SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ? NUMBER 89 Avian Paleontology at the Close of the 20th Century: Proceedings of the 4th International Meeting of the Society of Avian Paleontology and Evolution, Washington, D.C, 4-7 June 1996 Storrs L. Olson EDITOR Peter Wellnhofer, Cecile Mourer-Chauvire, David W. Steadman, and Larry D. Martin ASSOCIATE EDITORS Smithsonian Institution Press Washington, D.C. 1999 FRONTISPIECE.?Participants in the Fourth International Meeting of the Society of Avian Paleontology and Evolution, Washington, D.C, 5 June 1996. 1. Peter Wellnhofer (Munich, Germany) 2. Ingrid Wellnhofer (Munich, Germany) 3. Helen James (Washington, D.C.) 4. Storrs Olson (Washington, D.C.) 5. Cecile Mourer-Chauvire (Lyon, France) 6. Kenneth Campbell (Los Angeles, California) 7. Stefan Peters (Frankfurt, Germany) 8. Zhonghe Zhou (Beijing, China) 9. Lianhai Hou (Beijing, China) 10. John Ostrom (New Haven, Connecticut) 11. Virginia Naples (DeKalb, Illinois) 12. Lawrence Witmer (Athens, Ohio) 13. Peter Houde (Las Cruces, New Mexico) 14. Paul Buhler (Gschwend, Germany; deceased, 16 July 1996) 15. Joseph McKee (Palmerston North, New Zealand) 16. Sylvia Hope (San Francisco, California) 17. Jean Martin (Lawrence, Kansas) 18. Gregory Paul (Baltimore, Maryland) 19. Bartomeu Segui (Ciutat de Mallorca, Spain) 20. David Parris (Trenton, New Jersey) 21. Evgeny Kurochkin (Moscow, Russia) 22. Larry Martin (Lawrence, Kansas) 23. Herculano Alvarenga (Sao Paulo, Brazil) 24. Kenneth Parkes (Pittsburg, Pennsylvania) 25a. Ewan Fordyce (Dunedin, New Zealand) 25b. Hiroshige Matsuoka (Kyoto, Japan) 26. Charlie Magovern (Boulder, Colorado) 27. Luis Chiappe (Los Angeles, California) 28. Kraig Derstler (New Orleans, Louisiana) 29. Craig Jones (Dunedin, New Zealand) 30. Ralph Chapman (Washington, D.C.) 31. Diego Rassman-Gutman (Washington, D.C.) 32. Felix Dzerzhinsky (Moscow, Russia) 33. Per Ericson (Stockholm, Sweden) 34. Richard Zusi (Washington, D.C.) 35. Phil Millener (Chattanooga, Tennessee) 36. Mary Root (Albuquerque, New Mexico) 37. Marco Pavia (Turin, Italy) 38. Jorge Noriega (Diamante, Argentina) 39. Antoni Alcover (Ciutat de Mallorca, Spain) 40. Trevor Worthy (Nelson, New Zealand) 41. Walter Boles (Sydney, Australia) 42. Tom Stidham (Berkeley, California) 43. William Hilgartner (Baltimore, Maryland) 44. Richard Benson (St. Paul, Minnesota) 45. Andrzej Elzanowski (Wroclaw, Poland) 46. Darrel Tanke (Drumheller, Canada) 47. Pamela Rasmussen (Washington, D.C.) 48. Tommy Tyrberg (Kimstad, Sweden) 49. Pippa Haarhoff (Capetown, South Africa) 50. Gerald Mayr (Berlin, Germany) 51. Steven Emslie (Wilmington, North Carolina) 52. John Stewart (Cambridge, United Kingdom) 53. Jenna Boyle (Gunnison, Colorado) 54. David Steadman (Gainesville, Florida) 55. Joanne Cooper (London, United Kingdom) 56. John Becker (Edmonton, Canada) 57. Claudia Tambussi (La Plata, Argentina) 58. Alexandr Karhu (Moscow, Russia) Registrants not shown: Michael Braun (Washington, D.C.) Rachel Burke (New Brunswick, New Jersey) Sankar Chatterjee (Lubbock, Texas) Adele Conover (Huntington, Maryland) Alan Cooper (Washington, D.C.) Catherine Forster (Stony Brook, New York) Steve Gatsey (Providence, Rhode Island) G.E. Goslow, Jr. (Providence, Rhode Island) Michael Gottfried (Solomons, Maryland) Kevin Middleton (Providence, Rhode Island) Eleni Paxinos (Washington, D.C.) Samuel O. Poore (Providence, Rhode Island) Paul Sereno (Chicago, Illinois) Wesley Sutton (New York, New York) Patricia Wainwright (New Brunswick, New Jersey) Luvia Zusi (Washington, D.C). ABSTRACT Olson, Storrs L., editor. Avian Paleontology at the Close of the 20th Century: Proceedings of the 4th International Meeting of the Society of Avian Paleontology and Evolution, Washington, D.C, 4-7 June 1996. Smithsonian Contributions to Paleobiology, number 89, 344 pages, fron tispiece, 169 figures, 49 tables, 1999.?The 32 papers collected herein reflect the great diver sity and interest that the study of fossil birds has generated in recent years. The first seven papers (Mourer-Chauvire et al., Worthy and Jouventin, Segui and Alcover, Steadman and Hil- gartner, Millener, Worthy, Pavia) relate to late Quaternary birds from islands, where human intervention in the last few thousand years has caused many heretofore unrecorded extinctions. Three papers on Quaternary avifaunas of continental Europe deal with distributional changes and cultural use of birds by humans in Siberia (Potapova and Panteleyev), the utility of patterns of seabird distribution in determining former marine climatic conditions (Tyrberg), and tempo ral changes in morphology of ptarmigans (Lagopus) through the late Pleistocene (Stewart). Three papers deal with late Cenozoic raptors (Campbell et al., Tambussi and Noriega, Emslie and Czaplewski). New genera from Paleogene deposits are described by Boles and Ivison, Karhu, and Peters. Five papers deal with ancient waterfowl. Alvarenga describes the first fossil screamer (Anhimidae) from the Oligocene of Brazil. Olson provides the first fossil records of the Anseranatidae, with the description of a new species from the early Eocene of England, which is referred to Anatalavis from the Paleocene/Cretaceous of New Jersey. Ericson provides the means to distiguish Eocene fossils of the duck-like Presbyornis from the flamingo-like Jun- citarsus and gives new records of the latter. Benson shows that the Paleocene Presbyornis isoni once ranged from Maryland to North Dakota, and he gives records of other Paleocene birds from North Dakota. Hope names a new, larger species of Graculavus, extending the range of the genus from New Jersey to the Cretaceous of Wyoming. The early history and evolution of birds receives great attention. Dzerzhinsky expands upon the significance of cranial morphology in paleognathous birds. Kurochkin relates the early Cre taceous genus Ambiortus to the Chinese Otogornis, which are supposed to be on a line with modern birds, as opposed to the Enantiornithes. Bochenski uses paleogeography to suggest that the Enantiornithes must antedate Archaeopteryx. Zhou and Martin show that the manus of Archaeopteryx is more bird-like than previously realized. Martin and Stewart use bird teeth to argue against dinosaurian origins for Aves, whereas Elzanowski diverges on various aspects of dinosaurian cranial morphology and that of early birds that may have evolutionary signifi cance. Witmer, Chiappe, and Goslow present summaries of three sessions of a roundtable dis cussion on avian origins, early evolution of birds, and the origins of flight, which was held on June 7, the last day of the meeting, and which covered much controversial territory. 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: The trilobite Phaecops rana Green. Library of Congress Cataloging-in-Publication Data Society of Avian Paleontology and Evolution. Symposium (4th : 1996 : Washington, D.C.) Avian paleontology at the close of the 20th century : proceedings of the 4th International Meeting of the Society of Avian Paleontology and Evolution, Washington, D.C, 4-7 June 1996 / Storrs L. Olson, editor ; Peter Wellnhofer ... [et al.], associate editors. p. cm. ? (Smithsonian contributions to paleobiology ; no. 89) Includes bibliographical references. 1. Birds, Fossil Congresses. I. Olson, Storrs L. II. Wellnhofer, Peter. III. Title. IV. Series. QE871.S63 1996 586?dc21 99-33292 CIP ? 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 Preface vii QUATERNARY INSULAR BIRDS The Avifauna of Reunion Island (Mascarene Islands) at the Time of the Arrival of the First Europeans, by Cecile Mourer-Chauvire, Roger Bour, Sonia Ribes, and Francois Moutou 1 The Fossil Avifauna of Amsterdam Island, Indian Ocean, by Trevor H. Worthy and Pierre Jouventin 39 Comparison of Paleoecological Patterns in Insular Bird Faunas: A Case Study from the Western Mediterranean and Hawaii, by Bartomeu Segui and Josep Antoni Alcover 67 A New Species of Extinct Barn Owl (Aves: Tyto) from Barbuda, Lesser Antilles, by David W. Steadman and William B. Hilgartner 75 The History of the Chatham Islands' Bird Fauna of the Last 7000 Years?A Chronicle of Change and Extinction, by Philip R. Millener 85 The Role of Climate Change Versus Human Impacts?Avian Extinction on South Island, New Zealand, by Trevor H. Worthy Ill The Middle Pleistocene Avifauna of Spinagallo Cave (Sicily, Italy): Preliminary Re port, by Marco Pavia 125 QUATERNARY AVIFAUNAL STUDIES IN CONTINENTAL EUROPE Birds in the Economy and Culture of Early Iron Age Inhabitants of Ust' Poluisk, Lower Ob' River, Northwestern Siberia, by Olga R. Potapova and Andrei V. Panteleyev 129 Seabirds and Late Pleistocene Marine Environments in the Northeast Atlantic and the Mediterranean, by Tommy Tyrberg 139 Intraspecific Variation in Modern and Quaternary European Lagopus, by John R. Stewart 159 LARGE RAPTORS FROM THE LATE CENOZOIC OF THE NEW WORLD A New Genus for the Incredible Teratorn (Aves: Teratornithidae), by Kenneth E. Campbell, Jr., Eric Scott, and Kathleen B. Springer 169 The Fossil Record of Condors (Ciconiiformes: Vulturidae) in Argentina, by Claudia P. Tambussi and Jorge I. Noriega 177 Two New Fossil Eagles from the Late Pliocene (Late Blancan) of Florida and Ari zona and Their Biogeographic Implications, by Steven D. Emslie and Nicholas J. Czaplewski 185 THREE NEW GENERA OF PALEOGENE BIRDS A New Genus of Dwarf Megapode (Galliformes: Megapodiidae) from the Late Oli gocene of Central Australia, by Walter E. Boles and Tessa J. Ivison 199 A New Genus and Species of the Family Jungornithidae (Apodiformes) from the Late Eocene of the Northern Caucasus, with Comments on the Ancestry of Humming birds, by Alexandr A. Karhu 207 Selmes absurdipes, New Genus, New Species, a Sandcoleiform Bird from the Oil Shale of Messel (Germany, Middle Eocene), by D. Stefan Peters 217 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY EARLY WATERFOWL (ANSERIFORMES) AND RELATIVES A Fossil Screamer (Anseriformes: Anhimidae) from the Middle Tertiary of South eastern Brazil, by Herculano M.F. Alvarenga 223 The Anseriform Relationships of Anatalavis Olson and Parris (Anseranatidae), with a New Species from the Lower Eocene London Clay, by Storrs L. Olson 231 New Material of Juncitarsus (Phoenicopteriformes), with a Guide for Differentiating that Genus from the Presbyornithidae (Anseriformes), by Per G.P. Ericson 245 Presbyornis isoni and Other Late Paleocene Birds from North Dakota, by Richard D. Benson 253 A New Species of Graculavus from the Cretaceous of Wyoming (Aves: Neornithes), by Sylvia Hope 261 MESOZOIC BIRDS AND EARLY AVIAN EVOLUTION Implications of the Cranial Morphology of Paleognaths for Avian Evolution, by Felix Y. Dzerzhinsky 267 The Relationships of the Early Cretaceous Ambiortus and Otogornis (Aves: Ambior- tiformes), by Evgeny N. Kurochkin 275 Enantiornithes: Earlier Birds than Archaeopteryx?, by Zygmunt Bocheriski 285 Feathered Dinosaur or Bird? A New Look at the Hand of Archaeopteryx, by Zhonghe Zhou and Larry D. Martin 289 Implantation and Replacement of Bird Teeth, by Larry D. Martin and J.D. Stewart 295 Humeral Rotation and Wrist Supination: Important Functional Complex for the Evo lution of Powered Flight in Birds?, by John H. Ostrom, Samuel O. Poore, and G.E. Goslow, Jr 301 A Comparison of the Jaw Skeleton in Theropods and Birds, with a Description of the Palate in the Oviraptoridae, by Andrzej Elzanowski 311 Early Evolution of Birds: Roundtable Discussion 325 New Aspects of Avian Origins: Roundtable Report, by Lawrence M. Witmer 327 Early Avian Evolution: Roundtable Report, by Luis M. Chiappe 335 The Origin of Bird Flight: Roundtable Report, by G.E. Goslow, Jr 341 Preface The 4th International Meeting of the Society of Avian Paleontology and Evolution (SAPE) was held at the Smithsonian Institution, Washington, D.C, 4-7 June 1996, as an official part of the celebration of the 150th anniversary of the Smithsonian. Sessions for papers were held at the S. Dillon Ripley Center the first two days of the meeting. A very successful workshop organized by Sylvia Hope was held during the afternoon of the second day at the National Museum of Natural History. Participants examined and compared fossils of latest Cretaceous and early Tertiary birds, resulting in numerous valu able insights and revelations. A field trip on June 6th to the Miocene exposures of Calvert Cliffs along Chesapeake Bay was followed by a visit to the Calvert Marine Museum (CMM) at Solomon's, Mary land, and culminated with an outdoor crab feast. All of this took place under the most ideal imaginable conditions, thanks to fine weather and the careful planning of our hosts for the day at CMM. The final day was devoted to a symposium and roundtable on Mesozoic birds and their origins, organized by Peter Wellnhofer. The roundtable brought out animated discussion of the more intractable issues that invest this subject today, but these discussions were con ducted totally without rancor or animosity and in a spirit of genuine collegiality. The SAPE meeting's tradition of international composition was fully upheld by the fourth meeting, in which there were registrants from at least 14 countries and 18 states of the United States. To maintain international participation, the Society has successfully been able to alter the venue of its quadrennial meetings among continents. At the business meeting in Washington, D.C, an offer was extended by the Institute of Vertebrate Paleon tology of the Academica Sinica to hold the fifth SAPE meeting, in the year 2000, in Beijing, China. After some thoughtful discussion, the invitation was accepted. No matter whether the 21st century begins in the year 2000 or 2001, the Washington SAPE meeting was the last to be held in a year beginning with "19." Thus, the title of this volume suggested itself. It is worth reflecting on the fact that during the last quarter of the 20th century there was probably as much or more learned about the fossil history of birds than there was throughout the rest of history. The papers that are collected herein reflect the continued vigor and diversity of this line of investigation around the world. Now the legacy of SAPE is passed to a new continent in a new century. May its light be undimin ished. ACKNOWLEDGMENTS.?The local committee on arrangements for the Washington meeting consisted of Helen James, Storrs Olson, Michael Gottfried, Pamela Rasmussen, and Ralph Chapman. That the meeting took place at all is due entirely to the persistence, optimism, and insistence of Helen James, who, through an oppressive winter of govern ment shutdowns and dreadful weather, and with little prospect of finding sufficient fund ing, kept communications open and plans progressing when others only despaired. Direct funding for the meeting came from the 150th Anniversary Program Committee and the Office of the Director, National Museum of Natural History, Smithsonian Institu tion. Travel subsidy for some of the participants was provided by the Office of Fellow ships and Grants, Smithsonian Institution; the International Science Foundation; and the University of Kansas. Arrangements for the field trip and the crab feast were made by Michael Gottfried of the Calvert Marine Museum, cosponsor of the meeting. In order to spread some of the burden of the editorial process for the present volume, I divided the majority of the submitted papers among four associate editors, who were re sponsible for reading and commenting on manuscripts and soliciting additional reviews. Of necessity, much of the task of refereeing fell largely to members of SAPE. The follow- vn vm SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ing is a list of referees who commented on one or more of the papers, and I am very grate ful to all of them for their help: Antoni Alcover, S. Christopher Bennett, Zygmunt Bochen- ski, Alan Brush, Eric Buffetaut, Kenneth Campbell, Luis Chiappe, Charles Collins, Miguel Elorza, Andrzej Elzanowski, Steven Emslie, Per Ericson, Alan Feduccia, G.E. Goslow, Peter Houde, Helen James, Denes Janossy, Larry Martin, Gary Morgan, Cecile Mourer-Chauvire, Storrs Olson, John Ostrom, Kevin Padian, David Parris, Steven Parry, Stefan Peters, Gregory Pregill, J.H. Reichholf, Dale Serjeantson, David Steadman, Burkhard Stephan, Tommy Tyrberg, David Unwin, Kenneth Warheit, Paul Weldon, Peter Wellnhofer, Lawrence Witmer, Trevor Worthy, Zhonghe Zhou, and Richard Zusi. For assistance with transmission of manuscripts by electronic mail I am grateful to James Dean, Craig Ludwig, Chris Milensky, and Brian Schmidt. Mary Parrish repeatedly assisted with problems concerning illustrations and provided the outline key for the fron tispiece. I also thank Sharon Jones, who cheerfully wielded her computer to render more readable several of the manuscripts that were more heavily scribbled by the editor. Storrs L. Olson The Avifauna of Reunion Island (Mascarene Islands) at the Time of the Arrival of the First Europeans Cecile Mourer-Chauvire, Roger Bour, Sonia Ribes, and Francois Moutou ABSTRACT The excavations of five fossil bird localities on Reunion Island have yielded the remains of (1) five species still present on Reunion, which are mainly marine; (2) one extant species no longer on Reunion, the Greater Flamingo (Phoenicopterus ruber Linnaeus); and (3) 11 extinct species, namely, a night heron (Nyc- ticorax duboisi (Rothschild)), an ibis {Threskiornis solitarius (Selys-Longchamps)), a sheldgoose (Alopochen kervazoi (Cowles)), a teal {Anas theodori Newton and Gadow), a falcon (Falco duboisi Cowles), a rail (Dryolimnas augusti, new species), a coot {Fulica newtonii Milne-Edwards), a pigeon (Nesoenas duboisi Rothschild), a parrot {Mascarinus mascarinus Linnaeus), an owl {Mascarenotus grucheti Mourer-Chauvire et al.), and a starling {Fregilupus varius (Boddaert)). Representatives of extinct endemic Mascarene taxa, such as the Raphidae, Aphanapteryx, Erythromachus, and large parrots of the genera Lophopsittacus and Necropsittacus, are so far unknown from Reunion. Except for Fulica newtonii, which probably colonized Reunion from Mauri tius, and Dryolimnas augusti, all the other forms appear to have had normal or nearly normal flying ability. It is possible that Reunion was colonized by the ancestors of the same forms that colonized Mauritius and Rodrigues, but these forms may have been exterminated during the very explosive events of the last phase of volcanic activity of Piton des Neiges, which took place between 300,000 and 200,000 years ago. Reunion would then have been colonized again by forms that perhaps have not had enough time to lose their ability to fly. After the arrival of the first Europe ans, all the larger endemic land birds became extinct, with the exception of Circus maillardi Verreaux. Although most of them were not morphologically flightless, and although the topography Cecile Mourer-Chauvire, Centre de Paleontologie Stratigraphique et Paleoecologie de I'Universite Claude Bernard-Lyon I, ERS 2042 du CNRS, 27-43 boulevard du II Novembre 1918, 69622 Villeurbanne Cedex, France. Roger Bour, Museum National d'Histoire Naturelle, Laboratoire des Reptiles et des Amphibiens, 25 rue Cuvier, 75005 Paris, France. Sonia Ribes, Museum d'Histoire Naturelle, 1 rue Poivre, 97400 Saint-Denis, La Reunion, France. Franqois Moutou, Minist'ere de I'Agriculture, CNEVA, Laboratoire Central de Recher ches Veterinaires, 22 rue Pierre Curie, BP n? 67, 94703 Maisons-Al- fort Cedex, France. of the island was very different, these birds became extinct on Reunion as rapidly as on the other Mascarene islands. Introduction The Mascarene Islands were "discovered" by Portuguese navigators as early as 1500 (North-Coombes, 1979). They were shown on more ancient maps of Arabian navigators but were uninhabited prior to the sixteenth century. Thus, Europeans were the first settlers to observe the very unusual fauna of these islands, one that was almost entirely exterminated within just two centuries. Numerous historical accounts of Reunion have been gath ered into two publications by Albert Lougnon (1970, 1992), with the oldest account being that of Samuel Castleton, who landed on the island in 1613. The parts concerning the birds are given in extenso in the books of Barre and Barau (1982) and Barre et al. (1996). Particular points of these accounts have been discussed by Cheke (1987). The most complete and de tailed report is that written by Dubois (1674), also known as Sieur D.B., who spent 16 months on the island in 1671 and 1672 (English translation by Oliver, 1897). These accounts are invaluable because they make it possible to follow the demise of the endemic fauna of Reunion from the beginning of the sev enteenth century to the disappearance of the two most recently extinct species, Mascarinus mascarinus Linnaeus, in 1834, and Fregilupus varius (Boddaert), between 1838 and 1858. Fossil bird remains were found very early on Rodrigues (in 1786) and on Mauritius (in 1865), making it possible to have good knowledge of their vanished faunas. Until recently, how ever, the bird fauna of Reunion was known only from the ac counts of early explorers. The first fossil remains were un earthed in 1974 by B. Kervazo during archaeological excavations in the Grotte des Premiers Francais. Subsequently, four other fossil sites containing bird remains also were discov ered. The fossil birds of the present study came from these five SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY localities, all of them situated in the territory of Saint-Paul, on the northwest portion of the island (Figure 1). The caves are not karstic cavities but are developed inside the lava flows, or at the junction between successive lava flows. ACKNOWLEDGMENTS.?We thank the Conseil general de la Reunion, the Conseil regional de la Reunion, and the Societe Reunionaise des Amis du Museum, whose generous coopera tion made it possible to carry out excavations in the Marais de l'Ermitage, and we thank all the friends who helped us with these excavations and who provided material from other locali ties, in particular, Jean Jacques Argoud, Pierre Brial, Jean- Pierre and Colette Esmiol, Jean-Michel Probst, and the mem bers of the Societe d'Etudes scientifiques des Cavernes de la Reunion. The work at the Smithsonian Institution was made possible by a Short Term Visitor Grant from the Office of Fel lowships and Grants. For the loan of comparative modern and fossil material we thank Daniel Goujet and Christine Lefevre (Museum National d'Histoire Naturelle, Paris), Storrs L. Ol son, (National Museum of Natural History, Smithsonian Insti tution), Robert Prys-Jones (The Natural History Museum, Lon don), and Janet Hinshaw (University of Michigan). We thank Alain Dubois for his advice concerning nomenclatural prob lems and our two referees, Storrs Olson and David Steadman, for their very constructive criticism. Steve Goodman provided unpublished measurements of Madagascan anatids. For the x- ray pictures of the two specimens of Mascarinus mascarinus, we thank Jean Dorst, Paris, and Herbert Schifter, Vienna. The photographs are by Noel Podevigne, and the drawings are by Arlette Armand (Centre des Sciences de la Terre, Universite Claude Bernard-Lyon 1). ABBREVIATIONS AND ACRONYMS.?The following museum acronyms are used: AMNH, American Museum of Natural History, New York; BMNH, The Natural History Museum, London (formerly, British Museum (Natural History)); FSL, Faculte des Sciences de Lyon; LAC, Laboratoire d'Anatomie Comparee du Museum National d'Histoire Naturelle, Paris; MNHN, Museum National d'Histoire Naturelle, Paris; MHNR, Museum d'Histoire Naturelle, Saint-Denis, La Reunion; UCB, Universite Claude Bernard-Lyon 1; USNM, collections of the former United States National Museum, now in the National Museum of Natural History, Washington, D.C. In the listing of material, the following abbreviations are used: d., distal; j., juvenile; 1., left; p., proximal; r., right; s., shaft. Fossil Localities Grotte des Premiers Francais (Grande Caverne).?This is a very large cavity, situated 1.5 km from the center of Saint-Paul. Details concerning the stratigraphy and the location of excava tions have been given by Kervazo (1979) and Bour (1979, 1980a). The bird material comes from layers 4 and 5 of pit number 2, in the most northeastern cave, which now includes a statue of the Virgin Mary. The top of layer 4 was situated about 80 cm below the floor of the cave at the time of the excavation, and layers 4 and 5 together were 25 cm thick. These layers did not include any trace of human occupation or any remains of introduced mammals; therefore, the vertebrate material can be considered as having been deposited prior to the occupation of the island by humans. In this cave, as well as in Grotte "au sable," most of the bird remains belong to the two species of shearwater, Puffinus paci- ficus Gmelin and P. Iherminieri Lesson. The remains come from all growth stages, from very young individuals, the bones of which are simple sticks without articulations, to fully grown adults. With the exception of a single bone of Fregilupus vari us, no passerine remains were found. The sediments were not sieved (Kervazo, pers. comm., 1996), but a large number of very small bones, such as pedal phalanges of shearwaters, were collected, and, if there had been passerine remains, they would have been collected. The shearwater remains probably come from individuals that were nesting in burrows in the vicinity of the cave; the sediments of their nesting site might have been washed into the cave during a cyclonic episode. The other ver tebrate remains probably come from animals that took refuge in the cave during such an episode (Bour, 1979). Bird material from this site was first described by Cowles (1987, 1994). All this material is in the MNHN and has the pre fix LAC. Grotte de l'Autel.?This cave was discovered in 1980 by R. Bour and F. Moutou (Bour, 1980a). It is situated along the Nl road, 2 km south of Saint-Gilles, to the south of and a little higher up than a shrine dedicated to Saint Expedit. It is a small cavity, about 2.50 m long, 1.20 m wide, and 1 m high at its highest part. The sediment, completely removed in 1980, was not sieved. Grotte "au sable."?This cave also was discovered and exca vated in 1980 by R. Bour and F. Moutou (Bour, 1980a); it was excavated again in 1987 by F. Moutou, R. Mourer, and C Mourer-Chauvire. It is situated close beside the Nl road, to the north and at the same level as the Saint Expedit shrine. Its di mensions are about 2 m2 and 1.50 m high at its highest part. Its sediment also was completely excavated in 1987 and was sieved with 1.5 mm mesh screens. Very tiny lizard bones were found, which are smaller than those of recent endemic passe rines, as well as isolated mouse teeth. It is therefore likely that the absence of passerines actually reflects their absence in the deposits and is not an artifact of collecting. The vertebrate re mains from Grotte de l'Autel and Grotte "au sable" are proba bly contemporaneous with the occupation of the caves by the first settlers, from the middle of the seventeenth century, be cause they include some bones of introduced mammals. All the material from Grotte de l'Autel and Grotte "au sable" is in the UCB and has the prefix FSL. Marais de l'Ermitage.?This swamp is situated a little fur ther to the south of the above-mentioned sites, along the Nl road, between Saint-Gilles and La Saline-les-Bains, at the lo cality of l'Ermitage-les-Bains. The first fossil bones were dis- NUMBER 89 55?20 EST 55?40 EST LePort St- Paul Grotte des premiers frangais St-G Mies' Grotte "au sable" Grotte de lAutel Marais de l'Ermitage Caverne de la Tortue 21?SUD St-Pierre FIGURE 1.?Map of Reunion Island showing the five localities in the northwest where fossil birds have been recovered. Contour intervals are in meters. 21? 20 covered in 1989-1990, during earthworks for the construction of the Jardin d'Eden. Excavations have been carried out in this same swamp, to the south of the Jardin d'Eden, under the responsibility of the Museum d'Histoire Naturelle de La Reunion, since 1992, and are still conducted every year. The area excavated so far is about 80 m2 The filling of the swamp consists of about 80 cm of organic soil with very few fossils. This overlies a layer 30 to 40 cm deep made up of countless bones and shell fragments of the large, extinct land-tortoise Cylindraspis borbonica Bour, mixed with volcanic rocks and blocks of coral. Among this extraordinary accumulation of tortoise remains occur very rare bird and bat bones. The fos- siliferous layer rests on marine sediments made up of coral line sands, fragments of corals, and marine molluscs. A very few remains of domestic mammals, introduced by humans, have been found associated with the extinct tortoises, al though most remains of domestic mammals are found in the upper sediments. The sediments of the marsh were systematically sieved using 2.5 mm and 1.5 mm mesh screens, and very small bones of ju venile Cylindraspis borbonica as well as tiny bones of bats were found. Here also, the absence of passerines probably re flects their absence in the sediments. Radiocarbon dates have been obtained for material from the Marais de l'Ermitage. The first one, on several bones of Cylin draspis borbonica, yielded an age of 915? 120 BP (Lyon 5551); interval in real years after calibration: 883-1273 AD (confidence interval 90%). Two other dates on single bones by the accelerator (AMS) method, gave the following results: Bone of Cylindraspis borbonica, from the base of the filling: 1755 ?40 BP (OxA-5994 (Lyon 201)). Interval in real years af ter calibration: 186-391 AD (confidence interval 95%). SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Bone of Sus scrofa Linnaeus, introduced by humans, from the fossiliferous layer, associated with bones of Cylindraspis borbonica: 365?35 BP (OxA-5993 (Lyon 200)). Interval in real years after calibration: 1458-1633 AD (confidence inter val 95%). This date agrees very well with historical accounts indicating that pigs were released on the island in or before 1629(Lougnon, 1970, 1992). The three dates obtained for the Marais de 1'Ermitage indi cate that the remains of tortoises, birds, and bats accumulated over a period of at least 1400 years, from about 300 AD to about 1700 AD. All the bird material from Marais de l'Ermitage is in MHNR and has the prefix MHN-RUN-FE-O. Caverne de la Tortue.?This cave is situated in the territory of Saint-Paul, near Chemin Bruniquel, between the localities of La Saline, the upper village, and La Saline-les-Bains, on the ocean side, approximately halfway between Ravine de la Sa line, at the northwest, and Ravine Tabac, at the southeast, at an altitude of about 250 m above sea level. It is a complex cavern, including several galleries, and extends over a maximum length of 322 m (Brial, 1996). The entrance is through the roof of a tunnel, the floor being situated 5 m below. The bird re mains were collected on 23 March 1996 by Jean-Michel Probst and Pierre Brial, in Salle des Tortues Geantes. They were situ ated on the surface of the floor, along the walls, and associated with remains of the extinct tortoise (C borbonica), introduced mammals {Capra hircus Linnaeus, Lepus nigricollis Cuvier, Rattus sp., Tenrec ecaudatus (Schreber)), and molluscs (Brial, 1996). The bones are poorly preserved and appear recent. The tortoise disappeared from this part of the island between 1717 and 1732 (Bour, 1980a), and Lepus nigricollis was introduced between 1868 and 1887 (Cheke, 1987), so the age of the animal remains is unknown. The material is in MHNR and has the pre fix MHN-RUN-CT. A list of species identified in the different localities is given in Table 1. Systematic Paleontology Family PROCELLARHDAE Genus Puffinus Brisson Puflinus pacificus (Gmelin, 1789) Wedge-tailed Shearwater MATERIAL.?Grotte des Premiers Francais: Very numer ous remains from all parts of the skeleton, corresponding at least to 5 individuals (4 adults, 1 juvenile) in layer 4 and to 43 individuals (21 adults, 22 juveniles) in layer 5. Grotte "au sable": Very numerous remains, corresponding at least to 14 individuals (5 adults, 9 juveniles): 330549-330571, 330589-330590, 330592-330595, 330598, 330600- 330627, 330639-330641, 330644, 330647-330650, 330653-330660, 330665-330667, 330669, 330671-330683, 330700-330711, 330727-330729. Grotte de l'Autel and Grotte "au sable": Very numerous re mains, corresponding at least to 7 individuals (5 adults, 2 juve niles): 330763-330800. Marais de l'Ermitage: 1. p. ulna, 1836; 1. d. ulna, very abraded, 1919; r. d. tarsometatarsus, 1835; 2 1. tarsometatarsi, 1833, 1834; pedal phalanx, 1837. Caverne de la Tortue: r. p. humerus, 28; fragment of r. car- pometacarpus, 29; 1. tarsometatarsus, 30. TABLE 1.?List of the bird species found in the different fossil localities. The numbers correspond to the mini mum numbers of individuals. Species Puffinus pacificus Puffinus Iherminieri Phaethon lepturus Nycticorax duboisi Threskiornis solitarius Phoenicopterus ruber Alopochen (A/.) kervazoi Anas theodori cf. Ay thy a sp. Falco duboisi Dryolimnas augusti, new species Fulica newtonii Numenius phaeopus Nesoenas duboisi Streptopelia piclurata Mascarinus mascarinus Mascarenotus grucheti Fregilupus varius Total Grotte des Premiers Francais 48 28 1 1 1 - 2 -- 2 -1 1 1 - 1 I 1 89 Grotte de l'Autel T 4* 1 1 2 - 2 -- 1 1 - 1 - 1 2 - 23 Grotte "au sable" 14 11 - -- 1 -- - -- - - 1 2 1 - 30 Marais de l'Ermitage 2 -- 2 3 8 3 1 1 - - 3 - _ - _ 1 - 24 Caveme de la Tortue 1 -- -_ - -- - - 2 _ _ _ - _ - - 3 'Combined collections from Grotte de l'Autel and Grotte "au Sable.' NUMBER 89 Puffinus Iherminieri Lesson, 1840 Audubon's Shearwater MATERIAL.?Grotte des Premiers Francais: Very numer ous remains from all parts of the skeleton, corresponding at least to 1 adult in layer 4 and to 27 individuals (14 adults, 13 juveniles) in layer 5. Grotte "au sable": Very numerous remains, corresponding at least to 11 individuals (2 adults, 9 juveniles): 330572-330588, 330591, 330596-330597, 330599, 330628-330638, 330642-330643, 330645-330646, 330651-330652, 330661-330664, 330668, 330670, 330684-330699, 330712-330726. Grotte de l'Autel and Grotte "au sable": Very numerous re mains, corresponding at least to 4 individuals (2 adults, 2 juve niles): 330745-330762. REMARKS.?The high proportion of juveniles indicates that both species of Puffinus were probably nesting in these cavities or their surroundings. Puffinus pacificus now nests at Reunion only on a small islet, accessible only with difficulty for humans or introduced mammals, whereas P. Iherminieri nests in nu merous parts of the island (Jouanin, 1987). In the three caves, P. pacificus is more numerous than P. Iherminieri, but at present P. Iherminieri is much more common on the island. Two other species of Procellariidae, Pseudobulweria aterri- ma (Bonaparte) and Pterodroma baraui (Jouanin), now nest on Reunion, at high altitudes. Subfossil remains of P. baraui have been found in the Caverne a Cotte, at about 1800 m elevation (Jouanin and Gill, 1967). In our fossil localities, which are situ ated at low elevation, we did not find remains that can be at tributed to either of these species. The limited fossil evidence therefore suggests that they may always have nested at higher elevations than Puffinus. Family PHAETHONTIDAE Genus Phaethon Linnaeus Phaethon lepturus Daudin, 1802 White-tailed Tropicbird FIGURE 4/ MATERIAL.?Grotte des Premiers Francais: Part of skull, 1993-51. Grotte de l'Autel: 1. coracoid, 330515 (Figure At). REMARKS.?The internal length of the coracoid is 39.0 mm, which falls within the range of variation of modern P. lepturus (36.2^10.1, ?=10) and is clearly smaller than in the other two species, P. rubricauda Boddaert, the Red-tailed Tropicbird (47.0-53.9, /i=24), and P. aethereus Linnaeus, the Red-billed Tropicbird (47.0-50.4, ?=7). Phaethon lepturus still nests on Reunion. Family ARDEIDAE Genus Nycticorax Forster Megaphoyx Hachisuka 1937b: 148 [type by original designation, Ardea mega- cephala Milne-Edwards, 1874]. Each of the three Mascarene islands sustained an extinct spe cies of Nycticorax. The first of these species was described from Rodrigues by Milne-Edwards (1874) under the name Ardea megacephala. Giinther and Newton (1879) showed that the Rodrigues heron belonged in the genus Nycticorax and that, although its size was not very different from a large modern N. nycticorax (Linnaeus), its wing bones were proportionally shorter and its femur, tarsometatarsus, and pedal phalanges proportionally longer. Actually, when compared with typical N. nycticorax nycticorax, the wings are not very short, but the femora, tibiotarsi, and tarsometatarsi are wider, longer, and more robust (Cowles, 1987). Newton and Gadow (1893) de scribed a second, Mauritian species, Butorides mauritianus, which is smaller than N. megacephalus, and then Rothschild (1907) described a third species, from Reunion, under the name of Ardea duboisi. Rothschild placed the three species in the ge nus Ardea but wrote (1907:115): "From the short, stout legs and general build, I am inclined to think that all three of these Herons belong to the genus Nycticorax." Later, other authors (Lambrecht, 1933; Hachisuka, 1953; Brodkorb, 1963) placed the three species in different genera, and it was Cowles (1987) who first formally united them in the genus Nycticorax. Nycticorax duboisi (Rothschild, 1907), new combination Reunion Night Heron FIGURE 4a-h "Butors ou Grands Gauziers" Dubois, 1674:169. Ardea duboisi Rothschild, 1907:114 [based on birds described by Dubois (1674) from Bourbon (= Reunion Island)]. Megaphoyx duboisi.?Hachisuka, 1953:175. Nycticorax n. sp. Cowles, 1987:94. Nycticorax borbonensis Cowles, 1994:90, fig.ld,e [new synonymy; holotype, distal half of left tibiotarsus MNHN, LAC 1993-35, from bed 4, Grotte des Premiers Francais (Grande Caverne), Reunion Island]. MATERIAL.?Grotte des Premiers Francais: Holotype of N. borbonensis (see below). Grotte de l'Autel: r. scapula, 330516. Marais de l'Ermitage: r. scapula, 1866; 1. p. coracoid, 1831; 1. d. humerus, 1826; r. ulna, 1828; 1. p. ulna, 1832; r. fe mur, 1827; 2 r. d. tibiotarsi, 1829, 1830; r. tarsometatarsus, 1916; 2 1. d. tarsometatarsi, 1865, 1917. DESCRIPTION AND COMPARISONS.?The material from Reunion agrees perfectly with the genus Nycticorax. The dif ferent species of this genus show much variation in size, the largest being N. caledonicus (Gmelin), which lives in Indone sia, Australia, New Zealand, New Caledonia, and in some Pa cific archipelagos (Mayr and Cottrell, 1979). The remains of N. SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY duboisi are larger than the largest individuals of N. nycticorax or N. caledonicus in the USNM collection. The supratendinal bridge is incompletely ossified in the three tibiotarsi of N. duboisi. The condition is unknown in N. mauritianus, whereas in N. megacephalus the supratendinal bridge is completely ossified (Milne-Edwards, 1874, pl. 14: fig. 7). The dimensions of Mascarene Nycticorax are given in Table 2. Most of the remains ofN. duboisi are larger than those of either N. megacephalus or N. mauritianus (Figure Ai-j), ex cept for the tarsometatarsus, which is almost the same size as in N. megacephalus, and the femur, which is smaller in most di mensions than that in N. megacephalus. The ratio-diagram (Figure 2) shows the differences in the rel ative proportions of the bones in the three Mascarene species, compared with the modern species N. nycticorax and N cale donicus. In N. mauritianus and N. megacephalus the wings (humeri, ulnae, carpometacarpi) are considerably reduced, and the legs (particularly the femora) are longer than in modern species. In contrast, N. duboisi is larger than living species, but the proportions are almost the same. The tarsometatarsus is slightly longer than in N. nycticorax or N. caledonicus, but only two are available, and for one of them the length is esti mated. The ratio-diagram clearly demonstrates that the Rod rigues and Mauritius species had a reduced flying ability, whereas the Reunion form had a flying ability quite similar to that of living species. By the proportions of the tarsometatarsus, which is short and thick, the Mascarene night herons are more similar to Nyctico rax nycticorax than to other congeners, particularly N. cale- TABLE 2.?Dimensions (mm) of the long bones of Nycticorax duboisi, from Reunion, N. megacephalus, from Rodrigues, and TV. mauritianus, from Maritius. (a=data from Milne-Edwards (1874), Giinther and Newton (1879), and material at MNHN; b=data from Newton and Gadow (1893) and from material at MNHN; c=from base of glenoid facet to top of acromion; d=from the most internal part to the most external part; est.=estimated; /j=number of specimens.) Measurement Scapula length art. part (c) width art. part (d) width shaft depth shaft Coracoid internal length width midshaft depth midshaft Humerus total length distal width width midshaft depth midshaft Una total length proximal width proximal depth distal width depth ext. cond. width midshaft depth midshaft Femur total length distal width distal depth width midshaft depth midshaft Tibiotarsus total length distal width distal depth width midshaft depth midshaft Tarsometatarsus total length proximal width proximal depth distal width distal depth width midshaft depth midshaft N. dubo Mean (n) 11.10 (2) 12.65 (2) 4.90 (2) 2.90 (2) - -4.8 (1) -5.4 (1) est. 137 (1) 18.4 (1) 8.1 (1) 7.2 (1) est. 155 (1) 13.5 (1) 11.5 (1) -10.3 (1) -9.0 (1) 6.20 (2) 5.65 (2) est. 85 (1) 14.8 (1) 14.2 (1) 6.6 (1) 6.2 (1) -144.5 (2) 12.87 (3) 13.2 (1) 6.80 (3) 6.03 (3) est. 97.55 (2) -13.5 (1) -13.5 (1) 14.2 (1) 7.35 (2) 6.53 (3) 4.73 (3) isi Range 10.2-12.0 12.4-12.9 4.7-5.1 2.8-3.0 -- - -- -- -- -- - 6.2-6.2 5.6-5.7 -- -- - -139?150 12.4-13.2 - 6.8-6.8 5.9-6.1 95.1-est. 100 -- - 7.2-7.5 6.4-6.8 4.4-5.0 N. megacephalus Range (a) -- -- est. 53.0-57.0 -- 114-119 16.5-16.8 6.9-7.0 5.9 121 -- -- -- 86-90 15.0 - 6.2 - 136-140 13.0 - 6.0 _ 93-95.5 13.7-14.0 13.5 13.5-13.8 7.4 6.0-6.5 5.0 N. mauritianus Range (b) -- -- est. 45.5 4.3 4.2 -- -- 111-112 -- _ _ -_ 79.3 13.8 11.7 6.4 5.8 _ _ _ _ _ 79.5-87 12.7-12.8 12.4-12.6 11.7-12.1 7.1-7.3 5.8-5.9 4.7-4.8 NUMBER 89 donicus. In the case of N. megacephalus and N. mauritianus, however, the robustness of the tarsometatarsus is probably ac centuated by the reduced flying ability (Table 3). The Reunion night heron had green feet and had gray plum age flecked with white, a description that fits very well with the juvenile plumage of Nycticorax nycticorax. REMARKS.?Cowles (1994) thought that the species name Ardea duboisi, created by Rothschild (1907), was a nomen nu dum, but actually this name, published with a description, is valid. It has been used several times (Hachisuka, 1953; Green- way, 1967) and therefore must be retained, in conformity with the law of priority. Many other accepted scientific names of Mascarene birds are based on similar descriptions. The description given by Dubois is as follows: "Bitterns or Great throats, large as big capons [domestic fowl, Gallus gallus (Linnaeus)], but fat and good [to eat]. They have grey plumage, each feather tipped with white, the neck and beak like a heron and the feet green, made like the feet of the 'Poullets dTnde' [domestic turkey, Meleagris gallopavo (Linnaeus)]. That lives on fish" (Barre and Barau, 1982:30, our translation). Dubois' words "Butors ou Grands Gauziers" were left in French by 01- TABLE 3.?Robustness index of the tarsometatarsus in different modem and extinct species of the genus Nycticorax. (Robustness index = midshaft width x 100/ total length; ?=number of specimens.) Species Nycticorax nycticorax nycticorax Nycticorax caledonicus Nycticorax melanolophus Nycticorax (Gorsachius) leuconotus Nycticorax megacephalus Nycticorax mauritianus Nycticorax duboisi Mean 6.25 5.87 5.44 5.02 6.57 7.23 -6.77 Range 5.60-6.59 5.49-6.20 - - 6.32-6.99 7.17-7.29 6.73--6.80 n 11 4 1 1 3 2 2 Cor. Hum. Uln. Cpm. Fern. Tbt. Tmt. 0 Nycticorax caledonicus, standard ^ N. nycticorax C^mean cf N. nycticorax cf, mean ? Nycticorax mauritianus # Nycticorax megacephalus O Nycticorax duboisi -80 I 60 I 40 -20 20 40 I 60 FIGURE 2.?Ratio-diagram of the dimensions of the main long bones of the three species of Nycticorax of the Mascarenes. The standard is a male N. caledonicus from New Caledonia (USNM 561542). For iV. nycticorax the dimensions are the means of three females (USNM 292037, 319467, 430526) and nine males (USNM 289884, 292036, 430527, 432698, 488680, 489903, 499390, 501991, 610609). For N. megacephalus the data are from Milne-Edwards (1874), and for N. mauritianus the data are the means of the dimensions given by Newton and Gadow (1893) plus those of the fossil material in the MNHN collection. Coracoid measurement is of internal length; for other bones, measurement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, Fem.=femur, Hum.=humerus, Tbt.=tibiotarsus, Tmt.=tarsometatarsus, Uln.=ulna.) 8 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY iver (1897) but afterward were translated into English by Roth schild (1907:114) as "Bitterns or Great Egrets." The word gos- ier, in old French gauzier, does not mean "egret" but "throat," and the words grand gosier designated both pelicans and the "argalas" of India {Leptoptilus dubius (Gmelin)). Family THRESKIORNITHIDAE Genus Threskiornis Gray Apterornis Selys-Longchamps, 1848:293 [not Apteromis Owen, 1848:1, a se nior homonym; type by subsequent designation of Gray, 1855:154, Apteror nis solitarius Selys-Longchamps, 1848]. [New synonymy] Ornithaptera Bonaparte, 1854:139 [new name for Apterornis Selys-Long champs, 1848, not Apterornis Owen, 1848:1]. [New synonymy] Borbonibis Mourer-Chauvire and Moutou, 1987:419 [type by original desig nation, Borbonibis latipes Mourer-Chauvire and Moutou, 1987]. [New syn onymy] Threskiornis solitarius (Selys-Longchamps, 1848), new combination Reunion Ibis FIGURES 4k-s, la-e Apterornis solitarius Selys-Longchamps, 1848:293 [based on birds described by Tatton in S. Castleton, 1613 (1625), D.B. (^Dubois), 1671-1672 (1674), and Abbe Carre, 1667 (1699), from Bourbon (= Reunion Island)]. Didus apterornis Schlegel, 1854:244 [based on birds described by Tatton in S. Castelton, 1613 (1625) and in Abbe Carre, 1667 (1699), from Bourbon (=Reunion Island)]. Ornithaptera borbonica Bonaparte, 1854:2 [based on birds described by Tatton in S. Castleton, 1613 (1625) and in Abbe Carre, 1667 (1699), from Bourbon (= Reunion Island)]. Victoriornis imperialis Hachisuka, 193 7a: 71 [based in part on descriptions of the Reunion Ibis by Tatton in S. Castleton, 1613 (1625) and in W.Y. Bon- tekoe, 1646, but mainly on illustrations by Holsteyn and Withoos that likely pertain to the dodo {Raphus) of Mauritius, so that the disposition of the name must depend on future lectotypification]. Borbonibis latipes Mourer-Chauvire and Moutou, 1987:419 [holotype, right juvenile tarsometatarsus, FSL 330512 (UCB), Grotte de l'Autel, Saint-Gilles, commune of Saint-Paul, Reunion Island]. MATERIAL.?Grotte des Premiers Francais: r. d. j. tar sometatarsus, 1993-37. Grotte de l'Autel: r. coracoid, 330510; 1. p. coracoid, 330527; r. carpometacarpus 330511; r. d. j. tibiotarsus, 330513; r. j. tarsometatarsus, 330512; r. d. j. tarsometatarsus, 330514; metatarsal I, 330536; j. pedal phalanx 1 of digit I, 330530; j. pedal phalanx 1 of digit II, 330529; j. pedal phalanx 1 of digit III, 330532; j. pedal phalanx 2 of digit III, 330533; j. pedal pha lanx 1 of digit IV, 330535. Marais de l'Ermitage: Anterior part of mandible, 1872; 1. quadrate, 1913; sacrum, 1918; fragment of pelvis, r. side, 1912; 1. scapula, 1909; 1. p. humerus, 1908; r. p. ulna, 1806; ulna, s., 1910; p. radius, 1871; 3 d. radii, 1808, 1875, 1911; r. car pometacarpus, 1809; 3 r. d. tibiotarsi, 1804, 1805, 1807; 1. ti biotarsus, 1867; 1. d. tibiotarsus, 1868; r. and 1. tarsometatarsi, same individual, 1801, 1803; r. j. tarsometatarsus, 1870; 1. tar sometatarsus, 1802; 1. j. tarsometatarsus, 1869; 2 pedal phalan ges 1 of digit II, 1873, 1874. REMARKS.?It had generally been thought that a representa tive of the family Raphidae (Columbiformes), equivalent to the Mauritius Dodo {Raphus cucullatus (Linnaeus)) or to the Rod rigues Solitaire {Pezophaps solitaria (Gmelin)), used to live on Reunion. For the following reasons, however, we believe that the "solitaire" described by the early explorers was an ibis and not a dodo (Mourer-Chauvire et al., 1995a, 1995b). First, al though the early accounts speak of a solitaire, we did not find any remains of dodo-like birds. On the other hand, we found relatively abundant remains of an ibis, which had never been mentioned in the historical reports. This begged the question, was the solitaire of Reunion an ibis? Second, the morphologi cal and behavioral characteristics given by eyewitnesses agree better with an ibis than with a dodo. Dubois said that the soli taire had a beak like a woodcock {Scolopax) but larger, and Feuilley mentioned that "their food is but worms and filth tak en on or in the soil" (Cheke, 1987:39). The first taxonomic authors to refer to the solitaire of Reunion (Strickland and Melville, 1848; Selys-Longchamps, 1848; Bonaparte, 1854; Schlegel, 1854) regarded it as different from the Mauritius Dodo. Schlegel, although placing it in the same genus, presented a restoration that was quite different from the dodo, showing a bird with a longish beak, probably reflecting Dubois' description of it being like a woodcock. Nevertheless, from the time when paintings of a white dodo were considered to depict the Reunion solitaire (Newton, 1869), this bird was regarded as a species of Raphidae. Storer (1970) was the only person to suggest that it could have be longed to a different family. In addition, as pointed out by Cheke (1987:39), "none of the existing paintings ascribed to Reunion birds has supporting documentation." DESCRIPTION AND COMPARISONS.?We previously indicat ed that the Reunion Ibis, then described under the new name of Borbonibis latipes, was more closely related to the genus Geronticus (Mourer-Chauvire and Moutou, 1987). Now, with more fossil remains and more comparative material, we con clude that this was an error. The remains of the Reunion Ibis have been compared with specimens of the different extant genera of Threskiornithidae in the USNM collection, and they agree perfectly with the ge nus Threskiornis. They are most similar to the Sacred Ibis, T. aethiopicus (Latham), and to the Straw-necked Ibis, T. spini- collis (Jameson), from Australia, which is sometimes ascribed to a separate genus, Carphibis. The ratio-diagram of total bone length (Figure 3) shows that the curve obtained for T. solitarius is practically identical to that of T. aethiopicus and is parallel to that of T. spinicollis. Various skeletal dimensions of Threskiornis are given in Ta ble 4. The differences between T. solitarius, on the one hand, and T. aethiopicus and T spinicollis on the other, are mainly in robustness. The total length of the bones as yet known of T. solitarius is almost the same as in a large male of T. aethiopi cus and is slightly greater than in T. spinicollis, although the NUMBER 89 Cor. Cpm. o Tbt. Tmt. Ph. DI Ph. 1 DM Ph. 1 Dill Ph. 2 Dill Ph. 1 DIV ? Geronticus eremita, standard O Threskiornis spinicollis ? Threskiornis aethiopicus % Threskiornis solitarius 50 I 100 I 150 I 200 FIGURE 3.?Ratio-diagram of dimensions of long bones of Threskiornis solitarius compared with those of a male T. aethiopicus (USNM 558412), and of T. spinicollis (mean of two males, USNM 347785,429720). The standard is a Geronticus eremita (UCB Lyon 1974-1). Coracoid measurement is of internal length; for other bones, mea surement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, DI-DIV=digits I-IV, Ph.=pedal phalanx, Tbt.=tibiotarsus, Tmt.= tarsometatarsus.) width and depth of the proximal and distal ends and of the shaft are almost always larger than in those two species. This indi cates that the Reunion Ibis must have been of comparable size but was much heavier. The quadrate (Figure As, Table 4) is much stronger than in living forms; thus, the head of the bird must have been larger. The two mandibular rami are wider at the level of the symphysis (Figure An-o), and the bill must have been more robust. In Threskiornis solitarius the acrocoracoid is wider on the anterior face (Figure lk-l). Only a proximal part of the humer us is known (Figure 4p), i.e., where the head forms a lobe above the capital groove; the distal outline of this area is much more rectilinear in T. aethiopicus. The two known carpometa- carpi have an accessory foramen at the distal part of the sym physis between the alular and the major metacarpals (Figure Aq), and this foramen is absent in T. aethiopicus. The pisiform apophysis is preserved on only one specimen; it projects fur ther internally than it does in T aethiopicus or T. spinicollis. The alular metacarpals and the pisiform apophyses end in rough protuberances, as can be seen in birds that use their wings to fight. In T. solitarius the minor and major metacarpals are fused over a longer distance, at both proximal and distal ex tremities, than in T. aethiopicus, but the same is true in T. spin icollis. On the tarsometatarsus, the trochleae are more splayed and are disposed on a less-curved line than in T. aethiopicus. The tarsometatarsi of T. solitarius (Figure Ic-e) look very 10 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY NUMBER 89 11 FIGURE 4 (opposite).?Fossils of herons, ibis, and tropicbird from the Mas carene Islands. Nycticorax duboisi: a, left humerus, Marais de l'Ermitage, 1826, palmar view; b. right ulna, Marais de l'Ermitage, 1828, internal view; c, right tibiotarsus, Marais de l'Ermitage, 1829, anterior view; d, right femur, Marais de l'Ermitage, 1827, posterior view; e, left tarsometatarsus, distal part, Marais de l'Ermitage, 1865, anterior view;/ right tarsometatarsus, Marais de rErmitage, 1916, anterior view; g, same, posterior view; h, right scapula, Grotte de l'Autel, 330516, dorsal view. Nycticorax mauritianus: i, left femur, Mare aux Songes, MNHN MAD-6563, posterior view;/ right tarsometatarsus, Mare aux Songes, MNHN MAD-7080, anterior view. Threskiornis solitarius: k, right coracoid, Grotte de l'Autel, 330510, anterior view; /, same, posterior view; m, left scapula, Marais de l'Ermitage, 1909, ventral view; n, mandible, anterior part, Marais de l'Ermitage, 1872, right lateral view; o, same, dorsal view; p, left humerus, proximal part, Marais de l'Ermitage, 1908, anconal view; q, right carpometacarpus, Grotte de l'Autel, 330511, internal view; r, right ulna, proximal part, Marais de l'Ermitage, 1806, palmar view; s, left quadrate, Marais de l'Ermitage, 1913, posterior view. Phaethon lepturus: t, left coracoid, Grotte de l'Autel, 330515, posterior view. All figures are natural size. much like those of a specimen of T. spinicollis (USNM 429720), in which the tarsometatarsi are shorter, with more splayed trochleae than in another individual of the same spe cies examined, although in T. solitarius the trochleae are still more splayed. Indices for the tarsometatarsus (Table 5) show that its distal part is proportionally wider, and its shaft width proportional to depth is greater, in T. solitarius than in T aethi opicus or T. spinicollis. The various pectoral elements so far known, with one excep tion, do not indicate any reduction in flying ability; the cora coid is elongated and the proximal parts of the humerus and ulna are very robust. The only possible indication of a reduc tion in flying ability is the occurrence of an accessory foramen in the symphysis between the alular and the major metacarpal (Mourer-Chauvire et al., 1995b). To our knowledge this fora men exists only in flightless forms. It is present in Palaeotis weigelti Lambrecht, a flightless fossil ratite from the Eocene of Germany, and it also is regularly present in Struthio, in the Spheniscidae, and occasionally in Rhea (Houde and Haubold, 1987). It also exists in Sylviornis neocaledoniae Poplin, a gi ant, flightless, extinct galliform from New Caledonia (Poplin and Mourer-Chauvire, 1985). The ibis of Reunion was probably much heavier than the liv ing members of the genus Threskiornis. It was perhaps flight less in its behavior, but, apart from this accessory foramen, this had not yet led to osteological consequences. Family PHOENICOPTERIDAE Genus Phoenicopterus Linnaeus Phoenicopterus ruber Linnaeus, 1758 Greater Flamingo MATERIAL.?Marais de l'Ermitage. Male-sized fossils: Ar ticular part of 1. mandible, 1906; r. p. tarsometatarsus, 1840; r. s. tarsometatarsus, 1904; 1. j. tarsometatarsus, 1907; 1. d. tar sometatarsus, 1842. Female-sized fossils: r. s. humerus, 1856; 1. d. humerus, 1877; r. d. ulna, 1878; 1. p. ulna, 1905; r. car pometacarpus, 1880; r. p. tibiotarsus, 1898; r. d. tibiotarsus, 1839; r. d. tibiotarsus, 1900; 1. d. tibiotarsus, 1897; 3 1. d. tibio tarsi (2 j.), 1838, 1847, 1899; 1. d. tarsometatarsus with medul lary bone, 1902; r. d. tarsometatarsus, 1901; 5 1. d. tarsometa tarsi, 1844, 1860, 1861, 1862, 1903. Male- or female-sized fossils: 1. quadrate, 1887; fragment of sternum, 1854; fragment of pelvis, 1855; r. scapula, 1876; 1. scapula, 1852; 3 s. ulna, 1845, 1849, 1879; p. radius, 1843; 2 d. radii, 1848, 1881; 1. d. carpometacarpus, 1864; r. s. tibiotarsus, 1857; 1. s. tibiotarsus, 1846; 2 1. d. tibiotarsi, 1858, 1859; r. s. tarsometatarsus, 1863; r. d. j. tarsometatarsus, 1841; 6 fragments of tarsometatarsi, 1851, 1882-1886; j. pedal phalanx, 1853. REMARKS.?The size and shape of the fossils correspond to Phoenicopterus ruber, the Greater Flamingo, and differ from Phoeniconaias minor Geoffroy Saint-Hilaire, the Lesser Fla mingo. Among the tarsometatarsi of female size, one that is broken into two pieces shows a deposit of medullary bone in side the shaft. Such medullary bone tissue develops in practi cally all the bones of the skeleton of a female bird 10 to 14 days before egg laying. It constitutes a reserve of calcium that is used during the laying period to produce the eggshell, and it is very quickly resorbed as soon as egg laying is over (Simkiss, 1967; Rick, 1975). The presence of medullary bone indicates that Greater Flamingos formerly nested on Reunion, probably at or near the spot where the Marais de l'Ermitage is now, which, on some old maps of the island, is shown as a pond that persisted at least until the eighteenth century. The flamingo material also includes bones from juveniles. Greater Flamingos are mentioned several times in historical accounts on Reunion (Lougnon, 1970, 1992), and Feuilley in dicated that there were 3000 to 4000 of them in 1704 on the Et- ang du Gol (Barre and Barau, 1982). They disappeared be tween 1710 and 1730 (Cheke, 1987). They also have been found as fossils on Mauritius (the distal end of an ulna from Mare aux Songes is in the MNHN, MAD 6573), whence the resident population disappeared around 1758 (Cheke, 1987). Family ANATIDAE Genus Alopochen Stejneger Alopochen (Mascarenachen) kervazoi (Cowles, 1994), new combination Kervazo's Egyptian Goose FIGURE lf-n MATERIAL.?Grotte des Premiers Francais. Holotype: Frag ment of rostrum, 1993-19. Paratypes: Sternum, 2 anterior parts, 1993-22, 1993-24; sternum, posterior part, 1993-25; fur- cula, 1993-21; cervical vertebra, 1993-27; r. carpometacarpus, with thickened bony knob, 1993-20; 1. carpometacarpus, 1993- 12 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 4.?Dimensions (mm) of different bones of the extinct Reunion Ibis. aethiopicus (USNM, MNHN, and UCB), and modern T. spinicollis (USNM). out external tubercle; ?=number of specimens.) Threskiornis solitarius, modern T. (a=proximal width measured with- Measurements T. solitarius mean (?) range T. aethiopicus range (?) T. spinicollis range (?) Quadrate length from squamosal articulation to mandibular articulation Mandible width at level of symphysis dorsoventral length at same level Coracoid maximum length internal length proximal width proximal depth head width sternal-facet length sternal-facet depth midshaft width midshaft depth Humerus proximal width Ulna proximal width proximal depth midshaft width midshaft depth Carpometacarpus total length proximal width proximal depth distal width distal depth width metacarpale majus depth metacarpale majus Tibiotarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Tarsometatarsus total length proximal width (a) proximal depth distal width distal depth midshaft width midshaft depth Metatarsal I total length width distal articulation Pedal phalanges phalanx 1 digit I total length proximal width phalanx 1 digit II 18.9(1) 6.4(1) 5.2(1) -58.2(1) 53.5(1) 15.8(1) 13.9(1) 8.5(1) -25.2(1) 5.4(1) 9.2(1) 7.8(1) 32.3(1) 14.6(1) 13.6(1) 7.1(1) 7.2(1) 73.95 (2) 8.8(1) 17.95 (2) 8.6(1) 11.3(1) 5.65 (2) 5.0 (2) 73.0-74.9 17.3-18.6 5.5-5.8 4.8-5.2 15.0-15.5(4) 4.7-6.3 (4) 5.0-6.8 (4) 51.5-59.0(12) 46.2-52.3 (12) 12.1-14.4(12) 11.3-13.0(12) 5.6-7.0(12) 19.4-23.4(12) 4.2-5.9(12) 6.7-9.0(12) 5.4-6.3(12) 26.4-30.0(12) 12.2-13.8(12) 11.7-13.5(12) 5.6-6.6(12) 5.7-6.7(12) 64.3-75.6(12) 6.5-7.5(12) 14.6-17.0(12) 6.2-7.7(12) 9.19-10.3(12) 4.8-5.5 (12) 3.9^t.6(12) 55.8-57.0 (2) 49.6-51.8(2) 13.1-13.7(2) 12.3-12.3(2) 6.0-7.5 (2) 20.0-23.0 (2) 4.7-6.0 (2) 8.5-8.8 (2) 6.3-6.4 (2) 28.2-30.4 (2) 14.2-14.2 (2) 13.1-13.5(2) 6.3-6.7 (2) 6.2-6.8 (2) 73.6-75.2 (2) 7.4-7.4 (2) 16.6-17.2(2) 7.3-8.1 (2) 9.2-9.7 (2) 5.1-5.7(2) 4.2^1.8 (2) 144.5(1) 16.0(1) 21.6(1) 15.78(6) 15.80(3) 9.63 (4) 6.45 (4) 105.10(6) 17.85(6) 17.83 (3) 20.72 (5) 13.90(4) 7.98 (6) 5.65 (6) 16.8(1) 7.3(1) 27.4(1) 7.4(1) - - - 15.0-16.1 15.2-16.6 9.1-10.3 6.3-6.6 103.4-106.9 17.2-18.5 17.0-18.4 19.4-21.3 13.7-14.5 7.2-8.4 5.5-5.8 - - - 129.5-156.0(12) 12.6-15.5(12) 16.4-19.1 (12) 11.8-14.4(12) 12.3-14.7(12) 6.4-8.0(12) 4.6-6.5(12) 89.1-115.2(12) 13.4-16.5(12) 12.1-14.7(12) 13.7-16.2(12) 10.1-12.5(12) 5.3-6.7(12) 4.3-5.4(12) 13.4-15.8(12) 4.2-5.6(12) 25.5-29.4(12) 4.4-5.4(12) 133.4-137.0(2) 13.5-15.0(2) 18.0-18.8(2) 12.8-14.1 (2) 13.3-14.0(2) 7.0-7.5 (2) 5.5-6.0 (2) 89.0-96.7 (2) 14.9-16.4(2) 13.5-14.2(2) 15.3-16.5(2) 11.5-12.2(2) 6.0-6.8 (2) 5.0-5.1 (2) 15.0-15.8(2) 4.5-5.2 (2) 24.0-24.3 (2) 5.3-5.5 (2) NUMBER 89 13 TABLE 4.?Continued. total length proximal width phalanx 1 digit III total length proximal width phalanx 2 digit III total length proximal width phalanx 1 digit IV total length proximal width T. solitarius mean {n) 29.97 (3) 7.40 (3) 29.7(1) 8.2(1) 25.3(1) 5.9(1) 20.8(1) 7.4(1) range 29.5-30.8 7.3-7.5 - - - - - T. aethiopicus range (?) 24.7-30.8(12) 4.9-6.4(12) 24.3-31.2(12) 5.5-5.8(12) 21.9-26.4(12) 4.2-5.4(12) 19.0-24.1 (12) 5.0-6.5(12) T. spinicollis range (?) 25.1-25.6(2) 5.6-5.9 (2) 25.6-27.2 (2) 6.4-6.6 (2) 20.1-22.0(2) 4.9-5.1 (2) 18.0-19.8(2) 5.7-6.2 (2) 23; phalanx 1 of the major digit of wing, 1993-26. Additional Specimen: r. d. radius, 1993-54. Grotte de l'Autel: Cranium, 330525; sternum, anterior part and fragment, 330523; 1. coracoid, 330519; 1. scapula, 330524; r. humerus, 330517; 1. p. ulna, 330518; 2 r. carpometacarpi, 330520, 330522; 1. femur, 330526; r. tarsometatarsus, 330521. Grotte "au sable": Fragment of furcula, 330735; sacrum, r. lateral part of pelvis, 1. lateral part of pelvis from same bone, 330730-330732; d. radius, 330733; phalanx 1 of major digit of wing, 330736; phalanx 2 of major digit of wing, 330734. Marais de l'Ermitage: Sacrum, 1914; fragment of pelvis (without number); 2 1. anterior scapulae, 1825, 1889; 2 1. cora- coids, 1822, 1888; r. s. humerus, 1821; r. d. ulna, 1892; r. d. carpometacarpus, 1891; 1. p. carpometacarpus, 1823; 1. s. car pometacarpus, 1890; r. d. tibiotarsus, 1824; r. s. j. tibiotarsus, 1820; 1. d. j. tibiotarsus, 1915; r. d.j. tarsometatarsus, 1893. REMARKS.?The new genus and species Mascarenachen kervazoi was created by Cowles (1994) for an extinct sheld- goose from Reunion. Comparison of a larger quantity of mate rial from the Grotte de l'Autel and Marais de l'Ermitage, in ad dition to that from the Grotte des Premiers Francais, shows that the Reunion form is very close to the living Alopochen aegypti- acus (Linnaeus), the Egyptian Goose, which lives in many parts of Africa. Extinct species of Alopochen also are known from Madagascar and Mauritius. Cowles (1994) gave two sets of characteristics in the diagno sis of the genus Mascarenachen. First, the bill is shorter and is dorsoventrally deeper at the level of the cranial junction than in other Tadorninae, and the rostral tip forms a true semicircle, whereas in the other Tadorninae it is more pointed and forms a semiellipse (Figure Ig). Second, the sternal carina has an al most straight, not concave, anterior margin and a pronounced ventral manubrial spine. Dubois described the geese of Reunion as being "wild geese, slightly smaller than European geese. They have the same feathering, but with the bill and the feet red. They are very good [to eat]" (Barre and Barau, 1982:30, our translation). This description applies well to a goose related to A. aegyptiacus, TABLE 5.?Tarsometatarsus distal width and shaft indices in the extinct Reunion Ibis Threskiornis solitarius and in modern T. aethiopicus and T. spini collis. (Distal-width index=distal width x 100/total length; shaft index=width midshaft x 100/ depth midshaft; ?=number of specimens.) Species T. solitarius T. aethiopicus T spinicollis Distal width index Shaft index mean (w) range mean (w) range 19.92 (4) 14.65(11) 17.18(2) 19.81-20.04 13.70-15.94 15.82-18.54 141.33(5) 125.45(11) 127.00(2) 129-151 119-137 118-136 which has a pink bill and bright pink legs and feet (Brown et al., 1982). COMPARISON WITH LIVING FORMS.?After examining 11 specimens of A. aegyptiacus in the USNM and MNHN collec tions, we found the cranium and all the elements of the post-cranial skeleton of Mascarenachen to be morphologically very similar to those in the genus Alopochen, although differ ing from those in all the other genera of Tadorninae. The pre maxilla of the Reunion sheldgoose differs from all specimens of A. aegyptiacus, however, by its shorter length (Table 6) and by the semicircular shape of the tip. The rostrum is longer and anteriorly sharper in A. aegyptiacus. In the Reunion sheld goose, the part of the nasal above the nostrils bulges; the nos trils are not narrow and elongated, as in A. aegyptiacus, but are more rounded. The dorsal outline of the premaxilla is almost straight, whereas in A. aegyptiacus it is clearly upturned at the tip. The premaxilla also is much deeper than in A. aegyptiacus. The anterior carinal margin of the sternum is almost straight in the specimen described by Cowles (MNHN LAC-1933-22), but it is strongly concave in specimen UCB FSL-330523 (Fig ure In), which also possesses a ventral manubrial spine that is narrow but very projecting. The shapes of the anterior carinal margin and of the ventral manubrial spine are highly variable in A. aegyptiacus; some individuals have a straight anterior margin, others a slightly incurved one, and still others a deeply incurved one. Likewise, some individuals have a very project ing ventral manubrial spine, others a very short one, and others no ventral manubrial spine at all. 14 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 6.?Dimensions (mm) of the skull and long bones of the extinct Alopochen (Mascarenachen) kervazoi, from Reunion, mod em A. aegyptiacus (USNM), and extinct Alopochen sirabensis (MNHN), from Madagascar. (a= maximal width of frontal at level of processus supraorbitalis of prefrontal bone; b= minimal width of frontal above orbits; c=cranium width at level of insertion of pos- torbital processes; d=bill length, from frontonasal hinge to tip of premaxillae; e=bill depth, from dorsal surface at frontonasal hinge to ventral surface at proximal end of maxillary; f= sternum-keel depth, from dorsal surface of sternum to ventral tip of keel; g= pelvis length, from first synsacral vertebra to last synsacral vertebra; h=length of anterior scapula, from base of glenoid facet to top of acro mion; ?=number of specimens.) Measurement Skull frontal width (a) frontal width (b) cranium width (c) bill length (d) bill depth (e) ratio e x 100 : d Sternum keel depth (f) Pelvis length (g) Coracoid maximum length internal length proximal width proximal depth stemal-facet length stemal-facet depth midshaft width midshaft depth Scapula anterior length (h) Humerus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Ulna total length proximal width proximal depth distal width depth external condyle midshaft width midshaft depth Carpometacarpus total length proximal width proximal depth distal width distal depth width metacarpale majus depth metacarpale majus Phalanx 1, major wing digit total length Phalanx 2, major wing digit total length Femur A. (M.) kervazoi mean (n) 25.3(1) 13.0(1) 30.7(1) 45.1(1) 18.6(1) 41.2(1) 37.20 (2) 100.6(1) -63(1) 56.90 (3) 13.8(1) 12.03 (3) 23.5(1) 5.4(1) 6.87 (3) 4.97 (3) 14.70 (3) 126.0(1) 25.7(1) 14.2(1) 19.2(1) 10.9(1) 9.65 (2) 8.25 (2) -119(1) 13.1(1) 11.9(1) 12.3(1) 10.8(1) 6.65 (2) 6.50 (2) 72.35 (4) 7.70 (4) 19.80(4) 9.05 (4) 6.23 (4) 6.07 (7) 4.78 (6) 31.60(2) 24.1(1) range - - - - - - 36.7-37.7 - - 53.7-60.0 - 11.1-13.0 - - 6.2-7.4 4.6-5.2 14.2-15.5 - - - - - 9.0-10.3 8.0-8.5 - - - - - 6.3-7.0 6.0-7.0 70.7-75.7 7.4-8.2 18.3-22.4 8.6-9.5 5.8-6.5 5.3-6.5 4.2-5.3 30.9-32.3 - A. ae mean (n) 24.17(7) 12.58(8) 30.46 (8) 50.91 (8) 19.16(8) 37.66 (8) 37.17(9) 104.71 (9) 67.71 (9) 60.56(11) 14.17(9) 13.71 (9) 25.30 (9) 6.10(9) 7.52(10) 5.33 (9) 17.11(9) 134.10(11) 28.79(10) 14.86(10) 20.63 (10) 11.63(10) 9.62(10) 8.89 (9) 128.92(11) 13.80(10) 12.53(10) 12.01 (10) 11.47(9) 7.13(10) 6.84 (9) 79.06(10) 8.21 (10) 22.61(11) 9.76 (9) 6.92 (9) 6.06 (9) 5.09(8) - 26.06 (7) gyptiacus range 21.8-26.3 11.1-14.2 29.4-32.3 47.6-54.1 17.5-21.0 34.3-39.9 34.1^*0.2 99.0-112.0 62.0-75.1 55.0-65.6 12.4-16.0 12.2-15.1 23.7-27.9 5.0-7.3 6.5-8.8 4.6-6.2 15.7-19.0 122.8-148.4 25.7-32.7 13.2-16.4 18.3-22.5 10.5-12.5 8.6-11.0 8.1-10.0 118.1-144.1 11.8-15.4 11.5-13.7 10.6-13.5 10.2-12.7 6.4-7.9 5.9-7.5 72.0-86.0 7.2-9.0 19.3-27.4 8.8-10.6 6.3-7.5 5.5-6.7 4.5-6.0 - 24.1-28.1 A. sirabensis mean (ri) - - - - - - - - - 63.11 (16) - - - - 7.34(16) - - 140.26(30) - - - - 10.07 (30) - 128.56(10) - - - - 6.96(10) - 80.69(21) - _ _ _ 6.52(21) _ _ - range - - - - - - - - - 56.7-68.3 - - - - 6.6-7.9 - - 127.3-152.4 - - - - 8.8-11.8 _ 119.5-139.1 _ _ _ _ 6.3-7.5 _ 73.2-90.0 _ _ _ _ 5.9-7.1 - NUMBER 89 15 TABLE 6.?Continued. Measurement total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Tibiotarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Tarsometatarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth A. (M.) kervazoi mean (n) 69.0(1) 18.8(1) 11.8(1) 16.9(1) 12.6(1) 6.9(1) 7.8(1) - - - - - 6.0(1) 5.8(1) 80.5(1) 14.8(1) 13.4(1) -16(1) -11(1) 5.95 (2) 5.30(2) range - - - - - - - - - - - - - - - - - - 5.2-6.7 4.8-5.8 A. aegyptiacus mean (n) 71.04(11) 17.31(10) 12.01 (10) 17.60(10) 13.38(10) 7.24(10) 7.70 (9) 133.30(11) 14.52 (9) 17.42(9) 14.01 (10) 15.33(10) 7.44(10) 6.14(9) 86.42(11) 15.25(10) 13.23(10) 16.17(10) 11.96(9) 6.13(10) 5.90(10) range 65.0-78.0 14.9-20.6 11.0-13.7 15.5-20.1 12.1-15.0 6.4-7.9 6.1-9.0 124.2-146.2 12.8-17.0 16.0-18.7 12.6-15.6 13.7-17.6 6.5-8.3 5.3-6.9 77.0-96.6 14.3-16.1 12.1-14.6 14.6-18.7 10.9-13.4 5.3-7.0 5.2-6.6 A. sirabensis mean (n) 72.64(12) - - - - 8.03(12) - 136.97(19) - - - - 7.55 (19) - 86.03(17) - - - - 6.01 (17) - range 66.8-78.4 - - - - 7.3-9.0 - 123.7-144.8 - - - - 6.2-8.7 - 76.5-94.3 - - - - 5.3-7.3 - Humerus UCB FSL-330517 (Figure li) also shows a slight difference compared with A. aegyptiacus. On the anconal face, below the head, and on the medial side of the pneumatic fossa, there is a proximodistally elongated depression that does not exist in A. aegyptiacus. Moreover, on this humerus, the pneu matic foramen, which opens at the bottom of the pneumatic fossa, has a small surface compared with that of the fossa, whereas in A. aegyptiacus the pneumatic foramen occupies all the surface of the pneumatic fossa. On the whole, because of the great similarities between the Reunion sheldgoose and the living Egyptian goose, we think that Mascarenachen must be considered a subgenus, and that the Reunion form can be designated as Alopochen {Mascaren achen) kervazoi (Cowles), new combination. This subgenus in cludes only the type species, kervazoi. Although the size of A. aegyptiacus is highly variable (Table 6), the mean dimensions of A. (M) kervazoi are smaller than those of A. aegyptiacus. The size of the Reunion sheldgoose ei ther is at the lower limit of variation in A. aegyptiacus or is slightly smaller. COMPARISON WITH FOSSIL FORMS.?An extinct sheldgoose, Chenalopex sirabensis, was described by Andrews (1897) from Holocene deposits in Madagascar. Because Chenalopex is a synonym of Alopochen, it is now known as Alopochen siraben sis (Brodkorb, 1964). It varies widely in size, so Andrews di vided the material into two sets that he attributed to males and females. In a large amount of material at MNHN, the size of the bones also varies according to the locality, the material from Antsirabe being larger than that from Ankazoabo. We present measurements of specimens from both sites (Table 6) without trying to separate males and females. Bones of A. sira bensis on average are larger than those of living A. aegyptiacus and are thus distinctly larger than A. (M) kervazoi. Another extinct species, Sarkidiornis mauritianus, was de scribed by Newton and Gadow (1893). The holotype is a left carpometacarpus with a very projecting alular metacarpal end ing in a callosity. Andrews (1897) showed that this bone corre sponded to the genus Chenalopex and did not differ from the species Chenalopex sirabensis that he was describing from Madagascar (although the former name has priority). The length of this carpometacarpus (77 mm) is within the variation of A. sirabensis (Table 6) and is slightly larger than the largest individuals of A. {M.) kervazoi. The Reunion form shows some characteristics that are probably related to insularity (see be low). The Mauritian form also may be an endemic insular form. With more fossil material from Mauritius, it might be possible to discern distinctive characteristics compared with the Madagascar and Reunion forms. The ratio-diagram (Figure 5) drawn for the species Alo pochen aegyptiacus, A. sirabensis, and A. (M) kervazoi shows a slight lessening of flying ability in A. sirabensis compared with A. aegyptiacus, as indicated by shortening of the ulna and carpometacarpus and by the slight lengthening of the femur. The curve of A. (M) kervazoi closely parallels that of A. sira bensis, but it also shows a slight reduction of the ulna and car pometacarpus and a slightly longer femur. For comparison, an- 16 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Cor. Hum. Uln. Cpm. Fern. Tbt. Tmt. ? Alopochen aegyptiacus standard 5)c Branta hylobadistes O Alopochen (M.) kervazoi + Alopochen mauritianus O Alopochen sirabensis 120 -100 I I 80 -60 I 40 I -20 I 20 40 I FIGURE 5.?Ratio-diagram of the mean dimensions of the long bones of Alopochen (Mascarenachen) kervazoi, compared with those of Alopochen sirabensis, A. mauritianus, and Branta hylobadistes (USNM 322632, see Olson and James, 1991). The standard is the mean of the dimensions of 11 A. aegyptiacus (USNM 19003, 291415, 346399, 346854, 430829, 431686 431687, 488145, 488713; MNHN-LAC 1874-154, 1911-39). Cora coid measurement is of internal length; for other bones, measurement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, Fem.=femur, Hum.=humerus, Tbt.=tibiotarsus, Tmt.=tarsometatarsus, Uln.=ulna.) other extinct species of goose, Branta hylobadistes Olson and James (1991) from the Hawaiian Islands, shows a much more reduced flying ability, with a considerable shortening of the carpometacarpus and lengthening of the femur (Olson and James, 1991). Thus, it can be concluded that A. (M) kervazoi had only slightly reduced flying ability compared with the con tinental form A. aegyptiacus. The other typical characteristic of A. (M) kervazoi is the shortening of the bill, with a depth/length ratio larger than in the continental form A. aegyptiacus. Unfortunately, the bill is unknown in A. sirabensis (none in MNHN, see also Cowles, 1994) as well as in A. mauritianus. Such shortening is very conspicuous in the extinct genera from Hawaii, Thambetochen and Chelychelynechen (Olson and James, 1991). The similarity of morphology between the species from Reunion and Hawaii is undoubtedly a convergent adaptation to an insular environ ment. Genus Anas Linnaeus Anas theodori Newton and Gadow, 1893 Sauzier's Teal FIGURE lq,r MATERIAL.?Marais de l'Ermitage: Sternum, anterior part, 1895; 1. ulna, 1810; 1. tibiotarsus, 1894. REMARKS.?Anas theodori was described by Newton and Gadow (1893) from material including the anterior part of a sternum, two coracoids, eight humeri, and two tarsometatarsi from Mauritius. The material preserved in MNHN includes a cast of this sternum, a coracoid, four humeri, an incomplete ju venile carpometacarpus, a tibiotarsus, and two juvenile tar sometatarsi. The anterior part of the sternum from Reunion is poorly preserved and does not allow detailed comparison. The Reunion tibiotarsus (Figure Ir), however, is absolutely identi- NUMBER 89 17 cal in shape and dimensions to that from Mauritius and there fore must belong to the same species. Compared with living forms of Anas from Madagascar, A. theodori is larger than A. bernieri (Hartlaub) (Figure 5, Table 7) and smaller than A. melleri P.L. Sclater (Howard, 1964). We compared it in the USNM with various species of Anas. The one it resembles most is Anas gibberifrons Muller, which lives in the east and northeast of the Indian Ocean, from the Andaman Is lands to Indonesia and Australia, and on islands in the western Pacific Ocean (Mayr and Cottrell, 1979). The common charac teristics and the differences between A. theodori and A. gibberi frons are as follows. If the anterior part of the sternum illustrated by Newton and Gadow (1893, pl. 34: figs. 11, 12) is perfectly preserved, the carina projects further anteriorly and the ventral manubrial spine (spina externa of Newton and Gadow) is narrow and elongated in A. gibberifrons and is more elongated than in A. theodori. The coracoid of A. theodori looks very similar to that of A. gibberifrons. On the internal side of the posterior face, above the sternal facet, there is a small, proximally directed spike, followed by a small depression, which also is present in A. gibberifrons. On the anterior face, near the internal side of the acrocoracoid, an almost circular muscular scar occurs in A. the- TABLE 7.?Dimensions (mm) of the main bones of the extinct Anas theodori, from Mauritius and Reunion, com pared with modem Anas gibberifrons (USNM) and Anas bernieri (two specimens; Steve Goodman, pers. comm., 1995). For Anas theodori from Mauritius, the dimensions are from material at MNHN (prefix MAD), from New ton and Gadow (N. & G., 1893), or measured from their figures; Reunion dimensions are from material at MHNR (prefix FE-O). (a=width of front part of sternum, measured between the two lateral intermuscular lines; b=stemum-keel depth, from dorsal surface of sternum to ventral tip of keel; c=measured without cnemial crests; d=measured with cnemial crests; ?=number of specimens.) Measurement Sternum width (a) keel depth (b) Coracoid internal length proximal width proximal depth stemal-facet length stemal-facet depth minimum shaft width depth at same level Humerus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Ulna total length proximal width proximal depth distal width depth external condyle midshaft width midshaft depth Tibiotarsus total length (c) proximal width (c) proximal depth (c) distal width distal depth minimum shaft width depth at same level Tarsometatarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Reunion FE-O-1810 -63 7.5 7.0 6.0 7.1 4.4 3.9 FE-O-1894 71.0 8.5 9.7 8.0 - 3.8 3.0 Anas theodori Mauritius MAD-4057 (cast) 27.7 21.7 MAD-4993 41.1 -9 8.3 -16 3.6 4.3 3.5 MAD-4988 73.0 -16 -11.2 6.6 5.2 4.5 MAD-7161 71.0 8.6 9.2 8.0 9.0 3.7 3.2 MAD-4992* -42 -- -- 4.2 3.6 Mauritius N. &G., 1893 42.5 -8 -17.2 -4.1 - MAD-4989 69.6 15.5 7.9 11.3 6.5 5.2 4.2 N. &G., 1893 42.0 9.2 8.3 9.2 - 4.3 - Mauritius MAD-4991 73.1 16.1 8.2 11.6 6.5 5.6 4.6 Anas bernieri -30.9,-31.2 - -- -- 64.5, 66.5 13.3, 13.5 -10.0, 10.3 - -- 58.0, 60.6 7.4, 7.6 -6.0,6.1 -- - 65.1, 67.6(d) 6.7, 6.8 -6.7, 6.8 -- - 38.7, 39.3 7.0, 7.3 -7.3, 7.5 - -- Anas mean (n=4) 24.25 20.78 37.55 8.03 7.20 15.33 2.98 3.80 2.98 69.93 15.18 7.68 10.75 6.50 5.33 4.50 61.55 7.45 6.88 5.98 6.50 3.95 4.00 64.38 7.50 8.53 6.78 7.58 3.08 2.68 36.15 7.48 6.83 7.70 5.48 3.35 2.93 gibberifrons range 23.2-25.8 19.4-21.8 35.1-38.9 7.4-8.4 6.8-7.6 14.5-16.3 2.7-3.2 3.7^t.O 2.6-3.2 -68-71.4 14.5-15.5 7.4-8.0 9.9-11.2 6.2-6.7 5.1-5.6 4.2^1.8 59.5-63.2 7.1-7.6 6.2-7.5 5.7-6.1 6.2-6.8 3.6-^.2 3.6-4.4 59.5-67.4 -7-7.9 -8-9.3 6.6-7.1 7.0-8.0 3.0-3.1 2.5-2.8 33.3-38.9 7.1-8.1 6.3-7.2 7.1-8.3 5.1-5.9 3.0-3.9 2.7-3.1 'Specimen very juvenile. 18 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY odori as well as in A. gibberifrons. The humerus of A. theodori also looks very similar to that of A. gibberifrons. The humeral head forms a well-developed rim above the tricipital fossa, and the pectoral attachment is well developed and elongated. On the distal part, the brachialis anticus impression and the olecranal fossa are deep. The ulna of A. theodori (Figure Iq) is short and stout, exactly as in A. gibberifrons, and its size is the same as in a male of that species (USNM 610562). The humero-ulnar de pression is pronounced proximally. Distally, the external condyle is extended by a narrow and well-defined lip, which ris es proximad along the shaft, as in A. gibberifrons. The tibiotar sus and tarsometatarsus, compared with those of A. gibberifrons, are longer, more robust, and their proximal and distal parts are proportionally wider (Table 7). The proportions of A. theodori are very close to those of A. gibberifrons (Figure 6), but the coracoid, tibiotarsus, and tar- FlGURE 7 (opposite).?Fossils of ibis and waterfowl from Reunion Island. Threskiornis solitarius: a, left tibiotarsus, Marais de l'Ermitage, 1867, ante rior view; b, same, posterior view; c, right tarsometatarsus, Marais de 1'Ermitage, 1801, anterior view; d, same, posterior view; e, right tarsometa tarsus, distal part, Grotte des Premiers Francais, 1993-37, anterior view. Alo pochen (Mascarenachen) kervazoi: f, cranium, Grotte de l'Autel, 330525, forsal view; g, rostrum, holotype, Grotte des Premiers Francais, 1993-19, dorsal view; h, left coracoid, Grotte de l'Autel, 330519, posterior view; i, right humerus, Grotte de l'Autel, 330517, anconal view;y, right tarsometatar sus, Grotte de l'Autel, 330521, anterior view; k, right carpometacarpus, Grotte de l'Autel, 330522, internal view; /, right carpometacarpus, paratype, Grotte des Premiers Francais, 1993-20, internal view; m, left femur, Grotte de l'Autel, 330526, posterior view; n, sternum, anterior part, Grotte de l'Autel, 330523, right lateral view. Aythya sp.: o, right carpometacarpus, Marais de l'Ermitage, 1924, internal view; p, left carpometacarpus, proximal part, Marais de l'Ermitage, 1925, internal view. Anas theodori: q, left ulna, Marais de l'Ermitage, 1810, internal view; r, left tibiotarsus, Marais de l'Ermitage, 1894, anterior view. All figures are natural size. Cor. Hum. Uln. Cpm. Fern. Tbt. Tmt. ? Anas hottentota, standard ? Anas marecula )k Anas bernieri A Anas aucklandica + Anas gibberifrons cj O Anas theodori -150 -100 -50 50 100 150 ! 200 I FIGURE 6.?Ratio-diagram of the mean dimensions of Anas theodori compared with A. gibberifrons (male, USNM 610562) and to the mean dimensions of two Anas bernieri (Steve Goodman, pers. comm., 1995). The standard is Anas hottentota (male, USNM 430832). For comparison, the dimensions of two flightless forms are indicated, Anas aucklandica (USNM 612796), and the extinct Anas marecula, after Martinez (1987). For A. bernieri the internal length of the coracoid has been estimated from its total length. Coracoid measurement is of internal length; for other bones, measurement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, Fem.=femur, Hum.=humerus Tbt.=tibiotarsus, Tmt.=tarsometatarsus, Uln.=ulna.) NUMBER 89 19 20 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY sometatarsus are proportionally longer. These proportions do not indicate a reduction in flying ability. They are very differ ent from those of flightless teals, such as A. aucklandica (Gray) or the extinct form Anas marecula Olson and Jouventin (1996) from Amsterdam Island (Martinez, 1987; Olson and Jouventin, 1996). Thus it is possible that A. theodori had normal flying ability and could fly between Mauritius and Reunion, which explains how the same species could occur on both islands. The only mention of small anatids on Reunion is by Dubois: "River ducks, smaller than European ones, feathered like teals. They are good [to eat])" (Barre and Barau, 1982:30, our trans lation). ANATIDAE, cf. Aythya Boie cf. Aythya sp. FIGURE lo.p MATERIAL.?Marais de l'Ermitage: r. carpometacarpus, 1924; 1. p. carpometacarpus, 1925. REMARKS.?Two carpometacarpi from the Marais de l'Er mitage differ from the genus Anas in that the alular metacarpal does not project very far anteriorly, does not rise very much proximally, and, on the internal face, the posterior outline of the carpal trochlea is less rounded. Thus, they look more simi lar to the genus Aythya, but the pisiform process is broken, making the generic attribution uncertain. In their dimensions (total length 46.8 mm), the carpometa carpi from Reunion correspond to the Northern Hemisphere species A. marila (Linnaeus) or A. ferina (Linnaeus). Around the Indian Ocean the genus Aythya is represented by three spe cies, A. baeri (Radde) (Southeast Asia and China), A. australis (Eyton) (Australia), and A. innotata (Salvadori) (Madagascar). The length of the carpometacarpus is known for one individual of A. innotata (42.9 mm; Goodmann, pers. comm., 1995). Giv en the wide range of variation in this species (Hoyo et al., 1992), it is possible that the Reunion specimens belong to A. innotata. Family FALCONIDAE Genus Falco Linnaeus Falco duboisi Cowles, 1994 Reunion Kestrel FIGURE 13a-e MATERIAL.?Grotte des Premiers Francais. Holotype: 1. tarsometatarsus, 1993-28. Paratypes: Mandible, left part and symphysis, 1993-34; furcula, 1993-32; 1. coracoid, 1993-31; 1. femur, 1993-29; 1. tibiotarsus, large size, 1993-30; r. tibiotar sus, small size, 1993-33. Additional Specimen: r. ulna, 1993- 53. Grotte de l'Autel: r. coracoid, 330547; 1. coracoid, 330548. REMARKS.?This species was described by Cowles (1994) on the material from the Grotte des Premiers Francais. An addi tional ulna was found later, and two additional coracoids were found in the Grotte de l'Autel. The species is characterized by a size generally comparable to that of Falco tinnunculus Lin naeus but with more robust leg bones. In the holotypical tar sometatarsus in particular, the proximal and distal ends are wider and deeper. The two tibiotarsi (Figure I3b,c) are very different in size (57.9 mm compared with 66.8 mm). Cowles indicated that the smaller one is immature, but we disagree because the external aspect of the bone is that of an adult. The difference between the two bones is similar to the range of variation between ex tremes in Falco punctatus Temminck from Mauritius (Table 8). This size variation is due to sexual dimorphism, with the largest individuals being the females, as in other species of Falco (Jones, 1987). As a whole, Falco duboisi is much larger than F. araea (Oberholser) (Seychelles) or F. newtoni (Gurney) (Madagascar and Aldabra) and is slightly larger than F. punctatus. Jones (1987:209) indicated that in F. punctatus the wings are short and rounded at their tip, which is an adaptation for for est-dwelling raptors and also is found in other insular forms of the genus Falco, such as F. novaeseelandiae Gmelin, from New Zealand, and F. araea. The shape of the wings is conver gent with that of the hawks of the genus Accipiter, which also live in forests. Compared with Falco tinnunculus, the coracoid (Figure 13/) and wing bones of F. punctatus are much shorter, the femur and tibiotarsus are slightly shorter, and the tarsometatarsus is almost the same size (Figure 8). In F. duboisi the humerus and carpometacarpus are unknown, but the coracoid and the ulna are not reduced compared with the femur and the tarsometatar sus. The coracoid, the femur, and the tarsometatarsus have the same relative proportions as in F. tinnunculus, whereas in F punctatus the coracoid is very reduced compared with the leg bones. Thus, it can be concluded that the wings of F. duboisi were not as shortened as in F. punctatus. In the historical accounts, Dubois noted three birds of prey. The first were the papangues, Circus maillardi Verreaux, which are still living. "The second ones are named yellow-feet, with the size and shape of falcons. They do harm to the fowls of the inhabitants and the game of the island. The third ones are emerillons, which, although small, do not fail to carry away chickens and eat them" (Barre and Barau, 1982:31, our transla tion). The word emerillon is the French name of Falco colum- barius, the merlin, which is a small form, whereas F. duboisi was almost the same size as F. tinnunculus, the kestrel. We think that it is perhaps the term "yellow-feet" that corresponds to F. duboisi. NUMBER 89 21 TABLE 8.?Dimensions (mm) of the main bones of the extinct Falco duboisi (Reunion) (MNHN, prefix LAC; UCB, prefix FSL) compared with modem Falco punctatus (Mauritius) and modem Falco araea (Seychelles). Dimensions of F punctatus come from an incomplete skeleton in the Mauritius Institute, the measurements given by Cowles (1994) for material at BMNH, and fossil material from Mare aux Songes and Montagne du Pouce, Mauritius (MNHN). Dimen sions off: araea come from a trunk skeleton (female, USNM 488428) and from Cheke and Jones (1987:412) (tarsus length from study skins and living examples). (?=number of specimens.) Measurement Coracoid internal length proximal width proximal depth stemal-facet length stemal-facet depth midshaft width midshaft depth Humerus total length Ulna total length proximal width proximal depth distal width depth external condyle midshaft width midshaft depth Carpometacarpus total length Femur total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Tibiotarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Tarsometatarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Falco duboisi LAC 1993-31 (paratype) 29.6 7.5 6.5 - 2.6 3.0 3.4 - LAC 1993-53 62.7 6.7 6.8 6.2 5.2 3.8 3.8 -LAC 1993-29 (paratype) 49.0 9.0 6.7 9.2 8.1 4.8 4.3 LAC 1993-30 (paratype) 66.8 7.8 8.8 8.2 6.4 4.2 3.7 LAC 1993-28 (holotype) 45.7 8.6 7.2 8.7 5.6 4.4 3.7 FSL 330547, FSL 330548 30.2, 29.6 - 6.6, 6.6 11.6,12.5 2.5,2.6 3.3,3.0 3.3,3.6 LAC 1993-33 (paratype) 57.9 -7.6 7.0 5.4 3.6 3.2 Falco mean (n) 23.55 (4) - 4.63 (4) 10.0(1) 2.0(1) 2.23 (4) 2.43 (4) 44.33 (6) 46.40 (2) 5.3(1) 4.9(1) 4.8(1) 3.8(1) 2.8(1) 3.0 28.23 (4) 41.20(2) 7.1(1) 4.8(1) 7.0(1) 5.3(1) -3.0(1) -3.1(1) 56.80 (4) -- 6.50 (4) 4.77 (3) 2.97 (3) 2.65 (2) 40.54 (5) 6.80 (5) 5.60 (2) 6.33 (6) 4.28 (4) 3.25 (4) 2.33 (3) punctatus range 21.8-25.3 - 4.2^1.9 -- 2.0-2.5 2.3-2.6 42.0-46.6 45.3^17.5 -- -- -- 25.3-29.6 37.7^4.7 -- -- -- 51.7-60.4 -- 6.3-6.9 4.6-5.1 2.8-3.1 2.5-2.8 -38-42.6 6.4-7.1 5.4-5.8 5.7-7.0 3.7-4.7 2.9-3.5 2.2-2.5 Falco araea 20.0 4.7 4.4 8.5 1.3 2.0 1.7 - -- -- -- - - 32.6 5.9 4.4 5.8 4.9 2.7 2.7 --5.0 5.6 -- -- (tarsus length) 27.7?3.3(M=17) Family RALLIDAE Genus Dryolimnas Sharpe Dryolimnas augusti, new species Reunion Rail FIGURE 9 HOLOTYPES.?Right and left tarsometatarsi, MHNR (prefix MHN-RUN-CT) 13, 14, from the same individual. TYPE LOCALITY.?Caverne de la Tortue, locality of Saint-Paul, Reunion Island, Indian Ocean. HORIZON.?Holocene. MEASUREMENTS.?See Table 9. PARATYPES.?All at MHNR (prefix MHN-RUN-CT): Frag ment of left mandible, CT 1; sacrum, CT 2; 1. s. coracoid, CT 3; r. p. humerus, CT 4; 1. d. humerus, CT 5; r. ulna, CT 6; 1. p. fe mur, CT 7; 1. d. femur, CT 8; 2 r. d. femur, CT 9, CT 10; r. p. ti biotarsus, CT 11; 1. d. tibiotarsus, CT 12; pedal phalanx (p. p.) 1 of digit II, CT 15; 2 p. p. 2 of digit II, CT 16-17; p. p. 1 of digit III, CT 18; 2 p. p. 2 of digit III, CT 19-20; p. p. 3 of digit III, CT 21; p. p. 1 of digit IV, CT 22; 5 vertebrae, CT 23-27. ETYMOLOGY.?This species is dedicated to Auguste de Villele, whose interest for the history of his island and tireless 22 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Cor. Hum. Uln. Cpm. Fern. Tbt. Tmt. ? Falco tinnunculus, standard ? Falco punctatus, min. and max. O Accipiter nisus A Falco duboisi -150 I -100 50 50 100 FIGURE 8.?Ratio-diagram of the dimensions of the bones of Falco duboisi compared with the minimum and maximum dimensions of F. punctatus. For F. duboisi the dimensions of the two known tibiotarsi have been indi cated. The standard is Falco tinnunculus (UCB, Lyon 119-8). Accipiter nisus (male, UCB, Lyon 96-5) is included for comparison. Coracoid measurement is of internal length; for other bones, measurement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, Fem.