. Newcomb, Wendell Oswalt, G. M. Henderson, andJ. W. Van Stone assisted in the excavations. Hopkins's visit wasmade possible by the cooperation of the U. S. Geological Survey.The writers gratefully acknowledge the assistance of H. N. Michael,who translated pertinent passages of Okladnikov (1950), and R. J.Hackman, who assisted in locating air photos of the Anaktuvuk Riversite. REFERENCESBartlett, H. H.195 1. Radiocarbon datability of peat, marl, caliche, and archaeological ma-terials. Science, n.s., vol. 114, pp. 55-56.Brooks, C. E. P.1949. Climate through the ages, 395 pp. New York.Bryan, Kirk.1946. Cryopedology—the study of frozen ground and intensive frost action,with suggestions on nomenclature. Amer. Journ. Sci., vol. 244,pp. 622-642.Flint, R. F., and Deevey, E. S., Jr.1951. Radiocarbon dating of late Pleistocene events. Amer. Journ. Sci.,vol. 249, pp. 257-300.
32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121GiDDINGS, J. L.1949. Early flint horizons on the north Bering Sea coast. Journ. Washing-ton Acad. Sci., vol. 39, pp. 85-90.1950. Early man on the Bering Sea coast. Ann. New York Acad. Sci.,ser. 2, vol. 13, pp. 18-21.195 1. The Denbigh flint complex. Amer. Antiquity, vol. 16, pp. 193-203.Grigoriev, N. F.1950. A geological account of the Kullata site. In Okladnikov, A. P., Lenaantiquities, vol. 3, pp. 163-177. (Translated by H. N. Michael.)Hopkins, D. M., and Sigafoos, R. S.1 95 1. Frost action and vegetation patterns on Seward Peninsula, Alaska.U. S. Geol. Surv. Bull. 974-C, pp. 51-100.Irving, William.195 1. Archaeology in the Brooks Range of Alaska. Amer. Antiquity, vol.17, P- 52.Johnson, Frederick (editor).1951. Radiocarbon dating. Soc. Amer. Archaeol., Mem. 8 (suppl. to Amer.Antiquity, vol. 17), 65 pp.Kellogg, C. E., and Nygard, I. J.195 1. Exploratory study of the principal soil groups of Alaska. U. S. Dept.Agr., Agr. Mon. 7, 138 pp.Larsen, Helge.1950. Archaeological investigations in southwestern Alaska. A.mer. An-tiquity, vol. 15, pp. 177-186.Larsen, Helge, and Rainey, F. G.1948. Ipiutak and the Arctic whale hunting culture. Amer. Mus. Nat. Hist.,Anthrop. Pap., vol. 42, 276 pp.Lawrence, D. B.1950. Glacier fluctuation for six centuries in southeastern Alaska and itsrelation to solar activity. Geogr. Rev., vol. 40, pp. 191-223.Manley, Gordon.1 95 1. Climatic fluctuations: A review. Geogr. Rev., vol. 41, pp. 656-660.Matthes, F. E.1942. Glaciers. In Physics of the earth, vol. 9, Hydrology, pp. 149-219.New York.Okladnikov, A. P.1950. Lena antiquities, vol. 3, 195 pp. U.S.S.R. Acad. Sci., Moscow-Leningrad.Pewe, T. L.1950. Origin of the upland silt in the Fairbanks area, Alaska (abstr.).Bull. Geol. Soc. Amer., vol. 61, p. 1493.1951. Recent history of Black Rapids Glacier, Alaska (abstr.). Bull. Geol.Soc. Amer., vol. 62, p. 1558.1952. Preliminary report of late Quaternary history of the Fairbanks area(abstr.). Bull. Geol. Soc. Amer., vol. 63, pp. 1289-1290.Schultz, C. B., Lueninghoener, G. C., and Frankforter, W. D.1951. A graphic resume of the Pleistocene of Nebraska (with notes on thefossil mammalian remains). Bull. Univ. Nebraska State Mus., vol.3, No. 6, 41 pp.
NO. 1 1 lYATAYET SITE, ALASKA—HOPKINS AND GIDDINGS 33Sears, P. B.1951. Climate and culture—new evidence. Science, n.s., vol. 114, pp. 46-47.Sharp, R. P.1951. Glacial history of Wolf Creek, St. Elias Range, Canada. Journ. Geol.,vol. 59, pp. 97-117-SiGAFOOS, R. S., and Hopkins, D. M.1952. Soil instability on slopes in regions of perennially frozen ground. JnFrost action in soils, Highway Res. Board Spec. Rep. 2, Nat. Res.Counc. Publ. 213, pp. 176-192.Smith, P. S., and Eakin, H. M.1911. A geological reconnaissance in southeastern Seward Peninsula andthe Norton Bay-Nulato region. U. S. Geol. Surv. Bull. 449, 146 pp.SOLECKI, R. S.195 1. Notes on two archaeological discoveries in northern Alaska, 1950.Amer. Antiquity, vol. 17, pp. 55-57.SoLECKi, R. S., and Hackman, R. J.1951, Additional data on the Denbigh flint complex in northern Alaska.Journ. Washington Acad. Sci., vol. 41, pp. 85-88.United States Weather Bureau.1943. Climatic atlas for Alaska. Rep. 444 (229 pp.) and Suppl. (72 pp.)of the Weather Information Branch, Headquarters, Army Air Force.

SMITHSONIAM MISCELLANEOUS COLLECTIONS VOL. 121, NO. 11, PL. 1
o
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121, NO. 11, PL. 2
^W
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121, NO. 11, PL. 3
^ "- '
Surface of Denbigh Culture Layer exposed on floor of Pit PE-4Kid.t^L-s near lianinicr arc anticlinal fuld.s in the Denhi.^h culture layer; "caves"are iiiiderside.s of folds from wliicli infolded sterile silt has heen removed.
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 121, NO. 11, PL. 4