=femur, Hum.=humerus, Tbt.=tibiotarsus, Tmt.=tarsometatarsus, Uln.=ulna.) activity and hospitality made it possible for numerous natural ists to discover the caves of Reunion and, in particular, the Caverne de la Tortue. DIAGNOSIS.?Species larger and with stouter tarsometatar sus than the recent species of the genus. COMPARISONS WITH LIVING FORMS.?We tentatively place the extinct rail of Reunion in the genus Dryolimnas, which tra ditionally includes only the species D. cuvieri (Pucheran). This species is represented by two living subspecies, the flying D. cuvieri cuvieri from Madagascar and the flightless D. cuvieri aldabranus (Gunther) from Aldabra, and by one extinct sub species, D. cuvieri abbotti (Ridgway) from Assumption Island. The two living forms show great differences in the propor tions of their skeletons and in their morphological characteris tics; the coracoid and wing bones are strongly reduced, and the sternal carina is lower and more posteriorly situated in the Al dabra subspecies. The extinct Reunion Rail shows the following morphological similarities to D. cuvieri. On the coracoid, the coracoidal fenes tra is situated along the middle axis of the shaft and there is a well-pronounced stemocoracoidal fossa; on the coracoid from Reunion, only the middle part of the shaft is preserved, but it is possible to see the top of this stemocoracoidal fossa (Figure 9/'). On the humerus, the shaft is thin and sinuous, and there is an elongated, narrow depression on the anconal face, distally below the dorsal pillar of the internal tuberosity (cms dorsale fossae; Baumel, 1979) (Figure 9c). The ulna is relatively elon gate (Figure 9d,e). The femur is very elongate and incurved. It has two curvatures in two different planes; the proximal and distal extremities are incurved both posteriorly and internally (Figure 9f-h). The main differences between the Reunion Rail and D. cu vieri are in size, the former being larger, and in the shape of the tarsometatarsus (Figure 9a, b), which in the Reunion Rail is much more robust. The internal trochlea, however, is posteri orly displaced and is only slightly splayed internally, as in D. cuvieri. The ossified tendinal loop (retinaculum extensorium tarsometatarsi; Baumel, 1979) is broken on the two tarsometa- NUMBER 89 23 FIGURE 9.?Fossils of the Reunion Rail, Dryolimnas augusti, new species, from Caverne de la Tortue: a, right tarsometatarsus, holotype, CT 13, anterior view; b, left tarsometatarsus, holotype, CT 14, posterior view; c, right humerus, proximal part, paratype, CT 4, anconal view; d, right ulna, paratype, CT 6, palmar view; e, same, inter nal view;/ proximal part and shaft of left femur, paratype, CT 7, posterior view; g, distal part and shaft of right femur, paratype, CT 9, posterior view; h, same, anterior view; i, right tibiotarsus, paratype, CT 11, proximal view;/ left coracoid, shaft, paratype, CT 3, posterior view. All figures x 1.5. tarsi, but it was present, as in D. cuvieri. It also is present in Gallirallus australis (Sparrman), whereas it is absent in Apha- napteryx and Erythromachus. On the internal side of the hypo- tarsus there are three ridges and two open grooves, as in Dry olimnas, Aphanapteryx, and Erythromachus, whereas in Gallinula and Fulica, for example, the most internal groove is closed. The accurate lengths of the pectoral and wing bones are un known, but the proportions of the wing bones compared to those of the leg bones are similar to those of the subspecies D. cuvieri aldabranus, so it is likely that the Reunion Rail also was flightless. This hypothesis is corroborated by the robust ness of the tarsometatarsus. The modem species Lewinia pectoralis (Temminck) is con sidered by Olson (1973) to belong to the genus Dryolimnas. The Reunion form differs from it by its much larger size. COMPARISON WITH FOSSIL FORMS.?For the extinct rail of Rodrigues, Milne-Edwards (1874) created the genus Erythro machus as distinct from the extinct genus Aphanapteryx of Mauritius. Gunther and Newton (1879) transferred the Rod rigues species to Aphanapteryx and were followed by Olson (1977), who, however, indicated that these two species have numerous differences in their osteology as well as in their plumage, which is known from historical accounts. Piveteau (1945) had already mentioned strong differences in the cranial morphology. The main osteological differences between Aphanapteryx bonasia (Selys-Longchamps), from Mauritius, and Erythroma chus leguati Milne-Edwards, from Rodrigues, are as follows. In Aphanapteryx the skull is narrower and longer, the temporal fossae are deeper, and the nostrils are shorter and higher (Ol son, 1977). On the sternum, the sternal carina is much lower, and the gap between the coracoidal facets is much wider. The humems is longer, its shaft is more incurved, and its bicipital surface is proportionally shorter, whereas in Erythromachus the humeral shaft is straighten The carpometacarpus is un known because the bone illustrated by Gunther and Newton (1879, pl. 43: fig. G), and referred to E. leguati, does not be long to the Rallidae. The femur is elongated and anteroposteri orly incurved in Aphanapteryx but is shorter and stouter in Erythromachus. The tarsometatarsus is proportionally more ro bust, and its trochleae are less splayed in Erythromachus, whereas in Aphanapteryx the shaft of the tarsometatarsus is proportionally narrower, and the trochleae are more splayed, particularly the internal trochlea. For these reasons we consider the rails of Mauritius and Rodrigues to be two different genera. The Reunion Rail differs from Aphanapteryx by the shape of the tarsometatarsus, which is stouter, with proportionally nar- 24 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 9.?Dimensions (mm) of the bones of the Reunion Rail, Dryolimnas augusti, new species, compared with recent Dryolimnas cuvieri cuvieri from Madagascar (BMNH 1897.5.10.47), and D. cuvieri aldabranus from Aldabra (BMNH s/1989.38.5, BMNH s/1993.6.2). (?=number of specimens.) Measurement Coracoid midshaft width midshaft depth Humerus total length proximal width head width midshaft width midshaft depth Ulna total length proximal width proximal depth midshaft width midshaft depth Femur total length distal width distal depth midshaft width midshaft depth Tibiotarsus proximal width proximal depth distal width distal depth Tarsometatarsus total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Dryolimnas augusti, n. sp. (ri) 3.2(1) 2.3(1) -48(1) 10.0(1) 3.0(1) 3.3 (2) 3.3 (2) -40(1) 4.8(1) 5.4(1) 2.7(1) 3.0(1) -61(1) 10.1-11.3(3) 9.0-9.1(2) 4.5^.9 (4) 4.4-5.0 (4) 9.7(1) 12.1(1) -7.8(1) -8.0(1) 53.0-53.1 (2) 8.7-8.8 (2) 9.2 (2) 9.5(1) 7.2(1) 4.3-4.4 (2) 3.8-3.9 (2) Dryolimnas cuvieri cuvieri (n=\) 3.1 1.9 47.9 9.7 2.8 3.1 2.9 41.2 4.7 5.1 2.3 2.6 50.5 8.3 7.3 3.6 3.6 7.4 10.1 6.7 7.0 47.6 7.2 7.1 7.2 5.9 3.0 2.6 aldabranus (M=2) 2.1,2.2 1.5,1.5 37.7,39.0 8.1,8.7 2.2,2.2 2.2,2.3 2.1,2.2 31.0,32.5 3.8,4.0 3.8,4.0 1.6,2.1 2.1,2.3 42.8,44.8 7.1,7.5 6.6,6.6 2.9,3.3 3.1,3.3 6.3, 6.4 8.9, 9.4 5.8,5.9 6.0, 6.5 40.2, 44.2 6.1,6.5 6.3, 6.4 6.6, 6.6 5.3, 5.4 2.7,2.7 2.2,2.3 rower proximal and distal parts. In Aphanapteryx the hypotar- sus projects more posteriorly and the external calcaneal ridge is situated closer to the external side; in the Reunion Rail this ridge is situated more medially. The Reunion Rail differs from Erythromachus by the characteristics of the humems (shaft thin and incurved), femur (more elongated and incurved), and tar sometatarsus (trochleae less splayed). In the ratio distal width x 100: total length, the Reunion Rail occupies an intermediate position between Dryolimnas cuvieri, which has less splayed trochleae, and more terrestrial rails, such as Erythromachus, Aphanapteryx, and Gallirallus, which have more splayed trochleae. This ratio varies between 14.9 and 16.4 inD. cuvieri, is 17.9 in D. augusti, varies between 19.2 and 21.2 in E. leguati (after the measurements given by Gunther and Newton, 1879), varies between 20.6 and 20.9 in A. bonasia (MNHN), and reaches 20.2 and 22.8 in two specimens of Gallirallus australis (MNHN). REMARKS.?The Reunion Rail is likely to correspond to a bird that was mentioned only by Dubois (1674) as Rale des Bois. It cannot correspond to the Oiseau Bleu, which must have been larger, being the same size as a solitaire, according to Dubois, or the size of a large capon, according to Feuilley (Cheke, 1987). The Reunion Rail was smaller, approximately the size of a Common Moorhen {Gallinula chloropus (Linnae us)), with reduced wings. A fossil rail from Mauritius was recently identified as Dry olimnas cuvieri by Cowles (1987). Genus Fulica Linnaeus Palaeolimnas Forbes, 1893:544 [type by monotypy, Fulica newtonii Milne-Ed wards, 1867]. Paludiphilus Hachisuka, 1953:154 [type by monotypy, Fulica newtonii Milne- Edwards, 1867]. Fulica newtonii Milne-Edwards, 1867 Newton's Coot FIGURE \li-m Fulica newtonii Milne-Edwards, 1867:203, pl. 10. Fulica newtoni.?Anonymous [=A. Newton], 1868:482. Palaeolimnas newtoni.?Forbes, 1893. Paludiphilus newtoni.?Hachisuka, 1953. MATERIAL.?Grotte des Premiers Francais: Rostrum, ante rior part, 1993-44; sternum, 1993-39; incomplete pelvis, 1993- 38; 2 vertebrae, 1993-46; r. tibiotarsus, 1993-40; r. tibiotarsus, 1993-41; fibula, 1993-43; r. tarsometatarsus, 1993-42; 5 pedal phalanges, 1993-45. Grotte de l'Autel: Pedal phalanx 1 of digit III, 330528; pedal phalanx 2 of digit III, 330531. Marais de l'Ermitage: Fragments of pelvis, 1819; r. cora coid, 1814; r. p. coracoid, 1922; I. d. ulna, 1815; 1. carpometac arpus, 1921; r. d. tibiotarsus, 1816; 2 tarsometatarsi, r. and 1., from same individual?, 1811, 1812; r. tarsometatarsus, 1920; r. p. tarsometatarsus, 1813; pedal phalanx 1 of digit II, 1817; ped al phalanx 1 of digit II, 1818; 2 pedal phalanges, 1896, 1923. REMARKS.?Remains of Fulica newtonii in MNHN from the Mare aux Songes, Mauritius, were compared with those from Reunion and were found to be identical, so both populations must have belonged to a single species. The tarsometatarsi show a great range of variation (Table 10), which probably re lates to sexual dimorphism, with the males being larger than the females. Newton and Gadow (1893:292) wrote: "The sternum of F. newtoni resembles in several points that of Aphanapteryx, Erythromachus, and Ocydromus, and differs from Tribonyx, Fulica proper, and Porphyrio, first in the configuration of the whole anterior margin of the sternum, especially in the double or basally divided spina externa, which is moreover broad and flat, while in the other genera this spine is single and furnished with a ventral longitudinal sharp ridge; secondly, by the reced ing and broad anterior margin of the keel, which, however, is well developed, although less than in Tribonyx and Fulica atra, but the tendency towards a reduction of the keel is apparent." The sternum of Fulica newtonii from Reunion (Figure 13/) is NUMBER 89 25 TABLE 10.?Dimensions (mm) of the main bones of the extinct Fulica newtonii from Mauritius (MNHN) and Reunion. (a = maximum length in median plane, b=width between sterno-coracoidal processes, c=width between ventral labial prominences, d=measured with cnemial crests, e=measured without cnemial crests; ?=number of specimens.) Measurement Sternum (Reunion) length (a) width (b) width (c) keel depth Coracoid (Reunion) internal length proximal width proximal depth midshaft width midshaft depth Humerus (Mauritius) total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Ulna (Reunion) distal width external condyle depth midshaft width midshaft depth Carpometacarpus (Reunion) total length proximal width proximal depth distal width distal depth Mean (ri) 66.7(1) 36.6(1) 27.7(1) 20.7(1) 39.4(1) 10.60(2) 6.70 (2) 5.1(1) 3.8(1) 85.40 17.15(2) 9.6(1) 12.15(2) 7.10(2) 5.50(2) 4.70 (2) 7.2(1) 7.3(1) 4.6(1) 5.1(1) 49.1 (1) 4.6(1) 9.3(1) 3.5(1) 4.8(1) Range 10.4-10.8 6.5-6.9 83.5-87.3 16.7-17.6 11.7-12.6 6.9-7.3 5.3-5.7 4.5-4.9 Measurement width metacarpale majus depth metacarpale majus Tibiotarsus (Mauritius+Reunion) total length (d) total length (e) proximal width proximal depth distal width distal depth midshaft width midshaft depth Tarsometatarsus (Mauritius+Reunion) total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Phalanx 1 digit II (Reunion) total length proximal width Phalanx 1 digit III (Mauritius+ Reunion) total length proximal width Phalanx 2 digit III (Reunion) total length proximal width Mean (n) 3.9(1) 3.2(1) 129.24(5) 122.54(5) 12.56(5) 18.60(5) 12.06(9) 11.67(3) 6.75 (8) 5.14(8) 84.15(6) 13.53 (7) 12.35(2) 13.95(6) 10.80(6) 6.51 (7) 5.09 (7) 37.25 (2) 5.27 (3) 34.30 (2) 7.20 (2) 27.3(1) 6.5(1) Range 126.8-131.3 120.2-124.6 12.1-13.0 17.7-19.4 11.2-12.8 11.5-11.8 6.2-7.0 4.6-5.7 76.7-89.3 -12.5-14.2 11.4-13.3 12.4-14.8 9.7-11.7 6.0-6.8 4.7-5.3 36.2-38.3 5.1-5.5 34.3-34.3 -7-7.4 identical to that from Mauritius illustrated by Newton and Gad ow (1893, pl. 35: figs. 5-7). It presents a ventral manubrial spine (spina externa) that is wide, with two small lateral points separated by a shallow notch. This characteristic cannot be considered as different from Fulica, however, for in numerous living species of that genus the ventral manubrial spine is very variable among individuals. Some individuals have a narrow, short point prolonged by a median ventral ridge, others have a short, wide point, and others have two points separated by a notch, as in F. newtonii. We have observed that the shape of the ventral manubrial spine is very variable in F. cristata Gme lin (7 individuals), F. americana Gmelin (20), F. caribaea Ridgway (3), and F. leucoptera Vieillot (7) from the USNM. In F. ardesiaca Tschudi (1) and F. rufifrons Philippi and Land- beck (3), the ventral manubrial spine is wide and short. Three of seven individuals of F. cristata and two of seven individuals of F. leucoptera have two points separated by a notch, rather than a single point. In Fulica newtonii the anterior carinal margin is more poste riorly displaced than in F. cristata, but the carina is still well developed. The shape of this carina is very different from that of the flightless species Tribonyx mortierii Du Bus, in which the carina is low and the anterior carinal margin is wide and is formed by two ridges separated by a median groove. It is still more different from that of Aphanapteryx bonasia, in which the anterior carinal margin is very wide, with two ridges separated by a wide groove, and the carina itself is strongly displaced posteriorly and is very low (Newton and Gadow, 1893, pl. 35: figs. 14-16). In conclusion, the characteristics of the sternum indicate only a slight reduction in flying ability. Milne-Edwards (1867, 1867-1871) wrote that the shape of the posterior iliac crests of the pelvis of F. newtonii was more similar to that of F. atra Linnaeus than to that of F cristata. Although the two pelves from Reunion are not complete enough to see if this characteristic is constant in F. newtonii, we think that by the marked widening of the pelvis at the level of the acetabula and by the strong projection of the antitro- chanters externally, F. newtonii is more similar to F. cristata than to F. atra (see Milne-Edwards, 1867-1871, pl. 99: figs. 1-5, pl. 107: figs. \^). In Fulica newtonii the bones of the scapular girdle and the wing (coracoid, humerus, carpometacarpus) are the same size as in a large male of F. cristata, the tibiotarsus and the phalanx of pedal digit II are slightly larger, and the femur and the tar sometatarsus are much larger (Figure 10). Compared to F. cris tata, F. newtonii was a poorer flier as indicated by the reduc tion of the coracoid and wing bones. For comparison we present the curves of Tribonyx mortierii, the flightless Tasma- nian Native-Hen, and of Aphanapteryx bonasia, the extinct, flightless Mauritius Red Rail. In these two species, the cora- 26 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Cor. Hum. Cpm. Fern. Tbt. Tmt. Phal. 1 post, digit II * ? Fulica atra, standard ? Aphanapteryx bonasia )|c Tribonyx mortierii + Fulica newtonii, min. and max. o Fulica cristata 100 50 50 100 II 150 200 I FIGURE 10.?Ratio-diagram of long bones of Fulica newtonii compared to those of F cristata (male, USNM 430843), Tribonyx mortierii (UCB, Lyon 1975-1), and Aphanapteryx bonasia. The standard is Fulica atra (UCB, Lyon 147-2). For Fulica newtonii the minimum and maximum dimensions of the fossil material from Mauritius and Reunion are indicated and include measurements given by Newton and Gadow (1893). For the extinct Apha napteryx bonasia, the dimensions are from the fossil material at MNHN (MAD 6501, 6502, 6561, 6565, 6566, 6579, 6580, 6818, 6937, 6967) and from Newton and Gadow (1893). Coracoid measurement is of internal length; for other bones, measurement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, Fem.=femur, Hum.=humerus, Phal.= phalan ges, post.=posterior, Tbt.=tibiotarsus, Tmt.=tarsometatarsus.) coids are the same size as in F. newtonii, and the proportions of the leg bones are very similar to those of F. newtonii, but the reduction of the wing is much more advanced. The pedal pha langes are much shorter in T. mortierii. So it can be concluded that F. newtonii had some reduction in flying ability but was still able to fly, which explains how the same species could be present on both Mauritius and Reunion. Several early authors have mentioned the presence of "moorhens" on Reunion, but the most detailed description was given by Dubois: "Moorhens, which are as big as hens. They are completely black and have a big white crest on the head" (Barre and Barau, 1982:30, our translation). On these grounds, Milne-Edwards (1867-1871) said that F. newtonii must have been very different from F. cristata, the forehead shield of which is dark red; actually, the forehead shield in F. cristata is white, sometimes tinged with pink, and is topped by two more or less developed red tubercles in the adult (Langrand, 1990). Keith (in Urban et al., 1986:129) wrote that the Red-knobbed Coot is "not easy to tell from the Eurasian Coot, F. atra. At close range red knobs at top of shield distinguish it, but during non-breeding season they are small and very hard to see." In conclusion we think that F. newtonii was probably derived from F. cristata, which lives mainly in southern and East Afri ca and on Madagascar. It is not possible, however, to exclude the possibility that it could be derived from F. atra, for this species is widely distributed in the Palearctic region as well as in India, Indonesia, Australia, Tasmania, and New Zealand (Cramp and Simmons, 1979). NUMBER 89 27 Family SCOLOPACIDAE Genus Numenius Brisson Numeniusphaeopus (Linnaeus, 1758) Whimbrel MATERIAL.?Grotte des Premiers Francais: 1. tarsometa tarsus, 1993-52. REMARKS.?The Whimbrel is a palearctic migrant that still occurs regularly, between September and March, on the mud flats and beaches of the west coast of Reunion. It was known to the early explorers, who mentioned it using its old French name of corbigeau (Barre and Barau, 1982). Family COLUMBIDAE Two species of pigeons are known from Mauritius, Alectroe- nas nitidissima, the extinct Blue Pigeon, or Pigeon hollandais, and Nesoenas mayeri (Prevost), the Pink Pigeon, which still survives (Jones, 1987). Milne-Edwards (1874) described a very special sternum from Rodrigues, which did not correspond to any living genus, under the new name of "Columba" roderica- na, and he attributed a tarsometatarsus to the living species Streptopelia picturata (Temminck), of Madagascar. Gunther and Newton (1879) said there was no reason to put these two elements into two different species and listed them both under Columba rodericana. Shelley (1883:258) wrote: "Columba ro- dericana Milne-Edwards, is only known by a few bones. It was a native of Rodriguez, and probably belonged to the genus Alectroenas." Later, Rothschild (1907) referred to it as Alectro- enas (?) rodericana, and Hachisuka (1953) referred to it as Alectroenas rodericana, without a question mark. Although Cowles (1987:97) also placed C. rodericana in the genus Alec troenas, he recognized that the sternum was "quite unlike that of any living genus known today." The sternum described and illustrated by Milne-Edwards (1874, pl. 12: figs. 1, la-c) is quite different from that of the genus Alectroenas. Among the living genera of Columbidae that we have been able to examine, it is most similar to that of the genus Gallicolumba, the present distribution of which ex tends from the Philippines, New Guinea, and Celebes and adja cent islands to Polynesia (Peters, 1937; Steadman, 1992). The tarsometatarsus (Milne-Edwards, 1874, pl. 12: fig. 2f) appears inseparable from Streptopelia picturata. There is no fossil evi dence that the genus Alectroenas was present on Rodrigues. Rather, the island probably was occupied by an extinct genus, including "'Columba''' rodericana, and by Streptopelia pictura ta, the Madagascar Turtledove. On Reunion, Bontekoe indicated "ramiers of the species with blue wings" (Barre and Barau, 1982:26), and Dubois men tioned two kinds of wild pigeons, in addition to ramiers and turtledoves, thus apparently indicating four species of colum- bids. Dubois described the wild pigeons as "some with slaty-coloured feathering, the others msset-red. They are a little larger than the European pigeons, and have a stronger bill, red at the end close to the head, the eyes bordered by the colour of fire, like the pheasants" (Barre and Barau, 1982:30, our transla tion). The slaty or blue-winged birds are generally considered to belong to the genus Alectroenas, and the msset-red birds are considered a form related to Nesoenas mayeri from Mauritius (Cheke, 1987). On the basis of Dubois' description, Rothschild (1907) named the red form Nesoenas duboisi, citing for this species only the characteristics of the bill and of the border of the eyes, which in Dubois' account apply not only to the red form but also to the blue one. Because one of the fossil species found on Reunion belongs to the genus Nesoenas, the name created by Rothschild must be used for it, as in the case of Nyc ticorax duboisi. Genus Nesoenas Salvadori Nesoenas duboisi Rothschild, 1907 Reunion Pink Pigeon FIGURE \3g,h MATERIAL.?Grotte des Premiers Francais: r. d. humems, 1993-55. Grotte de l'Autel: r. humems, 330546. REMARKS.?The humems from Grotte de l'Autel (Figure I3g,h) is similar to that of Nesoenas mayeri, the Pink Pigeon of Mauritius. It differs from the genus Alectroenas by the follow ing characteristics. On the anconal face, in Nesoenas as in Columba, there is a slightly indicated tubercle situated distally compared with the humeral head, more or less on the median axis of the bone, at the place where the capital groove ends on its medial side. This tubercle does not exist in Alectroenas. The humeral head is more proximodistally elevated in Alectroenas and is more flattened in Nesoenas. The bicipital surface is more internally projecting in Nesoenas. The pectoral attachment is narrow and elongated, only slightly protruding, and oriented along the axis of the bone in Alectroenas, whereas it is more protmding, with a more triangular shape, and obliquely orient ed in Nesoenas. The distal part is more mediolaterally elongat ed in Alectroenas. On the palmar face, the impression of M. brachialis anticus is much wider and more diffuse in Alectroe nas, whereas it is smaller and with a more discrete outline in Nesoenas. The internal condyle, more globular in Alectroenas, is more elongate in Nesoenas. The attachment of the anterior ligament is more protmding in Nesoenas. We have compared humeri of Nesoenas from Reunion with three humeri of captive N. mayeri from the USNM collection and with a series of 39 fossil humeri from the caves of Le Pouce Mountain, Mauritius (MNHN). Most of the dimensions of the Reunion humeri fall within the range of variation of N. mayeri except for the total length, which is a little larger (Table 11). Because we have a good sample of comparative material, we think that the Reunion Nesoenas belongs to a different spe cies, characterized by slightly larger size than the Mauritian one. With the hope of finding more fossil material, we refer it for now to Nesoenas duboisi Rothschild. 28 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 11.?Dimensions (mm) of the humerus of Nesoenas mayeri, modem and fossil, from Mauritius, and N. duboisi, extinct, from Reunion. The modem skeletons of N. mayeri are from USNM, and the fossil humeri are from the caves of Montagne du Pouce, Mauritius (MNHN). (a=from external tuberosity to bicipital crest, with out deltoid crest; b=from internal tuberosity to bicipital surface, without deltoid crest; ?=number of specimens; 5=standard deviation.) Measurement Humems total length proximal width (a) proximal depth (b) distal width distal depth midshaft width midshaft depth midshaft widthxlO/total length mean (n) 45.58 (34) 13.63(29) 8.38 (27) 10.54(35) 6.54(31) 4.95 (39) 3.88 (39) 1.09(33) Nesoenas mayeri s 1.64 0.54 0.37 0.46 0.18 0.20 0.24 0.05 range 42.5-19.2 12.7-14.4 7.7-9.1 9.3-11.4 6.2-6.9 4.6-5.3 3.4-4.4 1.01-1.20 Nesoenas duboisi 50.4 14.2 -10.4, 11.1 6.6, 6.6 5.1 4.0 1.01 Genus Streptopelia Bonaparte Streptopelia picturata (Temminck, 1813) Madagascar Turtledove FIGURE 13? MATERIAL.?Grotte "au sable": r. p. humems, 330738; 1. d. ulna, 330739. REMARKS.?Both remains correspond to the living Strep topelia picturata (one skeleton in MNHN). They are similar to fossil remains from the caves of Le Pouce Mountain, Mauri tius. The Madagascar Turtledove lives now in Madagascar and on other islands of the western Indian Ocean (Glorioso, Anj- ouan in the Comoros, Aldabra, Assumption, the Amirantes, some of the Seychelles, and Diego Garcia; Peters, 1937). Al though assumed to have been introduced to Reunion, Mauri tius, and Rodrigues, its presence as a fossil indicates that it was living on these three islands before humans arrived, disap peared, and was then reintroduced (Cheke, 1987). Dubois spoke of ramiers and turtledoves in Reunion, and ac cording to Cheke (1982), the name Pigeon ramier is still used in Mauritius and Reunion to designate S. picturata. Family PSITTACIDAE Genus Mascarinus Lesson Mascarinus mascarinus (Linnaeus, 1771) Mascarene Parrot FIGURE 13O-JC MATERIAL.?Grotte des Premiers Francais: r. d. coracoid, 1993-56. Grotte de l'Autel (bones probably from one individual): 1. coracoid, 330545; r. and 1. humeri, 330539, 330540, respec tively; 1. ulna, 330541; 1. carpometacarpus, 330544; 1. femur, 330543; 1. tibiotarsus, 330542. Grotte "au sable": 1. scapula, 330810; 2 1. coracoids, 330741, 330742; r. p. coracoid, 330743; r. s. carpometacarpus, 330740; r. d. tarsometatarsus, 330744. REMARKS.?We think the few remains of a large parrot are from Mascarinus mascarinus, a genus and species endemic to Reunion that became extinct between 1750 and 1800 (Barre and Barau, 1982). Unfortunately, no skeleton has been pre served for this species, which is known from two mounted specimens, one in MNHN (Paris, 1998-1725) and the other in the Natural History Museum of Vienna (Austria, 50.688). X-ra- diographs made it possible to take the measurements of some bones, which show that the fossil remains are intermediate be tween those of the two modem specimens (Table 12). We have compared the fossil remains from Reunion with the species Coracopsis nigra (Linnaeus), the Lesser Vasa Parrot, which lives on Madagascar, the Comoros, and on Praslin in the Seychelles. Coracopsis nigra was introduced to Reunion very early, circa 1780 (Cheke, 1987), and is about the same size as M. mascarinus; in both species the total length is 35 cm (Lan- grand, 1990; Forshaw, 1973). The lengths of the coracoid, fe mur, and tibiotarsus of the Reunion parrot fall within the range of variation of C. nigra, whereas the humems, ulna, and car pometacarpus are somewhat smaller (Table 12). The parrot of Reunion also shows morphological differences compared with C. nigra. The distal part of the humems is more laterally com pressed in the Reunion form, and the olecranal fossa is narrow er, whereas in Coracopsis nigra, as well as in C. vasa (Shaw), the distal part of the humems is mediolaterally wider (Figure 11). The ratio distal depth x 100:distal width is 65.6 in the Reunion parrot, whereas it ranges from 57.3 to 63.0 in six mod em C. nigra and from 57.8 to 59.7 in two modem C. vasa. On the distal part of the tibiotarsus from Reunion (Figure 13x), on the anterior face, the internal condyle is narrow, not flattened, and is proximodistally oriented, whereas in C. nigra it is wide, flattened, and oriented proximointernally. In five of the six C. nigra examined, the tendinal groove is situated almost in the median plane of the bone; in the Reunion form, it is situated al most on the internal side. On the Reunion tibiotarsus there is a depression on the anterior face, above the external condyle, but this depression does not exist in C. nigra. Other differences probably occur in the proximal part of the tarsometatarsus, but this is unknown in the Reunion form. At the distal end, on the posterior face, the accessory trochlea is less anteroposteriorly developed in the Reunion parrot than in C. nigra. NUMBER 89 29 TABLE 12.?Dimensions (mm) of the main long bones of M. mascarinus fossil (UCB, prefix FSL) and mounted speci mens, compared with modem Coracopsis nigra (USNM, AMNH, MNHN). Measurement Coracoid internal length proximal width proximal depth stemal-facet length stemal-facet depth midshaft width midshaft depth Humems total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Ulna total length proximal width proximal depth distal width external condyle depth midshaft width midshaft depth Carpometacarpus total length proximal width proximal depth distal width distal depth width metacarpale majus depth metacarpale majus Femur total length proximal width proximal depth distal width distal depth midshaft width midshaft depth Tibiotarsus total length proximal width proximal depth midshaft width midshaft depth Tarsometatarsus total length midshaft width width middle trochlea depth middle trochlea depth external trochlea Grotte de l'Autel FSL-330545 29.2 7.2 4.3 -- 2.9 2.5 FSL-33053940 48.8-48.8 13.0 7.6 9.3-9.3 6.1-6.1 4.7-4.7 3.9-3.9 FSL-330541 57.8 7.6 6.0 6.0 5.4 3.5 3.6 FSL-330544 35.9 4.0 9.0 3.8 5.6 3.0 2.6 FSL-330543 39.7 7.5 4.4 7.6 6.2 3.1 2.9 FSL-330542 57.8 6.0 7.2 3.0 2.7 Mascarinus mascarinus Grotte "au sable" FSL-330741-743 28.9 7.6-7.2 -4.4-4.3 -8.2 2.0-2.0 3.2-3.2 2.4-2.4 FSL-330740 -- -3.9 - 3.1 2.6 FSL-330744 -- 3.6 2.8 4.0 mounted specimens Paris Vienna 60.5 -- -- 3.8 - 37.2 35.0 -9.0 -- 2.7 2.7 2.5 18.5 2.7 -- - Coracopsi mean (n) 30.05 (8) 8.08 (6) 4.88 (6) 8.83 (6) 2.48 (6) 3.50(6) 2.43 (3) 51.88(6) 13.27(6) 7.67 (6) 10.43(6) 6.23 (6) 4.85 (6) 3.97(6) 61.98(5) 7.52(5) 6.32(5) 6.20 (5) 5.62(5) 3.72 (6) 3.80(3) 38.22 (6) 4.06 (5) 9.34(5) 4.06 (5) 6.02 (5) 3.10(3) 2.87 (6) 38.43 (7) 7.46 (7) 4.81 (7) 7.60 (7) 5.99 (7) 3.08 (7) 2.93 (3) 56.84 (7) 6.40 (6) 7.18(6) 2.89 (7) 2.63 (3) 22.30 (7) 3.32 (6) 3.50 (6) 3.03 (6) 4.47 (6) 5 nigra range 28.7-30.9 7.7-8.3 4.6-5.4 7.4-10.1 2.2-2.7 3.1-3.8 50.2-53.6 12.5-14.2 7.3-8.1 10.0-11.0 5.9-6.4 4.5-5.4 3.8-4.1 59.3-64.4 7.5-7.6 6.2-6.6 6.0-6.5 5.5-6.0 3.5-4.0 3.6-4.0 36.7-40.0 3.8-4.3 9.0-10.0 3.8-4.2 5.8-6.3 3.0-3.2 2.6-3.1 37.0-40.1 6.9-8.2 4.6-5.0 7.1-8.3 5.6-6.4 2.9-3.4 2.8-3.0 54.6-59.5 5.9-7.0 6.9-7.9 2.4-3.4 2.4-3.0 21.1-23.8 3.0-3.8 3.2-3.9 2.9-3.3 4.2^1.7 FIGURE 11.?Drawing of the distal part of the right humems in the genus Cora copsis (a) (Coracopis nigra, MNHN LAC 1883-507), compared with a humerus from Reunion referred to Mascarinus mascarinus (b) (FSL 330539). In conclusion, the parrot remains of Reunion present some differences in their dimensions and morphological character istics compared with C. nigra, and because their dimensions are compatible with those available for M. mascarinus, we think that they can be referred to that species. Compared to modem C. nigra, M. mascarinus had the femur and tibiotarsus of the same size, but the wing bones and tarsometatarsus were shorter (Figure 12). The main difference is the length of the tarsometatarsus from Vienna, which is much smaller than the tarsometatarsus of C. nigra from Madagascar, although the 30 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Cor. ? Hum. Uln. Cpm. < i Fem. Tbt. Tmt. ? Psittacula krameri, standard * Mascarinus mascarinus, fossil D M. mascarinus, Recent, Paris and Vienna + Necropsittacus rodericanus O Coracopsis nigra, min. and max. 80 I -100 120 I 140 I 160 I 180 200 I FIGURE 12.?Ratio-diagram of the dimensions of bones of fossil Mascarinus mascarinus from Reunion, com pared with those taken from x-radiographs of the two mounted specimens (MNHN, Natural History Museum of Vienna) and with Coracopsis nigra. The standard is Psittacula krameri (Pierce Brodkorb 27712). For C. nigra the dimensions are the minimum and maximum of the specimens at AMNH (3571, 4399), MNHN (LAC 1883- 507-1883-509), and USNM (224810, 292917, 432236). Coracoid measurement is of internal length; for other bones, measurement is of total length. When measurements are not known, successive points are united by dashed lines. (Cor.=coracoid, Cpm.=carpometacarpus, Fem.=femur, Hum.=humerus, Tbt.=tibiotarsus, Tmt.= tarsometatarsus, Uln.=ulna.) subspecies from the Comoros and Praslin are smaller (For- shaw, 1973). According to the accounts of the early explorers, there were at least four species of parrots on Reunion. The description giv en by Dubois of "parrots a little bigger than pigeons, the feath ering of the color of petit-gris, a black hood on the head, the beak very strong and the color of fire" is considered to apply to M. mascarinus (Barre and Barau, 1982:31, our translation). Petit-gris is the name given to the fur of the Eurasian Red Squirrel {Sciurus vulgaris Linnaeus) in its dark phase. Two other species were a grey parrot, which according to Feuilley was smaller than M. mascarinus, and a green parakeet with a black ring, the Reunion Ring-necked Parakeet, Psittacula eques (Boddaert), which probably was conspecific with the still-living Mauritian Echo Parakeet (Cheke, 1987). A fourth species, which was the same size as P. eques, was described by Dubois as a "green parrot with head, upper parts of wings, and tail color of fire" (Barre and Barau, 1982:31, our translation). FIGURE 13 (opposite).?Fossils of falcons, coots, pigeons, and parrots from the Mascarenes. Falco duboisi: a, left coracoid, Grotte de l'Autel, 330548, posterior view, xl.5; b, left tibiotarsus, paratype, Grotte des Premiers Francais, 1993-30, anterior view; c, right tibiotarsus, paratype, Grotte des Premiers Francais, 1993-33, anterior view; d, left femur, paratype, Grotte des Premiers Francais, 1993-29, posterior view; e, left tarsometatarsus, holotype, Grotte des Premiers Francais, 1993-28, anterior view. Falco punctatus: f left coracoid, Montagne du Pouce caves, Mauritius, MNHN, not numbered, posterior view, xl.5. Nesoenas duboisi: g, right humems, Grotte de l'Autel, 330546, anconal view; h, same, pal mar view. Fulica newtonii: i, sternum, Grotte des Premiers Francais, 1993-39, ventral view;/ right tibiotarsus, Grotte des Premiers Francais, 199341, anterior view; k, right tarsometatarsus, Marais de l'Ermitage, 1811, anterior view; /, right coracoid, Marais de l'Ermitage, 1814, posterior view; m, left ulna, distal part, Marais de l'Ermitage, 1815, internal view. Streptopelia picturata: n, right humems, proximal part, Grotte "au sable," 330738, anconal view, xl.5. Mascar inus mascarinus: o, right humems, Grotte de l'Autel, 330539, anconal view;p, same, palmar view; q, left carpometacarpus, Grotte de l'Autel, 330544, internal view, xl.5; r, left coracoid, Grotte de l'Autel, 330545, posterior view, xl.5; 5, same, anterior view, xl.5; /, left ulna, Grotte de l'Autel, 330541, palmar view; u, same, internal view; v, left femur, Grotte de l'Autel, 330543, posterior view; w, same, anterior view; x, left tibiotarsus, Grotte de l'Autel, 330542, anterior view. Mascarenotus grucheti: y, right tarsometatarsus, holotype, Grotte des Premiers Francais, 199349, anterior view. (a,f n,q,r,s xl.5; others natural size.) NUMBER 89 31 32 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Without any justification, Rothschild (1907) placed this form in the genus Necropsittacus of Rodrigues and described it as N. borbonicus. Family STRIGIDAE Genus Mascarenotus Mourer-Chauvire, Bour, Moutou, and Ribes Mascarenotus grucheti Mourer-Chauvire, Bour, Moutou, and Ribes, 1994 Gmchet's Mascarene Owl FIGURE 13^ MATERIAL.?Grotte des Premiers Francais. Holotype: r. tarsometatarsus, 1993-49. Paratypes: 1. humems, 1993-50; r. p. tibiotarsus, 1993-48; r. d. tibiotarsus, 1993-47. Grotte de l'Autel: r. tarsometatarsus, 330537; r. d. tar sometatarsus, 330538. Grotte "au sable": 1. quadrate, 330809; 1. p. femur, 330737. Marais de l'Ermitage: r. tarsometatarsus, 1800. REMARKS.?We have not found any additional remains oth er than those used in the original description of Mascarenotus grucheti (Mourer-Chauvire et al., 1994). We have placed all the Mascarene strigiforms in the extinct genus Mascarenotus, which resembles the genus Otus but presents some distinctive morphological characteristics. This genus includes one species on Mauritius, Mascarenotus sauzieri, based on Strix sauzieri Newton and Gadow, 1893 (synonyms, Otus commersoni Oust- alet (1896), Strix newtoni Rothschild (1907)), one species on Rodrigues, Mascarenotus murivorus based on Strix {Athene) murivora Milne-Edwards (1874) (synonym, Bubo (?) leguati Rothschild (1907)), and one species on Reunion, Mascarenotus grucheti Mourer-Chauvire et al., 1994. The tarsometatarsi of M. grucheti are very close in size to those of M. sauzieri, although the only humems known from Reunion is clearly smaller than those of M. sauzieri, for which we have a good sample of comparative material (12 speci mens). The ratio of element lengths (Mourer-Chauvire et al., 1994, fig. 1) shows M. sauzieri to be strikingly parallel to mod em insular species of the genus Otus, such as Otus lawrencii (Sclater and Salvin), or O nudipes (Daudin), from the West In dies, in which the legs are much more elongated than in conti nental forms. The same adaptation is found in the four extinct insular species of the genus Grallistrix, from Hawaii, which is derived from the genus Strix (Olson and James, 1991). In the Strigiformes, as well as in the genus Accipiter, this lengthening of the legs corresponds to an adaptation for catching birds on islands lacking terrestrial mammals. This species must have had a very secretive life in the forests or remote areas, because the early explorers of Reunion never spoke of nocturnal raptors, although eared owls were noted his torically on the other Mascarene Islands. Family STURNIDAE Genus Fregilupus Lesson Fregilupus varius (Boddaert, 1783) Reunion Starling MATERIAL.?Grotte des Premiers Francais: 1. d. femur, 1993-57. REMARKS.?This passerine femur agrees in size and mor phological characteristics with what is known of the Reunion Starling. On the posterior face, the proximal edge of the inter nal condyle ends internally with a point that projects proximad. On the internal face, the internal condyle is anteroposteriorly compressed. This characteristic accentuates the shape of the external condyle, which looks very protmding, and the rotular groove on the anterior face is deep, as was indicated by Murie (1874), who described the only known skeleton of this species. The distal width (7.2 mm) and the distal depth (5.8 mm) agree with a femur the total length of which ranges from 31.6 mm (Berger, 1957) to 35.6 mm (Murie, 1874) or 37 mm (Gunther and Newton, 1879). The Reunion Starling became extinct between 1838 and 1858 (Barre and Barau, 1982). Relationships, Origin, and Fate of the Reunion Avifauna The avifauna found as fossils on Reunion differs from that of the other two Mascarenes (Table 13) in that, with the exception of Dryolimnas augusti, none of the species had lost their ability to fly. Lacking on Reunion are the most distinctive Mascarene birds, namely, the dodo and solitaire {Raphus, Pezophaps), the large flightless rails {Aphanapteryx, Erythromachus), and the large parrots with enormous bills and atrophied wings {Lo- phopsittacus mauritianus, Necropsittacus rodericanus) (New ton and Newton, 1876; Gunther and Newton, 1879). Among the forms that perhaps had lost the ability to fly is the Oiseau bleu, placed by Olson (1977) in the genus Porphyrio, and which was either an extinct species of that genus or a popula tion of the modem species Porphyrio porphyrio (Linnaeus). Dubois (1674) said that it could not fly, but in 1724 Father Brown said that it was able to fly, but rarely and just above the ground (Barre and Barau, 1982, our translation). The authentic ity of Father Brown's report has been questioned (Lougnon, 1970, 1992), but Cheke (1987) thinks that his report comes from an unidentified but authentic source. Remains referable to Oiseau bleu are yet to be found, so we know nothing more about it. The other genera represented on Mauritius and Rodrigues by species with reduced flying ability are represented on Reunion by species with normal flying ability. This is the case for Nycti corax, with the flying species N. duboisi, and for Falco duboi si, the coracoid and ulna of which are not reduced, unlike that of F. punctatus, of Mauritius. The other extinct species, namely NUMBER 89 33 TABLE 13.?Native resident land birds of the Mascarene Islands. L=species still living on the island today, E=species completely extinct, X=species now extinct but known by modem specimens, F=species known from fossils found on the specified island, H=species known by historical accounts (when not known from fossils). (Egretta dimorpha on Mauri tius is after Milne-Edwards (1874, pl. 33: fig. 3). There is no indication that "Necropsittacus" borbonicus belongs to the same genus as N. rodericanus. Hypsipetes (species undescribed) and Timaliinae (genus and species undescribed) on Rod rigues from Cowles (1987). A supposed grebe from Mauritius was based on a fossil of the migratory Whimbrel, Nume nius phaeopus (Cowles, 1987).) Family PHALACROCORACIDAE ARDEIDAE THRESKIORNITHIDAE PHOENICOPTERIDAE ANATIDAE ACCIPITR1DAE FALCONIDAE TURNICIDAE RALLIDAE RAPHIDAE COLUMBIDAE PSITTACIDAE STRJGIDAE APODIDAE HlRUNDINIDAE CAMPEPHAGIDAE PYCNONOTIDAE MUSCICAPIDAE TURDIDAE TlMALIIDAE SYLVIIDAE ZOSTEROPIDAE PLOCEIDAE STURNIDAE Reunion Phalacrocorax africanus? (as "Cormoran"), H Nycticorax duboisi, E,F Egretta dimorpha? (as "Aigrette blanche et grise"), H Threskiornis solitarius, E,F Phoenicopterus ruber, F Alopochen (M.) kervazoi, E,F Anas theodori, E, F Aythya sp., F Circus maillardi, L Falco duboisi (as "Pieds jaunes"), E,F "Emerillons," E,H" Turnix nigricollis (as "Petites perdrix"), L,H Dryolimnas augusti, n. sp. E,F - Porphyrio caerulescens (as "Oiseau bleu"), E,H Fulica newtonii, E, F - Alectroenas sp.? (as "Pigeon couleur d'Ardo- ise"), E,H Nesoenas duboisi, E, F - Streptopelia picturata, L, F - "Perroquet gris," E,H "Necropsittacus" borbonicus (as "Perroquet vert a tete, dessus des ailes et queue couleur de feu"), E,H Mascarinus mascarinus, E,X,F Psittacula eques/echo? (as "Perroquet vert a col lier noir"), H Mascarenotus grucheti, E,F Collocalia francica, L Phedina borbonica, L Coracina newtoni, L Hypsipetes borbonicus, L Terpsiphone bourbonnensis, L Saxicola tectes, L - - Zosterops borbonicus, L Zosterops olivaceus, L Foudia sp. (as "Moineaux"), E,H Fregilupus varius, E,X,F Mauritius Phalacrocorax africanus, F Nycticorax mauritianus, E, F Egretta dimorpha, F - Phoenicopterus ruber, F Alopochen mauritianus, E,F Anas theodori, E,F - Circus alphonsi, E,F Falco punctatus, L,F -- Dryolimnas cuvieri, X,F Aphanapteryx bonasia, E,F - Fulica newtonii, E, F Raphus cucullatus, E,X,F Alectroenas nitidissima, E,X,F Nesoenas mayeri, L, F - Streptopelia picturata, L, F, H Lophopsittacus mauritianus, E, F "Lophopsittacus" bensoni, E, F - Psittacula eques/echo, L, F Mascarenotus sauzieri, E,F Collocalia francica, L Phedina borbonica, L Coracina typica, L Hypsipetes olivaceus, L Terpsiphone bourbonnensis, L - ? - Zosterops borbonicus, L Zosterops chloronothos, L Foudia rubra, L - Rodrigues - Nycticorax megacephalus, E,F - - possibly historical "geant," H - - - - - -- - Erythromachus leguati, E,F - - Pezophaps solitaria, E, F - - "Columba" rodericana, E,F Streptopelia picturata, L, F - - Necropsittacus rodericanus, E,F - Psittacula exsul, E,X,H Mascarenotus murivorus, E,F - - - Hypsipetes, sp. undescribed, E,F - TIMALIINAE, genus and sp. un described, E,F Acrocephalus rodericanus, L - - Foudia flavicans, L Necropsar rodericanus E,F Threskiornis solitarius, Alopochen (M) kervazoi, Anas the odori, Nesoenas duboisi, and Mascarinus mascarinus, do not show reduction in the scapular girdle or wing bones. Alopochen (M) kervazoi shows a very slight reduction compared with the recent African forms, but this also exists in the extinct form from Madagascar. We suggest that this reduction had already occurred in the Malagasy forms before they colonized Reunion. Fulica newtonii shows a reduction of the coracoid and wing bones, but it is the same species as in Mauritius, and it probably colonized Reunion from that island. Mascarenotus grucheti has a more reduced humems than does M. sauzieri, but it is the only exception. Reunion Island dates back 3 Ma (Molnar and Stock, 1987), a length of time amply sufficient for birds to lose their flying 34 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ability in such a way that it can be perceived in the skeleton. For example, the flightless ibises from Hawaii are known only from islands dating to less than 1.8 Ma (Olson and James, 1982). Thus it is necessary to find a cause for this absence of flightless forms. We have good reason to believe that this cause may be found in the volcanic past of the island. Reunion is situated on a hotspot that gave rise to the Deccan traps during the Cretaceous, then the Chagos-Laccadive Ridge, the Mascarene Plateau, and Mauritius (Bonneville, 1990). It is made up of two volcanoes: the oldest is Piton des Neiges, in the northwest, the volcanic activity of which now consists only of thermal springs, whereas the more recent Piton de la Foumaise, in the southeast, is still active. Piton des Neiges is a strato-volcano, made up of hundreds of lava flows spread out on gentle slopes. Built up on a 4000 m deep seafloor, it emerged about three million years ago (m.y.a.), but the oldest dated rocks are 2.1 Ma in age. Its con struction includes two main phases, one with tholeiitic and transitional basalts, dating back 2.1 to 0.43 Ma, and one with differentiated alkaline lavas, dating from 0.35 Ma to less than 30,000 years ago (Kieffer, 1990; Deniel et al., 1992; Kieffer et al., 1993). The alkaline products appeared in the course of very explosive emptions that resulted in several calderas. Piton de la Foumaise appeared above the ocean about one m.y.a. but collapsed several times on its eastern side. After each collapse a new cone appeared farther westward. The first collapse is marked by the caldera of Riviere des Remparts, which is very close to the massif of Piton des Neiges. The second phase of Piton des Neiges, with differentiated al kaline lavas, shows three major explosive episodes. The first one, before 230,000 years ago, emitted pyroclastites that fell down on the whole western flank and probably also on the northern flank. The Quenched Bombs Formation, which be longs to this episode, is found from the Dos d'Ane, in the northwest, to Saint-Pierre, in the south of the Piton des Neiges massif. Thus at this time, the massif must have been almost en tirely covered by pyroclastites. The second one, known as the dalle soudee (welded slab) episode, dates back to 230,000 years. It was produced by huge lava fountains that shot up sev eral km into the air and came down while still fluid, welding after falling. After this explosive event, the center of the volca no collapsed, resulting in a caldera, the diameter of which was eight to 10 km. The third explosive episode, dated back to 188,000 years, produced ignimbrites that flowed down mainly on the eastern flank. Other explosive phenomena occurred sub sequently, but the material emitted remained mainly inside the caldera. Thus, during these explosive volcanic emptions, al most all the massif of Piton des Neiges was covered by incan descent products. In the meantime, the Piton de la Foumaise, which had no explosive episodes, collapsed several times into the ocean. It may be supposed that Reunion Island was colonized by the same birds that colonized Mauritius and Rodrigues, that is, a pigeon ancestor of the Dodo and Solitaire, a rail ancestor of the genera Aphanapteryx and Erythromachus, and a parrot ances tor of the genera Lophopsittacus and Necropsittacus, and that they evolved on Reunion as on the other islands, progressively losing their ability to fly. These flightless or almost flightless forms disappeared during the explosive episodes of the second phase of activity of Piton des Neiges, between 300,000 and 180,000 years ago. Either they disappeared instantaneously or their environment was so depleted that they were not able to survive. The island was colonized again by forms from Africa or Madagascar, such as the ibis, Alopochen, falcon, and night heron, or by forms from Mauritius, such as Anas theodori and Fulica newtonii, and none of these forms had enough time to become flightless. Two genera, however, Mascarinus and Fregilupus, are en demic to Reunion. It is probable that if they had arrived on Reunion after 180,000 years ago, they would not have had enough time to become generically distinct. One may propose the hypothesis that the ancestral forms of these genera arrived on Reunion at a more ancient period and survived the holo caust. The presence of a flightless rail of the genus Dryolimnas is compatible with this hypothesis because we have the example of D. cuvieri aldabranus, which has become flightless, whereas the Madagascan subspecies, D. cuvieri cuvieri, is still able to fly. Aldabra Island has undergone several cycles of emergence and submergence, and its most recent emergence occurred about 80,000 years BP (Braithwaite et al., 1973). The Aldabra White-throated Rail must have arrived on the island after the last emersion, so 80,000 years or less were enough for its skel eton to be considerably modified. The study of bats and of the Reunion land tortoise, Cylin draspis borbonica, does not refute the foregoing hypothesis. The bats include two extinct species, Pteropus niger (Kerr) and Pteropus subniger (Kerr), another that is probably extinct, Sc- otophilus borbonicus (E. Geoffroy) {=S. leucogaster), and two species that are still present, Mormopterus acetabulosus (Her mann) and Taphozous mauritianus E. Geoffroy. All but Scoto- philus once lived, or are still living, on Mauritius (Cheke and Dahl, 1981; Moutou, 1982). Scotophilus borbonicus also oc curs in Africa and Madagascar and probably colonized Reunion in recent times (Cheke, 1987). The genus Cylindraspis, endemic to the Mascarenes and now extinct, is represented on Reunion by a single species, whereas two species are known on Mauritius, and another two are known on Rodrigues. The two Mauritian species, Cylindraspis neraudii (Gray) (=C. inepta; =?C. indica) and Cylindraspis triserrata (Gunther) (=?C. graii), can be distinguished from each other by their skeleton and particularly by their skull, the latter showing a specialization of the triturating surfaces (sup plementary ridges). These two sympatric and synchronous spe cies indicate either a long in situ evolution, with speciation, or two successive colonizations from a common ancestral stock. On Rodrigues the two species Cylindraspis rodericensis (Gunther) (=?C. vosmaeri) and C. pel tastes (Dumeril and Bi- bron) (the smallest form in the genus) exhibit a striking syna pomorphy (predominance of the palatine arterial circulation NUMBER 89 35 over the stapedio-temporal circulation; Bour, 1985), which in dicates that their common ancestor evolved independently from the Mauritian populations before it was subjected to the environmental constraints that led to the present-day situation. On the other hand, the Reunion land-tortoise, although imme diately identifiable by its extremely robust dentary and maxil lary alveolar surfaces, remained close to one of the Mauritian species (C. neraudii) and did not show advanced specializa tions as in the other three species. Therefore it could have colo nized Reunion in a relatively recent period from an ancestor re lated to that Mauritian species. Its distribution on the island, restricted as far as we know to the leeward regions (the western part), reinforces the hypothesis of a recent immigration. It is not possible to know, however, if it reached the island before or after the explosive phenomena of the Piton des Neiges, or if Reunion housed in far-off days a previous population of tor toises that was exterminated together with other representatives of the original fauna. We must remember that the first vestiges of the Bourbon tortoise, their whereabouts unfortunately un known, were found by L. Maillard in 1854, at Cap La Hous- saye, under four meters of lava (Bour, 1980b). Were they the representatives of an ancient population, wiped out a long time ago by a cataclysm, or were they the ancestors of the recently extinct form? As a result of the above considerations, we propose the hy pothesis that most of the birds that were present on Reunion colonized the island only after the explosive episodes of Piton des Neiges. At the time when the Europeans were colonizing the island, the Reunion avifauna included about 33 species of resident land birds (Table 13). Of these 33 species, 14 have been found as fossils. Among those that were certainly present but have not been found as fossils are members of the genera Circus, Alectroenas, Psittacula, Collocalia, and all the small passe rines, Phedina, Coracina, Hypsipetes, Terpsiphone, Saxicola, Zosterops, and Foudia. We think that the small passerines were not present in the fossiliferous sites that we have exploited. Among these 33 species, 17 (52%) are extinct, four or five are no longer present on Reunion, and 11 are still present there. Among the 11 surviving species, eight are very small birds (Apodiformes and Passeriformes), and one species, Coracina newtoni (Pollen), is very endangered. The extinctions took place very rapidly, over a period of two centuries from 1646. A first group of species, reported by the early visitors and by Dubois (1674), in 1671-1672, became ex tinct almost immediately because they are not mentioned after wards. Species that disappeared at that time are Nycticorax duboisi, Alopochen {M.) kervazoi, Anas theodori, Falco duboi si, a smaller falcon known as Emerillon, Dryolimnas augusti, Fulica newtonii, a parrot known as perroquet vert a tete...couleur defeu, and Foudia sp. In Dubois' time the island had only 314 human inhabitants (Bour, 1980a). Rats were ab sent in 1671, as indicated in the log of Le Breton and by Dubois (1674) in 1671-1672, but they invaded the island in 1675 (Cheke, 1987). This first wave of extinction mainly included aquatic forms that were living in the ponds and marshes of the west coast, the area that was first settled. There was no other refuge possible for these species. They were still able to fly but they had become flightless in their behavior. Numerous writers emphasize how tame they were and how easily they allowed themselves to be killed without fleeing. Moreover, anseriforms are very vulnerable during their period of molt. The two fal cons were perhaps adapted for capturing prey living in the dry west coast palm savanna and the western lowland dry forest, environments that were completely cleared for cultivation. Ac cording to Cheke (1987), Foudia was exterminated by rats. The second wave included species mentioned by Feuilley in 1704 (Barre and Barau, 1982) but not recorded thereafter. Spe cies that disappeared at that time are a cormorant (probably Phalacrocorax africanus (Gmelin)), an egret (probably Egretta dimorpha Hartert), Phoenicopterus ruber, Threskiornis solitar ius, and four species of pigeons, Alectroenas sp., Nesoenas duboisi, Streptopelia picturata, and another dove. Flamingos were known to be breeding on the island, and they are vulnera ble during their colonial nesting. The Reunion Ibis, known as "Solitaire," survived for a short time, taking refuge in the mountains, but it was decreasing and was not reported after 1708. Cats were introduced into the island in 1703 to fight rats and must have played a large part in the destmction of birds be cause in 1704 Feuilley wrote: "Ramiers have not been seen for some time, either they have deserted the island, or they have been destroyed by the cats," and, concerning the huppes {Fregilupus) and the merles {Hypsipetes): "Marron cats de stroy many of them. These birds let the cats get very close and they are caught without getting out of their places" (Barre and Barau, 1982:38, our translation). Then, between 1734 and 1740, Oiseau bleu, a grey parrot, and a parakeet, Psittacula eques/echo, disappeared, followed about 1780 by Mascarinus mascarinus, and lastly, between 1838 and 1858, by Fregilupus varius. According to Cheke (1987:24) the Solitaire and Oiseau bleu "were probably victims of feral cats but the parrots may have been dependent on the lowland habitats." It is difficult to evaluate accurately the role played by differ ent factors in the extinction of these birds. But these factors? excessive hunting by humans, the action of introduced preda tors (e.g, pigs, rats, cats), and habitat destmction?cannot ac count for the disappearance of the Reunion Starling, for which the introduction of a disease or parasite has been invoked (Cheke, 1987). The example of Reunion shows that, in an insular environ ment, flying birds disappear almost as quickly as flightless ones. The relief of Reunion is very mgged, much more than that of Mauritius or Rodrigues. It has many places that are dif ficult to reach and where birds may have been protected, but this mgged topography did not prevent the tame, naive birds of Reunion from disappearing. 36 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Literature Cited Andrews, Charles W. 1897. On Some Fossil Remains of Carinate Birds from Central Madagas car. Ibis, 7(3):343-359. Anonymous [=Alfred Newton] 1868. [Review of Recent Literature on Mascarene Birds.] Ibis, series 2, 4:479^182. Barre, Nicolas, and Armand Barau 1982. Oiseaux de La Reunion. 196 pages. La Reunion: Imprimerie Arts Graphiques Modernes. Barre, Nicolas, Armand Barau, and Christian Jouanin 1996. Oiseaux de La Reunion. 208 pages. Paris: Editions du Pacifique. Baumel, Julian J. 1979. Osteologia. In J.J. Baumel, editor, Nomina Anatomica Avium: An Annotated Dictionary of Birds, pages 53-121. London and New York: Academic Press. Berger, Andrew J. 1957. On the Anatomy and Relationships of Fregilupus varius, an Extinct Starling from the Mascarene Islands. Bulletin of the American Mu seum of Natural History, 113(3):225-272. Boddaert, Pietr 1783. Table des planches enlumineez d'histoire naturelle de M. d'Auben- ton. xv+58 pages. Utrecht. Bontekoe, Willem Ysbrantsz 1646 ("1619"). Voyage de Guillaume Ysbrabtsz Boutekou de Hoom [=Willem Ysbrantsz Bontekoe], ecrit par lui-meme. First edition. Holland. [Date on title page is 1619; actually published in 1646.] [Republished, 1725, in Constantin de Renneville, Recueil des voy ages qui ont servi a I 'etablissement et aux progrez de la Compagnie des Indes orientates, New edition, volume 8. Rouen: Pierre Cailloue.] Bonaparte, Charles L. 1850-1857. Conspectus Generum Avium. 232 pages. Lugduni Batavorum [Leiden]: E.J. Brill. Bonneville, Alain 1990. Structure de la lithosphere. In Jean-Francois Lenat, editor, Le Volca- nisme de la Reunion?Monographic pagesl-18. Clermont-Ferrand: Centre de Recherches Volcanologiques. Bour, Roger 1979. Premiere decouverte de restes osseux de la Tortue terrestre de la Reunion, Cylindraspis borbonica. Comptes Rendus de I'Academie des Sciences de Paris, series D, 288:1223-1226. 1980a. Histoire de la tortue terrestre de Bourbon. Bulletin de I'Academie de I'lle de La Reunion, 25:97-147. 1980b. Systematique des Tortues terrestres des Ties Mascareignes: genre Cy lindraspis Fitzinger, 1835 (Reptilia, Chelonii). Bulletin du Museum National d'Histoire Naturelle, Paris, series 4, section A, 2(3):895-904. 1985. Les Tortues terrestres geantes des lies de l'ocean Indien occidental: donnees geographiques, taxinomiques et phylogenetiques. Studia Palaeochelonologica, Salamanca [Studia Geologica Salmanticen- sia, special volume 1], 1:17-76. Braithwaite, C.J.R., J.D. Taylor, and W.J. Kennedy 1973. The Evolution of an Atoll: The Depostional and Erosional History of Aldabra. Philosophical Transactions of the Royal Society of Lon don, series B, 266:307-340. Brial, Pierre 1996. La Caverne de la Tortue. Bulletin de la Societe d'Etudes Scienti- fiques des Cavernes de La Reunion, 1:8-l 1. Brodkorb, Pierce 1963. Catalogue of Fossil Birds: Part 1 (Archaeopterygiformes through Ardeiformes). Bulletin of the Florida State Museum, 7(4): 179-293. 1964. Catalogue of Fossil Birds: Part 2 (Anseriformes through Galli- formes). Bulletin of the Florida State Museum, 8(3): 195-335. Brown, Leslie H., Emil K. Urban, and Kenneth Newmann 1982. The Birds of Africa. Volume 2, xiii+521 pages. London and New York: Academic Press. Carre, Abbe 1699. Voyage des Indes orientales mile de plusieurs histoires curieuses. Two volumes. Paris. Castleton, Samuel 1625. A Joumall of a Voyage Made by the Pearle to the East India Wherein Went as Captain Master Samuel Castleton of London, and Captain Gerorge Bathurst as Lieutenent: Writtten by John Tatton, Master. In S. Purchas, editor, Hakluytus posthumus, or Purchas His Pilgrimes: Contayning a History of the World in Sea Voyages and Lande Travels by Englishmen and Others, 3:343-354. [Reprinted 1905-1907]. Glasgow: James MacLehose and Sons. Cheke, Anthony S. 1982. Les noms Creoles des oiseaux dans les iles francophones de VOcean Indien (essai ethno-ornithologique). 64 pages. Paris: Collection l'Homme et son Milieu, Institut International d'Ethnosciences. 1987. An Ecological History of the Mascarene Islands, with Particular Reference to Extinctions and Introduction of Land Vertebrates. In A.W. Diamond, editor, Studies of Mascarene Islands Birds, pages 5-89. Cambridge: Cambridge University Press. Cheke, Anthony S., and J.F. Dahl 1981. The Status of Bats on Western Indian Ocean Islands, with Special Reference to Pteropus. Mammalia, 45(2):205-238. Cheke, Anthony S., and Carl G. Jones 1987. Measurements and Weights of the Surviving Endemic Birds of the Mascarenes and Their Eggs. In A.W. Diamond, editor, Studies of Mascarene Islands Birds, pages 403?422. Cambridge: Cambridge University Press. Cowles, Graham S. 1987. The Fossil Record. In A.W. Diamond, editor, Studies of Mascarene Islands Birds, pages 90-100. Cambridge: Cambridge University Press. 1994. A New Genus, Three New Species and Two New Records of Extinct Holocene Birds from Reunion Island, Indian Ocean. Geobios, 27(l):87-93. Cramp, Stanley, and K.E.L. Simmons 1979. The Birds of the Western Palearctic. Volume 2, 695 pages. Oxford and London: Oxford University Press. Daudin, Francois Marie 1802. Histoire naturelle par Buffon dediee au citoyen Lacepede. Volume 14. Paris. Deniel, C, G. Kieffer, and J. Lecointre 1992. New 230Th-238U and 14C Age Determinations from Piton des Neiges Volcano, Reunion?A Revised Chronology for the Differentiated Series. Journal of Volcanology and Geothermal Research, 51:253-267. Dubois 1674. Les voyages fails par le Sieur D. B. aux isles Dauphine ou Madagas car et Bourbon ou Mascarenne, es annees 1669, 70, 71 et 72. 234 pages. Paris: Claude Barbin. Forbes, Henry Ogg 1893. A List of the Birds Inhabiting the Chatham Islands. Ibis, 6(5): 521-546. Forshaw, Joseph M. 1973. Parrots of the World. 584 pages. New York: Doubleday and Com pany. Gmelin, Johann Friedrich 1789. Carol i a Linn6... Systema Naturae per regna tria... Editio decima tertia... Cura J.F. Gmelin. Volume 1, part 2:501-1032. Lugduni [Lyon]: J.B. Delamolliere. NUMBER 89 37 Gray, G.R. 1842. Appendix to a List of the Genera of Birds. iv+16 pages. London: R. and J.E. Taylor. 1855. Catalogue of Three Genera and Subgenera of Birds Contained in the British Museum. 192 pages. London: British Museum. Greenway, James C, Jr. 1967. Extinct and Vanishing Birds of the World. Second revised edition, xvi+520 pages. New York: Dover. Gunther, A., and Edward Newton 1879. The Extinct Birds of Rodriguez. Philosophical Transactions of the Royal Society of London. 168:423-437, plates 41-43. Hachisuka, Masauji 1937a. Revisional Note on the Didine Birds of Reunion. Proceedings of the Biological Society of Washington, 50:69-71. 1937b. On the Flightless Heron of Rodriguez. Proceedings of the Biological Society of Washington, 50:145-150. 1953. The Dodo and Kindred Birds, or The Extinct Birds of the Mascarene Islands. 450 pages. London: Witherby Ltd. Houde, Peter, and Hartmut Haubold 1987. Palaeotis weigelti Restudied: A Small Middle Eocene Ostrich (Aves: Struthioniformes). Palaeovertebrata, 17(2):27?42. Howard, Hildegarde 1964. Fossil Anseriformes. In Jean Delacour, The Waterfowl of the World, 4:233-236. London: Country Life Ltd. Hoyo, Josep del, Andrew Elliott, and Jordi Sargatal 1992. Handbook of the Birds of the World. Volume 1, 696 pages. Barce lona: Lynx Edicions. Jones, Carl G. 1987. The Larger Land-Birds of Mauritius. In A.W. Diamond, editor, Stud ies of Mascarene Islands Birds, pages 208-300. Cambridge: Cam bridge University Press. Jouanin, Christian 1987. Notes on the Nesting Procellariiformes in Reunion. In A.W. Dia mond, editor, Studies of Mascarene Islands Birds, pages 358-363. Cambridge: Cambridge University Press. Jouanin, Christian, and Frank B. Gill 1967. Recherche du Petrel de Barau, Pterodroma baraui. L'Oiseau et la Revue Francaise d'Ornithologie, 37:1-19. Kervazo, Bertrand 1979. Fouilles de la grotte dite des "Premiers Francais." Info-Nature, 17:47-52. Kieffer, Guy 1990. Grands traits morphologiques de I'ile de La Reunion (Ocean In dien). In Jean-Francois Lenat, editor, Le Volcanisme de la Reunion? Monographie, pages 75-185. Clermont-Ferrand: Centre de Recher ches Volcanologiques. Kieffer, Guy, Pierre-Yves Gillot, Yves Cornette, Christian Germanaz, and Pierre Nativel 1993. Une phase eruptive exceptionnelle dans l'histoire recente du Piton des Neiges (lie de la Reunion): l'eruption de la "dalle soudee." Comptes Rendus de I'Academie des Sciences de Paris, series 2, 317:835-842. Lambrecht, Kalman 1933. Handbuch der Palaeornithologie. 1024 pages. Berlin: Bomtraeger. Langrand, Olivier 1990. Guide to the Birds of Madagascar. 364 pages. New Haven and Lon don: Yale University Press. Lesson, Rene Primevere 1840 ("1839"). Description de treize oiseaux nouveaux, suivies [sic] de rectifictions sur quelques especes deja publiees. Revue Zoologique, 2:100-104. [Date on title page is 1839; actually published in 1840.] Linnaeus, Carolus 1758. Systema Naturae. Tenth edition, 823 pages. Holmiae [Stockholm]: Laurentius Salvius. 1771. Caroli a Linne Mantissa Plantarum altera Generum editionis VI et Specierum Editionis II; Regni Animalis appendix, pages 144-588. Holmiae [Stockholm]: Laurentii Slavii. Lougnon, Albert 1970. Sous le signe de la tortue; Voyages anciens a File Bourbon (1611-1725). Third edition, 285 pages. Saint-Denis, La Reunion. 1992. Sous le signe de la tortue; Voyages anciens a File Bourbon (1611-1725). Fourth edition, 286 pages. Saint-Denis, La Reunion: Azal6es Editions. Martinez, Joseph 1987. Un nouveau eas probable d'endemisme insulaire: Ie canard de l'ile Amsterdam. In Cecile Mourer-Chauvire\ editor, L'Evolution des oiseaux d'apres le temoignage des fossiles. Documents des Labora- toires de Geologie de Lyon, 99:211 -219. Mayr, Ernst, and G. William Cottrell 1979. Check-List of Birds of the World. Volume 1, second edition, xvii+547 pages. Cambridge, Masschusetts: Museum of Compara tive Zoology. Milne-Edwards, Alphonse 1867. Memoire sur une espece eteinte du genre Fulica qui habitait autre fois l'ile Maurice. Annates des Sciences Naturelles, series 5 (Zoolol- ogie), 8:195-220, plates 10-13. 1867-1871. Recherches anatomiques et paleontologiques pour servir a l'histoire des oiseaux fossiles de la France. 4 volumes. Paris: Victor Masson et Fils. 1874. Recherches sur la faune ancienne des iles Mascareignes. Annates des Sciences Naturelles, series 5 (Zoologie), 19:1-31, plates 11-15. Molnar, Peter, and Joann Stock 1987. Relative Motions of Hotspots in the Pacific, Atlantic and Indian Oceans since Late Cretaceous Time. Nature, 327:587-591. Mourer-Chauvire, Cecile, Roger Bour, Francois Moutou, and Sonia Ribes 1994. Mascarenotus nov. gen. (Aves, Strigiformes), genre endemique eteint des Mascareignes et M. grucheti n. sp., espece eteinte de La Reunion. Comptes Rendus de I'Academie des Sciences de Paris, se ries 2, 318:1699-1706. Mourer-Chauvire, Cecile, Roger Bour, and Sonia Ribes 1995a. Was the Solitaire of Reunion an Ibis? Nature, 373(6515):568. 1995b. Position systematique du Solitaire de la Reunion: nouvelle interpre tation basee sur les restes fossiles et les recks des anciens voy- ageurs. Comptes Rendus de I'Academie des Sciences de Paris, series 2a, 320:1125-1131. Mourer-Chauvire, Cecile, and Francois Moutou 1987. Decouverte d'une forme recemment eteinte d'ibis endemique insu laire de l'ile de La Reunion: Borbonibis latipes n. gen. n. sp. Comptes Rendus de I'Academie des Sciences de Paris, series 2, 305:419-423. Moutou, Francois 1982. Note sur les chiropteres de l'ile de la Reunion (Ocean Indien). Mam malia, 46(1):36-51. Murie, James 1874. On the Skeleton and Lineage of Fregilupus varius. Proceedings of the Zoological Society of London, 1874:474-488. Newton, Alfred 1869. On a Picture Supposed to Represent the Didine Bird of the Island of Bourbon (Reunion). Transactions of the Zoological Society of Lon don, 6:373-376. Newton, Alfred, and Edward Newton 1876. On the Psittaci of the Mascarene Islands. Ibis, series 3, 6:281-289. Newton, Edward, and Hans Gadow 1893. On Additional Bones of the Dodo and Other Extinct Birds of Mauri tius Obtained by Mr. Theodore Sauzier. Transactions of the Zoologi cal Society of London, 13, part 7(1 ):281-302, plates 33-37. North-Coombes, Alfred 1979. La decouverte des Mascareignes par les Arabes et les Portugais. 175 pages. Port-Louis, Mauritius. 38 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Oliver, Samuel Pasfield, editor 1897. The Voyages Made by the Sieur D. B. to the Islands Dauphine or Madagascar and Bourbon or Mascarenne in the Years 1669-70-71 and 72. xxxv+160 pages. London: David Nutt. Olson, Storrs L. 1973. A Classification of the Rallidae. Wilson Bulletin, 85(4):381^U6. 1977. A Synopsis of the Fossil Rallidae. In S. Dillon Ripley, Rails of the World: A Monograph of the Family Rallidae, pages 339-379. Bos ton: David R. Godine. Olson, Storrs L., and Helen F. James 1982. Prodromus of the Fossil Avifauna of the Hawaiian Islands. Smithso nian Contributions to Zoology, 365: 59 pages. 1991. Description of Thirty-two New Species of Birds from the Hawaiian Islands, Part I: Non-passeriformes. Ornithological Monographs, 45: 88 pages. Olson, Storrs L., and Pierre Jouventin 1996. A New Species of Small Flightless Duck from Amsterdam Island, Southern Indian Ocean (Anatidae: Anas). Condor, 98(1): 1-9. Oustalet, E. 1896. Notice sur la faune ornithologique ancienne et modeme des lies Mascareignes et en particulier de l'ile Maurice d'apres des docu ments inedits. Annates des Sciences Naturelles, series 8, (Zoologie), 3:1-128. Owen, Richard 1848. On the Remains of the Gigantic and Presumed Extinct Wingless and Terrestrial Birds of New Zealand (Dinornis and Palapteryx), with Indications of Two Other Genera (Notornis and Nestor). Proceed ings of the Zoological Society of London, 1848:1-11. Peters, James Lee 1937. Check-List of Birds of the World. Volume 3, 311 pages. Cambridge, Masschusetts: Museum of Comparative Zoology. [Reprinted 1961.] Piveteau, Jean 1945. Etude sur VAphanapteryx, oiseau eteint de l'ile Maurice. Annates de Paleontologie, 31:31-37. Poplin, Francois, and Cecile Mourer-Chauvire 1985. Sylviornis neocaledoniae (Aves, Galliformes, Megapodiidae), oiseau geant eteint de Tile des Pins (Nouvelle-Caledonie). Geobios, 18(l):73-97. Rick, Anne M. 1975. Bird Medullary Bone: A Seasonal Dating Technique for Faunal An alysts. Bulletin of the Canadian Archaeologists Association, 7:183-190. Rothschild, Walter 1907. Extinct Birds, xxix+244 pages. London: Hutchinson. Schlegel, H. 1854. Ook een woordje over den Dodo (Didus ineptus) en zijne verwanten. Verslagen en Mededeelingen der Koninklijke Akademie van Weten- schappen, 2:232-256. S61ys-Longchamps, Edmond de 1848. Resume concemant les Oiseaux brevipennes mentionnes dans l'ou- vrage de M. Strickland sur le Dodo. Revue Zoologique, Octobre 1848:292-295. Shelley, G.E. 1883. On the Columbidae of the Ethiopian Region. Ibis, series 5, 1: 258-331. Simkiss, K. 1967. Calcium in Reproductive Physiology: A Comparative Study of Verte brates. 264 pages. London: Chapman and Hall, Ltd. Steadman, David W. 1992. New Species of Gallicolumba and Macropygia (Aves: Columbidae) from Archeological Sites in Polynesia. In K.E. Campbell, editor, Pa pers in Avian Paleontology Honoring Pierce Brodkorb. Sciences Se ries, Natural History Museum of Los Angeles County, 36:329-348. Storer, Robert W. 1970. Independent Evolution of the Dodo and the Solitaire. Auk, 87(2):369-370. Strickland, H.E., and A.G. Melville 1848. The Dodo and Its Kindred, or the History, Affinities and Osteology of the Dodo, Solitaire, and Other Extinct Birds of the Islands Mauri tius, Rodriguez, and Bourbon. 141 pages. London: Reeve, Benham, and Reeve. Temminck, Coenraad Jacob 1813. Histoire naturelle generate des Pigeons et des Gallinaces. Volume 1, pages 1-16, 1-499, errata, 1 plate. Amsterdam: J.C. Sepp et Fils; Paris: G. Dufour. Urban, Emil K., C. Hilary Fry, and Stuart Keith 1986. The Birds of Africa. Volume 2, xvi+552 pages. London: Academic Press. The Fossil Avifauna of Amsterdam Island. Indian Ocean Trevor H. Worthy and Pierre Jouventin ABSTRACT The fossil avifauna of Amsterdam Island is described from 23,562 identifiable bones representing 2060 individuals from 30 sites. Twenty species of seabirds and one land bird are represented by the fossils. This may underestimate the prehuman species rich ness because two of the 10 indigenous species that now breed on the island are not represented among the fossils. Lengths of bones for all common species on Amsterdam are compared with those of populations elsewhere. On Amsterdam, Great-winged Petrel {Pterodroma macroptera (Smith)) and Grey Petrel (Procellaria cinerea Gmelin) populations were composed of individuals of rel atively small mean size. The Macgillivray's Prion (Pachyptila macgillivrayi (Mathews)) is shown to be specifically distinct from the Broad-billed Prion (P. vittata (Forster)) and Salvin's Prion (P. salvini (Mathews)), based on osteological measurements. Introduction Amsterdam (37?50'S, 77?31'E) and St. Paul islands lie 80 km apart in the middle of the Indian Ocean, more than 3,000 km from any continent (Figure 1); both islands are of volcanic origin. Amsterdam is roughly circular, 9.2 km long and 7.4 km wide, and rises to 881 m above sea level (a.s.l.). The western coastal cliffs are spectacularly high (400-700 m), but cliffs are 20-80 m high over most of the remaining coastline. Average air temperature varies between 11.2? C in August and 17? C in February. The climate is windy and humidity is high, with rain fall (annual mean 1114 mm) usually falling as a light drizzle on 239 days of the year (Jouventin, 1994). This high humidity and frequent rain has promoted peat development over much of the island. Prior to human disturbance, lowland areas less than 250 m that were above the coastal cliffs, and where the soil was wet Trevor H. Worthy, Palaeofaunal Surveys, 43 The Ridgeway, Nelson, New Zealand. Pierre Jouventin, Centre d'Etudes Biologiques de Chize, Centre National de la Recherche Scientifique, 79360 Beauvoir surNiort, France. and deep, were covered by a thick, six- to seven-meter-high forest of the tree Phylica nitida (Rhamnacae). In the highlands, and on the central plateau, peatlands were dominated by an as sociation of clubmoss {Lycopodium trichiatum) and a fern {Gleichenia polypodioides), with some grasses, sedges, and forbs {Uncinia brevicaulis, Poa fuegiana, Trisetum insulare, Acaena seurguisarbae, and Scirpus aucklandicus) (Jouventin, 1994; Micol and Jouventin, 1995). HISTORY OF EXPLORATION.?The island was discovered on 18 March 1522, but the first landing was not until 1696. Draw ings made then showed that 27% of the island was forested, but by 1875 forest cover was reduced to 5%, and by 1990 only 12 ha, or 0.2%, remained (Jouventin, 1994). Deforestation was caused by ships stopping en route from South Africa to Austra lia to collect wood, and by repeated fires. Fur seals {Arctocephalus tropicalis (Gray)) were harvested in great numbers from about 1790, but they disappeared by 1893. A small colony was found in 1905 and is making a good recovery (Jouventin, 1994). Other human visitors to the island killed penguins, albatrosses, and petrels for food or for bait for lobster pots. Many animals were introduced. The dog {Canis familiaris Linnaeus), pig {Sus scrofa Linnaeus), and goat {Ca- pra hircus Linnaeus) died out, whereas the house mouse {Mus musculus Linnaeus), Norway rat {Rattus norvegicus (Berken- hout)), cat {Felis catus Linnaeus), and cow {Bos taunts Linnae us) remain (Jouventin, 1994; Micol and Jouventin, 1995). Habitat degradation caused by repeated fires and cattle graz ing, plus predation by other mammals and active hunting by people, combined to decimate the avifauna, which consisted mainly of seabirds. The breeding fauna on Amsterdam at first human contact has been estimated to include at least 20 species (Jouventin, 1994), of which only the following 10 species re main (Micol and Jouventin, 1995) (nomenclature for modem avifauna discussed herein follows Marchant and Higgins, 1990). The Northern Rockhopper Penguin {Eudyptes chryso- come) and the Yellow-nosed Albatross {Diomedea chlororhyn- chos) are common. The Sooty Albatross {Phoebetria fused) is rare, with about 240 pairs, and only about 20 breeding pairs of the endemic Amsterdam Albatross {Diomedea amsterdamen- 39 40 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Mauritius Reunion Pr. Edward I Marion I Crozet Is ha ._ Amsterdam I X St. Paul I Kerguelen Is Heard I AUSTRALIA FIGURE 1.?Location of Amsterdam Island and St. Paul Island in the Indian Ocean. NUMBER 89 41 sis) survive. Both the Soft-plumaged Petrel {Pterodroma mol lis) and the Grey Petrel {Procellaria cinerea) are now very rare and are believed to be breeding on cliffs away from cattle tram pling. Macgillivray's Prion, listed as Pachyptila salvini macgillivrayi by Micol and Jouventin (1995), is rare and en dangered with 100-200 pairs, and the Diving Petrel {Pele- canoides sp.) is very rare, with only one record of a breeding pair. The Antarctic Tern {Sterna vittata tristanensis) breeds on coastal cliffs, and a few Brown Skua {Catharacta skua hamil- toni) nest inland (Jouventin, 1994; Micol, 1995). Fortunately, St. Paul has a small rock stack, Roche Quille, just offshore, that has remained predator free, allowing the pre carious survival of the following species of procellariiforms that no longer breed on Amsterdam or on St. Paul: Fairy Prion {Pachyptila turtur, 10-20 pairs), Great-winged Petrel {Ptero droma macroptera, 40-60 pairs), and Little Shearwater {Puffi nus assimilis, 25 pairs) (Micol, 1995). The Flesh-footed Shear water {Puffinus carneipes), previously reported from Roche Quille (Tollu, 1984; Jouventin, 1994), is absent there now, but it is present on St. Paul, where Micol (1995) reported 532 pairs. The White-bellied Storm-petrel {Fregetta grallaria) survives on both Roche Quille and St. Paul in small numbers, and Wil son's Storm-petrel {Oceanites oceanicus) has been found breeding in small numbers on St. Paul (Micol, 1995). Roche Quille has the only other known population of Macgillivray's Prion (100-200 pairs) apart from that on Amsterdam. The survival of several species on Roche Quille and not on St. Paul or Amsterdam suggests that several species of storm petrels, prions, shearwaters, and petrels have disappeared from Amsterdam. Of the remaining 10 breeding species, eight are now rare. Apart from the loss of several procellariiforms from Amsterdam, the only known native terrestrial species, a minute, endemic, flightless duck, Anas marecula (Bourne et al. 1983; Martinez, 1987; Olson and Jouventin, 1996), is extinct. PREVIOUS INVESTIGATIONS OF THE FOSSIL FAUNA.?The volcanic origin of Amsterdam Island has resulted in the forma tion of many lava caves, the collapsed roofs of which have in many places formed pitfall traps into which birds have fallen. Fossil bones have been found to be abundant in these sites. The first fossils were collected in 1955 by Jouanin and Paulian (1960), who identified several bones from one individual as a Wandering Albatross {Diomedea exulans). They noted two siz es of Pterodroma (listed as Bulweria): they identified the larger one as P. neglecta, primarily on the basis of size, and the small er one as P. mollis. Other species identified included Puffinus assimilis, Pachyptila vittata, Pelagodroma marina, Pele- canoides urinatrix, and Anas sp. Lastly, they reported a mum mified rail that "crumbled to dust," which they tentatively re ferred to Crex crex (Linnaeus). Bourne et al. (1983) examined the duck bones reported by Jouanin and Paulian (1960) and suggested that they had some similarity to those of a Garganey {Anas querquedula). Martinez (1987) briefly described the collection of fossils that he made in 1983 and 1984. He listed 17 species: one albatross, two Pro cellaria, three Pterodroma, two Puffinus, two Pachyptila, a Pelecanoides, three storm petrels, a Eudyptes, a Catharacta, and a flightless duck whose bones he described and compared to other ducks, but which he did not name. This collection is the subject of this paper. The fossil sites are all lava caves (Figure 2) that vary in length from a few meters to about 200 m (Figures 3, 4). No ra diocarbon dates for the fossils are available, and without dating each site these would not establish the relative ages of all sites. In their absence, however, we make the observation that all bones were found on the surface, and many had organic re mains, including dried tissue and feathers on them, suggesting that most are between a few hundred and a few thousand years old. The material was collected without retention of skeletal asso ciations, and, later, Martinez made a preliminary sort of it into species by element. After most Anas and Diomedea bones were removed for independent analysis, the collection was shipped to the Museum of New Zealand Te Papa Tongarewa (MNZ) (formerly National Museum of New Zealand) for analysis. We describe herein the fossil fauna from Amsterdam Island, detailing the species represented, their relative abundance, and their size (using bone lengths) relative to that of modem popu lations. Extensive comparative descriptions and mensural com parisons were necessary to justify our specific determinations. ACKNOWLEDGMENTS We thank the Institut Francais pour la Recherche et Technol ogic Polaires and the administration of Terres Australes et Ant- arctiques Francaises for providing logistic and financial sup port. We are grateful to J. Martinez and other researchers for collecting the material and sketching maps of the caves, to J. Palmer for photography, and to J.A. Bartle, MNZ, for encour agement, support, and advice during the course of the project. The following curators generously made available specimens in their care: J. Wombey, Australian National Wildlife Collec tion, CSIRO, Canberra, Australia; J. Bailey, The Natural Histo ry Museum (BMNH), Tring, England; and R. Coory and J.A. Bartle, MNZ, Wellington, New Zealand. We thank A. Tenny son for commenting on an earlier draft of the text, and S. Emslie, D. Steadman, and S. Olson for substantive comments that greatly improved the text. METHODS The Martinez collection is cataloged under numbers MNZ S34560-S35079 in the Museum of New Zealand Te Papa Ton garewa. One of us (THW) examined it in 1994 and 1995 and is responsible for all identifications, measurements, and compari sons. Material that remained in France, including the majority of the albatross and duck bones, was not included in the analy sis. The duck bones are now at the National Museum of Natu ral History, Smithsonian Institution (USNM, housing the col- 42 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 2km FIGURE 2.?Distribution of fossil sites on Amsterdam Island; numbers indicate locations of lava caves (see Figures 3, 4). Site 1 > | 6- u. 4- 2- 0 ? ? u ? ill 38 39 40 41 42 43 44 45 Length mm FIGURE 7.?Histograms of lengths of Pterodroma macroptera bones from Amsterdam Island. measured across the lacrimals, crania of these two species dif fer as follows. The temporal fossae in P. baraui are anteropos teriorly much broader than those of P. arminjoniana (9.0-9.4 mm (H=3) vs. 6.8 mm), and the width across the zygomatic processes in P. baraui (Table 2) is greater (28.2-29.7 mm vs. 26.8 mm). In both P. baraui and P. arminjoniana the os lacri mal has a robust ventral portion with a facet that is longer than high, articulating with the quadratojugal. This facet is much TABLE 2.?Measurements (mm) of modern specimens of Pterodroma baraui. Abbreviations are defined in the text. Catalog number MNZ 23831 Paris, number 2603 Paris, number 2601 Paris, number 772 Fern L 33.4 34.93 33 32.88 TibAL 58.4 63.3 59.17 59.44 TmtL 38.3 39.8 37.34 36.82 Hum L 96.5 98.51 97.4 92.04 UlnaL 99.7 102.96 98.75 96.18 Cmc L 49.2 50.27 48.08 47.72 CorL 26.9 28.08 26.86 27.7 Skull TL 83.6 - 82 80.2 LacW 25.8 - 25.7 24.6 POW 32.9 - 31.24 31.08 ZPW 29.7 _ 28.5 28.26 PmxL 43 _ 41.2 40.5 PmxW 17.4 17.3 16.94 50 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Humeri Femora Si 5 ?i.llllll. 73 75 77 79 81 83 85 Length mm 15- 10- JUV 26.5 27.5 28.5 29.5 30.5 31.5 Length mm 15i 10- 5- Ulnae I.I.I 73 75 77 79 81 83 85 Length mm a 4 I Tibiotarsi ? -III. ? i 48 50 52 54 56 Length mm 20 J 15 I 10 f 5 Carpometacarpi ???????I.I i? i? i ? i r*~i???i 36 37 38 39 40 41 42 15-, g 10 o 3 | 5 u. 0 Tarsometatarsi ? -Mil.. 31 32 33 34 35 36 Length mm FIGURE 8.?Histograms of lengths of Pterodroma mollis bones from Amsterdam Island more robust than in the similar-sized P. neglecta. Pterodroma baraui differs from P. arminjoniana in that the ventral process of the os lacrimale, where it descends from the os ectethmoi- dale, is much shorter, with a different shape of the articular fac et. The lower fonticulus orbitocranale is more elongate in P baraui {n=3, 5.8 mm high x 9.1 mm long, 6.0 x 8.6 mm, 6.0 x 8.8 mm, vs. 6.5 x 7.7 mm in P. arminjoniana). The fossil crani um from site 8.11 has a basicranial length of 37 mm but is too broken to measure further; however, the extensive temporal fossae, lacrimal shape, and shape of the fonticulus orbitocrani- ale clearly identify it as P baraui. The posterior portion of the cranium from site 8.3 (width at zygomatic processes 25.5 mm) has wide temporal fossae like P. baraui and is referred to that species. The postcranial bones in P. baraui and P. arminjoni ana are very similar, and subtle differences, if present, between them would not be of use in identifying fossil material that is worn, weathered, or broken. Because all cranial material from site 8 is referred to P. baraui, all the postcranial bones of mid sized Pterodroma from this site also are referred to that species (Table 3). This is further supported by the tarsometatarsi from site 8, which are similar to those of P. baraui (Table 4) and are shorter and stouter than those of P. macroptera. TABLE 3.?Summary statistics for bones attributed to fossil Pterodroma baraui and measurements (mm) of intermediate-sized fossil Pterodroma species, Amsterdam Island. Species P. baraui mean standard deviation minimum maximum sample size P. arminjoniana/ baraui site 18.1 site 18.27 site 18.26 P. arminjoniana site 18.2 site 18.8 site 18.11 Fern L 32.87 0.05 32.83 32.9 2 - - - -- - TmtL 38.14 0.06 38.1 38.18 2 - - - -- - HumL 91.82 1.82 90.4 94.32 4 - 95.4 95.8 - 89.7 91.5 UlnaL 97.42 1.99 95 102.5 13 96.8 - - 97.3 - - Cmc L 47.62 1.36 46 50.1 9 - - - -- - CorL 26.07 0.42 25.6 26.43 3 - - - -- - Pterodroma sp. Two premaxillae (MNZ S34852, MNZ S34859) are very short and stout and are not referable to any of the above spe cies. Bourne (1968) gave measurements for a premaxilla of the NUMBER 89 51 TABLE 4.?Measurements (mm) of fossil tarsometatarsi of Pterodroma baraui compared with modem specimens. Catalog number Fossil site 8.8 site 8.5 Modem MNZ 23831 Paris, number 772 Paris, number 2601 Paris, number 2603 Length 38.2 38.1 38.2 36.8 37.3 39.8 Proximal width 6.3 _ 6.8 6.3 6.9 6.4 Shaft width 2.8 3.2 2.8 2.8 3.2 2.8 Distal width 6.1 6.6 6.7 5.8 6.3 6.2 Mascarene Petrel {Pterodroma aterrima) that compare well with those of the more complete specimen (MNZ S34852, measurements given second): length from nostrils 20 mm vs. -22 mm; length of nostril 10 mm vs. 12 mm; length to angle gape 34 mm vs. 35 mm; depth at proximal end of nostril 11 mm vs. 10 mm. The width of MNZ S34852 at the angle gape is 16.5 mm. Pachyptila macgillivrayi FIGURES 9,10 One species of prion that now breeds on Amsterdam Island is currently referred to Pachyptila salvini macgillivrayi (Roux et al., 1986; Jouventin, 1994; Micol and Jouventin, 1995), al though previously it was referred to as a subspecies of the Broad-billed Prion (P vittata) (e.g., Jouanin, 1953; Paulian, 1960; Roux and Martinez, 1987), presumably because of its similarly broad bill. It was allied with Salvin's Prion (P. salvi ni) on account of its similar size, its breeding cycle (which is more aligned with that of P. salvini), and its blue bill, as op posed to the "steel-grey bills" of P. vittata (Roux and Martinez, 1987). Prion bones are very numerous in most sites. Their measure ments exhibit an apparently normal, unimodal size distribution (Figure 10, Appendix 3), so we refer the bones in this size range to a single taxon with which we associate the numerous cranial remains characterized by very broad bills that are found in the same sites. The bones of this Amsterdam prion cannot be referred to either P. vittata or P. salvini for the following rea- FlGURE 9.?Photographs of Pachyptila skulls in dorsal (left) and ventral (right) views. A,C, modem P. vittata, MNZ 12667, New Zealand; B,D, fossil P. macgillivrayi from site 18.7. Scale bar=5 cm. 52 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Humeri Femora 52 40-i 5* 30 I 20 f 1. 54 56 58 60 Length mm Ulnae 1 111 52 54 56 58 60 Length mm 62 15 10 f 5 Length mm Tibiotarsi .Mil. 45 47 49 51 53 Length mm 30 i Carpometacarpi 15-i 10 4> 3 Sf 5 Tarsometatarsi l.llllll Length mm Length mm FIGURE 10.?Histograms of lengths of Pachyptila macgillivrayi bones from Amsterdam Island. i ? i i i 30 31 32 33 34 35 36 sons. Adult cranial and long bones of the Amsterdam prion were compared with those of the Broad-billed Prion {Pachypti la vittata) and Salvin's Prion (P. salvini) from the New Zealand region (Table 5; Appendix 3). The Amsterdam species is significantly smaller than P. vittata {t test, P<0.001) in all measurements. With the exception of lacrimal width, which is narrower (P<0.001), there is no significant difference in size between the Amsterdam prion and P. salvini, based on the postcranial and cranial measurements; however, although the length of premaxillae of P. salvini and the Amsterdam prion are not significantly different, those of Salvin's Prion are nar rower (P<0.001), and the premaxillae width/length ratio is sig nificantly different (P<0.001). In contrast, although premaxil lae of P. vittata are longer and wider (P<0.001), the width/ length ratio is not significantly different from that of the Am sterdam prion. Therefore, whereas the Amsterdam prion is of a size similar to the Salvin's Prion, it has a significantly narrower lacrimal width, and its bill is absolutely and relatively much wider than that of P. salvini. The Amsterdam prion's bill is rel atively as wide as that of P. vittata. Because of these differenc es we think that the Amsterdam prion should be removed from P. salvini and reinstated as a distinct species, Macgillivray's Prion (P. macgillivrayi). Pachyptila desolata A single premaxilla identical to that of the Antarctic Prion (P. desolata) is present, but unfortunately it has no site data with it. Pachyptila desolata has not previously been recorded from Amsterdam Island or St. Paul Island, but it breeds on Crozet, Heard, and Kerguelen islands (Jouventin et al., 1984; Marchant and Higgins, 1990). Pachyptila turiur A few postcranial elements below the size range of P. macgillivrayi are present in some sites. These are referred to the Fairy Prion (P. turtur), which breeds on Roche Quille (<10 pairs) (Roux and Martinez, 1987; Micol, 1995). NUMBER 89 53 TABLE 5.?A comparison of the significance of the difference between means of various measurements for the Amsterdam Island prion and for Pachyptila vittata and P. salvini by /-tests (Microsoft Excel, ver. 5), assuming unequal variances. Shown are /-statistic, degrees of freedom, and significance level (P>0.05=NS, P<0.001 = ***). Measurement Femur length Tibiotarsus length Tarsometatarsus length Coracoid length Humerus length Ulna length Carpometacarpus length Premaxilla length Premaxilla width Premaxilla width/length Amsterdam prion vs. P. vittata -11.381, 29, *** -7.956, 38, *** -8.678, 30, *** -8.163, 21, *** -14.727, 24, *** -15.331, 21, *** -12.015, 21, *** -10.446, 27, *** -6.060, 17, *** -1.210, 20, NS Amsterdam prion vs. P. salvini -1.36, 11, NS -0.97, 11, NS -0.38, 11, NS -1.21, 11, NS -1.56, 10, NS 0.22, 9, NS -0.49, 10, NS 0.527, 9, NS 4.987, 8, *** 8.773, 18, *** Procellaria cinerea FIGURES 11,12 Grey Petrels {Procellaria cinerea) are regularly seen off shore around Amsterdam and breed in small numbers on the main island (Roux and Martinez, 1987; Jouventin, 1994; Mi col, 1995). White-chinned Petrels (P. aequinoctialis) are regu lar visitors offshore but are not known to breed on Amsterdam (Roux and Martinez, 1987). The Procellaria crania indicate a single species that is identi cal in form to P. cinerea from New Zealand but is generally smaller (Figure 12; Appendix 4), except for a female (MNZ 24432) that is very similar in size. No specimens of P. cinerea from the Indian Ocean were available for comparison. Bones of P. cinerea from the New Zealand region are similar in size to those of the Black Petrel (P parkinsoni) and are smaller than those of P. aequinoctialis or the Westland Petrel (P. westlandica) (Worthy and Holdaway, 1993). Because the Amsterdam Procellaria is even smaller than P. cinerea from New Zealand it is not referable to P. westlandica or P. aequi noctialis. Humeri of P. parkinsoni are much more gracile than those of P. cinerea. Skulls of P. cinerea differ from those of P. parkinsoni as follows: the premaxilla is about 5-6 mm longer in the region anterior to the nares in P. cinerea, although both have the same depth and width at the anterior end of the nares; the gap between the temporal fossae is narrower in P. cinerea than in P. parkinsoni; and the dorsal margins of the fossae are subparallel, in contrast to widely diverging in P. parkinsoni; on the posterior margin of the temporal fossae above the tympanic cavity there is a fossa that is small in P. cinerea (1-2 mm) but relatively large in P. parkinsoni {2-A mm); the lateral process of the os ectethmoidale that abuts the os lacrimale is much more squared in P. cinerea, in contrast to rounded in P. parkin soni; and the orbitocranial fonticulus is longer and narrower (dorsoventrally) in P. cinerea than in P. parkinsoni. The fossil crania of Procellaria are like P. cinerea in all these respects. Puffinus assimilis FIGURE 13 The majority of the bones attributable to Puffinus are readily identifiable as Little Shearwaters (P. assimilis) because of their small size (Appendix 5). This species breeds in small numbers on Roche Quille (Tollu, 1984; Roux and Martinez, 1987; Mi col, 1995). Puffinus carneipes Twenty-one subadult bones of an individual much larger than Puffinus assimilis were in sites 15.4 and 15.5. The size of the bones, the shape of the skull and premaxilla, and the unusu al features (for Puffinus) of having a humems with a rounded shaft with a deep fossa M. brachialis and a femur with little dorsoventral curvature, allowed ready identification of these bones as those of a Flesh-footed Shearwater {Puffinus car neipes). This species breeds on St. Paul (Micol, 1995). Puffinus griseus A sacmm from site 18.11 and an imperfect right femur from site 18.7 were identified by their size and shape as a Sooty Shearwater {Puffinus griseus), matching the comparative series well. This species has been seen regularly around Amsterdam (Roux and Martinez, 1987). Pelagodroma marina At least three species of storm-petrels are represented. The bulk of the leg bones are readily identifiable as those of White- faced Storm-petrels {Pelagodroma marina). Fregetta grallaria Several very stout tarsometatarsi are immediately referable to Fregetta. White-bellied Storm-petrels {F. grallaria) have been seen, rarely, off Amsterdam, and they breed in low num bers on Roche Quille and St. Paul (Micol, 1995). Although Black-bellied Storm-petrels {F. tropica) do not breed in the ar ea, they have been seen at sea nearby (Roux and Martinez, 1987). Compared to Pelagodroma marina, tarsometatarsi of Freget ta spp. are relatively wide for their length, are absolutely short er, with the distal ends twisted medially (vs. not twisted), and have a convex posterior distal surface (vs. bounded by ridges and concave). Tarsometatarsi of F. tropica are relatively nar rower than in F. grallaria (Appendix 6). The Amsterdam bones are most similar to F. grallaria, and so we refer them to that species. Because other skeletal elements of P. marina and F. grallaria are of similar size, the following distinguishing char acters are listed. Femora of F. grallaria are longer, are not as dorsoventrally curved, and have a deeper fossa poplitea than in P. marina. Fregetta grallaria and F. tropica have ulnae with a prominent ridge descending from under the facies articularis 54 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 11.?Photographs of Procellaria cinerea skulls in lateral (top) and dorsal (lower) views. A,D, fossil spec imen from site 5.2; c, fossil specimen from unrecorded site; B,E, modem specimen MNZ 16486, New Zealand region. Scale bar=5 cm. NUMBER 89 55 Femora Humeri !> f 2 :i ?II ??! I ? 44.5 45.5 46.5 47.5 48.5 Length mm Tibiotarsi c 3 ? I 2- 1 - ? ll.l. 86 88 90 92 94 124 126 128 130 132 134 136 Length mm Ulnae ? ?II Uln Length mm Tarsometatarsi 122 124 126 128 130 132 Length mm Carpometacarpi 3-i i 2 3 ? 1 LU ll 1 5 * 4 o I 3 I 2 "? 1 r^"-i i 56 57 58 59 60 61 ? 111 di 59 60 Length mm 61 62 63 Length mm 64 65 FIGURE 12.?Histograms of lengths of Procellaria cinerea bones from Amsterdam Island. FIGURE 13.?Photographs of Puffinus assimilis skulls in dorsal (left) and lateral (right) views. A,D, modem spec imen of P. a. elegans, MNZ 21865, Antipodes Island; B,C, fossil specimen from site 18.2. Scale bar=5 cm. 56 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY radiocarpalis (not so in Pelagodroma). Humeri of P. grallaria are longer, and the tuberculum dorsale is broader, not as raised above the margo caudalis, and not as elongate distally as those of P. marina. In F. grallaria the tuberculum ventrale is as high as wide rather than higher than wide as in P. marina. On the cranial surface, the sulcus at the base of the crista deltopectora- lis is shallower than in P. marina. The sulcus ligamentis trans versus ends ventrally beside a shallow sulcus on the ventral margin of the crista bicipitalis in F. grallaria, but not in P. ma rina, where there is no shallow sulcus. Oceanites sp. A few bones of a very small storm-petrel are smaller than those found in any genus except Oceanites. Comparisons were made with Grey-backed Storm-petrels {Oceanites nereis) and Wilson's Storm-petrels {O oceanicus). The few measurements available suggest O nereis is smaller, but although lengths of the tarsometatarsi overlap {O. nereis, mean=33.47 mm, range=31.1-34.6 mm, n=4; O oceanicus, mean=35.21 mm, range=33.0-37.2 mm, n=5), those of the ulnae do not {O. nereis, mean= 17.67 mm, range= 16.7-18.4 mm, ?=4; O oce anicus, mean= 19.73 mm, range= 19.1-20.4 mm, n-5). Ulnae of two of the fossil specimens are 18.1 mm and 18.7 mm long, suggesting that they may belong to the smaller taxon. Ocean ites oceanicus is regularly seen offshore, and a specimen of O o. parvus was collected ashore on Amsterdam (Roux and Mar tinez, 1987). In 1995 a small colony of O. oceanicus was found breeding on St. Paul (Micol, 1995). Because the specimens of O. oceanicus measured were mainly from Heard Island, they may be bigger than birds breeding on more northern islands (see discussion in Marchant and Higgins, 1990), so it is possi ble that O oceanicus from the Amsterdam group could be sim ilar in size to O nereis. The fossils are referred only to Ocean ites sp. Pelecanoides urinatrix Bones of a diving-petrel are present in many sites. At present, two diving-petrels are found in the Indian Ocean. The smaller South Georgian Diving-petrel {Pelecanoides geor- gicus) breeds on subantarctic islands, and the larger Common Diving-petrel (P. urinatrix) has a more widespread distribu tion, between 35?S and 55?S (Marchant and Higgins, 1990). Both exhibit size variation that is to some extent clinal, with southern populations being larger (Marchant and Higgins, 1990), but this is not the case in New Zealand (Appendix 7). The fossils, particularly the ulnae and carpometacarpi, are larger than bones of P georgicus from Heard Island, smaller than bones of P. urinatrix exsul from both Auckland and Heard islands, and smaller than bones of P. u. urinatrix from beaches around Cook Strait in New Zealand (Appendix 7). They are, however, similar in size to P. u. chathamensis from Southeast Island in the Chatham Group, and so they are referred to P. uri natrix. There are two records of P. urinatrix from Amsterdam (Roux and Martinez, 1987; Micol and Jouventin, 1995). Catharacta skua Skua bones were recovered from several sites, but most were in poor condition. They are much smaller than those of the Subantarctic (Brown or Southern Great) Skua {Catharacta skua lonnbergi) from the New Zealand subantarctic islands and are bigger than than those of the Antarctic Skua (C. maccor- micki) (Appendix 8). The carpometacarpus is longer than the tarsometatarsus in C. maccormicki, in contrast to C. skua lonn bergi, in which the opposite is true. Unfortunately, none of the fossils were complete enough to use this feature. The Tristan Skua (C. skua hamiltoni), however, for which no comparative material was available, is smaller than C. skua lonnbergi and breeds on Gough Island and in small numbers on Amsterdam (Micol, 1995). Because the fossils are markedly smaller than bones of C. skua lonnbergi but are bigger than bones of C. maccormicki, it seems probable that the Amsterdam skua fos sils are referable to C. skua hamiltoni. Anas marecula The bones of a small duck found in the collection were obvi ously missed when duck material was extracted for the study that resulted in the description of a new species, Anas marecu la, by Olson and Jouvenin (1996) (see Table 1). The series they analyzed, now in the USNM, was composed of at least 33 indi viduals. Those listed herein are almost certainly parts of the same individuals. Discussion COMPOSITION OF THE FOSSIL FAUNA We consider some records based on fossils from Amsterdam Island to be of questionable validity; these are as follows: 1. Wandering Albatross {Diomedea exulans). The record of this species is of bones of a single individual identified by Jouanin and Paulian (1960) before D. amsterdamensis was de scribed, and which Jouventin et al. (1989) reported to be of similar size to the bones they referred to D. amsterdamensis. 2. Kermadec Petrel {Pterodroma neglecta). Jouanin and Paulian (1960) identified this species from a few bones that were smaller than those herein described of Pterodroma mac roptera and bigger than those of P. mollis. The given lengths for the bones are in the size range of P. baraui and P. arminjo niana. Pterodroma baraui was only described in 1964, and nei ther of these two species was compared to the fossils. Pending reexamination of the bones, the record of P. neglecta from Am sterdam is suspect. 3. Broad-billed Prion {Pachyptila vittata). Jouanin and Paulian (1960) identified the prion bones they had as this spe cies and thought it probable that they were of the subspecies P. NUMBER 89 57 vittata macgillivrayi. Data herein show that P. vittata is distinct from P. macgillivrayi, and both are distinct from P. salvini, so the Amsterdam prion should not be listed as Pachyptila vittata. The 20 species of seabird recorded herein as fossils from Amsterdam Island underestimate the total because the Sooty Albatross {Phoebetria fusca) and the Antarctic Tem {Sterna vittata) presently breed on Amsterdam (Jouventin et al.,1984) but are not represented among the fossils. Micol and Jouventin (1995) noted that at least one Phoebetria fusca had been identi fied in fossil material, but no material was seen by THW to substantiate this. Jouventin (1994) and Micol and Jouventin (1995) reported the supposed presence of two extinct species of Pterodroma and two extinct storm-petrels. This study finds no evidence for any extinct procellarid having previously existed on Amster dam Island. Procellaria cinerea, Pterodroma macroptera, and P. mollis are common as fossils and undoubtedly bred there. The large number of P. baraui bones from site 8 suggests this species also was breeding on Amsterdam; in contrast, the few bones of P. arminjoniana could be from nonbreeding visitors, or possibly from skua kills. SIZE RANGES OF SPECIES Pterodroma macroptera: Several species of seabirds repre sented in the fossil fauna of Amsterdam Island are smaller than conspecific populations in the New Zealand region. Data in Appendix 1 show that P. macroptera from Amsterdam are smaller than Australasian and South Atlantic specimens. The specimens available from Eclipse Island in Western Australia are of similar size to the Amsterdam specimens, but the one specimen from Coffin Island is larger than any Amsterdam specimen and is within the size range for P. macroptera gouldi. Measurements given in Marchant and Higgins (1990) show that P. m. macroptera is a little smaller than P. m. gouldi. They also show that males are slightly larger than females in most measurements, but there is no detectable dimorphism in the fossil sample, which has apparently normal, unimodal, size dis tributions. Pterodroma mollis: The limited evidence suggests that P. mollis from Amsterdam is smaller than Australasian and Atlan tic birds. Measurements in Marchant and Higgins (1990), how ever, indicate no geographical or sexual size variation. The fos sil samples have mainly unimodal size distributions (Figure 8) except for humeri, where the distribution is bimodal, suggest ing that there may be some sexual dimorphism. Procellaria cinerea: The Amsterdam Procellaria cinerea have bone lengths 4%-6% smaller than birds from the New Zealand region. Measurements of birds from the Crozet and Kerguelen islands and from New Zealand (Marchant and Hig gins, 1990) also indicate that Indian Ocean birds are smaller than New Zealand ones. Some sexual dimorphism is apparent in external measurements in Marchant and Higgins (1990), al though their two data sets are contradictory: in one the males are larger and in the other the females are larger. Measurements of sexed skeletons in our comparative series (3 males, 7 fe males) suggest that females are smaller, although the small sample size precludes meaningful statistical comparison. The fossil bones have a length distribution (Figure 12) that trends in most cases toward bimodality, compared to the unimodal (ap parently normal) distributions of Pterodroma macroptera and Pachyptila macgillivrayi (Figures 7, 10). If the modem sample is representative, then the larger bones probably represent males. Pachyptila macgillivrayi: The differences detailed above show that the Amsterdam prion is a distinct species, P. macgillivrayi, that is endemic to the Amsterdam-St. Paul group. The very large samples of long bones are apparently normally distributed and unimodal, suggesting there is no sexu al size dimorphism. Sexual dimorphism in prions is generally slight, although Genevois and Bretagnolle (1995) found male Thin-billed Prions (P. belcheri) to be larger overall and to have larger bills than do females. Puffinus assimilis: There are several subspecies of P. assi milis in the southern oceans region. The nominate race from Norfolk Island (P. assimilis assimilis) is the smallest (Appen dix 5). The two New Zealand subspecies, P. a. haurakiensis and P. a. kermadecensis, are of similar size and are larger than P. a. assimilis but are smaller than P. a. elegans, from farther south in the Antipodes (Appendix 5). None of the beach-cast specimens from New Zealand west-coast beaches are large enough to be P. a. elegans, and they are considered to be P. a. kermadecensis (J.A. Bartle, pers. comm., 18 June 1995; veri fied in many cases by the species of lice present). Size varia tion therefore appears to be clinal, with larger birds in the south. Puffinus assimilis bones from Amsterdam are on average slightly smaller than those of P. a. elegans from the Antipodes but are much larger than those of the other subspecies consid ered above, and so they are probably referable to P. assimilis elegans. This subspecies ranges from the Antipodes and Chatham islands in the New Zealand region to Tristan da Cun ha and Gough Island (Marchant and Higgins, 1990). Pelecanoides urinatrix: As discussed above, there is con siderable variation in mean size of individuals between popula tions in this species. Although some of the variation may be clinal, with more southern populations comprising larger indi viduals, this is not the case in New Zealand, where populations in Cook Strait and farther north are distinctly larger than those around Stewart Island and on the Chatham Islands. The limited data presented herein suggest that the Amsterdam birds are smaller than those from Heard Island, significantly smaller than those from the lower latitude populations in the Cook Strait-Wellington region of New Zealand, and similar to those from slightly higher latitudes on Southeast Island, in the Chatham Group. Size variation in Pelecanoides is thus not well explained by clinal factors. 58 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY COMPARISONS WITH OTHER ISLANDS The 20 species of seabirds recorded as fossils from Amster dam Island do not include Phoebetria fusca or Sterna vittata, which currently breed there, suggesting that up to 22 species had breeding populations. Species that are very rare in the fos sil record, however, such as Pachyptila desolata, Puffinus car- neipes, and P. griseus (each with one individual), are likely to have been nonbreeding vagrants or the result of skua kills, so a more realistic estimate is that no more than 19 species of sea- birds bred on Amsterdam Island in the recent past. Amsterdam Island is similar to the Crozet and Kerguelen is lands in that most breeding species are seabirds; however, with at least 34 and 30 breeding species of seabirds, respectively, these subantarctic island groups have far richer communities than Amsterdam did (Jouventin et al., 1984). Although the fos sil faunas of these islands have not been studied, it is likely that neither island group has suffered species extinctions as hap pened on Amsterdam because some islands in each group have remained free of predators (Jouventin et al., 1984). The Mascarene islands Mauritius and Rodriguez are well known for their extinct, endemic land birds (Newton, 1888; Hachisuka, 1953; Gill, 1967), but the fossil seabirds have largely been ignored (Bourne, 1968). These islands are much farther north than Amsterdam, and various tems and boobies dominate the fauna, with a few species of petrel present. Bourne (1968) reported Wedge-tailed Shearwaters {Puffinus pacificus), a larger shearwater (possibly P. carneipes), and the Mascarene Petrel {Pterodroma aterrima) in addition to numer ous remains of the White-tailed Tropic Bird {Phaethon leptu rus) in a small collection of fossils from Rodriguez. The dis covery of bones of Barau's Petrel {Pterodroma baraui) on Amsterdam extends the range of this species, previously known only from Reunion, and provides a faunal link with the Mascarene Islands. Pterodroma aterrima, also known only from Reunion Island, probably bred on Rodriguez (Bourne, 1968), and it may also have reached Amsterdam occasionally. Studies of the fossil avifauna of St. Helena Island in the At lantic Ocean (16?S) have revealed remains of 21 species, among which the greatest loss of species and individuals was five of the six resident petrels, three of which were endemic (Olson, 1975). Farther north in the Atlantic, the fossil avifauna of the eastern Canary Islands, which lie off the African coast between about 28?N and 29?N, is dominated by remains of two extinct species of Puffinus (Alcover and McMinn, 1995). In the tropical Pacific Ocean the fossil faunas of numerous islands have been studied, and all are rich in terrestrial species: seabirds are dominated by procellariids, but toward the equa tor, the species diversity of tems, boobies, tropicbirds, and frig- atebirds often equals or exceeds that of procellariids (Olson and James, 1991; Steadman, 1995). On every island studied, several species of land birds and populations of seabirds were exterminated following the arrival of humans in the last few thousand years (Steadman, 1989, 1995; Olson and James, 1991). Often losses, particularly of land birds, have exceeded 50% of the original species diversity, and, among seabirds, some of the greatest losses have been petrels and shearwaters, especially in eastern Polynesia (Steadman, 1989). In the South Pacific the fossil faunas of the North and South islands of New Zealand have been studied extensively: Millen- er (1990) listed 34 species that became extinct in the late Ho- locene following the arrival of humans, to which Scarlett's Shearwater {Puffinus spelaeus), South Island Adzebill {Aptor- nis defossor), Bush Wren {Xenicus longipes), and Long-billed Wren {Dendroscansor decurvirostris) should be added, for a total of 38 species. In the New Zealand region, seabirds have not suffered the same degree of loss as elsewhere because, al though numerous populations of seabirds have been extirpated from the main islands, colonies on offshore islands have en sured their species survival, with the exception of Puffinus spelaeus (Holdaway and Worthy, 1994). The fossil faunas of the New Zealand subantarctic islands have not been studied. Conclusions Like most islands where numerous extinctions followed the arrival of humans and commensal mammals, Amsterdam Is land had a naive avifauna (Milberg and Tyrberg, 1993) unable to cope with predation by people and rats. Its location far from other landmasses was no doubt responsible for the paucity of land-bird species, and its relatively southern position resulted in a seabird fauna composed mainly of petrels. Studies of fossil faunas from islands throughout the world, reviewed above, show that petrels are particularly susceptible to the effects of humans and to predation by introduced commensal mammals. Amsterdam Island conforms to this generality in that only 10 of at least 19 species that formerly bred there survive; in other words, 47% of the original species have been extirpated. Five of these 10 surviving species, however, are critically endan gered on Amsterdam. Only three species are endemic to Am sterdam and St. Paul, and of these the land bird Anas marecula is extinct, whereas Diomedea amsterdamensis and Pachyptila macgillivrayi survive. NUMBER 89 59 Appendix 1 Summary statistics for Pterodroma macroptera (Measurements are in mm. Abbreviations are defined in "Methods.") Fossil Pterodroma macroptera specimen from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Coefficient of variation Fern L 36.20 0.86 34.20 37.83 47 2.37 TibAL 66.67 2.08 63.03 71.17 25 3.12 TmtL 41.83 1.08 38.63 44.26 41 2.58 Hum L 102.21 2.01 98.20 106.80 59 1.97 UlnaL 104.52 2.53 99.10 110.40 45 2.42 Cmc L 50.01 1.35 47.00 53.10 62 2.70 CorL 26.82 0.75 24.90 28.13 43 2.80 Skull TL 83.29 2.47 80.0 86.35 10 2.96 LacW 26.68 0.98 25.4 29.1 16 3.67 POW 33.08 0.90 31.38 34.7 13 2.72 ZPW 27.53 0.81 26.05 29.1 22 2.94 PmxL 37.78 2.49 35.5 44.3 11 6.59 Pmx W 15.52 0.77 14.4 16.84 11 4.96 BCL 39.33 0.53 38.0 40.1 16 1.35 Pmx W/L 0.40 0.04 0.35 0.45 7 10.0 Catalog number CSIRO PROS 240 CSIRO PROS 237 CSIRO PROS 961 BMNH S/1964.14.19 BMNH S/1964.14.18 BMNH 1848.8.31.39 BMNH 1848.8.31.40 Modem Pterodroma m. macroptera specimens from the Indian and Atlantic oceans (WA= Fern L 36.18 35.57 - 39.52 38.6 38.35 38.5 TibAL 67.98 65.4 69.33 71.67 68.1 68.7 67.72 TmtL - 42.5 44.12 45.0 43.27 43.0 43.16 Hum L 102.7 100.35 107.3 110.46 106.65 106.3 107.36 UlnaL - 103.96 111.18 114.81 109.33 109.43 111.76 Cmc L - 50.9 49.86 54.34 52.73 53.26 52.4 CorL 27.52 25.8 27.86 - 27.98 29.04 29.31 Skull TL 82.95 84.05 - 89.1 86.74 91.1 88.28 LacW 27.6 26.63 - 28.67 27.7 28.43 28.88 POW 34.9 34.72 - 34.4 35.44 34.07 35.48 ZPW 28.0 27.7 - 30.5 29.5 28.5 28.86 =Western Australia PmxL 41.9 42.8 - 49.56 45.5 47.05 45.56 . Pmx W Pmx W/L 14.72 14.58 - 15.35 16.3 14.9 16.12 0.35 0.34 - 0.31 0.36 0.32 0.35 Loc Eclipse Id., WA Eclipse Id., WA Coffin Id., WA Discovery Bay, Victoria Discovery Bay, Victoria South Atlantic South Atlantic Recent Pterodroma macroptera gouldi from the New Zealand region. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 39.24 1.05 38.0 40.7 14 TibAL 70.56 1.34 68.4 73.3 14 TmtL 43.8 0.80 42.7 45.6 14 Hum L 109.93 2.42 105.5 114.5 14 UlnaL 114.34 2.10 110.7 118.9 14 Cmc L 53.83 1.21 51.5 56.1 14 CorL 28.46 0.90 27.2 30.0 14 Skull TL 89.47 1.99 86.2 91.8 13 LacW 28.79 0.62 27.7 29.9 13 POW 34.66 0.97 33.4 36.1 11 ZPW 29.86 0.66 28.9 30.9 12 PmxL 45.92 1.72 43.1 48.5 13 Pmx W 16.43 0.88 15.0 18.0 13 Pmx W/L 0.36 0.02 0.33 0.39 13 Appendix 2 Summary statistics for Pterodroma mollis (Measurements are in mm. Abbreviations are defined in "Methods.") Fossil Pterodroma mollis from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Coefficient of variation Fern L 28.20 0.83 26.53 31.5 54 2.94 TibAL 53.14 1.44 48.48 55.22 20 2.71 TmtL 33.35 0.84 31.23 35.3 39 2.52 Hum L 78.60 2.25 73.1 84.0 78 2.86 UlnaL 80.95 2.42 74.7 84.2 75 2.99 Cmc L 39.06 0.92 36.5 41.4 74 2.35 CorL 21.53 0.56 20.2 23.0 87 2.60 Skull TL 69.70 1.82 67.5 73.0 6 2.61 LacW 22.59 0.83 21.2 24.4 15 3.67 POW 27.99 0.64 27.0 28.9 7 2.29 ZPW 23.01 0.69 21.6 24.2 19 3.00 PmxL 32.72 0.73 31.3 34.3 13 2.23 Pmx W 11.68 0.48 11.0 12.8 15 4.11 BCL 33.33 1.00 31.3 35.3 23 3.00 Pmx W/L 0.36 0.02 0.33 0.41 12 5.56 Modem Pterodroma mollis from Gough Island and New Zealand (NZ). Catalog number MNZ 22424 MNZ 22423 MNZ 22419 MNZ 16583A MNZ 21454 Fern L Tib AL Tmt L Hum L Ulna L Cmc L Cor L Skull TL Lac W PO W ZP W Pmx L Pmx W Pmx W/L Locality 30.1 30.7 30.4 29.5 29.8 57.5 58.5 58.3 57.6 57.5 36.3 35.3 35.1 34.8 35.3 83.1 83.4 82.1 83.1 82.2 85.7 86.8 83.3 86.6 85.3 40.9 41.6 39.9 40.6 41.7 23.2 23.0 22.3 22.8 23.3 72.0 72.7 73.7 73.5 - 22.6 23.3 23.3 23.3 - 27.7 29.7 28.9 29.7 - 22.9 23.5 24.5 24.2 - 35.0 35.4 35.6 34.3 - 11.5 10.8 12.3 11.7 - 0.33 0.31 0.36 0.34 - Gough Id. Gough Id. Gough Id. NZ, Rangitikei River NZ, Petone Beach 60 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Appendix 3 Summary statistics for Pachyptila (Measurements are in mm. Abbreviations are defined in "Methods.") Fossil Pachyptila macgillivrayi from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Coefficient of variation Fern L 24.35 0.81 22.58 26.21 51 3.33 TibAL 49.53 1.51 45.60 52.04 36 3.05 TmtL 32.83 1.20 30.40 35.50 71 3.65 Hum L 57.26 1.59 53.30 60.80 120 2.78 UlnaL 56.94 1.41 52.30 61.70 127 2.48 Cmc L 30.41 0.89 28.60 32.80 110 2.93 CorL 19.08 0.65 17.50 20.90 90 3.41 Skull TL 65.98 1.47 62.9 69.3 18 2.23 LacW 17.33 0.86 15.8 18.6 27 4.96 POW 22.09 0.55 21.1 23.2 20 2.49 ZPW 20.05 0.74 18.7 22.5 31 3.69 PmxL 36.25 1.40 33.1 39.4 32 3.86 PmxW 16.30 0.74 14.9 17.7 31 4.54 BCL 27.05 0.63 25.8 28.2 19 2.33 Pmx W/L 0.45 0.02 0.40 0.52 28 4.44 Modem Pachyptila vittata' from the New Zealand region. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 26.86 0.78 25.40 28.30 17 TibAL 52.65 1.24 50.80 55.10 17 TmtL 35.16 0.94 33.90 37.20 17 Hum L 61.92 1.12 60.00 64.70 16 UlnaL 61.78 1.16 60.20 64.30 16 Cmc L 33.04 0.81 31.50 34.60 16 CorL 20.53 0.65 19.50 21.90 16 Skull TL 71.88 1.54 68.50 74.40 16 LacW 21.54 0.96 19.8 23.10 14 POW 23.78 0.83 22.20 24.70 15 ZPW 21.66 0.59 20.60 22.40 16 PmxL 40.89 1.43 37.90 43.40 15 PmxW 18.99 1.64 15.50 20.50 15 Pmx W/L 0.46 0.036 0.382 0.507 15 'Included within the above statistics for P. vittata are data from MNZ 12481 and MNZ 12482, from Phillip Island, Australia, and the west coast of Wellington, New Zealand, respectively. These specimens are similar in size to the rest of the sample but have markedly narrower bills. Pmx W/L=0.40 mm and 0.38 mm, respectively, com pared to data for the other 13 specimens: mean=0.48, standard deviation=0.022, minimum=0.45, maximum=0.51, w=13. Modem Pachyptila salvini from the New Zealand region. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 24.87 1.15 23.6 27.9 10 TibAL 50.25 2.21 48.3 56.3 10 TmtL 33.00 1.36 31.6 36.1 10 Hum L 58.38 2.22 56.0 64.1 10 UlnaL 56.75 2.71 52.6 62.8 10 Cmc L 30.61 1.25 28.5 32.6 10 CorL 19.34 0.63 18.3 20.5 10 Skull TL 65.46 3.47 61.2 73.1 9 LacW 19.01 1.05 17.7 21.2 9 POW 22.62 0.87 21.1 23.7 10 ZPW 20.21 0.88 18.6 21.6 10 PmxL 35.78 2.58 33.1 41.6 9 PmxW 14.06 1.21 12.5 16.0 8 Pmx W/L 0.39 0.01 0.37 0.41 8 Appendix 4 Summary statistics for Procellaria cinerea (Measurements are in mm. Abbreviations are defined in "Methods.") Fossil Procellaria cinerea from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Coefficient of variation Fern L 46.60 1.21 45.00 48.75 16 2.60 TibTL 104.73 2.41 100.30 107.70 9 2.30 TibAL 90.88 1.62 87.70 93.30 15 1.78 TmtL 58.85 1.35 56.20 60.70 13 2.29 Hum L 130.60 2.61 124.50 136.60 24 2.00 UlnaL 126.89 2.54 122.90 131.50 19 2.00 Cmc L 61.03 1.38 58.90 64.10 19 2.26 Cor L Skull TL Preorb W 34.02 1.02 32.40 36.00 17 3.00 101.50 3.40 97.8 106.0 6 3.35 17.04 0.71 15.82 18.3 14 4.17 POW 37.88 0.98 36.3 38.8 5 2.59 ZPW 32.70 1.02 31.08 34.0 10 3.12 PmxL 52.08 1.46 50.3 54.5 9 2.80 PmxW 16.33 0.61 15.4 17.3 10 3.73 BCL 46.45 1.16 44.5 48.5 9 2.50 Pmx W/L 0.31 0.01 0.30 0.34 9 3.22 Modem Procellaria cinerea from the New Zealand region. Statistic Mean Standard deviation Minimum Maximum Sample size FemL TibAL Tmt L Hum L Ulna L Cmc L CorL Skull TL Preorb W POW ZPW PmxL PmxW Pmx W/L 48.44 1.65 43.7 51.2 21 95.67 1.78 92.2 98.8 22 62.23 1.19 60.0 64.3 21 136.59 2.45 132.6 141.5 24 134.93 2.43 131.5 140.1 23 64.12 1.19 61.8 66.1 23 35.94 0.58 34.7 36.8 23 106.78 2.89 101.1 113.0 22 16.70 1.16 12.5 18.1 22 39.04 1.16 36.7 40.9 21 33.75 0.71 32.4 35.2 21 56.70 1.89 53.3 60.4 23 16.25 1.14 13.2 18.2 23 0.29 0.02 0.24 0.32 23 NUMBER 89 61 Appendix 5 Summary statistics for Puffinus assimilis (Measurements are in mm. Abbreviations are defined in "Methods.") Fossil Puffinus assimilis from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Coefficient of variation Fern L 25.58 0.94 23.78 27.30 32 3.67 TibTL 71.31 0.43 71.00 71.61 2 0.60 TibAL 56.73 1.30 54.45 58.60 21 2.29 TmtL 40.39 1.38 35.16 43.46 36 3.42 Hum L 62.78 1.56 59.60 65.60 45 2.48 UlnaL 55.20 1.28 52.10 57.80 45 2.32 Cmc L 33.36 0.94 31.60 35.10 33 2.82 CorL 22.62 0.64 21.20 24.10 43 2.83 Skull TL Preorb W 65.25 0.64 64.8 65.7 2 0.98 8.57 0.45 7.6 9.4 21 5.25 POW 25.92 0.95 24.3 27.7 13 3.66 ZPW 20.40 0.69 19.2 22.0 15 3.38 PmxL 30.77 0.95 29.3 32.0 7 3.09 PmxW 9.17 0.26 8.8 9.4 6 2.84 BCL 31.50 0.58 30.6 33.2 20 1.84 Pmx W/L 0.30 0.02 0.28 0.32 6 6.67 Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 26.34 0.74 25.1 27.3 7 Vlodern Puffinus assimilis elegans TibAL 58.56 1.45 55.6 59.9 7 TmtL 40.83 0.92 39.2 41.7 7 from the Hum L 64.76 1.47 62.3 66.5 7 Antipodes. UlnaL 56.53 1.47 53.6 57.9 7 Cmc L 34.01 0.99 32.0 34.8 7 CorL 23.09 0.57 21.9 23.6 7 Modem Puffinus assimilis assimilis from Norfolk Island. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 23.20 - 23.2 23.2 1 TibAL 49.90 1.98 48.5 51.3 2 TmtL 35.15 2.05 33.7 36.6 2 Hum L 55.90 2.40 54.2 57.6 2 UlnaL 49.45 3.46 47.0 51.9 2 Cmc L 29.55 1.06 28.8 30.3 2 CorL 18.76 - 18.76 18.76 1 Recent Puffinus assimilis haurakiensis from Northland, New Zealand. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 23.80 0.85 23.2 24.4 2 TibAL 54.00 3.68 51.4 56.6 2 TmtL 37.95 2.19 36.4 39.5 2 Hum L 58.25 0.92 57.6 58.9 2 UlnaL 51.65 0.92 51.0 52.3 2 Cmc L 31.55 1.77 30.3 32.8 2 CorL 21.95 0.49 21.6 22.3 2 Recent Puffinus assimilis kermadecensis from the Kermadec Islands. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 24.02 0.57 23.1 25.0 10 TibAL 53.95 1.71 51.9 56.8 10 TmtL 38.25 1.46 36.1 40.7 10 Hum L 59.82 1.79 57.0 62.9 10 UlnaL 54.16 1.85 51.8 56.4 10 Cmc L 32.00 0.96 30.7 33.4 10 CorL 21.16 0.87 20.0 22.6 10 Recent Puffinus assimilis ssp. from New Zealand west-coast beaches. Statistic Mean Standard Deviation Minimum Maximum Sample size Fern L 23.48 0.53 22.6 24.2 10 TibAL 52.14 1.40 49.7 55.0 10 TmtL 36.61 0.57 35.5 37.5 10 Hum L 57.97 1.25 56.0 61.1 12 UlnaL 52.68 1.36 51.0 56.3 12 Cmc L 30.93 0.71 30.2 32.8 12 CorL 20.74 0.60 20.0 21.7 11 62 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Appendix 6 Summary statistics for storm-petrels (Measurements are in mm. Abbreviations are defined in "Methods.") Fossil Pelagodroma marina from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Coefficient of variation Fern L 17.23 0.41 15.92 17.93 37 2.38 TibTL 59.48 2.07 56.30 65.73 31 3.48 TibAL 54.19 1.66 51.86 60.32 46 3.06 TmtL 40.91 1.18 37.30 43.36 87 2.88 Hum L 25.02 0.81 22.84 26.59 49 3.24 UlnaL 22.32 0.57 21.01 23.80 73 2.55 Cmc L 16.88 0.51 16.03 18.13 40 3.02 CorL 14.17 0.61 13.25 15.10 15 4.30 Modem Pelagodroma marina from the New Zealand region. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 16.86 0.19 16.5 17.1 12 TibAL 53.92 1.16 52.1 56.5 14 TmtL 40.90 0.99 39.33 42.68 14 Hum L 25.09 0.46 24.12 25.8 14 UlnaL 22.93 0.58 21.6 23.96 14 Cmc L 16.82 0.47 15.76 17.72 14 CorL 13.72 0.49 13.05 14.58 14 Fossil Fregetta grallaria from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 18.94 0.06 18.9 18.98 2 TibTL 50.98 1.91 49.61 53.16 3 TibAL 47.77 1.66 46.18 49.4 4 TmtL 37.57 1.09 36.15 39.13 12 Hum L 27.03 0.38 26.6 27.31 3 UlnaL 23.17 0.54 22.43 23.76 5 Cmc L 17.76 - 17.76 17.76 1 Modem Fregetta from the New Zealand region. Specimen Fregetta grallaria MNZ 16071 Fregretta tropica MNZ 18963 MNZ 19277 MNZ 22254 MNZ 23798 Fern L 17.6 19.02 18.26 18.02 18.62 TibTL 53.0 57.73 57.05 57.52 58.70 TibAL 49.06 53.40 52.70 52.70 54.32 TmtL 37.4 42.45 40.69 41.42 42.54 Hum L 25.72 26.06 24.94 25.02 25.54 UlnaL 23.58 23.15 22.26 22.56 23.28 Cmc L 17.82 17.5 16.9 17.14 17.60 Tmt SW as % length 5.08 4.55 4.67 4.68 4.61 NUMBER 89 63 Appendix 7 Summary statistics for Pelecanoides species and subspecies (Measurements are in mm. Abbreviations are defined in "Methods.") Modem Pelecanoides georgicus from Heard Island. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 21.60 0.23 21.2 21.9 8 TibTL 43.75 0.69 42.6 44.8 8 TmtL 23.44 0.50 22.5 23.9 7 Hum L 40.18 0.67 39.3 41.0 8 UlnaL 30.30 0.70 29.4 31.1 8 CmcL 21.79 0.47 21.1 22.5 8 Cor L 22.54 0.65 21.6 23.7 7 Modem Pelecanoides urinatrix exsul from Auckland and Heard islands. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 23.46 0.49 22.9 24.2 7 TibTL 47.41 0.74 45.9 48.2 7 TmtL 25.74 0.63 24.6 26.5 7 Hum L 42.43 0.64 41.8 43.3 7 UlnaL 33.23 0.29 32.9 33.7 7 Cmc L 24.03 0.64 23.0 24.9 7 CorL 23.85 0.99 22.6 25.1 6 Modem Pelecanoides urinatrix chathamensis from Southeast Island, Chatham Group. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 22.74 0.67 21.5 23.4 7 TibTL 45.99 1.35 43.7 48.1 7 TmtL 24.99 0.89 23.5 26.3 7 HumL 40.80 0.68 39.8 41.9 7 UlnaL 32.31 0.81 30.8 33.3 7 Cmc L 23.18 0.50 22.5 23.8 6 CorL 22.80 0.41 22.0 23.2 7 Modem Pelecanoides u. urinatrix from beaches near Wellington, New Zealand. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 24.01 0.61 22.9 24.7 8 TibTL 48.45 0.89 46.9 49.6 8 TmtL 26.16 0.90 25.3 27.7 8 Hum L 44.43 0.73 43.1 45.7 8 UlnaL 35.66 0.61 34.7 36.4 8 Cmc L 25.24 0.67 24.3 26.3 8 CorL 23.83 0.61 22.8 24.8 8 Fossil Pelecanoides urinatrix from Amsterdam Island. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 22.51 0.35 22.05 23.13 7 TibTL 43.43 1.38 42.45 44.4 2 TmtL 24.42 0.40 23.8 24.8 5 Hum L 40.77 1.09 39.6 42.14 6 UlnaL 31.91 0.50 31.1 32.36 9 Cmc L 22.57 0.11 22.46 22.7 4 CorL 22.95 0.90 22.1 24.1 4 64 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Appendix 8 Summary statistics for Catharacta spp. (Measurements are in mm. Abbreviations are defined in "Methods.") Modem Catharacta maccormicki from the Antarctic. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 60.56 2.03 58.5 64.1 7 TibAL 103.01 3.05 100.3 108.1 7 TmtL 66.41 1.81 64.0 69.8 7 Hum L 135.54 4.15 130.6 141.6 7 UlnaL 139.56 4.45 133.4 144.8 7 Cmc L 69.60 2.36 66.4 73.7 7 Modem Catharacta skua lonnbergi from New Zealand Subantarctic islands. Statistic Mean Standard deviation Minimum Maximum Sample size Fern L 73.55 3.76 69.6 78.8 6 TibAL 126.18 4.22 122.0 132.5 5 TmtL 81.50 1.99 79.2 84.6 5 Hum L 152.02 5.15 145.1 158.0 6 UlnaL 153.88 4.40 148.4 159.5 5 Cmc L 76.40 1.80 74.1 78.4 5 Fossil Catharacta bones from Amsterdam Island. Location Hum L UlnaL Cmc L Fern L TibAL TmtL Site 8.4 Site 9 Site 13.5 Site 19 Site 15.2 71 144.1 74.4 76.5 75.9 NUMBER 89 65 Literature Cited Alcover, J.A., and M. McMinn 1995. Fossil Birds of the Canary Islands. Courier Forschungsinstitut Senckenberg, 181:207-213. Bourne, W.R.P. 1968. The Birds of Rodriguez, Indian Ocean. Ibis, 110:338-345. Bourne, W.R.P, A.C.F. David, and C. Jouanin 1983. Probable Garganey on St. Paul and Amsterdam Islands, Indian Ocean. Wildfowl, 34:127-129. Genevois, F., and V. Bretagnolle 1995. Sexual Dimorphism of Voice and Morphology in the Thin-billed Prion (Pachyptila belcheri). Notornis, 42:1-10. Gill, F.B. 1967. Birds of Rodriguez Island (Indian Ocean). Ibis, 109:383-390. Hachisuka, M. 1953. The Dodo and Kindred Birds. 250 pages. London: H.F. & G. With- erby, Ltd. Holdaway, R.N., and TH. Worthy 1994. A New Fossil Species of Shearwater Puffinus from the Late Quater nary of the South Island, New Zealand, and Notes on the Biogeogra phy and Evolution of the Puffinus Gavia Superspecies. Emu, 94: 201-215. Jouanin, C. 1953. Le materiel ornithologique de la mission "Passage de Venus sur le soleil" (1874), Station de L'ile Saint-Paul. Bulletin du Museum Na tional d'Histoire Naturelle, Paris, series 2, 25(6):529-540. Jouanin, C, and P. Paulian 1960. Recherches sur des ossements d'oiseaux provenant de file Nou- velle-Amsterdam (ocean Indien). In G. Bergman, K..O Donner, and Lars von Haartman, editors, Proceedings of the XHth International Ornithological Congress, Helsinki 5-12 VI 1958, pages 368-372. Jouventin, P. 1994. Past, Present, Future of Amsterdam Island (Indian Ocean) and Its Avifauna. Birdlife Conservation Series, 1:122-132. Jouventin, P., J. Martinez, and J.P. Roux 1989. Breeding Biology and Current Status of the Amsterdam Albatross Diomedea amsterdamensis. Ibis, 131:171-182. Jouventin, P., J.C. Stahl, H. Weimerskirch, and J.L. Mougin 1984. The Seabirds of the French Subantarctic Islands and Adelie Land, Their Status and Conservation. In J.P. Croxall, P.G.H Evans, and R.W. Schreiber, editors, Status and Conservation of the World's Seabirds, pages 609-625. Cambridge: International Council for Bird Preservation, Technical Publication number 2. Marchant, S., and P.J. Higgins, coordinators 1990. Handbook of Australian, New Zealand and Antarctic Birds, IA: Ra- tites to Petrels. 735 pages, 53 plates. Melbourne: Oxford University Press. Martinez, J. 1987. A New Probable Case of Insular Endemism: The Garganey of Am sterdam Island. Documents des Laboratoires de Geologie de la Fac- ulte des Sciences de Lyon, 99:211-219. Micol, T. 1995. Projet pilote de rehabilitation ecologique de l'ile Saint-Paul per eradication des rats et des lapins. [Unpublished report at Centre d'Etudes Biologiques de Chize, Centre Natioanl de la Recherche Scientifique, Administration du Territoire des Terres Australes et Antarctiques Francaises, 34 me des Renandes 75017 Paris, France.] Micol, T, and P. Jouventin 1995. Restoration of Amsterdam Island, South Indian Ocean, Following Control of Feral Cattle. Biological Conservation, 73:199-206. Milberg, P., and T. Tyrberg 1993. Naive Birds and Noble Savages?A Review of Man-Caused Prehis toric Extinctions of Island Birds. Ecography, 16:229-250. Millener, P.R. 1990. Evolution, Extinction and the Subfossil Record of New Zealand's Avifauna. In B.J. Gill and B.D. Heather, editors, A Flying Start, pages 93-100. Auckland: Random Century in association with Or nithological Society of New Zealand, Inc. Newton, E. 1888. List of Birds of the Mascarene Islands Including the Seychelles. Transactions of the Norfolk and Norwich Naturalist's Society, 4: 548-554. Olson, Storrs L. 1975. Paleornithology of St. Helena Island, South Atlantic Ocean. Smith sonian Contributions to Paleobiology, 23: 49 pages, 10 figures, 6 plates, 8 tables. Olson, Storrs L., and H.F. James 1991. Descriptions of Thirty-Two New Species of Birds from the Hawai ian Islands, Part 1: Non-Passeriformes. Ornithological Monographs, 45:1-88. Olson, Storrs L., and P. Jouventin 1996. A New Species of Small Flightless Duck from Amsterdam Island, Southern Indian Ocean (Anatidae: Anas). Condor, 98(1): 1-9. Paulian, P. 1960. Quelques donnees sur L'avifaune ancienne des lies Amsterdam et Saint-Paul. L 'Oiseau et la Revue Francaise d'Ornithologie, 30(1): 18-23. Roux, J.P, and J. Martinez 1987. Rare, Vagrant and Introduced Birds at Amsterdam and Saint Paul Is lands, Southern Indian Ocean. Cormorant, 14:3-19. Roux, J.P, J.-L. Mougin, and J.A. Bartle 1986. Le prion de Macgillivray; donnees taxinomiques. L 'Oiseau et la Re vue Francaise d'Ornithologie, 56(4):379-383. Steadman, D.W. 1989. Extinction of Birds in Eastern Polynesia: A Review of the Record, and Comparisons with Other Pacific Island Groups. Journal of Ar chaeological Science, 16:17 7-2 05. 1995. Prehistoric Extinctions of Pacific Island Birds: Biodiversity Meets Zooarchaeology. Science, 267:1123-1131. Tollu, B. 1984. La Quille (He Saint Paul, ocean Indien), sanctuaire de populations relictes. L 'Oiseau et la Revue Francaise d'Ornithologie, 54(1): 79-85. Worthy, T.H., and R.N. Holdaway 1993. Quaternary Fossil Faunas from Caves in the Punakaiki Area, West Coast, South Island, New Zealand. Journal of the Royal Society of New Zealand, 23(3): 147-254. Comparison of Paleoecological Patterns in Insular Bird Faunas: A Case Study from the Western Mediterranean and Hawaii Bartomeu Segui and Josep Antoni Alcover ABSTRACT A comparison among the Pleistocene avifaunas from the west- em Mediterranean islands (Menorca, Mallorca, and Cabrera (Gymnesic Islands); Eivissa (=Ibiza) and Formentera (Pityusic Islands); Corsica; Sardinia) shows great differences between those islands with terrestrial mammals (Gymnesic Islands, Corsica, Sar dinia) and those lacking them (Pityusic Islands). A close parallel is found between the late Pleistocene avian communities of the Pity- usics and the prehuman avifauna of the Hawaiian Islands, whereas the communities from the remaining western Mediterranean islands were similar ecologically to those of the eastern Mediterra nean islands. A key factor for the understanding of avian commu nity structure is the presence or absence of middle-sized herbivorous mammals (Myotragus balearicus Bate on the Gymne sic Islands and Megaceros cazioti (Deperet) on Corsica and Sar dinia). Introduction Since the works of Darwin (1859) and Wallace (1881), the study of islands has been a powerful tool for the advancement of many fields of biology (Vitousek et al., 1995). The recent in clusion of fossils in the study of insular faunas has conferred a historical dimension to previous analyses of insular biogeogra phy and ecology. This new component enables us to test the validity of faunal composition and biogeographic concepts de rived from the study of present faunas, allowing new kinds of analyses of the components and the structure of paleocommu- nities (Alcover and McMinn, 1994; James, 1995). The study of insular fossil faunas has already changed sever al paradigms of insular biogeography and ecology. Thus, the Bartomeu Segui, Departament de Ciencies de la Terra, Universitat de les Hies Balears, Carretera de Valldemossa, km 7.5, 07071 Ciutat de Mallorca, Balears (Spain). Josep Antoni Alcover, Institut Mediterrani d'Estudis Avancats, Carr et era de Valldemossa, km 7.5, 07071 Ciutat de Mallorca, Balears (Spain). actual existence of ecological processes of "faunistic turnover" for the establishment of an "insular equilibrium" (MacArthur and Wilson, 1967) has been seriously questioned (e.g., Olson and James, 1984; Steadman, 1986, 1995; James, 1995). Recent findings also have questioned ecological processes, such as the so-called "taxon cycle" (see Pregill and Olson, 1981), or postu lates to the effect that insular communities harbor species of predators exhibiting lesser body size than those from the clos est mainland (Blondel and Frochot, 1976). Typical insular communities are characterized by their dis harmony, by so-called "insular poverty," and by their high de gree of endemism (Wallace, 1895). These factors imply that ecological communities assemble on islands in a peculiar way, with an organization of ecological relationships between spe cies that is different from that on continents. Before human col onization, island communities were even more peculiar than at present (James, 1995). The western Mediterranean islands (Figure 1) have provided moderately rich Quaternary avifaunas (Alcover et al., 1992) as well as prehuman faunas of mammals, reptiles, and amphibians (see Table 1). To date, the late Pleistocene paleornithological record consists of 55 or 56 species in Mallorca, eight in Menor ca, 60 to 65 in Eivissa, two in Formentera, 74 in Corsica, and 49 in Sardinia. The fossil avifauna of Eivissa is considered to be the best represented of prehuman bird communities of any island in the Mediterranean region. The late Pleistocene record of Eivissa is based on more than 150,000 fossil bird specimens, mainly coming from a single deposit, Es Pouas (McMinn, in prep.), as well as from another six sites of lesser importance. In Mallorca, more than 3000 bird bones have been exhumed to date from 14 late Pleistocene and Holocene sites (Segui, 1997). The fossil record from Corsica and Sardinia, in spite of not be ing as representative of its Quaternary avifaunas, displays sev eral features that allows the characterization of its prehuman native faunas. The late Pleistocene fossil record of Menorca and Formentera is still very limited. Because we do not consid er the fossil records of these islands to be as representative of 67 68 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY EUROPE Balearic Islands FIGURE 1.?Map of the Mediterranean region showing the position and nomenclature of the Balearic Islands used in this paper. Eivissa is the Catalan name for the island known in other languages as Ibiza. native, prehuman avifaunas as those of the other islands, we have excluded Menorca and Formentera from our analysis. The numbers of fossil bones presented above, together with data on species distribution in the different deposits, allow a rough esti mate of relative species abundance on the islands, taphonomic biases included. The very different vertebrate paleofaunas of the Mediterra nean-basin archipelagos are of interest for comparison with other insular faunas worldwide. Herein we present a prelimi nary analysis of the paleoecology of the late Pleistocene bird communities of the western Mediterranean islands based on the known trophic ecology of the different species recorded on the islands. We emphasize parallels with other known prehu man native-bird communities worldwide. The Pleistocene bird communities of the western Mediterranean mainly include ex tant species, even though many have recently vanished from several or all the islands considered. They also include at least four or five extinct species: Tyto balearica Mourer-Chauvire, Alcover, Moya, and Pons, a nonendemic species found on Mallorca, Menorca, Corsica (Mourer-Chauvire, pers. comm., 1996), and in some mainland deposits from France and the Iberian Peninsula; Bubo insularis Mourer-Chauvire and Wees- ie, endemic to Corsica and Sardinia; an undescribed species of Athene endemic to Corsica (Mourer-Chauvire, pers. comm., 1996); an undescribed species of Rallus endemic to Eivissa; and Grus primigenia Milne-Edwards, a nonendemic species widely distributed during the European Pleistocene, also present on Eivissa and Mallorca, which is probably conspecif- ic with Grus grus (J.R. Stewart, pers. comm., 1996). Some taxa previously considered as possible endemics (e.g., Fring- illidae, species undescribed, and Corvidae, species unde scribed; see Alcover et al., 1992) have been deleted after taxo nomic reappraisal. Nomenclature for binomials of modern birds follows Sibley and Monroe (1990). Even though the determination of the feeding habits and spe cializations of extinct species may be imprecise or speculative, the allocation of these species to broad trophic categories can allow us to undertake a more general analysis. These kinds of inferences are applicable to all the extinct Mediterranean spe cies considered above, but they cannot be extended indiscrimi nately, especially in the case of bizarre small Passeriformes without known analogs (e.g., Vangulifer from Maui, Hawaiian Islands; James and Olson, 1991). ACKNOWLEDGMENTS.?We are indebted to Storrs Olson, Helen James, Dave Steadman, and Anna Traveset for their comments, ideas, and useful discussions. Damia Jaume and NUMBER 89 69 TABLE 1.?Vertebrate fauna, excluding birds, from the late Pleistocene and the Holocene of the main western Mediterranean islands. Fauna Amphibians Reptiles Mammals Eivissa Podarcis pityusensis bats Mallorca Alytes (B.) muletensis Podarcis lilfordi Nesiotites hidalgo Hypnomys morpheus Myotragus balearicus bats Corsica and Sardinia Euproctus spp. Speleomanthes spp. Discoglossus spp. Hyla sarda Phyllodactylus europaeus Podarcis tiliguerta Archaeolacerta bedriagae Algyroides filzingeri Nesiotites spp. Rhagamys orthodon Microtus henseli Prolagus sardus Lutrini spp. Megaceros (D.) cazioti bats Eva Bonner helped us with the English text. This paper is in cluded in the Research Projects of the Direccion General de In- vestigacion Cientifica y Tecnica (Madrid), project PB97-1173, and the Comision Interdepartmental de Ciencia y Technologia (Madrid), project AMB96-0843. Bird Trophic Type (BTT) Initially, we defined the main Bird Trophic Types (BTTs) present on the western Mediterranean islands. Afterwards, we identified the taxa present in the fossil record of each island and assigned them to their respective BTT. The distribution of the taxa among BTTs was analyzed at different taxonomic lev els (species, genus, or family) to enable comparisons with other insular faunas, even those from other biogeographical regions. In Table 2 the basic BTTs considered in this work are de fined. These do not include all the actual trophic types of birds, especially those of passerines, and they exclude others, such as aerial planctophagous species (e.g., swifts) or coastal general- ists (e.g., gulls). Nevertheless, the majority of bird species of the Mediterranean islands fit well in these BTTs (see Table 3). Trophic types considered are as follows: predators of verte brates (superpredators, nocturnal predators, omithophagous di urnal predators, generalist diurnal predators, fish-eaters); scav engers; sea birds; herbivorous nonpasserines; malacophagous, insectivorous, and herbivorous nonpasserines; and three BTTs within the Passeriformes (large-sized omnivores, small-sized omnivores, granivores). The assignment of each taxon to a BTT was based on the main components of their diet, according to various general TABLE 2.?Main BTTs (Bird Trophic Types) in the Pleistocene-Holocene fossil record from the western Mediterranean Islands. Nonpasseriforms Passeriforms Vertebrate predators Superpredators Nocturnal predators Diurnal predators specialized omi thophagous generalist Fish-eaters Scavengers Sea birds Herbivores Malacophages, insectivores, and herbivores Omnivores large-sized small-sized large-sized small-sized large-sized small-sized Granivores large-sized birds that prey mainly on middle- and large-sized vertebrates nocturnal birds that prey mainly on small-sized vertebrates during the night and twilight diurnal birds specialized for preying on middle- and small-sized birds, mainly small Passeriformes diumal birds for which birds are just a part of the diet, which consists mainly of small-sized verte brates, insects, and carrion specialized fish-catchers that feed in inland and littoral waters diumal birds that feed totally or partially on carrion and bone fragments of large carcasses pelagic birds that feed mainly on crustaceans, cephalopods, and fish from the upper layers of the water column middle- and large-sized browsers, grazers, and frugivores small-sized browsers, grazers, and frugivores large-sized birds that feed mainly on terrestrial snails and insects as well as on varied vegetable sources medium- and small-sized birds that feed mainly on terrestrial snails and insects as well as on varied vegetable sources large-sized passeriformes that feed on varied animal and vegetable sources middle- and small-sized passeriformes that feed on varied animal and vegetable sources small passeriformes that feed mainly on seeds 70 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 3.?The main BTTs (Bird Tropic Types) on islands with terrestrial mammals (Corsica, Sardinia, Mallorca) compared with islands without terrestrial mammals (Eivissa, Hawaiian Islands). Taxa in parentheses are known only from very scarce remains. On islands with terrestrial mammals, some of the BTTs are not present because the niches that they represent are practically monopolized by mammals (see shaded boxes). Data are from the authors; Alcover et al., 1992; Olson and James, 1991; and James and Olson, 1991. Nonpasseriforms Passeriforms Bir Vertebrate predators d trophic types Superpredators Nocturnal predators Diumal predators specialized ornithopha- gous generalists Fish-eaters Scavengers Sea birds Herbivores Malacophages, insectivores, detritivores, herbivores Omnivores large-sized small-sized large-sized small-sized small-sized large-sized Granivores Islands with terrestrial mammals Corsica and Sardinia Aquila Bubo, Asio, Tyto alba Accipiter Falco, Milvus, Buteo, Circus Haliaeetus Gypini Procellaridae spp. Cervidae Rodentia, Lagomorpha (Otis) (Rallus) (Porzana) (Gallinula) (Crex) Sturnus, Pyrrhocorax (Corvus corax) several Fringillidae, Passeridae, Emberizidae Mallorca Aquila Tyto alba/ Tyto balearica Accipiter Falco - Gypini - Bovidae Rodentia (Grus) (Porzana) (Sturnus), (Corvus), Pyrrhocorax - several Fringillidae, Passeridae, Emberiz idae Eastern Mediterranean islands Aquila Tyto alba, T melitensis, Athene ere tens is, Aegolius, (Ketupa) (A. flammeus) Accipiter Buteo Haliaeetus Gypini (Puffinus) Cervidae, Elephantidae, Hippopotamidae Rodentia (Otis) Grus melitensis (Porzana) (Gallinula) (Fulica) (Sturnus), Corvus, Pyrrhocorax, Garrulus (Corvus corax) several Fringillidae and Emberizidae Islands without terrestrial mammals Eivissa Hawaiian Islands Haliaeetus Asio flammeus Accipiter Falco (Circus) Haliaeetus, Pandion Haliaeetus Procellaridae spp. Anser/Branta - Otis, Grus Rallus, Crex Pyrrhocorax Corvus several Fringillidae, Passeridae, Emberizidae Haliaeetus Grallistrix Circus dossenus Buteo Haliaeetus Haliaeetus Procellaridae spp. Branta, moa-nalos - Apteribis Porzana Aidemedia Corvus viriosus, Corvus impluviatus, Corvus hawaiiensis several Drepanidini works (Cramp and Simmons, 1977, 1980; Cramp, 1985; Hoyo et al., 1992, 1994, 1996). Usually, each taxon was referred only to a single BTT, although a small group of species occupied more than one type (e.g., Haliaeetus on Eivissa, which is con sidered to have been the superpredator on the island but was si multaneously the main fish predator and the main scavenger). The absence of medium-sized herbivorous mammals on Eivissa is related to the presence of an abundant goose, proba bly derived from Anser erythropus. The tribe Anserini is usual ly underrepresented, due to taphonomic conditions, in nonan- thropogenic cave sites, excepting on some islands. Because of the great number of remains that have been exhumed from spe leological sites on Eivissa, the extinct goose must have been very abundant in Pityusic paleornithological communities. This species was probably the main medium-sized herbivorous ver tebrate on the island and must be considered to be the key spe cies in understanding Pityusic paleoecosystems. The occupation of the browsing- or grazing-herbivore trophic type by mammals (Bovidae and Cervidae) on all the Mediterranean islands has further implications for trophic webs. Thus, the islands with ungulates have eagles of the genus Aquila as superpredators and have vultures {Aegypius, Gyps) as the chief large scavengers. On the other hand, on islands with out these mammals but with abundant Anseriformes and Pro- cellariiformes, the role of superpredators is taken by sea-eagles of the genus Haliaeetus, which also must be considered the main scavengers there. This relationship, confirmed on the Mediterranean islands, also applies to other islands in the world (Alcover and McMinn, 1992, 1994). Sea-eagles must have been very abundant in the Pityusic Islands before human colo nization. In just one fossiliferous site, remains of at least nine individuals have been found, including several practically com plete associated skeletons. On Corsica and Sardinia, on the oth er hand, very few remains of Haliaeetus have been recovered. As far as vertebrate nocturnal predators are concerned, the absence of Tytonidae in the fossil record of Eivissa, as is the case in the Hawaiian Islands, is particularly noteworthy. Ty tonidae are present in the early Pleistocene and the Holocene of NUMBER 89 71 Mallorca, probably in prehuman levels, and in the late Pleis tocene from Corsica-Sardinia (Alcover et al., 1992). Their presence is probably related to the presence of small mammals on these islands {Hypnomys and Nesiotites in the Gymnesic Is lands, Nesiotites, Rhagamys, Microtus, and Prolagus in Corsi ca-Sardinia), whereas small mammals are absent from Eivissa. It is not possible, however, to infer in a strict sense the absence of Tytonidae on islands without small mammals. For example, in the Galapagos and Canary islands different small mammals have been recorded on some of the islands but not on all of them. In these archipelagos, fytonids have even been obtained in prehuman deposits on islands without small mammals (Steadman, 1986; Alcover, unpublished). Strigid owls that prey on vertebrates are unknown in the fos sil record of Mallorca. Although Otus scops and Athene noctua do occur in the fossil record, these species feed mainly on in sects. Asio flammeus and Bubo insular is are present in the fos sil record on Sardinia, and Bubo insularis, Asio otus, and a new, undescribed species of Athene have been found as fossils on Corsica (Mourer-Chauvire, pers. comm., 1996). Several re mains of Asio flammeus and those of a single individual of a large species of Strigidae, probably Bubo bubo, have been found in the late Pleistocene of Eivissa (Alcover et al., 1992; Sondaar et al., 1995). Asio flammeus, the more abundant strigid on Eivissa, is more diurnally active than other Palearctic spe cies of owls of the same size, like Asio otus or Strix aluco (Cramp, 1985). This behavioral pattern may have been advan tageous for preying on birds, which were the primary food available on Eivissa before human colonization. Circadian rhythms might differ between insular and continental systems (Granjon and Cheylan, 1990, 1993, gave an example of change of circadian rhythms in a mammalian species). Asio flammeus may have been more diurnal on Eivissa than in other areas of its range. The endemic genus Grallistrix lived in the Hawaiian Islands (Olson and James, 1991). This genus shows anatomical spe cializations for bird-catching (elongated legs and shortened wings). Its ornithophagous habits have been confirmed by fos sil pellets (Olson and James, 1991). The main activity period of Grallistrix is unknown, but its ability to catch and kill birds in flight suggests that the species of this genus may have been more diurnal than other Strigidae. Hawks and falcons are widely distributed on western Medi terranean islands and have been found as fossils on all those is lands studied herein. In the Hawaiian Islands this BTT is occu pied by Circus dossenus, a species convergent with bird-eating hawks of the genus Accipiter (Olson and James, 1991). On Mediterranean islands with mammals, species of Buteo (with a diet of small mammals and small- and medium-sized reptiles), species of Circus and Milvus (each with a varied diet), and small species of the genus Falco (mainly herpetophagous and entomophagous) have been found (Alcover et al., 1992). This BTT is occupied by Buteo solitarius in the Hawaiian Islands (Olson and James, 1991), whereas in Eivissa it is represented by Circus cyaneus, which is scarce in the fossil record, and by small species of the more abundant Falco (Sondaar et al., 1995). Pelagic sea birds are not found on the large western Mediter ranean islands with mammals, being unknown in Corsica and Sardinia (Alcover et al., 1992). On the other hand, the small is land of Tavolara, near Sardinia, has yielded fossil remains of Calonectris diomedea and Puffinus yelkouan (Mayaud and Schaub, 1950). Significantly, Procellariiformes were abundant in the Pityusic Islands. Many fossil bones of Puffinus maure- tanicus and scarcer remains of Calonectris diomedea and Hy- drobates pelagicus have been found in Eivissa (Sondaar et al., 1995). Similarly, abundant remains of procellariforms, such as the extinct, small-sized gadfly petrel Pterodroma jugabilis (Ol son and James, 1991), have been found in Oahu and Hawaii, and many other Procellariidae are commonly encountered. The BTT that includes the nonpasseriform species with a wide diet (malacophagous, insectivorous, detritivorous, and herbivorous birds) has a better representation on the Pityusic and Hawaiian islands than on the other western Mediterranean islands. Two or more flightless species of ibis belonging to the extinct genus Apteribis lived on the Hawaiian islands Molokai, Maui, and Lanai, which constituted the Pleistocene island of Maui Nui (Olson and James, 1991). The genera Otis (bustards) and Grus (cranes) can be considered to be the Palearctic eco logical analogs of these species. Both are common in the Eivis- san fossil record but are not as common on the other western Mediterranean islands: so far, only a few remains of one speci men of Grus have been obtained in Mallorca (Mourer-Chau vire et al., 1975), and a few remains of Otis have been found on Corsica (Mourer-Chauvire, pers. comm., 1996). Among the rails, only a few remains of Porzana have been recorded on Mallorca, and no rails have been found in the Sar dinian fossil record. Four species of rails are infrequently en countered in the fossil record of Corsica (Alcover et al., 1992), whereas rails are well represented in the late Pleistocene of Eivissa as well as on the Hawaiian Islands. Endemic species have evolved on the Hawaiian Islands and on Evissa but not on the other western Mediterranean islands. An endemic unde scribed form of Rallus lived on Eivissa (McMinn, in prep.), whereas the genus Porzana underwent a significant radiation in the Hawaiian Islands (Olson and James, 1991). The BTTs occupied by Passeriformes are not as easily de fined as those occupied by nonpasserines. Thus, only three BTTs were considered for the Passeriformes. Some patterns, coming mainly from corvids, emerged from our analysis. On all the islands studied, corvids played an important role. Pyr rhocorax was present during the late Pleistocene on all the western Mediterranean islands. At least one small-sized species of Corvus also was present in Corsica, Sardinia, and Mallorca. Large-sized species of Corvus are scarce in the fossil record of Mallorca, Corsica, and Sardinia. In contrast, on Eivissa abun dant remains of a large Corvus, initially identified as C. ante- corax (Florit et al., 1986), but whose identity is now under re vision (McMinn, in prep.), have been recorded. In the Pleistocene of the Hawaiian Islands three large-sized species of 72 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Corvus have been found: the extinct C. impluviatus James and Olson (1991) and C viriosus James and Olson (1991) and the extant C. hawaiiensis. Probably all the large Hawaiian and Mediterranean species of Corvus represent roughly the same BTT. The scarcity of large crows on the Mediterranean islands with terrestrial mammals, in contrast with their abundance in Eivissa and the Hawaiian Islands, reinforces the parallels be tween the two archipelagos. The BTT that includes small-sized corvids on the Mediterra nean islands is tentatively paralleled by the Hawaiian genus Aidemedia, which on the basis of its jaws was thought to have dietary habits similar to those of Sturnus (James and Olson, 1991). Sturnus is present in the fossil records of Mallorca (Seg ui et al., 1997), Corsica, and Sardinia (Alcover et al., 1992), al though it is considerably scarcer than Pyrrhocorax. The living representatives of both genera have opportunistic dietary habits and eat lots of fleshy fruits. The small Passeriformes (perching birds) are included in a very wide range of BTTs that are more difficult to characterize than the larger species, compounded by the fact that they often change their BTT through the year. Because of these and other considerations we have omitted them from Table 3, except for small granivorous species, which are included to show that the members of a variety of families occupy the same BTT on the Mediterranean islands as do some granivorous species of Drepanidini in the Hawaiian Islands. A similar pattern among granivorous passerines in the Galapagos Islands (all belonging to a single radiation) and in the Canary Islands (where species belonging to different families occupy this BTT) has been re corded. Discussion The peculiarity of the Eivissan paleoavifauna within the Mediterranean region is clearly supported by the preceding analysis. The late Pleistocene fauna of Eivissa was structured along similar lines to that of the Hawaiian Islands: Anseri formes as the most important middle-sized grazers, sea-eagles as superpredators, mammal-eating tytonids absent, bird-catch ing strigids with more diurnal behaviour present, diurnal birds of prey (bird-catching specialists or more generalists) present, ground-dwelling species with varying food habits (i.e., medi um- and small-sized malacophagous, insectivorous, detritivo- rous, and herbivorous species), along with flying, omnivorous, medium-sized passerines (large corvids). In the faunas of the other Mediterranean islands, where land mammals were present, the paleornithological communities were structured in very different ways. One of the main points to be emphasized is that the greatest parallel to the late Pleistocene bird community of Eivissa is not on the other Mediterranean islands, the Atlantic or Indian oce anic islands that lack terrestrial mammals, or on the majority of Pacific islands, but on the very distant and isolated Hawaiian Islands. Conversely, the most accurate parallel to the Hawaiian paleoavifauna, according to the assignment of the different BTTs, is not to found among the other Pacific Islands, or among the Atlantic Islands, but rather in the late Pleistocene fauna from Eivissa. The significance of such an ecological rela tionship is currently unknown. Nevertheless, there is no doubt that an accurate study of the Eivissan fossil avifauna will be useful for the understanding of the development of the Hawai ian avifauna, and vice versa. In any case, the striking similarity between the avian paleocommunities of the Hawaiian Islands and the Pityusic Islands reinforces criticisms of the randomness of the ecological processes of immigration and extinction in the development of insular communities. One point of interest of the present analysis is its predictive character. We have an accurate knowledge of the late Pleis tocene fauna from Eivissa, contrasting with scarce data on late Pliocene/early Pleistocene bird fauna of the island (Alcover, 1989, and unpublished data). During this earlier period, a verte brate fauna, including a giant tortoise and at least two terrestrial mammals, lived on Eivissa. In addition, a varied mollusk fauna of at least 22 species also was present. These faunas suffered a dramatic, early or middle Pleistocene extinction event (Alcover et al., 1994). The avifauna associated with the giant tortoise ep isode is as yet poorly known, although it likely was substantial ly different from the late Pleistocene avifauna. The mass ex tinction on Eivissa must have forced a change in the composition and structure of the bird communities. Study of the fossil bird fauna from Menorca also should be enlightening. This island, which is about same size as Eivissa, was occupied by Myotragus balearicus Bate, a terrestrial ungu late that must be considered a key species in the ecosystem. If the size of the island alone was the key factor in determining its fauna, the late Pleistocene fossil avifauna from Menorca would have been similar to that of Eivissa. But if the island's ecology has greater importance in determining its fauna, as we postu late, the late Pleistocene avifauna from Menorca will cluster with that of Mallorca, where Myotragus also was present. CATALAN SUMMARY La comparacio entre les avifaunes pleistoceniques de les illes de la Mediterrania occidental (Gimnesies, Pitiuses, Massis cirno- sard) documenta l'existencia de dues castes de comunitats orni- tiques en el passat: una a les illes habitades per mamifers terrestres (Gimnesies, Corsega, i Sardenya) i raltra a les illes que no en con- tenen (Pitiuses). Les comunitats ornitiques pitiuses troben el seu paral.lelisme mes evident a les comunitats ornitiques prehumanes de les Hawaii, mentre que les comunitats ornitiques de les altres illes de la Mediterrania Occidental s'addiuen mes amb les de les illes de la Mediterrania Oriental. El factor clau per entendre I'estructuracio de les comunitats ornitiques insulars mediterxanies del Pleistoce sembla esser la presencia/absencia de mamifers her- bivors de talla mitjana {Myotragus balearicus Bate a les Gimne sies i Megaceros caziotti (Deperet) al massis cimo-sard). NUMBER 89 73 Literature Cited Alcover, J.A. 1989. Les aus fossils de la Cova de Ca Na Reia. Endins, 14-15:95-100. Alcover, J.A., F. Florit, C. Mourer-Chauvire\ and P.D.M. Wessie 1992. The Avifaunas of the Isolated Mediterranean Islands during the Middle and Late Pleistocene. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:273-283. Alcover, J.A., and M. McMinn 1992. Presencia de I'aguila marina Haliaeetus albicilla (Linnaeus 1758) al jaciment espeleologic quatemari d'Es Pouas (Sant Antoni de Portmany, Eivissa). Endins. 17-18:81-87. 1994. Predators of Vertebrates on Islands. BioScience, 44(1): 12-18. Alcover, J.A., M. McMinn, and CR. Altaba 1994. Eivissa: A Pleistocene Oceanic-like Island in the Mediterranean Sea. Research and Exploration, 10:236-238. Alcover, J.A., S. Moya-Sola, and J. Pons-Moya 1981. Les quimeres del passat; Els vertebrats fossils del Plio-Quatemari de les Balears i Pitiuses. Monografies Cientifiques, 1:1-260. Palma de Mallorca: Editorial Moll. Blondel, J., and B. Frochot 1976. Caracteres generaux de l'avifaune Corse: effets de l'insularit? et influence de l'homme sur son evolution. Bulletin de la Societe des Sciences Historiques et Naturelles de la Corse, 619-620:63-74. Cramp, S. 1985. The Birds of the Western Palearctic. Volume 4, 960 pages. Oxford: Oxford University Press. Cramp, S., and K..E.L. Simmons 1977. The Birds of the Western Palearctic. Volume 1, 722 pages. Oxford: Oxford University Press. 1980. The Birds of the Western Palearctic. Volume 2, 695 pages. Oxford: Oxford University Press. 1983. The Birds of the Western Palearctic. Volume 3, 913 pages. Oxford: Oxford University Press. Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, ix+502 pages. London: John Murray. Florit, R, C. Mourer-Chauvire, and J.A. Alcover 1986. Els ocells pleistocenics d'Es Pouas, Eivissa; nota preliminar. Bol leti Institucio Catalana d'Histdria Natural, 56:35?46. Granjon, L., and G. Cheylan 1990. Adaptations comportementales des rats noirs Rattus raltus des iles Ouest-Mediterraneennes. Vie et Milieu, 40:189-195. 1993. Differenciation genetique, morphologique et comportementale des populations de rats noirs Rattus rattus (L.) des lies d'Hyeres (Var, France). Scientific Report of the Port-Cros National Park, 15: 153-170. Hoyo, J. del, A. Elliott, and J. Sargatal 1992. Handbook of the Birds of the World. Volume 1, 696 pages. Barce lona: Lynx Editions. 1994. Handbook of the Birds of the World. Volume 2, 638 pages. Barce lona; Lynx Editions. 1996. Handbook of the Birds of the World. Volume 3, 821 pages. Barce lona: Lynx Editions. James, H.F. 1995. Prehistoric Extinctions and Ecological Changes on Oceanic Is lands. In P.M. Vitousek, L.L. Loope, and H. Adsersen editors, Is lands; Biological Diversity and Ecosystem Function. Ecological Studies, 115:87-102. Berlin: Springer-Verlag. James, H.F., and S.L. Olson 1991. Descriptions of Thirty-Two New Species of Birds from the Hawai ian Islands, Part II: Passeriformes. Ornithological Monographs, 46:1-88. MacArthur, R.H., and E.O. Wilson 1967. The Theory of Island Biogeography. 203 pages. New Jersey: Princ eton University Press. Mayaud, N., and S. Schaub 1950. Les puffins subfossiles de Sardaigne. Verhandlungen der Naturfor- schenden Gesellschaft in Basel, 61:19-27. McMinn, M. In prep. La omitofauna fosil de Eivissa: No Passeriformes. Doctoral manuscript, Universitat Illes Balears. Mourer-Chauvire, C, R. Adrover, and J. Pons 1975. Presence de Grus antigone (L.) dans L'"Avenc de Na Coma" a Ma- jorque (Espagne). Nouvelles Archives du Museum d'Histoire Na turelle de Lyon, 13:45-50. Olson, S.L., and H.F. James 1984. The Role of Polynesians in the Extinction of the Avifauna of the Hawaiian Islands. In PS. Martin and R.G. Klein, editors, Quater nary Extinctions: A Prehistoric Revolution, pages 768-780. Tuc son: University of Arizona Press. 1991. Descriptions of Thirty-Two New Species of Birds from the Hawai ian Islands, Part I: Non-Passeriformes. Ornithological Mono graphs, 45:1-88. Pregill, G.K., and S.L. Olson 1981. Zoogeography of West Indian Vertebrates in Relation to Pleis tocene Climatic Cycles. Annual Review of Ecology and Systemat ics, 12:75-98. Segui, B. 1997. Les avifaunes fossils dels jaciments carstics del Plioce, Plistoce i Holoce de les Gimnesies. Bolleti de la Societal d'Histdria Natural de les Balears, 39:25-42. Segui, B., C. Mourer-Chauvire, and J.A. Alcover 1997. Upper Pleistocene and Holocene Fossil Avifauna from Moleta Cave (Mallorca, Balearic Islands). Bolleti de la Societal d'Histdria Natural de les Balears, 40:223-252. Sibley, Charles G., and Burt L. Monroe, Jr. 1990. Distribution and Taxonomy of Birds of the World. 1111 pages. New Haven: Yale University Press. Sondaar, P.Y., M. McMinn, B. Segui, and J.A. Alcover 1995. Interes paleontologie dels jaciments carstics de les Gimnesies i les Pitiuses. Endins, 20:155-170. [Also published in the series Monografies de la Societal d'Histdria Natural de les Balears, vol ume 3.] Steadman, D.W. 1986. Holocene Vertebrate Fossils from Isla Floreana, Galapagos. Smith sonian Contributions to Zoology, 413: 103 pages. 1995. Prehistoric Extinctions of Pacific Islands Birds: Biodiversity Meets Zooarchaeology. Science, 267:1123-1131. Vitousek, P.M., L.L. Loope, and H. Adsersen, editors 1995. Islands; Biological Diversity and Ecosystem Function. Ecological Studies, 115:1-238. Berlin: Springer-Verlag. Wallace, A.R. 1881. Island Life, xvi + 522 pages. New York: Harper and Brothers. A New Species of Extinct Barn Owl (Aves: Tyto) from Barbuda, Lesser Antilles David W. Steadman and William B. Hilgartner ABSTRACT A new species of extinct barn owl, Tyto neddi, is described from six bones discovered in a late Quaternary cave deposit on Barbuda, Lesser Antilles, West Indies. Tyto neddi is the first extinct species of bam owl known from the Lesser Antilles. It appears to be most closely related to the several large, extinct species of Tyto known from late Quaternary cave deposits in the Greater Antilles and Bahamas. By far, the most abundantly represented species of ver tebrate at the type locality of T. neddi is a large, extinct oryzomy- ine rodent (genus and species undescribed) that probably was the primary prey item of T. neddi. A single pedal phalanx from the type locality of T. neddi represents a much smaller species of Tyto that is the size of the extant T. alba (Scopoli). Bones of the Bur rowing Owl, Athene (Speotyto) cunicularia (Molina), also occur commonly in Barbudan caves. Thus Barbuda, where no species of owl occurs today, once supported at least three species. Introduction During 16-24 January 1983, DWS was part of a team (see Acknowledgments) that surveyed modern and prehistoric ver tebrates on Barbuda, Leeward Islands, Lesser Antilles (Figure 1). Although no deep, stratified bone deposits were found, we did collect hundreds of vertebrate fossils from several small, shallow, sediment accumulations in a limestone cave system at Gun Shop Cliff on Barbuda's northeastern coast. At least 42 in digenous species of reptiles, birds, and mammals are known from these and other prehistoric sites on Barbuda, which has a fossil vertebrate fauna exceeded in the Lesser Antilles only on nearby Antigua (Pregill et al., 1994). The most common species in the Barbudan sites by far is a large, undescribed, extinct cricetid rodent (Oryzomyini, genus and species undescribed), which would suggest that a large owl was responsible for the bone deposits in these caves. Bones of David W. Steadman, Florida Museum of Natural History, University of Florida, P.O. Box 117800, Gainesville, Florida 32611, United States. William B. Hilgartner, Department of Geography and Environ mental Engineering, The Johns Hopkins University, Baltimore, Mary land 21218, United States. owls, however, are often scarce in Quaternary fossil deposits on islands, even in deposits that the owls helped to create (Steadman, 1986). In spite of the abundance of cricetid rodent bones in prehistoric sites from both Barbuda and Antigua (Ray, 1962; Steadman et al., 1984; Watters et al., 1984), evidence of any large species of owl was lacking on these islands and else where in the Lesser Antilles until we found six bones of a large species of Tyto in a small solution cavity at Gun Shop Cliff known as "Rat Pocket." Five extinct species of barn owls (Tytonidae: Tyto) are known from bone deposits in the Greater Antilles and Baha mas. Four of these are larger than any extant West Indian spe- 1 63'W ?. Sombrero 1 62? W Atlantic ? . , Ocean "' ^0Anguilla ? IB'N ^St. Marti . *s. -?'Saba %.St.E * . .. St. Kitts^ n */ ..?? Bartholomew 1 18?N ? d^BARBUDA jstatius " '. tk. "? S, ^ Naw.s.0, dKfc-,Ant ? 17?N % Caribbean ?-:Redonda'-- ?'' MONTSERRAT'^' Sea 0 25 50 0 25 A A Island Bank (200m c ^ 75 50 isobath) Guadeloupe ^| r lies des Saintes m-f 62?W 1 ?? ', igua 17?N ? I . La Desirade ? ^L- Marie ^^ Galante 61?W 1 ..- FIGURE l.?The Leeward Islands of the Lesser Antilles, West Indies. 75 76 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY cies of Tyto (see "Discussion"). In this paper we describe from Barbuda a fifth large West Indian species of Tyto. We note as well the bones of two other species of owls from Barbuda, an island that lacks owls today. MATERIALS AND METHODS.?The partially mineralized bones of Tyto from Barbuda are housed in the Vertebrate Pale ontology Collection of the National Museum of Natural Histo ry, Smithsonian Institution (housing the collections of the former United States National Museum (USNM)). Modern comparative skeletons are from the USNM, the Florida Muse um of Natural History (UF), and the New York State Museum. Prehistoric bones of T ostologa Wetmore (1922b) from His- paniola are from the USNM (especially St. Michel, Cave 1, Haiti, collected in 1928 by A.J. Poole; see Wetmore, 1922b, 1959; Miller, 1926, 1929; Wetmore and Swales, 1931) and the UF (sorted by DWS from various sites excavated from 1978 to 1984 by CA. Woods and colleagues; see Woods et al., 1985). Specimens of T. punctatissima (G.R. Gray) from Holocene cave deposits in the Galapagos Islands are housed at the USNM (Steadman, 1986; Steadman and Zousmer, 1988). Cer tain osteological nomenclature follows Baumel et al. (1993). ACKNOWLEDGMENTS.?We thank the government and peo ple of Barbuda, especially Hillborn Frank, Morris Nedd, and Johnny DiSouza for their generous cooperation during our vis it. Able field associates on Barbuda were R.I. Crombie, J.P. Dean, G.K. Pregill, CL. Watters, and D.R. Watters. Field work was sponsored by the National Science Foundation (grant DEB-8207347 to G.K. Pregill), the Smithsonian Institution (Scholarly Studies Program, Alexander Wetmore Fund), and the Carnegie Museum of Natural History. Museum research was supported in part by the National Science Foundation (grant BSR-8607535 to DWS). S.L. Olson (USNM) and CA. Woods (UF) kindly provided access to collections. L.J. Justice and E.S. Wootan helped to prepare the manuscript. For com ments on the manuscript, we thank M.R. Browning, S.L. Ol son, G.K. Pregill, D.R. Watters, and CA. Woods. The photo graphs are by V.E. Krantz (USNM). Systematics Class AVES Order STRIGIFORMES Family TYTONIDAE The bones are referred to the Tytonidae and the genus Tyto rather than to the Strigidae because of the following characters. Femur with proximal portion of cms condylus lateralis rather pointed (not squared) in posterior aspect; condylus lateralis ex tends farther posteriorly beyond trochlea fibularis in lateral as pect. Coracoid with foramen nervus supracoracoidei relatively small; processus procoracoideus sterno-humerally elongate; fa des articularis clavicularis nonpneumatic and confined to hu- meralmost portion of bone. Pedal phalanges with condyles (distal articulations) narrow relative to overall size of bones. Tyto neddi, new species HOLOTYPE.?Right femur, USNM 359240 (Figure 2), col lected on 19-20 January 1983 by D.W. Steadman, G.K. Pregill, D.R. Watters, R.I. Crombie, and J.P. Dean. TYPE LOCALITY.?Rat Pocket, Gun Shop Cliff, Two Foot Bay, Barbuda. HORIZON AND AGE.?Unstratified; late Quaternary, probably late Pleistocene or ear ly Holocene. Because the deposit lacked organic materi als, such as charcoal or unmineralized bone for radiocarbon dating, we were unable to re fine the chronology of this site beyond being late Quaternary. PARATYPES.?Left coracoid, USNM 359245 (Figure 3); left pedal digit I, phalanx I, USNM 359242 (Figure 4); left pedal digit II, phalanx 1, USNM 359243 (Figure 5); left pedal digit III, phalanx 2, USNM 359241 (Figure 6); juvenile pedal phalanx (either dig it II, phalanx 1; digit III, phalanx 3; or digit IV, phalanx 4), USNM 359244 (not figured). FIGURE 2.?Right femur of Tyto in caudal aspect: A, T. ostologa, St. Michel (Cave 1), Haiti, USNM uncataloged; B, T neddi, holotype, Barbuda, USNM 359240; C, T. alba furcata, male, Jamaica, USNM 553575. (Scale=10 mm.) NUMBER 89 77 All material collected 19-20 January 1983 by D.W. Steadman, G.K. Pregill, D.R. Watters, R.I. Crombie, and J.P. Dean. ETYMOLOGY.?We name this species after Mr. Morris Nedd, a resident of Barbuda generally known by his nickname "Tomac." Barbuda's premier naturalist, Tomac guided us to many caves in 1983, including the type locality of Tyto neddi. DIAGNOSIS.?A large species of Tyto that is slightly smaller than T. ostologa, slightly larger than T. noeli Arredondo (1972a), and either much larger or much smaller than all other New World insular species of Tyto (Tables 1-5). Compared to Tyto ostologa, femur with larger and deeper impressio ansae musculo iliofibularis, deeper depression on medial side of condylus medialis, and tuberculum musculo gas trocnemius lateralis placed more proximally; coracoid with larger foramen nervus supracoracoidei; and pedal phalanges proportionately more robust. Discussion BIOGEOGRAPHY.?Tyto neddi is part of a late Quaternary Barbudan avifauna that included 15 species of birds that no longer occur on the island (Pregill et al., 1994). Antigua and Barbuda were coalesced into a single, large island during Qua ternary glacial intervals (Pregill et al., 1988). Thus it is likely that T. neddi also occurred on Antigua. Tyto neddi also is the sixth extinct species of bam owl to be described from West In dian Quaternary fossil deposits and is the first described from the Lesser Antilles. Arredondo (1976) reviewed the large ex tinct tytonids in the Greater Antilles and Bahamas, particularly those from Cuba, where two large species once existed (T. noe li and the extremely large T. riveroi Arredondo, 1972b). Single large species of Tyto are known from Hispaniola {T. ostologa, which is geographically the closest to T. neddi of all large con geners (see Figure 8)) and the Bahamas {T. pollens Wetmore, 1937). A smaller extinct species {T. cavatica Wetmore, 1920) inhabited Puerto Rico (see also Wetmore, 1922a). Although no extinct species of Tyto are known from Jamaica (Olson and Steadman, 1977; Pregill et al., 1991), this may be an artifact re flecting how few avian fossils have been recovered and studied from Jamaica. From the same bone deposit as the type material of Tyto ned di is an ungual phalanx (digit II, phalanx 3), USNM 453559 FIGURE 3.?Left coracoid of Tyto in dorsal aspect: A, T. ostologa, St. Michel (Cave 1), Haiti, USNM uncataloged; B, T neddi, paratype, Barbuda, USNM 359245; C, T. alba furcata, male, Jamaica, USNM 553575. (Scale=10 mm.) (Figure 7), that we cannot distinguish quantitatively (Table 6) or qualitatively from digit II, phalanx 3, of modem specimens of T. alba (Scopoli). We assign the specimen to digit II, pha lanx 3, based on the uniform roundness of the dorsal surface, the relatively flattened ventral surface, and the relatively rounded (less oblong) articular surface. USNM 453559, which represents an adult bird, is not an adequate basis for species- level identification. Its similarity in size to the same phalanx in T alba means that USNM 453559 is smaller than in T neddi, even though digit II, phalanx 3, is not available for T neddi. The taxa of Tyto that reside today in the Greater Antilles are TABLE 1.?Measurements (mm) of the coracoid in New World species of Tyto, with mean (x), range, and sample size (n). (Glenoid facet=facies articularis humeralis; coracoidal foramen - foramen nervus supracoracoidei.) Species T. neddi T. ostologa T. alba pratincola T. alba furcata T. glaucops T. punctatissima Locality Barbuda Hispaniola North America Jamaica, Haiti, Cuba Hispaniola Galapagos Islands Length of glenoid x 9.4 10.7 7.3 7.8 6.2 5.2 range 10.6-10.7 6.7-8.2 7.0-8.8 5.6-6.8 4.9-5.4 facet n 1 4 11 8 2 4 Width at coracoidal foramen X 7.4 8.9 6.2 6.4 6.1 4.9 range 8.7-9.3 5.5-6.6 5.9-7.0 6.0-6.2 4.8-5.0 n 1 3 11 8 2 2 Depth at coracoidal foramen x 3.9 5.2 3.7 4.1 3.4 2.9 range 4.6-5.7 3.1^t.O 3.7^1.4 3.3-3.5 2.8-3.0 n l 6 11 8 2 4 78 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 4.?Left pedal digit I, phalanx 1, of Tyto in dorsal (A-C), plantar (D-F), and lateral (G-l) aspects: A,D,G, T ostologa, St. Michel (Cave 1), Haiti, USNM uncataloged; B,E,H, T neddi, paratype, Barbuda, USNM 359242; C,F,I, T. alba furcata, male, Jamaica, USNM 553575. (Each scale=10 mm.) FIGURE 5.?Left pedal digit II, phalanx 1, of Tyto in dorsal (A-C) and plantar (D-F) aspects: A,D, T ostologa, St. Michel (Cave 1), Haiti, USNM uncataloged; B,E, T neddi, paratype, Barbuda, USNM 359243; C,F, T. alba fur cata, male, Jamaica, USNM 553575. (Scale=10 mm.) FIGURE 6.?Left pedal digit III, phalanx 2, of Tyto in dorsal (A-c), plantar (D-F), and lateral (G-l) aspects: A,D,G, T. ostologa, St. Michel (Cave 1), Haiti, USNM uncataloged; B,E,H, T neddi, paratype, Barbuda, USNM 359241; C,F,I, T. alba furcata, male, Jamaica, USNM 553575. (Each scale=10 mm.) NUMBER 89 79 T. alba furcata (Temminck) in the Bahamas, Cuba, Jamaica, and Cayman Islands, T. alba niveicauda Parkes and Phillips (1978) on Isle of Pines, and T. glaucops (Kaup), endemic to Hispaniola (Ridgway, 1914:612-613; Wetmore and Swales, 1931; Parkes and Phillips, 1978). The only race of T. alba cur rently recognized for North American populations, T a. prat- incola (Bonaparte), has been recorded in the nonbreeding sea son on Cuba (Garrido, 1978) and Hispaniola (Schwartz and Klinikowski, 1965). In the Lesser Antilles, T. alba may be rep resented by T. "a." nigrescens (Lawrence) on Dominica and T "a." insularis (Pelzeln) on St. Lucia, St. Vincent, Bequia, Car- riacou, Union, and Grenada (Peters, 1940; Bond, 1956, 1980; Evans, 1990). These two forms are smaller and darker than other Antillean subspecies of T. alba and, like T. glaucops, may deserve recognition as a single distinct species (Ridgway, 1914:613-615). The current absence of Tyto on the northern (leeward) islands of the Lesser Antilles (including Barbuda) is without obvious ecological or biogeographic explanation. Thus it is not surprising that a smaller species of Tyto (but larger than T "a." nigrescens or T. "a." insularis) once occurred on Barbuda. The West Indian species of Tyto can be arranged into a size progression (see references above and Tables 1-6, herein) from smallest to largest, as follows (*=extinct): (1) T glaucops: (2) *T. cavatica; (3) T. [alba!] nigrescens, T. [alba!] insularis; (4) T. alba furcata, T. alba pratincola, T. sp. (Barbuda); (5) *T. noeli; (6) *T neddi; (7) *T ostologa; (8) *T. pollens; and (9) *T. riveroi. We have not seen skeletons of T. nigrescens or T. insularis, although skins of these taxa are consistently smaller than those of T. alba furcata or T. a. pratincola in all external measurements (Ridgway, 1914:601-615). Also, whereas mea surements of the skeletal elements in Tables 1-6 are not avail able for T. riveroi, the measurements and photographs of other elements of T. riveroi reveal a size about 10% larger than that of T. ostologa and 30%-40% larger than that of T. noeli (Arredondo, 1972a, 1972b, 1976, 1982). The only other New World species of tytonid is T. punctatissima, which is endemic to the Galapagos Islands (Steadman, 1986) and is smaller than any of the West Indian species (Tables 1, 2). Large extinct species of Tyto are not confined to the West In dies. Tyto balearica Mourer-Chauvire, Alcover, Moya, and Pons (1980) and T melitensis (Lydekker, 1891) were described from middle and late Pleistocene deposits on the Mediterra nean islands of Mallorca and Menorca and of Malta, respec tively (Alcover et al., 1992). The geochronologic and geo graphic range of T. balearica has been extended to the late Miocene and Pliocene of mainland Spain (Mourer-Chauvire and Sanchez, 1988; Cheneval and Adrover, 1995). From the Miocene of the Gargano Peninsula in Italy are two species of Tyto {robusta Ballmann, 1973, gigantea Ballmann, 1976) that are each larger than any living congeners (Ballmann, 1973, 1976; Olson, 1978; Mourer-Chauvire et al., 1980). In spite of a TABLE 2.?Measurements (mm) of the femur in New World species of Tyto, with mean (x), range, and sample size (?). Values for T noeli are from Arredondo (1976). Species T. neddi T ostologa T. noeli T pollens T alba pratincola T. alba furcata T glaucops T punctatissima Locality Barbuda Hispaniola Cuba Bahamas North America Jamaica, Haiti, Cuba Hispaniola Galapagos Islands x 15.4 17.3 14.4 18.1 11.8 11.9 11.0 8.5 Distal width range 16.5-18.3 14.0-14.8 10.9-12.9 11.4-12.8 10.8-11.1 8.4-8.6 n 1 9 2 1 11 8 2 4 Depth X 11.1 12.9 -- 8.4 7.7 6.6 5.4 of inner condy range 12.5-13.6 -- 7.8-9.4 7.1-8.9 6.5-6.7 5.2-5.6 e n 1 5 -- 11 8 2 4 Least X 5.1 6.4 -- 4.0 4.7 4.2 3.6 depth between condyles range n 1 6.0-6.7 5 -- 3.6-4.7 11 4.4-5.1 8 4.1^*4 2 3.5-3.8 5 TABLE 3.?Measurements (mm) of pedal digit I, phalanx 1, in New World species of Tyto, with mean (x), range, and sample size (?). Species T. neddi T. ostologa T. alba pratincola T. alba furcata T. glaucops Locality Barbuda Hispaniola North America Jamaica, Cuba Hispaniola Statistic X range X range ri X range n X range n X range n Total length 18.3 1 23.5 23.1-24.3 5 14.0 13.4-14.9 11 15.1 14.0-15.8 4 12.0 1 Proximal width 7.3 1 7.7 7.5-7.9 5 4.9 4.6-5.2 11 5.2 4.9-5.3 4 4.6 1 Proximal depth 7.6 1 9.1 8.7-9.6 5 5.4 4.9-5.9 11 5.8 5.5-6.0 4 4.5 1 Least width shaft 4.6 1 5.3 5.2-5.3 5 3.0 2.8-3.3 11 3.3 3.1-3.4 4 2.7 1 Least depth shaft 3.7 1 4.5 4.4-4.9 5 2.5 2.2-2.8 11 2.7 2.4-2.9 4 2.1 1 Distal width 5.5 1 6.1 6.0-6.3 5 3.6 3.1^.1 11 3.7 3.6-3.8 4 3.2 1 Distal depth 6.1 1 6.0 6.8-7.1 5 4.4 4.1-4.8 11 4.7 4.4-4.9 4 3.8 1 80 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 4.?Measurements (mm) of pedal digit II, phalanx 1, in New World species of Tyto, with mean (x), range, and sample size (n). Species T. neddi T. ostologa T alba pratincola T alba furcata T. glaucops Locality Barbuda Hispaniola North America Jamaica, Cuba Hispaniola Least shaft width x range n 5.1 6.0 3.5 3.7 3.2 5.9-6.0 3.3-3.8 3.6-3.8 1 3 11 4 1 Least shaft depth x range n 4.3 5.4 3.1 3.3 2.6 5.3-54 2.6-3.6 3.2-3.4 1 3 11 4 1 X 6.2 7.0 4.1 4.5 3.7 Distal width range n 1 6.9-7.1 3 3.8^1.5 11 4.3^1.7 4 1 TABLE 5.?Measurements (mm) of pedal digit III range, and sample size (?). phalanx 2, in New World species of Tyto, with mean (x), Species T. neddi T ostologa T. alba pratincola T. alba furcata T. glaucops Locality Barbuda Hispaniola North America Jamaica, Cuba Hispaniola Statistic X range n X range n X range n X range n X range n Total length 18.0 1 23.8 22.8-244 11 14.6 14.2-15.6 11 15.6 14.9-16.3 4 12.4 1 Proximal width 7.0 1 7.5 7.2-8.0 13 4.5 4.3-5.1 11 4.9 4.7-5.1 4 4.1 1 Proximal depth 6.5 1 7.9 7.5-8.5 11 5.0 4.9-5.3 11 5.4 5.2-5.5 4 4.2 1 Least shaft Least shaft width 4.4 1 5.0 4.7-5.5 13 3.0 2.7-3.3 11 3.2 3.1-3.3 4 2.8 1 depth 3.8 1 4.6 4.4-4.8 13 2.7 2.5-3.0 11 2.8 2.7-2.9 4 2.3 1 Distal width 6.1 1 6.4 5.9-6.7 13 3.8 3.(5-4.2 11 4.0 3.7-4.3 4 3.2 1 Distal depth 5.6 1 7.3 7.0-7.6 11 4.4 4.1?4.7 11 4.6 4.5-4.1 4 3.4 1 TABLE 6.?Measurements (mm) of the ungual phalanx (digit II, phalanx 3) in New World species of Tyto, with mean (x), range, and sample size (n). Species T. ostologa T alba pratincola T. alba furcata Tyto sp. T. glaucops Locality Hispaniola North America Jamaica, Cuba Barbuda Haiti Articulation width x range n 7.6 7.2-8.1 4.7 4.4-5.0 4.8 4.5-5.0 5.3 4.4 6 8 4 1 1 Articulation depth x range n 9.6 9.4-9.8 6 5.6 5.2-5.9 8 5.7 5.2-6.1 4 5.9 1 5.0 1 FIGURE 7.?Ungual phalanx (digit II, phalanx 3) of Tyto in lateral aspect: A, T. ostologa, St. Michel (Cave 1), Haiti, USNM uncataloged; B, Tyto sp., Barbuda, USNM 453559; C, T. alba furcata, male, Jamaica, USNM 553575. (Scale = 10 mm.) fairly rich fossil record of birds, no large extinct species of Tyto have been reported from the Canary or other island groups in the North Atlantic (Baez, 1992; Alcover and McMinn, 1995). No species of owls live on Barbuda today, where the prehis toric bones now reveal the former occurrence of at least three owl species (two tytonids and one strigid). Nowhere in the West Indies today does more than a single resident species of tytonid owl survive. Prehistorically, however, three species of Tyto are known from Cuba (Arredondo, 1976, 1982) and two from Hispaniola (Wetmore and Swales, 1931). The single spe cies of strigid owl from the Barbudan caves, Athene cunicular- ia (Molina), is the only species of strigid recorded anywhere in the Lesser Antilles, whereas individual major islands in the Greater Antilles once sustained at least three to seven species of strigid owls in the genera Otus, Gymnoglaux, Bubo, Orn- imegalonyx, Glaucidium, Athene, Pulsatrix, Asio, and Pseudo- scops (Arredondo, 1976, 1982; Arredondo and Olson, 1994). NUMBER 89 FIGURE 8.?The West Indies. EVOLUTION AND PALEOECOLOGY.?The available speci mens of T neddi are not adequate to evaluate whether it is de rived from a Greater Antillean large species of Tyto or whether it evolved autochthonously from a smaller Lesser Antillean species. The evolution of tytonid owls in the West Indies may be linked to their prey (primarily rodents and insectivores, to a lesser extent ground sloths, primates, bats, amphibians, rep tiles, and birds) and perhaps as well to interactions with strigid owls. This can be evaluated only through the fossil record be cause most indigenous species of West Indian barn owls and nonvolant mammals became extinct in the late Quaternary. West Indian insectivores, primates, and ground sloths were confined to the Greater Antilles (Morgan and Woods, 1986; MacPhee and Iturralde-Vinent, 1994, 1995). The short toes of T. neddi (relative to those of T ostologa) may reflect a diet fo cused more exclusively on rodents. The longer toes of the His- paniolan T. ostologa may have been advantageous when hunt ing arboreal primates or thick-skinned edentates. Two great radiations of rodents occurred in the West Indies: caviomorphs, mainly in the Greater Antilles and Bahamas (Woods, 1989), and oryzomyines, primarily in the Lesser Anti lles (Ray, 1962; Woods, 1989; Pregill et al., 1994). Nearly all of the species in the former radiation and every species in the latter one are now extinct. All of the extinct caviomorph ro dents and many of the extinct oryzomyine rodents are larger than the typical prey of Tyto alba today. Thus an approximate correlation with size of prey, especially of rodents, might ac count for the large size of most of the extinct West Indian spe cies of Tyto. The undescribed, extinct oryzomyine from Barbu da (and Antigua) was larger than a large packrat {Neotoma spp.) but smaller than a muskrat {Ondatra zibethicus (Linnae us)). Considering age-related size variation, adults of this ro- 82 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY dent seem well suited as prey for T. neddi, whereas the juve niles would be appropriate for the smaller species of Tyto that lived on Barbuda. The extinction of the various large species of Tyto was prob ably related to loss of their preferred prey species. Because the stratigraphy and/or chronology of West Indian fossil Tyto are so poorly documented, it is uncertain whether most of the ex tinct tytonids survived into the Holocene or became extinct in the late Pleistocene. Extensive anthropogenic change has oc curred in the terrestrial habitats of Barbuda and Antigua, both in prehistoric and in historic times (Harris, 1965; Steadman et al., 1984; Pregill et al., 1988). On both Barbuda and Antigua, the extinct oryzomyine rodent survived into the late Holocene, being recorded commonly in archaeological sites (Watters et al., 1984, 1992; Pregill et al., 1994). This would suggest that Tyto neddi also may have survived into the last millennium or two. The Burrowing Owl is too small to have fed upon the large extinct cricetid rodent that dominates the Barbudan fossil as semblages. In the West Indies, Athene cunicularia seems to have eaten mainly insects, amphibians, and reptiles (Steadman et al., 1984). Conversely, predation from the much larger Tyto neddi and Tyto sp. may explain why bones of A. cunicularia occur commonly in the bone deposits on Barbuda and Antigua. Literature Cited Alcover, J.A., F. Florit, C. Mourer-Chauvire, and P.D.M. Weesie 1992. The Avifaunas of the Isolated Mediterranean Islands during the Middle and Late Pleistocene. In K.E. Campbell, editor, Papers in Avian Biology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:273-283. Alcover, J.A., and M. McMinn 1995. Fossil Birds from the Canary Islands. In D.S. Peters, editor, Acta Palaeornithologica: 3 Symposium SAPE; 5 Internationale Sencken- berg-Konferenz, 22-26 Juni 1992. Courier Forschungsinstitut Senckenberg, 181:207-213. Frankfurt am Main, Germany. Arredondo, O. 1972a. Nueva especie de ave fosil (Strigiformes: Tytonidae) del pleistoceno superior de Cuba. Boletin de la Sociedad Venezolana de Ciencias Naturales, 29415-431. 1972b. Especie nueva de lechuza (Strigiformes: Tytonidae) del pleistoceno Cubano. Boletin de la Sociedad Venezolana de Ciencias Naturales, 30:129-140. 1976. The Great Predatory Birds of the Pleistocene of Cuba. In S.L. Olson, editor, Collected Papers in Avian Biology Honoring the 90th Birth day of Alexander Wetmore. Smithsonian Contributions to Paleobi ology, 27:169-187. 1982. Los Strigiformes fosiles del pleistoceno Cubano. Boletin de la So ciedad Venezolana de Ciencias Naturales, 37:33-55. Arredondo, O., and S.L. Olson 1994. A New Species of Owl of the Genus Bubo from the Pleistocene of Cuba (Aves: Strigiformes). Proceedings of the Biological Society of Washington, 107436-444. Baez, M. 1992. Zoogeography and Evolution of the Avifauna of the Canary Islands. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:425-431. Ballmann, P. 1973. Fossile vogel aus dem Neogen der Halbinsel Gargano (Italien). Scripta Geologica, 17:1-75. 1976. Fossile vogel aus dem Neogen der Halbinsel Gargano (Italien) zweiter Teil. Scripta Geologica, 38:1-59. Baumel, J.J., A.S. King, J.E. Breazile, H.E. Evans, and J.C. Vanden Berge, editors 1993. Handbook of Avian Anatomy: Nomina Anatomica Avium. Second edition, 779 pages. Cambridge, Massachusetts: Nuttall Ornithology Club. Bond, J. 1956. Check-list of Birds of the West Indies. 214 pages. Philadelphia: Academy of Natural Sciences of Philadelphia. 1980. Birds of the West Indies. Fourth edition, 256 pages. Boston: Hough ton-Mifflin Co. Cheneval, J., and R. Adrover 1995 ("1993"). L'avifaune du Miocene superior dAljezar B (Los Aljeza- res, Province de Teruel, Espagne), systematique et paleoecologie. Paleontologia i Evolucio, 26-27:133-144. [Date on title page is 1993; actually published in 1995.] Evans, P.G.H. 1990. Birds of the Eastern Caribbean. 162 pages. London: Macmillan Educational Ltd. Garrido, O. 1978. Nuevo record de la lechuza norteamericana, Tyto alba pranticola [sic] (Bonaparte), en Cuba. Misceldnea Zoologica, 7:1-4. Harris, D.R. 1965. Plants, Animals, and Man in the Outer Leeward Islands, West In dies. University of California Publications in Geography, 18: 1-164. Lydekker, R. 1891. Catalogue of the Fossil Birds in the British Museum (Natural His tory). 368 pages. London: British Museum (Natural History). MacPhee, R.D.E., and M.A. Iturralde-Vinent 1994. First Tertiary Land Mammal from Greater Antilles: An Early Mi ocene Sloth (Xenarthra, Megalonychidae) from Cuba. American Museum Novitates, 3094:1-13. 1995. Earliest Monkey from Greater Antilles. Journal of Human Evolu tion, 28:197-200. Miller, G.S., Jr. 1926. Exploration of Haitian Caves near St. Michel, Haiti. Smithsonian Miscellaneous Collections, 78:36-40. 1929. A Second Collection of Mammals from Caves near St. Michel, Haiti. Smithsonian Miscellaneous Collections, 81:1-30. Morgan, G.S., and CA. Woods 1986. Extinction and Zoogeography of West Indian Land Mammals. Bio logical Journal of the Linnean Society, 28:167-203. Mourer-Chauvird, C, J.A. Alcover, S. Moya, and J. Pons 1980. Une nouvelle forme insulaire d'effraie geante, Tyto balearica n. sp. (Aves, Strigiformes) du Plio-Pleistocene des Baleares. Geobios, 13:803-811. Mourer-Chauvire\ C, and A. Sanchez M. 1988. Presence de Tyto balearica (Aves, Strigiformes) dans des gisements continenteaux du Pliocene de France et d'Espagna. Geobios, 21:639-644. NUMBER 89 83 Olson, S.L. 1978. A Paleontological Perspective of West Indian Birds and Mammals. Academy of Natural Sciences of Philadelphia Special Publication, 13:99-117. Olson, S.L., and D.W. Steadman 1977. A New Genus of Flightless Ibis (Threskiornithidae) and Other Fossil Birds from Cave Deposits in Jamaica. Proceedings of the Biological Society of Washington, 90:447?457. Parkes, K.C., and A.R. Phillips 1978. Two New Caribbean Subspecies of Bam Owl (Tyto alba), with Re marks on Variation in Other Populations. Annals of the Carnegie Museum, 47:479-492. Peters, J.L. 1940. Check-list of Birds of the World. Volume 4, 291 pages. Cambridge, Massachusetts: Harvard University Press. Pregill, G.K., R.I. Crombie, D.W. Steadman, L.K. Gordon, F. Davis, and W.B. Hilgartner 1991. Modern and Late Holocene Fossil Vertebrates and Vegetation of the Cockpit Country, Jamaica. Atoll Research Bulletin, 353:1-19. Pregill, G.K., D.W. Steadman, S.L. Olson, and F.V. Grady 1988. Late Holocene Fossil Vertebrates from Burma Quarry, Antigua, Lesser Antilles. Smithsonian Contributions to Zoology, 463: 27 pages. Pregill, G.K., D.W. Steadman, and D.R. Watters 1994. Late Quaternary Vertebrate Faunas of the Lesser Antilles: Historical Components of Caribbean Biogeography. Bulletin of Carnegie Mu seum of Natural History, 30:1-51. Ray, CA. 1962. Oryzomyine Rodents of the Antillean Subregion. 356 pages, 36 fig ures, 41 tables. Doctoral dissertation, Harvard University, Cam bridge, Massachusetts. Ridgway, R. 1914. The Birds of North and Middle America, Part VI. Bulletin of the United States National Museum, 50: 882 pages, 36 plates Schwartz, A., and R.F. Klinikowski 1965. Additional Observations on West Indian Birds. Notulae Naturae, 376:1-16. Steadman, D.W. 1986. Holocene Vertebrate Fossils from Isla Floreana, Galapagos. Smith sonian Contributions to Zoology, 413: 103 pages. Steadman, D.W., G.K. Pregill, and S.L. Olson 1984. Fossil Vertebrates from Antigua, Lesser Antilles: Evidence for Late Holocene Human-Caused Extinctions in the West Indies. Proceed ings of the National Academy of Sciences of the United States of America, 81:4448-4451. Steadman, D.W., and S. Zousmer 1988. Galapagos: Discovery on Darwin's Islands. 208 pages. Washington, D.C: Smithsonian Institution Press. Watters, D.R., J. Donahue, and R. Stuckenrath 1992. Paleoshorelines and the Prehistory of Barbuda, West Indies. In L.L. Johnson, editor, Paleoshorelines and Prehistory: An Investigation of Method, pages 15-52. Boca Raton, Florida: CRC Press. Watters, D.R., E.J. Reitz, D.W. Steadman, and G.K. Pregill 1984. Vertebrates from Archaeological Sites on Barbuda, West Indies. An nals of the Carnegie Museum, 53:383?412. Wetmore, A. 1920. Five New Species of Birds from Cave Deposits in Porto Rico. Pro ceedings of the Biological Society of Washington, 33:7 7-81. 1922a. Bird Remains from the Caves of Porto Rico. Bulletin of the Ameri can Museum of Natural History, 46:297-333. 1922b. Remains of Birds from Caves in the Republic of Haiti. Smithsonian Miscellaneous Collections, 74:1?4. 1937. Bird Remains from Cave Deposits on Great Exuma Island in the Ba hamas. Bulletin of the Museum of Comparative Zoology, Harvard College, 80427^441. 1959. Birds of the Pleistocene in North America. Smithsonian Miscella neous Collections, 138:1-24. Wetmore, A., and B.H. Swales 1931. The Birds of Haiti and the Dominican Republic. Bulletin of the United States National Museum, 155:1?483. Woods, CA. 1989. The Biogeography of West Indian Rodents. In CA. Woods, editor, Biogeography of the West Indies: Past, Present, and Future, pages 741-798. Gainesville, Florida: Sandhill Crane Press, Inc. Woods, CA., J.A. Ottenwalder, and W.L.R. Oliver 1985. Lost Mammals of the Greater Antilles: The Summarized Findings of a Ten Weeks Field Survey in the Dominican Republic, Haiti and Pu erto Rico. Dodo, Jersey Wildlife Preservation Trust, 22:23?42. The History of the Chatham Islands' Bird Fauna of the Last 7000 Years?A Chronicle of Change and Extinction Philip R. Millener ABSTRACT Over the past 150 years, thousands of fossil bones of extinct and living species of birds have been collected from the Chatham Islands, an isolated island group some 860 km east of New Zealand. Recent field research (1988-1993) has dramatically aug mented the earlier collections and has provided, for the first time, a sound stratigraphic and radiometric chronology for this rich sub- fossil avifauna. Most of these bones have been found naturally deposited in the buried soil horizons of coastal sand dunes or in limestone caves, but some are of archeological origin, deposited by human agency in coastal dune middens or, occasionally, in dwelling caves. Most, if not all, of the avian remains are of early Holocene or younger age, as indicated by a series of some 65 accelerator-mass-spectometry radiocarbon dates ranging from ca. 7000 yrs BP to ca. 300 yrs BP. The Holocene fossil record for, and patterns of evolution and extinction within, the Chatham Islands' avifauna are documented and discussed. Taxonomic studies indi cate that several taxa, all extinct, can no longer be considered inseparable from their mainland counterparts; among these are as yet undescribed species of Eudyptes, Tadorna, Mergus, and Nestor. Introduction Dune sands and cave sediments on the Chatham Islands (Fig ure 1) have yielded thousands of Holocene fossil bones of ex tinct and living species of birds (Forbes, 1892a, 1892b, 1892c, 1893a, 1893b; Andrews, 1896a, 1896b; Scarlett, 1955; Daw son, 1957, 1958, 1959, 1960, 1961a, 1961b; Simmons, 1964; Olson, 1975, 1984, 1990; Sutton and Marshall, 1977; Sutton, 1979,1982; Millener, 1981, 1991, 1996; Tennyson and Millen er, 1994). The Museum of New Zealand, Wellington, for ex ample, holds more than 250,000 Chatham Islands fossil speci- Philip R. Millener, Geology Department, University of Tennessee at Chattanooga, 615 McCallie Avenue, Chattanooga, Tennessee 37403- 2598, United States. mens, whereas extensive additional material is housed at the Canterbury Museum, Christchurch, and in the Natural History Museum, London. The diversity of this fossil avifauna, which included many endemic land birds (among them several flight less forms), numerous waterfowl, and colonies of breeding subtropical and subantarctic seabirds, contrasts markedly with that of the present. The total number of bird species, living and extinct, recorded from the Chatham Islands is more than 100. The majority of them, some 60 species, are marine birds; among them are albatrosses, petrels and shearwaters, penguins, cormorants, waders, and gulls and terns (see Table 1). Only about 25 of these marine species currently breed in the Chathams, although formerly perhaps as many as seven more did so (Bourne, 1967). The remaining seabirds are summer mi grants, occasional visitors, or vagrants. The rest of the birds (those of terrestrial and freshwater habitat) can be divided into three groups: (1) those present prior to first human contact; (2) those that self-colonized more recently, some of them within the historic period (e.g., Welcome Swallow {Hirundo tahitica), White-faced Heron {Ardea novaehollandiae), Spur-winged Plover {Vanellus miles)); and (3) those that have been deliber ately introduced by Europeans (e.g., Black Swan {Cygnus atra- tus), Weka {Gallirallus australis), House Sparrow {Passer do- mesticus), Blackbird {Turdus merula), Song Thrush {Turdus philomelos), Starling {Sturnus vulgaris)). This paper is primari ly concerned with the composition of the prehistoric Chatham Island avifauna, so the recent self-colonists and the deliberately introduced species are not considered further. Nomenclature for species' binomials and English names of modern birds fol lows Turbott (1990) unless otherwise noted. METHODS Radiometric ages quoted in this paper were determined by the Rafter Radiocarbon Laboratory, Institute of Geological and Nuclear Sciences, Gracefield, New Zealand, using accelerator 85 86 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 1.?Avian taxa identified from fossil deposits in the Chatham Islands. Fossil bones often cannot be identi fied to subspecies. Nevertheless, the subspecific epithets given in this table are those of the subspecies known to occur, or to have occurred, in the Chathams region. Nomenclature follows Turbott (1990) and Olson (1973; Ral lidae) except where more recent revisions apply. For common names, "Chatham Island" is used for taxa endemic, fossil or living, to any of the islands in the Chathams' group. (* = extinct in Chathams in prehistoric period; *c.=extinct in Chathams within historic period, with date of last sighting or specimen; B=breeding; FB= formerly breeding; V= visitor or vagrant.) Species PR0CELLAR1IF0RMES Diomedea exulans/D. e. epomophora D. e. sanfordi D. cauta eremita D. bulleri platei Phoebetria palpebrata Puffinus carneipes P. bulleri P. griseus P. tenuirostris P. gavia/huttoni P. assimilis elegans Pelecanoides u. urinatrix Procellaria ?cinerea P. parkinsoni P. ?westlandica P. ?aequinoctialis Daption capense australe Fulmarus glacialoides Macronectes Phalli Pachyptila turtur P. crassirostris pyramidalis P. vittata Pterodroma nigripennis P. axillaris P. cf. inexpectata P. macroptera gouldi P. magentae P. lessonii Oceanites nereis Pelagodroma marina maoriana Fregetta tropica SPHENISCIFORMES Aptenodytes patagonicus Megadyptes antipodes Eudyptula minor Eudyptes species undescribed PELECANIFORMES Morus serrator Sula dactylatra Phalacrocorax carbo novaehollan- diae Leucocarbo onslowi Stictocarbo featherstoni Fregata ariel ClCONIIFORMES Botaurus poiciloptilus Ixobrychus novaezelandiae Common Name Wandering/Southern Royal Albatross Northern Royal Albatross Chatham Island Mollymawk Northern Buller's Molly mawk Light-mantled Sooty Al batross Flesh-footed Shearwater Buller's Shearwater Sooty Shearwater Short-tailed Shearwater Fluttering/Hutton's Shearwater Subantarctic Little Shear water Common Diving-petrel Grey Petrel Parkinson's Petrel Westland Petrel White-chinned Petrel Snares Cape Pigeon Antarctic Fulmar Northern Giant Petrel Fairy Prion Chatham Island Fulmar Prion Broad-billed Prion Black-winged Petrel Chatham Petrel cf. Mottled Petrel Grey-faced Petrel Magenta Petrel (Taiko) White-headed Petrel Grey-backed Storm Petrel New Zealand White-faced Storm Petrel Black-bellied Storm Petrel King Penguin Yellow-eyed Penguin Blue Penguin "Chatham Island Crested Penguin" Australasian Gannet Masked Booby Black Cormorant Chatham Island Shag Pitt Island Shag Lesser Frigatebird Australasian Bittern New Zealand Little Bittern Status V/V,?FB/ ?FB B B B V V V B V V/V B B V V V V V V B B B B B B ?FB V B V B B V V V B *FB V V B B B V 'c. 1910, FB *?FB Species ANSERIFORMES Cygnus sumnerensis Tadorna, species undescribed Anas s. superciliosa A. chlorotis Anas, species undescribed A. rhynchotis variegata Pachy anas chathamica Aythya novaeseelandiae Mergus, species undescribed FALCONIFORMES Circus approximans Haliaeetus australis Falco novaeseelandiae GRUIFORMES Gallirallus dieffenbachii G. modestus Diaphorapteryx hawkinsi Porzana tabuensis P. pus ilia Fulica chathamensis CHARADRIIFORMES Haematopus chathamensis Thinornis novaeseelandiae Arenaria interpres Coenocorypha pus ilia C. chathamica Calidris canutus Numenius phaeopus ?hudsonicus Catharacta skua lonnbergi Larus d. dominicanus L. novaehollandiae scopulinus Sterna albostriala S. striata S. vittata/paradisea S. nereis COLUMBIFORMES Hemiphaga chathamensis PSITTACIFORMES Nestor, species undescribed Cyanoramphus novaezelandiae chathamensis C. auriceps forbesi CUCULIFORMES Chrysococcyx 1. lucidus PASSERIFORMES Anthus novaeseelandiae chathamensis Bowdleria rufescens Gerygone albofrontata Rhipidura fuliginosa penita Common Name New Zealand Swan Chatham Island Shelduck Grey Duck Brown Teal Chatham Island Teal New Zealand Shoveler Chatham Island Duck New Zealand Scaup Chatham Island Merganser Australasian Harrier Chatham Island Sea- eagle New Zealand Falcon Dieffenbach's Rail Chatham Island Rail Hawkins' Rail Spotless Crake Marsh Crake Chatham Island Coot Chatham Island Pied Oystercatcher Shore Plover Turnstone Chatham Island Snipe Extinct Chatham Island Snipe Lesser Knot American Whimbrel Brown Skua Southern Black-backed Gull Red-billed Gull Black-fronted Tern White-fronted Tem Antarctic/Arctic Tern Fairy Tern Chatham Island Pigeon "Chatham Island parrot" Chatham Island Red- crowned Parakeet Chatham Island Yellow- crowned Parakeet Shining Cuckoo Chatham Island Pipit Chatham Island Fembird Chatham Island Warbler Chatham Island Fantail Status *FB *FB B *c. 1915, FB *FB *c. 1925, ?FB *FB *FB *FB B *?FB *c.l890,FB *c. 1840, FB ?c. 1900, FB *FB B V *FB B B V B *FB V V B B V B V/V V B ?FB B B B B *c. 1900, FB B B NUMBER 89 87 Species Petroica macrocephala chathamensis P. traversi Anthornis melanura melanocephala Common Name Chatham Island Tomtit Black Robin Chatham Island Bellbird TABLE 1.?Continued Status B B "c. 1906, FB Species Prosthemadera novaeseelandiae chathamensis Palaeocorax moriorum Common Name Chatham Island Tui New Zealand Crow Status B *FB mass spectrometry (AMS) techniques on avian-bone collagen or marine-shell carbonate. Within the text, the reported age given for a specific sample (assigned a Rafter Radiocarbon Laboratory reference number, prefixed by NZA) is the conven tional radiocarbon age before present (Stuiver and Polach, 1977). Such ages are expressed in the form "age ? standard de viation (SD) yrs BP." Calibrated ages to which the appropriate terrestrial or marine calibrations have been applied are ex pressed in the form "CAL BP" (see Stuiver and Braziunas, 1993; Stuiver and Pearson, 1993). Within the Appendix, both conventional and calibrated ages are given for each of the sam ples listed. Locality names (see Figure 1) and grid references for sampled sites are from New Zealand Topographical Map, New Zealand Map Series (NZMS) 260, 1:50000 series, Chatham Islands, Edition 1, 1981 (Chatham Island, sheet 1; Pitt Island, sheet 2). ACKNOWLEDGMENTS This research was carried out during my tenure as Curator of Birds at the Museum of New Zealand (MNZ), Wellington. Funding for field research and for some radiocarbon dating was provided by the Museum, and substantial additional funding for the latter came from several grants from the New Zealand Lottery Grants Board, through the Lottery Science Research Committee. All AMS radiocarbon dating was carried out in the Rafter Radiocarbon Laboratory at the Institute of Geological and Nuclear Sciences, Gracefield. I am indebted to Rodger Sparks, Joe McKee, Nicola Redvers-Newton, and Jocelyn Tumbull for so efficiently processing my numerous radiocar bon samples. I am grateful to members of the Collection Man agement staffs of the Museum of New Zealand, Wellington (Mike Rudge, Sandy Bartle, Noel Hyde, Raymond Coory), the Canterbury Museum, Christchurch (Geoff Tunnicliffe), and the National Museum of Natural History, Smithsonian Institution, Washington, D.C. (Phil Angle, James Dean) for arranging loans of, and for access to, specimens in their care. Frank Cli- mo (MNZ) identified the land snails listed in Table 2.1 thank Allan Munn and his staff at the Department of Conservation (DOC) Chatham Island Field Centre, Bruce McFadgen (DOC, Wellington), and Noel Hyde for assistance with field research. I also thank Norm Heke (MNZ Photographic Unit) for taking the photographs. Institutional abbreviations for registration numbers listed in the Appendix are as follows: Natural History Museum, London (BMNH; formerly British Museum, Natural History); National Museum of New Zealand, Wellington (mod em specimens, MNZ; fossil specimens, MNZ S). Site Descriptions and Avifaunal Analysis Our knowledge of the Chatham Islands' prehistoric bird fau na comes from the detailed examination of the abundant fossil bones naturally deposited in coastal sand dunes and limestone caves as well as the archeological material deposited by human agency in coastal dune middens or dwelling caves. The AMS radiocarbon dates for more than 60 bone and shell samples (see Appendix for data and localities) have provided, for the first time, a sound stratigraphic and radiometric chronology for a broad selection of avian remains from a variety of depositional environments. Comparison of these fossil assemblages with the recent fauna indicates that 21 of the original 36 species of land birds or waterfowl have become extinct since human settle ment began about 450 years ago (McFadgen, 1994) and that breeding populations of several seabirds have been reduced or eliminated. Of the original 100 or so avian taxa recorded, fossil or living, from the Chathams, only 25 marine and 15 terrestrial species (a total of 40) now breed there (see Table 1). Fossil bird bones are known only from Chatham, Pitt, and Mangere islands. The remaining islands and rock stacks are typically steeply cliffed and lack extensive sand dunes, swamps, or caves that could have acted as repositories for bones. SAND DUNE SITES Coastal dune belts, in the form of a series of rows of progra- dational sand dunes, generally running parallel to the shore, and sometimes extending several hundred meters inland of it, are important physiographic features fringing all but the south ern coasts of Chatham and Pitt islands. These dunes now are eroded into sequences of discontinuous ridges and hillocks. They began to form in their present positions only about 6500-7000 years ago, after the sea reached its approximate current level following the last (Otira) glacial low sea level of perhaps -120 meters (Hay et al., 1970). At least four deposi tional episodes, consisting of unstable phases with high rates of deposition followed by stable phases with the establishment of vegetative cover and soil formation, seem to have taken place over the last 7000 years. It is clear that periodic denudation, followed by erosion, must have removed parts of the strati graphic sequence. Buried soils are frequently exposed as undu lating bands, following the surface contours of the dunes upon whose surfaces they were formed (Figures 2-5). These soils consist of variously pale yellow, orange, or chocolate-brown/ black- stained sand up to two meters thick, usually overlain by unconsolidated drift sand, rapidly deposited and marking the onset of the first (unstable) phase of the next depositional cy- 88 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Waitangi West wa Point 43? 45' S CHATHAM ISLAND (Rekohu: Wharekauri) 44? S NEW -35?S -40?S South -45?9/ 0 Stewai 1 \ ZEALAND _5 \ Island^ j t I. 170?E h )/ w 500 k -p I Chatham n 180? Islands 10 km Motutapu Point Tarawhenua Point PITT ISLAND (Rangiauria) SOUTH EAST1. (Rangitira) 176? 30' W FIGURE 1.?Map of the Chatham Islands, showing their location relative to New Zealand (inset), and the posi tions of the 21 numbered sites from which radiocarbon dates (see "Appendix") were obtained. cle. The older soils often exhibit quite complex soil profiles, in dicative of lengthy stable periods. Such soils typically grade from a dark-stained, indurated erosion surface at the top, un derlain by a grey-colored, strongly leached horizon, through variably consolidated, pale- to chocolate-brown sand, often markedly orange-stained through the formation of an incipient iron pan, and all underlain by unconsolidated, unstained sand at the base (see Wright, 1959; McFadgen, 1994). It seems likely that coastal forest vegetation clothed the slopes and ridges of these developing dunes for much of this period (especially dur ing the stable, soil forming phases), whereas swamps and ephemeral brackish-water lakes filled interdune hollows, espe cially in low-lying areas behind the active foredunes. It was this mosaic of coastal forest, scrub, and swampland that pro vided habitat suitable for the various land birds, waterfowl, breeding seabirds, and land snails whose remains are now pre served within the sands. The abundant fossil bones are found in the sands and soils, often in situ as complete associated skele tons in (or recently eroded from) buried soil sequences on the flanks of hillocks or in lag deposits on the floors of the often NUMBER 89 89 NZA 797,1930,1934 Unconsolidated yellow-brown drift sand /? *_ Chocolate/purple-brown, consolidated, paleosol surface grading down to paler, creamy-brown, more friable sand * NZA 1931, 2602, 3245 NZA 3239,3246 * Older paleosol of dark brown consolidated sand. Radiocarbon dates indicate lateral equivalence with ephemeral dune-lake deposits NZA 1929,1935, 2632,3238 Evaponte deposits (carbonate-cemented sands) Ephemeral dune-lake deposits (-2 m a.s.l.) 1 m FIGURE 2.?Schematic stratigraphic cross section of dune sands at Long Beach (localities 3, 4). Locations of radiocarbon-dated samples (including those from comparable stratigraphic horizons at other sites within the same dune series) are marked by asterisks (*) and are identified by their Rafter Radiocarbon Laboratory refer ence (NZA) numbers. Unconsolidated pale gray drift sand Compact midden shell layer (20-40 cm thick) Black-brown, peaty soil with some midden shell, fish ? NZA 2614 bones, bird bones and occasional pebbles, grading down to brown dune sand at its base (-70-80 cm depth) Orange-brown sand/paleosol grading down to paler, less consolidated, fine sand * NZA 1982, 2609, 3189,3287 * NZA 3190 Erosional deflation surface exposed laterally (especially in areas where dune sands have not been protected by overlying midden shell layers) 1 m FIGURE 3.?Schematic stratigraphic cross section of the Eastern Maunganui Dunes (locality 11), drawn from a photograph. Locations of radiocarbon-dated samples (including those from comparable stratigraphic horizons at other sites within the same dune series) are marked by asterisks (*) and are identified by their Rafter Radiocar bon Laboratory reference (NZA) numbers. extensive blow-out deflation hollows. These noncultural death assemblages represent birds whose remains were buried rela tively quickly after death by drift sand and were thus preserved as fossils, but no doubt the bones of many more, which were not so fortuitously interred, have decayed completely. In low-lying areas, where ephemeral dune lakes probably once existed (Figure 2), the most common bird remains are those of the extinct Chatham Island Coot {Fulica chathamen sis; see Andrews, 1896c; Millener 1980, 1981), the extinct swan {Cygnus sumnerensis), and various species of duck. A 90 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Inland dune ridge - see Figure 3 'NZA 2614 NZA 1947,1982, 2585, 2609, 3189, 3190, 3287 Surface-consolidated, dark- brown sand forming the present deflation surface Older foredune ri Younger foredune ridge Seaward margin of inland dune ridge Seaward margin of older foredune sea-level 10 m FIGURE 4.?Schematic stratigraphic cross section of dunes at Tahatika Creek in the Eastern Maunganui Dunes (locality 11). Locations of radiocarbon-dated samples are marked by asterisks (*) and are identified by their Rafter Radiocarbon Laboratory reference (NZA) numbers. Samples listed for the inland dune ridge include those from comparable stratigraphic horizons at several sites along the ridge. Pale gray drift sand, in places stabilized by dune vegetation. Yellow semi-consolidated, cross-bedded sand NZA 3427 * NZA2587 Consolidated, dark-gray sand/occupation-soil surface, with la pebbles and midden debris, underlain by leached gray sand Chocolate-brown consolidated sand grading down to orange-brown with incipient iron-pan at base \j \y ^ Yellow-brown sand grading down to finer, cream sand - with abundant avian remains found in situ, or as lag deposits in the bottoms of dune hollows FIGURE 5.?Schematic stratigraphic cross section of the Taupeka inland dunes (locality 12), drawn from a pho tograph. Locations of radiocarbon-dated samples are marked by asterisks (*) and are identified by their Rafter Radiocarbon Laboratory reference (NZA) numbers. specimen of Fulica from one of these lake beds yielded the old est radiocarbon age (NZA 3238, locality 4; 6879 ? 68 yrs BP) so far obtained for bird bones on main Chatham Island. The un dulating dune slopes and buried soils in the broad zone be- species, both marine and terrestrial, but with forest-dwelling birds such as the Chatham Island Pigeon {Hemiphaga chatha mensis), Dieffenbach's Rail {Gallirallus dieffenbachii), the ex tinct Chatham Island Snipe {Coenocorypha chathamica), tween foredune and ridge crests yielded the greatest variety of Nestor {Nestor, species undescribed, Figure 10), parakeets (Cv- NUMBER 89 91 anoramphus spp.), Chatham Island Tui {Prosthemadera novae seelandiae chathamensis), and Chatham Island Bellbird {An- thornis melanura melanocephala) most often represented. Radiocarbon dates from this complex physiographic zone cov er a broad range, from ca. 700 to ca. 6500 CAL BP, the oldest dates coming from the eroded older inland dunes, and the youngest dates from younger foredune sites, typically only a short distance inland of the present, unconsolidated, active dune/beach zone (see Appendix: locality 2, NZA 1930, NZA 1931; locality 8, NZA 3285, NZA 3426; locality 9, NZA 3608; locality 12, NZA 2587; locality 14, NZA 2588). On the higher slopes and ridge crests, and especially in blow outs in pasture land up to several hundred meters inland, bones of many of these same forest birds occur, but they are outnum bered by bones of seabirds, especially Taikos {Pterodroma ma- gentae), Sooty Shearwaters {Puffinus griseus), Common Div ing-petrels {Pelecanoides urinatrix), and various prions {Pachyptila spp.) and storm-petrels {Oceanites, Pelagodroma, Fregetta). The presence of eggshell fragments and the bones of nestlings of all these seabirds (Bourne, 1967) indicates that these widespread fossil sites, dated at between ca. 700 and ca. 3300 CAL BP, mark the locations of former breeding colonies (see Appendix: locality 4, NZA 794; locality 5, NZA 1932, NZA 1933; locality 6, NZA 795; locality 17, NZA 2777). MIDDEN SITES The prehistoric Moriori left extensive faunal remains at nu merous, widely spread sites in the dunes and as surface scat ters on them (Simmons, 1964; see also Figures 3-5). These kitchen middens are generally dominated by marine shells, but bones of sea lions, seals, fish, and birds also occur, partic ularly in the oldest sites (those dating between ca. 400 and ca. 450 CAL BP). Some earlier archeological workers, who col lected much of their material from lag deposits in deflation hollows and who did not have access to radiocarbon dating, failed to distinguish between natural and midden deposits and tended to ascribe a midden origin to virtually all the bird re mains they found (e.g., Coutts, 1969). Recent stratigraphic studies, supported by a large number of radiocarbon dates on bird bones, demonstrate that much of this so-called midden material has been eroded from naturally accumulated deposits that considerably predate human occupation by hundreds to thousands of years. An example is the abundant material from Sutton's main Waihora site (see Sutton, 1976, 1979, 1981, 1982; Marshall et al., 1987), which, apparently, was obtained by excavation of intact strata and was assumed to be entirely of midden origin, but which has yielded dates of ca. 5750 and ca. 5950 CAL BP (for Dieffenbach's Rail bones from Sut ton's Layers I (NZA 3193) and III (NZA 3194), respectively; see Appendix, locality 1). In a few sites, notably Sutton's CHA and CHB sites at Waihora, and at Tupuangi and Waipaua on Pitt Island, bird remains of genuine midden prov enance certainly do occur, often in great abundance. No dates of greater than ca. 450 CAL BP have yet been obtained for in situ midden material at these and other sites. My data corrob orate McFadgen's (1994) suggestion that first Polynesian set tlement of the Chathams group did not occur until about 450 years ago. The assemblages at these sites indicate that the Mo riori hunted a wide range of species but that certain species (e.g., Taiko, Chatham Island Pigeon, Common Diving-petrel, Dieffenbach's Rail) were sought more intensively than the rest. There can be little doubt that prehistoric hunting had a profoundly deleterious effect on the Chatham Islands bird fauna. It appears, however, that the Moriori neither construct ed permanent dwellings nor lived in long-term encampments, and perhaps it is because they led a rather more itinerant life style that they have left less evidence of their hunting (in the form of extensive bird bone middens) than did, for example, the Maori hunters of mainland New Zealand (see Trotter and McCulloch, 1984). CAVE SITES In several places on Chatham and Pitt islands are limestone crevices and cavities that contain bone deposits because they were used as shelters and nest sites by terrestrial and marine birds. There also are larger caves that acted as pit-fall traps, and these have yielded far more abundant fossil remains. The most significant of these is a small, single-chambered cave on the western edge of the Te Whanga Lagoon, Chatham Island. This cave, Te Ana a Moe (see Simmons, 1964), is developed near the base of a 15 m high cliff of Eocene Te One Lime stone (typically creamy yellow in color, relatively soft, and rich in bryozoan fragments), immediately above its contact with the underlying Te Whanga Limestone (typically grey white, hard, crystalline and, here, where it forms the raised shore-platform, strongly karstic; see Hay et al., 1970; Camp bell, 1996). The cave has a single walk-in entrance about 3 m above present lagoon level and is filled in with stratified sedi ment to a depth of at least 2 m (see Figure 6). Above the al most unfossiliferous basal layers (sands overlain by angular limestone slabs), a distinctive layer of water-rounded lime stone cobbles is overlain by more than a meter of stratified sediment containing shells of at least 17 species of land snails (Table 2) and an extraordinary abundance of avian bones of some nine marine and 21 terrestrial or freshwater species (Ta ble 3). Radiocarbon dates range from ca. 1150 CAL BP (NZA 1948, locality 16, at only 15 cm below the base of the dis turbed surface soil) to ca. 3900 CAL BP (NZA 1989, locality 16, at a depth of 1.3 m). Bird remains were found to be partic ularly concentrated within several short (~2 m), narrow (-0.5 m diameter), blind tunnels leading off the main chamber, at depths of 0.9-1.5 m. Faunal material in these tunnels has yielded radiocarbon ages within the range of 2300-3900 CAL BP (NZA 801, NZA 1989, locality 16). Although a wide 92 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY DEPTH (cm) 0 50 60 100 120 140 150 180 190 200 210 * NZA 1948 * NZA798 * NZA 2778 * NZA 801 * NZA 800 * NZA 1989 ggbil Uneven cave sediment-fill surface 10-50 cm dark-brown organic soil, much disturbed, with bones of sheep (Ovis) and rat (Rattus) 10 cm orange-brown sandy silt, much disturbed, with bones of sheep and rat Surface level of undisturbed sediment 10 cm brownish-cream, homogeneous detrital sand rich in avian bones 30 cm creamy-white detrital sediment, predominantly bryozoan fragments and echinoid spines. Rich in cream- colored avian remains and landsnail shells 20 cm white detrital sediment with fewer 'fines' than the layer above. Rich in avian bones - pale whitish-cream except near walls and intruding roots where they have become orange stained 20 cm darker brown/cream sediment with greater clay fraction. In main cave fewer bones than in the layer above, but dense accumulations in several side tunnels 10 cm rounded, creamy-white, limestone cobbles(up to 70 mm diam.), interspersed with marine shells (Cliione) 5 cm hard, moist, fine, brown clay/mud with an orange (iron- stained) upper surface 20-30 cm large, angular slabs of Te Whanga Limestone with an intermediate layer of cobbles and rounded boulders (to 150 mm diam.) 10 cm - air gap where percolating water has washed away finer sediments 20 cm of friable white-yellow-brown sand, with lenses of orange-brown sand, and a moist water-scoured upper surface Solid rock floor of Te Whanga Limestone (essentially at the contact between the Te Whanga Lst and the overlying Te One Lst in which the cave has developed) FIGURE 6.?Schematic stratigraphic section of sedimentary deposits within Te Ana a Moe Cave (locality 16). Locations of radiocarbon-dated samples are marked by asterisks (*) and are identified by their Rafter Radiocar bon Laboratory reference (NZA) numbers. range of taxa are represented in the deposits, in total many thousands of bones from hundreds of individuals, the follow ing species predominate in the assemblages: Chatham Island Rail {Gallirallus modestus), Dieffenbach's Rail, a merganser {Mergus, species undescribed, Figures 12-14), Chatham Is land Fernbird {Bowdleria rufescens), Magenta Petrel, or Tai- ko, Fairy Prion {Pachyptila turtur), and Common Diving-pe trel. One particularly important find was that of an almost complete individual skeleton of the flightless Chatham Island Duck, Pachyanas chathamica Oliver (1955) (Figure 7). Bones were more abundant in the lower levels (particularly in the 2300-3900 CAL BP strata), but species composition varied little with depth. The fact that the youngest dates obtained for in situ faunal material were ca. 1150 CAL BP (e.g., NZA 1948, locality 16) is taken to indicate that at about this time, when the infilling sediment reached the level of the single walk-in entrance, the cave ceased to be an effective pit-fall trap for birds. NUMBER 89 93 TABLE 2.?Land snails (Gastropoda: Pulmonata) identified from Te Ana a Moe Cave, Chatham Island (Locality 16 (Figure 1)). Taxonomy follows Climo (pers. comm., 1993). Sample 1 (PRM sample 155/91) is from brown sand/soil in the uppermost 20 cm of undisturbed sediment; sample 2 (PRM sample 156/91) is from creamy white bryozoan detrital sand, at 1 m depth (-40 cm below upper surface of undisturbed sediment); sample 3 (PRM sample 120A/91) is from brown gray bryozoan detrital sand at 1.2-1.4 m depth (60-80 cm below upper surface of undisturbed sediment) in a side tunnel in S W quadrant. For each sample, abundance of each species is ex pressed as a percentage of sample size (n). (See also Figure 6.) Family TORNATELLINIDAE PUNCTIDAE CHAROPIDAE FLAMMULINIDAE ROTADISCIDAE CAMAENIDAE Species Lamellidea novoseelandica Serratopunctum serratocostata Litopunctum rakiura Punctum lateumbilicata Alexaoma chathamensis Dellopsis stewartensis Pryhina chathamensis Phenacharopa pseudanguicula Charopa coma Flammocharopa mayhillae Sinployea parva Huonodon hectori Mitodon wairarapa Basimocella maculata Discocharopa eta Cavellia buccinella Thalassohelix sp. 1 (?=164) 0.6 - 74.4 - - - 10.4 1.2 - - 0.6 - - 5.5 0.6 2.4 4.3 Sample 2(?=110) - 0.9 18.2 4.5 - - 21.8 - 6.4 - - 2.7 2.7 35.5 - 8.2 1.8 3(?=128) 1.6 - 40.6 4.7 0.6 0.8 13.3 1.6 2.3 0.8 1.6 1.6 - 18.8 - 7.0 3.9 TABLE 3.?Bird species identified from Te Ana a Moe Cave, Chatham Island (Locality 16 (Figure 1)). See "Appendix" and Table 1. (A=abundant (mini mum number of individuals (MNI) >100); C=common (MNI > 10); R=rare (MNK 10); *=extinct.) Taxon MARINE SPECIES Pelecanoides urinatrix Pachyptila turtur Pachyptila crassirostris Pterodroma nigripennis Pterodroma axillaris Pterodroma magentae Pelagodroma marina Eudyptula minor Larus dominicanus TERRESTRIAL AND FRESHWATER SPECIES *Tadorna, species undescribed *Anas, species undescribed *Pachyanas chathamica *Mergus, species undescribed Falco novaeseelandiae *Gallirallus dieffenbachii *Gallirallus modestus "Diaphorapteryx hawkinsi *Fulica chathamensis *Coenocorypha chathamica Hemiphaga chathamensis *Nestor, species undescribed Cyanoramphus novaezelandiae Cyanoramphus auriceps Anthus novaeseelandiae *Bowdleria rufescens Gerygone albofrontata Rhipidura fuliginosa Petroica traversi *Anthornis melanura Prosthemadera novaeseelandiae Abundance A (at all levels) A (at all levels) R R C A (at all levels) R C R (upper level only) R (MNI=2, in upper level only) R R (MNI=1, in lower level) A (at all levels) R A (at all levels) A (at all levels) R R C R R C R R C (at all levels) R R C (at all levels) R R SWAMP SITES Although peat deposits and more recent swamps are wide spread on Chatham Island, the conditions in them appear to have been unsuitable for the preservation of bones. Peat fires burning to considerable depths have occurred frequently on the island. Furthermore, both peats and more recent swamps seem typically to have been too acidic to allow long-term preserva tion of bone. Paleogeography, Ancestral Immigration, and Avifauna Change It is possible that the Chatham Islands, although isolated by a broad oceanic gap from the New Zealand mainland since the Late Cretaceous, some 80 million years ago, have provided a land mass capable of supporting viable bird populations more or less continuously for perhaps many millions of years (Fleming, 1962, 1975; Cooper and Millener, 1993). There is evidence, however, that during the late Eocene (40 Ma) and again during the Pliocene (5-2 Ma) the only emergent land in the entire Chathams group would have been a few volcanic peaks (Campbell, 1996:36). Thus colonization by the forerun ners of the Holocene species may postdate the Pliocene. Fur ther, there is some suggestion that later, in the Pleistocene, high interglacial sea-levels during the Castlecliffian (ca. 1.4-0.32 Ma) may have inundated all but the highest points of Chatham Island (see Hay et al., 1970). Therefore much, if not all, of any earlier established terrestrial avifauna may have been eliminated during this period. Any ancestral avian colo nists that reached the Chathams since the Cretaceous, includ ing those still arriving today, could have done so only by fly- 94 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 7.?Skeletal elements of P'achyanas chathamica (MNZ S29475, PRM sample #92/91) from Te Ana a Moe Cave, Te Whanga Lagoon, Chatham Island. Total length of cranium+premaxilla is 113.8 mm. ing over water. In the absence, however, of any fossil record of birds on the Chathams beyond the last 7000 years, one can only speculate on the numbers and variety that may have reached the Chathams in earlier times and subsequently died out, leaving no trace of their passing. Possibly through extinc tion of early colonists, but more likely as a consequence of the serendipitous nature of transoceanic colonization ("sweep stakes dispersal"), some significant avian groups are absent from the known avifauna of the Chathams, including ratites (moas and kiwis), Podicipedidae, Coturnix, Strigidae, Alce- dinidae, Acanthisittidae, Callaeidae, and Turnagra. No native frogs, tuatara, or geckos are known, fossil or living, from the Chathams. The one extant species of skink is possibly a geo logically relatively recent arrival. Both extant and extinct Chatham birds are presumably de rived from the same ancestral stocks as are comparable species on the New Zealand mainland. They have, however, evolved in isolation and exhibit many of the same evolutionary features found in bird faunas on other small, isolated, oceanic islands, such as New Caledonia (Balouet and Olson, 1989), the Hawai ian Islands (Olson and James, 1982, 1991; James and Olson, 1991), and many other islands of Australasia and the southwest Pacific (van Tets et al., 1981; Meredith, 1991; Steadman, 1995). Typically, on oceanic islands such as these, which prior to human colonization lacked mammalian predators, birds ex hibit increased body size and may lose their powers of flight (McNab, 1994). The evolutionary history of Chatham Islands birds has followed this same path. Most of the Chatham Islands land birds and waterfowl tend to be larger than their mainland NUMBER 89 95 counterparts, with some very much more so, such as Hawkins' Rail {Diaphorapteryx hawkinsi), the flightless Chatham Island Duck, the Chatham Island Pigeon, and the Chatham Island Bellbird. Further, of the 36 prehistorically known species, at least seven were flightless (two ducks, four rails, and the Chatham Island Fernbird; see Olson, 1990), and three more were weak fliers (a duck, a snipe, and a parrot). With the exception of an apparently new species of crested penguin {Eudyptes, species undescribed; Figures 8, 9), no ma rine taxa are known to have become extinct. The seabirds have been affected, nevertheless, by the changes that occurred fol lowing human settlement. Fossil remains of numerous species of seabirds occur in large concentrations in many places, par ticularly along dune ridges and promontories, on Chatham Is land and on Pitt. Few species of seabirds still breed on these two larger islands; most current breeding colonies are restricted to rugged, small, offshore islands or stacks. Fossil eggshells and bones of very immature chicks indicate the presence of former breeding colonies of a wide range of species. Radiocar bon dating of bones indicates that some of these colonies were occupied from at least 4500 years ago (NZA 1906, locality 18; 4300 ? 150 yrs BP=4545 CAL BP), and presumably much ear lier. Some colonies continued to be occupied into the period of first human settlement, but there is no evidence of their persis tence into the European era and, indeed, little evidence of via ble colonies beyond about 300 years ago. These Chatham Island/Pitt Island breeding species included albatrosses {Di omedea epomophora sanfordi and at least one larger species); mollymawks; shearwaters; diving-petrels; several prions; gad fly petrels, especially the Taiko and another species, apparently similar but not identical to the Mottled Petrel {Pterodroma in- expectata); and storm-petrels. FIGURE 8.?Skulls of Eudyptes spp. (dorsal views). Left to right: Eudyptes pachyrhynchus (MNZ 24546); Eudyptes, species undescribed (MNZ S26908, PRM sample #6/89), from Maunganui, Chatham Island; Eudyptes sclateri (MNZ 18897). (Scale bar=30 mm.) 96 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 9.?Lower mandibles of Eudyptes spp. (left lateral views). Top to bottom: Eudyptes pachyrhynchus (MNZ 24546); Eudyptes, species undescribed (MNZ S30440, PRM sample # 15/92), from Kaingaroa, Chatham Island; Eudyptes sclateri (MNZ 18897). Total length of MNZ S30440 is 119.1 mm. THE COMPOSITION OF THE CHATHAMS AVIFAUNA It is not possible to produce an entirely accurate catalog of the living and extinct birds of the Chathams because in many cases doubts have been raised (and may always exist) over re ported occurrences of several taxa. The exact composition of the terrestrial and freshwater bird fauna is particularly diffi cult to ascertain. Several of the 45 recorded species appear to have been included by various authors as a result of errors of identification or locality (see van Bemmelen, 1993:32), whereas others, regarded by some as members of the indige nous Chathams fauna, seem more likely to have been intro duced by humans. Taxa that I exclude from the analysis of living and extinct fauna are considered below. Apteryx sp. (kiwi): Travers (1866:358) stated that "former ly an apterix [sic], said by the Maoris to have been identical with a New Zealand species, and... a smaller species of the same bird...were found [in the Chathams], but have become extinct...." Hutton (1904), as had Wallace (1893), accepted this statement and included Apteryx in his catalog. Travers (1873:213), however, commented that his son "has now reason for believing that the weka..., the kakapo..., and the kiwi... were erroneously assigned to [it]" and later (Travers, 1883:183) that he himself was "not disposed to accept [these records]." No fossil bones of Apteryx have been found in the Chathams, and the record must remain unsubstantiated. Gallirallus australis (Weka): Bones, reportedly identical to those of this species, have been recovered from several sites on Chatham Island, leading to the possibility that wekas were indigenous in the Chathams prior to the introduction of G. aus tralis hectori in 1905 (Turbott, 1990). There is no unequivocal evidence that any of these bones predate the European intro duction. A mounted specimen of a Weka in the Rijksmuseum van Natuurlijke Historie, Leiden (Temminck collection, acces sioned in 1823), is said to be from the Chatham Islands, but this locality information may be spurious (van Bemmelen, 1993). On biogeographic grounds it seems highly unlikely that a pop ulation of Gallirallus australis, identical to the mainland form, would have evolved independently on the Chathams. If associ ation with pre-European middens could be demonstrated, the most likely explanation would be that the species was taken to the Chathams from mainland New Zealand by the original Mo riori inhabitants. Because there are no reports of Europeans having seen the species prior to 1905, it would seem that the NUMBER 89 97 population, if such existed, must have been eliminated in pre historic times. Gallirallus minor (extinct Weka): Several authors have re ported bones of this species. Olson (1975:76) remarked that the presence of this species in the Chathams would be "a highly unlikely occurence" and indicated that at least the bones men tioned by Falla (1960) are well within the size range for G. di- effenbachii. Gallirallus minor is a species that has never been satisfactorily defined and may eventually prove to be no more than a smaller variant of G. australis. Strigops habroptilus (Kakapo): Travers (1873) mentioned that kakapos (as well as kiwis and wekas, see above) were known to the Maori on the Chathams, and the record was fol lowed by several subsequent authors. Although this statement was later disavowed, Forbes (1892c, 1893a) continued to ac cept the unsubstantiated myth. Dawson (1959), in a detailed analysis, concluded that the only material evidence for the former presence of Strigops rested with a single bone in the Travers collection, which perhaps had not come from the Chathams at all. Subsequently, Dawson (1960) discovered two further bones of Strigops, among uncataloged material in the British Museum, that had allegedly been collected by or for Forbes in the 1890s. Because these three bones are the only ones ever identified among the many thousands of bones ob tained from the Chathams, it seems most likely that their local ity, too, was incorrectly recorded, although there remains a possibility (in my view, highly implausible) that Kakapo were at some time taken to the islands from New Zealand by the Maori. Nestor notabilis (Kea): The occurrence of fossil bones identified as those of the Kea was mentioned in several papers by Forbes (1892c, 1893a). Oliver (1955:542) obviously fol lowed Forbes when remarking that "in pre-European times [the kea was present] on the Chatham Islands." Dawson (1959), however, considered all the Chatham Island bones of Nestor obtained by Forbes to belong to N. meridionalis (Kaka). My own research has shown that the Chatham Islands Nestor (Fig ure 10) is a new, undescribed species, poorly volant and now extinct, but it is structurally more like N. meridionalis than like N. notabilis. Sceloglaux albifacies (Laughing Owl): Forbes (1892c) not ed that among his Chatham Island material were bones that he identified as those of the Laughing Owl. Neither Dawson (in 1958) nor I (in 1984), however, recognized any bones attribut- FlGURE 10.?Pelves of Nestor spp. (dorsal views). Left to right: Nestor notabilis (MNZ 23161); Nestor, species undescribed (NMZ S29990, PRM sample #152/91), from Te One, Chatham Island; Nestor m. meridionalis (NMZ 22504). (Scale bar=30 mm.) 98 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY able to Sceloglaux in Forbes' material in the Natural History Museum, London. It seems that this record is a case of mistak en identity. More recently (1993) I have examined a tarsometa tarsus from Forbes' collection upon which is written in Forbes' distinctive hand "Sceloglaux, Ch.I., HOF." This specimen is clearly attributable to Falco novaeseelandiae. Several other bones of Falco have been recognized in Forbes' collection (see Dawson, 1961b) and many more have been collected from the Chathams in recent years, but none of Sceloglaux. The Chatham Island Sea-eagle, originally described as Ich- thyophaga australis (Harrison and Walker, 1973) but removed to Haliaeetus by Olson (1984), is a particular enigma, although it seems prudent to list it among the indigenous Chatham Is lands birds until unequivocal evidence shows otherwise. Housed in the Natural History Museum, London, is a collection of about a dozen eagle bones found or otherwise acquired by H.O. Forbes in the 1890s, and labelled as being from the Chatham Islands. These bones, parts of four individual birds, certainly belong to the genus Haliaeetus (see Dawson, 1961a; Olson, 1984) and have been considered to differ somewhat from any existing species (although I could not distinguish them from bones of the Alaskan race of Bald Eagle, H leuco- cephalus alascanus Townsend). There is reason to doubt that these bones came from the Chathams (some other bones in Forbes' collections are wrongly labelled) because it is extraor dinary that among all the hundreds of thousands of bones col lected subsequent to Forbes, not one of a sea-eagle has been re covered. Even the appearance of the bones is unlike that of other Chatham Islands fossils, as they seem to have a surface texture more like that of modern material. Another element of doubt about their authenticity is raised by a radiocarbon date (NZA 1548, locality uncertain) obtained from one of the paratypical bones (BMNH A3732). Depending on the calibra tion given, an age as young as ca. 1836 AD can be calculated. The enigma remains unresolved. Perhaps Forbes chanced upon the only bones yet known of an endemic Chatham Sea-eagle, or perhaps the bones are from an existing species, acquired as part of Forbes' reference collection of modern specimens. NOTES ON SELECTED EXTINCT SPECIES FROM THE CHATHAM ISLANDS Dieffenbach's Rail {Gallirallus dieffenbachii): This spe cies is closely related to the widely distributed Banded Rail {G. philippensis) and is presumed to have evolved from the same ancestral stock (Olson, 1975). The anatomical differences be tween Dieffenbach's Rail and the Banded Rail (of which it has previously been considered only a subspecies; see Turbott, 1990) indicate that they are separate species (pers. obs.). Only one live specimen was ever collected, by Ernst Dieffenbach in 1840, who stated (1841:195) that the species "was formerly very common, but since cats and dogs have been introduced it has become very scarce" (see Forbes, 1893b). Dieffenbach's Rail was flightless, weighed perhaps 340-400 g (about twice as much as the Banded Rail, which is a capable flyer), and had a rather dull plumage and a distinctly down-curved bill (Atkin son and Millener, 1991). An inhabitant of forest and scrub, it probably consumed a wide range of foods, including soil inver tebrates probed from soft earth, insects, seeds, and even eggs of ground-nesting birds. Chatham Island Rail {Gallirallus modestus): This diminu tive (body weight 50-70 g; Atkinson and Millener, 1991), flightless species may have evolved from the same stock as Di effenbach's Rail (Olson, 1975). The type specimen was ob tained by H.H. Travers from Mangere Island in 1871, but Oliv er (1955:355-356) stated that "through the work of collectors... aided by cats... [it] was exterminated about twen ty-five years after it was discovered." The bill of the Chatham Island Rail is very long and delicate and must surely have been used as a probe to capture small invertebrates in soft soil or leaf litter. The only observations of its habits are those of Hawkins (in Forbes, 1893b:532): "They nest in holes in the ground... [and] live on insects, principally the sandhoppers which travel into the bush a long way." Until recently, fossil bones of the Chatham Island Rail had not been commonly found, but they have now been recorded from Pitt and Mangere islands and from the Te Ana a Moe Cave beside the Te Whan ga Lagoon on Chatham Island. From this one small cave, many thousands of bones, representing hundreds of individual birds (Table 3), have been excavated from sediments dating from about 1150 years to almost 4000 years ago. Hawkins' Rail {Diaphorapteryx hawkinsi): Fossil bones of this species were first collected by W. Hawkins who sent them to H.O. Forbes in 1892 (see Forbes, 1892a, 1892b, 1892c, 1893a). This large (body weight ~2 kg; Atkinson and Millener, 1991), flightless rail is so distinct from the Gallirallus group that it is placed in its own genus. Its wings were greatly re duced, its legs robust, and its toes elongate. Its long, decurved bill may have been an adaptation for probing into soft earth for soil invertebrates. No living specimen of Hawkins' Rail was ever seen or collected by Europeans, but the substantial num bers of its bones in Moriori middens indicates that it was fre quently hunted for food. Chatham Islands waterfowl: The Chatham Islands former ly supported a wide range of waterfowl, including the extinct Chatham Island Swan and perhaps eight species of duck. The Chatham Island Duck {Pachyanas, Figure 7) and Merganser {Mergus, species undescribed, Figures 12-14) were flightless, and several of the other species (e.g., the Chatham Island Shel- duck, Tadorna, species undescribed, Figure 11) show indica tions of having been weaker flyers than their mainland counter parts. All but the Grey Duck {Anas superciliosa) became extinct in the Chathams following human settlement. The ex tinct swan {Cygnus sumnerensis) would seem to have evolved from the same stock as the Australian Black Swan, which was introduced to the Chathams about 1890 and now flourishes on NUMBER 89 99 FIGURE 11.?Skulls of Anatidae (dorsal views). Left to right: Anas, species undescribed (MNZ S33298, ex. H.O. Forbes collection), from "Chatham Islands"; Anas chlorotis (MNZ 14978); Tadorna variegata (MNZ 16473, male); Tadorna, species undescribed (MNZ S32830, presumed male), from Maunganui, Chatham Island. (Scale bar=30 mm.) the Te Whanga Lagoon (Turbott, 1990). The extinct species, known also from mainland New Zealand, differs in being somewhat larger, with a stouter bill and relatively shorter wings. Its bones have been found in greatest abundance at Waitangi West, at Tioriori, and at Te One Beach, near Red Bluff, in deposits ranging in age from ca. 7600 (cf. NZA 3238, locality 4) to ca. 700 (NZA 2603, locality 11) years old, but they also occur commonly in middens. The Chatham Island Duck, whose bones have been found at a considerable number of sites on Chatham Island, but not elsewhere, was robust, weighing as much as 2.5 kg (more than twice the weight of a Grey Duck), and yet with wings perhaps even a little smaller than in that species (pers. obs.). The new merganser {Mergus, species undescribed), whose bones have been found in abun dance only in Te Ana a Moe Cave, was a little smaller than the recently extinct (ca. 1905), flightless Auckland Island Mergan ser {Mergus australis; see Kear and Scarlett, 1970) and had a shorter bill and even more reduced wings (Millener, pers. obs.). EXTINCTIONS The indigenous bird faunas of remote oceanic islands, hav ing evolved in isolation from most or all vertebrate predators, are extraordinarily vulnerable to the impact of humans (see Milberg and Tyrberg, 1993; Steadman, 1995). The land birds and waterfowl of the Chatham Islands were well adapted to their island environment. Most became larger than their main land counterparts, many became completely or nearly flight less, some probably laid smaller clutches, and none developed or retained a fear of mammalian predators. First human settle ment was almost certainly accompanied by accidental or delib erate forest clearing, and on such small islands there would 100 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 12.?Bones of Mergus, species undescribed (MNZ S30049, PRM sample #113/91; all from 20 liters sed iment), from Te Ana a Moe Cave, Te Whanga Lagoon, Chatham Island. (Scale bar= 100 mm.) have been few refuges for species whose habitat was destroyed. Hunting by humans and predation by rats {Rattus exulans) must have depleted the most vulnerable species, particularly the flightless and ground-nesting ones. Thirteen species (36% of the original terrestrial avifauna) were exterminated between about 450 years ago, the estimated date of first human arrival, and about 300 years ago, after which few, if any, bones of pre- historically extinct species are to be found in natural or midden deposits (Table 4). This group included nine species endemic to the Chathams: four ducks (three of which are undescribed, NUMBER 89 101 FIGURE 13.?Lower mandibles (left) and skulls (right) of Mergus spp. (dorsal views). Top to bottom: Mergus, spe cies undescribed (MNZ S30049, PRM sample #113/91) from Te Ana a Moe Cave, Te Whanga Lagoon, Chatham Island; Mergus australis (BMNH 1904.8.4.3, male); Mergus serrator (MNZ 12707, female). Scale bar=30 mm. see Figures 11-14), two rails, a snipe, the undescribed species of Nestor (Figure 10), and the putative Chatham Island Sea-ea gle. Of the four other species rendered extinct on the Chathams in pre-European times, the swan and the crow {Palaeocorax moriorum) may have survived on the New Zealand mainland a little longer. The New Zealand Little Bittern {Ixobrychus no vaezelandiae) survived on the mainland until early in the twen tieth century. The New Zealand Scaup {Aythya novaeseelandi ae) is the only one of this group of 13 species exterminated on the Chathams that still survives on the mainland. European settlement brought further problems for the sur vival of the remaining Chathams avifauna; habitat destruction continued as forest and scrub were cleared for farming, two more species of rats, house mice, cats, and dogs were intro duced, and human hunting no doubt continued. The four en demic species that became extinct between 1840 and 1906 were all of small size, three were flightless (and thus obligate ground nesters), and even the bellbird was a weak flyer. There can be little doubt their demise was hastened by a combina tion of habitat destruction and predation. Adults would have been particularly vulnerable to cat predation, and their eggs and young vulnerable to rats. The Brown Teal {Anas chloro- TABLE 4.?Extinction of Chatham Islands terrestrial and freshwater birds. Of the 36 former breeding species, 13 were exterminated during the pre-European era, and eight were exterminated during the European era; total extinctions=21 (58%). There are now 15 breeding terrestrial species (excluding historic colo nists and introductions) on the Chathams Islands. Era Pre-European European Total extinction (endemics) Mergus, species undescribed Pachyanas chathamica Tadorna, species undescribed Anas, species undescribed Haliaeetus australis Diaphorapteryx hawkinsi Fulica chathamensis Coenocorypha chathamica Nestor, species undescribed Gallirallus dieffenbachii (1840) Gallirallus modestus (1900) Bowdleria rufescens (1900) Anthornis melanura melano- cephala (\906) Local extirpation Cygnus sumnerensis Aythya novaeseelandiae Ixobrychus novaezelandiae Palaeocorax moriorum Botaurus poiciloptilus (1910) Anas chlorotis (1915) Anas rhynchotis (1925) Falco novaeseelandiae (1900) tis) and New Zealand Shoveler {Anas rhynchotis), which were last seen in the Chathams in 1915 and 1925, respectively, would seem to have been the victims of recreational hunting. 102 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 14.?Sterna of Mergus spp. (right lateral views). Top to bottom: Mergus, species undescribed (MNZ S30049, PRM sample #113/91), from Te Ana a Moe Cave, Te Whanga Lagoon, Chatham Island; Mergus austra lis (BMNH 1904.8.4.3, male); Mergus serrator (MNZ 12707, female). (Scale bar=30 mm.) The Australasian Bittern {Botaurus poiciloptilus) may have suffered the same fate, although there is some doubt whether this species was ever really established in the Chathams. The New Zealand Falcon {Falco novaeseelandiae), from its sub- fossil record formerly abundant and reportedly seen as late as the 1890s, may have been vulnerable to rats and cats, particu larly when nesting. The elimination of these eight species during the historic pe riod (34% of the 23 species that had survived through the Polynesian period) meant that since first human contact just over 400 years before, at least 58% of the Chatham Islands' original complement of land birds and waterfowl had become extinct. As a more positive adjunct to this sad record, it should be noted that without enlightened human intervention several more species, notably the Black Robin {Petroica traversi), the New Zealand Shore Plover {Thinornis novaeseelandiae), and NUMBER 89 103 the Chatham Island Pied Oystercatcher {Haematopus chatham ensis), also might have become extinct. Conclusions Analysis of the abundant and well-dated fossil avian mate rial found in sand-dune and cave deposits on several of the larger islands of the Chatham group indicates that these is lands have supported a diverse, highly endemic avifauna since at least 7000 years ago. This avifauna, some 100 species in all, including many endemic land and freshwater birds, as well as a wide variety of seabirds, survived apparently un scathed until shortly after the first human arrival about 450 years ago. Many birds of the Chatham Islands exemplify the evolution ary trend toward larger body size and diminished flying ability so typically found in small, isolated oceanic-island groups that, prior to human colonization, lacked mammalian predators. Most of the Chatham Islands land birds and waterfowl are larg er than their mainland counterparts, and of the 36 prehistorical- ly known species, at least seven were flightless and three more would have been weak fliers. The land birds of the Chatham Islands were clearly no ex ception to the general rule that insular species tend to be "na ive" toward humans and introduced predators (Milberg and Tyrberg, 1993:229). The lethal combination of weak flight and trusting attitude predisposed them to an extraordinary vulnera bility to human interference. The fossil record of the last 7000 years gives no indication that any of the prehistorically known species became extinct, or even less abundant, prior to human arrival. All of the flightless and weak-flying species, and a fur ther 11 flying species, however, became extinct within a few hundred years of first human settlement through the combined effects of human perturbations. Appendix This appendix provides an annotated listing of both conven tional and calibrated radiocarbon ages for samples from Chatham and Pitt islands. Samples (bone collagen or marine- shell carbonate) are identified by their Rafter Radiocarbon Laboratory reference numbers (prefixed by NZA). Age data are presented as follows: conventional age based on the old (Lib- by) half-life of 5568 yrs (as "age ? standard deviation yrs BP"); calibrated (corrected) age, given as median age ("CAL BP") where possible; and calibrated (corrected) age as a range ex pressed in terms of the 95% confidence interval (age ? two standard deviations). Locality names and grid references (GR) given for sampled sites (see Figure 1, numbered sites 1-21) are from New Zealand Topographical Map, NZMS 260, 1:50000 metric series, Chatham Islands, edition 1, 1981 (Chatham Is land (localities 1-17), sheet 1; Pitt Island (localities 18-21), sheet 2 ). (a.s.l.=above sea level.) CHATHAM ISLAND Locality 1, Waihora, Point Durham NZA 3193: Gallirallus dieffenbachii, Sutton coll. WH/VII/2 Layer 1; GR 358470; Waihora Mound site, supposedly mid den material; 5030 ? 68 yrs BP; ca. 5750 CAL BP; 5894-5605 CAL BP. NZA 3194: Gallirallus dieffenbachii, Sutton coll. WH/VII/ 23 Layer 3; GR 358470; Waihora Mound site, supposedly midden material; 5237 ? 72 yrs BP; ca. 5950 CAL BP; 6173-5753 CAL BP. Locality 2, Red Bluff NZA 2610: Hemiphaga chathamica, MNZ S31065; GR 469603; back-beach sequence in embayment, -800 m SSE of Te Whenuhau Trig, from uppermost stratum of coarse, yellow, shell sand (beneath windblown fine sand). Indicates minimum age for >2 m thick sequence; 985 ? 80 yrs BP; 843 CAL BP; 975-689 CAL BP. Locality 3, Long Beach, S of Henga limestone bluffs (GR 450655) NZA 1930: Fulica chathamensis, MNZ S27821; GR 465626; Long Beach, in typical, stratified, consolidated, dune hum mock (-4 m a.s.l.), immediately beneath eroded surface of brown sand/soil horizon (30 cm thick), now overlain by drift sand; 2560 ? 145 yrs BP; 2254 CAL BP; 2666-1924 CAL BP. NZA 1931: Hemiphaga chathamensis, MNZ S27822; GR 465626; Long Beach, in pale brown sand at 1.5 m depth (di rectly underlying sample NZA 1930); 3790 ? 150 yrs BP; 4109 CAL BP; 4511-3697 CAL BP. NZA 1929: Diaphorapteryx hawkinsi, MNZ S27820; GR 456646; Long Beach, Milton's Gully site, embedded in gray brown, hard-pan deflation surface (former interdune lake deposit?); 6660 ? 150 yrs BP; 7456 CAL BP; 7693-7207 CAL BP. NZA 3246: Hemiphaga chathamensis, MNZ S33158; GR 456646; Long Beach, Milton's Gully site, in horizontally bedded brown sand at inland margin of and stratigraphically 104 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY above hard-pan deflation surface; 5355 ? 70 yrs BP; 6091 CAL BP; 6271-5935 CAL BP. NZA 3239: Gallirallus dieffenbachii, MNZ S33146; GR 456646; Long Beach, Milton's Gully site, in consolidated brown sand at seaward margin of and stratigraphically above hard-pan deflation surface; 5779 ?71 yrs BP; 6543 CAL BP; 6721-6405 CAL BP. NZA 2632: Hemiphaga chathamensis, MNZ S32545; GR 454653; Long Beach, -500 m N of Milton's Gully site. Or ange brown bones in iron-stained sand. Site of former ephemeral dune lake? 4720 ? 87 yrs BP; ca. 5450 CAL BP; 5594-5057 CAL BP. Locality 4, Long Beach, N of Henga limestone bluffs (GR 450655) NZA 797: Pterodroma magentae, NMZ S26679; GR 447663; Long Beach, -350 m N of Henga bluffs. From up permost 20 cm of typical brown sand/soil exposed in erod ing hillocks -5 m a.s.l.; 3190 ? 130 yrs BP; 3025 CAL BP; 3321-2749 CAL BP. NZA 1934: Hemiphaga chathamensis, MNZ S27825; GR 447663; Long Beach, -350 m N of Henga bluffs. In situ in typical brown sand/soil exposed in eroding, seaward-ex tending ridge -5 m a.s.l.; 2200 ? 150 yrs BP; 2139 CAL BP; 2702-1739 CAL BP. NZA 2602: Hemiphaga chathamensis, MNZ S30584; GR 445669; Long Beach, in situ at -5 m a.s.l. in extensive, con solidated brown-sand horizon exposed on large, sea ward-sloping deflation surface; 4165 ? 92 yrs BP; ca. 4700 CAL BP; 4835-4418 CAL BP. NZA 3245: Hemiphaga chathamensis, MNZ S33047; GR 445669; Long Beach, in situ at -7 m a.s.l. in extensive, con solidated, gray brown sand horizon exposed on large, sea ward-sloping deflation surface; 4493 ? 66 yrs BP; ca. 5050 CAL BP; 5282^1872 CAL BP. NZA 3238: Fulica chathamensis, MNZ S33035; GR 445669; Long Beach, in situ at -3 m a.s.l. in consolidated brown- sand horizon exposed on lower slopes of large, sea ward-sloping deflation surface; 6879 ? 68 yrs BP; 7626 CAL BP; 7759-7529 CAL BP. NZA 1935: Fulica chathamensis, MNZ S27819; GR 443675; Long Beach, associated skeleton on extensive deflation sur face, -2 m a.s.l. (former interdune lake?); 4990 ? 150 yrs BP; ca. 5737 CAL BP; 5988-5320 CAL BP. NZA 794: Pterodroma magentae, NMZ S25362; GR 436683; Long Beach, on 30 m high dune ridge, -200 m in land of shoreline. Site of former nesting colony? 3150 ? 150 yrs BP; 2982 CAL BP; 3351-2685 CAL BP. Locality 5, Lake Marakapia NZA 1932: Pteodroma magentae, MNZ S27823; GR 447677; Lake Marakapia, in situ (complete individual skele ton) in brown surface sand/soil of hilltop deflation area in farm pasture, -15 m a.s.l. Site of former nesting colony? 2390 ? 140 yrs BP; 2047 CAL BP; 2351-1702 CAL BP. NZA 1933: Pteodroma magentae; MNZ S27824: GR 447677; Lake Marakapia, in situ (complete individual skele ton) in yellow sand beneath brown surface sand/soil of hill top deflation area in farm pasture. Site of former nesting colony?; 3310 ? 140 yrs BP; 3163 CAL BP; 3472-2814 CAL BP. Locality 6, Tennant's Lake NZA 795: Pterodroma magentae, MNZ S27008; GR 436693; Tennant's Lake, in situ (complete individual skele ton), -200 m inland of shoreline, -15 m a.s.l., on mo saic-cracked pan. Site of former nesting colony? 3420 ?210 yrs BP; 3299 CAL BP; 3790-2798 CAL BP. Locality 7, Ohira Bay NZA 3430: Hemiphaga chathamensis, MNZ S (uncata loged); GR 384717; Ohira Bay, on eastern slope in brown sand, -6 m a.s.l.; 4608 ? 66 yrs BP; ca. 5190 CAL BP; 5452^989 CAL BP. Locality 8, Waitangi West NZA 2611: Pterodroma magentae, MNZ S31189; GR 252752; Waitangi West, in older consolidated dune series, -50 m inland of shoreline, from dark brown sand/soil hori zon, in places overlain by midden debris, -3 m a.s.l.; 1114 ? 82 yrs BP; 690 CAL BP; 856-547 CAL BP. NZA 3425: Pachyanas chathamica, MNZ S32638; GR 252752; Waitangi West, associated skeleton from low-lying deflation area, -2 m a.s.l., between foredunes and inland consolidated dune series; 1913 ? 62 yrs BP; 1792 CAL BP; 1935-1625 CAL BP. NZA 3285: Gallirallus dieffenbachii, MNZ S32639; GR 247764; Waitangi West, in inland dune series -3-4 m a.s.l., in brown sand/soil immediately beneath occupation-midden shell stratum; 887 ? 59 yrs BP; ca. 730 CAL BP; 906-667 CAL BP. NZA 3426: Cygnus sumnerensis, MNZ S (uncataloged); GR 247764; Waitangi West, in inland dune series -3-4 m a.s.l., in pale brown sand/soil 30 cm beneath sample NZA 3285; 2306 ? 63 yrs BP; ca. 2250 CAL BP; 2358-2077 CAL BP. NZA 2612: Gallirallus dieffenbachii, MNZ S31299; GR 250765; Waitangi West, from fine-grained, consolidated, pale brown sand (weathering to gray white), at -3-4 m a.s.l. on N bank of Waihi Creek, -200 m from outlet; 3625 ? 84 yrs BP; ca. 3880 CAL BP; 4088-3639 CAL BP. NUMBER 89 105 Locality 9, Western Maunganui Dunes (Cape Pattison E-Maunganui) NZA 1950: Paphies subtriangulatum, MNZ S27830; GR 277761; Maunganui, Midden Site, 50 m E of Moravian Mis sion Stone Cottage. Marine midden shell from uppermost 10 cm of chocolate brown soil horizon (-35 cm thick) overly ing natural dune sequence; 760 ? 140 yrs BP; 395 CAL BP; 634-121 CAL BP. NZA 1949: Gallirallus dieffenbachii, MNZ S27831; GR 277761; Maunganui, Midden Site, 50 m E of Moravian Mis sion Stone Cottage. From brown, humic-stained dune sand, -40 cm below NZA 1950; 1340 ? 150 yrs BP; ca. 1200 CAL BP; 1501-931 CAL BP. NZA 1981: Gallirallus dieffenbachii, MNZ S27832; GR 277761; Maunganui, Midden Site, 50 m E of Moravian Mis sion Stone Cottage. From pale brown/gray, consolidated dune sand, -70 cm below NZA 1950; 1830 ? 150 yrs BP; 1702 CAL BP; 2008-1353 CAL BP. Locality 10, Mid-Maunganui Dunes (Maunganui E-Washout Creek) NZA 1947: Diaphorapteryx hawkinsi, MNZ S27828; GR 286760; Maunganui, in situ in pale brown, semiconsoli- dated foredune sand, 150 m E of Maunganui Bluff, -3 m a.s.l.; 1860 ? 150 yrs BP; 1734 CAL BP; 2102-1400 CAL BP. NZA 3608: Gallirallus dieffenbachii, MNZ S32892; GR 288760; Maunganui, -4 m a.s.l., in back-beach dune face, from upper level of -2 m thick, black brown (humic- stained) dune sand, here overlain by 50 cm thick layer of oc cupation-midden shell; 677 ? 60 yrs BP; ca. 600 CAL BP; 667-535 CAL BP. NZA 2585: Gallirallus dieffenbachii, MNZ S29026; GR 289758; Maunganui, in brown sand of inland dune series, -6 m a.sl, -300 m W of distinctive Basalt Knob; 4282 ? 89 yrs BP; ca. 4750 CAL BP; 5024-4523 CAL BP. NZA 3189: Hemiphaga chathamensis, MNZ S32846; GR 289758; Maunganui, in brown sand of inland dune series, -6 m a.s.l., -300 m W of Basalt Knob. From fine, pale brown sand that seen laterally is overlain by thin (-10 cm) orange-colored sand layer (iron-stained) then 50 cm thick stratum of gray/black sandy soil. This sand/soil sequence here overlain by compact, 20 cm thick midden-shell layer, topped by recent drift sand; 4113 ? 67 yrs BP; ca. 4600 CAL BP; 4820^1411 CAL BP. NZA 3191: Diaphor apteryx hawkinsi, MNZ S32834; GR 292758; Maunganui, immediately E of Basalt Knob, within inland dune series, from brown sand near base of N wall of -8 m deep, nearly circular, steep-sided deflation hollow; 3857 ? 65 yrs BP; ca. 4200 CAL BP; 4406-3991 CAL BP. NZA 3190: Tadorna, species undescribed, MNZ S32830; GR 292758; Maunganui, immediately E of Basalt Knob, from pale brown sand layer, 2 m below (stratigraphically) brown sand from which NZA 3191 was obtained; 5291 ? 66 yrs BP; ca. 6030 CAL BP; 6187-5902 CAL BP. Locality 11, Eastern Maunganui Dunes (Washout Creek E-Tahatika Creek) NZA 1982: Hemiphaga chathamensis, MNZ S27832; GR 307758; Maunganui, in inland dune series, -1.2 km E of Washout Creek (50 m W of Big Midden Site). From brown sand on south-facing deflation surface, -5-6 m a.s.l. (strati graphic equivalent of NZA 3189, NZA 3287); 3760 ? 160 yrs BP; 4067 CAL BP; 4510-3635 CAL BP. NZA 2614: Gallirallus dieffenbachii, MNZ S32031; GR 308758; Maunganui, in inland dune series, -1.25 km E of Washout Creek (Big Midden Site). From dark-stained sand -30 cm beneath compact, 20 cm thick midden-shell hori zon; 1390 ? 80 yrs BP; 1247 CAL BP; 1386-1069 CAL BP. NZA 3287: Diaphorapteryx hawkinsi, MNZ S (uncataloged); GR 308758; Maunganui, in inland dune series, -1.25 km E of Washout Creek (Big Midden Site). In situ in brown sand on south-sloping deflation surface, -5-6 m a.s.l. (strati graphic equivalent of NZA 1982, NZA 3189); 3966 ? 60 yrs BP; 4350 CAL BP; 4517^152 CAL BP. NZA 2609: Hemiphaga chathamensis, MNZ S30778; GR 347771; Maunganui, on ridge crest of inland dune series, -500 m W of Tahatika Creek. In situ skeleton from pale brown sand at -1.2 m, beneath -10 cm thick, orange, iron-stained zone and brown black sandy soil, 30 cm thick, with rounded lag pebbles on surface; 3264 ? 84 yrs BP; ca. 3450 CAL BP; 3631-3218 CAL BP. NZA 3286: Pachyanas chathamica, MNZ S32634; GR 349773; Tahatika, in situ in pale, gray brown sand, on low foredune (-2 m a.s.l.), -250 m W of Tahatika Creek; 1529 ? 57 yrs BP; 1373 CAL BP; 1502-1293 CAL BP. NZA 3284: Fulica chathamensis, MNZ S (uncataloged); GR 351772; Tahatika, from gullied, older red brown (iron-stained) sand, -4 m a.s.l., on seaward (N) slopes of in land dune series, -150 m W of Tahatika Creek; 3296 ? 59 yrs BP; ca. 3450 CAL BP; 3621-3361 CAL BP. NZA 796: Cygnus sumnerensis, MNZ S26482; GR 354773; Tahatika, from seaward foredune slope, -2 m a.s.l., 250 m E of Tahatika Creek (stratigraphic equivalent of NZA 3286); 1490 ? 130 yrs BP; 1351 CAL BP; 1606-1068 CAL BP. NZA 1937: Cygnus sumnerensis, MNZ S27827; GR 354773; Tahatika, from brown sand forming nearly level deflation surface inland of foredune slope (cf. NZA 796), -3 m a.s.l., 250 m E of Tahatika Creek; 1420 ? 140 yrs BP; 1276 CAL BP; 1533-985 CAL BP. NZA 2603: Cygnus sumnerensis, MNZ S30709; GR 354773; Tahatika, from brown sand ridge upon main defla tion surface (cf. NZA 1937), -3 m a.s.l., -300 m E of Taha tika Creek ; 792 ? 77 yrs BP; ca. 700 CAL BP; 891-549 CAL BP. 106 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Locality 12, Taupeka Inland Dunes NZA 2587: Pelagodroma marina, MNZ S30314; GR 517783; Taupeka inland dune series, -2.5 km E of Taupeka. Excavated from face of large blowout hollow in consoli dated, yellow brown sand -3 m below present ground sur face. As elsewhere along Taupeka inland dunes, fossilifer- ous sand lies beneath well-developed dune-soil sequence 1 m or so thick (see Figure 5); 6632 ? 98 yrs BP; 7127 CAL BP; 7330-6926 CAL BP. NZA 3427: Pterodroma cf. inexpectata, MNZ S (uncata loged); GR 531778; Taupeka inland dune series, -3 km E of Taupeka. From consolidated, pale brown sand -2-3 m below present ground surface (cf. NZA 2587); 4935 ? 73 yrs BP; 5289 CAL BP; 5462-5048 CAL BP. Locality 13, Lake Pateriki NZA 3429: Fulica chathamensis, MNZ S (uncataloged); GR 660779; from pale, creamy white sand near base (-2-3 m a.s.l.) of sloping dune face at seaward end of E shore of Lake Pateriki; 2278 ? 70 yrs BP; ca. 2250 CAL BP; 2350-2053 CAL BP. Locality 14, Kaingaroa NZA 1983: Fulica chathamensis, MNZ S27834; GR 683787; SW of Kaingaroa on high dune ridge (-30 m a.s.l.). From depth of 50 cm in black peaty sand/soil forming sea ward-sloping, lag-pebble strewn deflation surface (cf. NZA 1988, taken from 50 cm below); 2620 ? 160 yrs BP; 2623 CAL BP; 3062-2311 CAL BP. NZA 1988: Paleocorax moriorum, MNZ S27835; GR 683787; SW of Kaingaroa on high dune ridge (-30 m a.s.l.). From depth of 1 m in brown-sand horizon (2 m+ thick), be neath 60-cm-thick black, peaty sand/soil that forms sea ward-sloping, lag-pebble strewn deflation surface (cf. NZA 1983, taken from 50 cm above); 3410 ? 150 yrs; 3285 CAL BP; 3635-2898 CAL BP. NZA 2588: Cygnus sumnerensis, MNZ S30421; GR 684796; W of Kaingaroa, at -4 m a.s.l. in steep, beach-cut face of foredune. From yellow, unconsolidated sand, -1 m below more compacted brown sand/soil horizon (buried occupa tion soil?), with midden shell debris upon it. This sequence exposed for >500 m southward along 2-10 m high, eroding, back-beach face; 1325 ? 84 yrs BP; ca. 1210 CAL BP; 1318-992 CAL BP. Locality 15, Okawa Point NZA 3428: Hemiphaga chathamensis, MNZ S (uncata loged); GR 706763; from extensive shallow deflation area (3-4 m a.s.l.) at S end of inland dune series, -1.4 km N of Okawa Point. In pale brown sand beneath orange-colored incipient iron-pan, chocolate brown sand/soil, leached gray sand and, finally, loose yellow-white surface sand; 3938 ? 68 yrs BP; 4312 CAL BP; 4512^1093 CAL BP. Locality 16, Te Ana a Moe Cave, W Shore of Te Whanga Lagoon NZA 1948: Gallirallus dieffenbachii, MNZ S27829; GR 480652; Te Ana a Moe Cave, in SE quadrant, from -60 cm depth immediately below surface of undisturbed sediments (Figure 6); 1250 ? 145 yrs BP; 1149 CAL BP; 1414-795 CAL BP. NZA 798: Gallirallus modestus, MNZ S23708; GR 480652; Te Ana a Moe Cave, in SE quadrant, from -65 cm depth in brownish cream bryozoan detrital sand, 5 cm below surface of undisturbed sediments (Figure 6); 1270 ? 120 yrs BP; ca. 1163 CAL BP; 1606-1068 CAL BP. NZA 2778: Tadorna, species undescribed, MNZ S (uncata loged); 480652; Te Ana a Moe Cave, in NE quadrant, in brownish cream bryozoan detrital sand, -15 cm below sur face of undisturbed sediments (Figure 6); 1534 ? 62 yrs BP; ca. 1410 CAL BP; 1495-1338 CAL BP. NZA 801: Gallirallus modestus, NMZ S27409; GR 480652; Te Ana a Moe Cave, in SW quadrant, in fine, pale cream, bryozoan detrital sand at -90 cm depth (-30 cm below sur face of undisturbed sediments; see Figure 6); 2290 ? 140 yrs BP; 2242 CAL BP; 2364-2003 CAL BP. NZA 800: Gallirallus modestus, NMZ S27501; GR 480652; Te Ana a Moe Cave, in SW quadrant, in coarser, white, bry ozoan detrital sand at -1 m depth (-40 cm below surface of the undisturbed sediments; see Figure 6); 2950 ? 140 yrs BP; 3057 CAL BP; 3357-2768 CAL BP. NZA 1989: Pterodroma magentae, MNZ S27836; GR 480652; Te Ana a Moe Cave, in SW quadrant at 1.25 m depth, in creamy brown sand, overlying rounded limestone cobbles (Figure 6), near end of short (-2 m long), blind-end ing side tunnel; 3900 ? 150 yrs BP; 3904 CAL BP; 4308-3524 CAL BP. Locality 17, Thomas Property, opposite Limestone Quarry, 4.5 kmNofTeOne NZA 2589: Hemiphaga chathamensis, MNZ S30516; GR 469630; Lower Limestone Cave, excavated from sediments beneath overhang at base of exposed E face of outcrop; 2239 ? 87 yrs BP; 2188 CAL BP; 2348-1982 CAL BP. NZA 2590: Hemiphaga chathamensis, MNZ S30516; GR 469630; Upper Limestone Cave, excavated from sediments beneath overhang 3-5 m above and 10 m W of NZA 2589; 2434 ? 88 yrs BP; ca. 2550 CAL BP; 2734-2187 CAL BP. NZA 2777: Pterodroma magentae, MNZ S32593; GR 467630; -400 m inland of S Long Beach, from dune sands mantling W portions of extensive outcrop of karstic Te Whanga Limestone. In situ skeleton from brown, sur face-consolidated sand/soil overlying coarser, more friable NUMBER 89 107 yellowish sand (site of former nesting colony?); 2898 ? 64 yrs BP; 2691 CAL BP; 2814-2482 CAL BP. slope of Motutapu Point promontory; 4419 ? 92 yrs BP; ca. 5000 CAL BP; 5292-4726 CAL BP. PITT ISLAND Locality 18, Tarawhenua Peninsula NZA 2613: Gallirallus dieffenbachii, MNZ S31558; GR 690225; Tarawhenua Peninsula, from brown sands on W slope of narrow "neck" connecting peninsula to mainland; 2994 ? 83 yrs BP; 3104 CAL BP; 3336-2897 CAL BP. NZA 1549: Diomedea epomophora; MNZ S27811; GR 679233; Tarawhenua Peninsula. Specimen from Lindsay collection (see Falla, 1960); 4440 ? 150 yrs BP; 4625 CAL BP; 4998^224 CAL BP. NZA 1906: Diomedea epomophora, MNZ S27817; GR 679233; from cliff-top exposure, at W extremity of Tarawhenua Peninsula. Bones of fledgling (indicative of former presence of nesting colony) from orange brown sandy soil (upon karstic Te Whanga Limestone), which un derlies present pasture-grass surface; essentially same site as NZA 1549; 4300 ? 150 yrs BP; 4545 CAL BP; 4824-4059 CAL BP. Locality 19, Motutapu Point NZA 1907: Diomedea epomophora, MNZ S27818; GR 721248; from consolidated brown sand forming 50 cm thick surface layer, on steep S slope of Motutapu Point promon tory, -6-8 m a.s.l. Bones of fledgling (indicative of former presence of nesting colony); 1070 ? 150 yrs BP; 664 CAL BP; 935^136 CAL BP. NZA 2586: Fulica chathamensis, MNZ S30949; GR 721248; from coarse, yellow sand, beneath 50 cm thick consolidated, brown-sand surface layer, -15-20 m a.s.l., high on steep S Locality 20, Tupuangi NZA 2631: Hemiphaga chathamensis, MNZ S31464; GR 741234; older dunes forming S bank of Tupuangi Creek es tuary. From charcoal-blackened, greasy soil among oven stones at Moriori camp site. Indicates minimum date of first settlement of Pitt Island; 491 ? 80 yrs BP; ca. 450 CAL BP; 631-305 CAL BP. NZA 3431: Hemiphaga chathamensis, MNZ S (uncataloged); GR 742235; older dune ridge, 200 m S of Tupuangi Creek estuary. From blowout hollow, in distinctive consolidated brown sand stratum, 1 m below present drift-sand surface; 1235 ? 60 yrs BP; ca. 1100 CAL BP; 1253-974 CAL BP. Locality 21, Near Kokope Island NZA 3461: Hemiphaga chathamensis, MNZ S (uncata loged); GR 756218; foredunes of Tupuangi dune series, just N of Kokope Island. Specimen in situ in brown-sand stra tum stratigraphically below paler gray-sand horizon from which human (Moriori) skeletal remains were eroding. Indi cates maximum age for human burial; 876 ? 62 yrs BP; 749 CAL BP; 906-663 CAL BP. Locality Uncertain, Chatham Islands? NZA 1548: Haliaeetus australis; BMNH A3732; "Chatham Islands," of uncertain provenance, from H.O. Forbes collec tion in The Natural History Museum, London; 1025 ? 51 yrs BP. Southern Hemisphere marine calibration, median 615 CAL BP; 679-533 CAL BP. Northern Hemisphere marine calibration, median 258 CAL BP; 406-114 CAL BP. Literature Cited Andrews, CE. 1896a. Note on the Skeleton of Diaphorapteryx hawkinsi, Forbes, a Large Extinct Rail from the Chatham Islands. Geological Magazine, 43:337-338. 1896b. On the Extinct Birds of the Chatham Islands, Part 1: The Osteology of Diaphorapteryx hawkinsi. Novitates Zoologicae, 3(1): 73-84. 1896c. On the Extinct Birds of the Chatham Islands, Part 2: The Osteology of Palaeolimnas chathamensis and Nesolimnas dieffenbachii. Novi tates Zoologicae, 3(3):260-271. Atkinson, I.A.E., and PR. Millener 1991. An Ornithological Glimpse into New Zealand's Pre-human Past. Acta XX Congressus International is Ornithologici: 129-192. Well ington, New Zealand: New Zealand Ornithological Congress Trust Board. Balouet, J.C, and S.L. Olson 1989. Fossil Birds from Late Quaternary Deposits in New Caledonia. Smithsonian Contributions to Zoology, 469: 38 pages. Bourne, W. 1967. Subfossil Petrel Bones from the Chatham Islands. Ibis, 109(1): 1-7. Campbell, H.J. 1996. Geology. In The Chatham Islands?Heritage and Conservation, pages 34?48. Christchurch: Canterbury University Press. Cooper, R.A., and PR. Millener 1993. The New Zealand Biota: Historical Background and New Research. Trends in Ecology and Evolution, 8:429?433. Courts, P. 1969. Analysis of Archaeological Material from Eroding Midden Sites. Journal of the Royal Society of New Zealand, 2:407?412. Dawson, E.W. 1957. Falcon in the Chatham Islands. Notornis, 7:113. 1958. Re-discoveries of the New Zealand Subfossil Birds Named by H.O. Forbes. Ibis, 100:232-237. 108 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 1959. The Supposed Occurrence of Kakapo, Kaka and Kea in the Chatham Islands. Notornis, 8(4):106-115. 1960. New Evidence of the Former Presence of Kakapo (Strigops habrop- tilus) in the Chatham Islands. Notornis, 9(3):65-67. 1961a. An Extinct Sea-eagle in the Chatham Islands. Notornis. 9(5): 171. 1961b. The Former Existence of a Species of Falco in the Chatham Islands, New Zealand: Some New Evidence. Emu, 61:209-210. Dieffenbach, E. 1841. An Account of the Chatham Islands. Journal of the Royal Geo graphical Society, 11:195. Falla, R.A. 1960. Notes on Some Bones Collected by Dr. Watters and Mr. Lindsay at Chatham Islands. Notornis, 8(7):226-227. Fleming, CA. 1962. History of the New Zealand Land Bird Fauna. Notornis, 9: 270-274. 1975. The Geological History of New Zealand and Its Biota. In A. Kus- chel, editor, Biogeography and Ecology in New Zealand, pages 1-86. The Hague: W. Junk. Forbes, H.O. 1892a. Aphanapteryx in the New Zealand Region. Nature, 45(1173): 580-581. 1892b. Preliminary Notice of Additions to the Extinct Avifauna of New Zealand. Transactions of the New Zealand Institute, 24:185-189. 1892c. Aphanapteryx and Other Remains in the Chatham Islands. Nature, 46(1185):252-253. 1893a. The Chatham Islands and Their Story. The Fortnightly Review, 53(317):669-690. 1893b. A List of the Birds Inhabiting the Chatham Islands. Ibis, 6(5): 521-546. Harrison, C.J.O., and CA. Walker 1973. An Undescribed Extinct Fish-Eagle from the Chatham Islands. Ibis, 115:274-277. Hay, R.F., A.R. Mutch, and WA. Watters 1970. Geology of the Chatham Islands. New Zealand Geological Survey Bulletin, new series, 83:1-86. Hutton, F.W. 1904. Index Faunae Novae Zealandiae. viii + 372 pages. London: Dulau and Co. Published for the Philosophical Institute of Canterbury. James, H.F, and S.L. Olson 1991. Description of Thirty-two New Species of Birds from the Hawaiian Islands, Part II: Passeriformes. Ornithological Monographs, 46: 88 pages. Kear, J., and R.J. Scarlett 1970. The Auckland Islands Merganser. Wildfowl, 21:78-86. Marshall, Y.M., R.J. Scarlett, and D.G. Sutton 1987. Bird Species Present on the Southwest Coast of Chatham Island in the Sixteenth Century A.D. Working Papers in Anthropology, Ar chaeology, Linguistics, Maori Studies, University of Auckland, De partment of Anthropology, 76:1-25. McFadgen, B.G. 1994. Archaeology and Holocene Dune Stratigraphy on Chatham Island. Journal of the Royal Society of New Zealand, 24( 1): 17?44. McNab, B.K. 1994. Energy Conservation and the Evolution of Flightlessness in Birds. The American Naturalist, 144(4):628-642. Meredith, C. 1991. Vertebrate Fossil Faunas from Islands in Australasia and the South west Pacific. In P. Vickers-Rich, J.M. Monaghan, R.F. Baird, and T.H. Rich, editors, Vertebrate Palaeontology of Australasia, pages 1345-1382. Melbourne: Pioneer Design Studio. Milberg, P., and T. Tyrberg 1993. Naive Birds and Noble Savages?A Review of Man-Caused Prehis toric Extinctions of Island Birds. Ecography, 16:229-250. Millener, PR. 1980. The Taxonomic Status of Extinct New Zealand Coots, Fulica chathamensis subspp. (Aves: Rallidae). Notornis, 27(4):363-367. 1981. The Subfossil Distribution of Extinct New Zealand Coots, Fulica chathamensis subspp. (Aves: Rallidae). Notornis, 28(1): 1-9. 1991. The Quaternary Avifauna of New Zealand. In P. Vickers-Rich, J.M. Monaghan, R.F. Baird, and T.H. Rich, editors, Vertebrate Palaeon tology of Australasia, pages 1317-1344. Melbourne: Pioneer De sign Studio. 1996. Extinct Birds. In The Chatham Islands?Heritage and Conserva tion, pages 113-120. Christchurch: Canterbury University Press. Oliver, W.R.B. 1955. New Zealand Birds. Second edition, 661 pages. Wellington: Reed. Olson, S.L. 1973. A Classification of the Rallidae. Wilson Bulletin, 85(4):381-416. 1975. A Review of the Extinct Rails of the New Zealand Region (Aves: Rallidae). National Museum of New Zealand Records, 1(3):63?79. 1984. The Relationships of the Extinct Chatham Island Eagle. Notornis, 31(4):273-277. 1990. Osteology and Systematics of the Fernbirds (Bowdleria: Sylviidae). Notornis, 37(3): 161-171. Olson, S.L., and H.F. James 1982. Prodromus of the Fossil Avifauna of the Hawaiian Islands. Smithso nian Contributions to Zoology, 365: 59 pages. 1991. Descriptions of Thirty-two New Species of Birds from the Hawaiian Islands, Part I: Non-Passeriformes. Ornithological Monographs, 45: 88 pages. Scarlett, R.J. 1955. Falcon in Chatham Islands. Notornis, 6(3):57. Simmons, D.R. 1964. Chatham Island Archaeological Survey. New Zealand Archaeologi cal Association Newsletter, 5(l):238-244. Steadman, D.W. 1995. Prehistoric Extinctions of Pacific Island Birds: Biodiversity Meets Zooarchaeology. Science, 267:1123-1131. Stuiver, M., and T.F. Braziunas 1993. Modeling Atmospheric 14C Influences and 14C Ages of Marine Sam ples to 10,000 BC. In M. Stuiver, M.A. Long, and R.S. Kra, editors, Calibration 1993. Radiocarbon, 35(1): 137-189. Stuiver, M., and G.W. Pearson 1993. High-Precision Bidecadal Calibration of the Radiocarbon Time Scale, AD 1950-500 BC and 2500-6000 BC. In M. Stuiver, A. Long, and R.S. Kra, editors, Calibration 1993. Radiocarbon, 35 (l):l-23. Stuiver, M., and H.A. Polach 1977. Discussion: Reporting of l4C Data. Radiocarbon, 19(3):353-363. Sutton, D.G. 1976. Radiocarbon Dates from the Waihora Mound Site. New Zealand Ar chaeological Association Newsletter, 19:195-196. 1979. Inland and Coastal Fowling Strategies of the Prehistoric Moriori. In A. Anderson, editor, Birds of a Feather. BAR International Series, 62:123-129. New Zealand Archaeological Association Monograph 11. 1981. Polynesian Coastal Hunters in the Chatham Islands. Doctoral disser tation, Anthropology Department, University of Otago, Dunedin, New Zealand. 1982. The Chatham Islands. In N.J. Prickett, editor, The First Thousand Years: Regional Perspectives in New Zealand Archaeology, pages 160-178. New Zealand Archaeological Association Monograph 13. Sutton, D.G., and Y.M. Marshall 1977. Archaeological Bird Bone Assemblages from Chatham Island: An Interpretation. In B.F. Leach, editor, Working Papers in Chatham Islands Archaeology, 12:1-15. Dunedin: Anthropology Department, University of Otago. Tennyson, A.J.D., and PR. Millener 1994. Bird Extinctions and Fossil Bones from Mangere Island, Chatham Islands. In Richard N. Holdaway, editor, Chatham Islands Omithol- NUMBER 89 109 ogy. Notornis, supplement, 41:165-178. Travers, H.H. 1866. On the Destruction of the Aborigines of Chatham Island. Transac tions of the Ethnological Society of London, new series, 4: 352-360. Travers, W.T.L. 1873. On the Birds of the Chatham Islands, by H.H. Travers, with Intro ductory Remarks on the Avi-fauna and Flora of the Islands in Their Relation to Those of New Zealand. Transactions and Proceedings of the New Zealand Institute, 5:212-222. 1883. Remarks on the Distibution within the New Zealand Zoological Sub-region of the Birds of the Accipitres, Passeres, Scansores, Gall- inae, Struthiones, and Grallae. Transactions and Proceedings of the New Zealand Institute, 15:178-187. Trotter, M.M., and B. McCulloch 1984. Moas, Men and Middens. In P.S. Martin and R.G. Klein, editors, Quaternary Extinctions: A Prehistoric Revolution, pages 708-727. Tucson: University of Arizona Press. Turbott, E.G., convener 1990. Checklist of the Birds of New Zealand and the Ross Dependency, Antarctica. Third edition, 247 pages. Auckland: Random Century in association with the Ornithological Society of New Zealand, Inc. van Bemmelen, J.F. 1993. The Extinct Birds of the Chatham Islands. Occasional Papers of the Hutton Foundation, New Zealand, 3: v+73 pages, van Tets, G.F., P.V. Rich, P.J. Fullagar, and P.M. Davidson 1981. Fossil, Sub-fossil and Early Historic Birds of Lord Howe and Nor folk Islands. In H.F. Rechter and W.F. Ponder, editors, Lord Howe Island, pages 29-31. Sydney: Australian Museum. Wallace, A.R. 1893. Mr. H.O. Forbes's Discoveries in the Chatham Islands. Nature, 48(1228):27-28. Wright, A.C.S. 1959. Soils of Chatham Island (Rekohu). New Zealand Soil Bureau Bulle tin, 19:1-61. The Role of Climate Change Versus Human Impacts?Avian Extinction on South Island, New Zealand Trevor H. Worthy ABSTRACT The Late Quaternary avifaunas of South Island, New Zealand, reveal discrete faunal assemblages for the contrasting environ ments offered by wet, closed forest and open, grassland-shru- bland-forest mosaics. These faunal associations are recognizable in deposits dating from the last glacial (Otiran) and the Holocene periods. Sites in western regions of South Island exhibit significant differences in the fauna's species composition between deposits formed in the last glacial period and those from the present inter glacial period, but sites in the east do not. Several species became regionally extinct at the end of the glacial period, but all survived in the east until the present millennium. Although climate change caused the redistribution of species, all Late Quaternary extinc tions in New Zealand were ultimately caused by humans during the last 1000 years. Introduction New Zealand has three main islands and numerous smaller ones, and it occupies the southernmost comer of Polynesia in the South Pacific Ocean. It is of continental origin but has been separated from other land masses for the last 80 million years and is now 1500 km from Australia. Its long isolation has re sulted in a unique avifauna with a high degree of endemism and many flightless species (Fleming, 1979; Millener, 1990; Bell, 1991). Fossil deposits have been known from New Zealand since the early nineteenth century; they are rich in material and are widely distributed (Atkinson and Millener, 1991; Worthy and Holdaway, 1993). Most early work sought to describe the unique elements of the fossil fauna, notably the various species of moa (Aves: Dinornithiformes; see references in Anderson, Trevor H. Worthy, Palaeofaunal Surveys, 43 The Ridgeway, Nelson, New Zealand. 1989), whereas paleoecological studies were lacking. Although fossil deposits in caves, swamps, and dunes provided extreme ly abundant remains, as recently as 1979 fossil avifaunas older than the Holocene in New Zealand were considered rare and limited in size (Fleming, 1979). Since then, extensive investi gations of cave deposits combined with the intensive use of ra diocarbon dating have shown that faunas of the last glacial age are common (Worthy, 1993a; Worthy and Holdaway, 1993, 1994a, 1995). Analysis of moa faunas throughout New Zealand showed that there was a pattern to the distribution of species that was related to habitat (Worthy, 1990). This paper summarizes some of the important new informa tion arising out of these and other recent studies of the Quater nary avifauna of New Zealand by the author and R.N. Hold away. The primary purpose of the research has been to document fossil avifaunas and to describe the faunal changes brought about by climate during the last glacial-interglacial cy cle, mainly during oxygen isotope stages 1 and 2. Study areas around South Island, New Zealand, were chosen for the range of climatic conditions each now has. Each area was kept small, usually 10-20 km across, to minimize geographic and present climate variation. These factors are assumed to have been in strumental in the control of vegetation physiognomies, so a rel atively homogenous vegetation structure within each area is as sumed and is related to the faunal composition. The distribution of birds was most affected by whether the vegeta tion was a closed-canopy forest or a mosaic of shrubland and grassland. The floral composition of the forest seems to have been of secondary importance to its structure, because closed forests, whether dominated by beech or by one of several podocarps, all had the same moa assemblage in the late Ho locene. Also, grassland-shrubland associations in the subalpine zone have markedly different floras from those in lowland en vironments, but the same birds characterize both areas. METHODS.?For each study area, all available fossil faunas of Holocene or Pleistocene age were examined, and extensive new collections were made. Efforts were made to obtain faunas 111 112 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY with diverse taphonomic origins to offset the biases inherent in any one depositional environment. For example, pitfalls overly represent ground-dwelling species, and small, volant passe rines are rare. The chronology of sites is based on 99 new and 27 preexisting radiocarbon dates obtained from bones (Worthy, 1993a, 1997; Worthy and Holdaway, 1993, 1994a, 1995, 1996). Most dates are based on accelerator-mass-spectrometry analysis of collagen or gelatin extractions from single bones as detailed in Worthy (1993a, 1997) and Worthy and Holdaway (1993, 1994a, 1995, 1996). Geologic ages cited hereafter are conventional radiocarbon ages. Nomenclature for species' bi nomials and English names of modern birds follows Turbott (1990). ACKNOWLEDGMENTS.?This work was supported by a grant from the New Zealand Foundation for Research, Science, and Technology, and with funds from the New Zealand Lottery Grants Board for some of the radiocarbon dates. The generous support of the Museum of New Zealand Te Papa Tongarewa, Canterbury Museum, and Otago Museum, and their curators, is gratefully acknowledged. Much of the initial work for this study was done with Richard Holdaway, with whom numerous discussions have enabled the development of ideas contained herein. Lastly, the work would not have been possible without the support of the many land owners and the Department of Conservation permitting access to fossil sites. A critical review by Storrs Olson led to a greatly improved text. The Study Areas GEOLOGY AND CLIMATE FIGURE 1 WEST COAST.?The lowland (0-300 m) karst region be tween Punakaiki in the south and Charleston to the north, on the west coast of South Island, was studied by Worthy and Holdaway (1993). The area has a mild, humid climate, with mean monthly temperatures of 10?-18? C, and 2800-4000 mm of rainfall annually. A tall, closed-canopy, mixed beech {Nothofagus)/podocavp (dominated by Dacrydium cupressi- num Lambert) rainforest characterizes the unmodified vegeta tion. The fossil faunas are from sites in 42 caves and are up to 25,000 years old. Most faunas are from pitfalls. Some sites contained single skeletons lying on the surface; their origin is attributed to "vagrants" (individuals that entered the cave for any number of reasons and that often traveled 10-100 m from the entrance before dying). Often vagrants entered via large, horizontal entrances. The age of all such skeletons was consid ered individually because adjacent skeletons varied in age by thousands of years and may have been deposited during either the last glacial period or the present interglacial. There was only one fauna accumulated by the predatory Laughing Owl {Sceloglaux albifacies) (Worthy and Holdaway, 1994b). HONEYCOMB HILL CAVE SYSTEM.?This cave system lies inland of Karamea, in the northwest part of South Island, in a valley at an altitude of about 300 m that receives 3000-4000 mm of rainfall annually. The present vegetation and tempera tures are similar to those of the west-coast study area, although winter frosts occur. There are more than 50 discrete fossil sites in this complex cave, which has about 14 km of passages and 70 entrances. The Graveyard and the Eagles Roost are the two most important sites (Worthy, 1993a). The fossil deposits are up to 20,000 years old. TAKAKA HILL AND TAKAKA VALLEY.?Takaka Hill and Takaka Valley are in the northern part of South Island. Because southwesterly airflows prevail over New Zealand, and Takaka is east of a tract of mountains, it receives considerably less rainfall than the two previous study areas, about 2000-2500 mm annually. Temperatures are similar to those of the western sites (mean annual temperature for Takaka Valley is 12.7? C), although there is greater seasonality, with summer drought common, and on Takaka Hill snowfalls can be expected during winter. Sites in the valley (0-200 m) were compared to hill fau nas (600-800 m) to detect altitudinal effects. The late Holocene vegetation of the valley was a tall, multitiered, closed-canopy mixed podocarp (dominated by Podocarpus totara D. Don) and broadleaf forest, compared with a closed-canopy (10-15 m), primarily beech {Nothofagus spp.) forest, with some Hall's Totara {Podocarpus hallii Kirk) and cedar {Libocedrus spp.) on the hill. Fossil faunas were obtained from 43 caves in the combined hill and valley areas (Worthy and Holdaway, 1994a), although most were from the more extensive karst areas on the hill. De posits are up to 30,000 years old and are mainly pitfalls, with only two significant faunas from Laughing Owl prey accumu lations. None are in alluvial contexts. NORTH CANTERBURY-MT. COOKSON.?The Mt. Cookson study area is a karst plateau at 400-600 m in the province of North Canterbury. Because it lies just east of the high Amuri Range and is about 20 km from the east coast, there is a marked rain-shadow effect. It has an annual rainfall of about 700 mm and a markedly seasonal climate; summers are hot (tempera tures often >30? C), with drought common, and in winter snow lies on the ground for several weeks. The late Holocene vegeta tion was a closed-canopy beech forest. Fossil faunas were from several pitfall sites and from three deposits accumulated by falcons {Falco novaeseelandiae) (Worthy and Holdaway, 1995). The sites are up to 38,000 years old. NORTH CANTERBURY-WAIKARL?All sites in the Waikari study area are at an altitude of 200-400 m and are within 10 km of Waikari in North Canterbury. The climate is dry, with annual rainfall of about 660 mm, and warm, with a mean an nual temperature of 10.8? C. It is markedly seasonal; summer droughts are common, and in winter frosts and occasional snowfalls are normal. The late Holocene vegetation (5000-1000 years ago) was a tall podocarp (dominated by Prumnopitys taxifolia (Banks & Solander ex Lamb.) de Laub.) forest on the valley floors, with beech forest above this on the higher slopes (to 800 m). Areas of shrublands and grasslands were present on ridges and along river beds. The NUMBER 89 113 FAR NORTH DUNES Dunedin 300km FIGURE 1.?Map of New Zealand showing the study areas on South Island, the four main cities (Auckland, Well ington, Christchuch, and Dunedin), major geographic features, and degrees of south latitude (numbered bars). Waitomo Caves and the Far North dunes are areas with important fossil deposits. vegetation was characterized by its mosaic nature and dif fered markedly from the closed-canopy forests of western re gions. The faunas are from sites with diverse taphonomic histories: 10 sites are Laughing Owl deposits, five are swamps, one is a cave pitfall, two are archaeological, and several are rockshel- ters (Worthy and Holdaway, 1996). Most faunas are of Ho locene age, but one is late glacial (10,000-12,000 years old). The oldest is from Otiran deposits (>24,000-< 100,000 years old) in alluvial beds in the Waipara Valley. SOUTH CANTERBURY.?The South Canterbury study area is at an altitude of 200-400 m and is located between Fairlie, Raincliff, and Bluecliffs Station (25 km southwest of Timaru). The climate is very similar to that of the Waikari study area, and the late Holocene vegetation is assumed also to have been very similar. The 59 fossil sites include 27 Laughing Owl deposits, two falcon deposits, two swamps, and 28 cave pitfall or rockshelter deposits (Worthy, 1997). The deposits are up to 38,000 years old, although most are of late-Holocene age. 114 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY OTAGO.?Three study areas were chosen in Otago because this large region has a very varied geography and climate, in cluding the driest regions of New Zealand. 1. North Otago downlands (<300 m; between Oamaru and Duntroon just south of the Waitaki River): The climate is mild, with an annual temperature of 11? C and low rainfall (500-550 mm annually). Summer drought is common but is not as extreme as in more inland regions of Otago. The late- Holocene vegetation was a tall podocarp {Prumnopitys taxifo- lia dominated) forest on the downlands, but in the alluvial val ley floors Dacrycarpus dacrydioides (Rich.) de Laub. probably prevailed. The faunas are from fossil sites of diverse taphono mies: seven swamps, 11 Laughing Owl deposits, one falcon de posit, and 11 cave or rockshelter deposits of pitfall or vagrant derivation. 2. Wanaka, western Otago: Several sites are at altitudes between 300 m and 600 m in the lee of the Southern Alps near Wanaka. Most are pitfall deposits in fissures formed between large blocks of schist, but three have faunas accumulated by Laughing Owls, and two are swamps. Annual rainfall at Wana ka is low (419-952 mm, mean 682 mm). The late-Holocene vegetation of the hillslopes around most sites was a closed-can opy beech forest, but alluvial river flats had podocarp forests. Adjacent areas of higher altitude, and recently stable river flats, had shrubland and grassland. 3. Alexandra to Cromwell, central Otago: Fossil faunas are from isolated sites in fissures within schist, at altitudes from 300 m to 600 m. Only a few fissures had extensive faunas. Two swamp deposits are present, but no deposits accumulated by predators were found. Alluvial deposits at Chatto Creek contain the only fauna of Otiran age, although it is small. The incised gorges of the Clutha and Manuherikia rivers are up to 300 m deep and are notable features. Broad valleys at 200-300 m are surrounded by rounded ranges rising to 1700 m. In central Ota go, rainfall varies from about 330 mm to 560 mm, with a mean of 409 mm. Mean annual temperature is about 11 ? C at Crom well and Alexandra but rapidly decreases with altitude on the nearby ranges. There is marked seasonality; summer tempera tures often exceed 30? C, with drought common, and in winter frosts are severe, and snowfalls are usual. CLIMATIC AND VEGETATIONAL HISTORY There are many data on the vegetation of the late Holocene but fewer for progressively older time periods. The following is an attempt to place successive fossil faunas in their contempo rary habitat. The period of time considered herein is roughly the last 50,000 years (oxygen isotope stages 1 to 3) and, more impor tantly, the last 20,000 years (stages 1 and 2). In the first half of oxygen isotope stage 3 (ca. 57,000-35,000 years ago (Nelson et al., 1993)), the climate in New Zealand was 2?-3? C cooler than at present. Average annual temperatures then decreased and reached a glacial minimum at 18,000 years ago, during the Kumara-2 glacial advance (Suggate and Moar, 1970; Suggate, 1990), of 4?-5? C cooler than the present (McGlone, 1988; Mildenhall, 1995). There were some minor retreats and ad vances of glaciers, but full glacial conditions lasted at least un til about 14,000 years ago (Suggate, 1965, 1990; Suggate and Moar, 1970), and it has even been suggested that the retreat of glaciers did not begin until 12,500 to 13,000 years ago in some areas (Mabin, 1983). Temperatures approaching those of the present were achieved about 10,000 years ago. An increase in precipitation was associated with warming, whereas the glacial periods were cold and dry. The coldest periods of the glaciation saw the treeline lowered by an average of 800 m to 830 m below that of the present (Soons, 1979; McGlone, 1985, 1988). Trees and shrubs de clined in importance from about 30,000 years ago and by 18,000 years ago were in low percentages, if present at all, in most sites. A corresponding rise in the representation of pollen of Poaceae and various shrub taxa shows that a mosaic of grassland and shrubland dominated the landscapes. In western areas, at lower altitudes, it is probable that some stands of for est survived because it is unlikely that rainfall could have dropped to sufficiently low levels to have prevented the growth of forest, as it did in eastern areas (McGlone, 1988). In the west-coast study area, the vegetation during the cold est times of the last glacial period (Otiran) was a mosaic of tall shrubland and beech forest, with grassland and shrubland in the river valleys, induced by localized cold air drainage off the mountains. The pollen record indicates that a tall, closed-cano py podocarp forest became established shortly after 12,000 years ago and has been the vegetation ever since (McGlone, 1988; Worthy and Holdaway, 1993). Around Honeycomb Hill Cave, a tall shrubland, including tree ferns, persisted in the valley floor throughout the glacial period, but shorter, shrubland-grassland mosaics occupied ad jacent slopes. The present tall, closed-canopy, mixed podocarp forest became established between 10,000 and 12,000 years ago (Worthy and Mildenhall, 1989). On Takaka Hill, the sites were well above the depressed treeline of the last glacial period. Although a glacial pollen record for this region is lacking, it seems reasonably certain that there was a mosaic of shrubland and grassland around the fossil sites at this time. With the warming at the end of the gla cial period, a tall shrubland became established, which was re placed about 8000 years ago by a beech forest that has persist ed to the present day (McGlone, 1988; Worthy and Holdaway, 1994a). A similar history is envisaged for Mt. Cookson, al though it is probable that this more eastern area was even drier and that the Otiran vegetation was even more open and sparse and was dominated by grassland (McGlone, 1988; Worthy and Holdaway, 1995). The vegetational history of the study areas in North and South Canterbury and in North Otago is inferred to have been similar (McGlone, 1988). Shrublands and grasslands were the primary vegetation types throughout the Otiran period. In the late glacial, about 12,000 years ago, a taller shrubland commu nity became established, and areas of grassland were reduced. NUMBER 89 115 Tall shrublands persisted until 5000-6000 years ago, when in creased precipitation allowed a tall podocarp forest to develop, at least in valley floors and on lower slopes. Beech forest spread in higher altitudes, but shrubland remained a significant component of regional vegetation. Between 800 and 1000 years ago all forest and tall shrubland was destroyed by anthro pogenic fires, and grassland and short shrubland again became widespread. McGlone et al. (1995) described pollen profiles from central Otago sites that record the vegetation from the late Pleistocene to the late Holocene. Between about 12,000 and 9000 years ago, a low scrub (1-2 m high) of small-leaved and xerophytic species formed a mosaic with Chionochloa grassland. Al though tall podocarp forest was established in coastal areas of Southland and Otago by 9500 years ago, such forest is unlikely to have existed in other than small isolated stands in the interior before about 7500 years ago. The delay in forest establishment there may be explained by a decrease in available water at that time, either by lower rainfall or by a combination of increased evapotranspiration (resulting from higher temperatures) and decreased rainfall. The absence of, or very slow, peat deposi tion during this interval seems to support a prevailing water deficit. An abundance of tree-fern spores in the same period, however, suggests water was not limiting, at least in sheltered gullies. Frequent fires enabled the continued presence of tree ferns by stopping the establishment of slower growing, fire- sensitive podocarps, thus maintaining serai conditions. About 7500 years ago a coniferous forest of Prumnopitys taxifolia, Dacrycarpus dacrydioides, and Podocarpus abruptly replaced lower altitude grassland communities, whereas Phyl- locladus alpinus Hook.fi and Halocarpus bidwilli (Hook.f. ex Kirk) Quinn formed the upper treeline. The afforestation has been attributed to increased precipitation and a slight decrease in temperature. Nothofagus menziesii (Hook.f.) Oerst. became established in the area about 6000 years ago, followed shortly by a Nothofagus fusca type and Dacrydium cupressinum, al though Phyllocladus dominated pollen assemblages. Signifi cant percentages of pollen of shrub taxa such as Coprosma, Asteraceae, and Poaceae indicate the continued presence of grassland-shrubland communities above the treeline. After 3000 years ago, episodic destruction of podocarp forests by fire resulted in a reduction in the frequency of some tree pollen, es pecially Prumnopitys taxifolia, and an increase in grass pollen. Forests were widely destroyed by anthropogenic fires, resulting in a sudden proliferation of Pteridium esculentum (Forster f.) Nakai spores and vastly increased amounts of charcoal about 600 years ago. Avifaunal Changes The composition of the Late Quaternary and Holocene avi fauna of terrestrial and inland wetland habitats is summarized in Table 1 from data in Millener (1990), with alterations as elu cidated in Worthy (1993a, 1997) and Worthy and Holdaway (1993, 1994a, 1996). The following notes support the numbers of species listed herein as inhabiting such inland areas. Pelecaniformes Only Pelecanus, Phalacrocorax carbo, and P. melanoleucos are inland taxa. Ciconiiformes Only Egretta alba, Botaurus stellaris, and Ix obrychus novaezelandiae are inland taxa. Anseriformes I do not accept Anas rhynchotis or Oxyura as part of the prehuman fauna, Mergus was coastal, and only Cnemiornis had endemic species on each island. Falconiformes Circus approximans is a recent immigrant, so the prehuman fauna comprised one eagle, one harrier, and one falcon. Gruiformes The North and South Island pairs Aptornis otidiformislA. defossor and P. mantellil Porphyrio hochstetteri each consist of sepa rate species (Trewick, 1996); Porphyrio p. melanotus is considered a recent immigrant; Gallirallusphilippensis is recorded from gla cial deposits on South Island; and Porzana tabuensis and P. pusilla, although rare as fos sils, are assumed to have been on both is lands. Charadriiformes Only Haematopus unicolor, Charadrius bicinctus, Thinornis novaeseelandiae, Ana- rhynchus frontalis, Coenocorypha auckland ica, Himantopus novaezelandiae, Larus do- minicanus, L. bulleri, and Sterna albostriata used inland areas habitually. Passeriformes There are seven acanthisittid wrens, among which Pachyplichas had discrete species on each island and Dendroscansor decurviros- tris was endemic to South Island. The three TABLE 1.?The number of species in each order of birds that inhabited prehuman inland wetlands and/or terrestrial habitats on North and South islands, New Zealand. Data is from Millener (1990), Worthy (1993a, 1997), and Worthy and Holdaway (1993, 1994a, 1996). Order DiNORNITHIFORMES APTERYGIFORMES PODICIPEDIFORMES PELECANIFORMES CICONIIFORMES ANSERIFORMES FALCONIFORMES GALLIFORMES GRUIFORMES CHARADRIIFORMES COLUMB1FORMES PSITTACIFORMES CUCULIFORMES STRIGIFORMES CAPRIMULGIFORMES CORACIFORMES PASSERIFORMES Total North Island 7 2 2 3 3 11 3 1 9 9 1 5 2 2 1 1 20 82 South Island 9 3 2 3 3 11 3 1 8 9 1 5 2 2 1 1 20 84 Total 11 3 2 3 3 12 3 1 11 9 1 5 2 2 1 1 24 94 116 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY species of Mohoua are distributed one on North Island and two on South Island. North Island has two endemic monotypic genera, Notiomystis and Heterolocha. Procellariids are not listed here, although at least 15 spe cies nested on one island or the other. A total of 94 species lived in inland habitats in mainland ar eas of New Zealand prior to human colonization: 82 on North Island and 84 on South Island. The birds recorded in fossil fau nas from South Island are listed in Table 2. CHANGES CAUSED BY THE LOWERED TREELINE Some faunal changes are directly explicable as the result of downslope movement of faunal groups that in the Holocene are associated with the subalpine zone. The recovery of faunas from the numerous fossil sites in the extensive areas of karst in subalpine areas of northwest Nelson has provided considerable data (Worthy, 1989, unpublished data). Fossil avifaunas from caves now above the treeline (>1200 m) are assumed to have accumulated in the last few thousand years of the Holocene be cause the karst was glaciated in the last glacial period. More- TABLE 2.?The fossil avifauna of South Island, New Zealand. West coast data is from Worthy and Holdaway (1993), with glacial faunas derived particularly from Babylon Cave, Hermits Cave, and Honeycomb Hill (Worthy, 1993). Takaka faunas are from Worthy and Holdaway (1994a), with sites in Irvines Tomo, the cave in the Golden Bay Cement Co. silica quarry, Hawkes Cave, Kairuru Cave, and Hobsons Tomo, the most important site for the Otiran fauna. Data for North Canterbury is mainly from the Waikari study area (Worthy and Holdaway, 1996), but the last glacial faunas are derived from Merino Cave, Mt. Cookson (Worthy and Holdaway, 1995), fluvial sites at Omihi Stream, Waipara (Worthy and Holdaway, 1996), and from loess sites (Worthy, 1993b). Data for South Canterbury and for Otago is from Worthy (1997, unpublished data, respectively). (Ab=species at least locally abundant, coastal=species present only in coastal sites, rare=species rare, valley=species present only in Takaka Valley, Y = species present in fossil record.) Taxon Megalapteryx didinus Anomalopteryx didiformis Pachyornis elephantopus Pachyornis australis Euryapteryx geranoides Emeus crassus Dinornis struthoides Dinornis novaezealandiae Dinornis giganteus Apteryx australis/haastii Apteryx owenii Poliocephalus rufopectus Procellaria parkinsoni Procellaria westlandica Pterodroma inexpectata Pterodroma cookii Puffinus griseus Puffinus spelaeus (s) or gavia/huttoni (g/h) Pelecanoides urinatrix Oceanites nereis Fregetta tropica ssp. Pelagodroma marina Pachyptila turtur Anas chlorotis Anas gracilis Anas superciliosa Aythya novaeseelandiae Euryanas finschi Hymenolaimus malacorhyn- chos Malacorhynchus scarletti Biziura delautouri Tadorna variegata Cnemiornis calcitrans Cygnus sumnerensis West Coast (Glacial) Y(Ab) Y(Ab) Y1 Y (Ab)2 Y Y Y Y Y(s)(Ab) Y (Ab)3 ? Y(Ab) Y(Ab) ? Y Y (Ab)4 West Coast (Holocene) Y Y(Ab) Y Y(Ab) Y(Ab) Y Y Y Y Y Y(s) (Ab) Y ? Y Y Y Takaka (Glacial) Y Y(Ab) Y Y(Ab) Y Y Y (valley) Y Y(s) Y(Ab) Y (valley) Y (valley) Takaka (Holocene) Y(Ab) Y(Ab) Y(Ab) Y(Ab) Y(Ab) Y Y(s) Y(Ab) Y North Canterbury (Glacial) Y Y(Ab) Y(Ab) Y(Ab) Y Y Y Y Y Y Y(Ab) Y Y Y North Canterbury (Holocene) Y(rare) Y(Ab) Y(Ab) Y(Ab) Y Y(Ab) Y Y Y Y Y Y (coastal) Y(g/h) Y Y Y(Ab) Y Y Y Y(Ab) Y Y(Ab) Y South Canterbury (Glacial) Y(Ab) Y(Ab) Y Y Y(Ab) Y(Ab) South Canterbury (Holocene) Y (rare) Y (rare) Y(Ab) Y(Ab) Y(Ab) Y Y (rare) Y Y Y Y Y Y(g/h) Y Y Y Y Y Y(Ab) Y Y Y(Ab) North Otago (Holocene) Y(Ab) Y(Ab) Y(Ab) Y Y Y Y Y Y(g/h) Y Y Y Y Y Y(Ab) Y Y Y Y(Ab) Y Central Otago (Holocene) Y(Ab) Y(Ab) Y(Ab) Y(Ab) Y Y (rare) Y Y Y Y Y Y(Ab) Y Y Y(Ab) NUMBER 89 117 Taxon Strigops habroptilus Nestor meridionalis Nestor notabilis Cyanoramphus spp. Ninox novaeseelandiae Sceloglaux albifacies Aegotheles novaezealandiae Hemiphaga novaeseelan diae Eudynamys taitensis Falco novaeseelandiae Circus eylesi Harpagornis moorei Gallinula hodgenorum Gallirallus australis Gallirallus philippensis Porphyrio hochstetteri Fulica prisca Aptomis defossor Larus dominicanus Coenocorypha aucklandica Charadrius bicinctus Thinornis novaeseelandiae Sterna albostriata Himantopus novaezelandiae Egretta alba Phalacrocorax carbo Phalacrocorax varius Coturnix novaezelandiae Acanthisitta chloris Xenicus sp. Traversia lyalli Pachyplichas yaldwyni Dendroscansor decurvirostris Bowdleria punctata Prosthemadera novae seelandiae Anthornis melanura Petroica australis Petroica macrocephala Mohoua ochrocephala Mohoua novaeseelandiae Gerygone igata Corvus moriorum Rh ipidura fuliginosa Anthus novaeseelandiae Callaeas cinerea Philesturnus carunculatus Turnagra capensis TOTAL6 West Coast (Glacial) Y(Ab) Y Y Y Y Y Y Y Y Y Y(Ab) Y Y Y Y (Ab)5 Y Y Y Y Y ? Y Y Y Y Y Y Y Y 42-45 West Coast (Holocene) Y(Ab) Y(Ab) Y Y Y(Ab) Y Y Y Y(Ab) Y Y Y(Ab) Y(Ab) Y(Ab) Y(Ab) Y Y Y(Ab) Y(Ab) Y Y Y Y Y(Ab) Y(Ab) Y 42 Takaka (Glacial) Y Y Y Y Y Y Y Y ? Y Y Y Y Y Y Y 27-28 TABLE 2 Takaka (Holocene) Y(Ab) Y Y Y(Ab) Y Y Y(Ab) Y Y Y Y(Ab) Y Y Y Y (rare, valley) Y(Ab) Y(Ab) Y(Ab) Y(Ab) Y Y(Ab) Y(Ab) Y Y(Ab) Y Y Y Y rare Y Y(Ab) Y rare 40 ?Continued. North Canterbury (Glacial) Y Y Y Y Y Y Y Y Y Y Y 25 North Canterbury (Holocene) Y Y(Ab) Y Y(Ab) Y Y(Ab) Y(Ab) Y(Ab) Y Y Y Y(Ab) Y(Ab) Y Y(Ab) Y(Ab) Y(Ab) Y Y Y Y Y Y(Ab) Y Y Y Y Y Y(Ab) Y Y(Ab) Y Y Y Y Y Y Y Y(Ab) Y Y(Ab) 64 South Canterbury (Glacial) Y Y Y Y Y Y Y 13 South Canterbury North Otago (Holocene) Y Y Y Y(Ab) Y Y(Ab) Y Y(Ab) Y Y Y Y(Ab) Y Y Y Y Y Y Y(Ab) Y Y Y Y Y Y Y Y(Ab) Y Y Y Y Y (rare) Y(Ab) Y(Ab) Y(Ab) Y Y(Ab) 59 (Holocene) Y Y Y Y(Ab) Y Y(Ab) Y Y(Ab) Y Y Y Y Y(Ab) Y Y Y Y Y Y Y Y Y(Ab) Y Y Y Y Y Y Y(Ab) Y Y Y Y Y Y(Ab) Y(Ab) Y Y(Ab) 58 Central Otago (Holocene) YiAb) Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y 16 1 Pachyornis australis was abundant in Honeycomb Hill (300 m above sea level (a.s.l.)) Otiran faunas. 2Euryapteryx geranoides was common at low levels on the west coast but was rare at Honeycomb Hill. 3Pelecanoides urinatrix was abundant in some coastal sites in the Otiran, e.g., Road Cave (Worthy, unpublished data). ACnemiornis was abundant at low levels on the west coast but was rare at Honeycomb Hill in the Otiran. 5Aptornis was abundant at Honeycomb Hill (300 m a.s.l.) but was rare at low levels on the west coast in the Otiran. 6This total species diversity does not include recent self-introduced species, listed below, or any of the European introductions that are incorporated in the youngest faunas. The following species are not found in any deposits demonstrably older than 1000 years and so are assumed to have colonized New Zealand following the habitat disruptions caused by Polynesians: shoveler (Anas rhynchotis), pukeko (Por phyrio p. melanotus), and Australasian Harrier (Circus approximans). 118 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY over, they did so in much the same environment as is now there (Worthy, 1989). The dominant moa in subalpine sites is the Upland Moa {Megalapteryx didinus), with the Crested Moa {Pachyornis australis) the only other emeid. The dinomithids are most com monly represented by the Slender Moa {Dinornis struthoides) and the Large Bush Moa {D. novaezealandiae); the Giant Moa {D. giganteus) has never been found at these altitudes. Associ ated birds included Finsch's Duck {Euryanas finschi), Great- spotted Kiwi {Apteryx haastii), Little-spotted Kiwi {Apteryx owenii), New Zealand Coot {Fulica prisca), South Island Taka- he {Porphyrio hochstetteri), Weka {Gallirallus australis), Ey- les's Harrier {Circus eylesi), Haast's Eagle {Harpagornis moorei), New Zealand Falcon {Falco novaeseelandiae), Kaka po {Strigops habroptilus), Kea {Nestor notabilis), New Zealand Pipit {Anthus novaeseelandiae), Rock Wren {Xenicus gilviven- tris), and Stephens Island Wren {Traversia lyalli) (Worthy, 1989; unpublished data). Downslope from the subalpine zone there is a gradual change in the species composition of moa faunas. For example, in sites of Holocene age in northwest Nelson near Mt. Arthur, Megalapteryx didinus dominates assemblages between 700 m and 900 m, but Anomalopteryx didiformis also is present. Be low 700 m, A. didiformis is the only emeid present. Similar al- titudinal changes in species composition are known for Fiord- land (Worthy, 1989). On Takaka Hill, the Holocene moa fauna is dominated by A. didiformis, whereas M. didinus is unknown (Worthy and Hold away, 1994a). In deposits of the last glacial age, however, M. didinus is present, and A. didiformis is absent, a difference best explained as the result of altitudinal depression of the subalpine ecosystems. The deposits in Honeycomb Hill Cave, to the west, record a similar pattern: M. didinus and Pachyornis australis dominate deposits 14,000-20,000 years old. REGIONAL CHANGES IN AVIFAUNAS The most significant result of the recent studies of the South Island Quaternary avifaunas is that faunas from sites separated by as little as a few meters may differ markedly in species com position because of different ages. As a result, where in the past such associations were used as evidence for the coexistence of various species (e.g., Atkinson and Millener, 1991), they are now known to be the result of deposition at different times with markedly different environments. Graham and Lundelius (1984) expressed the opinion that most individual stratigraphic units are deposited over too short a time period for them to have accumulated through periods of environmental change. In New Zealand, unconformities separating deposits of glacial, late glacial, and Holocene age are the exception rather than the rule, and many sites have faunal remains essentially on the cave floor that range in age from modern to 20,000-30,000 years old, such as Hawkes Cave (Worthy and Holdaway, 1994a). Articulated skeletons of all ages indicate continuous deposition throughout this time. In the caves where many dates on individual bones are available, such as Madonna Cave (Worthy and Holdaway, 1993), Hawkes Cave, Kairuru Cave, Irvines Cave (Worthy and Holdaway, 1994a), and Honeycomb Hill Cave (Worthy, 1993a), the association of the moas Pachy ornis elephantopus and Euryapteryx geranoides with Anoma lopteryx didiformis is shown to be the result of deposition at different time periods. Many other undated talus accumulations beneath cave entrances have essentially unstratified deposits, with these same species found together, indicating that the de posits were accumulated and mixed over a significant time pe riod, for example, Ngarua Cave and Commentary Cave (Wor thy and Holdaway, 1994a). Graham (1993) described deposits such as these as time-averaged sequences and detailed numer ous methods, with examples, whereby disharmonious associa tions could form by various time-averaging processes. In New Zealand, the factors that promote time-averaged sequences are constant humidity (deposits are always wet), low temperatures (most sites average <10? C, so weathering is slow, bones last longer, and weathering of cave surfaces provides little sedi ment), and low rates of alluvial sedimentation. The fossil faunas from New Zealand caves in the regions now characterized by wetter climates all record major changes in the living faunas between the Otiran glacial period and the Holocene, especially the late Holocene (<6000 years ago). Thus, there are two distinct faunas in the fossil deposits, as the data in Tables 2 and 3 show. THE WESTERN OTIRAN FAUNA.?The Otiran (last glacial) lowland fauna of western regions is characterized by the pres ence of the Stout-legged Moa {Euryapteryx geranoides), the Heavy-footed Moa {Pachyornis elephantopus), and a large morph of Megalapteryx didinus. Associated carinates include the South Island Goose {Cnemiornis calcitrans), South Island Adzebill {Aptornis defossor), Fulica prisca, New Zealand Gallinule {Gallinula hodgenorum), Harpagornis moorei, and Euryanas finschi. The members of this group, hereafter termed the Euryapteryx assemblage (Table 3), are all unknown from, or are very rare in, Holocene deposits in these areas. There are differences in the relative frequency of species be tween the study areas within this western region. For example, the Otiran fauna at Honeycomb Hill differs from those in the lower-altitude west coast area in that Pachyornis australis is abundant, P. elephantopus is rare, Cnemiornis is rare, and Ap tornis is common. In the Punakaiki karst, P. australis is rare, and Cnemiornis is more common than Aptornis. THE WESTERN HOLOCENE FAUNA.?In contrast to Otiran faunas, Holocene deposits have a distinctive suite of species termed the Anomalopteryx assemblage (Table 3). The small emeid Anomalopteryx didiformis dominates this fauna and is usually associated with Dinornis struthoides and D. no vaezealandiae. If Megalapteryx didinus is present, it is only as small morphs. Kakapo {Strigops habroptilus), Weka {Galliral lus australis), and kiwis {Apteryx spp.) are common, and the Brown Teal {Anas chlorotis) is the duck most often encoun- NUMBER 89 119 TABLE 3.?Lists of characteristic species of the Anomalopteryx and Euryapteryx assemblages that respectively characterize the closed-canopy wetter forests typical of western areas in the Holocene and the forest- shrubland-grassland mosaics of the drier eastern areas, when these species occur together in abundance. Anomalopteryx assemblage Euryapteryx assemblage Anomalopteryx didiformis Dinornis novaezealandiae Apteryx australis Strigops habroptilus Anas chlorotis Gallirallus australis Callaeas cinerea Philesturnus carunculatus Pachyplichas jagmi/yaldwyni Xenicus longipes Petroica australis Euryapteryx geranoides Euryapteryx curtus Emeus crassus Dinornis giganteus Pachyornis mappini/elephantopus Pachyornis australis (uplands only) Megalapteryx didinus (uplands only) Cnem iorn is gracilis/calcitrans Euryanas finsch i Aptornis otidiformis/defossor Gallinula hodgenorum Fulica prisca Harpagornis moorei Coturnix novaezelandiae tered. Acanthisittid wrens of several species are abundant in deposits (when conditions of preservation permit), and the New Zealand Robin {Petroica australis) and Saddleback {Philestur nus carunculatus) are abundant. It is important to note that these species are found in faunas deposited under vegetational mosaics but are relatively rare and infrequent, in contrast to their numerical abundance and dominance of faunas from areas where the vegetation was a closed-canopy forest. It is therefore not presence or absence so much as relative frequency that is important in the definition of this assemblage. In contrast, the mere presence of species listed in the Euryapteryx assemblage indicates the presence of open habitat. NORTH CANTERBURY.?The Mt. Cookson study area occu pies an intermediate zone between the wetter west and the drier east and, as expected, does not exhibit the same degree of fau nal turnover. The drier climate of eastern areas is probably re sponsible for the delay of the establishment of closed-canopy forest until about 6000 years ago, much later than in the west. During the latter stages of oxygen isotope stage 3 and the early part of stage 2, Pachyornis elephantopus dominated moa fau nas on Mt. Cookson but was associated with rare remains of Dinornis giganteus, D. struthoides, Megalapteryx didinus, Emeus crassus, and Euryapteryx geranoides. During the cold est period of the last glacial, however, P. elephantopus was vir tually the only moa living in these and other eastern land scapes, where loess was being deposited. But, as in the west, with warming temperatures and forest establishment in the late Holocene, Anomalopteryx didiformis and Dinornis no vaezealandiae came to dominate the faunas. Unlike in the west, however, Euryanas finschi and Aptornis defossor remained common in the late Holocene fauna, perhaps because the for ests remained much drier. THE TYPICAL EASTERN FAUNAS.?In contrast to the above faunas, those of the lowlands in North and South Canterbury and North Otago provide no evidence for any faunal turnover in the period spanning the last glacial to the late Holocene. The greatest change detected is a reversal of dominance roles with in the same suite of species: whereas Emeus crassus, Euryap teryx geranoides, and P. elephantopus were the main emeids throughout this time, Pachyornis elephantopus dominated gla cial faunas, but Emeus crassus was the most common during the Holocene. Dinornis giganteus and D. struthoides were the dominant dinornithids during the glacial and the Holocene. In these eastern lowlands, there is no evidence of local extir pation of taxa at the end of the glacial period, as there is for western regions. The lack of glacial faunas containing smaller birds limits comparisons of faunal turnover to moas. Because there are no drastic changes in the moa faunas, however, and because Cnemiornis, Harpagornis, and Aptornis are known to have frequented the Otiran landscapes, and continued to do so in the Holocene, we can speculate on the composition of the as sociated fauna. These larger birds are part of the Euryapteryx assemblage that frequented Otiran western landscapes. So, as in the west, Euryanas finschi, Gallirallus australis, Nestor no tabilis, New Zealand Crow {Corvus moriorum), New Zealand Pipit {Anthus novaeseelandiae), New Zealand Quail {Coturnix novaezelandiae), Piopio {Turnagra capensis), Fulica prisca, and Gallinula hodgenorum were probable associates of Cnemi ornis, Harpagornis, and Aptornis in Canterbury. The relative abundance of species in late Holocene faunas of the east differs from that of western areas. This partly results from eastern areas having a much greater diversity resulting from the continued presence of all the presumed Otiran species during the Holocene, when other species typical of forest habi tats became established. Such species include Kakapo {Stri gops habroptilus), Philesturnus carunculatus, and Petroica australis. These species, however, are all relatively rare com pared to their abundance in western faunas. Differences in the frequency of a species can usually be explained by the avail ability of the preferred habitat of that species. For example, the grassland inhabiting quail, Coturnix novaezelandiae, had little or no habitat in the west during the Holocene and is rare there, but in the east it is common. The most common passerine in late Holocene deposits in eastern areas was Turnagra capensis, which suggests that the preferred habitat of this extinct bird was the shrubland mosaics of drier areas. This is supported by the observation that Turnagra is present in fossil deposits of western areas only in Otiran deposits, when shrubland habitats were widely available. THE CENTRAL OTAGO FAUNAS.?The few data available for the Otiran fauna of central Otago indicate Pachyornis elephan topus and Dinornis giganteus were the most common moas, with Cnemiornis calcitrans being the only known carinate. The abundance of the last species in these deposits and in Otiran loess deposits in Canterbury illustrates its preference for the open, short shrubland/grassland habitats prevailing at that time. The Holocene moa fauna of central Otago is most similar to eastern ones, but it is influenced by altitude. In the broad val leys, Emeus crassus is common, and E. geranoides and P. ele- 120 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY phantopus are common associates. Megalapteryx didinus, however, also is common, especially in hill sites, which reflects the presence of upland shrubland habitats, as this species domi nates subalpine Holocene deposits. In central Otago, the small er carinates are better represented in late-Holocene deposits than in older deposits, but, even so, they are not easily com pared with those of other eastern areas because they do not in clude faunas accumulated by predators. As in Eastern faunas, Euryanas finschi is abundant, Sceloglaux albifacies, parakeets {Cyanoramphus spp.), New Zealand Snipe {Coenocorypha aucklandica), Gallinula hodgenorum, New Zealand Pigeon {Hemiphaga novaeseelandiae), and Coturnix novaezelandiae are common, Cnemiornis calcitrans is present, and Gallirallus australis and Apteryx spp. are relatively rare. Nestor notabilis is more abundant than in other regions, which also reflects the presence of substantial areas of upland habitat. The Anomalopteryx assemblage characterized rimu-dominat- ed podocarp forests of the west coast and central North Island and the beech forests of Takaka Hill and Mt. Cookson during the Holocene. This observation supports the contention of Gra ham (1992) that vegetation structure may be more important to some animals than the species composition of the vegetation. Although these areas differed markedly floristically, they pre sented a common structure of a continuous closed canopy that excluded significant areas of grassland and shrubland. THE QUESTION OF PLEISTOCENE EXTINCTIONS OF MEGAFAUNA The faunal turnover at the end of the glacial period in New Zealand has considerable international relevance to the world wide debate on the cause or causes of Pleistocene extinctions of megafauna, whether climate-induced or attributable to over kill by humans (Martin and Klein, 1984; Graham, 1986). The term megafauna has been defined in various ways, although most definitions encompass the larger species (Martin and Klein, 1984). In Australia, megafauna is often used for all spe cies that went extinct in the late Pleistocene, regardless of size (Murray, 1991). In New Zealand, as elsewhere in the world, larger species were more susceptible to extinction, with all ter restrial birds greater than two kilograms becoming extinct (Cassels, 1984). The faunal changes documented for western areas of the South Island (Worthy, 1993a; Worthy and Hold away, 1993) demonstrate that in New Zealand there were ex tinctions at the end of the Pleistocene, but they were only re gional in extent. These are equivalent to the faunal shifts commonly related to past climatic changes in Quaternary fau nas of Australia (Lundelius, 1983) and North America (Gra ham, 1987, 1992; Graham and Grimm, 1990). The warming climate in New Zealand produced habitats characterized by continuous tracts of closed-canopy forest for which the members of the Euryapteryx assemblage were not adapted, and so they were displaced by the Anomalopteryx as semblage. Such faunal shifts support the environmental change/habitat destruction hypothesis advocated as causal for North American Pleistocene extinctions (Graham, 1986). In New Zealand, however, all known species survived into the last millenium, with local adjustments in range. The Euryap teryx assemblage was restricted to the areas of grassland, shru bland, and forest mosaics that persisted east of the Southern Alps. Their continued survival in these areas could be consid ered an accident of geography because the Alps create a rain- shadow, hence the dry conditions necessary to maintain this vegetational mosaic. But this is not the only reason because in the North Island and along the Southland coast, the ecotonal dunelands present a fundamentally similar vegetation structure, albeit in very small areas, sufficient for the survival of the dominant Otiran species alongside members of the Anomalop teryx assemblage. Also, the past survival of these species through several glacial-interglacial cycles suggests that they were not at risk of extinction in this last cycle. All moas, indeed all of New Zealand's Late Quaternary terrestrial species that eventually became extinct, did so only after humans arrived, about 800 to 1000 years ago (Anderson, 1991). In North America, greater habitat heterogeneity during the glacial and late glacial is associated with faunas of higher spe cies diversity than those of the Holocene, so the loss of this habitat variety may have contributed to megafaunal extinctions (Graham, 1985, 1986). In New Zealand, although the members of the Euryapteryx assemblage lived in the areas of most heter ogenous habitat, the greatest species diversity was achieved not in glacial times but rather during the late Holocene, when the warm-temperate forest element populated the forest segments of this mosaic. It may be equally valid to argue that, in a land scape that was otherwise forested, the species with require ments for grassland and shrubland habitats found refuge in these mosaics. That all members of the Euryapteryx assem blage became extinct seems to support the concept that the loss of habitat heterogeneity was important in the extinction event. Countering this, however, is the fact that many members of the Anomalopteryx assemblage also became extinct. The presence of heterogenous habitats are not implicitly a glacial/late glacial phenomenon, as inferred for North America by Graham (1992), but are rather a function of water availabili ty and ecotonal habitats. That species with preferences for open areas, such as grassland, closed forests, or forest margins, can all find available habitat in such areas contributes to high spe cies diversity. In New Zealand, forest remnants in the vegeta tional mosaics probably provided the source from which the members of the Anomalopteryx assemblage spread to dominate the faunas of the new closed-canopy forests of the Holocene. Conversely, at an earlier stage of the glacial-interglacial cycle, the Euryapteryx assemblage spread from remnant mosaic habi tats existing in the last interglacial to dominate the open glacial landscapes. NUMBER 89 121 It therefore seems unlikely that the last of many phases of al ternate constriction and expansion of areas of particular vege tation physiognomies caused by climatic shifts contributed to megafaunal extinction in New Zealand. New Zealand differs fundamentally from North America or Australia in that hu mans were not present 10,000 years ago. The New Zealand da ta, demonstrating a combination of regional extirpations at the end of the Pleistocene and total extinction when humans ar rived, suggest that neither overkill nor changing environments are wholly explanatory hypotheses. Both are contributing fac tors in a complex interaction. Murray (1991) reviewed the evi dence on megafaunal extinction in Australia and concluded that although many factors were involved, the megafauna would have survived until European arrival without the influ ence of aboriginal humans. The New Zealand data fit with Murray's conclusion. In summary, climatic change at the end of the Pleistocene led to widespread habitat change in continental Australia and North America, vastly reducing the available habitat for the megafauna and their associates inhabiting Pleistocene vegeta tional mosaics. These species were thus compressed into small areas, where, in the absence of humans, they are likely to have survived as they did in New Zealand. Humans entered both North America and Australia several thousand years before the end of the last ice age and its associated climatic and vegeta tional changes. It seems probable that the continued exploita tion of megafaunal species as food resources, after environ mental changes had severely constricted the ranges of these species, then led to their extinction. The concept of the community in the evolving biota can be examined by studying the pattern of changes in the range of species. The individualistic model suggests communities are an amalgam of species that respond to changes in their environ ment in accordance with individual tolerances. As a result, communities are continually evolving, and modern associa tions do not necessarily represent analogs for previous time pe riods (Graham, 1985, 1992; Graham and Grimm, 1990). North American fossil faunas provide much support for this concept (Graham, 1992). An alternative hypothesis is that individuals are constrained by their ecological requirements and form rec ognizable associations that move across the landscape follow ing available habitat. Alroy (MS) reanalyzes the North Ameri can data and finds much support for this concept of ecological tracking. The homogeneity of the Euryapteryx assemblage throughout Pleistocene and Holocene landscapes of South Is land suggests that this suite of species was inextricably linked by habitat requirements; hence, it supports the ecological track ing model. The presence of Euryanas and Aptornis in Holocene forests of Mt. Cookson is an apparent contradiction of the dis creteness of the Anomalopteryx and Euryapteryx assemblages, but it is explained by this area being an ecotonal zone between dry and wet areas and thus supporting a mix of species. COMPARISON OF CLIMATIC VERSUS HUMAN EFFECTS.?The small, relatively unmodified area of Takaka Hill invites com parison of the relative importance of climatic and human ef fects on the fauna. At the end of the last glacial period, the moas Megalapteryx didinus, Pachyornis elephantopus, P. aus tralis, and Euryapteryx geranoides were displaced from Taka ka Hill. With them went their main predator, Harpagornis moorei, and at least the associated species Euryanas finschi, Aptornis defossor, Fulica prisca, and Gallinula hodgenorum. Dendroscansor decurvirostris was probably lost from the area at the same time. During the Holocene, the faunal composition remained constant from the time of forest reestablishment, about 9000 to 10,000 years ago, until humans arrived in New Zealand. Major changes in the composition of the fauna result ed from various human activities and from predation by newly introduced mammals (Cassels, 1984; Anderson, 1989; Hold away, 1989; Bell, 1991). Initially, only humans and the Pacific Rat {Rattus exulans (Peale)) preyed on the native fauna, but with the coming of Europeans another wave of predators swept through the forest (King, 1984), and much of the land was cleared for farming. The result has been that in the last 1000 years, 10 species on Takaka Hill have become locally extinct: Apteryx owenii, A. haastii, A. australis, Anas chlorotis, Strigops habroptilus, Porphyrio hochstetteri, Xenicus gilviventris, Mohoua ochro- cephala, Callaeas cinerea, and Philesturnus carunculatus. A further 11 species became globally extinct: Anomalopteryx didiformis, Dinornis struthoides, D. novaezealandiae, Scel oglaux albifacies, Aegotheles novaezealandiae, Circus eylesi, Coturnix novaezelandiae, Xenicus longipes, Traversia lyalli, Pachyplichas yaldwyni, and Turnagra capensis. Therefore, at least 21 bird species living on Takaka Hill about 1000 years ago have been extirpated by the activities of humans and vari ous introduced mammals, with most losses being among the terrestrial browser and nocturnal guilds (Holdaway and Wor thy, 1996). Of the remaining birds, Nestor meridionalis and Cyanoramphus spp. are in serious decline in the area. In the first wave of extinctions associated with Polynesian arrival, the species that became extinct were large and flightless (moas, Aptornis, Cnemiornis), and thus susceptible to human hunting, or were very small and flightless or ground nesting (e.g., flightless acanthisittid wrens and procellariids), and thus subject to predation by the Pacific Rat. With the arrival of mustelids, other rats, and cats, other mainly flightless or weak- flying species became extinct or endangered {Sceloglaux, Stri gops, Nestor). The impact of humans caused the regional or total extinction of more than double the number of birds that were only displaced from Takaka Hill at the Otiran/Holocene transition and so was considerably worse than the effect of major climate change. Note Added In Press: In 1998, the Weka Gallirallus aus tralis went extinct on Takaka Hill. The losses continue. 122 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Literature Cited Alroy, J. MS. Non-individualistic Dynamics in Late Quaternary Mammalian Fau nas. Anderson, A. 1989. Prodigious Birds: Moas and Moa-Hunting in New Zealand. 238 pages. Cambridge: Cambridge University Press. 1991. The Chronology of Colonization in New Zealand. Antiquity, 65:765-795. Atkinson, I.A.E., and PR. Millener 1991. An Ornithological Glimpse into New Zealand's Pre-human Past. Acta XX Congressus Internationalis Ornithologici, 1:127-192. Wellington: New Zealand Ornithological Congress Trust Board. Bell, B.D. 1991. Recent Avifaunal Changes and the History of Ornithology in New Zealand. Acta XX Congressus Internationalis Ornithologici, 1: 195-230. Wellington: New Zealand Ornithological Congress Trust Board. Cassels, R. 1984. The Role of Prehistoric Man in the Faunal Extinctions of New Zealand and Other Pacific Islands. In P.S. Martin and R.G. Klein, editors, Quaternary Extinctions: A Prehistoric Revolution, pages 741-767. Tuscon: University of Arizona Press. Fleming, CA. 1979. 77ie Geological History of New Zealand and Its Life. 141 pages. Auckland, New Zealand: Auckland University Press. Graham, R.W. 1985. Diversity and Community Structure of the Late Pleistocene Mam mal Fauna of North America. Acta Zoologica Fennica, 170: 181-192. 1986. Plant-Animal Interactions and Pleistocene Extinctions. In D.K. El liott, editor, Dynamics of Extinction, pages 131-154. New York: John Wiley and Sons, Inc. 1987. Late Quaternary Mammalian Faunas and Paleoenvironments of the Southwestern Plains of the United States. In R.W. Graham, H.A. Semken, Jr., and M.A. Graham, editors, Late Quaternary Mamma lian Biogeography of the Great Plains and Prairies. Scientific Papers (Illinois State Museum), 22:24-86. Springfield: Illinois State Mu seum. 1992. Late Pleistocene Faunal Changes as a Guide to Understanding Ef fects of Greenhouse Warming on the Mammalian Fauna of North America. In R.L. Peters and T.E. Lovejoy, editors, Global Warming and Biological Diversity, pages 76-87. New Haven: Yale Univer sity Press. 1993. Processes of Time-Averaging in the Terrestrial Vertebrate Record. In S.M. Kidwell and A.K. Behrensmeyer, editors, Taphonomic Ap proaches to Time Resolution in Fossil Assemblages. Short Courses in Paleontology, 6:102-124. Knoxville: Department of Geological Sciences, University of Tennessee. Graham, R.W., and E.C. Grimm 1990. Effects of Global Climatic Change on the Patterns of Terrestrial Bi ological Communities. Tree, 5(9):289-292. Graham, R.W., and E.L. Lundelius, Jr. 1984. Coevolutionary Disequilibrium and Pleistocene Extinctions. In P.S. Martin and R.G. Klein, editors, Quaternary Extinctions?A Prehis toric Revolution, pages 223-249. Tuscon: University of Arizona Press. Holdaway, R.N. 1989. New Zealand's Prehuman Avifauna and Its Vulnerability. In M.R. Rudge, editor, Moas, Mammals and Climate in the Geological His tory of New Zealand. New Zealand Journal of Ecology, supplement, 12:11-25. Holdaway, R.N., and T.H. Worthy 1996. Diet and Biology of the Laughing Owl Sceloglaux albifacies (Aves: Strigidae) on Takaka Hill, Nelson, New Zealand. Journal of Zool ogy, London, 239:545-572. King, CM. 1984. Immigrant Killers: Introduced Predators and the Conservation of Birds in New Zealand. 224 pages. Auckland: Oxford University Press. Lundelius, E.L., Jr. 1983. Climatic Implications of Late Pleistocene and Holocene Faunal As sociations in Australia. Alcheringa, 7:125-149. Mabin, M.CG. 1983. Late Otiran Sedimentation and Glacial Chronology in the Warwick Valley, Southeast Nelson. New Zealand Journal of Geology and Geophysics, 26:189-195. Martin, PS., and R.G. Klein, editors 1984. Quaternary Extinctions: A Prehistoric Revolution, x + 892 pages. Tuscon: University of Arizona Press. McGlone, M.S. 1985. Plant Biogeography and the Late Cenozoic History of New Zealand. New Zealand Journal of Botany, 23(4):723-749. 1988. New Zealand. In B. Huntley, and T. Webb III, editors, Vegetation History, pages 557-599. Amsterdam: Kluwer Academic Publishers. McGlone, M.S., A.F. Mark, and D. Bell 1995. Late Pleistocene and Holocene Vegetation History, Central Otago, South Island, New Zealand. Journal of the Royal Society of New Zealand. 25(1): 1-22. Mildenhall, D.C. 1995. Pleistocene Palynology of the Petone and Seaview Drillholes, Pe- tone, Lower Hurt Valley, North Island, New Zealand. Journal of the Royal Society of New Zealand, 25(2):207-262. Millener, PR. 1990. Evolution, Extinction and the Subfossil Record of New Zealand's Avifauna. In B.J. Gill and B.D. Heather, editors, A Flying Start. No tornis, supplement, 37:93-100. Murray, P. 1991. The Pleistocene Megafauna of Australia. In P.V. Rich, J.M. Mon aghan, R.F. Baird, and T.H. Rich, editors, Vertebrate Palaeontology of Australasia, pages 1071-1164. Victoria: Pioneer Design Studio Pty. Ltd., in association with Monash University, Melbourne. Nelson, C.S., P.J. Cooke, C.H. Hendy, and A.M. Cuthbertson 1993. Oceanographic and Climate Changes over the Past 160,000 Years at Deep Sea Drilling Project Site 594 off Southeastern New Zealand, Southwest Pacific Ocean. Paleoceanography, 8(4): 435-458. Soons, J.M. 1979. Late Quaternary Environments in the Central South Island of New Zealand. New Zealand Geographer, 35:16-23. Suggate, R.P. 1965. Late Pleistocene Geology of the Northern Part of the South Island, New Zealand. Bulletin of the New Zealand Geological Survey, new series, 77: 91 pages. 1990. Late Pliocene and Quaternary Glaciations in New Zealand. Quater nary Science Reviews, 9:175-197. Suggate, R.P., and N.T. Moar 1970. Revision of the Chronology of the Late Otira Glacial. New Zealand Journal of Geology and Geophysics, 13:742-746. Trewick, S. 1996. Morphology and Evolution of Two Takahe: Flightless Rails of New Zealand. Journal of Zoology, London, 238:221-237. NUMBER 89 123 Turbott, E.G., convener 1990. Checklist of the Birds of New Zealand and the Ross Dependency, Antarctica. Third edition, 247 pages. Aukland: Random Century in Association with the Ornithological Society of New Zealand, Inc. Worthy, T.H. 1989. Moas of the Subalpine Zone. Notornis, 36:191-196. 1990. An Analysis of the Distribution and Relative Abundance of Moa Species (Aves: Dinornithiformes). New Zealand Journal of Zool ogy, 17:213-241. 1993a. Fossils of Honeycomb Hill. 56 pages. Wellington, New Zealand: Museum of New Zealand Te Papa Tongarewa. 1993b. A Review of Fossil Bird Bones from Loess Deposits in Eastern South Island, New Zealand. Records of the Canterbury Museum, 10(8):95-106. 1997. Quaternary Fossil Fauna of South Canterbury, South Island, New Zealand. Journal of the Royal Society of New Zealand, 27(1): 67-172. Worthy, T.H., and R.N. Holdaway 1993. Quaternary Fossil Faunas from Caves in the Punakaiki Area, West Coast, South Island, New Zealand. Journal of the Royal Society of New Zealand, 23(3): 147-254. 1994a. Quaternary Fossil Faunas from Caves in Takaka Valley and on Takaka Hill, Northwest Nelson, South Island, New Zealand. Jour nal of the Royal Society of New Zealand, 24(3):297-391. 1994b. Scraps from an Owl's Table?Predator Activity as a Significant Taphonomic Process Newly Recognised from New Zealand Fossil Deposits. Alcheringa, 18:229-245. Quaternary Fossil Faunas from Caves on Mt. Cookson, North Can terbury, South Island, New Zealand. Journal of the Royal Society of New Zealand, 25(3):333-370. Quaternary Fossil Faunas, Overlapping Taphonomies, and Palaeo- faunal Reconstruction in North Canterbury, South Island, New Zealand. Journal of the Royal Society of New Zealand, 26(3): 275-361. Worthy, T.H., and D.C Mildenhall 1989. A Late Otiran-Holocene Paleoenvironment Reconstruction Based on Cave Excavations in Northwest Nelson, New Zealand. New Zealand Journal of Geology and Geophysics, 32:243-253. 1995. 1996. The Middle Pleistocene Avifauna of Spinagallo Cave (Sicily, Italy): Preliminary Report Marco Pavia ABSTRACT A preliminary study of the middle Pleistocene birds from Spina gallo Cave (Siracusa, Sicily) shows an avifauna composed of 61 species (28 Passeriformes and 33 non-Passeriformes), including Anseriformes, Falconiformes, Gruiformes, Charadriformes, and Strigiformes. Three extinct taxa, probably new to science, include a large Tyto, a long-legged Athene, and a small species of Corvidae to be described later. Paleoenvironmental reconstruction of the site indicates a temperate climate, like the present or slightly colder. Introduction In 1959 and 1960, many fossil bones were collected from Pleistocene cave deposits in Spinagallo Cave, near Siracusa, southeastern Sicily, Italy (Accordi et al., 1959; Accordi and Colacicchi, 1962) (Figure 1). The fossil association contains bones of mammals, especially dwarf elephants, reptiles, am phibians, and birds. The age, determined by Bada et al. (1991) from amino-acid racemization analysis of mammal bones, is about 500,000 years, or middle Pleistocene. There are no signs of human activities on the bones or in the cave, so the accumu lation is not artificial. The specimens have been stored in the Museum of the Dipartimento di Scienze della Terra (Universita "La Sapienza") di Roma. The Pleistocene vertebrate fauna of mammals, reptiles, and amphibians from Spinagallo has been described by various au thors (Accordi, 1962; Ambrosetti, 1968, 1969; Petronio, 1970; Kotsakis, 1977, 1984; Kotsakis and Petronio, 1980). The east ern part of Sicily, during the middle Pleistocene, was inhabited by two species of dwarf elephants (Ambrosetti, 1968), a giant species of Gliridae (Ambrosetti, 1969; Petronio 1970), and an exctinct lizard (Kotsakis, 1977, 1984; Delfino, pers. comm., 1995). The mammal and reptile faunas seem to indicate that Sicily was isolated during most of the Pleistocene and was colo nized by a typically mainland fauna only in the late Pleistocene. Marco Pavia, Dipartimento di Scienze della Terra, Via Accademia delle Scienze 5, 10123 Torino, Italy. The fossil avifauna consists of almost 1000 bones that have been identified by comparison with recent skeletons in the Mu seo Regionale di Scienze Naturali di Torino, the Regalia Col lection stored in the Institut de Paleontologie Humaine de Par is, and the collections of the Departement des Sciences de la Terre de 1'Universite de Lyon. Systematic List The avifauna of Spinagallo is composed of 61 taxa, which are listed according to the nomenclature of Voous (1973, 1977). Non-passeriformes Passeriformes Geronticus eremita Anser erythropus Branta sp. Anas penelope Anas querquedula Marmaronetta angustirostris Accipiter gent His Accipter nisus Falco tinnunculus Falco cotumbarius Falco subbuteo Falco eleonorae Coturnix coturnix Rallus aquaticus Grus sp. Recurvirostra avosetta Scolopax rusticola Larus minutus Larus ridibundus Columba livia Columba livia/oenas Columba palumbus Streptopelia turtur Cuculus canorus Tyto, species undescribed'' Otus scops cf. Surnia ulula Athene, species undescribedt Asio otus Caprimulgus cf. europaeus Apus apus/pallidus Apus melba Picus viridis Dendrocopos leucotos Calandrella brachydactyla Lullula arborea Hirundo sp. Anthus sp. Prunella modularis Erithacus nubecula Oenanthe cf. hispanica Monticola solitarius Turdus sp. 1 Turdus sp. 2 Sylvia sp. Phylloscopus sibilatrix/ collybita Lanius senator Pica pica Pyrrhocorax graculus Corvidae genus and species indet.''' Sturnus sp. Petronia petronia Fringilla coelebs/montifringilla Serinus sp. Carduelis chloris Carduelis sp. Pyrrhula pyrrhula Coccothraustes coccothraustes Emberiza sp. 1 Emberiza sp. 2 Emberiza sp. 3 Passeriformes indet. 125 126 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?Map of Sicily (shaded); inset shows the position of Spinagallo cave (*). Remarks The Pleistocene avifauna contains two new extinct species that are probably endemic to Sicily: a giant Tyto, similar in size to Tyto robusta (Ballmann, 1973), and a new species of Athene, characterized by having the legs longer than in Athene noctua but shorter than in Athene cretensis (Weesie, 1982). Descrip tions of both are in preparation. Another extinct species, an un determined Corvidae, probably the same as found in the Bale aric Islands (Alcover et al., 1992), was found in Spinagallo and in another cave in Sicily of the same age (Alcover, pers. comm., 1995). Bones of a large crane, similar in size to the liv ing Grus antigone, also were found. The presence of apparently endemic forms, combined with other typical features of insular avifaunas (Alcover et al., 1992), seems to confirm the isolation of Sicily during the mid dle Pleistocene, as previously suggested by the mammalian fauna. One of the most evident characteristics of fossil island avifaunas is the absence of Galliformes, with the exception of Coturnix coturnix (Alcover et al, 1992), which is true of Spina gallo. This is in contrast to mainland cave avifaunas, which are dominated by members of this order. On the Mediterranean is lands, remains of C. coturnix are common, doubtless because of the migratory habits of the species. The absence of partridg es of the genus Alectoris also is typical, although they are now present on Mediterranean islands, probably due to human intro duction, and are very common. The presence of Laridae differs from the normal composition of insular avifaunas (Alcover et al., 1992) but can be explained by the short distance between Sicily and the mainland, where fossil and recent gulls are both recorded. The composition of the avifauna suggests a coastal environ ment with a cliff close to the sea; the same Miocene cliff in which the cave was formed. This physiographic feature sup ported many species, such as Geronticus eremita, Falco ele- onorae, F. tinnunculus, Tyto (species undescribed), Columba livia, Apodidae, and Pyrrhocorax graculus. Inland, on top of the cliff, it is supposed that there was an extension of Mediter ranean forest with large trees and dense undergrowth, appropri ate habitat for Accipiter gentilis, A. nisus, Falco subbuteo, Scolopax rusticola, Strigidae (except the probable vagrant Sur- nia ulula), Columba palumbus, Streptopelia turtur, Caprimul- gus europaeus, all the Picidae, and many Passeriformes. Along the sea, wetland is indicated by the Anseriformes and other waterbirds such as Laridae. The records of Falco columbarius, Caprimulgus europaeus, and many Passeriformes, such as Alaudidae, Anthus sp., Lanius senator, Oenanthe hispanica, Carduelis sp., and the Emberizidae, suggest that open, dry country with scattered bushes also was present. The number of birds of prey in the Spinagallo fauna is high, possibly because many raptors lived in or near the cave. NUMBER 89 127 Literature Cited Accordi, B. 1962. La grotta ad elefanti nani di Spinagallo (Siracusa). Atti della Acca- demia delle Scienze di Ferrara, 37:9-15. Accordi, B., B. Campisi, and R. Colacicchi 1959. Scoperta di un giacimento pleistocenico a elefanti nani e ghiro gi- gante nella grotta di Spinagallo (Siracusa). Atti della Accademia Gioenia Scienze Naturali di Catania, 12:167-182. Accordi, B., and R. Colacicchi 1962. Excavations in the Pigmy Elephants Cave of Spinagallo (Siracusa). Geologica Romana, 1:217-229. Alcover J.A., F. Florit, C. Mourer-Chauvire, and P.D.M. Weesie 1992. The Avifaunas of the Isolated Mediterranean Islands during the Middle and Late Pleistocene. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:273-283. Ambrosetti, P. 1968. The Pleistocene Dwarf Elephants of Spinagallo (Siracusa, South Eastern Sicily). Geologica Romana, 7:277-398. 1969. Rappresentanti del genere Leithia nel Pleistocene della Sicilia. Memorie del Museo Civico di Storia Naturale di Verona, 75-80. Bada, J.L., G. Belluomini, L. Bonfiglio, M. Branca, E. Burgio, and L. Delitala 1991. Isoleucine Epimerisation Ages of Quaternary Mammals from Sic ily. // Quaternario, 4(la):49-54. Ballmann, P. 1973. Fossile Vogel aus dem Neogen der Halbinsel Gargano (Italien). Scripta Geologica, 17:1-75. Kotsakis, T. 1977. I resti di anfibi e rettili pleistocenici della grotta di Spinagallo (Sir acusa, Sicilia). Geologica Romana, 16:211-229. 1984. Crocidura esui n.sp. (Soricidae, Insectivora) du Pleistocene de Spinagallo (Sicile orientale, Italie). Geologica Romana, 23:51-64. Kotsakis, T, and C Petronio 1980. I chirotteri del Pleistocene superiore della grotta di Spinagallo (Si racusa, Sicilia). Bollettino del Servizio Geologico d'ltalia, 101: 49-76. Petronio, C. 1970. I roditori pleistocenici della grotta di Spinagallo (Siracusa). Geo logica Romana, 9:149-194. Voous, K.H.F. 1973. List of Recent Holarctic Bird Species, Non-Passerines. Ibis, 115: 612-638. 1977. List of Recent Holarctic Bird Species, Passerines. Ibis, 119: 223-250, 376-406. Weesie, P.D.M. 1982. A Pleistocene Endemic Island Form within the Genus Athene: Athene cretensis n.sp. (Aves, Strigiformes) from Crete. Proceed ings of the Koninklijke Nederlandse Akademie van Wetenschap- pen, section B, 85(3):323-336. Birds in the Economy and Culture of Early Iron Age Inhabitants of Ust' Poluisk, Lower Ob' River, Northwestern Siberia Olga R. Potapova and Andrei V. Panteleyev ABSTRACT The archaeological settlement of Ust' Poluisk, located in the lower Ob' River basin in northwestern Siberia (66?33'N, 66?35'E), yielded a rich vertebrate fauna with a high ratio of bird remains (1996 bones). Cultural remains were deposited over sev eral centuries and were dated by association with archaeological artifacts to 400-100 BC. Thirty-nine species were identified in the bird assemblage. Among these species, 10 are represented by rare breeding and rare vagrant birds, indicating a somewhat warmer climate at the Ob' River mouth than at present. The remains of Golden Eagles and White-tailed Eagles excavated from sacrificial areas of the settlement are of special interest. These findings indi cate special cultural attention to and attitude toward eagles, which may have been kept in captivity. Based on bird remains, the site has provided the earliest evidence of eagle worship in Siberia. Introduction Humans have been dependent upon nature throughout their history. The economic lifeways of ancient peoples were mainly determined by natural conditions. The best evidence of this is found in the north, where agriculture was absent, the possibili ties of plant gathering were limited, and subsistence activities were based primarily on hunting and fishing. In northern lati tudes, fowling was an important means of survival. Besides being a source of sustenance and of feathers (for fletching ar rows and myriad other uses), birds played a significant role in the cultures of many peoples. Cultural roles included cult cere monies, decorations, and subjects in tales, legends, and tradi tions. The abundant remains of birds from Ust' Poluisk settle ment in northwestern Asia (Figure 1) provides data on the Olga R. Potapova and Andrei V. Panteleyev, Zoological Institute, Russian Academy of Sciences, St. Petersburg, 199034 Russia. economic, subsistence, and cultural systems of its early Iron Age human inhabitants. MATERIALS AND METHODS.?The Ust' Poluisk archaeologi cal site was discovered in 1932. Bones examined in this study were collected during excavations conducted in 1935 and 1936 by V.S. Adrianov (Museum of Anthropology and Ethnography (MAE)), St. Petersburg, Russia. The size of the settlement was estimated to be about 4000 m2. It was surrounded by a kremlin wall and a trench. About 10% of the site (410 m2) was excavat ed. Thirty-six thousand artifacts and bones were recovered (Adrianov, MSa). For unknown reasons, the excavations were not completed, and almost all documents from the 1936 exca vations were lost. The cultural layer of the site is generally 20-30 cm thick, but it widens to 40-50 cm thick in hollows, indicating its homoge neity and the long duration of site occupation (Moshinskaya, 1953). Artifacts were dated to 400-100 BC based on artifacts and tools of the Anan'inskaya and Tagarskaya cultures (Cher- netsov, 1953). Discovery of metal knives at the site indicate an early Iron Age settlement (T.A. Popova, pers. comm., 1997). Bird bones were apparently excavated from trench number 5, where remains of two or three dwellings and a sacrifice area were found. Also located were a hearth with a pile of dog skulls, reindeer bone fragments, isolated human bones, and nu merous ceramic, bone, and some bronze artifacts (Moshin skaya, 1953). Other faunal remains included squirrel, beaver, hare, fox, arctic fox, sable, moose, pinnipeds, and some large fish (Adrianov, MSc). Among nonavian, partly identified bone remains, reindeer were predominant at the site (Kosintsev, 1997). Beavers were represented by 23 specimens, among which were 20 young animals and one juvenile (O.R. Potapo va, pers. obs., 1997). Descriptions of excavation methods are lacking in Adri- anov's 1935 report (MSa, MSb, MSc). The authors believe the deposits were excavated using shovels and knives but were not screened. The deposits were, however, most likely subjected to 129 130 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?Map showing location of Salekhard and Ust' Poluisk settlement (diamond). thorough visual examination after removal from the excavation units. This excavation methodology was in general use by Rus sian archaeologists in the 1930s (D.N. Praslov, Institute of Ma terial Culture, St. Petersburg, pers. comm., 1996). The authors believe excavation methods were quite thorough, judging from the small size of many of the artifacts collected and of some bird bones (quadratum, premaxilla, and others). The lack of small passerine birds at the site may be due to taphonomic conditions, similar to the other open-air sites on the Russian Plain (i.e., Kostenki) (D.N. Praslov, pers. comm., 1996) and North Caucasus (i.e., Ilskaya 2) (O.R. Potapova, pers. obs., 1997). The deposits from these sites were screened but yielded no small passerine bird remains and only extremely rare fossilized rodents. The hunting preferences of aboriginal humans, oriented to prey larger than passerine birds, could be another reason. There is, however, the possibility that some of the smallest bone material was not located or was lost. All bird-bone collections from the site are deposited in the Zoological Institute, Ornithological Section, St. Petersburg, Russia. Bird-bone artifacts are deposited in the MAE. ACKNOWLEDGMENTS.?We express our appreciation to T.A. Popova, Gennadii Khlopachev, and L.N. Gizha (MAE Archae ology Department) for their help and for making bird-bone ar tifacts available for the study. We also thank M. McGrady, Royal Society for the Preservation of Birds (RSPB, Scotland), and Eugene R. Potapov, Institute of Biological Problems of the North (IBPN, Magadan, Russia), for the translation and pre liminary editing of the first manuscript draft, and we thank Anne Karin Hufthammer, University of Bergen, Norway, for valuable discussions of Holocene assemblages of Norway. We are very grateful for the essential suggestions, reviews, and critical editing by S. Olson, C. Mourer-Chauvire, and D. Ser- jeantson. We express special thanks to Lance W Rom (U.S. Department of Agriculture, Natural Resources Conservation Service) for the substantial contribution he made in editorial and stylistic revisions and essential additions to the final manuscript. Discussion Thirty-nine bird species were indentified from excavations at Ust' Poluisk settlement (Table 1). Many of these species are now restricted to more southern regions and only occasionally visit the lower Ob' River as vagrants or, rarely, as breeders, this being the northern boundary of their current range. It is possi ble that 2000 years ago the breeding ranges of these birds ex tended farther to the north, because the paleobotanical record suggests that at that time northwestern Siberia had a warmer climate than at present (Volkova et al., 1989). Two thousand years ago the timber line was north of Salekhard, and there were pine forests at the mouth of the Polui River (Moshin- skaya, 1953). The presence of a forest ecosystem during the time the site was occupied is supported by relatively numerous archaeological findings of birch bark and of remains of forest- dwelling animals, such as squirrels, beavers, sable, and moose (Kosinstsev, 1997), at the site. The bird-species assemblage includes individuals from seven groups. More than 92% are grouse and waterfowl. Diumal birds of prey and owls represent 6.1% (Table 2). Birds from these groups were probably hunted for food or may have played a role in the cultural traditions of the population. Passe rine birds are represented by corvids, which might have been attracted to the settlement by garbage. At Ust' Poluisk, grouse account for 51.4% of all bones, with most of these belonging to Willow Ptarmigans {Lagopus lago- pus (Linnaeus)). This species is numerous in Paleolithic-age sites of the northern and middle Urals (Potapova, 1990, 1991) and in forest and forest-steppe zone sites on the Russian Plain, such as Kostenki, Novgorod-Severskii, Mezin (Zubareva, 1950), and Afontova Gora-3 in southern Siberia (Tugarinov, 1932). In Holocene archaeological sites, the remains of Willow Ptarmigans are rare. At Mayak 2, an early Bronze Age site on the Kola Peninsula, the remains of Willow Ptarmigans com- NUMBER 89 131 TABLE I.?Bird species from Ust' Poluisk (NISP=number of bones; MNI=minimum number of individuals; *=rare vagrant or rare breeding spe cies at the Ob' River mouth). Taxon Gavia stellata Gavia arctica Podiceps cristatus Cygnus cygnus Cygnus bewickii Anser cf. albifrons Anser cf. fabalis Branta spp. Anas platyrhynchos Anas crecca Anas penelope Anas acuta Anas querquedula Anas clypeata Anas spp. Aythya fuligula Aythya marila Melanitta fusca Melanitta nigra Clangula hyemalis Bucephala clangula Mergus albellus Mergus merganser Anatidae indeterminate Haliaeetus albicilla Accipiter gentilis Buteo lagopus Aquila chrysaetos Circus cyaneus Accipitridae indeterminate Falco peregrinus Lagopus mutus Lagopus lagopus Lagopus indeterminate Tetrao urogallus Lyrurus tetrix Grus grus Grus leucogeranus Larus argentatus Common name Red-throated Loon Arctic Loon * Great Crested Grebe Whooper Swan Bewick's Swan Greater White-fronted Goose Bean Goose small geese Mallard Green-winged Teal Eurasian Wigeon Northern Pintail Garganey Northern Shoveler teal Tufted Duck Scaup White-winged Scoter Black Scoter Oldsquaw Common Goldeneye * Smew * Common Merganser ducks White-tailed Eagle Northern Goshawk * Rough-legged Hawk * Golden Eagle Northern Harrier kites, hawks, eagles * Peregrine Falcon Rock Ptarmigan Willow Ptarmigan ptarmigan Capercaillie Black Grouse * Common Crane Siberian White Crane Herring Gull Charadriiformes indeterminate shorebirds Bubo bubo Nyctea scandiaca Strix nebuloza Corvus cornix Corvus corax Aves indeterminate TOTAL * Eagle Owl Snowy Owl * Great Gray Owl Hooded Crow * Common Raven birds NISP 25 48 8 28 10 132 88 3 12 50 40 24 6 12 6 4 1 1 2 14 7 1 1 46 143 4 1 28 2 1 2 134 653 318 16 1 1 6 65 1 4 12 1 11 11 12 1996 MNI 7 10 2 5 2 41 10 1 5 23 10 4 2 3 2 4 1 1 1 5 2 1 1 18 10 1 1 7 1 1 1 36 144 30 6 1 1 1 10 1 1 2 1 3 2 422 TABLE 2.?Bird remains analyzed by groups from Ust' Poluisk (MNI=minimum number of individuals). Groups Loons, grebes, shorebirds, gulls Ducks, geese, swans Diurnal raptors Grouse Cranes Owls Crows TOTAL Number of 5 19 6 4 2 3 2 41 species MNI 30 142 22 217 2 4 5 422 MNI% 7.1 33.7 5.2 51.4 0.5 0.9 1.2 100 prise only 1% of all bird bones (Potapova and Sablin, 1994; O.R. Potapova, pers. obs., 1997). In the northern Urals, Willow Ptarmigan bones were reported only from the Kaninskaya cave Bronze-Iron Age site on the upper Pechora River (Kuzmina, 1971). All skeletal elements of Willow Ptarmigans and Rock Ptar migans {Lagopus mutus (Montin)) were found at Ust' Poluisk, with humeri and femora being the most common (see Figure 2A). The relative abundance of skeletal elements is similar to that of skeletal elements found at Abri Fontales, Ebbou, and La Colombiere archaeological sites in France (Mourer-Chauvire, 1983). The numbers of skulls, sterna, and pelvises found at the Ust' Poluisk site suggest that the birds were delivered to the site intact, with processing occurring at the site. All the bones are very well preserved, with unbroken bones constituting al most 88% of the total. The majority of bones belong to adults, with only 1.5% belonging to young individuals, based on spongy tissue at the ends of the long bones. Many tibiae and femora have distinct tooth marks (13% of all Lagopus bones) that appear as small holes and dents on their ends (these fall into three size classes: 1.6 x 3.0 mm; 2.0 x 3.7 mm; 2.7 x 2.9 mm). Some of the bones (0.4%) have signs of cuts on the shafts (femora) and have cuts through the entire articular ends (humeri). Willow Ptarmigans probably attracted prehistoric hunters be cause their seasonal abundance and behavior patterns made them relatively easy to obtain. The hunting process required no special equipment, such as bows or arrows, because nooses, traps, or nets could be used successfully (Silantyev, 1898; Ko- losov and Shibanov, 1957). Not long ago these methods were still widely used in the Russian north. In the nineteenth century Silantyev (1898:367) wrote: "Local hunters get grouse without guns." Catching grouse with nets in the spring was used in the tundra of the Lena-Khatanga depression (Romanov, 1934), and net hunting is still widely applied in the Lower Kolyma throughout the year (E.R. Potapov, pers. comm., 1997). In North America, it is still possible to catch large numbers of grouse in the Great Plains using nets or clubs; nets were used by nineteenth century Shoshone Indians during rabbit drives, and fiber nets dating from thousands of years ago have been lo cated in archaeological sites in the Great Plains and Great Basin (L.W Rom, pers. comm., 1997). The technology of pro ducing nets from nettles or willow bast was known to all fish ing peoples from the Neolithic-Bronze Age (Kosarev, 1987a, 1991; Krushanov, 1989). It is possible that the same nets were employed for fishing and catching molting geese in summer and for hunting grouse in other seasons. Although not required, it also is possible that some special equipment was used for hunting grouse. Besides the more productive hunting in the fall, winter, and spring, inhabitants could catch grouse during other seasons by beating or by catching birds by hand in remote areas, where birds were not afraid of humans (Potapov, 1985). Except during the breeding period, Willow Ptarmigans live in flocks. On the Lower Ob' River, the young form small flocks in August, and flocks of up to 100 birds have been observed in 132 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Cranium Mandibula Scapula Clavicula Coracoideum Sternum Humerus Ulna wzm 3 WA'AVJ JWJVJW} Radius Carpometacarpus Vertebrae Pelvis Femur Tibiotarsus Tarsometatarsus jfc&j&r-j Phalanx 1 ifn cog t ? E Lagopus lagopus and L. mutus KflBMCWWH StfWWBSB^iHIWWWBm .?:???????:?, 10 15 20 25 30 35 40 Cranium Mandibula Scapula Clavicula Coracoideum Sternum Humerus Ulna Radius Carpometacarpus Vertebrae Pelvis Femur Tibiotarsus Tarsometatarsus Phalanx B ffenlr niiiniiii llllllllllllllmii iiiiiiiiiiiiiiiiiii iiniiiiiiiiiiiiiii IIUIIIIIllllllllll IIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIII DIM IIIIIIIIIIIIIIIIIII IIIIIIIII ? Cygnus cygnus M Cygnus bewickii IIIIIIIIIIIIIIIIIII II IIIIIIIIIIIIIIIIIII mum IIIIIIIIIIIIIIIIIII IIIIIIIII 11 ITTTITTTOTTTI mi mill mmm "iiniii IIIIIIIIIIIIIIIIIII iiiiiiiiiiiinmii IIIIIIIIIIIIIIIIIII 10 15 20 25 30 FIGURE 2.?Diagram of relative representation of main skeletal elements for some bird species from the Ust' Poluisk settlement. The column for "cranium" includes maxillare. A, Ptarmigans (Lagopus); B, Bewick's and Whooper Swan (Cygnus); c, Eurasian Wigeon (Anas penelope), Green-winged Teal (Anas crecca), Mallard (Anas platyrhynchos), and Northern Pintail (Anas acuta). Cranium Mandibula Scapula Clavicula Coracoideum Sternum Humerus Ulna Radius Carpometacarpus Vertebrae Pelvis Femur Tibiotarsus Tarsometatarsus Phalanx sssssssnggg mam. 1MWYVYH inJm \.mm.*.v*\v^t.*.VMlv .MMII S Anas platyrhynchos ? Anas crecca ED Anas penelope O Anas acuta 10 20 30 40 50 60 70 80 90 October (Boikov, 1965). In late autumn Willow Ptarmigans start to migrate toward the timberline, at times reaching such a high density that they are commercially hunted (Potapov, 1985). Thus, late autumn through spring might have been the main ptarmigan-hunting period for the ancient Polui dwellers. The bones of Capercaillie {Tetrao urogallus Linnaeus) and Black Grouse {Lyrurus tetrix (Linnaeus)) found at the Ust' Poluisk site are from adults of both sexes. As in many archaeo logical sites on the northern Russian Plain, their remains are much less numerous than those of Lagopus. Geese and ducks also were important groups at the Ust' Poluisk site, where their remains can be easily explained by the presence of the Polui River. Waterfowl could have been hunted during both the breeding and the migration seasons. Of the waterfowl, it appears that the hunters from ancient Ust' Poluisk preferred teals {Anas spp.), Greater White-fronted NUMBER 89 133 {Anser albifrons (Scopoli)) and Bean geese {Anser fabalis (Latham)), and Eurasian Wigeons {Anas penelope Linnaeus). These species each make up 8% to 34% of all individuals. Geese and swans were presumably hunted during the very short molting season (mid-July) or during migration. Hunting molting geese was a common practice for local tribes at the Lena River mouth (Kosarev, 1987b) and the Kolyma River (Wrangel, 1848) in the eighteenth century. The relative numbers of each of the skeletal elements of wa terfowl are similar to those for Willow Ptarmigans (Figure 2B,C), although few or no sterna of swans or of Greater White- fronted or Bean Geese were found. The sternum, as would be the case for other inedible parts of the skeleton, was presum ably used for a variety of purposes. The 17 spoons examined from the site were made of bird sterna belonging to loons {Ga via sp., 11 spoons), a goose (1) a Greater White-fronted Goose (1), eagles {Aquila chrysaetos (Linnaeus) or Haliaeetus albicil- la (Linnaeus)) (2), a Mallard {Anas platyrhynchos Linnaeus) (1), and an Oldsquaw {Clangula hyemalis (Linnaeus)) (1). Moshinskaya (1953) considered the spoons made of large-wa terfowl sterna to be the most archaic elements among the spoon-like tools that might have been used by inhabitants in rit uals at the site. Production of various types of spoons from bird sterna has been well documented in recent times. The Mansi used bone spoons during sacrificial and burial rituals, and in Nenets folklore, the main hero Pornene, half woman and half bear, used the sternum of a swan as a spoon (Moshinskaya, 1953). Northern Pintails {Anas acuta Linnaeus) and Green-winged Teal {Anas crecca Linnaeus) are represented primarily by wings. Northern Pintails are rare in the bone remains from this site, although they have been very numerous in the lower Ob' region in recent times. People of northern Russia still use duck wings with brightly colored feathers as a decoration for cloth ing and housewares. Today, bird wings are used for applying cooking oil when preparing pancakes or other foods in rural houses and in Russia's urban areas. The people of the lower Kolyma region commonly use wings of geese as rubbish brush es (brooms) in their cabins (E.R. Potapov, pers. comm., 1997). In the past, wings of Northern Shovelers {Anas clypeata Lin naeus), Northern Pintails, and Green-winged Teals were proba bly used for similar purposes. Loons and gulls, which are migratory species, can be ob tained in the lower Ob' River region between May and August (Flint, 1988; Yudin and Firsova, 1988). The relatively few re mains of loons and Herring Gulls {Larus argentatus Pontoppi- dan) at Ust'Polusik, species common in the lakes and rivers of the tundra-forest zone today, could suggest that these species were rarely hunted and were less desirable. It also could indi cate that they were less common 2000 years ago. The skeletal- element representation of these species is similar to that of the Willow Ptarmigan and suggests full utilization by ancient hunt ers (Figure 3A). The skins of loons and Great Crested Grebes {Podiceps cristatus (Linnaeus)) now are greatly valued for their use in clothing (Kolosov et al., 1975), and these species might have been hunted in the site area for the same purpose. Loons could have been used for fat, which generally is of great value for northern peoples. Between 1000 BC and medieval times, inhabitants of the Udal and Buckquoy sites, in the Outer Hebrides and Orkney Islands, respectively, hunted loons exclu sively for fat (Serjeantson, 1988); however, gull remains at Me- solithic and early medieval sites in northern Scandinavia and Scotland are rare (A.K. Hufthammer, pers. comm., 1991; Ser jeantson, 1988). Therefore, it is probable that rather than being less desirable species, loons and gulls were less common in the Ust' Poluisk area when the site was occupied. The remains of owls at Ust' Poluisk belong to three large species: the Snowy Owl {Nyctea scandiaca (Linnaeus)), which is the most abundant; the Eagle Owl {Bubo bubo (Linnaeus)); and the Great Gray Owl {Strix nebulosa Forster). Among the bird images on tools and kitchenware found at the site, there was one bronze, stylized owl (Adrianov, MSc). Because of the frequency of their remains, it appears that owls were specifical ly hunted. Derugin (1898) reported that the local people hunted owls in autumn and winter, when owls accumulated large quan tities of fat and were considered a delicacy. Their wings were subsequently used as fans against mosquitoes. The Samoeds (Nentsy), of the Yamal Peninsula, Russia, hunted owls using nooses fixed on high poles (Shukhov, 1915) and using traps at nests (Zhitkov, 1912). Indians of the North American Great Plains treated owls as a superstitious power and sometimes even as a medicine. They kept owls in captivity for soothsaying and used their feathers (especially those of the Great Horned Owl, Bubo virginianus (Gmelin)) for ceremonies and dances. Some tribes utilized certain species of owls for food, and there is evidence that the Arikara Indians ate Great Horned Owls (Parmalee, 1977a). At the sacrificial area of the Ust' Poluisk site, a number of bones of Golden and White-tailed Eagles were found. Golden Eagles {Aquila chrysaetos) were represented by disproportion ately high numbers of skulls, one of which has had the upper part completely cut off (Figure 4c,D). White-tailed Eagles {Haliaeetus albicilla) were represented by full sets of bones (Figures 3B,C), and, unlike Golden Eagles, they were buried in tact. The different proportions of skeletal elements suggest dif ferent uses for these species. Among the bones of White-tailed Eagles, one tibia and one ulna (out of a minimum of 10 individuals found at the site) had been broken and had grown back together (Figure 4A,B). Birds with healed broken bones are rare in the wild. In the remains from the Ust' Poluisk site there were only two other cases of knitted fractures found: one femur of a Willow Ptarmigan (0.5% of individuals) and one fibula of a Greater White-fronted Goose (2.0% of individuals). Among several thousand Pleis tocene bones from the Binagady asphalts (eastern Caucasus), only two were found with knitted fractures: one mallard and one Steppe Eagle {Aquila rapax Temminck) (Burchak-Abram- ovich, 1949, 1968). In ducks, healed fractures may be found 134 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Cranium Mandibula Scapula Clavicula Coracoideum a Sternum B5ZS5^ E3 Larus argentatus ? Gavia stellata Humerus fT ' ' 1 f ~ -" Ulna Radius V///////////7A SSSSSS///SJ///S//S////S//S/S/SS///S/S///SS/SS/ Carpometacarpus Vertebrae Pelvis Femur Tibiotarsus gggjL^ Tarsometatarsus f&w* Phalanx sssssm Cranium Mandibula Scapula Clavicula Coracoideum Sternum Humerus Ulna Radius Carpometacarpus Vertebrae Pelvis Femur I 1 V//////A ri H E3 Aquila chrysaetos '////////////A 2222a MMW/////////////A (^^W^^W^^^J^ Tibiotarsus V//////A Tarsometatarsus '////MA Phalanx 0 10 15 20 25 30 35 40 45 10 15 20 25 30 FIGURE 3.?Diagram of relative representation of main skeletal elements for some bird species from the Ust' Poluisk settlement. The column for "cranium" includes maxillare. A, Red-throated Loon (Gavia stellata) and Herring Gull (Larus argentatus); B, Golden Eagle (Aquila chrysaetos); C, White-tailed Eagle (Hali aeetus albicilla). Cranium Mandibula Scapula Clavicula Coracoideum Sternum Humerus ? Ulna ? Radius ? Carpometacarpus Vertebrae Pelvis Femur Tibiotarsus Tarsometatarsus Phalanx iiiiiiiiiiiiiiiiiiiiiiiniiiiiinii i iiiiiiniiiiiiiiiu iiiiiiiiiiiiiiiiiiiiiii mmnniMi mnnnnmnnnni nnnmifflfflimii i ?111111 nnnnimi] IIIIIIIIIIIIIIIII inn i llllllllllllllllllllll llllllllllllllllllllll llllllllllllllllllllll llllllllllllllllllllll llllllllllllllllllllll llllllllllllllllllllll IIIIIIIIIIIIIIIIIIIIIII mil IIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII nnnDii IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII ID nunnmni IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII n ED Haliaeetus albicilla DU IIIIIIIIIIIIIIIIIIIIIII ?11111 nnuniii IIIIIIIIIIIIIIIIIIIIIII gunnnnii IIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII mi iiiii'i IIIIIIIII iiiimiiiiiiiniiiii IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIII mini DDHIID IIIIIIIIIIIIIIIIIIIIIII im mi 10 12 14 more often (13% of 256 wild duck skeletons examined by Tie- mier, 1941). A bird of prey with a broken leg or wing has very little chance of survival in the wild. It is possible that eagles, at least White-tailed Eagles, were kept captive at the settlement. Brothwell (1993:37) noted that "excluding injury received in the wild, birds are most likely to display evidence of trauma as a result of hunting, catching, keeping, or handling." Interest ingly, among 145 skeletons of captive macaws {Ara spp.), birds of ritual and trade value for Pueblo Indians found in archaeo logical sites in Arizona and New Mexico, 13% have healed bones (Brothwell, 1993). Furthermore, one ulna from a Mexi can macaw found at an archaeological site in Arizona (Broth well, 1993, fig. 2B) has a similar pattern of trauma as the eagle ulna from the Ust' Poluisk site. Keeping captive birds, including birds of prey, was wide spread in various groups of people. Eagles taken from nests were kept by the Ayny, Selkups, and Kets (Sokolova, 1972; Kosarev, 1981, 1991). The Ayny kept eagles in cages for sacri fices (Sokolova, 1972). Mikhail Litvin, traveling in southwest ern Russia during the second half of the sixteenth century, NUMBER 89 135 FIGURE 4.?Bones of the White-tailed Eagle (Haliaeetus albicilla). a, broken and healed distal part of ulna, b, broken and healed tibiotarsus, c, lateral and d, ventral views of the upper part of a cranium showing clear evidence of having been cut. (Scale=3 cm.) noted that eagles were kept to provide feathers for arrows (Aristov, 1866). In some California Indian tribes, birds of prey were used as decoys for hunting eagles; this also was a ritual event among the Great Plains Indians (Parmalee, 1977b). Re mains of Golden and Bald Eagles {Haliaeetus leucocephalus (Linnaeus)) were discovered in 60% of 51 Arikara tribe sites in South Dakota and constituted 9% of the bone remains from these sites (Parmalee, 1977b). Golden and Bald Eagles, along with other birds of prey, still play important roles in Great Plains Indian material and spiritu al culture, as they have in the past. They were, and are, repre sented in rock art, ledger art, and all types of decoration. They form integral components of legends and visions. Archaeologi cal sites specifically devoted to capturing eagles are known throughout the North American Great Plains (L.W Rom, pers. comm., 1997). Many Siberian people worshipped eagles, and some of them call March the month of the Eagle. Eagles were associated with the sun god, were equated with the sun god, or were the sun's owner, or creator. The eagle was the supreme god, the benefac tor of individuals, peoples, or clans, or a bird of fortune (Stern berg, 1925:718). In the Urges tribe of the Ob' River, the totem- ic cult of the eagle was linked to the image of the soul-bird (the fourth soul of a person), which dwelled in the hair (Chernetsov, 1959; Kosarev, 1981). Sculptural images of eagles are very common on kitchenware and cult weapons from the Ust' Polu isk site. They are engraved on combs, buckles, suspenders, spoons, klevtsy (ritual axes), and knives (Moshinskaya, 1953; Chernetsov, 1953). Among 12 bone carvings of birds found at the site in 1935-1936, nine were eagles. Six of the carved ea gles, on combs and bone tools, are depicted sitting on and pecking either a moose head or the head of another stylized bird having a heavy beak. Two other artifacts have stylized pecking eagles, and one has a stylized eagle with outstretched wings (Adrianov, MSb, figs. 2, 3, 18, 51, 70, 176, 177, 190, 260). Chernetsov (1953) believed that forest hunters of the Ob' River basin, who practiced the cult of moose or bear, later adapted the eagle cult from the south-steppe Skiph-Sarmatian tribes who inhabited the steppe zone north to the Kama River basin. The bird from the sun, "Kars" (eagle), was the most impor tant part of worship dedicated to the "Upper" world. This wor ship is very ancient and is believed to have originated in India or southern Iran (Sternberg, 1925; Chernetsov, 1947). Kars, from heaven, was believed to be seated in a sacred tree, where the sun and the moon grew. In Siberia this tree was either a birch or a larch (Sternberg, 1925). Chernetsov (1959) observed the initiation procedure in the clan of the Winged Old Man (Eagle) of the Urges in the Ob' River region. According to his account, young men that had reached the age of initiation walked to a special, sacred place where they climbed a sacred tree, home of the clan's "winged" ancestor. In the twentieth century the Urges from Ob' perform a similar ritual, but without a real bird in the tree; however, per haps in their shamanistic past there was a live eagle in the tree. The eagle, as a totem bird, was considered untouchable by many Siberian peoples. Yakuts buried dead eagles, and a com munity member who killed an eagle by mistake was expected to roast it on a fire and eat all but its head (Sternberg, 1925:723). This possibly explains why there were so many skulls of eagles at the Ust' Poluisk site. Conclusions The species composition of the bird-fauna remains collected at the Ust' Poluisk settlement indicates more favorable envi ronmental conditions in that area and a warmer climate at the time of deposition, ca. 400-100 BC, than at present. This con clusion is supported by findings at the site of remains of forest- animal species with ranges that now stop at the timberline, which is south of the site. Avifaunal remains indicate that set tlement inhabitants hunted fatty and/or meaty birds that could have been obtained easily, like grouse and waterfowl, as well as birds of ritual importance. The latter included eagles, which were hunted or which might have been kept alive at the settle ment, as evidenced by healed, broken eagle bones. Additional 136 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY evidence of an eagle cult at the site comes from the numerous bone artifacts with engravings or carvings of eagles. Interest ingly, at the place of sacrifice there also were found about 40 broken skulls of dogs that were killed for a ceremonial purpose (Adrianov, 1936; Moshinskaya, 1953). The eagle cult still ex ists and is practiced by local tribes today (Sternberg, 1925), and the bone remains of eagles reported herein provide the earliest evidence of eagle worship in the Siberian region. Literature Cited Adrianov, VS. MSa. Materialy Nizhne-Obskoi arkheologicheskoi ekspeditsii IE AN SSSR v 1935 g; Raskopki u ustya reki Poluya; Dnevnik raskopok, zamechaniya, karty (k kollektsii MAE #5331) [Materials of the Lower Ob' Archaeological Expedition of the Institute of Ethnogra phy of the Academy of Sciences (IEAS) of the USSR in 1935; Exca vations of the Settlement in the Polui River Mouth; Excavation Diary, Notes, Maps (MAE collection number 5331)]. Unpublished manuscript, MAE archive number 726, K-I, list 2. [In Russian.] MSb. Materialy Nizhne-Obskoi arkheologicheskoi ekspeditsii IE AN SSSR v 1935-1936 g; Raskopki u ustya reki Poluya; Fotografii, karty, zapisi (k kollektsiyam MAE #5331 i 5455) [Materials of the Lower Ob' Archaeological Expedition of IEAS of the USSR in 1935-1936; Excavations at Polui River Mouth; Excavation Photo graphs, Maps, Notes (MAE collection numbers 5331, 5455)]. Un published manuscript, MAE archive numbers 728, 729, K-I, list 2. [In Russian.] MSc. Raskopki poseleniya v ustiye reki Polui. Orudiya okhoty, truda i veshchi razlichnogo naznacheniya iz kostei zhivotnykh i vsevozmozhnye izdeliya iz bronzy, zheleza, kamnya i keramiki; G. Salekhard, ustie reki Poluya [Excavations of the Settlement in the Polui River Mouth: Hunting and Trade Tools, Animal Bones and Other Different Bronze, Iron, Stone and Ceramic Artifacts; City of Salekhard, Polui River Mouth]. 394 pages. [Unpublished manu script, Department of Archaeology, MAE inventory collection de scription number 5331.] [In Russian.] Aristov, N. 1866. Promyshlennost' drevnei Rusi [Industry of Ancient Russia]. 324 pages. St. Petersburg. [In Russian.] Boikov, V.N. 1965. Materialy po fenologii ptits severnoi lesotundry (nizoviya r. Poluya) [Materials on Bird Phenology in the North Tundra Forest Zone, Polyi River Mouth]. Trudy Instituta Biologii, 38:111-140. [In Rus sian.] Brothwell, Don 1993. Avian Osteopathology and Its Evaluation. Archaeofauna, 2:33?43. Burchak-Abramovich, N.I. 1949. Iskopaemaya utka-invalid v binagadinskoi orniofaune [Fossil Dis abled Duck in Binagady Avifauna]. Doklady Academii Nauk Az- erbaidzhanskoi SSR.l .212-216. [In Russian.] 1968. K posnaniu yavlenii izmenchivosti u pleistotsenovykh binagadin- skikh ptits [Study of Variability of Pleistocene Binagady Birds]. In V.I. Strelkovsky, editor, Obshcie Voprosy Evolutsionnoi Paleobi- ologii [General Questions of Evolutionary Paleobiology], 4:3-96. Tbilsi: Metsniereba. [In Russian.] Chernetsov, V.N. 1947. K voprosu o proiskhozhdenii vostochnogo serebra v Priobye [On the Problem of Origin of the East Silver in the Ob' Region]. Trudy Insti tuta Etnografii, 1:113-134. [In Russian.] 1953. Bronza Ust-poluiskogo vremeni [Bronze Age in Ust' Polui Time]. Materialy i Issledovanya po Arkheologii SSSR, 35:221-241. [In Russian.] 1959. Predstavlenie o dushe u obskikh ugrov [The Idea of Soul in Ob' Urges]. Trudy Instituta Etnografii, 51:114-156. [In Russian.] Derugin, K.M. 1898. Puteshestvie v dolinu srednego i nizhnego techeniya reki Obi i fauna etoi oblasti [The Trip to the Middle and Lower Basin of Ob' River and the Fauna of the Area]. Trudy Imperatorskogo Sankt-Peterburg- skogo Obshestva Estestvoispytatelei, Otdel Zoolologii i Physiologii, 29(2):47-140. [In Russian.] Flint, V.E. 1988. Semeistvo Pomornikovye [Family Stercorariidae]. In V.D. Il'ichev and V.A. Zubakin, editors, Ptitsy SSSR: Chaikovye [Birds of the USSR: Suborder Lari], pages 11-38. Moscow: Nauka. [In Russian.] Kolosov, A.M., N.P. Lavrov, and A.V. Mikheyev 1975. Biologia promyslovo-okhotnichikh ptits SSSR [Biology of Game Birds of the USSR]. 320 pages. Moscow: Vysshaya Shkola. [In Rus sian.] Kolosov, M.R, and S.V. Shibanov 1957. Borovaya dich: eye promysel i zagotovka [Pine Wood Game: Its Trade and Storage]. 116 pages. Moscow: Tsentrosoiuz. [In Rus sian.] Kosarev, M.F. 1981. Bronzovyi vek zapadnoi Sibiri [Bronze Age in Western Siberia]. 279 pages. Moscow: Nauka. [In Russian.] 1987a. Epokha posdnei bronzy I perekhodnoye vremya ot bronzovogo veka k zheleznomy [The Epoch of the Late Bronze Age and the Transi tional Time between the Bronze and Iron Ages]. In O.N. Bader, D.A. Krainov, and M.F. Kosarev, editors, Epokha bronzy lesnoi po- losy SSSR. Arkheologiya SSSR, (7):289-304. Moscow: Nauka. [In Russian.] 1987b. Nekotorye obshchiye zakonomernosti I regionalnyie osobennosti sotsialno-ekonomicheskogo razvitiya Zapadnoi Sibiri [Some Gen eral Tendencies and Regional Peculiarities of Social-Economic De velopment of Western Siberia]. Archaeologiya SSSR, pages 305-313. [In Russian.] 1991. Drevnyaya istoriya Zapadnoi Sibiri: chelovek I prirodnaya sreda [The Ancient History of Western Siberia: Man and Nature]. 302 pages. Moscow: Nauka. [In Russian.] Kosintsev, PA. 1997. Krupnye mlekopitaushie lesotundrovoi zony Zapadnoi Sibiri v nachale pozdnego golotsena [Megamammals of Forest-Tundra Zone of the West Siberia in the Beginning of the Late Holocene]. Materi aly po Istorii i Sovremennomu Sostoyaniu Fauny Severa Zapadnoi Sibiri, pages 133-163. Chelyabinsk: Institut Ekologii Rastenii i Zhi votnykh. [In Russian.] Krushanov, A.I., editor 1989. Istoriya Vostoka SSSR s drevneishikh vremen do XYII veka [The His tory of the Far East from Ancient Time until the XVII Century]. 375 pages. Moscow: Nauka. [In Russian.] Kuzmina, I.E. 1971. Istoria teriofauny Severnogo Urala v posdnem Antropogene [The Evolution of Theriofauna of the North Urals in the Late Anthropo- gene]. Trudy Zoologicheskogo Instituta, Akademiya Nauk SSSR, 49:44-122. [In Russian.] NUMBER 89 137 Moshinskaya, V.I. 1953. Materialnaya kultura I khozyaistvo Ust-Poluya [Material Culture and Economy of Ust' Poluisk Settlement]. Materialy i Issledovanya po Arkheologii SSSR. 35:72-106. [In Russian.] Mourer-Chauvire, Cecile 1983. Les oiseax dans les habitats prehistoriques: Gibier des hommes ou proies des rapaces? In C. Grigson and J. Clutton-Brock, editors, An imals and Archaeology, 2: Shell Middens, Fish and Birds. British Archaeological Reports, International Series, 183:111-124. Parmalee, Paul W. 1977a. The Avifauna from Prehistoric Arikara Sites in South Dakota. Plains Anthropologist, 22(77): 189-222. 1977b. Avian Bone Pathologies from Arikara Sites in South Dakota. Wilson Bulletin, 89(4):628-632. Potapov, R.L. 1985. Otryad Kuroobraznyie (Galliformes), Chast' 2: Semeistvo Teterevi- nyie (Tetraonidae) [Order Galliformes, Part 2: Family Tetraonidae]. Fauna SSSR, Ptitsy [Fauna of the USSR, Birds], 3(1): 1-638. [In Russian.] Potapova, O.R. 1990. Ostatki ptits is pleistotsenovykh otlozhenii Medvezhyei peshchery [Bird Remains from the Pleistocene Deposits of Medvezhya Cave in the Northern Urals]. Trudy Zoologicheskogo Instituta, Akademiya NaukSSSR, 212:135-153. [In Russian.] 1991. Ostati ptits is paleoliticheskikh sloev grota "Bolshoi Glukhoi" na Srednem Urale [Bird Remains from Paleolithic Layers of Grotto Bolshoi Glukhoi in the Middle Urals]. In G.F. Baryshnikov and I.E. Kuzmina, editors, Tezisy VI Koordinatsionnogo Soveshchaniya po Isucheniu Mamontov i Mamontovoi Fauny, pages 41?43. [In Rus sian. Brochure of abstracts printed for the Zoological Institute, Rus sian Academy of Sciences, at printing office GPPO-3, St. Petersburg, Russia.] Potapova, O.R., and M.V. Sablin 1994. Hunting Specialization of the People in the Russian North in the Early Metal Period. In M. Kokabi, J. Wahl, T. Uldin, and J. Rehmet, editors, Abstracts of the 7th International Congress on Archaeozool- ogy. Konstanz: Archaeologisches Landesmuseum Baden-Wurttem- berg. [Pages unnumbered.] Romanov, A.A. 1934. O beloi kuropatke (Lagopus albus) Lensko-Khatangskogo raiona [On Willow Ptarmigan (Lagopus albus) in Lena-Khatanga Rivers Area]. Trudy Arkticheskogo Instituta, 2:45-54. [In Russian.] Serjeantson, Dale 1988. Archaeological and Ethnographic Evidence for Seabird Exploitation in Scotland. Archaeozoologia, 2:209-224. Shukhov, I.N. 1915. Ptitsy Obdorskogo Kraya [Birds of Obdorsk District], Yezhegodnik Zoologicheskogo Museya Imperatorskoi Akademii Nauk, 20: 167-238. [In Russian.] Silantyev, A.A. 1898. Obzor promyslovykh okhot v Rossii [Review of Trade Hunting in Russia]. 620 pages. St. Petersburg. [In Russian.] Sokolova, Z.P 1972. Kult zhivonykh v religiyakh [Cults of Animals in Religions]. 216 pages. Moscow: Nauka. [In Russian.] Sternberg, L.Y. 1925. Kult orla u sibirskih narodov [Cult of Eagle in Siberian People]. Sbornik Museya Antropologii i Etnografii, 4(2):717-740. [In Rus sian] Tiemeir, O.W. 1941. Repaired Bone Injuries in Birds. Auk, 58(3):350-359. Tugarinov, A.Y. 1932. K kharakteristike chetvertichnoi omitofauny Sibiri [On Characteris tics of Quaternary Avifauna of Siberia]. Trudy Komissii po Isuche niu Chetvertichnogo Perioda, 1:115-130. [In Russian.] Volkova, VS., V.A. Bakhareva, and T.P Levina 1989. Rastitelnost' I klimat golotsena Zapadnoi Sibiri [Vegetation and Cli mate of the Holocene in Western Siberia]. In N.A. Khotinskii, edi tor, Paleoklimaty pozdnelednikovya i golotsena [Paleoclimates of Late Ice Age and Holocene], pages 90-95. Moscow: Nauka. [In Russian.] Wrangel, F.P. 1848. Puteshestviye po severnym beregam Sibiri i po Ledovitomu mom sovershonnoye v 1820, 1821, 1822, 1823 i 1824 gg.: ekspeditsiei pod nachalstvom flota leitenanta F.P. Wrangelya [The Expedition of Navy Lieutenant F.P. Wrangel by Northern Coasts of Siberia and by Arctic Sea in 1820, 1821, 1822 1823 and 1824]. 455 pages. Mos cow: Glavsevmorput'. [In Russian.] Yudin, K., and L.V. Firsova 1988. Serebristaya chaika [Herring Gull]. In V.D. Il'ichev and V.A. Zubakin, editors, Ptitsy SSSR: Chaikovye [Birds of the USSR: Sub order Lari], pages 126-145. Moscow: Nauka. [In Russian.] Zhitkov, V.M. 1912. Ptitsy poluostrova Yamala [Birds of the Yamal Peninsula]. Yezhe godnik Zoologicheskogo Museya Imperatorskoi Akademii Nauk, 17:311-369. [In Russian.] Zubareva, V.I. 1950. Iskopaemye ptitsy is chetvertichnykh otlozhenii USSR, Sobshchenie 1 [Fossil Birds from Quaternary Deposits of Ukxanian SSR, Report 1]. Trudy Instituta Zoologii, Akademii Nauk Ukrainskoi SSR, 4:78-99. [In Russian.] Seabirds and Late Pleistocene Marine Environments in the Northeast Atlantic and the Mediterranean Tommy Tyrberg ABSTRACT The technique of reconstructing Pleistocene environments by finding present-day areas of sympatry for the taxa occurring in paleofaunas has been extensively used with micromammals in North America. For a number of reasons the method is not gener ally applicable to birds; however, most of these objections do not apply to obligate seabirds. This paper treats 31 West Palearctic late Pleistocene faunas from 22 sites containing two or more species of obligate seabirds. The analysis suggests that the waters around the British Isles during the last (Eemian) interglacial were slightly warmer than during the present interglacial, that conditions in the western Mediterranean during much of the last glaciation were similar to those found in the Bay of Biscay and around the British Isles at the present time, and that conditions in the Norwegian Sea during the warmest part of the mid-Weichselian interstadial were similar to those found off the west coast of Spitzbergen today. The stratigraphic position of a number of undated avifaunas containing seabirds is discussed based on the species composition of the sea- birds. Intoduction Reconstructing Pleistocene environments by identifying present-day areas of sympatry for the taxa occurring in paleo faunas is a method that has been extensively used in North America, particularly as applied to micromammals on the Great Plains (e.g., Graham et al., 1987). The theory and proce dures used are summarized by Graham and Semken (1987). The same method of analysis was independently applied by Ol son and Rasmussen (1986) to the Oligocene avifauna of Fayum in Egypt, but otherwise this technique does not seem to have been applied to birds. The method has been little used in Europe for several rea sons. The concept works best in large blocks of relatively ho mogenous territory without dispersal barriers, where taxa can migrate freely in response to climatic changes. This applies to Tommy Tyrberg, Kimstadsv. 37, S-610 20 Kims tad, Sweden. central North America but not to large parts of Europe, where mountain ranges and marine barriers strongly affect the distri bution of terrestrial animals. Most European Pleistocene faunas also are "disharmonious," that is, they contain taxa that are al- lopatric at the present time, either because there are no good modern analogs of the relevant Pleistocene habitats, or because the modern ranges of taxa have been affected by humans. Euro pean Pleistocene mammalian faunas also frequently contain a fairly large proportion of extinct taxa, the habitat requirements of which cannot be determined with certainty. These and similar factors also affect avifaunas. Although the proportion of extinct taxa is quite low in late Pleistocene Euro pean avifaunas, and dispersal barriers affect birds less than mammals, a large proportion of the European avifauna consists of long-distance migrants with total annual ranges so large that their occurrence provides very little constraint on environmen tal conditions. The high vagility of birds also means that there is always a risk that a fossil record may be from a vagrant and is outside the species' normal range. There is, however, one group of birds to which most of these problems do not apply, namely, strictly marine seabirds. These birds use a continuous habitat (the sea), without any dispersal barriers. Although their modern range has certainly been influ enced by humans, this would be more likely to result in the decimation or extirpation of breeding colonies than in changes in the total annual range of species. Late Pleistocene European avifaunas contain two extinct species of seabirds {Puffinus holei Walker et al. (the incorrect original spelling "holer used herein was emended to "holeae" by Michaux et al., 1991) and Pinguinus impennis), but this does not seriously compromise the usefulness of the sympatry method, particularly because it is possible to reconstruct the "present-day" range of Pinguinus impennis with fair precision from historical data and subfossil records (Figure 1). The range of Puffinus holei cannot be reconstructed, although it is known to have bred on Fuerteventura, Canary Islands (Walker et al., 1990). The composition of a local fauna of obligate seabirds should therefore be a reasonably good indicator of the condi tion of the nearshore waters off the site. 139 140 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?The reconstructed "modem range" of Pinguinus impennis in the East Atlantic used to determine areas of sympatry. (?=Holocene subfossil record, A=historical breeding site, x=historical nonbreeding record, ?=occurrence uncertain.) Unfortunately, despite the fact that well over 1000 sites with late Pleistocene avifaunas are known from the West Palearctic, obligate seabirds occur in only a very small pro portion of these sites. The main reason for this is the eustatic lowering of sea levels during glacial periods, which means that coastal sites of glacial age are now mostly submerged. Exceptions to this rule are mostly found on steep Mediterra nean coasts and in northern Europe (e.g., Norway), where the isostatic rebound of formerly glaciated areas have kept pace with the eustatic rise of the sea level. In one case (Archi, in Calabria, Italy), a glacial coastal site has been preserved through tectonic movements (Ascenzi and Segre, 1971a, 1971b). In principle, interglacial coastal sites should be acces sible, but avifaunas of interglacial age are unfortunately ex tremely rare. The reasons for this are not well understood, but extensive erosion and weathering (near the end of intergla- cials?), which have destroyed most interglacial cave deposits, are presumably at least a partial explanation. NUMBER 89 141 METHODS Only strictly marine species are considered in this study to ensure that the occurrence of the birds truly reflects marine conditions. For birds that regularly frequent freshwater or land habitats, occurrence at a site might indicate that suitable fresh water or terrestrial, rather than marine, habitat existed in the vi cinity. The species used herein are all Procellariidae, all Hydro- batidae, Moms bassanus, Phalacrocorax aristotelis, Somateria mollissima, Catharacta skua, Larus audouinii, Rissa tridacty- la, Pagophila eburnea, and all Alcidae (including Pinguinus). Nomenclature for species' binomials and English names of modern birds follows Sibley and Monroe (1990). With two ex ceptions, these species are never found inland except as rare vagrants. The exceptions are Somateria mollissima and Rissa tridactyla, which regularly fly over land on migration but rare ly stop at inland sites. Xema sabini also might have been in cluded, but it has not yet been recorded from the Pleistocene. Larus hyperboreus has not been included because it can proba bly not be reliably separated from Larus marinus on osteologi cal criteria. Puffinus puffinus and P. yelkouan have been treated as one species because they were not considered separate spe cies at the time when most of the determinations were made. Only late Pleistocene sites where at least two marine species were reported are included in this study. Data on relevant avi faunas and their dates were obtained through an extensive search of the literature. This yielded a total of 22 sites and 31 faunas that fulfill these criteria (Tables 1, 2). At stratified sites, each layer has been considered as a separate fauna. Only late Pleistocene records have been considered, both because dating of older records is usually quite uncertain and because the ecol ogy of the birds may have changed over a longer time interval. For modern distributions, the maps in Cramp (1977, 1983, 1985) were used. These are admittedly only approximate in offshore areas, but because the fossil record necessarily sam ples the coastal fauna, the nearshore distribution, which is bet ter known, is more significant. It should be noted that the rang es used are the total ranges of the species in question, breeding ranges usually being considerably more restricted. It is, howev er, only rarely possible to determine whether a fossil is from a breeding bird or not. It should be noted that the ranges in Cramp only define the main area of distribution and that most species are found more or less regularly in small numbers well outside the indicated range. This, of course, introduces a mar gin of error, but the probability of a rare or vagrant species be ing preserved as a fossil is certainly very low. For the recently extinct Pinguinus impennis, a "modern" distribution map was compiled from literary and subfossil data (Figure 1). Faunas The faunas (Table 1) are treated in approximately chronologi cal order below. TABLE 1.?Sites with late Pleistocene seabird faunas. Site Cyprus Akrotiri Aetokremnos Italy Archi Arene Candide Buca del Bersaglieri Cala Genovesi a Levanzo Grotta dei Fanciulli Grotta Pietro Tampoia Grotta Romanelli Norway Blomvag Skjonghelleren Portugal Grotte de Fuminha Gruta de Figueira Brava Spain Cova den Jaume Orat Cova Nova Cueva de Nerja Devil's Tower Es Pouas Gorham's Cave Great Britain Bacon Hole Creag nan Uamh Paviland Cave Potter's Cave Sources Mourer-Chauvire, in litt., 1996 Ascenzi and Segre, 1971a, 1971b; Cassoli and Segre, 1985 Cassoli, 1980 Lambrecht, 1933; Wolf 1938 Cassoli and Tagliacozzo, 1982 Del Campana, 1946 Lambrecht, 1933; Mayaud and Schaub, 1950; Newton, 1922; Wolf, 1938 Cassoli et al., 1979 Lie, 1986; Undas, 1942 Larsen, 1984; Larsen et al., 1987 Lambrecht, 1933; Roche, 1972; Villalta, 1964 Mourer-Chauvire and Antunes, 1991; Mourer Chauvire, in litt., 1995 McMinn etal., 1993 Florit and Alcover, 1987; McMinn and Alcover, 1992 Boessneck and von den Driesch, 1980; Eastham, 1986, 1988, 1989; Hernandez, 1993, 1994, 1995, and in litt. Garrodetal., 1928; Villalta, 1964 Alcover et al., 1981, 1992; Florit etal., 1989 Eastham, 1968, 1989; Vega Toscano, 1990 (dating only) Harrison, 1977, 1987; Stringer et al., 1986 Newton, 1917; Lambrecht, 1933; Wolf, 1938; Stuart, 1983 (dating only) Bell, 1922; Bowen, 1970 (dating only) David, 1991 142 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 2.?Composition of late Pleistocene seabird faunas. \=Fulmarus glacialis, 2=Puffinus puffinus, 3=Puffinus holei, 4=Puffinus gravis, 5=Puffmus griseus, 6=Puffmus spp., 1=Calonectris diomedea, %=Hydrobates pelagicus, 9=Phalacrocorax aristotelis, 10= Morus bassanus, \\=Somateria mollissima, \2=Catharactaskua, \3=Rissa tridactyla, \4=Pagophila eburnea, \5=Alle alle, \6=Uria aalge, \l=Uria lomvia, \8=Uria sp., \9=Cepphus grylle, 20=Alca torda, 2\=Pinguinus impennis, 22=Fratercula arctica. Site Cyprus Akrotiri Aetokremnos Italy Archi Arene Candide Buca del Bersaglieri Cala Genovesi a Levanzo Layer P9 P7 P5 P4 1. 2 t. 4 1. 2 t. 3 2 X X X X X 3 4 5 6 7 X X X 8 9 X X X 10 11 X 12 13 14 15 16 17 18 19 20 21 22 X X XX X X X X X Grotta Pietro Tampoia Grotta Romanelli Grotta dei Fanciulli Norway Blomvag Skjonghelleren Portugal Grotte de Fuminha Gruta de Figueira Brava Spain Cova den Jaume Orat Cova Nova Cueva de Nerja Devil's Tower Es Pouas Gorham's Cave Great Britain Bacon Hole Creag Nan Uamh Paviland Cave Potter's Cave C A6 EP UP UP/EP A2 Bl K D-F 5 DATED FAUNAS EEMIAN INTERGLACIAL.?Only one site can be confidently assigned to the last interglacial (Eemian/Ipswichian sensu stric- to, i.e., oxygen isotope stage (IS) 5e), Bacon Hole on the Gow- er Peninsula in Southwest Wales. Layers D-F at this site, which have been U/Th dated to 122?9 kilo annum (Ka) BP (Stringer et al., 1986), contain a temperate avifauna, including Calonectris diomedea and Alca torda (Harrison, 1977, 1987). The current area of sympatry of these two species is situated in an area stretching from the waters southwest of the British Isles to the western Mediterranean (Figure 2). There is no sympatric breeding of these two species today, but this might be due to the absence of suitable breeding sites for seabirds between Bretagne and Northwest Spain. The occurrence of these two species in Wales supports evidence from other sources (e.g., Mclntyre et al., 1972; Ruddiman and Mclntyre, 1976) that the Northeast Atlantic was somewhat warmer during the Eemian than during the present interglacial. THE EARLY WEICHSELIAN.?This is a climatically complex interval comprising two moderately cold stadials (IS 5b and IS 5d) and two interstadials of nearly interglacial magnitude (ISs 5a and 5c). The character of these interstadials was different in southern and northern Europe. In the south, IS 5a was, if any thing, warmer than the earlier IS 5c, whereas in Scandinavia the opposite was the case (e.g., Mangerud, 1991). The extent of glaciation during the stadials is uncertain, but in Scandinavia it seems to have been mainly restricted to the mountains except in the far north. No fauna can be definitely assigned to this pe riod, but it is possible that the fauna from Grotte de Furninha, Portugal (Tables 1, 2), for example, belongs here. NUMBER 89 143 FIGURE 2.?Area of sympatry of seabirds from Bacon Hole, Great Britain (Calonectris diomedea, Alca torda). (?=fossil site.) THE EARLY WEICHSELIAN PLENIGLACIAL.?This fairly short stadial (ca. 75-60 Ka BP) was quite cold, and the Scandinavian icecap expanded as far as eastern Denmark and northern Po land. The only fauna that might be assigned to this interval is layer K at Gorham's Cave (Gibraltar), which has been dated to "Wurm I" (Hernandez Carrasquilla, 1993). This is a rather "cold" fauna {Phalacrocorax aristotelis, Alle alle, Pinguinus impennis) (Figure 3), and the presence of at least Alle alle would certainly seem to indicate pleniglacial conditions, al though layer K also has been dated (Vega Toscano, 1990) to the somewhat milder mid-Weichselian (ca. 45 Ka BP). THE MID-WEICHSELIAN INTERSTADIAL COMPLEX.?This spans the period ca. 65-25 Ka BP, during which the climate was strongly cyclic. In very general terms, it consisted of two milder interstadials, ca. 60 Ka BP and ca. 30 Ka BP, separated by a colder stadial. For most of this interval, southern Scandi navia and at least parts of the Norwegian coast were ice-free. The most interesting site from this interval is Skjong- helleren (layer G) in Norway, dated to the Alesund interstadial 144 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 3.?Area of sympatry of seabirds from Gorham's Cave, Spain, layer K (Phalacrocorax aristotelis, Alle alle, Pinguinus impennis). (? = fossil site.) 30,000 yrs. BP) (Larsen, 1984; Larsen et al., 1987). This has a rich seabird fauna including Fulmarus glacialis, Rissa tridacty- la, Pagophila eburnea, Alle alle, Uria aalge, Uria lomvia, Cep- phus grylle, and Fratercula arctica. The site is a sea cave with out any trace of human presence, and the fauna likely samples mostly species breeding in the vicinity. This is the only fauna containing truly arctic taxa, and the area of sympatry is a rather narrow area stretching from the seas north of Iceland to the west coast of Spitzbergen plus the waters off Northwest Novaya Zemlya (Figure 4). The areas of "best fit" with regard to breed ing birds are Jan Mayen, Bear Island, Prince Charles' Foreland (all three with all species except Pagophila eburnea), and southern Spitzbergen (all species except Uria aalge). The pres ence of Fratercula arctica and Uria aalge as well as some non- avian taxa (e.g., Lutra lutra (Linnaeus), Pollachius virens (Lin naeus), Brosmius brosme (Ascanius)), however, indicates that water from the North Atlantic Current must have penetrated the Norwegian Sea for at least part of the Alesund interstadial. NUMBER 89 145 FIGURE 4.?Area of sympatry of seabirds from Skjonghelleren, Norway, layer G (Fulmarus glacialis, Rissa tridactyla, Pagophila eburnea, Alle alle, Uria aalge, Uria lomvia, Cepphus grylle, Fratercula arctica). (? = fossil site.) Within the rather wide margins of C14-dating, another fauna coeval with the Skjonghelleren fauna is that from the Gruta de Figueira Brava in Portugal. This is one of the youngest Moust- erian sites known and has been C14-dated to 30,930?700 yrs. BP (Mourer-Chauvire and Antunes, 1991). The fauna includes Puffinus holei, Morus bassanus, and Pinguinus impennis. The distribution of Puffinus holei outside the breeding season is un known, but the area of sympatry of the other two species are shown in Figure 5. Clearly this is a boreal fauna and suggests conditions approximating those found around the British Isles today. This is supported by the other seabirds found at the site {Gavia stellata, Melanitta nigra, Melanitta fusca, Clangula hy emails). A third site from this time interval is Archi in Calabria, Italy. This can probably be dated to ca. 40,000 yrs. BP (Cassoli and Segre, 1985). The fauna includes Morus bassanus and Pingui- 146 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 5.?Area of sympatry of seabirds from Figueira Brava, Portugal, and Archi, Italy (Morus bassanus, Pin guinus impennis). (?= fossil sites.) nus impennis, which yields the same area of sympatry as Gruta de Figueira Brava (Figure 5). This similarity in a fauna from the central Mediterranean is probably due to a somewhat colder climate at the time when the Archi fauna was deposited. THE LATE WEICHSELIAN PLENIGLACIAL.?Unfortunately, there are few sites that can be assigned to the glacial maximum, ca. 25,000-15,000 yrs. BP, probably both because this was the period of maximum eustatic lowering of the sea and because deposition of organic material virtually ceased over large areas during the coldest intervals. The only site that definitely falls within this interval is Arene Candide on the Italian Riviera. This site is very close to the sea today, and because the coast is quite steep, the sea was only ~3 km distant even during the gla cial maximum. The layers P4, P5, P7, and P9 at this site (the only ones to contain two or more species of seabirds) are all older than 18,500 yrs. BP but are probably younger than 25,000 yrs. BP (Bietti, 1987). The seabirds in these layers in clude Calonectris diomedea (P7), Puffinus puffinus (P5), Pha- NUMBER 89 147 FIGURE 6.?Area of sympatry of seabirds from Arene Candide, Italy, layer P9 (Phalacrocorax aristotelis, Uria aalge, Alca torda, Fratercula arctica). (? = fossil site.) lacrocorax aristotelis (P9), Uria aalge (P4,5,7,9), Alca torda (P4,9), and Fratercula arctica (P5,9). The areas of sympatry for the faunas in layers P4, P5, and P9 are rather similar and suggest that conditions approximating those around the British Isles at the present time prevailed off the Ligurian coast during the glacial. The area of sympatry for layer P9, which has the largest number of seabird species (4), is shown in Figure 6. Conditions may have been slightly milder when layer P7 was deposited (Figure 7). The Upper Paleolithic fauna from Cueva de Nerja (UP lay ers) near Malaga, Spain (Figure 8), is dated to the interval 16,520-13,350 yrs. BP (Hernandez, 1995) and falls within the later part of the pleniglacial, including the Lascaux interstadial and the Dryas 1 stadial. The unusually rich seabird fauna con sists of Calonectris diomedea, Puffinus puffinus, P. gravis, P. griseus, Phalacrocorax aristotelis, Morus bassanus, Uria aal ge, Alca torda, and Pinguinus impennis. This fauna has only a very small area of sympatry off Southwest Ireland. Such a 148 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 7.?Area of sympatry of seabirds from Arene Candide, Italy, layer P7 (Calonectris diomedea, Uria aalge). (#=fossil site.) small overlap in the elements of the fauna may indicate a con siderable degree of climatic change during the period of depo sition. In any case, the rich seabird avifauna indicates that the sea off Granada must have been cool and biologically rich at the time. THE LATE GLACIAL.?Sites with avifaunas from this interval (ca. 13,000-10,000 yrs. BP) are more common than for earlier intervals. Dating of the sites also is more exact, which makes it possible to divide late glacial records into three climatically distinct phases, the Boiling and Allerod interstadials and the Dryas 3 (Younger Dryas) stadial. The Boiling interstadial (ca. 13,000-12,000 yrs. BP) was a quite mild interval, when at least summer temperatures may have approached present values in some areas. Climatic conditions during the Allerod interstadial (ca. 11,800-10,800 yrs. BP) are somewhat controversial. The traditional view is that it was an interstadial comparable to, or even warmer than, Boiling and separated from it by a short but cold stadial (Dryas 2) ca. 12,000 yrs. BP. More recently the re- NUMBER 89 149 FIGURE 8.?Area of sympatry of seabirds from Cueva de Nerja, Spain (upper Paleolithic layer) (Puffinus puffi nus. Puffinus gravis, Puffinus griseus, Calonectris diomedea, Morus bassanus, Phalacrocorax aristotelis, Uria aalge, Alca torda, Pinguinus impennis). (?=fossil site.) ality of the Dryas 2 stadial has been questioned, and it has been argued that Allerod was actually colder than Boiling (Nilssom, 1983). There is, however, no doubt that during the Dryas 3 sta dial (ca. 10,800-10,100 yrs. BP), the last "cold snap" of the Wurmian glacial cycle, there was a return to fully glacial cli matic conditions lasting several centuries. The Blomvag site near Bergen in Norway is securely CI 4- dated to the Boiling interstadial (12,700-12,200 yrs. BP) (Lie, 1986). The seabird fauna consists of nine species {Fulmarus glacialis, Puffinus puffinus, Somateria mollissima, Rissa tri- dactyla, Alca torda, Uria aalge, U. lomvia, Cepphus grylle, Pinguinus impennis). The area of sympatry of these species (Figure 9) indicates conditions only slightly colder than at present. This implies a considerable contrast between condi tions in the Norwegian Sea and on land because most of Scan dinavia was still ice-covered at this time, and the ice-edge must 150 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 9.?Area of sympatry of seabirds from Blomvag, Norway (Fulmarus glacialis, Puffinus puffinus, Soma teria mollissima, Rissa tridactyla, Alca torda, Uria aalge, U. lomvia, Cepphus grylle, Pinguinus impennis). (?=fossil site.) have been quite close to Blomvag. Indeed, the site was tempo rarily overrun by the ice at some time after deposition of the fauna but before the Dryas 3 stadial. The fauna in layer B at Skjonghelleren in Sunnmore, Nor way, is C14-dated to either the end of the Allerod interstadial or the Dryas 3 stadial (11,510? 190-10,360? 170 yrs. BP) (Larsen, 1984; Larsen et al., 1987). The fauna consists of So materia mollissima, Alle alle, Uria sp., Cepphus grylle, and Fratercula arctica. Unfortunately, this is not a very informa tive set of taxa and only indicates boreal or low arctic condi tions, which might fit either a cool late phase of Allerod or the early part of the Dryas 3 stadial. There also is a fauna from Scotland that cannot be dated with any precision but is most likely from the late glacial. This is the fauna from layer 5 in Creag Nan Uamh cave, which has been considered to be either very late Pleistocene or earliest Ho- NUMBER 89 151 FIGURE 10.?Area of sympatry of seabirds from Creag Nan Uamh, Scotland (Somateria mollissima, Alle alle). (?=fossil site.) locene (Stuart, 1983). It contains Somateria mollissima and Alle alle and indicates conditions similar to or slightly colder than at present (Figure 10). Other faunal elements from the site, such as Rangifer tarandus (Linnaeus), support this conclusion. The fauna from layer C in Grotta Romanelli in Apulia, Italy {Puffinus spp., Phalacrocorax aristotelis, Rissa tridactyla, Pin guinus impennis), which is securely dated to the middle part of Dryas 3, has been published only in part (Cassoli et al., 1979), but it indicates that cool "Atlantic" conditions prevailed in the central Mediterranean even during this final cold snap of the Pleistocene (Figure 11). There are a few more Mediterranean faunas that are either late glacial or early Holocene: Gorham's Cave, Gibraltar, lay ers A2 {Phalacrocorax aristotelis, Fratercula arctica) and B1 {Puffinus puffinus, Fratercula arctica); Cueva de Nerja, Spain, epipaleolithic (EP) layers {Calonectris diomedea, Puffinus gri seus, Phalacrocorax aristotelis, Morus bassanus, Catharacta skua, Uria aalge, Pinguinus impennis) dated to 13,350-8770 152 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 11.?Area of sympatry of seabirds from Grotta Romanelli, Italy (Puffinus spp., Phalacrocorax aristote lis, Rissa tridactyla, Pinguinus impennis). (?=fossil site.) yrs. BP (Hernandez, 1995); Cala Genovesi on Levanzo Island off Sicily {Puffinus puffinus, Calonectris diomedea), which is younger than 11,180 yrs. BP (Cassoli and Tagliacozzo, 1982); and Akrotiri Aetokremnos on Cyprus {Puffinus puffinus, Pha lacrocorax aristotelis), which is dated to 11,700-9000 yrs. BP (Simmons, 1991; Mourer-Chauvire, in litt.). These faunas mostly indicate conditions more or less similar to the present day, which may indicate early Holocene age. The exception is the Cueva de Nerja epipaleolithic layer (Figure 12), which indi cates cool Atlantic conditions and may date largely to the cold Dryas 1 and/or Dryas 3 stadials. The fauna from Akrotiri Aetokremnos (Figure 13) also may indicate a slightly colder and more eutrophic sea around Cyprus than at present, perhaps because the fossils were deposited during the Dryas 3 stadial. UNDATED FAUNAS There are a number of faunas containing seabirds that are not at present satisfactorily dated. In these cases, consideration of the environmental information given by the seabirds may help NUMBER 89 153 FIGURE 12.?Area of sympatry of seabirds from Cueva de Nerja, Spain (epipalaeolithic layer) (Puffinus griseus, Calonectris diomedea, Morus bassanus, Phalacrocorax aristotelis, Catharacta skua, Uria aalge, Pinguinus impennis). (?=fossil site.) in defining the age of these faunas. These faunas can be divid ed into two groups: faunas that indicate significantly colder conditions than at present and faunas that indicate conditions approximately similar to the present day. GROTTA PIETRO TAMPOIA, ITALY.?This is a "warm" fauna {Puffinus puffinus, Calonectris diomedea, Hydrobates pelagi- cus), indicating conditions similar to today, which supports Mayaud and Schaub's (1950) view that the fossils from this site are largely or wholly of Holocene age. BUCA DEL BERSAGLIERI, ITALY.?This also is a warm fau na {Puffinus puffinus, Phalacrocorax aristotelis) (Figure 13), particularly because at least some of the shearwater fossils are from the warm-water species Puffinus {puffinus) yelkouan. Very little information on the date of the deposits is available, but it seems likely that the fauna is at least partly Holocene, a hypothesis that is supported by the composition of the avifau na in general and by the presence of Rattus rattus Linnaeus (Wolf, 1938). 154 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 13.?Area of sympatry of seabirds from Akrotiri Aetokremnos, Cyprus, and Buca del Bersaglieri, Italy (Puffinuspuffinus, Phalacrocorax aristotelis). (?=fossil sites.) PAVILAND CAVE, GREAT BRITAIN.?The seabirds from this site {Morus bassanus, Uria aalge) indicate temperate condi tions similar to the present, and the remains are therefore very unlikely to be contemporary with the famous Upper Paleolithic "red lady of Paviland," which has been C14-dated to 18,460?340 yrs. BP (Bowen, 1970), the coldest part of the pleniglacial. During the Pleniglacial, Paviland Cave would in any case have been situated a very considerable distance from the seashore. The seabird remains may well be Holocene. POTTER'S CAVE, GREAT BRITAIN.?It has been surmised that the seabirds from this site are younger than the other bird remains and are actually Holocene (David, 1991). The sympat- ric area of the species {Puffinus puffinus, Fratercula arctica) indicates temperate conditions and supports a Holocene date. Discussion Reconstructions of past conditions based on the composition of fossil faunas are based on the premise that the ecological re quirements of the species used has remained constant during NUMBER 89 155 the period studied. This is probably a safe assumption for most of the seabirds in this study. It should be noted that none of the 31 faunas contains seabird species with allopatric ranges (al though this might have been the case if Puffinus yelkouan had been treated as a distinct species). This absence of allopatric species is in marked contrast to late Pleistocene nonmarine avi faunas, where species that are today widely allopatric (e.g., Nyctea scandiaca and Alectoris spp.) frequently occur togeth er. This similarity with modern seabird faunas also is an argu ment that the ecology of the species concerned has remained fairly constant over the period studied. Most of the reasonably well-dated faunas described above are either from the western Mediterranean or from Norway. The western Mediterranean faunas mostly suggest cool, bio logically productive seas, and the western Mediterranean to gether with the waters around the Macaronesian islands may have been an important refugium for boreal seabirds during the coldest parts of the glaciation. Unfortunately, there is a dearth of sites from the mildest parts of the glaciation and from the eastern Mediterranean basin, although an isolated record shows that Morus bassanus occurred as far east as Crete during some part of the last glaciation (Suriano, 1980). The Norwegian faunas, of course, date only from the milder parts of the glaciation and are perhaps most interesting as illus trations of violent climatic and environmental shifts during the late Pleistocene. This is particularly striking when comparing the Blomvag fauna (Figure 9) from the mild Boiling interstadi al with the 2000 years younger Grotta Romanelli fauna (Figure 11) from the cold Dryas 3 stadial. The occurrence of Pinguinus impennis in both these faunas is particularly noteworthy and suggests that even this nonvolant species was capable of changing its distribution quite rapidly in response to changing conditions. This contradicts Bengtson's (1984) theory that the extinction of Pinguinus impennis was at least partly due to an inability to adapt to environmental changes and supports the traditional view that the species' extinction was directly caused by human action. Literature Cited Alcover, J.A., F. Florit, C. Mourer-Chauvire, and P.D.M. Weesie 1992. The Avifaunas of the Isolated Mediterranean Islands during the Middle and Late Pleistocene. In K.E. Campbell, editor, Papers In Avian Paleontology Honoring Pierce Brodkorb. Science Series, Nat ural History Museum of Los Angeles County, 36:273-284. Alcover, J.A., S. Moya-Sola, and J. Pons-Moya 1981. Les quimeres del passat. 260 pages. Ciutat de Mallorca: Editorial Moll. [Instirucio Catalana d'Histdria Natural filial de l'lnstitut d'Estudis Catalans, Memoria 11.] Ascenzi, A., and A.G. Segre 1971a. A New Neandertal Child Mandible from an Upper Pleistocene Site in Southern Italy. Nature, London, 233(5317):280-283. 1971b. II giacimento con mandibola neandertaliana di Archi (Reggio Cala bria). Accademia Nazionale dei Lincei, Rendiconti della Classe di Scienze Fisiche, Matematiche e Naturali, series 8, 50(6):763-771. Bell, A. 1922. Pleistocene and Later Birds of Great Britain and Ireland. Naturalist, Hull, 1922(Aug-Sept):251-253. Bengtson, S.-A. 1984. Breeding Ecology and Extinction of the Great Auk (Pinguinus im pennis): Anecdotal Evidence and Conjectures. Auk, 101:1-12. Bietti, A. 1987. Some Remarks on the New Radiocarbon Dates from the Arene Can dide Cave (Savona, Italy). Human Evolution, 2(2): 185-190. Boessneck, J., and A. von den Driesch 1980. Tierknochenfunde aus der sudspanischen hohlen. Studien uber Fruhe Tierknochenfunde von der Iberischen Halbinsel, 7:160-185. Bowen, D.Q. 1970. The Palaeoenvironment of the 'Red Lady' of Paviland. Antiquity, 44:134-136. Cassoli, P.F. 1980. L'Avifauna de pleistocene superiore delle Arene Candide (Liguria). Memorie dell'Istituto Italiano di Paleontologia Umana, new series, 3:155-234. Cassoli, P.F., and A.G. Segre 1985. L'Alca impenne Alca impennis del Pleistocene d'ltalia. In M. Fa- sola, editor, Atti 111 Convegno Italiano di Ornitologia, pages 251-254. Cassoli, P.F., A.G. Segre, and E. Segre 1979. Evolution morphologique et ecologique de la cote de Castro (Pouil- les) dans le Pleistocene final. In D. Sonneville-Bordes, editor, La fin des temps glaciaires en Europe. Colloques Internationaux du Centre National de la Recherche Scientifique, 271: 325-332. Cassoli, PR, and A. Tagliacozzo 1982. La fauna della grotta di Cala Genovesi a Levanzo. Rivista di Scienze Preistoriche, 37:124-133. Cramp, S., editor 1977. Handbook of the Birds of Europe, the Middle East and North Africa; The Birds of the Western Palearctic. Volume 1,714 pages. Oxford: Oxford University Press. 1983. Handbook of the Birds of Europe, the Middle East and North Africa; The Birds of the Western Palearctic. Volume 3, 913 pages. Oxford: Oxford University Press. 1985. Handbook of the Birds of Europe, the Middle East and North Africa; The Birds of the Western Palearctic. Volume 4, 936 pages. Oxford: Oxford University Press. David, A. 1991. Late Glacial Archaeological Residues from Wales: A Selection. In N. Barton, A.J. Roberts, and D.A. Roe, editors, The Late Glacial in Northwest Europe. Council for British Archaeology Research Re port, 77:141-159. Del Campana, D. 1946. Uccelli paleolitici degli strati basali della Grotta dei Fanciuli ai Balzi Rossi. Rivista di Scienze Preistoriche, 1:290-307. Eastham, A. 1968. The Avifauna of Gorham's Cave, Gibraltar. Bulletin of the Institute of Archaeology, 1:31-42. 1986. The Birds of the Cueva de Nerja. In J.F. Jorda Pardo, editor, Lapre- historia de la Cueva de Nerja (Malaga): Trabajos sobre la Cueva 156 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY de Nerja, 1: Paleolitico superior y Epipaleolitico, pages 109-130. Malaga: Patronato de la Cueva de Nerja. 1988. The Season and the Symbol: A Study of the Place of Swallows in Pleistocene Cave Deposits and Iconography. Archaeozoologia, 2(1/ 2):243-252. 1989. Cova Negra and Gorham's Cave: The Place of Birds in Mousterian Communities. In J. Clutton-Brock, editor, The Walking Larder: Pat terns of Domestication, Pastoralism, and Predation, pages 350-357. London: Unwin Hyman. [Papers delivered at the World Archaeologi cal Congress, Southampton, England, 1986.] Florit, X., and J.A. Alcover 1987. Els ocells del pleistoce superior de la Cova Nova (Capdepera, Mal lorca), I?II. Bolleti de la Societal d'Historia Natural de les Baleares, 31:7-44. Florit, X., C. Mourer-Chauvire, and J.A. Alcover 1989. Els ocells pleistocenics d'Es Pouas, Eivissa; Nota preliminar. Bull- let i de la Institucio Catalana d'Histdria Natural, section 5, Geolo gia, 56:35^46. Garrod, D.A.E., L.H. Dudley-Buxton, G. Elliot-Smith, and D.M.A. Bate 1928. Excavation of a Mousterian Rock-Shelter at Devil's Tower, Gibral tar. Journal of the Royal Anthropological Institute of Great Britain and Ireland. 58:33-110. Graham, R.W., and H.A. Semken, Jr. 1987. Philosophy and Procedures for Paleoenvironmental Studies of Qua ternary Mammalian Faunas. In R.W. Graham, H.A. Semken, Jr., and M.A. Graham, editors, Late Quaternary Mammalian Biogeog raphy and Environments of the Great Plains and Prairies, pages 1-17. Springfield: Illinois State Museum. Graham, R.W, H.A. Semken, Jr., and M.A. Graham, editors 1987. Late Quaternary Mammalian Biogeography and Environments of the Great Plains and Prairies, xiv+491 pages. Springfield: Illinois State Museum. Harrison, C.J.O. 1977. Non-Passerine Birds of the Ipswichian Interglacial from the Gower Caves. Transactions. British Cave Research Association, 4(4): 441-442. 1987. Pleistocene and Prehistoric Birds of South-West Britain. Proceed ings. University of Bristol Spelaeological Society, 18(1):81-104. Hernandez Carrasquilla, F. 1993. Catilogo provisional de los yacimientos con aves del Cuaternario de la Peninsula Iberica. Archaeofauna, 2:231-275. 1994. Addenda al catalogo provisional de yacimientos con aves del Cua ternario de la Peninsula Iberica. Archaeofauna, 3:77-92. 1995. Cueva de Nerja (Malaga): Las aves de las campanas de 1980 y 1982. In Trabajos sobre la Cueva de Nerja, 5:221-293. Malaga: Patronato de la Cueva de Nerja. Lambrecht, K. 1933. Handbuch der Palaeornithologie. 1024 pages. Berlin: Gebriider Bomtraeger. Larsen, E. 1984. Weichsel stratigrafi og glasialgeologi pa Nordvestlandet. [134 pages in total, comprising separate papers.] Doctoral dissertation, Univer- sitetet i Bergen, Geologisk Institutt, Avd. B. Larsen, E., S. Gulliksen, S.-E. Lauritzen, R. Lie, R. Lovlie, and J. Mangerud 1987. Cave Stratigraphy in Western Norway; Multiple Weichselian Glaci- ations and Interstadial Vertebrate Fauna. Boreas, 16:267-292. Lie, R. 1986. Animal Bones from the Late Weichselian in Norway. Fauna Nor- vegica, series A, 7:41-46. Mangerud, J. 1991. The Last Interglacial/Glacial Cycle in Northern Europe. In L.C.K. Shane and E.J. Cushing, editors, Quaternary Landscapes, pages 37-75. London: Belhaven Press. Mayaud, N., and S. Schaub 1950. Les Puffins subfossiles de Sardaigne. Verhandlungen der Natur- forschenden Gesellschaft in Basel. 61:19-27. Mclntyre, A., W.F. Ruddiman, and R. Jantzen 1972. Southward Penetration of the North Atlantic Polar Front: Faunal and Floral Evidence of Large-Scale Surface Water Movements over the Last 225,000 Years. Deep-Sea Research, 19:61-77. McMinn, M., and J.A. Alcover 1992. Els ocells del pleistoce superior de la Cova Nova (Capdepera, Mal lorca), III: Novas aportacions al registre. Bolleti de la Societal d'Historia Natural de les Baleares, 35:17-32. McMinn, M., CR. Altaba, and J.A. Alcover 1993. La fauna fossil de la Cova den Jaume Orat (Parroquia d'Albarca, Sant Antoni de Portmany, Eivissa). Endins, 19:49-54. Michaux, J., R. Hutterer, and N. Lopez-Martinez 1991. New Fossil Faunas from Fuerteventura, Canary Islands: Evidence for a Pleistocene Age of Endemic Rodents and Shrews. Comptes Rendus de I'Academie des Sciences, series 2, 312(7):801?806. Mourer-Chauvire, C, and M.T. Antunes 1991. Presence du Grand Pingouin, Pinguinus impennis (Aves Charadrii formes) dans le Pleistocene du Portugal. Geobios, 24(2):201-205. Newton, E.T. 1917. Notes on Bones Found in the Creag Nan Uamh Cave, Inchnadamff, Assynt, Sutherland. Proceedings of the Royal Society of Edinburgh, 1916-1917:344-348. 1922. Fossil Bird-Remains Collected by Dr. Forsyth Major in Sardinia, Corsica and Greece. Proceedings of the Zoological Society of Lon don, 1922:229-232. Nilsson, T. 1983. The Pleistocene?Geology and Life in the Quaternary Ice Age. 651 pages. Dordrecht: Reidel. Olson, S.L., and D.T. Rasmussen 1986 The Paleoenvironment of the Earliest Hominoids: New Evidence from the Oligocene Avifauna of Egypt. Science. 233:1202-1204. Roche, J. 1972. Faunes du Pleistocene superieur et final de rEstremadura, Portugal. Annates de Paleontologie (Vertebres), 58:229-242. Ruddiman, W.F., and A. Mclntyre 1976. Northeast Atlantic Paleoclimatic Changes Over the Past 600,000 Years. Memoirs, Geological Society of America, 145:111-146. Sibley, C.G., and B.L. Monroe, Jr. 1990. Distribution and Taxonomy of Birds of the World. 1111 pages. New Haven: Yale University Press. Simmons, A.H. 1991. Humans, Island Colonizations and Pleistocene Extinctions in the Mediterranean: The View from Akrotiri Aetokremnos, Cyprus. An tiquity, 65:857-869. Stringer, C.B., A.P Currant, HP Schwarz, and S.N. Collcutt 1986. Age of Pleistocene Faunas from Bacon Hole, Wales. Nature, Lon don, 320:59-62. Stuart, A.J. 1983. Pleistocene Bone Caves in Britain and Ireland, a Short Review. Studies in Speleology, 4:9-36. Suriano, F. 1980. Fossil Birds of Simonelli Cave. Accademia Nazionale dei Lincei, 249:123-126. Undas, I. 1942. Fossilfunnet i Blomvag. Naturen, Bergen, 66:97-107. Vega Toscano, L.G. 1990. La fin du paleolithique moyen au sud de l'Espagne i ses implications dans le contexte de la Peninsule Iberique. In C. Farizy, editor, Paleolithique moyen recent et Paleolithique superieur ancien en Eu rope; Ruptures et transitions: examen critique des documents archeologiques; Actes de Colloque international de Nemours 9-10-11 mai 1988. Memoires du Musee de Prehistoire d'lle de France, 3:170-176. Nemours. Villalta, J.F. 1964. Datos para un catilogo de las aves fosiles del cuaternario Espanol. Speleon, 15:79-102. NUMBER 89 157 Walker, C.A., G.M. Wragg, and C.J.O. Harrison 1990. A New Shearwater from the Pleistocene of the Canary Islands and Its Bearing on the Evolution of Certain Puffinus Shearwaters. His torical Biology, 3(3):203-224. Wolf, B. 1938-1941. Fauna Fossilis Cavernarum l-III, Parts 82, 89, 92. In W. Quennerstedt, editor, Fossilium Catalogus, I: Animalia. The Hague: W. Junk [parts 82, 89]; Neubrandenburg: Gustaf Feller [part 91]. [Complete pagination unknown, partial pagination as follows: sec tion 2 (Hohlenkatalog), part 82:1-192, part 89:193-240, part 91:241-288; section 3 (Tierkatalog), part 82:1-96, part 89:97-208, part 91:209-320. Intraspecific Variation in Modern and Quaternary European Lagopus John R. Stewart ABSTRACT Skeletal proportions of modem European populations of Lago pus lagopus (Linnaeus) and L. mutus (Montin) from Britain, Ice land, Scandinavia, northern Russia, and the Alps are compared with their Quaternary fossil counterparts from Britain, Poland, France, and Belgium. Lagopus lagopus and L. mutus from most pre-Holocene deposits are found to differ allometrically from modem samples. This difference is best seen in the tarsometatar sus, which often is more robust in both species in the Pleistocene and in turn may reflect greater body weight. Possible correlations between this phenomenon and the climatic and ecological condi tions of the past, as well as the possibility that these birds were less sedentary, are discussed. Introduction Previous workers, such as Newton (1924), Mourer-Chauvire (1975a, 1975b), Janossy (1974, 1976), Bochehski (1974, 1985, 1991), Harrison (1980), Potapova (1986), and Bo- cheriski and Tomek (1994) have described differently sized and proportioned postcranial bones of both Lagopus mutus (Montin) and Lagopus lagopus (Linnaeus) from the fossil record of Europe. Specifically, samples of both species from the Pleistocene were seen to differ in size and allometry from skeletons of their recent counterparts. These allometric differ ences were seen in the relative proportions of their tarsometa tarsi, carpometacarpi, phalanx 1 digit III pedis, and phalanx 1 digit II alae, as well as in cranial elements. Newton, who was probably the first to note the occurrence of anomalously proportioned Lagopus fossils of Pleistocene age, believed that a third species had existed (Newton, 1924). He re ferred to this species, found at Merlin's Cave, a late Pleistocene site in Britain, as a "small ptarmigan" and never named it as a distinct form, which is how many authors have dealt with these John R. Stewart, The McDonald Institute for Archaeological Re search, Downing Street, Cambridge CB2 3ER, England. anomalies since. Exceptions are the species Lagopus atavus Janossy (1974) from the late Pliocene of Poland (Janossy, 1976; Bocheriski, 1991) and the subspecies L. lagopus noail- lensis Mourer-Chauvire (1975a) and L. mutus correzensis Mourer-Chauvire (1975a) from La Fage, a Rissian (middle Pleistocene) site in France. Lagopus lagopus noaillensis and L. mutus correzensis were described by Mourer-Chauvire as being distinguished by the robustness of their tarsometatarsi as compared to modern popu lations. She also plotted the mean lengths and shaft widths of tarsometatarsi of both species from La Colombiere and Gigny, two assemblages of different ages from the last glacial of France, together with the two named subspecies from La Fage. This showed that they, too, had relatively robust tarsometatarsi, although their lengths varied, producing allometric-shape vari ation. The most detailed study of Lagopus fossils to date was that by Bocheriski (1974). In his book he compared a number of Polish fossil populations of both species from the last glaciation with samples from much of their modern European distribution. He concluded that the fossil birds possessed longer carpometa carpi and humeri but shorter tarsometatarsi than their modern counterparts, and that the articular ends of the humeri, tar sometatarsi, and coracoids were more massive. Bocheriski in terpreted the differences in the wing-bone lengths as indicating that the primary feathers of both species had become shorter, thus reducing wing-surface area over time. This, together with the change in bone robustness, especially that of the coracoid, indicated to him that the two species had maintained, or had only slightly reduced, their body size over time. Later, Bo cheriski (1985) showed that Pleistocene L. lagopus in Poland had longer distal-wing bones (carpometacarpi and phalanx 1 digit II alae) and shorter distal-leg elements (tarsometatarsi and phalanx 1 digit III pedis) when compared to modern birds. Sub sequently, Bocheriski and Tomek (1994) focused on the rela tive lengths of postcranial bones of Pleistocene L. mutus in Austria and showed that they differed from modern alpine birds in having shorter tarsometatarsi but longer carpometacarpi. 159 160 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Recent investigations by neontologists into intraspecific variation in size and allometry, seen across both space and re cent time, emphasize the need for a broader-based approach to variation seen in fossils. Zink and Remsen's (1986) review of the relevance of geographic variation to evolution, and case studies such as those by Johnston and Selander (1964), Grant (1971, 1986), Fleischer and Johnson (1982), James (1983), and Dennison and Barker (1991), demonstrate that within-species morphological differences in birds are frequently observed across space, and that these differences may evolve over rela tively short spans of time. Analyses of present-day intraspecific geographic variation therefore give an insight into the spectrum of variation that may have occurred over recent geological time in species. A notable exception to the lack of application of these neon tological studies to fossils or subfossils is the study by Ericson (1987), wherein subfossil Common Eider {Somateria mollissi ma Linnaeus) from Scandinavia were compared to a variety of different extant subspecies and were shown to differ signifi cantly. In this instance various environmental criteria were considered causal, including climatic change and anthropogen ic effects; the author preferred the latter as an explanation. MATERIALS AND METHODS The present study has attempted to bring together a widely distributed body of data on both species of Lagopus from the present-day western Palearctic. These data were used for bio- metric comparison with a broader range of geographically and temporally separated fossils than has been previously achieved. All but the British and Belgian fossil samples have been metri cally studied by others, although no one has examined all these samples together. Those that are newly analyzed herein are from Pin Hole Cave (Manchester City Museum), Merlin's Cave (University of Bristol Speleological Society Museum), Westbury-sub-Mendip (Paleontology Department, Natural His tory Museum, London), and Remouchamps (Musee Royaux d'Art d'Histoire, Bruxelle). The tarsometatarsus is invariably the most frequently occur ring skeletal element of medium-sized Galliformes in Europe an cave assemblages (Mourer-Chauvire, 1983) and is thus the main element dealt with herein. The relative abundance of this bone is fortunate because it is one of the elements most easily identified as either Lagopus lagopus or L. mutus (Kraft, 1972; Mourer-Chauvire, 1975a; Bocheriski, 1985). There appear to be no details of morphology that can aid determination, and the two species are identified simply on the basis of size: L. lago pus is consistently larger than L. mutus (see Figure 2). Many other postcranial bones present greater problems because they overlap considerably in size and therefore make analysis more complicated. Table 1 lists all the modern skeletal samples of both species, including two populations of L. lagopus scoticus (Latham), one from Scotland and one from Derbyshire, England; samples of TABLE 1.?Mensural data for the tarsometatarsus of modern and fossil samples of both Lagopus lagopus and L. mutus. See Table 2 for fossil site locations. Sites that have no mean and have question marks instead of a minimum value forZ,. lagopus and a maximum value fori, mutus are such because insufficient tarsometatarsal length difference was present between specimens to define the respective upper limits of L. lagopus and the lower limits of L. mutus (see Fig ure 2). (GL=greatest length; KB=shaft width; w=number of specimens.) Samples MODERN Lagopus lagopus lagopus (Scandinavia) Lagopus lagopus lagopus (Russia) Lagopus lagopus scoticus (Derbyshire, England) Lagopus lagopus scoticus (Scotland) Lagopus lagopus major (Kazakhstan) Lagopus mutus mutus (Scandinavia) Lagopus mutus millaisi (Scotland) Lagopus mutus helveticus (French Alps) Lagopus mutus islandorum (Iceland) FOSSIL Mamutowa Cave, Lagopus lagopus Mamutowa Cave, Lagopus mutus Remouchamps, Lagopus lagopus Merlin's Cave, Lagopus lagopus Merlin's Cave, Lagopus mutus La Balme-les-Grottes, Lagopus lagopus La Balme-les-Grottes, Lagopus mutus La Colombiere, Lagopus lagopus La Colombiere, Lagopus mutus Pin Hole Cave, Lagopus lagopus Pin Hole Cave, Lagopus mutus La Fage, Lagopus lagopus noaillensis La Fage, Lagopus mutus correzensis Westbury-sub-Mendip, Lagopus sp. Rebielice Krolewskie, Lagopus atavus Minimum-Maximum (n GL 37.04-42.1 (?=11) KB 2.96-3.34 (?= 11) GL 38.1-42.88 (n=6) KB 2.84-3.44 (n=5) GL 38.6-^3.58 (n= 19) KB 2.94-3.98 (?= 19) GL 38.38^14.06 (?=9) KB 3.1-3.6 (n=9) GL 45.32 (n=l) KB3.64(w=l) GL 29.44-34.08 (?=2) KB 2.66-2.74 (n=2) GL 33.22-35.9 (n=5) KB 2.88-3.12 (?=5) GL 31.8-35.7 (n= 16) KB 2.46-2.94 (K= 16) GL 30.18-34.82 (n=2) KB 2.58-2.86 (n=2) GL ?-39.84 KB? GL 32.1-? KB? GL 39.44-41.64 (?=3) KB 3.24-3.4 (n=3) GL ?-40.28 KB? GL 29.94-? KB? GL 36.36-38.8 (n= 10) KB 3-3.48 (n= 10) GL 32.24-32.3 (n=2) KB 3.2 (n=2) GL 35.2-40.6 (/i=30) KB 3.04-4.08 (?=30) GL 29.26-33.72 (n=30) KB 2.6-3.36 (w=30) GL 36.46-41.32 (n=22) KB 3.06-3.88 (?=22) GL 30.54-32.98 (?=27) KB 2.74-3.48 (n=27) GL 36.52-39.3 (n=l) KB 3.22-3.44 (?=7) GL 31.7-34.8 (n=9) KB 2.74-3.42 (n=9) GL- KB 3.43 (n=2) GL- KB3.96(?=1) Mean 38.95 3.18 40.61 3.19 41.1 3.24 41.1 3.36 - - 31.76 2.7 34.69 3 33.06 2.71 32.5 2.72 - -- - 40.19 3.3 - -- 37.75 3.22 32.27 3.2 38.21 3.4 31.39 3 38.66 3.41 31.85 3.01 38.07 3.33 33.38 3.09 - - - - L. lagopus lagopus from Scandinavia and Russia; and an indi vidual skeleton of L. lagopus major Lorenz. Lagopus mutus is represented by skeletons from Scotland {L. m. millaisi Hartert), Iceland {L. m. islandorum Faber), Scandinavia {L. m. mutus), NUMBER 89 161 and the French Alps {L. m. helveticus Thienemann). Fossil samples are from similarly scattered locations. Figure 1 gives the geographical position within Europe of the various fossil localities. An attempt also has been made to examine the spe cies through time, and the approximate ages of the samples are detailed in Table 2. A brief comment about the chronological framework is worth making because there are problems with correlating Pleistocene fossiliferous deposits across Europe that are too old for C-14 dating. The "Rissian" age quoted for the La Fage site (Chaline, 1975) is best interpreted as middle Pleistocene (oxy gen isotope stage 6 or 8) because there is no general agreement as to the correlation of the Alpine stages with the detailed oxy gen-isotope chronology derived from deep-sea cores (Shackle- ton and Opdyke, 1973). This more recently developed chronol ogy has shown the Alpine scheme to be oversimplified, and more interglacial and glacial phases are now recognized, im plying that sites described as Rissian, for example, include ones from different cold stages (Bridgland, 1994). Within the northern European scheme, the Westbury-sub-Mendip sample, from the rodent stratum (Andrews, 1990), is regarded as refer able to the early Anglian/Elsterian and probably equivalent to oxygen isotope stage 12 (A.P. Currant, The Natural History Museum, London, pers. comm., 1996). The Rebielice Krolews- kie material is late Pliocene in age according to the Mammal Neogene (MN) chronology (Mein, 1990; Bocheriski, 1991). The method of measurement for all skeletal elements of La gopus are as detailed in Kraft (1972) and were taken to the nearest 0.02 mm with slide calipers. ACKNOWLEDGMENTS I owe a great dept of gratitude to Adrian Lister, who has both helped and influenced my work. I also thank Fred Owen, Simon Parfitt, Fran Hernandez Carrasquilla, and Robert Moss for help ful discussions concerning this study. Andy Currant and Roger Jacobi provided useful information concerning the ages of sites that have not yet been fully published. Thanks also are due to FIGURE 1.?Distribution of sites containing important faunas of Lagopus mentioned in the text. (P=Pin Hole Cave; M=Merlin's Cave; W=Westbury-sub-Mendip.) 162 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 2.?Chronology and location of fossils samples (including ones discussed from literature). Site Remouchamps, Belgium Merlin's Cave, England Mamutowa Cave, Poland Pin Hole Cave, England La Balme-les-Grottes, France La Colombiere, France Gigny, France La Fage, France Westbury-sub-Mendip, England Rebielice Krolewskie, Poland Stratigraphic Position late glacial late glacial Vistulian (Upper pleniglacial) Devensian "Wurm IV" (late glacial) "Wurm IV" (late glacial) "Wurm II" "Rissian" Early Anglian (oxygen isotope stage 12) Biozone MNI6 Reference Hedges etal., 1994 Housley, 1991 Bocheriski, 1974 Jacobi, pers. comm., 1996 Mourer-Chauvire, 1975a Mourer-Chauvire, 1975a Mourer-Chauvire, 1975a Chaline, 1975 Currant, pers. comm Bocheriski, 1991 ., 1996 Age 10,330?110yrs. BP and 10,800?1I10 yrs. BP 10,020?120yrs. BP ca. 30-15 Ka ca.100-10 Ka ca. 15-10 Ka 13,390?300yrs. BP ca. 30-20 Ka ca. 300-150 Ka ca. 500-450 Ka ca. 3.6-2.4 Ma Laura Kaagan and Simeon Mellalieu for help with the final ver sion of this paper, and to Janet Stacey for much support and en couragement. The following people and institutions deserve thanks for allowing access to both modern and fossil Lagopus used in this analysis: Mark Adams, Don Smith, and Robert Prys-Jones (Subdepartment of Ornithology, Natural History Museum, Tring); Sandra Chapman and Andy Currant (Palaeon tology Department, The Natural History Museum, London); Per Ericson (Naturhistoriska Riksmuseet, Stockholm); Jon Fjeldsa (Zoologisk Museum, Copenhagen); Derek Yalden (University of Manchester); Cecile Mourer-Chauvire (Universite Claude- Bernard, Lyon); Michel Phillipe (Museum d'Histoire Naturelle de Lyon); Zygmunt Bocheriski (Institute of Systematics and Evolution of Animals, Kracow); John Nudds (Manchester City Museum); Chris Hawkes (University of Bristol Speleological Society Museum); the Musee Royaux d'Art d'Histoire, Bruxelle; and Ruth Charles (University of Oxford). Finally, I would like to acknowledge the support of the Natural Environ ment Research Council (NERC) for funding this research and to acknowledge the financial support from NERC as well as the Graduate School of University College London, which allowed me to present these results at the fourth international meeting of the Society of Avian Paleontology and Evolution. Results In Lagopus mutus, there is a noticeable difference in the mean size of the tarsometatarsi among modern subspecies in Europe (see Table 1). The Scottish subspecies {L. mutus millai si) is larger than populations from Scandinavia {L. m. mutus) or the Alps (L. m. helveticus). This difference is perhaps best seen in the length, although the bones appear to be isometric. This confirms the findings of Bocheriski (1974), although differ ences in the means calculated in our two studies exist. Kraft (1972) alleged differences between nominate L. m. mutus from Scandinavia and L. m. helveticus from the Alps, but this could not be confirmed in the present study due to small sample size. With Lagopus lagopus there appears to be greater overlap between the samples, although L. I. scoticus is slightly larger than nominate L. 1. lagopus. The single specimen of L. I. major is very much larger, although sample size prevents reliable consideration of this subspecies, which is, however, regarded as larger by ornithologists (Dement'ev and Gladkov, 1967). The most apparent difference between modern and fossil samples of both species of Lagopus is the difference in their tarsometatarsal-shaft widths (Figure 2). This is almost ubiqui tous, which is important because the modern samples are geo graphically widely spaced, and the fossils come from sites of significantly different ages as well as being widely spaced geo graphically. In addition, there is a tendency for fossils of both taxa from the last glaciation to have shorter tarsometatarsi, thus making these bones very robust (Figure 2). An exception to this pattern is provided by the study made by Mourer-Chauvire (1975a) on a sample from Gigny in France, which, although from the last (Wurm) glaciation, had relatively long tarsometa tarsi. Tarsometatarsi from La Fage (Rissian) also show this ten dency (Mourer-Chauvire, 1975a; Figure 2). Therefore, length is probably more variable among fossil populations than is shaft width. Given the consistency of greater robustness in Pleistocene tarsometatarsi of Lagopus, an explanation should be sought. Although it is conceivable that the species may have changed through evolution or replacement, both L. lagopus and L. mu tus can be traced from the Pleistocene, suggesting intraspecific adaptational changes. Research into bovid limbs and the vari ables affecting their morphology has shown that shaft width is closely correlated with body mass (Scott, 1985) because of weight-bearing constraints, so the greater robustness of the fos sil tarsometatarsi of Lagopus may reflect greater mean body weight. To test this hypothesis an attempt was made to assess the degree to which tarsometatarsal dimensions are correlated with body weight in modern Lagopus. Both species of Lagopus are included in this analysis to aug ment the size and range of samples. This approach is consid ered valid because the relationship should be strictly mechani cal and thus comparable between closely related taxa; it would seem unlikely that bones of L. lagopus and L. mutus possess significantly different mechanical properties. Figure 3 shows that there is a close, positive correlation between bird weights and their tarsometatarsal-shaft widths (correlation coefficient NUMBER 89 163 4 3.8 3.6 3.4 -- 5 j2 3.2 CO 3 2.8 2.6 2.4 5 ' 5 2 2 3 15 1 1 5 3 2 7 7 3 V145ap" * 55 r5s^r2 2 ;5 8 * 1 ?? 17 D 3 5 1 6? 6 A ? ? ? 27 29 31 33 35 37 39 Greatest length 41 43 45 ? L. lagopus scoticus - Derbyshire ? L. lagopus scoticus - Scotland A L. lagopus lagopus - Scandinavia ? L. lagopus lagopus - Russia x L lagopus maior o L. mutus millaisi - Scotland AL mutus islandorum - Iceland ? L. mutus mutus- Scandinavia ? L mutus mutus- Russia 0 L mutus helveticus- Alps 1 Pin Hole Cave 2 Merlin's Cave 3 Mamutowa Cave 4Remouchamps 5 La Colombiere 6 La Balme-les-Grottes 7L. Lagopus noaillensis - La Fage 8L. mutus correzensis- La Fage 47 FIGURE 2.?Scattergram of tarsometatarsal-shaft width versus length in Lagopus lagopus and Lagopus mutus. Includes both modem and fossil populations. Numbered symbols indicate fossil localities detailed in Table 2. (r)=0.951). Due to the positive correlation between the shaft widths and lengths, a regression also was performed for the tar sometatarsal lengths against bird weights. In this instance (see Figure 4), a positive correlation also is present (r=0.914). It may be significant that although high, the r value is lower than that produced for the tarsometatarsal-shaft width. This is not completely unexpected because similar trends were observed by Scott (1985) for bovids. Both tarsometatarsal width and length are positively correlat ed with body weight, which makes the interpretation of the more robust Pleistocene tarsometatarsi as representing larger birds less certain. Therefore, to test further the hypothesis that the Pleistocene birds were larger, various skeletal elements (co- racoids, ulnae, carpometacarpi) were considered from a similar perspective. These additional elements were not found to differ from the modern samples. Only when the humerus was subject ed to metrical analysis did a difference become apparent. Hu meri from La Colombiere and Pin Hole Cave were used for this purpose. A plot of humeral length versus humeral-shaft width merely showed that the Pleistocene birds conformed in these dimensions with the modern birds of the two species. When greatest length was plotted against proximal width, however, the fossil humeri proved to have wider proximal extremities than their modern counterparts (Figure 5). It was found that de spite the great overlap between the two species of Lagopus to day, the fossils plot as distinct clusters indicating the presence of L. lagopus and L. mutus for the larger and smaller humeri, respectively. The wider proximal extremities of the fossils are likely to be due to a need for larger areas of muscle attachment because this measurement takes in the degree of development of both the crista pectoralis and the crista bicipitalis. This in turn implies that there may have been a greater development of the pectoral and bicipital muscles that are needed for flight, supporting the hypothesis of their greater body weight. Discussion The hypothesis that both Lagopus lagopus and L. mutus were birds of greater body weight during the Pleistocene is support ed when compared with all modem populations and subspecies examined, except perhaps for L. lagopus major. Unfortunately, the exact timing of this change in mean body weight cannot be ascertained due to the lack of Holocene fossils available. If body size declined at the Pleistocene/Holocene boundary, a number of possible causes may be suggested, such as climatic change, vegetational changes, or interspecific competition. Interspecific competition in the form of character displace ment can be invoked as an explanation of size variation within 164 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 620 Vl M 570 ? 520 S 470f 420 ? L. lagopus scoticus - Derbyshire (weight from literature) ? L. lagopus scoticus - Scotland (weight from literature) * L. lagopus lagopus - Scandinavia (specimens of known weight) ? L. lagopus lagopus - Scandinavia (weight from literature) -fc L. mutus millaisi - Scotland (weight from literature) ? Lagopus mutus - Scotland? (weight from literature) -?- L. mutus mutus - Scandinavia (weight from literature) ? L. mutus helveticus - Alps (weight from literature) 2.6 2.7 2.8 2.9 3.0 3.1 KB 3.2 3.3 3.4 3.5 FIGURE 3.?Linear regression of mean weights of Lagopus versus mean shaft width (KB) of tarsometatarsi. The data plotted are based on mean weights of subspecies, taken from Cramp (1980), against mean tarsometatarsal- shaft widths and lengths of the corresponding subspecies measured in the present study. Mean weights were taken from the literature in all but one instance because in most cases skeletons in collections had no such data recorded for individual birds. The one exception was a sample of Lagopus lagopus from Scandinavia, where weights were recorded. ? L. lagopus scoticus - Derbyshire (weight from literature) ? L. lagopus scoticus - Scotland (weight from literature) ? L. lagopus lagopus - Scandinavia (specimens of known weight) ? L. lagopus lagopus - Scandinavia (weight from literature) -f( L. mutus millaisi - Scotland (weight from literature) ? Lagopus mutus - Scotland? (weight from literature) + L. mutus mutus - Scandinavia (weight from literature) - L. mutus helveticus - Alps (weight from literature) .2f s 620 570 520 470 420 ? * 31 33 35 37 GL 39 41 FIGURE 4.?Linear regression of mean weights ofLagopus versus their tarsometatarsal length. See legend to Fig ure 3 for source of data. NUMBER 89 165 18.5 -- 18 - 17.5 17 'i 16.5 16 15.5 15 ?- 14.5 14 52 5x ? ? ? 5 5 A x 5 ? A A ? ? o o ?<> ? A O ? ? ? % ? A X A a ?A ? * A ? ? A A 54 56 58 60 62 Greatest length 64 66 68 ? L. lagopus scoticus - Derbyshire ? L. lagopus scoticus - Scotland A L. lagopus lagopus - Scandinavia x L. lagopus lagopus - Russia x L. lagopus brevirostris o Lagopus mutus millaisi- Scotland o Lagopus mutus (Scotland?) A Lagopus mutus islandorum - Iceland D L mutus mutus - Scandinavia 0 L. mutus mutus - Russia A L. mutus helveticus - Alps 5 La Colombiere 1 Pin Hole Cave FIGURE 5.?ScattergTam of humerus length of Lagopus lagopus and Lagopus mutus versus proximal width. certain species. Studies such as that on the pygmy shrew Sorex minutus Linnaeus in northern Europe have shown that where two ecologically similar taxa occur in sympatry their sizes will be more divergent than when in allopatry (Malmquist, 1985). This does not appear to affect Lagopus today, and it could not affect the change in size seen through time because these changes are independent of sympatry or allopatry. Lagopus la gopus and L. mutus are presumably not ecologically similar enough for character displacement to take place. The most often-quoted hypothesis to account for change in body size during the Quaternary is that of climate and, in par ticular, temperature, which is the mechanism often invoked to account for Bergmann's Rule. Many Pleistocene mammals from glacial episodes were larger than today, and certain au thors have suggested that thermoregulation is the causal mech anism (Davis, 1981). Other paleontologists and biologists, however, have agreed that this mechanism has been applied where it may not be ap propriate, and that the subject is a much more complex one (Lister, 1992). A counterargument proposed by Guthrie (1984, 1990) and Geist (1986) is that it is not the climate that directly affects an animal's size but the consequences of the length and quality of the plant growing season, which in turn are affected by climate. The vegetational environment, called steppe-tundra or mammoth-steppe, has been described as very productive on the basis of the large herbivores it supported (Guthrie, 1990). The vegetation was a mosaic of high diversi ty, although predominated by grassland. It should be noted, however, that some palynologists have disagreed with the concept of the mammoth-steppe. They believe the vegetation was poor, a polar desert, based on the apparently low pollen influx at the time. The idea that the vegetational environment was a rich steppe-tundra has recently been expanded by Lister and Sher (1995), who have suggested that the steppe-tundra vegetation relied on a climatic regime that has vanished. They pointed out that detailed climatic records, such as studies of the Greenland ice cores, have shown that the Holocene is dis tinct from the late Pleistocene in having unusually stable con ditions. Pleistocene climatic instability may have allowed the mosaic vegetation of the steppe-tundra to persist. Once this climatic regime ceased to exist, the megafauna, which relied so heavily on the vegetation type the climate supported, changed along with it. Some animals became extinct, like the giant deer Megaloceros giganteus (Blumenbach) and the woolly rhinoceros Coelodonta antiquitatis (Blumenbach), or locally extinct, like the lion Panthera leo Linnaeus and spot ted hyena Crocuta crocuta Erxleben (Stuart, 1991). Others underwent a reduction in body size, such as the fox Vulpes vulpes Linnaeus and wild boar Sus scrofa Linnaeus (Davis, 1981). It is, therefore, an attractive hypothesis that certain 166 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY birds, such as the two species of Lagopus, which abounded in the steppe-tundra environment, also underwent changes upon its demise, such as reduced geographic ranges and body size. Interestingly, the largest subspecies of the genus, L. I. major, lives today on the steppes of Kazakhstan. In support of this idea is the fact that all the fossil popula tions examined, except that from Rebielice Krolewskie (the late Pliocene site), come from deposits considered to belong to cold phases of the Pleistocene. The oldest members of the ge nus Lagopus examined after those from Rebielice come from a cold horizon above the interglacial at Westbury-sub-Mendip in England, which is early middle Pleistocene (oxygen isotope stage 12). Unfortunately, the fossils are few in number and are fragmentary, which makes it difficult to assess to which species they belong. They do, however, possess relatively robust tar sometatarsi, so it may be that Lagopus was already adapted to the steppe-tundra and was larger in relation to today's birds. The next youngest assemblage examined in this survey is that from La Fage, which is late middle Pleistocene. Both species are definitely present, although it may be significant that they appear to be less divergent from each other in their tarsometa tarsal lengths than are modern birds (Figure 2). This may be support for Mourer-Chauvire's (1993) suggestion that the spe cies had diverged not long before. The greater areas of the crista pectoralis and crista bicipitalis may indicate that both L. lagopus and L. mutus were larger in the Pleistocene. If there were a primary selective force for large body size, so that birds were heavier, they would require great er muscle bulk to fly, which in turn adds further to body weight. Alternatively, the birds may have become larger be cause of selection for better-developed flight muscles under a different climatic regime when the birds may have been less sedentary. This hypothesis would be bolstered by the findings of Bocheriski (1974, 1985) and Bocheriski and Tomek (1994), who demonstrated that the distal-wing elements of Lagopus la gopus and L. mutus during the Pleistocene of Poland and Aus tria were relatively longer than in present-day birds, and that their legs were relatively shorter. This conclusion, however, was not confirmed by the samples analyzed in the present study. Bocheriski (1974) claimed that there was a clear, positive correlation between temperature and tarsometatarsus length, al though he pointed out that local vegetation type also was influ ential. The nature of the variability seen in tarsometatarsal length over both time and space implies that, unlike the shaft widths, local factors may have had an influence. This seems more likely than the variation being a reflection of the other thermoregulatory biogeographic rule (Allen's Rule), which would produce more uniform clines across the birds' former geographic ranges. Therefore, it may be that influences such as the local terrain are more important than the influence of tem perature because locomotion is generally regarded as important in determining leg length in mammals (Scott, 1985). Due to the small magnitude of the differences involved and the small size of the birds in relation to the ground relief, however, this con jecture is difficult to test. Conclusion Lagopus lagopus and L. mutus differ allometrically between the Pleistocene and the present, a consistent finding even when fossils from widely distributed areas and times are compared with greatly dispersed modern samples from across Europe. This difference is most readily identified in the tarsometatarsal- shaft width and implies a change that was due to a general or widespread effect, not a local adaptation similar to that pro posed for the Holocene evolution of the red grouse {L. I. scoti cus) in the British Isles (Voous, 1960; Tyrberg 1991). The changes in tarsometatarsal length appear to be reactions to vari ations in local conditions that remain unknown. The uniformity of change in tarsometatarsal width across Europe, however, implies regional rather than local effects. Global events that oc curred at the Pleistocene/Holocene boundary seem to be a like ly explanation; however, two problems arise. First, untangling cause and effect between climatic and vegetational factors, which are closely linked, and second, the possibility that a change occurred in the birds' vagility. It is perhaps easier to conceive that the two species of Lago pus changed size due to climate change and its direct effect on vegetation rather than to changes in the bird's degree of seden tariness. The birds appear to have reacted in much the same way as did many mammals that survived the Holocene/ Pleistocene boundary by becoming smaller. It is suggested herein that the change in seasonal length and vegetation type was the primary reason for this, and not temperature. The birds in the northern areas of Europe today, and particularly in moun tainous regions further south in the case of L. mutus, probably exist at much the same temperatures as they did in the Pleis tocene in southern Europe. This would eliminate temperature as the primary causal factor of size decrease in the genus over this period, because birds of comparable size to those of the Pleistocene are not present in northern Europe today. Dietary shifts caused by vegetational changes are proposed herein as the most significant factor leading to size reduction. It therefore seems reasonable to suggest that the birds in the genus Lagopus in the Palearctic today are relictual populations that originally evolved and diversified into Lagopus lagopus and L. mutus on the steppe-tundras of the late-middle and late Pleistocene. Bocheriski (1974) and Mourer-Chauvire (1975a) had previ ously demonstrated allometric trends in European Pleistocene Lagopus, but the suggestion of a major reduction in size at the Pleistocene/Holocene boundary over most of Europe is new, as is the suggestion that it was due to the vegetational changes de scribed above. This work has implications for the taxonomic use of allomet ric differences in skeletal elements. Often such differences are given greater significance than are mere size differences. The example described above clearly shows that a change in size NUMBER 89 167 may not be conferred to all elements equally. Therefore, di mensions thought to be significant should be considered in terms of their relationship to body size before taxonomic deci sions are taken. Furthermore, allometric changes have been ob served by neontologists to occur over short periods of time, for example, the increased bill size observed in Geospiza fortis Gould on the island of Daphne Major, Galapagos, as a result of strong selection through differential survival of large-seeded plants in response to a drought (Grant, 1986). These allometric changes have no taxonomic significance but simply represent organisms adapting to the complex pattern of natural selective pressures acting on any given skeletal dimension. Thus, cau tion should be used when interpreting allometric differences seen in fossils because the differences seen between Pleis tocene and Holocene Lagopus are of the type that have some times been considered to have taxonomic significance. Literature Cited Andrews, Peter 1990. Owls, Caves and Fossils: Predation, Preservation and Accumula tion of Small Mammal Bones in Caves, with Analysis of the Pleis tocene Cave Faunas from Westbury-sub-Mendip, Somerset, U.K. 231 pages. London: The Natural History Museum. Bertholt, P., and S.B. Terrill 1991. Recent Advances in Studies of Bird Migration. Annual Review of Ecology and Systematics, 22:357-378. Bocheriski, Z. 1974. Ptaki mlodszego czwartorzedu Polsk [The Birds of the Late Quater nary of Poland]. 205 pages. Warszswa-Krakow: Paristwowe Wy- dawnictwo Naukowe. 1985. Osteological Differentiation in Willow Grouse. Fortschritte der Zo ologie (Stuttgart), 30:69-72. 1991. Pliocene Grouse of the Genus Lagopus from Poland. Acta Zoologica Cracoviensia, 34(2):563-577. Bocheriski, Z., and T. Tomek 1994. Fossil and Subfossil Bird Remains from Five Austrian Caves. Acta Zoologica Cracoviensia, 37(2):347-358. Bridgland, D.R. 1994. Quarternary of the Thames. 441 pages. London: Chapman and Hall. Chaline, Jean 1975. Les rongeurs, Page et la chronologie climatique du remmplissage de l'aven de la Fage (Correze). Nouvelles Archives du Museum d'His toire Naturelle de Lyon, 13:113-117. Cramp, S., editor 1980. Handbook of the Birds of Europe, the Middle East and North Africa: The Birds of the Western Palaearctic, 2: Hawks to Bustards. viii+693 pages. Oxford: Oxford University Press. Davis, Simon, J.M. 1981. The Effects of Temperature Change and Domestication on the Body Size of Late Pleistocene to Holocene Mammals of Israel. Paleobiol ogy, 7(1):101-114. Dement'ev, G.P, and N.A. Gladkov, editors 1967. Birds of the Soviet Union. Volume 4. Jerusalem: Israel Program for Scientific Translation. Dennison, M.D., and A.J. Barker 1991. Morphometric Variability in Continental and Atlantic Island Popu lations of Chaffinches (Fringilla coelebs). Evolution, 45(l):29-39. Ericson, P.G.P. 1987. Osteology of the Eider Somateria mollissima (L.):A Study of Sexual, Geographic and Temporal Morphometric Variation in the Eider Skeleton. 142 pages. Stockholm: Statens Historiska Museum, Stat- ens Historiska Museum Studies 5. Fleischer, R.F., and R.F. Johnson 1982. Natural Selection on Body Size and Proportions in House Sparrows. Nature, 298:747-749. Geist, V. 1986. Bergmann's Rule is Invalid. Canadian Journal of Zoology, 65:1035-1038. Grant, PR. 1971. Variation in the Tarsus Length of Birds in Island and Mainland Re gions. Evolution, 25:599-614. 1986. Ecology and Evolution of Darwin's Finches. 458 pages. Princeton: Princeton University Press. Guthrie, R.D. 1984. Alaskan Megabucks, Megabulls, and Megarams; the Issue of Pleis tocene Gigantism. In H.H. Genoways and M.R. Dawson, editors, Contributions in Quaternary Vertebrate Paleontology: A Volume in Memorial to John E. Guilday. Carnegie Museum of Natural History Special Publication, 8:482-510. Pittsburgh. 1990. Frozen Fauna of the Mammoth Steppe: The Story of Blue Babe. 323 pages. London: The University of Chicago Press. Harrison, C.J.O. 1980. Pleistocene Bird Remains from Torneton Cave and the Brixham Windmill Hill Cave in South Devon. Bulletin of the British Museum (Natural History), Geology, 33(2):91-100. Hedges, R.E.M., R.A. Housley, C. Bronk Ramsey, and G.J. Klinken 1994. Radiocarbon Dates from the Oxford AMS System. Archaeometry, 36(2):337-374. Housley, Rupert A. 1991. AMS Dates from the Late Glacial and Early Postglacial in North west Europe: A Review. In N. Barton, A.J. Roberts, and D.A. Roe, editors, The Late Glacial in North-West Europe: Human Adaptation and Environmental Change at the End of the Pleistocene. Council for British Archaeology Research Report, 77:25-39. James, Frances C. 1983. Environmental Component of Morphological Differentiation in Birds. Science, 144:548-550. Janossy, D. 1974. Upper Pliocene and Lower Pleistocene Bird Remains from Poland. Acta Zoologica Cracoviensia, 19(21):531-566. 1976. Plio-Pleistocene Bird Remains from the Carpathian Basin, 1: Tetra onidae. Aquila, 82:13-34. Johnston, R.F., and R.K.. Selander 1964. House Sparrows: Rapid Evolution of Races in North America. Sci ence, 144:548-550. Kraft, E. 1972. Vergleichend morphologische Untersuchungen an Einzelknochen nord und mitteleuropa'ischer kleinerer Huhnervogel. 194 pages. Miinchen: Ludwig-Maximilians-Universitat. Lister, A.M. 1992. Mammalian Fossils and Quaternary Biostratigraphy. Quaternary Science Review, 11:329-344. 168 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Lister, Arian M., and Andree V. Sher 1995. Ice Cores and Mammoth Extinction. Nature, 378:23-24. Malmquist, M.G. 1985. Character Displacement and Biogeography of the Pygmy Shrew in Northern Europe. Ecology, 66(2):372-377. Mein, P. 1990. Updating of MN Zones. In E.H. Lindsay, V. Fahlbusch, and P. Mein, editors, European Neogene Mammal Chronology, pages 73-90. New York: Plenum Press. Mourer-Chauvire, Cecile 1975a. Les oiseaux du Pleistocene moyen et superieur de France. Docu ments des Laboratoire de Geologie de la Faculte de Sciences de Lyon, 64(2): 624 pages. 1975b. Les oiseaux (Aves) du gisement Pleistocene Moyen des Abimes da la Fage a Noailles (Correze). Nouvelles Archives du Museum d'His toire Naturelle de Lyon, 13:89-112. 1983. Les oiseaux dans les habitats paleolithiques: gibier des hommes ou proies des rapaces? In C. Grigson and J. Clutton-Brock, editors, An imals in Archaelogy, II: Shell Middens, Fishes and Birds. British Archaeological Reports, International Series (Oxford), 183: 111-124. 1993. The Pleistocene Avifaunas of Europe. In A. Morales Muniz, editor, Archaeornithology: Birds and the Archaeological Record; Proceed ings of the First Meeting of the ICAZ Bird Working Group, Madrid, October 1992. Archaeofauna, 2:53-56. Newton, E.T. 1924. Note on Bird's Bones from Merlin's Cave. Proceedings of the Uni versity of the Spelaeological Society, Bristol, 2:159-161. Potapova, OR. 1986. The Pleistocene Ptarmigans Lagopus from the Northern Urals. Pro ceedings of the Zoological Institute, Leningrad, 147:46-58. USSR Academy of Sciences. Scott, Kathleen M. 1985. Allometric Trends and Locomotor Adaptions in the Bovidae. Bulle tin of the American Museum of Natural History, 179(2): 197-288. Shackleton, N.J., and N.D. Opdyke 1973. Oxygen Isotope and Palaeomagnetic Stratigraphy of Equatorial Pa cific Core V28-238: Oxygen Isotope Temperatures and Ice Vol umes on a 105 and 106 Year Scale. Quaternary Research, 3:39-55. Stuart, A.J. 1991. Mammalian Extinctions in the Late Pleistocene of Northern Eurasia and North America. Biological Reviews, 66:453^462. Tyrberg, T. 1991. Arctic, Montane and Steppe Birds as Glacial Relicts in West Pale arctic. Ornithologische Verhandlungen, 25:29-49. Voous, K.H. 1960. Atlas of European Birds. 284 pages. London: Thomas Nelson and Sons, Limited. [Originally published as Atlas van de Europese Vo- gels, Amsterdam: Elsevier.] Zink, Robert M., and J.V. Remsen, Jr. 1986. Evolutionary Processes and Patterns of Geographic Variation in Birds. Current Ornithology, 4:1-69. A New Genus for the Incredible Teratorn (Aves: Teratornithidae) Kenneth E. Campbell, Jr., Eric Scott, and Kathleen B. Springer ABSTRACT A partial humerus of a giant flying bird from Blancan (upper Pliocene) deposits of California is determined to be a teratorn, although the humerus differs from those of other known genera of the family Teratornithidae in the position of the attachment of the M. latissimus dorsi and in the shape of the humeral shaft. The new specimen is referred to the Incredible Teratorn, Teratornis incredi- bilis Howard (1952), and a reexamination of all the specimens pre viously referred to this taxon reveals sufficient grounds to erect a new genus for this species. The size of the new partial humerus suggests that the bird had a wingspan of approximately 5 m, which is the same estimate previously given for the Incredible Teratorn. Introduction Teratorns are members of an extinct family of giant flying birds, the Teratornithidae (Miller, 1925), which currently is placed within the order Ciconiiformes (Jollie, 1976-1977; Rea, 1983; Olson, 1985; Emslie, 1988). Three genera have been rec ognized in the family: Teratornis L. Miller (1909), Cathartor- nis L. Miller (1910), and Argentavis Campbell and Tonni (1980). One species has been assigned to each of the latter two genera, whereas two species have been assigned to Teratornis: T. merriami L. Miller (1910) and T. incredibilis Howard (1952). Teratorns were the largest flying birds known, with the largest, Argentavis magnificens Campbell and Tonni (1980), reaching a wingspan of 6-8 m and a weight of 72-79 kg (Campbell and Tonni, 1980; Campbell and Marcus, 1992). Our current understanding of teratorns has been summarized by Campbell and Tonni (1980, 1982, 1983) and Campbell (1995). Kenneth E. Campbell, Jr., Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007, United States. Eric Scott and Kathleen B. Springer, San Bernardino County Museum, 2024 Orange Tree Lane, Redlands, California 92374, United States. The discovery of a partial humerus of a teratorn of a size similar to that estimated for Teratornis incredibilis allows us to clarify the status of that species, to which five specimens of widely different ages (late Pliocene to late Pleistocene) have previously been referred. The holotype of the species, an os carpi ulnare (cuneiform) (Howard, 1952), is the most diagnos tic specimen, whereas the four referred specimens are much less so. Although these specimens, namely, the proximal end of an ulna, the distal end of a radius, the fragmentary proximal end of a carpometacarpus, and the anterior portion of a beak, were not very diagnostic, they were identified as teratorns and were referred to T. incredibilis primarily on the basis of size (Howard, 1963, 1972; Emslie, 1995; Jefferson, 1995). A sixth specimen, the fragmentary distal end of a right carpometacar pus recently discovered in upper Pliocene deposits of central Mexico and described below, is assigned herein to this species. A seventh specimen, a vertebra, previously referred to T. in credibilis (Heaton, 1984) was later reassigned to T. merriami (Emslie and Heaton, 1987). This specimen was not seen by us. In spite of being rather fragmentary, the new specimen clear ly possesses characters that unite it with the teratorns. At the same time, other characters clearly distinguish it from humeri of the genus Teratornis, which suggests that it represents a ge nus distinct from Teratornis. We took the opportunity this specimen provided to reexamine all specimens referred to T. incredibilis, and we found grounds for placing all of them in a new genus. Osteological terminology is from Baumel (1993) and Howard (1980). ACKNOWLEDGMENTS.?Fossil mammals from Murrieta were identified by R.L. Reynolds, San Bernardino County Mu seum; additional identifications were provided by CA. Repen ning, United States Geological Survey. Paleontologie Resource Assessment Program (PRAP) mitigation was conducted in co operation with H. Meyers of the Kulberg Group, Temecula. We thank R. Chandler, Georgia College and State University, and O. Carranza-Castaneda, Instituto de Geologia, Universidad Na- 169 170 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY cional Autonoma de Mexico, for permission to include the par tial carpometacarpus from Mexico in this study; B. MacFad- den, Florida Museum of Natural History, for the loan of the Leisey Shell Pit specimens; and S. Emslie, S.J. Parry, and an anonymous reviewer for comments on an earlier version of this paper. Photographs are by R. Meier and D. Watson, Natural History Museum of Los Angeles County (LACM) staff pho tographers; the map in Figure 2 was produced by G.T. Braden, San Bernadino County Museum (SBCM). Systematics Order CICONIIFORMES Family TERATORNITHIDAE Aiolornis, new genus TYPE SPECIES.?Teratornis incredibilis Howard, 1952; type by original description. REFERRED SPECIES.?None. ETYMOLOGY.?Aiolos, Greek, masculine, god of the winds; ornis, Greek, masculine, bird. EMENDED DIAGNOSIS.?Placed in the family Teratornithidae and differing from genera of the Vulturidae by having the os carpi ulnare with the following characters (from Howard, 1952:51): (1) attachment for Lig. ulno-ulnocarpale long, diago nal, and ridge-like (short, almost papilla-like in the Vultu ridae); and (2) external prominence and attachment for Lig. ulno-ulnocarpale in close proximity (broad space separating the two in the Vulturidae). Differs from Teratornis by having the os carpi ulnare with (1) external prominence a long, prominent ridge forming one end of and extending from the facies articularis ulnaris to very near the attachment for Lig. ulno-ulnocarpale (short, rounded, not contacting facies articularis ulnaris in Teratornis); (2) at tachment for Lig. ulno-ulnocarpale proportionately longer, more prominently protruding from body of bone; (3) facies ar ticularis ulnaris narrowing slightly distad, more concave, with dorsal rim notably lower than ventral rim (narrows abruptly, less concave, with dorsal rim only slightly lower than ventral rim in Teratornis); (4) facies articularis metacarpals a slightly elongated oval, with long axis nearly aligned with long axis of facies articularis ulnaris, with external end very near facies ar ticularis ulnaris, and lying at greater angle to facies articularis ulnaris (markedly elongated oval with long axis at low angle to that of facies articularis ulnaris and external end at some dis tance from facies articularis ulnaris in Teratornis); and (5) with ventral surface of bone bordering and surrounding incisura metacarpalis nearly flat (markedly concave in Teratornis). Aiolornis incredibilis (Howard, 1952), new combination HOLOTYPE.?Complete right os carpi ulnare, LACM(CIT) 5067. TYPE LOCALITY.?Section 7-F-310 of LACM(CIT) locality 251, Smith Creek Cave, Snake Range, 54.4 km north of Baker, White Pine County, Nevada. AGE.?Rancholabrean NALMA (North American Land Mammal Age). DIAGNOSIS.?As for genus. REFERRED MATERIAL.?Left humerus: Proximal end and portion of shaft, missing caput and much of crista deltopectora- lis (Figure 1); SBCM A2239-2829, Section of Earth Sciences. The specimen was collected by Quintin Lake (PRAP), April 1993, from locality SBCM 05.006.399, which is located ap proximately 1 km northeast of Murrieta, Riverside County, California, at an approximate elevation of 368 m. The locality lies within the unsurveyed Temecula Land Grant (Figure 2), within the SWV4, NEtt, SWA section 9, T. 7S, R. 3W, San Ber nardino Base and Meridian. The specimen was exposed in situ in an erosional gully approximately 0.15 m below grade. The partial humerus came from sediments of an unnamed sandstone and conglomerate formation that unconformably un derlies the middle Pleistocene Pauba Formation and may un conformably overlie the lower Pliocene Temecula Arkose in the Elsinore Fault Zone (Kennedy, 1977). Two faunal compo nents have been recognized from the unnamed sandstone for mation, one dating to the late Blancan NALMA (late Pliocene, as interpreted by Lundelius et al., 1987) and the other to the Irvingtonian NALMA (early Pleistocene) (Scott and Cox, 1993). The partial humerus is derived from sediments that are late Blancan in age (pre-Olduvai subchron, 2.6-1.9 Ma), based on sites producing Blancan faunas in the area immediately ad jacent to SBCM 05.006.399. These localities (SBCM 05.006.156, 05.006.157, 05.006.158, 05.006.159, 05.006.397) are all located within 15 m of SBCM 05.006.399 and have yielded Hypolagus sp., Prodipodomys sp., Mimomys {Ophi- omys) parvus R. Wilson, and Sigmodon minor Gidley, but they lack any Pleistocene or later indicator taxa, such as Microtus sp. or Mammuthus sp. (Scott and Cox, 1993). The partial humerus is placed in the family Teratornithidae based on the following characters: (1) crista bicipitalis elongat ed and prominently bulbous for entire length; (2) planus inter- ruberculare smooth, broad, and symmetrically and deeply con vex in transverse section from caput to middle of crista deltopectoralis; (3) crista deltopectoralis low, very stout, curv ing ventrad on facies anterioris of corpus humeri just distal to end of crista bicipitalis, and tapering off gradually as it crosses to near midline of corpus humeri; and (4) distal insertion for M. pectoralis long, broad, and near midline of facies anterioris of corpus humeri. Other avian taxa, such as some of the orders Procellari- formes, Pelecaniformes, and Ciconiiformes, resemble teratorns somewhat in having a bulbous crista bicipitalis. In these taxa, however, the bulbous portion is more limited, being present only at the distal end of the crista bicipitalis. In Teratornis the bulbous portion occupies the entire length of the crista bicipi talis and is more pronounced; it is missing from the only known humerus of Argentavis. Although most of the bulbous NUMBER 89 171 M FIGURE 1.?New partial humerus views. The position of the line of bar=5 cm. referred to Aiolornis incredibilis: A, anterior, B, dorsal, and c, posterior insertion of M. latissimus dorsi (arrows) is unique among teratorns. Scale portion in the partial humerus under discussion is missing, enough of the base is present to document its presence and its size. Similarly, some other avian taxa have a proportionately broad and convex planus intertuberculare, but we know of none that approaches the symmetrical, smooth, deep convexity seen in teratorns. We know of no other group of birds in which the crista deltopectoralis curves ventrad onto the facies anterioris of the corpus humeri. Among the New World vultures, condors share the trait of having a large distal insertion for M. pectora- lis, but it is much smaller, more oval or tear-drop shaped, and positioned closer to the facies dorsalis of the humerus. The partial humerus differs from Argentavis and Teratornis, the only two of the three genera of the family for which the hu merus is known, by the following: (1) facies dorsalis fairly flat for length of attachment of M. latissimus dorsi, becoming slightly convex near its distal end (sloping anteriad proximally, changing to convex distally in Argentavis and Teratornis); (2) facies posterioris and facies dorsalis meet at near right angle, with line of insertion of M. latissimus dorsi following a well- defined "corner" of margo anteriodorsalis (line of insertion at angle to margo dorsalis, crossing facies posterioris from distal to proximal in Argentavis and Teratornis); (3) corpus humeri in 172 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 2.?Map showing location of locality SBCM 05.006.399 for the new partial humerus referred to Aiolornis incredibilis. Dashed line indicates unsur- veyed boundary of Temecula Land Grant. transverse section through nutrient foramen on margo ventralis at midshaft a flattened oval not quite twice as wide as deep (more egg-shaped, with margo ventralis slightly narrower than margo dorsalis in Argentavis; nearly round in Teratornis) (Fig ure 3); (4) corpus humeri less curved proximally in both hori zontal and vertical planes; and (5) attachment of M. pectoralis with distal portion a broad, flattened area (area damaged but appears similar in Argentavis; narrow, steeply sloping on distal end of crista deltopectoralis in Teratornis). These characters hold in comparison with all humeri of Teratornis from Rancho La Brea, California (n=>50). Although the position of the M. latissimus dorsi is so mark edly different from that seen in known teratorns, even to the oooO ABC D FIGURE 3.?Comparison of the size and shape of the cross section of the humeral shaft in the plane of the nutrient foramen in a vulturid (A) and teratorns (B-D): A, California Condor, Gymnogyps californianus; B, Merriam's Teratorn, Teratornis merriami; C, Aiolornis incredibilis; D, Argentavis magnificens. Scale bar=2 cm. point of leading one to question whether or not the humerus might even be from a new, unknown family, in the absence of additional characters the presence of the four teratorn autapo morphies listed above are sufficient grounds for placing the specimen with the teratorns. Measurements of this specimen are limited to shaft width and depth in the plane of the midshaft nutrient foramen on the facies ventralis: 37.8 mm and 22.0?1 mm, respectively. Com parable measurements for the single specimen of Argentavis are 52.2 mm and 32.8 mm, respectively, and measurements for Teratornis merriami {n=\6) are 23.2-28.9 mm (x=24.9 mm) and 18.3-23.0 mm (JC = 19.9 mm), respectively. Right radius: Distal end, Anza-Borrego Desert State Park Paleontological Collection, ABDSP(LACM) 1318/V3803, from Irvingtonian (middle Pleistocene, or 1.5-1.0 Ma (Savage and Curtis, 1970:223)) deposits at the 3600-ft level of the Vallecito Creek area, Anza-Borrego Desert State Park, referred to Tera tornis incredibilis by Howard (1963). Emended description, in comparison with T merriami, as follows: distal contour straight- er, less rounded at corners; ligamental prominence extending more proximad, significantly elevated above and more markedly set off from shaft and more elevated than central ligamental at tachment in anterior view (elevational differences slightly mag nified by loss of surface bone on central ligamental attachment; both areas of attachment at about same level in Teratornis) (bone slightly crushed along proximal edge of prominence, but this cannot account for sharp drop to shaft); shaft immediately above distal end flat in anterior view; internal edge of shaft prox imal to internal end of scapholunar facet broad, channeled, or slightly convex transversely, which gives way to knife-like ridge proximad, with interno-anterior edge formed by long, prominent attachment of Lig. radioulnare interosseum +Lig. ulno-radiocar- pale (edge very narrow in Teratornis, sloping steeply mediad with attachment of Lig. radioulnare interosseum+Lig. ulno-ra- diocarpale a broad, flat area halfway between edge and midline of bone that is slightly elevated proximally); depression occurs between attachment of Lig. radioulnare interosseum + Lig. ulno- radiocarpale and midline of bone (accentuated, but not caused by crushing) (absent in Teratornis); central ligamental prominence more elevated and extending closer to distal end before dropping off than in Teratornis; external edge of shaft proximal to liga mental prominence with linear convexity; and shaft significantly more curved in anterior view and slightly more curved in exter nal view (some of the curvature, but not much, may be a result of distortion in preservation). Premaxillary: Anza-Borrego Desert State Park Paleonto logical Collection, ABDSP(LACM) 6747/V26697, from Blan can (late Pliocene, or 3.5-3.2 Ma (Savage and Curtis, 1970:223; Evernden et al., 1964:164)) deposits at the 7000-8000-ft level of the Fish Creek beds in the Anza-Borrego Desert State Park, referred to Teratornis incredibilis by Howard (1972). Emended description as follows: markedly compressed, deep beak; palatal surface fairly flat (more concave than in Terator- NUMBER 89 173 nis but not highly vaulted as in vulturids), with prominent cen tral ridge, or septum (which is only slightly developed anterior ly in Teratornis), but lacking distinct median groove posterior to ridge (as present in Teratornis); crista tomialis wider in transverse section and more deeply grooved and more symmet rical anteriorly than in Teratornis but forming comparable en closure of deep, narrow channel in anteriormost portion of beak, in region of sharp curvature. In teratorns, the crista tomialis forms a sharp ridge external to the region of the anterior narial border, but farther anterior, about halfway to tip of beak, this ridge declines in prominence, and the median portion of the grooved crista tomialis forms a sharper ridge. The point of this transition represents the posteri- ormost portion of this specimen of beak that is preserved, which suggests that the specimen came from a bird with a pro portionately much deeper beak than seen in Teratornis. Left ulna: Proximal end and partial proximal shaft (Figure 4), Anza-Borrego Desert State Park Paleontological Collec tion, ABDSP (IVCM) 519/5660, from Vallecito member of the Palm Spring Formation (Woodward, 1963), June Wash area of the Vallecito-Fish Creek Basin. Specimen from Val lecito Creek local fauna (Jefferson, 1995), dating to about 1.8-0.9 Ma. This ulna is severely crushed, but enough of the original structure remains to determine that it differs from that of Tera tornis by having (1) olecranon long and very broad; (2) cotyla ventralis lying at a much steeper angle to the long axis of the shaft; and (3) cotyla dorsalis wider and relatively shallow. Al though the specimen is quite crushed, most of the bone of the proximal end is present so it is possible to approximate closely its original shape. The cotyla ventralis may be slightly rotated as a result of crushing, but the original shape, in ventral view, is clearly preserved. The amount of bone present suggests that the proximal end was less deep anterioposteriorly immediately distal to the cotyla ventralis, and the tuberculum Lig. collat. ventralis was much less bulbous, than in Teratornis. In addi tion, it appears that the olecranon was much more compressed anteroposteriorly, although some post-mortem crushing has taken place. A total of about 26 cm of the ulna is preserved, in four piec es, although only two of the four pieces actually fit together. Two of the shaft fragments have feather papillae, but the small est does not. The small diameter of the smallest fragment, which has no recognizable features, suggests that it may not even be part of the ulna, an idea supported by the presence of a fifth fragment that is unidentifiable to element, but which is definitely not part of an ulna. Right carpometacarpus: Fragmentary distal end, Univer sidad Nacional Autonoma de Mexico, Instituto de Geologia, Museo de Paleontologia, IGCU-6133, from locality GTO.31, Blancan, Rancho Viejo area, State of Gaunajuato, Mexico, about 240 km NNW of Mexico, Distrito Federal. This specimen is too fragmentary to provide much informa tion, but it is clearly a teratorn. It differs from carpometacarpi of Teratornis by having (1) the facies articularis digitalis major with the medial rise more elongated and elevated and more dis tinctly set off from the anterior portion; and (2) the os metacar pale minoris with the distal area of fusion with the os metacar pale majus proportionately shorter than in Teratornis and the distal end more massive, projecting distad more distinctly and at a greater angle from the os metacarpale majus. The single known partial carpometacarpus of Argentavis is too poorly pre served for comparison. For further details on the fauna associ- B FIGURE 4.?A,B, proximal end of the ulna referred to Aiolornis incredibilis (ABDSP(IVCM) 519/5660): A, anterior view; B, ventral view. C,D, ulna of Ter atornis merriami (George C. Page Museum, B124): c, anterior view; D, ventral view. Note the differences between the olecranons and the orientation of the cotylae ventralis. Scale bar=3 cm. 174 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ated with this specimen, see Miller and Carranza-Castaneda (1984). Right carpometacarpus: Fragmentary proximal end, in cluding only trochlea carpalis and area immediately distal to it, Florida Museum of Natural History, UF 123874, Leisey Shell Pit 3, Hillsborough County, Florida, Bermont Formation, Irv- ingtonian (between 1.66 and 1.4 Ma (MacFadden, 1995)). This specimen was originally described and figured by Emslie (1995:316-317), who referred it to Teratornis sp. cf. T. incredibilis on the basis of its similarity to Teratornis merriami and its size (-40% larger than T. merriami). We see no reason to doubt that this specimen is a teratorn, but little else can be said about it. Discussion Howard (1952) described Teratornis incredibilis on the basis of the complete os carpi ulnare (cuneiform) noted above, allo cating the species to Teratornis on the basis of its general simi larity to T. merriami. The much greater size of the specimen (43% larger than the same bone of T. merriami) and other char acters clearly established its distinction as a separate species (Howard, 1952). Howard (1963, 1972) also referred to Terator nis incredibilis the distal end of a radius and the anterior por tion of a beak discussed above on the basis of their general sim ilarity to T. merriami and the fact that both of these specimens were about 40% larger than comparable bones of the latter spe cies. Although Howard (1972:343) considered the possibility of generic separation of the larger species from Teratornis, she considered all three specimens she had assigned to T. incredibi lis too undiagnostic or too poorly preserved to justify establish ing a new genus. After restudying the holotype and three specimens previous ly referred to Teratornis incredibilis we concluded that the dif ferences seen between T. incredibilis and T. merriami are such that the erection of a new genus for the former was warranted. In reexamining the holotypical os carpi ulnare we found that the characters noted above for Aiolornis differ from those of Teratornis at a level comparable to the differences seen among the os carpi ulnare of the living genera of the family Vultu ridae, which are easily identified to genus, and the differences hold for all specimens of the element from Rancho La Brea ex amined (n=24). The unique characters of the humerus, ulna, ra dius, and carpometacarpus assigned to Aiolornis incredibilis, however, suggest that that species may have had different flight adaptations from Teratornis merriami. Because the os carpi ulnare is a bone integral to the flight of any bird, we would have expected greater character differences to be reflected in the os carpi ulnare of Aiolornis, which would serve to separate it more readily from Teratornis. For this reason we would not be surprised to find in the future that the older (Blancan and Irvingtonian) specimens herein referred to Aiolornis incredibi lis actually are referrable to another genus and species. The other genus and species of teratorn, Cathartomis gracil is, is known only from tarsometatarsi, one complete and one distal end, from the late Pleistocene asphalt deposits at Rancho La Brea, California. Although the length of the holotypical tar sometatarsus of C. gracilis falls within the size range of tar sometatarsi of Teratornis merriami, it is much more slender and has several features that distinguish it from its more heavi ly built contemporary. Miller and Howard (1938:169) reevalu ated the status of Cathartomis and concluded "that Cathartor- nis is markedly similar to Teratornis merriami, though it is undoubtedly a distinct species. We consider it also to be gener ically distinct." Some of the characters listed as separating the two genera are not particularly convincing, however, and with more specimens of teratorns available now than at the time of Miller and Howard's study, the case for maintaining Cathar tomis as a separate genus is weak. Resolution of the status of Cathartomis is deferred pending completion of the studies of the teratorns of Rancho La Brea by KEC. Given the marked similarity of Cathartomis to Teratornis, we considered assign ment to Cathartomis of the specimens now referred to Aiolor nis inappropriate. The size of the Incredible Teratorn, Aiolornis incredibilis, re mains its most remarkable known feature, even though we know there was at least one larger species of teratorn. Howard (1952) estimated the size of A. incredibilis to be 43% larger than Teratornis merriami based on the holotypical os carpi ulnare, which would give a wingspan of about 5 m. She esti mated (Howard, 1972:343) the radius she referred to A. incred ibilis to be "approximately 40% broader than a large radius of T. merriami." The beak she referred to A. incredibilis was esti mated to be "43% larger than the largest of four measurable specimens of T. merriami,'" based on what Howard (1972:343) considered the best available measurement. The ulna that Jef ferson (1995:94) referred to A. incredibilis was said to be "about 57% larger than the average (39.6) of five measured specimens of Teratornis merriami from Rancho La Brea." Un fortunately, it is not possible to draw a direct size comparison between the partial humerus of A. incredibilis and humeri of T. merriami. The only accurate measurements that can be taken from the holotype of A. incredibilis are the width and depth near midshaft, and the width to depth proportions of the shaft are so different among the known genera of teratorns (Figure 3) as to make such a size comparison meaningless. Neither is it possible to determine accurately the size of the bird from which came the carpometacarpi referred to A. incredibilis because of the poor state of preservation of those specimens. A reasonable "eyeball" estimate, however, suggests that both of these speci mens came from a bird with a wingspan intermediate between that of T merriami (3.5-4 m) and Argentavis magnificens (6-8 m). This gives a wingspan estimate of 5.0-5.5 m for A. incred ibilis, which conforms with that estimated from the other spec imens referred to this species. NUMBER 89 175 Literature Cited Baumel, Julian J., editor 1993. Handbook of Avian Anatomy: Nomina Anatomica Avium. Second edition, xxiv+779 pages. Cambridge, Massachusetts: Nuttal Orni thological Club. Campbell, Kenneth E., Jr. 1995. Additional Specimens of the Giant Teratorn, Argentavis magnifi- cens, from Argentina (Aves: Teratornithidae). Courier Forschungs- institut Senckenberg, 181:199-201. Campbell, Kenneth E., Jr., and Leslie Marcus 1992. The Relationship of Hindlimb Bone Dimensions to Body Weight in Birds. In K.E. Campbell, editor, Papers in Avian Paleontology Hon oring Pierce Brodkorb. Science Series. Natural History Museum of Los Angeles County. 36:395-412. Campbell, Kenneth E., Jr., and Eduardo P. Tonni 1980. A New Genus of Teratorn from the Huayquerian of Argentina (Aves: Teratornithidae). Natural History Museum of Los Angeles County, Contributions in Science, 330:59-68. 1982. Preliminary Observations on the Paleobiology and Evolution of Ter atorns (Aves: Teratornithidae). Journal of Vertebrate Paleontology, l(3^):265-272. 1983. Size and Locomotion in Teratorns (Aves: Teratornithidae). Auk, 100:390-^03. Emslie, Steven D. 1988. The Fossil History and Phylogenetic Relationships of Condors (Ciconiiformes: Vulturidae) in the New World. Journal of Verte brate Paleontology, 8(2):212-228. 1995. An Early Irvingtonian Avifauna from Leisey Shell Pits, Hillsbor ough County, Florida. In R.C. Hulbert, Jr., G.S. Morgan, and S.D. Webb, editors, Paleontology and Geology of the Leisey Shell Pits, Early Pleistocene of Florida. Bulletin of the Florida Museum of Nat ural History, 37(l):299-344. Emslie, Steven D., and Timothy H. Heaton 1987. The Late Pleistocene Avifauna of Crystal Ball Cave, Utah. Journal of the Arizona-Nevada Academy of Science, 21:53-60. Evemden, J.F, D.E. Savage, G.H. Curtis, and G.T. James 1964. Potassium-Argon Dates and the Cenozoic Mammalian Chronology of North America. American Journal of Science, 262:145-198. Heaton, Timothy H. 1984. Preliminary Report on the Quaternary Vertebrate Fossils from Crys tal Ball Cave, Millard County, Utah. Current Research in the Pleis tocene, 1:65-67. Howard, Hildegarde 1952. The Prehistoric Avifauna of Smith Creek Cave, Nevada, with a De scription of a New Gigantic Raptor. Bulletin of the Southern Cali fornia Academy of Sciences, 51:50?54. 1963. Fossil Birds from the Anza-Borrego Desert. Natural History Mu seum of Los Angeles County, Contributions in Science, 75:1-33. 1972. The Incredible Teratorn Again. Condor, 74(3):341-344. 1980. Illustrations of Avian Osteology Taken from "The Avifauna of Em eryville Shellmound." Natural History Museum of Los Angeles County, Contributions in Science, 330:xxvii-xxxviii. Jefferson, George T. 1995. An Additional Avian Specimen Referable to Teratornis incredibilis from the Early Irvingtonian, Vallecito-Fish Creek Basin, Anza-Bor rego Desert State Park, California. In P. Remeika and A. Sturz, edi tors, Paleontology and Geology of the Western Salton Trough Detachment, Anza-Borrego Desert State Park, California. Field Trip Guidebook and Volume for the 1995 San Diego Association of Geol ogist 's Field Trip to Anza-Borrego Desert State Park, 1:94?96. Jollie, Malcom 1976-1977. A Contribution to the Morphology and Phylogeny of the Fal coniformes. Evolutionary Theory, 1:285-298; 2:115-300; 3: 1-142. Kennedy, M.P. 1977. Recency and Character of Faulting along the Elsinore Fault Zone in Southern Riverside County, California. California Division of Mines and Geology, Special Report, 131:1-12. Lundelius, E.L., Jr., T. Downs, E.H. Lindsay, H.A. Semken, R.J. Zakrzewski, CS. Churcher, CR. Harington, G.E. Schultz, and S. David Webb 1987. The North American Quaternary Sequence. In M.O. Woodbume, editor, Cenozoic Mammals of North America: Geochronology and Biostratigraphy, pages 211-235. Berkeley: University of California Press. MacFadden, Bruce J. 1995. Magnetic Polarity Stratigraphy and Correlation of the Leisey Shell Pits, Tampa Bay, Hillsborough County, Florida. In R.C. Hulbert, Jr., G.S. Morgan, and S.D. Webb, editors, Paleontology and Geology of the Leisey Shell Pits, Early Pleistocene of Florida. Bulletin of the Florida Museum of Natural History, 37( 1): 107-116. Miller, Loye 1909. Teratornis, a New Avian Genus from Rancho La Brea. University of California Publications, Bulletin of the Department of Geology, 5:305-317. 1910. The Condor-like Vultures of Rancho La Brea. University of Califor nia Publications, Bulletin of the Department of Geology, 6: 1-19. 1925. The Birds of Rancho La Brea. Publications, Carnegie Institute of Washington, 349:63-106. Miller, Loye, and Hildegarde Howard 1938. The Status of the Extinct Condor-like Birds of the Rancho La Brea Pleistocene. Publications of the University of California at Los An geles in Biological Sciences, 1:169-176. Miller, Wade, and Oscar Carranza-Castafieda 1984. Late Cenozoic Mammals from Central Mexico. Journal of Verte brate Paleontology, 4(2):213-236. Olson, Storrs 1985. The Fossil Record of Birds. In D.S. Farmer, J.R. King, and K.C Parkes, editors, Avian Biology, 8:79-256. New York: Academic Press. Rea, Amadeo M. 1983. Cathartid Affinities: A Brief Overview. In S.R. Wilbur and J.A. Jackson, editors, Vulture Biology and Management, pages 26-54. Berkeley: University of California Press. Savage, D.E., and G.H. Curtis 1970. The Villafranchian Stage-Age and Radiometric Dating. Geological Society of America, Special Papers, 124:207-231. Scott, Eric, and Shelly M. Cox 1993. Arctodus simus (Cope, 1879) from Riverside County, California. PaleoBios, 15(2):27-36. Woodard, G.D. 1963. The Cenozoic Stratigraphy of the Western Colorado Desert, San Di ego and Imperial Counties, Southern California. 223 pages. Doc toral dissertation, University of California, Berkeley. The Fossil Record of Condors (Ciconiiformes: Vulturidae) in Argentina Claudia P. Tambussi and Jorge I. Noriega ABSTRACT At present, the fossil record indicates that condors probably originated in North America, and their fossil history in South America has been traced to the early?-middle Pliocene (Montehermosan?-Chapadmalalan Age) of the Pampean region (Argentina). The great diversity of condors that occurred in the late Cenozoic of this region comprises three genera and at least four species, namely, Dryornis pampeanus Moreno and Mercerat, Vultur gryphus Linnaeus, Geronogyps reliquus Campbell, and an indeterminate vulturid probably belonging to a new genus and spe cies. The presence of Geronogyps reliquus (up to now restricted to the Pleistocene of Peru) in the Pleistocene sediments of the Pam pean region extends considerably the geographic range of the spe cies. Introduction New World vultures (Vulturidae=Cathartidae auct.) are widely distributed in the Americas, ranging from Canada to Tierra del Fuego in Argentina, and are most diverse in tropical regions of South America. All species are adapted to feeding on carrion, but they present different and exceptional special izations in their habits of scavenging (Hertel, 1992, 1994). New World vultures are closely related in habits and appear ance to the Old World vultures (Accipitridae: Aegypiinae and Gypaetinae) due to convergent evolution. Vulturids show close phylogenetic relationships with the ciconiid storks (Ligon, 1967; Rea, 1983; Olson, 1985; Emslie, 1988a; Sibley and Ahl- quist, 1990), whereas the Old World Vultures show close affin ities to hawks and eagles (Brown and Amadon, 1968). Some analyses, however, do not agree with moving the Vulturidae Claudia P. Tambussi, CONICET, Departamento Cientifico Paleon- tologia Vertebrados, Museo de La Plata, Paseo del Bosque s/nro, 1900 La Plata, Argentina. Jorge I. Noriega, CONICET, Centro de In vestigaciones Cientificas y de Transferencia Technologica a la Pro- duccion, Materi y Espaha, 3105 Diamante, Entre Rios, Argentina. from the Falconiformes to the Ciconiiformes (Griffiths, 1994, and the literature cited therein). We do not discuss this contro versy but instead adopt the view that condors are Ciconii formes. Vulturids are represented by seven living species (Sibley and Monroe, 1990; nomenclature for species' binomials of modern birds mentioned herein follows Sibley and Monroe, 1990). The two largest species, the condors, are characterized by having long, broad wings and short tails and by certain osteological features of the cranium (Emslie, 1988a; Hertel, 1992) and in clude the California Condor, Gymnogyps californianus, and the Andean Condor, Vultur gryphus, from North America and South America, respectively. The earliest fossil vulture with affinities to condors is Hadrogyps aigialeus Emslie from the middle Miocene of Cali fornia. Two other pre-Pleistocene unequivocal condors were described from the late Miocene of Florida {Pliogyps charon Becker) and the middle Pliocene of Kansas {Pliogyps fisheri Tordoff). Based on the geographic distribution of recent and fossil species, Emslie (1988a, 1988b) suggested that condors probably originated in North America and may have partici pated in the Great American Biotic Interchange in the ear ly-middle Pliocene. In accordance with the latter hypothesis, condors in Argentina are restricted to the Pliocene to late Pleistocene of the Pampean region, far from the modern area of distribution of the recent species, with the earliest fossil his tory being the early?-middle Pliocene (Montehermo- san?-Chapamalalan Age) of Argentina, with the presence of Dryornis pampeanus Moreno and Mercerat. The purpose of the present study is to summarize the diversi ty of the previously described condors from Argentina and to describe new specimens. MATERIALS AND METHODS Specimens of fossil and extant vultures were examinated at the Division of Birds, Department of Vertebrate Zoology, Na tional Museum of Natural History, Smithsonian Institution, Washington, D.C, U.S.; Department of Ornithology, American 177 178 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Museum of Natural History, New York, New York, U.S; De partamento Cientifico Paleontologia Vertebrados and Departa mento Cientifico Zoologia Vertebrados of Museo de La Plata (MLP), La Plata, Argentina; Museo Municipal de Ciencias Nat urales de Monte Hermoso (MMH), Monte Hermoso, Argentina; Casa de Cultura de Medanos (CCM), Municipalidad de Villari- no, Argentina; Seccion Ornitologfa and Seccion Paleontologia Vertebrados, Museo Bernardino Rivadavia, Buenos Aires, Ar gentina, and Royal Ontario Museum (ROM), Toronto, Canada. The comparative material included skeletons of the following living vultures (number of specimens in parentheses): Vultur gryphus (8), Gymnogyps califomianus (5), Sarcoramphus papa (1), Coragyps atratus (3), Cathartes aura (3), Cathartes bur- rovianus (1) and the ciconiids Ciconia maguari (2), Mycteria americana (2), and Jabiru mycteria (1). Fossil specimens exam ined are discussed below; they included Geronogyps reliquus Campbell, Gymnogyps howardae Campbell, and Dryornis pam peanus Moreno and Mercerat. Comparisons were made with original material except Gymnogyps kofordi Emslie, for which illustrations and published descriptions were used. All measure ments were taken with vernier calipers to the nearest 0.1 mm and are given in millimeters. Anatomical terminology follows mainly Baumel and Witmer (1993) but also Fisher (1946). ACKNOWLEDGMENTS We acknowledge Storrs Olson and Helen James for making it possible to attend the fourth international meeting of the So ciety for Avian Paleontology and Evolution and for their kind attention in Washington, D.C. For supporting our visit, we are indebted to the Office of Fellowships and Grants of the Smith sonian Institution. One of us (CPT) thanks Gerry and Gina De Iulis for all their assistance during her visit to the ROM. Final ly, we acknowledge the comments on and suggestions to the manuscript given by Steven Emslie, Kenneth Campbell, Storrs Olson, and an anonymous reviewer. Systematic Paleontology Order CICONIIFORMES Family VULTURIDAE The fossil record of condors in the Pliocene-Pleistocene of the Pampean region (Figures 1, 2) comprises five taxa, which are discussed below. Dryornis pampeanus Moreno and Mercerat, 1891 LECTOTYPE.?Distal end of right humerus, MLP 20-169 (Figure 3). LOCALITY.?Farola de Monte Hermoso (39?S, 61?50'W), Coronel Rosales County, Buenos Aires Province. HORIZON AND AGE.?Early?-middle Pliocene (Monteher- mosan?-Chapadmalalan Age; see Cione and Tonni, 1995, for detailed stratigraphic analysis). 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 HOLOCEN LU Z LU O O I- (D UJ LU Z LU o o LU z LU o o SOUTH AMERICAN STAGE PLATAN LU J AN IAN ENSENADAN SANANDRESAN VOROHUEN BARRANCALOBAN Q < i o UPPER LOWER MONTEHERMOSAN HUAYQUERIAN FIGURE 1.?Chronostratigraphic units of the upper Cenozoic of South America. NUMBER 89 179 FIGURE 2.?Geographic location of fossiliferous localities in the Pampean region of Argentina: 1, 2, Farola de Monte Hermoso; 3, Cascada Grande on Quequen Salado River; 4, south of the Loberia stream; 5, Chasico stream; 6, Monte Hermoso City. REMARKS.?Moreno and Mercerat (1891) originally de scribed Dryornis pampeanus based on a distal end of a left hu merus (lectotype, MLP 20-169) and an abraded distal end of a right femur (MLP 20-170). As was pointed out previously by Patterson and Kraglievich (1960) and by Brodkorb (1967), the assignment of the femur (MLP 20-170) to a condor was errone ous. Our reexamination of this specimen agrees with prior revi sions in attributing it to a phorusrhacoid bird. Dryornis pampeanus is recognized by having a humerus with the following characters: (1) size similar to that of Vultur gryphus, but distal shaft width just proximal to the epicondylus dorsalis greater; (2) condylus dorsalis longer and straighter than in Vultur, Gymnogyps, or Geronogyps; (3) condylus ven tralis narrower than in Vultur or Gymnogyps, similar to Gero nogyps; slightly rotated anteriad, giving moderately flexed dis tal end as in Geronogyps, Gymnogyps, and Vultur, with marked flexion of distal end of humerus; (4) epicondylus dorsalis rounded as in Gymnogyps and less protrudent proximally than in Vultur, Gymnogyps, or Geronogyps; (5) insertions of M. bra- chialis anticus, M. pronator brevis (=M. p. superficialis), and M. flexor carpi ulnaris deeper than in Gymnogyps or Gerono gyps; (6) fossa M. brachialis less extended laterally than in all species compared; (7) sulcus scapulotricipitalis less marked than in Geronogyps or Vultur; (8) fossa olecrani relatively more excavated and proximal, and less extended laterally, than in Vultur, Gymnogyps, or Geronogyps; and (9) intercondylar furrow nearly straight. Historically there has been some agreement that Dryornis pampeanus is a valid genus within the Vulturidae (see Tonni, 1980; Emslie, 1988a). Our direct comparison of the material with other known fossil and living vultures supports the ana tomical distinctness of Dryornis pampeanus from the other members of the family. Thus, based on the set of characters listed above, we consider Dryornis pampeanus to be a valid ge nus and species and to be one of the earliest fossil condors in South America. Vultur gryphus Linnaeus, 1758 REFERRED MATERIAL.?Proximal end of humerus, MLP 48- XII-16-225 (Figure 4); Farola de Monte Hermoso (39?S, 61?50'W), Coronel Rosales County, Buenos Aires Province; early-middle Pliocence (Montehermosan?-Chapadmalalan Age. Shaft of Ulna, MLP 63-VI-10-15; Cascada Grande locality on the right margin of Quequen Salado River (38?30'S, 60?30'W), Coronel Dorrego County, Buenos Aires Province; middle Pliocene (Lower Chapadmalalan Age). REMARKS.?The humerus, which comes from the same lo cality as Dryornis pampeanus, was attributed to Vultur gryphus by Tambussi (1989), Tambussi et al. (1993), and Tambussi and Noriega (1996). The ulna shaft was referred to Vultur gryphus by Tambussi (1989) and shows 12 papillae remigiales ventrales and caudales (of secondary remiges). The cross section of the shaft is triangular and forms a smooth, sigmoid curve. cf. Vultur sp. REFERRED MATERIAL.?Distal end of femur, MMH 561; seashore close to Monte Hermoso city (39?S, 61?15'W); late Pleistocene (Lujanian Age). REMARKS.?Tonni (1984) pointed out that this femur is in distinguishable from that of Vultur, but it is too badly pre served for a more accurate identification. VULTURIDAE, genus and species indeterminate REFERRED MATERIAL.?Incomplete right ulna and articulat ed proximal end of radius, MLP 90-X-l-l (Figures 5, 6), col lected by Ulyses Pardinas and Maximiliano Lezcano; south of the mouth of Arroyo Loberia (38?15'S, 57?40'W); middle Pliocene (Upper Chapadmalalan Age). MEASUREMENTS (in mm).?Ulna: Total length as pre served, 253; width of proximal end, 40.8; width of midshaft, 180 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY / .':'???;?:?>'?? FIGURE 3.?Distal end of humerus of Dryornis pampeanus (MLP 20-169; lec totype): a, palmar view; b, anconal view. FIGURE 4.?Proximal end of humerus of Vultur gryphus (MLP 48-XII-l 6-225) in anconal view. Scale bar= 1 cm. impressio brachialis beginning more proximally than in Vultur and lacking foramina, unlike Vultur, Geronogyps, or Gymnog yps; (6) base of the processus cotylaris dorsalis palmarly exca vated; and (7) papillae remigiales situated on midline of dorsal surface, but situated more mediad in Vultur. The bone is dam aged at the level of the prominence for the tubercullum lig. col- lateralis ventralis. There are only 12 anconal papillae for at tachment of secondary remiges preserved. The radius is characterized by having (1) strong tuberculum bicipitali radii, as in Geronogyps; (2) deep and rounded bicipi tal attachment; (3) capital tuberosity (in the sense of Howard, 1929) prominent and limited internally by a deep notch, as in Geronogyps; and (4) capital tuberosity doubly pierced with large foramina. REMARKS.?The preserved bones of MLP 90-X-l-l are ap proximately 20% longer than those of either modern condors or Geronogyps. The features mentioned below make the fossil too different to be referred to any recent genus of fossil vultures; nevertheless, it would be imprudent to name a new genus and species until more complete material is recovered. 15.8; depth of midshaft, 16.4. Radius: Distance between head and capital tuberosity, 21. DESCRIPTION.?The ulna differs from those of other condor genera by having the following features: (1) olecranon devel oped as in Vultur, larger than in Gymnogyps or Geronogyps; (2) olecranon directed medially; (3) the humero-ulnar depres sion well excavated; (4) incisura radialis shallower than in Vul tur, Gymnogyps, or Geronogyps and located less medially; (5) Geronogyps reliquus Campbell, 1979 REFERRED MATERIAL.?Left humerus, partially broken, CCM Nro 95-VI-5-1 and 95-VI-5-4 (Figure 7); left margin of Arroyo Chasico, close to its mouth (38?30'S, 63?W), Villarino County, Buenos Aires Province; Pleistocene, sensu lato. MEASUREMENTS (in mm).?Depth of head, 19; distance be tween tuberculum dorsale and insertion of M. scapulohumera- NUMBER 89 181 FIGURE 5.?Vulturidae, probably a new genus and species (MLP 90-X-l-l). Proximal end of radius: a, anconal view; b, palmar view. Proximal end of ulna and shaft: c, external view; d, palmar view; e, anconal view. Scale bar=l cm. lis, 61; width of midshaft, 20.8; depth of midshaft, 16.3; distal width, 46.9. REMARKS.?Before now, Geronogyps was known only from the Pleistocene sediments of the Talara Tar Seeps, Peru (Camp bell, 1979). Campbell established this genus based on a com plete tarsometatarsus, with the distal end of a left humerus and the proximal and distal ends of two right humeri as paratypes. Because the paratypes are badly crushed, the specimens report ed herein bring additional information about the morphology of Geronogyps reliquus. The following characters of Geronogyps reliquus, given by Campbell (1979), are present in the Argentinian specimen: (1) margo caudalis dropping off sharply on both sides, whereas more rounded in Vultur and Gymnogyps; (2) attachment of M. proscapulohumeralis brevis more proximal than in Vultur; (3) crista deltopectoralis flaring distally, not flaring as much in Vultur or in Gymnogyps; (4) shaft ventral to M. pectoralis su- perficialis not depressed, unlike Vultur and similar to Gymno gyps; (5) condylus dorsalis wide and short, whereas narrow and long in Vultur and Gymnogyps; (6) condylus ventralis short (very long in Vultur and moderately long in Gymnog yps); (7) epicondylus dorsalis long and not angular, unlike Vultur; (8) condylus ventralis gently rotated anteriad, result ing in moderately flexed distal end, as in Gymnogyps (greatly rotated in Vultur); and (9) impression of M. brachialis shal lower than in Vultur or Gymnogyps. Discussion and Conclusions Condors, one of the primary scavenging lineages of birds, are large-sized vulturids with past and present distributions restricted to the New World. The fact that the temporal range of condors is less extensive in South America (middle-late Pliocene) than in North America (middle Miocene), and that their absence from the richly fossiliferous Paleogene and ear ly Neogene outcroppings of Argentina does not seem to be related to taphonomic problems, could probably be consid ered a reaffirmance of Emslie's idea (1988a) about a North American origin for the group. Condors may have participat ed in the Great American Biotic Interchange (GABI) during Pliocene times (Webb and Marshall, 1982; Emslie, 1988a), moving from north to south across open savanna environ ments; however, the presence of a fossil condor, Antillovultur varonai Arredondo (1971), from the late Pleistocene of Cu ba, proves that condors can cross water barriers. This fact, to gether with the evidence of a significant interchange of pre- Pliocene volant bird families between both Americas before the formation of a land connection (Rasmussen and Kay, 1992), seems to weaken the hypothesis that condors were part of the latest events of the GABI. Our analysis does not make it possible to bring arguments in favor of, or against, this hypothesis. The fossil record summarized above indicates that a great di versity of condors occurred in the late Cenozoic of the Pam pean region of Argentina, including three genera and at least four species: Dryornis pampeanus, Vultur gryphus, Gerono- 182 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 6.?Vulturidae, probably a new genus and species (MLP 90-X-l-l). Proximal end of ulna and shaft: a, external view; b, palmar view; c, anconal view. Scale bar=5 cm. gyps reliquus, and an indeterminate vulturid that probably rep resents a new genus and species. Dryornis pampeanus and Vultur gryphus, from the ear- ly?-middle Pliocene, constitute the earliest record of condors in South America. The remaining two taxa come from Pleis tocene and middle Pliocene sediments, respectively. The presence of Geronogyps reliquus (up to now restricted to the Pleistocene of Peru) in the Pleistocene sediments of the Pampean region extends considerably the geographic range of the species. Thus, condors were more widely distributed during the early?-middle Pliocene and Pleistocene than they are at present. Only one condor species now survives in South America, Vultur gryphus, and its distribution is restricted to Andean re gions and arid steppes of Patagonia. It has been hypothesized that the decline in condor diversity and the retraction of Vultur gryphus from the Pampean region is due to climatic changes and/or to the extinction of megaherbivorous mammals, which were likely their main food source (as carrion) (Emslie, 1987; Tonni and Noriega, 1998). NUMBER 89 183 Literature Cited FIGURE 7.?Referred humerus of Geronogyps reliquus (CCM 95-VI-5-1 (a), CCM 95-VI-5-4 (b,c)): a, proximal end, anconal view; b, distal end, anconal view; c, distal end, palmar view. Scale bars= 1 cm. Arredondo, O. 1971. Nuevo genero y especie de ave fosil (Accipitriformes: Vulturidae) del Pleistoceno de Cuba. Memorias de la Sociedad Cientifico Natu ral La Salle, Caracas, 31(90):309-323. Baumel, J.J., and L.M. Witmer 1993. Osteologia. In J.J. Baumel, A.S. King, J.E. Breazile, H.E. Evans, and J.C. Vanden Berge, editors, Handbook of Avian Anatomy: Nomina Anatomic Avium, second edition. Publications of the Nut- tall Ornithological Club, 23:45-132. Brodkorb, P. 1967. Catalogue of Fossil Birds, Part 3 (Ralliformes, Ichthyornithiformes, Charadriiformes). Bulletin of the Florida State Museum, Biological Sciences. 2(3): 106-218. Brown, L., and D. Amadon 1968. Eagles, Hawks and Falcons of the World. 945 pages. London: Coun try Life. Campbell, K. 1979. The Non-Passerine Pleistocene Avifauna of the Talara Piura Seeps, Northwestern Peru. Contributions, Life Sciences Division, Royal Ontario Museum, 118:1-203. Cione, A., and E. Tonni 1995. Los estratotipos de los pisos Montehermosense y Chapadmalalense (Plioceno) del esquema cronologico sudamericano. Ameghiniana, 32(4):369-374. Emslie, S. 1987. Age and Diet of Fossil Condors in Grand Canyon, Arizona. Science, 237:768-770. 1988a. The Fossil History and Phylogenetic Relationships of Condors (Ciconiiformes: Vulturidae) in the New World. Journal of Verte brate Paleontology, 8(2):212-228. 1988b. An Early Condor-like Vulture from North America. Auk, 105: 529-535. Fisher, H. 1946. Adaptations and Comparative Anatomy of the Locomotor Appara tus of New World Vultures. American Midland Naturalist, 35: 545-727. Griffiths, C. 1994. Monophyly of the Falconiformes Based on Syringeal Morphology. Auk, lll(4):787-805. Hertel, F. 1992. Morphological Diversity of Past and Present New World Vultures. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:413^118. 1994. Diversity in Body Size and Feeding Morphology within Past and Present Vulture Assemblages. Ecology, 75(4): 1074?1084. Howard, H. 1929. The Avifauna of Emeryville Shellmound. University of California Publications in Zoology, 32:301-394. Ligon, J. 1967. Relationships of the Cathartid Vultures. Occasional Papers of the Museum of Zoology, University of Michigan, 651:1?26. Linnaeus, C. 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis, 1: Regnum Animate, ii+824 pages. Tenth edition. [Facsimile reprinted in 1956 by the British Museum of Natural History.] Moreno, R, and A. Mercerat 1891. Catalogo de los pajaros fosiles de la Republica Argentina. Anales del Museo de La Plata, Paleontologia Argentina, 1:7?71. Olson, S. 1985. The Fossil Record of Birds. In D. Farner, J.R. King, and K.C. 184 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Parkes, editors, Avian Biology, 8:79-256. New York: Academic Press. Patterson, B., and J.L. Kraglievich 1960. Sistematica y nomenclature de las aves fororracoideas del Plioceno Argentine Publicaciones del Museo Municipal de Ciencias Natu rales y Tradicionales de Mar del Plata, 1 (1): 1 -51. Rasmussen, D.T., and R. Kay 1992. A Miocene Anhinga from Colombia, and Comments on the Zoogeo- graphic Relationships of South America's Tertiary Avifauna. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Sciences Series, Natural History Museum of Los Angeles County, 36:225-230. Rea, A. 1983. Cathartid Affinities: A Brief Overview. In S. Wilbur and J. Jackson, editors, Vulture Biology and Management, pages 26-54. Berkeley: University of California Press. Sibley, C, and R. Ahlquist 1990. Phylogeny and Classification of Birds: A Study in Molecular Evolu tion. xxii + 976 pages. New Haven, Connecticut: Yale University Press. Sibley, C, and B. Monroe 1990. Distribution and Taxonomy of Birds of the World. xxiv+1111 pages. New Haven and London: Yale University Press. Tambussi, C. 1989. Las aves del Plioceno tardio-Pleistoceno temprano de la Provincia de Buenos Aires, Argentina. 378 pages. Doctoral thesis, Univer sidad Nacional La Plata, Argentina. Tambussi, C, and J.I. Noriega 1996. Summary of the Avian Fossil Record from Southern South America. In G. Arratia, editor, Contributions of Southern South America to Vertebrate Paleontology. Munchner Geowissenschaftliche, Abhand lungen, series a, Geologie und Palaontologie, 30:245-264. Munchen: Verlag Dr. Friedrich Pfeil. Tambussi, C, J.I. Noriega, and E.P Tonni 1993. Late Cenozoic Birds of Buenos Aires Province (Argentina): An At tempt to Document Quantitative Faunal Changes. Paleogeography, Paleoclimatology, Paleoecology, 101:117-129. Tonni, E.P. 1980. The Present State of Knowledge of the Cenozoic Birds of Argentina. Contributions in Science, Natural History Museum of Los Angeles County, 330:105-114. 1984. Dos nuevas aves para el Pleistoceno del sur-sureste de la provincia de Buenos Aires. Resumenes: Primeras Jornadas Argentinas de Pa leontologia de Vertebrados. Buenos Aires: Comision de Investiga ciones Cientificas de la Provincia de Buenos Aires. Tonni, E.P, and J.I. Noriega 1998. Los condores (Ciconiiformes, Vulturidae) en el Cenozoico Superior de la Region Pampeana (Republica Argentina): distribucion, inter- acciones y extinciones. Ameghiniana, 35(2): 141-150. Webb, S.D., and L.G. Marshall 1982. Historical Biogeography of Recent South American Land-Mam mals. Special Publications of Pymatuning Laboratory of Ecology, 6:39-52. Two New Fossil Eagles from the Late Pliocene (Late Blancan) of Florida and Arizona and Their Biogeographic Implications Steven D. Emslie and Nicholas J. Czaplewski ABSTRACT Two new species of fossil eagles are described from the late Pliocene of Florida and Arizona, adding new information on the paleoecology of these regions. Aquila bivia, new species, is known from 33 skeletal elements from inglis IA, Citrus County, Florida, and from a partial skeleton from 111 Ranch, Graham County, Ari zona. It was a large eagle, approximately 10%-15% larger than females of modem A. chrysaetos (Linnaeus), and it is the first valid fossil species in this genus to be described from North Amer ica. Amplibuteo concordatus, new species, is known from 13 skel etal elements from Haile 7C, Alachua County, and Inglis 1C, Citrus County, Florida, and from three specimens from Duncan, Greenlee County, Arizona. It is the third species of the genus to be described, and it represents the earliest occurrence of this genus. These two new taxa add to a growing list of vertebrates with fossil distributions in both the Florida peninsula and western North America, which reflects a corridor of common habitat that once united these regions. This corridor initially developed during gla cial intervals in the late Pliocene, when numerous taxa of mam mals, birds, reptiles, and plants first appear in the fossil record of Florida. The corridor probably was composed largely of dry thorn- scrub and savannah communities, but it also may have had a mosaic of lakes, wetlands, and hammocks that allowed dispersal of a variety of species that reflect these communities. Introduction The fossil record in Florida and the southwestern United States indicates that numerous species of mammals, reptiles, and plants were shared between these regions during the Pliocene and Pleistocene (Neill, 1957; Blair, 1958; Marshall et al., 1982; Marshall, 1985; Meylan, 1982). This distributional Steven D. Emslie, Department of Biological Sciences, University of North Carolina, Wilmington, North Carolina 28403, United States. Nicholas J. Czaplewski, Oklahoma Museum of Natural History, 1335 Asp Avenue, University of Oklahoma, Norman, Oklahoma 73019, United States. pattern has been explained as the result of the Gulf Coast corri dor, a broad expanse of savannah and xeric thorn-scrub habitat that extended through Central America, Mexico, and the south ern portion of the United States (Blair, 1958; Mares, 1985; Webb, 1985). This corridor allowed the dispersal of taxa be tween the Florida peninsula and the western United States and between North and South America during the Great American Biotic Interchange, which began at about 2.5 Ma (Stehli and Webb, 1985). The fossil record of birds during this period is less well known, although similar dispersals have been documented (Vuilleumier, 1985; Emslie, 1996). For example, the phorus- rhacoid Titanis walleri Brodkorb is a representative of a South American group that reached Florida and Texas during the Plio-Pleistocene (Brodkorb, 1963; Baskin, 1995). An unusual group of condor-like vultures, the teratorns, also may have been from a South American lineage that entered North Ameri ca by the late Pliocene (Campbell and Tonni, 1981; Vuilleumi er, 1985; Emslie, 1988). Recent paleontological investigations in Florida and Arizona have indicated that considerably more avian taxa were shared between North and South America and the western and eastern United States during the Plio-Pleis tocene than previously have been documented (Emslie, 1998). Herein, we describe two new species of eagles from the late Pliocene of Florida and Arizona. As with mammalian taxa, these and other avian fossils indicate an extensive common habitat that once extended across the southern United States in the late Pliocene. METHODS.?Comparative analyses of fossil and recent skel etal material was completed at the Florida Museum of Natural History (FLMNH), University of Florida (UF), Gainesville, where the fossils are housed; the National Museum of Natural History (USNM; collections of the former United States Na tional Museum), Smithsonian Institution, Washington, D.C; the Museums of Paleontology (UMMP) and Zoology (UMMZ), University of Michigan, Ann Arbor; the Museum of Northern Arizona (MNA), Flagstaff; the University of California, Los 185 186 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Angeles (UCLA); and the George C. Page and Los Angeles County Museum of Natural History (LACM), California. Other fossils examined are in the collections of the American Muse um of Natural History (AMNH), New York, and the Oklahoma Museum of Natural History (OMNH), Norman. Specimens were measured with digital calipers to the nearest 0.1 mm; all measurements are self-descriptive except proximal depth of the coracoid, which was taken from the glenoid facet to the brachi al tuberosity. Comparative material for Amplibuteo hibbardi Campbell was not available, and character determinations for skeletal elements are based on the published descriptions and il lustrations in Campbell (1979). Osteological terminology fol lows that of Howard (1929), with certain modifications. No menclature for species' binomials and English names of modern birds follows Sibley and Monroe (1990). ACKNOWLEDGMENTS.?This research was funded by NSF Grant EAR 94-03206 to S. Emslie and by M. Mares, Director, OMNH, with funds to N. Czaplewski in support of field work. Assistance with museum collections was provided by K. Campbell and C. Shaw (LACM); D.W. Steadman and T. Web ber (UF); and P. Angle and S. Olson (USNM). G.S. Carr (for merly UF) and D.W. Steadman originally identified the Inglis 1A avifauna and recognized that an undescribed eagle was rep resented in the collection. Field assistance at Inglis 1C was pro vided by J. Boyle, Steve and Sue Hutchens, J. Mueller, M. Se- wolt, B. Shockey, Barbara and Reed Toomey, and T. Verry. M. Frank and B. MacFadden, UF, also provided support for the excavation of this site, and we thank E.H. Lindsay, Cheryl D. Czaplewski, and Lynn Saline (United States Bureau of Land Management) for their assistance to N. Czaplewski. We thank G. Morgan, S. Olson, S. Parry, and D.W. Steadman for their comments on an earlier version of this paper. AVES ACCIPITRIDAE Amplibuteo Campbell, 1979 Genus characterized by having the carpometacarpus with metacarpal III merging with the distal symphysis at an angle, a tendinal groove that curves proximally to the anterior surface of metacarpal II, the humerus with the shaft relatively straight and excavated at the distal end of the median crest, a broad pneumatic foramen, a deep excavation on the internal side of the impression for M. brachialis anticus, and a prominent del toid crest as described for Morphnus (Howard, 1932). Includes two species, A. woodwardi and A. hibbardi, formerly referred to Morphnus (see Campbell, 1979); A. concordatus, new spe cies; and one undescribed species. Amplibuteo concordatus, new species HOLOTYPE.?Right carpometacarpus, UF 159426 (Figure 1A; Table 1). TYPE LOCALITY.?Haile 7C, section 24, T. 9S, R. 17E, Ala chua County, Florida (FLMNH Vertebrate Paleontology Local ity number AL109; Figure 2). FIGURE 1.?A, Holotypical right carpometacarpus (UF 159426) of Amplibuteo concordatus from Haile 7C, Alachua County, Florida, in internal (left) and external (right) views; B, paratypical left car pometacarpus (AMNH 10395) from Duncan 4-19, Greenlee County, Arizona, in internal (left) and external (right) views. For each specimen, scale = xl, bar= 1 cm. NUMBER 89 187 TABLE 1.?Measurements (mm) of pectoral and wing elements of Amplibuteo woodwardi (sample size (n), mean (x), and standard deviation (s.d.)) from Rancho La Brea, California, compared with A. concordatus, new species. Measurements of the carpometacarpus for A. woodwardi are in Emslie (1995). (L=length, PW=proximal width, PD=proximal depth; LWS=least width of shaft, LDS=least depth of shaft, DW=distal width, DD=distal depth.) Element Coracoid Amplibuteo woodwardi n jc + s.d. range Amplibuteo concordatus UF 159404 UF 165529 AMNH 10399 Humerus Amplibuteo woodwardi n x?s.d. range Amplibuteo concordatus UF159406(right) UF 159407(left) UF 165542 Ulna Amplibuteo woodwardi n Jc?s.d. range Amplibuteo concordatus UF 121743 Radius Amplibuteo woodwardi LACM H7419 LACM H7414 LACM D2057 LACM D4872 Amplibuteo concordatus UF 159408 AMNH 10398 Carpometacarpus Amplibuteo concordatus UF 159426 AMNH 10395 L 5 76.4?2.3 74.5-80.7 - - 72.5 1 207.0 - 163.3 162.7 - -- - 177.4 -- - - - - 93.2 96.4 PW - - - - - - 6 40.0 ?2.4 37.4-42.7 33.3 34.2 32.6 5 24.3?0.6 23.9-25.0 17.9 11.6 11.1 - - - - 9.5 9.5 PD 5 21.1?1.2 20.2-23.2 18.4 16.4 17.9 6 12.3 ?0.5 11.5-12.8 9.0 9.0 8.8 5 19.0?0.7 18.2-19.5 14.6 8.5 8.2 - - - - 21.1 21.9 LWS 5 12.3 ?0.7 11.4-13.3 - 9.0 10.6 3 14.5?0.7 13.7-15.0 13.1 12.3 12.1 5 10.3?1.5 8.9-12.3 9.0 -- 6.0 6.2 5.4 - 7.9 8.0 LDS 5 7.8?0.4 7.3-8.4 - - 6.4 3 12.6?0.3 12.3-12.8 10.4 10.8 9.5 5 10.7?0.5 10.0-11.1 8.4 -- 4.9 5.3 4.0 - 5.2 5.2 DW 4 34.9?1.1 33.5-36.1 - - - 4 35.8? 1.8 34.1-37.5 26.7 26.7 - 4 16.3?1.4 13.9-17.4 14.7 -- 14.3 15.9 13.5 13.3 11.8 12.5 DD 4 9.6?0.7 8.8-10.3 - - 8.2 4 17.9 ?0.7 17.3-18.8 14.2 14.2 - 4 17.3 ?0.9 15.9-17.9 12.7 -- 6.7 7.6 5.8 6.1 13.1 13.9 HORIZON AND AGE.?Late Pliocene (late Blancan), 2.2-2.0 Ma (Morgan and Hulbert, 1995; Hulbert, 1997; Emslie, 1998). MEASUREMENTS OF HOLOTYPE.?See Table 1. PARATYPES.?Haile 7C: Humeral end of left coracoid, UF 159404; left furcula, UF 159403; carina of sternum, UF 159402; right and left humeri (matching) with shafts broken, UF 159406, 159407 (Figure 3B); right and left ulnae (match ing), UF 121743, UF 159405; distal end of left radius, UF 159408; partial synsacrum, UF 159427. Inglis 1C (section 10, T. 17S, R. 16E, Citrus County, Florida (FLMNH Vertebrate Paleontology Locality number CIO 19; Figure 2)): Associated right coracoid and fragmentary manu brium of sternum, UF 165529; left humerus missing distal end, UF 165542 (Figure 3A); left fibula, UF 165577. Locality dates to late Pliocene (early Irvingtonian, 2.0-1.6 Ma) based on ver tebrate biochronology similar to that of Inglis IA. Duncan 4-19 (south locality, west of Railroad Wash, Green lee County, Arizona (Figure 2)): Right coracoid, AMNH 10399; distal end of left radius, AMNH 10398; left car pometacarpus, AMNH 10395. Locality dates to late Pliocene, 3.7-3.2 Ma (Blancan III of Repenning, 1987). MEASUREMENTS OF PARATYPES.?See Table 1. ETYMOLOGY.?From Latin, concordatus, agree together or harmonize, in reference to the close similarity of the fossil ma terial both from the two Florida localities and from Arizona. DIAGNOSIS.?Differs from Amplibuteo woodwardi (L. Mill er) and A. hibbardi in the following characteristics. Car pometacarpus (UF 159426, AMNH 10395; Figure 1) with metacarpal I relatively large, with little or no proximal curva ture; deep, narrow fossa present on internal surface of metacar pal I below pisiform process; internal ligamental fossa moder ately to deeply excavated; and area between pisiform and 188 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY HAILE 7C FIGURE 2.?Locations of late Pliocene fossil localities discussed in the text. FIGURE 3.?Paratypical humeri of Amplibuteo concordatus from Florida: A, left humerus (UF 165542) from Ing lis 1C, Citrus County; B, left humerus (UF 159407) from Haile 7C, Alachua County. For each specimen, scale=xl, bar=l cm. NUMBER 89 189 carpal trochlea relatively small (carpometacarpus larger, metacarpal I relatively smaller, with distinct proximal curva ture, fossa below pisiform shallow and broad, area of internal ligamental fossa relatively greater in Amplibuteo woodwardi (?=4) and A. hibbardi). Coracoid with relatively short, narrow coraco-humeral surface (long and broad in A. woodwardi {n=5), short and broad in .4. hibbardi); pneumatic foramina be low brachial tuberosity small and indistinct (shallow to deep in A. woodwardi). Humerus (Figure 3) with internal tuberosity relatively narrow and less blunt (large and blunt in A. wood wardi (?=9), rounded in A. hibbardi); distinct excavation at distal end of medial bicipital crest (deep in A hibbardi; deep to shallow or absent in A. woodwardi); ligamental furrow rela tively narrow and deep (furrow broad, shallow to deep in A. woodwardi); bicipital crest merges with shaft immediately be low pneumatic foramen (merges with shaft more distal to pneu matic foramen, forming a distinct flange, in A. woodwardi and A. hibbardi); attachment for anterior articular ligament rela tively flat, long, and narrow (flat to convex, short and broad to long and narrow in A. woodwardi, short and rounded in A. hibbardi); impression of M. brachialis anticus relatively nar row (broad in A. woodwardi); and entepicondylar and ectepi- condylar prominences relatively small (larger and more robust in A. woodwardi). Ulna with bicipital attachment relatively high on shaft (attachment more distal on shaft in A. woodwardi (?=3) and A. hibbardi); prominence for anterior articular liga ment relatively small (large and robust in A. woodwardi and A. hibbardi); distal external condyle relatively long and narrow (short and broad in A. woodwardi and A. hibbardi); and carpal and radial tuberosities prominent (reduced in A. woodwardi, more prominent in A. hibbardi). Radius with ligamental prom inence relatively small and blunt (large with more distal exten sion in A. woodwardi (?=2) and ,4. hibbardi); scapholunar fac et in distal view relatively short and narrow (long and broad in A. woodwardi). COMPARATIVE MATERIAL.?Morphnus guianensis (Dau din), skeleton, USNM 432243, unsexed; LACM 91788, skele ton, female. Harpyhaliaetus solitarius (Tschudi), partial skele ton with sternum, coracoid, proximal humerus, and femur, UCLA 41971, unsexed. STATUS.?Extinct, known from fossils only. REMARKS.?The material from Haile 7C is probably from a single individual based on similarities in the size and features of matching elements. At least one individual is represented from Inglis 1C, and the humerus (UF 165542) is slightly po rous and incompletely ossified, indicating it is from a subadult. Amplibuteo concordatus was a small eagle compared to A. woodwardi ox A. hibbardi. A third possible species of Amplibu teo (UF 102550), reported by Emslie (1995) from the late-ear ly Irvingtonian Leisey Shell Pit 3B, is slightly larger than A. concordatus and smaller than A. woodwardi (see measure ments in Emslie, 1995, table 11). It compares most closely with A. woodwardi except for the carpometacarpus, which is shorter and differs in characters as described by Emslie (1995). The carpometacarpus (UF 102550) from Leisey was compared with UF 159426 and AMNH 10395 and was found to differ in hav ing a distinct proximal curvature of metacarpal I and having the fossa below the pisiform process broad and deep. These differ ences also suggest that the Leisey specimen represents a fourth, undescribed species of Amplibuteo. Of the two other fossil species described in this genus, A. woodwardi is known from the late Pleistocene of Rancho La Brea, California, and the middle and late Pleistocene of Florida (Miller, 1911; Howard, 1932; Emslie, 1995); A. hibbardi is known only from the late Pleistocene Talara Tar Seeps, Peru (Campbell, 1979). The addition of A. concordatus to this record extends the geologic range of the genus into the late Pliocene. This species also was relatively long-lived, with a po tential geologic range spanning 2.1 million years (between 3.7-1.6 Ma). The genus reached its greatest diversity in the Pleistocene of North and South America, when up to three spe cies may have existed, although none of these apparently were sympatric based on the available evidence. Campbell (1979) erected the genus Amplibuteo, with A. hib bardi of Peru as the type species, and at the same time trans ferred the fossil species Morphnus woodwardi from Rancho La Brea to the same new genus. His comparisons suggested that Amplibuteo is most closely related to Buteo and Geranoaetus, whereas Morphnus is closer to Harpia than to other genera. Campbell's comparisons included one specimen of Harpy haliaetus solitarius, from which only the tarsometatarsus was used in the generic diagnosis of Amplibuteo. We compared the sternum, coracoid, proximal end of the humerus, and femur of this species with those of Morphnus guianensis, Amplibuteo woodwardi, and A. concordatus. These comparisons indicate a closer similarity in skeletal characters between Harpyhaliaetus and Amplibuteo than previously has been recognized. Charac ters shared by these two genera include the sternum with the anterior carinal margin placed relatively far from the manubri um, the coracoid with the coracoidal fenestra positioned rela tively far from the scapular facet, the humerus with a bicipital crest that merges with the shaft distally, giving the crest a dis tinct flange in A. woodwardi (although the crest merges imme diately to a shaft with no flange in A. concordatus and Harpy haliaetus), and the femur with a high trochanter and a trochanteric ridge extending far down the shaft. These similarities raise the possibility that Amplibuteo may be closely related to or even congeneric with Harpyhaliaetus, of which additional modern skeletal specimens would be high ly desirable. The high diversity of Accipitridae and the lack of adequate comparative skeletons for some species (e.g., H. cor- onatus), however, compound the problems associated with de scribing new taxa in this group, and additional, more detailed comparisons are needed with all accipitrid genera. The two species of Harpyhaliaetus are today confined to the Neotro- pics. The Crowned Solitary Eagle {H. coronatus) is found in savannah habitats from Bolivia to southern Argentina, whereas the Black Solitary Eagle {H. solitarius) occurs on mountain 190 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY slopes from Mexico to Peru (Brown and Amadon, 1968). With little being known of the natural history of these species (Brown and Amadon, 1968), and comparative skeletal material being rare (Wood and Schnell, 1986), future additions to exist ing skeletal collections should provide further insight on the re lationships between Harpyhaliaetus and Amplibuteo. Given these affinities, Amplibuteo probably is of Neotropical origin, although this suggestion is not yet supported by the fossil record. Aquila Linnaeus These fossils represent a large eagle that is referable to Aqui la, and not Haliaeetus, by the following characters: cranium with relatively larger, more rounded foramen prooticum; man dible with symphysis broader and less tapered distally, and with relatively broader articular; scapula with external border of acromion more curved in proximal view; humerus with bi cipital crest smaller and extending less distally on shaft, with median crest forming a longer border to the pneumatic fossa distally; ulna with relatively more pronounced prominence for anterior articular ligament and with external condyle extending farther proximally on the shaft; radius with bicipital tubercle located relatively farther distally on shaft; carpometacarpus with relatively broader and deeper excavation below pisiform process, internal border of carpal trochlea long, extending clos er to metacarpal III; femur relatively longer, trochanter less pronounced; and tibiotarsus with tendinal bridge angled less medially and tendinal groove placed more internally than in Haliaeetus. Aquila bivia, new species HOLOTYPE.?Right carpometacarpus, UF 30015 (Figure 4; Table 2). TYPE LOCALITY.?Inglis IA, section 8, T. 17S, R. 16E, Cit rus County, Florida (FLMNH Vertebrate Paleontology Locali ty number CI001; Figure 2). HORIZON AND AGE.?Late Pliocene (early Irvingtonian), 2.0-1.6 Ma (Morgan and Hulbert, 1995). MEASUREMENTS OF HOLOTYPE.?See Table 2. PARATYPES.?Inglis IA: Fragment of left cranium, UF 159544; articular end of left mandible, UF 30028; two mandib ular symphyses, UF 30029, 30030; two proximal ends of left scapulae, UF 30026, 30027; shaft of left coracoid, UF 30034; anterior portion of carina of sternum, UF 30035; distal end (damaged) of right humerus, UF 30043; two proximal ends of left humeri, UF 30040, 30041; two proximal ends of left ulnae, UF 30023, 30024; distal end of left ulna, UF 30025; two proxi mal ends of right radii, UF 30038, 30039; distal end of left ra dius, UF 30036; proximal end of left radius, UF 30037; distal end of left carpometacarpus, UF 30014; three alar phalanges, TABLE 2.?Measurements (mm) of modem and late Pleistocene (maximum lengths only, Rancho La Brea, from Howard, 1932) Golden Eagle (Aquila chrysaetos) carpometacarpi, femora, and tibiotarsi (sample size (n), mean (x), and standard deviation (s.d.)) in comparison with fossils of A. bivia, new species, from Inglis IA, Florida. (L=length, PW=proximal width, PD=proximal depth; LWS=least width of shaft, LDS=least depth of shaft, DW=distal width, DD=distal depth.) Element Carpometacarpus Aquila chrysaetos w=10; ?,5? x?s.d range maximum Aquila bivia UF 30015 Femur Aquila chrysaetos n=12;7?,5? ic?s.d range maximum Aquila bivia UF 30019 Tibiotarsus Aquila chrysaetos ?=14;7?,7? x ?s.d. range maximum Aquila bivia UF 30012 L 103.0?4.4 97.6-112.5 112.9 119.6 129.7?4.5 120.8-134.6 135.8 140.1 164.2?5.1 155.6-172.7 172.8 180.6 PW 10.4?0.7 9.5-11.7 12.5 28.0?1.3 25.1-30.1 29.3 - - - PD 26.0?1.2 24.1-24.8 29.9 19.5?1.2 17.7-21.7 20.7 - - - LWS 8.4?0.4 7.9-9.2 10.3 13.5?0.8 12.5-15.0 14.1 11.3?0.5 10.4-12.1 11.6 LDS 6.1?0.2 5.7-6.5 6.7 I2.7?0.7 11.4-13.6 13.8 7.7?0.4 7.1-8.3 8.3 DW 13.7?0.6 12.8-14.8 15.1 30.1?1.5 27.3-32.3 29.6 22.9?1.3 20.7-26.0 21.3 DD 15.4?0.8 13.8-16.6 17.9 22.3 ?1.2 20.6-24.8 22.8 15.6?0.8 14.3-16.6 12.5 NUMBER 89 191 FIGURE 4.?Holotypical right carpometacarpus (UF 30015) of Aquila bivia from Inglis IA, Citrus County, Florida, in internal (left) and external (right) views. Scale=xl, bar=l cm. UF 30031-30033; two proximal ends of right femora, UF 30020, 30021; distal end of right femur, UF 30022; left femur, UF 30019 (Figure 5B; Table 2); right tibiotarsus, UF 30012 (Figure 5A; Table 2); proximal end of right fibula, UF 30013; right metatarsal I, UF 30018; right phalanx 1 of digit I, UF 30016; left ungual phalanx of digit II, UF 30017; right phalanx 2 of digit II, UF 30045; left phalanx 1 of digit III, UF 30046; phalanx 3 of digit III, UF 30044; phalanx 1 of digit IV, UF 30047; ungual phalanx of digit IV, UF 30048. At least two adults represented. Ill Ranch (East Ravine at Dry Mountain, Graham County, Arizona (OMNH Locality V818; Figure 2)): Partial associat ed skeleton (OMNH 50271) including proximal and distal ends of a right carpometacarpus, right and left ulnares, right radiale, proximal end of left tarsometatarsus; right metatarsal I, phalanx 1 and ungual of digit I, phalanx 2 and ungual of digit II, phalan ges 1-3 and ungual of digit III, phalanges 2-4 and ungual of digit IV; left metatarsal I and phalanx 1 of digit I, phalanx 2 and ungual of digit II, phalanx 3 (partial) of digit III, and pha langes 2-4 of digit IV. This locality is within greenish clay in the Gila Conglomer ate at the approximate level of paleomagnetic samples 112 and 113, East Ravine, of Galusha et al. (1984). This section has re versed polarity throughout and represents Chron C2r, or slight ly above the 2.47 Ma Dry Mountain Ash Bed (Izett, 1981; Ga lusha et al., 1984; Tomida, 1987). One adult (from associated material) is represented. MEASUREMENTS OF PARATYPES.?See Table 1. ETYMOLOGY.?From Latin, bivius, -a, -um, two-wayed, in reference to the distribution of the fossil specimens in Florida and Arizona. DIAGNOSIS.?The species is diagnosed by the following characters. Carpometacarpus (UF 30015, OMNH 50271; Fig ure 4) with relatively deep external ligamental attachment with pronounced proximal border, metacarpal I relatively long and robust, pit below pollical facet on metacarpal II, relatively large prominence for muscle attachment of proximal internal edge of metacarpal III (ligamental attachment shallow and proximal border less pronounced, metacarpal I relatively short er and less robust, absence of pit below pollical facet on metac arpal II, and prominence for muscle attachment small in Aquila chrysaetos, A. rapax, A. heliaca, A. verreauxii, and A. audax). Femur with relatively long and slender shaft (shaft shorter and more robust in A. chrysaetos) and with broad, deeply excavated popliteal area (area narrower and moderately excavated in A. chrysaetos; Figure 5B). Tibiotarsus with relatively long, nar row shaft (shaft shorter and more robust in A. chrysaetos; Fig ure 5A). Size relatively large compared with Aquila chrysaetos, A. rapax, A. heliaca, A. verreauxii, or A. audax. COMPARATIVE MATERIAL.?Haliaeetus leucocephalus (Lin naeus); USNM 611999, USNM 489276, 1 male and 1 female. Aquila rapax (Temminck), USNM 430406, 430532, 488147, 488148, 1 male, 3 females. Aquila heliaca Savigny, USNM 488808, female. Aquila chrysaetos (Linnaeus), UF 19399, 23961,23962,23964,1 male, 3 females; USNM 17721,17983, 18194, 18802, 19251, 19394, 19399, 19724, 19777,288513, 319967, 320978, 343130, 491476, 500354, 500355, 500367, 502292,612086, 5 males, 5 females, 9 unsexed; LACM 89953, female. Aquila audax (Latham), USNM 344883, unsexed. Aq uila verreauxii Lesson, USNM 612539, unsexed. STATUS.?Extinct, known from fossils only. REMARKS.?Aquila bivia was a large eagle that was closely related to the Golden Eagle {A. chrysaetos), but it was 10%-15% larger than females of that species and had limb ele ments that do not overlap with those of A. chrysaetos in length (Table 2). A large series of late Pleistocene eagle bones from Rancho La Brea that Howard (1947) recognized as a larger temporal form of A. chrysaetos also do not approach the length of the Inglis specimens (Table 2). Howard (1932) noted that the Rancho La Brea eagle had relatively longer wings and shorter legs than did recent A chrysaetos, but with a larger series of re cent skeletons she later found that the limb proportions of the 192 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 5.?Paratypes of Aquila bivia from Inglis IA, Citrus County, Florida: A, right tibiotarsus (UF 30012); B, left femur (UF 30019), in anterior (left) and posterior (right) views. For each specimen, scale= x 0.75, bar= 1 cm. fossil eagle did not differ significantly from those of the mod ern one (Howard, 1947). The ratio of carpometacarpus to femur length of recent Aq uila chrysaetos shows considerable variation (range, 0.79-0.85; n=\l). The maximum lengths of these elements given by Howard (1932) for the Rancho La Brea eagle give a ratio of 0.83, or within the range of the modern species. If the carpometacarpus (UF 30015) and femur (UF 30019) of A. biv ia are assumed to be from the same individual, they give a ra tio of 0.85 and are thus within the range of modern A. chrysa etos. The tibiotarsus (UF 30012) of A. bivia, however, is longer and proportionally more gracile compared to A. chry saetos (Table 2), suggesting that the fossil species had rela tively longer and more slender legs compared to other eagles of this genus. This large eagle is the first valid fossil species of Aquila to be described from North America. Two named species, A.ferox and A. lydekkeri, described from North America by Shufeldt (1915), are now recognized as synonyms of Minerva antiqua Shufeldt, an Eocene owl in the extinct family Protostrigidae (Olson, 1985). Aquila borrasi, described from the late Pleis tocene of Cuba by Arredondo (1970, 1976), was discussed by Olson and Hilgartner (1982), who suggested that it may be re lated to the large, extinct hawk Titanohierax gloveralleni Wet more (1937) of the Bahamas. They also suggested that none of the Cuban material was properly placed in the genus Aquila. NUMBER 89 193 We compared the paratypical femur of A. borrasi (Arredondo, 1970, fig. 7) to femora of A. bivia and found it to differ in its relatively greater size, shaft more distinctly flared toward the ends, and the relatively large proximal pneumatic foramen. This femur does not appear to represent any living genus of hawk, eagle, or vulture, and we agree with Olson and Hilgart ner (1982) that it probably is referable to Titanohierax. Bickart referred six specimens to Aquila sp. A and sp. B from the late Miocene/early Pliocene Big Sandy Formation, Arizona. These specimens are described as equal in size to or smaller than A. chrysaetos and probably do not represent A. bivia. Two other fossil species, Aquila delphinensis and A. penna- to'ides, described by Gaillard (1938), are known from the late Miocene of France, each only by the proximal end of the tar sometatarsus These specimens were not available for comparison in this study, but their geographic location and age suggest that the In glis fossils would not be referable to either of these species. Discussion The two new eagles described herein add to a growing list of living and extinct birds that indicate a former habitat corridor, extending from the Florida peninsula to western North Ameri ca, that probably developed in the late Pliocene when climatic changes allowed xerophytes from the south to move northward and those from the north to move southward (Blair, 1958; Ax- elrod, 1979; Simpson and Neff, 1985). The resulting habitat ap parently was a dry, thorn-scrub community and savannah as suggested by fossil and recent plant and animal distributions (Blair, 1958; Axelrod, 1979; Simpson and Neff, 1985) and was an important corridor for biotic dispersal during the Great American Biotic Interchange (Stehli and Webb, 1985). Based on topography, probable avenues of northward dis persal of Neotropical elements into the southwestern and southeastern United States were along the coastal lowland cor ridors on the eastern and western margins of mainland Mexico (i.e., below the Sierra Madre Oriental along the Gulf of Mexico and below the Sierra Madre Occidental along the Gulf of Cali fornia; Figure 6). From these areas of entry, late Blancan and early Irvingtonian invaders from the tropics spread into savan nah habitats on the Mexican Plateau and in the present-day southern United States. They moved especially into the south eastern United States (particularly the Florida peninsula, where the vertebrate fossil record is best known, but later as far north as South Carolina) but also into the southern Great Plains and to a lesser extent into present-day Sonora, Mexico, and south ern California, probably via the western corridor. Some of these taxa dispersed as far north as present-day Idaho, although the greatest diversity extends no farther north than the Texas panhandle (Figure 6). Before now, the greatest evidence for the Gulf Coast corridor was shown primarily by mammalian faunas of the late Blancan and early Irvingtonian Land Mammal Ages in North America (Webb and Wilkins, 1984; Morgan, 1991; Figure 6, Table 3). Florida fossil faunas are characterized more by Neotropical in fluences during this period than by northern or western faunal elements (Webb and Wilkins, 1984; Morgan, 1991). The fossil herpetofauna from Inglis IA, however, indicates greater influ ence from xeric habitats in the western United States during the early phase of the Plio-Pleistocene (15 of 31 species identified; Meylan, 1982). Less has been documented in relation to fossil avifaunas in the Plio-Pleistocene, but evidence so far indicates similar dis- peral routes and timing as for the mammals. Vuilleumeir (1985) found that representatives of only three South American groups, the teratorns {Teratornis spp.), caracaras {Caracara plancus (J.F. Miller) and Milvago readei (Brodkorb)), and pho- rusrhacids {Titanis walleri), are known from the fossil record of Florida, and that there was a greater influence of North American taxa on South American avifaunas than the reverse. New fossil records indicate that additional extant Neotropical taxa, representing lowland forest and aquatic habitats, inhabit ed Florida during the Plio-Pleistocene. These taxa include the Least Grebe {Tachybaptus dominicus (Linnaeus)) and Great Black-hawk {Buteogallus urubitinga (Gmelin)) from Inglis 1 A, Ringed Kingfisher {Ceryle torquata (Linnaeus)) from Haile 7C, and Gray-breasted Crake (cf. Laterallus exilis (Tem minck)) from Haile 16A (early Irvingtonian, 1.6-1.0 Ma; Carr, 1981; Emslie, 1998). Other extant species that first appear in the late Pliocene (Inglis IA) of Florida that have populations or closely related species in western North America are discussed by Emslie (1996, 1998). Extinct birds from the late Pliocene and early Pleistocene of Florida that reflect a common habitat between the peninsula and the western United States include the first record of an ex tinct cormorant {Phalacrocorax idahensis Marsh) from Flori da (Emslie, 1998) and two species of pygmy-owls {Glaucidi- um spp.), which represent the first occurrence of this genus in eastern North America (Carr, 1981; Emslie, 1998; Table 3). In addition, teratorns {Teratornis spp.) first appear in Florida and the western United States in the late Blancan and represent a group that probably originated in South America (Campbell and Tonni, 1981; Emslie, 1988). Other taxa that arrived in the peninsula during this period include condors {Gymnogyps spp.), an extinct accipitrid vulture {Neophrontops slaughteri Feduccia), Aquila bivia and Amplibuteo concordatus, de scribed herein, a tropical hawk-eagle {Spizaetus sp.), an unde scribed chachalaca (Cracidae, indet.), and an extinct turkey {Meleagris leopoldi A.H. Miller and Bowman/M anza Howard) (Steadman, 1980; Carr, 1981; Emslie, 1988, 1992, 1998; Table 3). As with the mammals, these extant and extinct taxa provide strong evidence that xeric, thorn-scrub and savannah habitats once existed between the Florida peninsula and western North America. Other species of mammals and birds that appeared in North America during this time, however, reflect aquatic and lowland tropical forest environments (Table 3). The presence of 194 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 3.?Occurrence of mammalian taxa of presumed South American origin, and birds with Neotropical and western affinities, in faunas of late Pliocene to early Pleistocene (late Blancan to early Irvingtonian; 3.7-1.0 Ma) age in western North America and the Florida peninsula. References used to compile this table include Akersten (1972), Carranza-Castaneda and Miller (1988), Con rad (1980), Dalquest (1975), Downs and White (1968), Emslie (1988, 1992, 1995, 1998), Frazier (1981), Galusha et al. (1984), Gillette and Ray (1981), Hager (1974), Hirschfeld and Webb (1968), Hulbert (1992, 1997), Hulbert and Morgan (1993), Jefferson (1989), Johnson et al. (1975), Johnston and Savage (1955), Lindsay (1978, 1984a, 1984b), Lindsay and Tessman (1974), Lundelius et al. (1987), Miller and Carranza-Castaneda (1984), Montellano-Ballesteros and Carranza-Castaneda (1986), Morgan (1991), Mor gan and Hulbert (1995), Opdyke et al. (1977), Robertson (1976), Schultz (1977, 1990), Schultz (1937), Seymour (1993), Skinner and Hibbard (1972), Tomida (1987), and Webb and Wilkins (1984). Inferred general habitat requirements for each taxon is indi cated by superscript numbers as follows: 'thorn-scrub and savannah, Rowland tropical forest and/or hammock, and 3aquatic or semiaquatic. Habitat assignment is based on that of living counterparts and/or paleoecological and paleobiological information provided in American Ornithologists' Union (1998), Brown and Amadon (1968), Campbell and Tonni (1981), Delacour and Ama don (1973), Downing and White (1995), Kurten and Anderson (1980), McDonald (1995), and Steadman (1980). (AZ=Arizona, CA=Califomia, CO=Colorado, ID=Idaho, MX=Mexico, TX=Texas.) Taxon Florida Western North America +* ?- 3 ~ N < Knolls, N < < U X 1, T :d Corra * speth, T X S ta Clara X 2 ilancan' ,CO Q igerman 75 (* e < P> (IN O ?. ^ ? ? = := o ac x ?s OH a H= X J CO T3 Mammals Dasypus bellus' Holmesinafloridanus' Glyptotherium arizonae3 Glyptotherium texanum3 Pachyarmatherium leiseyi '?2 Glossotherium chapadmalense' Glossotherium garbanii' Glossotherium sp.1 Paramylodon harlani' Megalonyx leptostomus2 Megalonyx wheatleyi2 Megalonyx sp.2 Eremotherium sp.2 Nothrotheriops texanus Nothrotheriops sp.1 Myrmecophaga tridactyla' Erethizon bathygnathum2 Erethizon kleini2 Erethizon dorsatum Erethizon poyeri2 Neochoerus dichroplax2,3 Neochoerus cordobai2'3 Neochoerus sp.2'3 Hydrochaeris holmesi2'3 Birds Phalacrocorax idahensis* Teratornis incredibilis^3 Teratornis merriami1-3 Gymnogyps kofordi' Gymnogyps sp.' Neophrontops slaughteri' Amplibuteo concordatus '?2 Aquila bivia' Spizaetus sp.2 Cracidae, indet.1 Meleagris leopoldi/anza1'2 Titanis walleriy Glaucidium explorator^2 Glaucidium sp.1,2 X X X X XXX xxxxxxxx X XXX X XXX X X X X X X XXX X XXX X X X X X X X X a Wolf Ranch, California Wash, Cal Tech, Benson, Mendevil Ranch, McRae Wash, Curtis Ranch b Upper Arroyo Seco, lower Vallecito Creek NUMBER 89 195 FIGURE 6.?Distribution of late Blancan to early Irvingtonian (ca. 2.5-1.0 Ma) local faunas in North America. Size of dot or circle indicates number of species of mammals of South American origin (Xenarthra, Caviomorpha) in the fauna. Local faunas include those listed in Table 3 plus Wellsch Valley, Saskatchewan; Delmont, South Dakota; Big Springs, Nebraska; Kentuck, Kansas; and Anita, Arizona. Arrows signify likely corridors of dis persal; the Gulf of Mexico terrestrial corridor would have been widened onto the continental shelf during glacial periods. these taxa suggests that the Gulf Coast corridor was composed of a mosaic of communities, including dry thorn-scrub, ham mocks, and aquatic zones (lakes and wetlands). Such a broad zone having patches of dry to moist habitats is unlike any such region today and probably developed in response to unusual cli matic conditions during glacial intervals in the Plio-Pleistocene. Conclusion The record of birds in North America during the Great American Biotic Interchange indicates patterns for timing and dispersals that are similar to those known for other vertebrates and for plants. As with the mammals, the birds are largely presumed to represent a xeric, thorn-scrub and savannah envi- 196 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ronment that developed from the southwestern United States and northern Mexico across to the Florida peninsula during glacial intervals in the Pilocene and Pleistocene. This corridor probably accounts for the distributions of extant and extinct birds, including the two new eagles described herein, between the western United States and Florida. Other bird and mam mal species suggest that a mosaic of habitats, including ham mocks, lakes, and wetlands, also existed in this corridor. Sub sequent climate change during interglacial intervals fragmented this broad habitat with concomitant isolation and extirpation of populations in the western United States and the Florida peninsula. Literature Cited Akersten, W.A. 1972. Red Light Local Fauna (Blancan) of the Love Formation, Southeast- em Hudspeth County, Texas. Bulletin of the Texas Memorial Mu seum, 20:1-53. American Ornithologists' Union 1998. The AO. U. Check-list of North American Birds. Seventh Edition, liv+829 pages. Lawrence, Kansas: American Ornithologists' Union. Arredondo, O. 1970. Neuva especie de ave Pleistocenica del orden Accipitriformes (Ac- cipitridae) y neuvo genero para las Antillas. Ciencias, series 4, 8:1-19. 1976. The Great Predatory Birds of the Pleistocene of Cuba. In Storrs L. Olson, editor, Collected Papers in Avian Paleontology Honoring the 90th Birthday of Alexander Wetmore. Smithsonian Contributions to Paleobiology, 27:169-187. Axelrod, D.I. 1979. Age and Origin of Sonoran Desert Vegetation. Occasional Papers of the California Academy of Sciences, 132:1-74. Baskin, J.A. 1995. The Giant Flightless Bird Titanis walleri (Aves: Phorusrhacidae) from the Pleistocene Coastal Plain of South Texas. Journal of Verte brate Paleontology, 15:842-844. Blair, W. Frank 1958. Distributional Patterns of Vertebrates in the Southern United States in Relation to Past and Present Environments. In CL. Hubbs, editor, Zoogeography, pages 433^168. Washington, D.C: American Asso ciation for the Advancement of Science. Brodkorb, P. 1963. A Giant Flightless Bird from the Pleistocene of Florida. Auk, 80:111-115. Brown, L., and D. Amadon 1968. Eagles, Hawks and Falcons of the World. 2 volumes, 946 pages. New York: McGraw-Hill. Campbell, K.E., Jr. 1979. The Non-Passerine Pleistocene Avifauna of the Talara Tar Seeps, Northwestern Peru. Contributions, Life Sciences Division, Royal Ontario Museum, 118:1 -203. Campbell, K.E., Jr., and E.P. Tonni 1981. Preliminary Observations on the Paleobiology and Evolution of Ter atorns (Aves: Teratornithidae). Journal of Vertebrate Paleontology, 1:265-272. Carr, Gail Speaker 1981. An Early Pleistocene Avifauna from Inglis, Florida. Doctoral disser tation, Department of Zoology, University of Florida, Gainesville. Carranza-Castaneda, O., and W.E. Miller 1988. Roedores caviomorfos de la mesa central de Mexico, Blancano tem- prano (Plioceno tardio) de la fauna local Rancho Viejo, estado de Guanajuato. Revista, Universidad Nacional Autonoma de Mexico, Insituto de Geologia, 7:182-199. Conrad, G.S. 1980. The Biostratigraphy and Mammalian Paleontology of the Glenns Ferry Formation from Hammett to Oreana, Idaho. Doctoral disserta tion, Idaho State University, Pocatello. Dalquest, W.W. 1975. Vertebrate Fossils from the Blanco Local Fauna of Texas. Occa sional Papers, The Museum, Texas Tech University, 30:10-52. Delacour, J., and D. Amadon 1973. Curassows and Related Birds. xv+247 pages. New York: American Museum of Natural History. Downing, K.F., and R. White 1995. The Cingulates (Xenarthra) of Leisey Shell Pit IA (Irvingtonian), Hillsborough County, Florida. Bulletin of the Florida Museum of Natural History, 37(2):375-396. Downs, T, and J.A. White 1968. A Vertebrate Faunal Succession in Superposed Sediments from Late Pliocene to Middle Pleistocene in California. Proceedings of the XXIII International Geological Congress, Prague, 10:41?47. Emslie, S.D. 1988. The Fossil History and Phylogenetic Relationships of Condors (Ciconiiformes: Vulturidae) in the New World. Journal of Verte brate Paleontology, 8:212-228. 1992. Two New Late Blancan Avifaunas from Florida and the Extinction of Wetland Birds in the Plio-Pleistocene. Natural History Museum of Los Angeles County Science Series, 36:249-269. 1995. An Early Irvingtonian Avifauna from Leisey Shell Pit, Florida. Bul letin of the Florida Museum of Natural History, 37(l):299-344. 1996. A Fossil Scrub-Jay Supports a Recent Systematic Decision. Condor, 98:675-680. 1998. Avian Community, Climate, and Sea-Level Changes in the Plio- Pleistocene of the Florida Peninsula. Ornithological Monographs, 50:1-113. Frazier, M.K. 1981. A Revision of the Fossil Erethizontidae of North America. Bulletin of the Florida State Museum, Biological Sciences, 27:1-76. Gaillard, C. 1938. Contribution a l'etude des oiseaux fossiles. Archives du Museum d'Histiore Naturelle de Lyon, 15: 100 pages. Galusha, T, N.M. Johnson, E.H. Lindsay, N.D. Opdyke, and R.H. Tedford 1984. Biostratigraphy and Magnetostratigraphy, Late Pliocene Rocks, 111 Ranch, Arizona. Bulletin of the Geological Society of America, 95:714-722. Gillette, D.D., and CE. Ray 1981. Glyptodonts of North America. Smithsonian Contributions to Paleo biology, 40: 255 pages. Hager, M.W. 1974. Late Pliocene and Pleistocene History of the Donnelly Ranch Verte brate Site, Southeastern Colorado. Contributions to Geology, Uni versity of Wyoming, Special Paper, 2:1-62. Hirschfeld, S.E., and S.D. Webb 1968. Plio-Pleistocene Megalonychid Sloths of North America. Bulletin of NUMBER 89 197 the Florida State Museum, Biological Sciences, 12:213-296. Howard, H. 1929. The Avifauna of Emeryville Shellmound. University of California Publications in Zoology, 32:301-394. 1932. Eagles and Eagle-like Vultures of the Pleistocene of Rancho La Brea. Contributions in Palaeontology from Carnegie Institution of Washington, 429: iii+82 pages. 1947. An Ancestral Golden Eagle Raises a Question in Taxonomy. Auk, 64:287-291. Hulbert, R.C, Jr. 1992. A Checklist of the Fossil Vertebrates of Florida. Papers in Florida Paleontology, 6:1-35. 1997. A New Late Pliocene Porcupine (Rodentia: Erethizontidae) from Florida. Journal of Vertebrate Paleontology, 17:623-626. Hulbert, R.C., Jr., and G.S. Morgan 1993. Quantitative and Qualitative Evolution in the Giant Armadillo Holmesina (Edentata: Pampatheriidae) in Florida. In R.A. Martin and A.D. Barnosky, editors, Morphological Change in Quaternary Mammals of North America, pages 134-177. Cambridge: Cam bridge 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, 86:5-12. Jefferson, G.T. 1989. Late Cenozoic Tapirs (Mammalia: Perissodactyla) of Western North America. Contributions in Science, Natural History Museum of Los Angeles County, 406:1-21. Johnson, N.M., N.D. Opdyke, and EH. Lindsay 1975. Magnetic Polarity Stratigraphy of Pliocene-Pleistocene Terrestrial Deposits and Vertebrate Faunas, San Pedro Valley, Arizona. Bulle tin of the Geological Society of America, 86:5-12. Johnston, C.S., and D.E. Savage 1955. A Survey of Various Late Cenozoic Vertebrate Faunas of the Pan handle of Texas, Part 1: Introduction, Description of Localities, Pre liminary Faunal Lists. University of California Publications in Geological Sciences, 31:27-50. Kurten, B., and E. Anderson 1980. Pleistocene Mammals of North America, xvii+442 pages. New York: Columbia University Press. Lindsay, E.H. 1978. Late Cenozoic Vertebrate Faunas, Southeastern Arizona. New Mex ico Geological Society Guidebook, 29th Field Conference, pages 269-275. 1984a. Late Cenozoic Mammals from Northwestern Mexico. Journal of Vertebrate Paleontology, 4:208-215. 1984b. Windows to the Past: Fossils of the San Pedro Valley. Fieldnotes from the Arizona Bureau of Geology and Mineral Technology, 14: 1-9, 12. Lindsay, E.H., and N.T. Tessman 1974. Cenozoic Vertebrate Localities and Faunas in Arizona. Journal of the Arizona Academy of Science, 9:3-24. Lundelius, E.L., T. Downs, E.H. Lindsay, H.A. Semken, R.J. Zakrzewski, CS. Churcher, CR. Harington, G.E. Schultz, and S.D. Webb 1987. The North American Quaternary Sequence. In M.O. Woodburne, editor, Cenozoic Mammals of North America; Geochronology and Biostratigraphy, pages 211-235. Berkeley: University of California Press. Mares, Michael A. 1985. Mammal Faunas of Xeric Habitats and the Great American Inter change. In F.G. Stehli and S.D. Webb, editors, The Great American Biotic Interchange, pages 489-520. New York: Plenum Press. Marshall, L.G. 1985. Geochronology and Land-Mammal Biochronology of the Tran- samerican Faunal Interchange. In F.G. Stehli and S.D. Webb, edi tors, The Great American Biotic Interchange, pages 49-85. New York: Plenum Press. Marshall, L.G., S. David Webb, J.J. Sepkoski, Jr., and D.M. Raup 1982. Mammalian Evolution and the Great American Interchange. Sci- ence, 215:1351-1357. McDonald, H.G. 1995. Gravigrade Xenarthrans from the Early Pleistocene Leisey Shell Pit 1 A, Hillsborough County, Florida. Bulletin of the Florida Museum of Natural History, 37(2):345-373. Meylan, PA. 1982. The Squamate Reptiles of the Inglis IA Fauna (Irvingtonian), Citrus County, Florida. Bulletin of the Florida State Museum, Biological Sciences, 27:1-85. Miller, L.H. 1911. A Series of Eagle Tarsi from the Pleistocene of Rancho La Brea. University of California Publications, Bulletin of the Department of Geology, 6:305-316. Miller, W.E., and O. Carranza-Castaneda 1984. Late Cenozoic Mammals from Central Mexico. Journal of Verte brate Paleontology, 2:95-107. Montellano-Ballesteros, M., and O. Carranza-Castaneda 1986. Descripcion de un milodontido del Blancano temprano de la mesa central de Mexico. Revista, Universidad Nacional Autonoma de Mexico, Instituto de Geologia, 6:193-213. Morgan, G.S. 1991. Neotropical Chiroptera from the Pliocene and Pleistocene of Flor ida. Bulletin of the American Museum of Natural History, 206: 176-213. Morgan, G.S., and R. Hulbert, Jr. 1995. Overview of the Geology and Vertebrate Biochronology of the Lei sey Shell Pit Local Fauna, Hillsborough County, Florida. Bulletin of the Florida Museum of Natural History, 37( 1): 1-92. Neill, W.T. 1957. Historical Biogeography of Present-Day Florida. Bulletin of the Florida State Museum, Biological Sciences, 2:175-220. Olson, S. 1985. The Fossil Record of Birds. In D.S. Farner, J.R. King, and K.C. Parkes, editors, Avian Biology, 8:80-256. New York: Academic Press. Olson, S., and W.B. Hilgartner 1982. Fossil and Subfossil Birds from the Bahamas. In Storrs L. Olson, ed itor, Fossil Vertebrates from the Bahamas. Smithsonian Contribu tions to Paleobiology, 48:22-56. Opdyke, N.D., E.H. Lindsay, N.M. Johnson, and T. Downs 1977. The Paleomagnetism and Magnetic Polarity Stratigraphy of the Mammal-Bearing Section of Anza-Borrego State Park, California. Quaternary Research, 7:316?329. Repenning, C 1987. Biochronology of the Microtine Rodents of the United States. In M.O. Woodburne, editor, Cenozoic Mammals of North America, Geochronology and Biostratigraphy, pages 236-268. Berkeley, Cal ifornia: University of California Press. Robertson, J.S. 1976. Latest Pliocene Mammals from Haile 15A, Alachua County, Flor ida. Bulletin of the Florida State Museum, Biological Sciences, 20:111-186. Schultz, G.E., editor 1977. Guidebook: Field Conference on Late Cenozoic Biostratigraphy of the Texas Panhandle and Adjacent Oklahoma, August 4-6, 1977. In West Texas State University, Department of Geology and Anthro pology, Killgore Research Center, Special Publication, 1: vii+160 pages. 1990. Stop 5: Blanco Local Fauna and the Blancan Land Mammal Age. In T.C. Gustavoson, editor, Tertiary and Quaternary Stratigraphy and Vertebrate Paleontology of Parts of Northwestern Texas and East- 198 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ern New Mexico, pages 44-51. Austin: University of Texas. Schultz, J.R. 1937. A Late Cenozic Vertebrate Fauna from the Coso Mountains, Inyo County, California. Contributions in Paleontology, from Carnegie Institution of Washington, 487:75-109. Seymour, K. 1993. Size Change in North American Quaternary Jaguars. In R.A. Martin and A.D. Bamosky, editors, Morphological Change in Quaternary Mammals of North America, pages 343-372. Cambridge: Cam bridge University Press. Shufeldt, R.W. 1915. Fossil Birds in the Marsh Collection of Yale University. Transac tions of the Connecticut Academy of Arts and Sciences, 19:1-110. Sibley C.G., and B.L. Monroe, Jr. 1990. Distribution and Taxonomy of Birds of the World. 1111 pages. New Haven: Yale University Press. Simpson, B.B., and J.L. Neff 1985. Plants, Their Pollinating Bees, and the Great American Interchange. In F.G. Stehli and S.D. Webb editors, The Great American Biotic Interchange, pages 427-452. New York: Plenum Press. Skinner, M.F., and C.W. Hibbard 1972. Early Pleistocene Pre-glacial and Glacial Rocks and Faunas of North-Central Nebraska. Bulletin of the American Museum of Natu ral History, 148:1-148. Steadman, D.W. 1980. A Review of the Osteology and Paleontology of Turkeys (Aves: Meleagridinae). Contributions in Science, Natural History Museum of Los Angeles County, 330:131-207. Stehli, F.G., and S.D. Webb, editors 1985. The Great American Biotic Interchange, xvii + 532 pages. New York: Plenum Press. Tomida, Y. 1987. Small Mammal Fossils and Correlation of Continental Deposits, Safford and Duncan Basins, Arizona, USA. ix+141 pages. Tokyo: National Science Museum. Vuilleumier, F. 1985. Fossil and Recent Avifaunas and the Interamerican Interchange. In F.G. Stehli and S.D. Webb, editors, The Great American Biotic In terchange, pages 387-424. New York: Plenum Press. Webb, S. David 1985. Late Cenozoic Mammal Dispersals Between the Americas. In F.G. Stehli and S.D. Webb, editors, The Great American Biotic Inter change, pages 357-386. New York: Plenum Press. Webb, S. David, and K. Wilkins 1984. Historical Biogeography of Florida Plistocene Mammals. Special Publications of the Carnegie Museum of Natural History, 8: 370-383. Wetmore, A. 1937. Bird Remains from Cave Deposits on Great Exuma Island in the Ba hamas. Bulletin of the Museum of Comparative Zoology, Harvard, 80:427-441. Wood, D.S., and G.D. Schnell 1986. Revised World Inventory of Avian Skeletal Specimens. 1986. 296 pages. Norman, Oklahoma: American Ornithologists' Union, Okla homa Biological Survey. A New Genus of Dwarf Megapode (Galliformes: Megapodiidae) from the Late Oligocene of Central Australia Walter E. Boles and Tessa J. Ivison ABSTRACT The earliest fossil record of the Megapodiidae is from the Pliocene; most records are from the Quaternary. A recently identi fied megapode from late Oligocene deposits at Lake Pinpa, central Australia, is older than any previously reported megapode taxon and is only about two-thirds the size of the smallest living species. It probably inhabited riparian forests bordered by tropical savanna woodland in the Oligocene environment of Lake Pinpa, occupying a role similar to that of the Orange-footed Scrubfowl, Megapodius reinwardt Dumont, in tropical Australia today. Although the new fossil confirms the presence of the Megapodiidae in Australia as early as the late Oligocene, it provides no information on the ori gins and relationships of the family, or on the evolution of its dis tinctive method of incubation. Introduction The megapodes, or mound-builders, are among the more in triguing families of birds because they exhibit the practice, unique among birds, of incubating their eggs through the use of external heat sources (sun, decaying vegetation, volcanic heat, etc.) rather than body heat. The young hatch in a hyperpreco- cial state, receive no parental care from the adults, and are ca pable of flight within a few hours. This family has a circum scribed distribution: with one exception, it occurs only east of Wallace's Line, from Sulawesi through Australo-Papua and to Tonga and Samoa (recently extinct), and north to the Philippine Islands and Micronesia. Several aspects of the biology and evolution of the Megapo diidae have been the subject of ongoing debate, including the relationships of the megapodes to other galliform families (e.g., Walter E. Boles and Tessa J. Ivison, Division of Vertebrate Zoology {Birds), Australian Museum, 6 College Street, Sydney, New South Wales 2000, Australia. Cracraft, 1973; Sibley and Ahlquist, 1990; Brom and Dekker, 1992; Jones et al., 1995), the origin of the family (either the Southern (Gondwanan) or the Northern Hemisphere (Cracraft, 1973; Olson, 1985)), and the factors responsible for its current distribution (whether competitive exclusion by pheasants (Pha- sianidae) (Olson, 1980) or predation by mammalian carnivores (Felidae, Viverridae) (Dekker, 1989)). It is now generally agreed that the megapodes' unusual incubation strategy was acquired secondarily, following the development of typical avi an incubation strategies in early birds (Clark, 1964a, 1964b; Dekker and Brom, 1992). For a current review of these and other topics on the biology, classification, and evolution of the Megapodiidae, see Jones et al. (1995). The 22 or so extant species are classified in six or seven gen era (Sibley and Monroe, 1990; del Hoyo et al., 1994; Jones et al., 1995), although relationships among genera of megapodes are unresolved. A division between the "scrubfowl" and the oth er taxa (Clark, 1964a, 1964b) has not been confirmed (Brom and Dekker, 1992), nor has the position of the fossil genus Progura De Vis, 1888, been examined. The brush-turkeys comprise three genera, Alectura (monotypic, endemic to Australia), Talegalla (three species), and Aepypodius (two species), with the last two restricted to New Guinea. The single species of malleefowl, Leipoa, occurs only in Australia. Among the scrubfowl and their relatives, there are two monotypic genera: Macrocepha- lon, restricted to Sulawesi, Indonesia, and Eulipoa (Moluccas and Misool Island), which is often merged with Megapodius. Megapodius has the widest distribution, occurring in Australia and New Guinea, east to Tonga (with prehistorically extinct Holocene forms from as far as Samoa), north to the Philippine Islands and the Palau and Mariana islands in Micronesia, and west through Sulawesi and Lombok, Indonesia, with isolated populations on the Nicobar Islands. Megapodius also is the most diverse genus, with nine to 13 living species, depending on taxonomy (e.g., Peters, 1934; Sibley and Monroe, 1990; del Hoyo et al., 1994; Roselaar, 1994; Jones et al., 1995). 199 200 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY The range of body sizes in the megapodes varies from the ex tinct Progura gallinacea De Vis (1888) of Australia, which reached the size of a turkey {Meleagris), to Megapodius laper- ouse of Micronesia, the smallest described species (28-30 cm). Steadman (1993a) reported an extinct, as yet undescribed spe cies of Megapodius from the late Quaternary of 'Eua, Tonga, which was smaller than any other known megapode. The earli est known fossil occurrence of the Megapodiidae is from the Pliocene (Boles and Mackness, 1994); most records are from the Pleistocene. Because of its age and diminutive size, a re cently identified megapode from late Oligocene deposits of central Australia, outside the current distribution of this family, is of considerable interest. FOSSIL RECORD OF MEGAPODES Mourer-Chauvire (1982) initially indicated the presence of the Megapodiidae in the Eocene-Oligocene deposits of Quercy, France. Later she regarded these as belonging to a more primi tive family of galliforms, the Quercymegapodiidae (Mourer- Chauvire, 1992), consisting of the single genus Quercymega- podius, which she created for Palaeocryptonyx depereti Gail- lard (1908), and a new species, Q. brodkorbi Mourer-Chauvire (1992). This family was interpreted as being the sister group of all the living Galliformes. Excluding the Quercymegapodiidae, there are no known occurrences of fossil megapodes outside Australia and the islands of the southwest Pacific. The Pleistocene record of megapodes is dominated by the fossil genus Progura. Progura gallinacea was originally de scribed by De Vis (1888) as a large pigeon. Van Tets (1974) recognized that the specimens belonged to a megapode larger than any living species. Additional material identified by De Vis as pigeons, as well as material he identified as storks or as bustards, also was included in this taxon (van Tets, 1974; van Tets and Rich, 1990). Specimens of a smaller but related form were described as P. naracoortensis by van Tets (1974), to which were referred fossils that previously had been attributed to Alectura lathami (Lydekker, 1891; Longman, 1945). It was later suggested (van Tets, 1984) that the two taxa of Progura actually represented a single, sexually dimorphic species. The only other Australian species in the fossil record is Leipoa ocellata, which has been recovered from late Pleistocene de posits in South Australia (van Tets, 1974). The only Tertiary records come from the Pliocene of Australia. De Vis (1889) de scribed the fossil megapode Chosornis praeteritus from Chin chilla, Queensland, which van Tets (1974) later placed in the synonymy of the Pleistocene species Progura gallinacea. Boles and Mackness (1994) reported on the presence of P. cf. naracoortensis at Bluff Downs, Queensland. A variety of fossil species of Megapodius, mostly extinct forms, are known from South Pacific islands from New Cale donia (Balouet and Olson, 1989) eastward into Polynesia (Steadman, 1989, 1993a, 1993b, 1995; see also Jones et al., 1995). Most of these sites are of Holocene age, and most or all of the extinctions were anthropogenic. The enigmatic Sylviomis neocaledoniae of New Caledonia was first described as a ratite (Poplin, 1980) but later was con sidered to be a large, flightless megapode (Poplin and Mourer- Chauvire, 1985). Balouet and Olson (1989) and C. Mourer- Chauvire (pers. comm., 1996) believe that that Sylviomis be longs to the Galliformes but is best placed in its own family. METHODS.?Measurements mostly follow those of Stead man (1980) and were made with digital calipers and rounded to the nearest 0.1 mm. Osteological nomenclature follows Baumel and Witmer (1993); taxonomic nomenclature for the Megapodiidae follows Jones et al. (1995). Comparisons were made with representatives of all extant genera of megapodes {Macrocephalon maleo, Eulipoa wallacii, Alectura lathami, Leipoa ocellata, Talegalla jobiensis, T. fuscirostris, Aepypodi- us arfakianus, A. bruijni, Megapodius reinwardt, M. freycinet, M. eremita, M. pritchardi, M. cumingii, M. nicobarensis) as well as with Progura gallinacea. ACKNOWLEDGMENTS.?We thank T. Rich (Museum of Vic toria) and P. Vickers-Rich (Monash University) for permitting us to work on this fossil; T. Rich for supplying information from his field notes on the discovery of this specimen; D. Steadman for discussion of this fossil and of his work with megapodes; S. Ol son, D. Steadman, and C. Mourer-Chauvire for valuable criti cisms of the manuscript; T. Wickey and C. Bento (Australian Museum) for taking the photographs; R. Schodde and W. Long- more (Australian National Wildlife Collection), M. LeCroy (American Museum of Natural History) and S. Olson and J. Dean (National Museum of Natural History, Smithsonian Insti tution) for access to comparative material; and the Australian Museum for providing a venue in which to carry out this study. Systematic Paleontology Order GALLIFORMES Family MEGAPODIIDAE Because of the gracility of the trochleae of the tarsometatar sus described herein compared with those of some other mega podes, the Pinpa fossil bears a superficial resemblance to the tarsometatarsus of a medium-sized pigeon, such as the Wonga Pigeon, Leucosarcia melanoleuca Latham. Rich et al. (1991) originally placed it in the Columbidae, which is understandable because there are superficial similarities between the tar sometatarsus in these two families, as shown by De Vis's (1888) original description of the fossil megapode Progura gallinacea as a relative of the crowned pigeons {Gourd). Van Tets (1974) discussed characters that differentiate the two groups. The tarsometatarsus in the Megapodiidae may be identified by the following combination of characters: three cristae hypo- tarsi (medialis large) and four sulci hypotarsi (medialis broad), only one of which is enclosed; distal extension of crista media- NUMBER 89 201 lis hypotarsi along the shaft flat and virtually absent; shaft dor- soplantarly compressed; depression on plantar face; fossa metatarsi I distinct; and trochleae metatarsi II and IV with equal distal extension. This suite of features separates the Megapodiidae from other living and most extinct Galliformes. The fossil and other Megapodiidae differ from the Quer cymegapodiidae by having the crista metatarsi medialis pro jecting further plantarly, trochlea metatarsi II not globular, and trochleae metatarsi II and IV with equal distal extension (char acters from Mourer-Chauvire, 1992). In species of the Gallinu- loididae the depression on the plantar surface is more extensive than in the Megapodiidae. The Megapodiidae also differ from the Columbidae by lack ing an indentation on the medial border of the shaft distal to the fossa metatarsi I and by having the sulcus hypotarsi in the same dorsoplantar line as the eminentia intercondylaris (in proximal view) rather than offset laterally; also, the trochlea metatarsi II is more in the same lateromedial line as trochleae metatarsi III and IV (in distal view) instead of being recessed plantarly and rotated medially. The fossil exhibits a suite of characters unlike those found in other genera of the Megapodiidae, for which reason it is recog nized as a new genus. Ngawupodius, new genus TYPE SPECIES.?Ngawupodius minya, new species, by origi nal designation and monotypy. ETYMOLOGY.?Ngawu, in South Australia an Aboriginal name for the Malleefowl Leipoa ocellata (see Peter and Peter, 1993), and -podius, Latinized Greek, "footed," in allusion to the similarities between the tarsometarsi of these taxa. DIAGNOSIS.?The proximal end is medially flared less than in Megapodius or Progura but is more so than in Alectura. The shape and size of the cotyla medialis are about the same as the cotyla lateralis in proximal view; the dorsal rim is not produced as far dorsally, and the medial rim is thin, unlike the conditions in Leipoa or Alectura. The sulcus hypotarsi is proportionally small in proximal view compared with all the modern taxa. The robustness of the lateral side of the hypotarsus in proximal view is blocky, unlike that in Megapodius. The plantar exten sion of the crista medialis hypotarsi is greater than in Alectura, Talegalla, Aepypodius, or Macrocephalon but is less than in Megapodius, Eulipoa, or Leipoa. The relative plantar extension of the crista lateralis hypotarsi differs from that in Alectura, Talegalla, Aepypodius, or Progura in being about one-half that of the crista medialis. The distally projecting process on the distal end of the hypotarsus is smaller than that in Megapodius. The dorsoventral compression of the shaft is greater than in Ta legalla or Aepypodius. The sides of the shaft are relatively par allel and do not widen toward the distal end, unlike Alectura, Leipoa, Aepypodius, Talegalla, Macrocephalon, or Progura. The relative development of the tuberositas M. tibialis cranialis is shorter in the fossil than in the modern forms. The fossa metatarsi I is not as distinct as in Megapodius, Leipoa, or Progura and has little medial extension. The trochleae are gracile and are not swollen as in the other genera, particularly Megapodius. The trochlea metatarsi II is at the same level as the trochlea metatarsi IV, rather than slightly above, when viewed either distally (unlike Megapodius, Leipoa, or Talegal la) or dorsally (unlike Megapodius, Leipoa, Aepypodius, or Progura). The trochlea metatarsi II is neither inflated nor glob ular as it is in Megapodius or Eulipoa, and it does not diverge strongly medially, unlike Alectura, Leipoa, Aepypodius, Tale galla, Macrocephalon, or Progura. The trochlea metatarsi IV does not project laterally, unlike Megapodius or Eulipoa. The articular groove of the trochlea metatarsi IV is moderately well developed, more so than in Alectura, Leipoa, Aepypodius, Ta legalla, or Progura, but is less distinct than in Megapodius. Ngawupodius minya, new species FIGURE 1A-C HOLOTYPE.?Complete right tarsometatarsus, Paleontology Collection of the Museum of Victoria (MV), Melbourne, P160493. Collected by I. Parker on 12 May 1979. TYPE LOCALITY.?South end of Lake Pinpa, South Australia (31?08'31"S, 140?12'47"E; T.H. Rich, pers. comm., 1996). HORIZON.?Namba Formation, Ericmas Fauna, late Oli gocene. ETYMOLOGY.?Minya, in South Australia an Aboriginal word for "small" (Reed, 1977); for the purposes of nomencla ture, minya should be considered to be without gender. DIAGNOSIS.?As for the genus. DESCRIPTION.?The fossil is from an adult bird because the surface lacks porosity, the tarsal cap is completely fused to the proximal end of the metatarsals, the hypotarsus is fully formed, and the metatarsals are fully fused over their entire length. In immature Leipoa ocellata, the tarsometatarsi may approach the size in that of adult birds yet still be highly porous and incom pletely fused. Total length 40.2 mm; proximal width 8.1 mm; distal width 9.5 mm; distal depth 5.7 mm. Eminentia intercondylaris low, about same height as rim of cotyla medialis; both higher than rim of cotyla lateralis. Depth of crista hypotarsi medialis about one-half that of proximal end of bone (proximal depth to hypo tarsus 4.4 mm; proximal depth with hypotarsus 8.0 mm), with laterally directed projection at end. In proximal view, sulcus medialis hypotarsi slightly laterad of eminentia intercondylaris, three shallow sulci lateralis hypotarsi and two low cristae latera lis hypotarsi on rather square hypotarsus. In lateral view, hypo tarsus short (5.1 mm). Ridge extending distally from hypotarsus on facies plantaris very low. Shaft broad (midshaft width, 4.3 mm; shaft width at proximal edge of fossa metatarsi I, 4.5 mm) but dorsoplantarly flattened (midshaft depth, 2.6 mm; shaft depth at proximal edge of fossa metatarsi I, 2.8 mm), markedly so on medial side; lateral margin straight, medial margin some what concave distally to fossa metatarsi I; distal to fossa pro duced more medially. Fossa metatarsi I elongate (5.8 mm). Fos- 202 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?Tarsometatarsus of fossil and recent megapodes. A-c, Ngawupodius minya (holotype, MV P160493), dorsal, plantar, and proximal views, respectively. D-H, dorsal views, recent species: D, Leipoa ocellata; E, Alec tura lathami; F, Aepypodius arfakianus; G, Talegalla jobiensis; H, Megapodius reinwardt. Bars= 10 mm. sa supratrochlearis plantaris shallow, extending from between trochleae on distal end of plantar surface proximally to about proximal end of facet. Trochlea metatarsi II strongly divergent, trochlea metatarsi IV less so; distal and plantar extensions about equal; articular grooves obsolete. Trochlea metatarsi III about twice as long as other trochleae, strongly grooved. NUMBER 89 203 The fossil is larger than the known tarsometatarsal speci mens of Quercymegapodius {Q. depereti: total length 30.0 mm, distal width 5.8 mm, distal depth 3.75 mm; Mourer-Chauvire, 1992). Discussion GEOGRAPHIC SETTING AND GEOLOGY Northeastern South Australia has produced a number of im portant fossil sites, ranging in age from Oligocene-Miocene to Pleistocene, many of which have yielded avian remains. The fossil megapode was recovered from Lake Pinpa, one of sever al localities in the Tarkarooloo Basin where outcrops of the Namba Formation are exposed. The formation is divisible into two members, the upper resting disconformably on the lower. Green claystones and dolomitic claystones at the top of the lower member have yielded vertebrate remains designated as the Pinpa Fauna. A sequence of thin-bedded, fine- to medium- grained sands cut into the lower member, and it is these basal sands of the upper member that have produced the Ericmas Fauna (type locality Ericmas Quarry, Lake Namba; 3ri2'S,140?14'E), from which the megapode bone was recov ered. For details of the geology, dating, and other vertebrate re mains, see Callen and Tedford (1976), Tedford et al. (1977), Woodburne et al. (1985), and references therein. The age of the Ericmas Fauna was originally placed at mid dle Miocene on the basis of its position above the Pinpa Fauna, and, where there are comparable species, species in the Eric mas Fauna appear less primitive. The lower member of the Namba Formation contains pollen floras similar to Balcombi- an-Batesfordian (middle Miocene) deposits in Victoria and South Australia. Tedford et al. (1977) put a maximum age of ca.14-16 Ma on vertebrates higher in the formation. Wood burne et al. (1985) considered the Ericmas Fauna to be middle Miocene in age. Subsequent work, however, has led to a revi sion of these dates. Studies on the central Australian Etadunna Formation (East Lake Eyre Basin), considered to be roughly contemporaneous with the Namba Formation, led Woodburne et al. (1993) to place its age at late Oligocene. These authors considered the Ditjimanka Fauna (Lake Palankarinna) from the Etadunna Formation and the Ericmas Fauna to be "approxi mate correlatives" (ca. 24-26 Ma). The Lake Pinpa site, like most others in central Australia, is characterized by lacustrine/fluviatile deposits. The bones have been disarticulated post-mortem and have been transported varying distances. The faunal summary by Rich et al. (1991) showed that there was a large aquatic component, represented by several species of lungfish {Neoceratodus), teleost fish, che- lid turtles, crocodiles, the primitive platypus Obdurodon insig- nis Woodburne and Tedford (1975), and a dolphin (Rhab- dosteidae). Terrestrial forms included marsupials of the families Dasyuridae, Phascolarctidae, Diprotodontidae, Pseu- docheiridae, and Petauridae. The only other bird thus far re ported was assigned to the Anseriformes (Rich et al., 1991) and has not yet been studied. In contrast, the Pinpa Fauna has abun dant bird remains, including grebes, pelicans, cormorants, wa terfowl, rails, burhinids, and flamingos. RECONSTRUCTION OF Ngawupodius AND ITS ENVIRONMENT Living megapodes were used as the basis of an attempt to re construct the general size and proportions of Ngawupodius. Be cause skeletons of certain taxa were not available, published tarsal measurements from skins were substituted; these give close approximations of the length of the tarsometatarsus. Weights and body lengths are less precise measurements but can serve as approximate indicators of size and permit some rough values to be obtained; data were taken from Marchant and Higgins (1993), Jones et al. (1995), and specimens. The tarsometatarsus of Ngawupodius is smaller in absolute terms than those of other described taxa (Table 1; Figure 1). It TABLE 1.--Measurements ( giving mean (x), range, and Species Ngawupodius minya Megapodius reinwardt Megapodius eremita Megapodius freycinet Megapodius pritchardi Megapodius cumingii Megapodius nicobarensis Eulipoa wallacii Macrocephalon maleo Alectura lathami Talegalla jobiensis Talegalla fuscirostris Aepypodius bruijni Aepypodius arfakianus Leipoa ocellata Progura gallinacea X 40.2 59.8 64.6 69.2 58.6 67.0 68.2 61.2 87.6 96.9 87.7 - 108.4 97.0 75.3 - Total length range - - - - 58.1-59.0 61.1-72.8 - - 86.6-88.6 88.2-103.3 85.0-90.4 - - - 71.2-76.7 - mm sarr n 1 1 1 1 2 2 1 1 2 5 2 - 1 1 3 - ) of the tarsometatarsus in Ngawupodius minya pie size (n). X 8.1 10.6 11.9 12.7 9.2 11.6 12.9 10.2 16.0 17.6 14.8 - 16.2 14.8 15.9 25.5 Proximal width range - - - - 9.0-9.4 10.3-12.9 - - 15.9-16.0 16.5-18.6 14.7-14.8 - - - 15.3-16.3 - n 1 1 1 1 2 2 1 1 2 5 2 - 1 1 3 1 X 9.5 12.2 13.7 14.0 10.8 13.5 14.3 12.3 16.7 18.1 16.3 16.0 17.9 16.0 17.5 28.4 and recent species of Distal width range - - - - 10.8 12.2-14.8 - - 16.1-17.2 16.5-18.7 15.4-17.1 - - - 16.1-18.3 - n 1 1 1 1 2 2 1 1 2 6 2 1 1 1 3 1 megapodes, X 4.3 4.5 5.2 5.4 3.9 5.4 5.9 4.5 6.2 7.3 6.3 5.3 6.8 6.2 6.8 - Midshaft width range - - - - 3.8-3.9 4.6-6.1 - - 6.0-6.3 7.1-7.7 5.8-6.7 - - - 6.5-7.1 - n 1 1 1 1 2 2 1 1 2 3 2 1 1 1 2 - 204 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY is about 80% of the external tarsal length of females of Mega podius laperouse laperouse (50-55 mm), the smallest living species of megapode. The proximal and distal widths of the tar sometatarsus of Ngawupodius are proportionally great com pared to its total length: about 21% for both measurements, at the top of the range among megapodes (15%-21%). Likewise, Ngawupodius has a proportionally wide middle shaft (10.6% of total length, range of megapodes 6%-ll%; 45% of distal width, 38%-45% in other species). In these proportions Ngawupodius resembles only Leipoa (Figure ID). Compari sons with Progura are not possible from existing material. The structure and length of the tarsometatarsus in Leipoa may be related to its more open habitat, and the resemblances to Ngawupodius could be coincidental. Given what is known of living species, a general range of size can be proposed for Ngawupodius. If it had disproportion- ally short legs, as in Leipoa, then it may have been similar in body length and weight to the smallest species of Megapodius. If the leg proportions were more typical of the other megapo des, then estimates of about 225-235 mm overall length and 230-330 g are reasonable. The small, undescribed species of Megapodius from 'Eua, Tonga, appears to have been of com parable size to Ngawupodius minya (D. Steadman, pers. comm., 1996). On the basis of pollen of grasses and subtropical rainforest flora from the lower member of the Namba Formation, Tedford et al. (1977:56) suggested a paleohabitat of "riparian forests with savannas on better drained fluviatiles." Some related but specifically distinct mammals in the faunas of the two mem bers were inferred by Tedford et al. (1977) to indicate that the division between them recorded both a change in the deposi tional environment and a significant time gap. If the general vegetation in the upper member remained much the same as that suggested by the pollen from the lower member, then it is tempting to envisage a situation much like that in northern Australia today. The Orange-footed Scrubfowl, Megapodius reinwardt, frequents riparian forests bordered by tropical sa vanna woodland. Ngawupodius may have occupied a similar habitat in the Oligocene environment of Lake Pinpa. The extinction of Ngawupodius was neither through compe tition with pheasants nor predation by felids. The only phasian- ids known from the Australian fossil record or occurring in a natural state in Australia today are small quail of the genus Coturnix (up to 120 g). These are too small to compete actively with megapodes, even Ngawupodius. Coturnix is known from other late-Oligocene-aged central Australian sites, although not Lake Pinpa. Pheasants, partridges, or other galliforms of a size comparable to any of the megapodes are not known from Aus tralia. Likewise, the Felidae and Viverridae do not occur in Australia, and there is no evidence that marsupial carnivores have a particularly deleterious effect on megapode populations. Marsupial carnivores were well represented in Australia during the Oligocene and also have been found in the Ericmas Fauna (Dasyuridae). If it left no descendants, the eventual extinction of Ngawupo dius may be related to a changing environment. With the late Miocene drying of Australia, the wet forest vegetation that ap parently was its habitat was lost from the center of the conti nent. Similar habitats are now restricted to parts of eastern and northern Australia. Other central Australian birds that were lost after these climatic alterations were aquatic forms (e.g., flamin gos, Phoenicopteridae). It is not yet possible to verify the loss of other terrestrial birds because these are not well represented, and existing material has yet to be studied in any depth. The morphology of Ngawupodius holds no clues to its rela tionships within the family; it shares osteological characters with most genera of megapodes. Until a complete phylogenetic analysis of the Megapodiidae is performed, the polarities of these characters will not be known. Likewise, whether Ngawupodius was ancestral to any of the living forms, and if so, how directly, cannot be determined. Similarities in propor tions and its occurrence in central Australia raise the possibility that Ngawupodius may have been a direct ancestor of Leipoa. Conversely, Ngawupodius may have represented a distinct lin eage within the family. There is general agreement that, regard less of their center of origin, megapodes were isolated in Aus- tralo-Papua for an extended period, which we know now to extend at least to the late Oligocene, Ngawupodius being the oldest known member of the Megapodiidae. Literature Cited Balouet, Jean Christophe, and Storrs L. Olson 1989. Fossil Birds from Late Quaternary Deposits in New Caledonia. Smithsonian Contributions to Zoology, 469:38 pages, 16 figures. Baumel, Julian J., and Lawrence M. Witmer 1993. Osteologia. In J.J. Baumel, A.S. King, J.E. Breazile, H.E. Evans, and C. Vanden Berge, editors, Handbook of Avian Anatomy: Nom ina Anatomica Avium. Publications of the Nuttall Ornithological Club, 23:45-132, figures 4.1?4.18. Boles, Walter E., and Brian Mackness 1994. Birds from the Bluff Downs Local Fauna, Allingham Formation. Records of the South Australian Museum, 27(2): 139-149,2 figures. Brom, T.G., and R.W.R.J. Dekker 1992. Current Studies on Megapode Phylogeny. In R.W.R.J. Dekker and D.N. Jones, editors, Proceedings of the First International Mega pode Symposium, Christchurch, New Zealand, December 1990. Zo- ologische Verhandelingen, 278:7-17, figures 1-5. Callen, Roger A., and Richard H. Tedford 1976. New Late Cainozoic Rock Units and Depositional Environments, Lake Frome Area, South Australia. Transactions of the Royal Soci ety of South Australia, 100(3): 125-168, 24 figures. NUMBER 89 205 Clark, George A., Jr. 1964a. Ontogeny and Evolution in the Megapodes (Aves: Galliformes). Postilla, 78: 37 pages, 11 figures. 1964b. Life Histories and the Evolution of Megapodes. The Living Bird, 3:149-167, 15 figures. Cracraft, Joel 1973. Continental Drift, Paleoclimatology, and the Evolution and Bioge ography of Birds. Journal of Zoology, London, 169(4):455-545, 20 figures, del Hoyo, Josep, Andrew Elliot, and Jordi Sargatal, editors 1994. Handbook of the Birds of the World, Volume 2: New World Vultures to Guineafowl. 638 pages, 60 plates. Barcelona: Lynx Edicions. Dekker, Ren? W.R.J. 1989. Predation and the Western Limits of Megapode Distribution (Mega podiidae; Aves). Journal of Biogeography, 16(4):317-321, 2 fig ures. Dekker, R.W.R.J., and T.G. Brom 1992. Megapode Phylogeny and the Interpretation of Incubation Strate gies. In R.W.R.J. Dekker and D.N. Jones, editors, Proceedings of the First International Megapode Symposium, Christchurch, New Zealand, December 1990. Zoologische Verhandelingen, 278:19-31, figures 6-11. De Vis, C.W. 1888. Australian Ancestry of the Crowned Pigeon of New Guinea. Pro ceedings of the Royal Society of Queensland, 5(4): 127-131. 1889. Additions to the List of Fossil Birds. Proceedings of the Royal Soci ety of Queensland, 6(l):55-58. Gaillard, C. 1908. Les oiseaux des Phosphorites du Quercy. Annates de I 'Universite de Lyon, new series, 1(23): 1-100. Jones, Darryl N., Rene W.R.J. Dekker, and Cees S. Roselaar 1995. The Megapodes. 262 pages, 8 plates. Oxford: Oxford University Press. Longman, H.A. 1945. Fossil Vertebrates from Gore Quarries. Memoirs of the Queensland Museum, 12(3): 164. Lydekker, Richard 1891. Catalogue of the Fossil Birds in the British Museum (Natural His tory). 368 pages, 75 figures. London: British Museum (Natural His tory). Marchant, Stephen, and Peter J. Higgins, editors 1993. Handbook of Australian, New Zealand and Antarctic Birds, Volume 2: Raptors to Lapwings. 984 pages, 68 plates. Melbourne: Oxford University Press. Mourer-Chauvire, Cecile 1982. Les oiseaux fossiles des Phosphorites du Quercy (Eocene superieur a Oligocene superieur): Implications paleobiogeographiques. In E. Buffetaut, P. Janvier, J.C. Rage, and P. Tassy, Phylogenie et Paleo- biogeographie: Livre jubilaire en l'honneur de Robert Hoffstetter. Geobios, special memoir, 6:413-426, 3 figures. 1992. The Galliformes (Aves) from the Phosphorites du Quercy (France): Systematics and Biostratigraphy. In K.E. Campbell, Jr., editor, Pa pers in Avian Paleontology Honoring Pierce Brodkorb. Science Se ries, Natural History Museum of Los Angeles County, 36:67-95, 14 figures. Olson, Storrs L. 1980. The Significance of the Distribution of the Megapodiidae. Emu, 80(l):21-24, 1 figure. 1985. The Fossil Record of Birds. In D.S. Farner, J.R. King, and K.C. Parkes, editors, Avian Biology, 8:79-238, 11 figures. New York: Academic Press. Peter, J.M., and M.B. Peter 1993. Appendix III: Aboriginal Names. In Stephen Marchant and Peter J. Higgins, editors, Handbook of Australian, New Zealand and Ant arctic Birds, Volume 2: Raptors to Lapwings, pages 971-973. Melbourne: Oxford University Press. Peters, James Lee 1934. Family Megapodiidae. In Check-list of Birds of the World, 2:3-9. Cambridge: Harvard University Press. Poplin, Francois 1980. Sylviomis neocaledoniae n.g. n.sp.; ratite eteint de la Nouvelle Cale- donie. Comptes Rendus de I'Academie des Sciences de Paris, series D, 290:489-492, 4 figures. Poplin, Francois, and Cecile Mourer-Chauvire 1985. Sylviomis neocaledoniae (Aves, Galliformes, Megapodiidae), oiseaux g6ant Eteint de l'ile de Pins (Nouvelle-Caledonie). Geo bios,\%(\):13-91, 8 figures, 4 plates. Reed, A.W. 1977. Aboriginal Words and Place Names. 286 pages. Adelaide: Rigby. Rich, Thomas H., Michael Archer, Suzanne J. Hand, Henk Godthelp, Jeannette Muirhead, Neville S. Pledge, Timothy F. Flannery, Michael O. Woodburne, Judd A. Case, Richard H. Tedford, William D. Turnbull, Ernest L. Lunde lius, Jr., Leallyn S.V. Rich, Michael J. Whitelaw, Anne Kemp, and Patricia V. Rich 1991. Australian Mesozoic and Tertiary Ten-estrial Mammal Localities. In P.V. Vickers-Rich, J.M. Monaghan, R.F. Baird, and T.H. Rich, edi tors, Vertebrate Palaeontology of Australasia, pages 1005-1070. Melbourne: Thomas Nelson. Roselaar, CS. 1994. Systematic Notes on Megapodiidae (Aves, Galliformes), Including the Description of Five New Subspecies. Bulletin, Zoologisch Mu seum, Universiteit van Amsterdam, 14(2):9-36, 8 figures. Sibley, Charles G., and Jon E. Ahlquist 1990. Phylogeny and Classification of Birds: A Study in Molecular Evolu tion. 976 pages, 385 figures. New Haven: Yale University Press. Sibley, Charles G., and Burt L. Monroe, Jr. 1990. Distribution and Taxonomy of Birds of the World. 1111 pages. New Haven: Yale University Press. Steadman, David W. 1980. A Review of the Osteology and Paleontology of Turkeys (Aves: Meleagridinae). In K.E. Campbell, Jr., editor, Papers in Avian Pale ontology Honoring Hildegard Howard. Contributions in Science, Natural History Museum of Los Angeles County, 330:131-207, 14 figures. 1989. New Species and Records of Birds (Aves: Megapodiidae, Colum bidae) from an Archaeological Site on Lifuka, Tonga. Proceedings of the Biological Society of Washington, 102(3):537?552, 7 figures. 1993a. Biogeography of Tongan Birds Before and After Human Impact. Proceedings of the National Academy of Science of the United States of America, 90:818-822, 1 figure. 1993b. Bird Bones from the To'aga Site: Prehistoric Loss of Seabirds and Megapodes. In P.V. Kirch and T.L. Hunt, editors, The To'aga Site: Three Millennia of Polynesian Occupation in the Manu'a Islands, American Samoa. Contributions of the University of California Archaeological Research Facility, Berkeley, 51:217-228, 2 fig ures. 1995. Prehistoric Extinctions of Pacific Island Birds: Biodiversity Meets Zooarchaeology. Science, 267:1123-1131,4 figures. Tedford, Richard H., Michael Archer, Allan Bartholomai, Michael Plane, Nev ille S. Pledge, Thomas H. Rich, Patricia Rich, and Rod T. Wells 1977. The Discovery of Miocene Vertebrates, Lake Frome Area, South Australia. BMR Journal of Australian Geology and Geophysics, 2:53-57, 3 figures, van Tets, G.F. 1974. A Revision of the Fossil Megapodiidae (Aves), Including a Descrip tion of a New Species of Progura De Vis. Transactions of the Royal Society of South Australia, 98(4):213-224, 5 figures. 1984. A Checklist of Extinct Fossil Australasian Birds. In M. Archer and G. Clayton, editors, Vertebrate Zoogeography and Evolution in 206 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Australasia, pages 469-475. Carlisle: Hesperian Press, van Tets, G.E, and Pat V. Rich 1990. An Evaluation of de Vis' Fossil Birds. Memoirs of the Queensland Museum, 28:165-168, 8 figures. Woodburne, Michael O., Bruce J. Macfadden, Judd A. Case, Mark S. Springer, Neville S. Pledge, Jeanne D. Power, Janice M. Woodburne, and Kathleen B. Springer 1993. Land Mammal Biostratigraphy and Magnetostratigraphy of the Eta dunna Formation (Late Oligocene) of South Australia. Journal of Vertebrate Paleontology, 13:483-515, 15 figures. Woodburne, Michael O., and Richard H. Tedford 1975. The First Tertiary Monotreme from Australia. American Museum Novitates, 2588:19 pages, 4 figures. Woodburne, M.O., R.H. Tedford, M. Archer, W.D. Tumbull, M.D. Plane, and E.L. Lundelius 1985. Biochronology of the Continental Mammal Record of Australia and New Guinea. Special Publications, Department of Mines and En ergy, South Australia, 5:347-363, 3 figures. A New Genus and Species of the Family Jungornithidae (Apodiformes) from the Late Eocene of the Northern Caucasus, with Comments on the Ancestry of Hummingbirds Alexandr A. Karhu ABSTRACT Argornis caucasicus, a new genus and species of the family Jungornithidae (Apodiformes), is based on an incomplete, articu lated skeleton of the shoulder girdle and wing from the late Eocene of the northern Caucasus. The holotype includes the manus, which was previously unknown in the Jungornithidae. In comparison with early Oligocene Jungornis, the new form is less advanced evolutionarily, and, in particular, it lacks certain characters shared by Jungornis and the Trochilidae. An emended diagnosis of the family Jungornithidae is given. Taking into account that both Jun gornis and Argornis possess an apodid-like deltopectoral crest, revealing their highly developed ability for gliding flight, the appearance of trochilid-like features in Jungornis demonstrates a real possibility that hovering specializations developed from glid ing adaptations. This conclusion conforms with the results of a comparative analysis of the transformation of forelimb muscles in three modem apodiform families. Introduction Jungornis tesselatus Karhu, 1988, a bizarre Paleogene apod iform referred to its own family, was described from an incom plete, articulated skeleton of the shoulder girdle and forelimb from the early Oligocene of the northern Caucasus. The second genus and species of the family Jungornithidae described, Palescyvus escampensis Karhu, 1988, was based on a single coracoid from the late Eocene of the Phosphorites du Quercy, France, that Mourer-Chauvire (1978) had previously assigned to Cypselavus gallicus Gaillard. Alexandr A. Karhu, Paleontological Institute, Russian Academy of Sciences, Profsoyuznaya Street 123, Moscow, 117868, Russia. The Jungornithidae possess features in common with such evolutionarily advanced families as the Apodidae and Trochil idae; the Jungornithidae also demonstrate a clear resemblance to the comparatively generalized apodiform family Hemiproc- nidae, and even to the Caprimulgidae. An unusual combination of characters in the Jungornithidae, some of which are shared separately either with the Apodidae or with the Trochilidae, has been considered evidence in favor of the common origin of these three families (Karhu, 1988, 1992a, 1992b). Further in vestigation of the apodiform flight apparatus has revealed, in particular, that an important part of the morphofunctional spe cializations in hummingbirds represents subsequent stages of development from the apodid-like adaptations (Karhu, 1992a). This contradicts the principal conclusion of Cohn (1968) that the similarity between the true swifts and hummingbirds is the result of convergence. In 1993, a new genus and species of jungornithid, described herein, was found in a late Eocene locality, Gorny Luch, north ern Caucasus, that has yielded an abundant marine ichthyofau na (Bannikov, 1993). This discovery provides important data concerning the morphological specialization of the flight appa ratus in the Jungornithidae. Taking into account the essential similarities between the Jungornithidae and Trochilidae, an older example of the former family may shed light on the early evolution of hummingbirds. METHODS.?Comparative study of the forelimb muscles in modern Apodiformes is very important for analysis of evolu tionary trends. The forelimb muscles of the following species were studied (number of specimens is in parentheses): Hemi- procnidae: Hemiprocne mystacea (1), Hemiprocne comata (1); Apodidae: Collocalia lowi (1), Hirundapus caudacutus (1), Chaetura pelagica (1), Chaetura brachyura (1), Apus apus (4), Apus pacificus (1); Trochilidae: Chlorostilbon ricordii (1), Chlorostilbon sp. (1), Heliomaster longirostris (1), and Papho- 207 208 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY sia helenae (1). Muscle dissections were stained by the method of Bock and Shear (1972), and magnifications from x8 to x50, mainly x 16 and x25, were used. The analysis of forelimb mus cles also utilizes data from Cohn (1968) and from Zusi and Bentz (1982, 1984). This paper presents a brief summary of the unpublished results of the investigation of the forelimb muscles in apodiforms (Karhu, 1992a) (see "Forelimb Muscles: Ten dencies of Transformation," below). Anatomical terminology used in the descriptions generally follows Baumel et al. (1993). ACKNOWLEDGMENTS.?I am very grateful to A.F. Bannikov (Paleontological Institute, Moscow) for donation of the speci men described in the present paper. I wish to acknowledge C. Mourer-Chauvire (Universite Claude Bernard, Lyon) for the opportunity to study the coracoid specimens of Aegialornis gallicus Lydekker. For the loan of skeletal and spirit specimens of Apodiformes I am very indebted to the following: CT. Col lins (California State University, Long Beach), F.Y. Dzerzhin sky (Moscow State University, Moscow), V.M. Loskot (Zoo logical Institute, St. Petersburg), G.F. Mees (Rijksmuseum van Natuurlijke Historie, Leiden), S.L. Olson, R.L. Zusi, J.P. An gle, and J. Dean (National Museum of Natural History, Smith sonian Institution, Washington, D.C), and F. Vuilleumier (American Museum of Natural History, New York). The manuscript was substantially improved by the thoughtful com ments of R.L. Zusi and S.L. Olson. I am very thankful to R.L. Zusi and CT. Collins for their reviews. This research was sup ported by grants from the Russian Foundation for Basic Re searches (number 96-04-50822) and the Frank M. Chapman Memorial Fund of the American Museum of Natural History. Systematic Paleontology Order APODIFORMES Suborder APODI Family JUNGORNITHIDAE TYPE GENUS.?Jungornis Karhu, 1988. EMENDED DIAGNOSIS.?Apex carinae moderately devel oped. Facies articulares coracoidei widely spaced and separat ed from rostri sterni. Sulci carinae well pronounced. Proc. acro- coracoideus claviculae placed along dorsal margin of scapus, being considerably narrower than scapus in dorsoventral di mension; concave facies articularis acrocoracoidea claviculae oriented caudolaterally. Facies articularis humeralis scapulae directed cranioventrally. Proc. acrocoracoideus between facies articularis clavicularis and impressio lig. acrocoracohumeralis stretched mediolaterally. Proc. lateralis coracoidei well pro nounced, protruding noticeably laterad beyond level of angulus lateralis. Facies articularis sternalis coracoidei wide, with dis tinctly outlined angulus medialis projecting sternally approxi mately to same level as angulus lateralis. Caput humeri direct ed caudally. Crista deltopectoralis high and proximally placed. Proc. supracondylaris dorsalis distally placed. Proximoventral border of cotyla ventralis ulnae not pronounced. Phalanx proxi- malis digiti majoris bifenestrated. INCLUDED GENERA.?Jungornis Karhu, 1988; Palescyvus Karhu, 1988; Argornis, new genus. STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION.?Upper Eocene and lower Oligocene, northern Caucasus, Russia; upper Eocene, Phosphorites du Quercy, France. REMARKS.?In comparison with the former diagnosis of the family Jungornithidae (Karhu, 1988), the emendation omits the presence of the distal enlargement of the middle of the caput humeri and the proximity of the tuberculum supracondylare ventrale and tuberculum M. pronator superficialis. It adds points concerning the development of proc. lateralis coracoi dei, orientation of the caput humeri, position of proc. supra condylaris dorsalis, development of the proximoventral border of the cotyla ventralis ulnae, and fenestration of the proximal phalanx of the major digit. Argornis, new genus TYPE SPECIES.?Argornis caucasicus, new species. DISTRIBUTION.?Upper Eocene; northern Caucasus, Russia. ETYMOLOGY.?From the Greek argos, swift, and ornis, bird; the gender is masculine. DIAGNOSIS.?Facies articularis acrocoracoidei claviculae lengthened mediolaterally. Acromion scapulae with cranial margin beveled laterally and crista lig. acrocoraco-acromiale well developed. Dorsal side of medial part of proc. acrococora- coideus forms high, caudally projecting crest, with mediolater ally narrow base. Facies articularis sternalis coracoidei saddle- shaped, with only medial part of crista ventralis protruding ventrad. Angulus lateralis of sternal facet projecting a little more distally than angulus medialis. Ventral part of caput hu meri oriented perpendicularly to long axis of bone, whereas dorsal part oriented obliquely to it, being placed more distally relative to ventral part; distal border of caput humeri clearly outlined. Tuberculum M. tensor propatagialis pars brevis well pronounced and placed just distal to proc. supracondylaris dor salis humeri. Tuberculum M. pronator superficialis clearly de tached from tuberculum supracondylare ventrale humeri. Tu berculum supracondylare ventrale adjoins condylus ventralis humeri. Epicondylus ventralis markedly prominent ventrad. Proc. flexorius projects slightly distally beyond condylus ven tralis. Tuberculum lig. collateralis ventralis ulnae relatively small and weakly protruding ventrad. Tuberculum bicipitale brachii radii with impression of M. biceps brachii. COMPARISON.?Clavicula: In Argornis the facies articu laris acrocoracoidei is more elongated mediolaterally, narrower dorsoventrally, and directed more caudally than it is in Jungor nis, in which it is a rounded, caudolaterally orientated facet. The overall configuration of the facies articularis acrocoracoi dei is very unusual in both Jungornis and Argornis in compari- NUMBER 89 209 son with other known Apodiformes, and it appears to be closest to that in the Caprimulgidae. Scapula: The cranial margin of the acromion is strongly beveled laterally in Argornis, whereas in Jungornis the margin is thickened. The crista lig. acrocoraco-acromiale is well devel oped in Argornis, but it is not pronounced in Jungornis. In Ar gornis the facies articularis humeralis is relatively longer cran- iocaudally and wider dorsoventrally in comparison with Jungornis. Coracoid: The overall configuration is close to that in Jun gornis. In these genera the shaft is relatively slender, and the mediolateral width of the processus acrocoracoideus exceeds the distance between the angulus medialis and the angulus lat eralis of the sternal facet (Figure 3c,D), whereas in Palescyvus the shaft is stouter, and the acrocoracoid is narrower than the sternal facet in frontal aspect (Figure 3E). In Argornis the dor sal crest of the medial portion of the acrocoracoid protrudes caudally almost to the level of the middle of the dorsal aperture of the canalis triosseus, and the base of this crest does not reach laterally to the level of the medial border of the impressio lig. acrocoracohumerale. In Jungornis the crest projects much less caudad, and its base is relatively wider, extending laterad to the level of the medial edge of the impression mentioned above. In Palescyvus the medial portion of the acrocoracoid curves strongly caudally but lacks a pronounced dorsal crest. The base of the proc. procoracoideus is relatively wider in Argornis in comparison with either Jungornis or Palescyvus. The facies ar ticularis sternalis is saddle-shaped in Argornis but is concave in Jungornis. In Argornis only the medial part of the crista ventra lis of the sternal facet protrudes ventrad, whereas in Jungornis the entire crista ventralis forms a ventral convexity. The ratio of the greatest dorsoventral width of the sternal facet to the dis tance between the angulus medialis and angulus lateralis is smaller in Argornis than it is in Jungornis. The angulus media lis is rounded in Argornis and in Palescyvus, whereas in Jun gornis it is moderately sharpened. The angulus lateralis projects slightly distad beyond the level of the angulus medialis in Argornis; in Jungornis and Palescyvus the angulus medialis and angulus lateralis are placed approximately on the same lev el (Figure 3C-E). Humerus: In Argornis the humeral shaft is more slender, and both the proximal and distal ends are relatively narrower in comparison with Jungornis (Figure 2E-G,R). The smaller, ven tral part of the caput humeri is situated perpendicularly to the long axis of the bone in Argornis, unlike in Jungornis, in which the entire caput humeri is transversely placed (Figure 2H,R). There is no distal enlargement of the middle of the caput hu meri on the facies caudalis in Argornis as there is in Jungornis. In Argornis the clearly pronounced tuberculum of M. tensor propatagialis pars brevis adjoins distally the base of the proc. supracondylaris dorsalis. This process in Jungornis is adjoined distally by a high thin crest, and the insertion of M. tensor pro patagialis pars brevis is not marked. Tuberculum M. pronator superficialis is low and is clearly separated from the tubercu lum supracondylare ventrale in Argornis, whereas in Jungornis the former protrudes strongly proximad and fuses with the proximal part of the tuberculum supracondylare ventrale. In Argornis the tuberculum supracondylare ventrale adjoins the base of the condylus ventralis ventroproximally, but in Jungor nis it is placed more proximally and is well separated from the base of the condylus ventralis. In Argornis the condylus ventra lis projects distad beyond the condylus dorsalis, whereas the distal borders of both condyli are placed on the same level in Jungornis. In Argornis the epicondylus ventralis protrudes strongly ventrad, is widened proximodistally, and is fused dis tally with the proc. flexorius, unlike the much smaller and de tached epicondylus ventralis of Jungornis. The proc. flexorius projects distad less in Argornis than in Jungornis. Ulna: In Argornis the tuberculum lig. collaterale ventrale is smaller and protrudes less ventrad in comparison with Jun gornis. Radius: In contrast to Jungornis, there is an impression of M. biceps brachii on the ventrocranial surface of the tubercu lum bicipitale brachii in Argornis. REMARKS.?An extremely peculiar morphology of the ster- nocoracoidal articulation is among the most characteristic fea tures of the Apodiformes sensu Wetmore, 1960 (Lucas, 1893; Lowe, 1939; Cohn, 1968; Karhu, 1988, 1992a). In the Apodi formes, the sternum possesses a weakly saddle-shaped or con vex facies articularis coracoidei instead of the coracoidal sul cus of most birds. Consequently, the coracoid has a more or less dorsoventrally widened facies articularis sternalis that is slightly saddle-shaped, or concave, and placed on the whole perpendicularly to the long axis of the bone. There is a single exception: in Aegialomis the coracoid has the sternal facet ven trally widened near the angulus medialis, whereas its greater part is wedge-shaped in dorsoventral section, as is typical for most birds. The structure of the sternal facet of the coracoid is consider ably more generalized in Argornis in comparison with Jungor nis. In Jungornis the overall configuration of the facies articu laris sternalis is apodid-like, with the entire ventral margin convex ventrad but to a lesser degree than in true swifts. In Ar gornis the coracoid possesses a facies articularis sternalis that is widened by a ventral prominence in its medial part only, which is similar to the most generalized type within Apodi formes as demonstrated by Aegialomis. Argornis differs from Aegialomis, however, in having the dorsal edge of the sternal facet situated much more sternally (Figure 3A,C). The other relatively generalized character of the coracoid in Argornis is the placement of the angulus lateralis slightly be yond the level of the angulus medialis (Figure 3C). In Jungor nis and Palescyvus the angulus lateralis and angulus medialis are on the same level relative to the long axis of the bone (Fig ure 3D,E), which was pointed out as a distinctive familial char acter in the former diagnosis of the Jungornithidae (Karhu, 1988). In the Hemiprocnidae and Apodidae, as well as in the Eocene genus Aegialomis, the angulus lateralis protrudes con- 210 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY siderably more sternally than does the angulus medialis (Figure 3A,B,F). All three known genera of the Jungornithidae are clearly dis tinguished from other fossil and modern Apodiformes by the presence of a well-developed proc. lateralis in the sternal part of the coracoid (Figure 3). In caudal aspect, the overall configuration of the caput hu meri in Argornis is most similar to that in Hemiprocne: the smaller ventral part of the head is placed approximately per pendicular to the long axis of the bone, whereas its greater part is placed obliquely and more distally relative to the ventral part of the head. At the same time, Argornis differs from Hemiproc ne in having the caput humeri directed mainly caudally, in con trast to Hemiprocne in which it is directed apically. Both Ar gornis and Jungornis have a relatively distally placed proc. supracondylaris dorsalis, similar to that in Aegialomis. In com parison with other Apodiformes, Argornis possesses the most distally situated tuberculum supracondylare ventrale, revealing a tendency toward the proximal displacement found in all apodiform families. In Argornis and Jungornis the cotyla ventralis ulnae has a slightly pronounced ventroproximal edge. The Trochilidae pos sess a similar structure of the cotyla ventralis in this regard, whereas in the Hemiprocnidae and Apodidae its ventroproxi mal edge is well marked. Argornis caucasicus, new species FIGURES 1,2A-P, 3C HOLOTYPE.?Incomplete, partially crushed articulated skele ton including the vertebral column, shoulder girdle, and fore- limbs; Paleontological Institute of the Russian Academy of Sciences, PIN 4425-18. TYPE LOCALITY.?Gorny Luch, left bank of Pshekha River, northern Caucasus, Russia. HORIZON.?Kuma (Kumsky) horizon, upper Eocene (Banni kov, 1993). MEASUREMENTS (in mm).?Clavicle, minimum width 0.5, maximum width 0.9; scapula, dorsoventral width of cranial end 2.4; coracoid, length 10.0, diameter of midshaft 0.9 by 1.0, me- diolateral width of sternal end 2.6, dorsoventral width of sternal end 1.2, distance between tips of angulus medialis and angulus lateralis 1.8; humerus, length 10.4, proximal dorsoventral width 4.1, distal dorsoventral width 3.3; ulna, length 16.0; radi us, length 15.1, diameter of midshaft 0.5 by 0.7; carpometacar pus, length 11.6, craniocaudal width through extensor process 3.8, craniocaudal width of midshaft of major metacarpal 1.1; proximal phalanx of major digit, length through articular sur faces 6.4, maximum length 7.2, craniocaudal width at middle 2.4; distal phalanx of major digit, length 6.3; phalanx of minor digit, length 3.0. Judging from the length of the coracoid, A. caucasicus may have been approximately the same overall size as Palescyvus escampensis, which has a coracoid 10.1 mm long (Harrison, 1984, table 2), although in A caucasicus the coracoid is some what more slender, suggesting smaller body size. Argornis caucasicus noticeably exceeds Jungornis tesselatus in all cor responding measurements. ETYMOLOGY.?After Caucasus, the geographic area of the type locality. DESCRIPTION.?Remains of the vertebral column, sternum, and ribs are too fragmentary and badly damaged for description of their features. Clavicula: The scapus claviculae is flattened mediolateral ly and smoothly widened toward the extremitas omalis. The transition between the clavicular shaft and the proc. acromialis is not pronounced. The proc. acrocoracoideus protrudes strong ly laterad, its base approximately one-half the dorsoventral width of the clavicular shaft. The facies articularis acrocoracoi deus is flattened dorsoventrally and is concave mediolaterally. Scapula: The acromion protrudes a little beyond the level of the cranial border of the tuberculum coracoideum. The dor sal margin of the acromial tip is curved laterad, with the crista lig. acrocoraco-acromiale short craniocaudally. The tubercu lum coracoideum passes gradually into the proc. glenoidalis. The dorsocaudal part of the proc. glenoidalis protrudes strongly laterad. The facies articularis humeralis is widened ventro- caudad, being directed cranioventrally and turned slightly later ally. Coracoid: The facies articularis clavicularis is convex. The well-marked cotyla scapularis is rounded and moderately con cave. The proc. procoracoideus is long, its base stretches caudad almost to the level of the cranial border of the impressio M. sternocoracoidei. The foramen supracoracoidei, situated in the middle of the base of the procoracoid, opens ventrally into a groove that extends along the base. The sulcus supracoracoi- deus is well developed craniad of the level of the foramen su pracoracoidei. The impression of M. supracoracoidei is deep and sharply outlined. The proc. lateralis is obtuse-angled in frontal aspect. The facies articularis sternalis is subdivided asymmetrically into a smaller medial and a larger lateral part by a saddle-like ridge that extends lateroventrad from the angu lus medialis. The larger lateral part is concave in both me- diolateral and dorsoventral dimensions. Humerus: The tuberculum dorsale is displaced distad from the dorsal part of the caput humeri. The crista deltopectoralis is high and tapering, with a concave proximal edge. The tip of the crista deltopectoralis is approximately on the level of the mid dle of the impression of the tendon of M. supracoracoideus. The proc. supracondylaris dorsalis occurs about one-quarter of the length of the humerus from its distal end. The tuberculum M. pronator superficialis is placed approximately on the level of the tuberculum M. tensor propatagialis pars brevis. The sul cus intercondylaris is narrow and shallow. The proc. flexorius is blunt and widened dorsoventrally. The cranial surfaces of the tuberculum supracondylare ventrale and epicondylus ventralis are fused continuously, as are the cranial surfaces of the epi condylus ventralis and proc. flexorius. The fossa of M. brachia- NUMBER 89 211 FIGURE 1.?Argornis caucasicus, new genus, new species, holotype (PIN 4425-18), partial skeleton: A,B, two complementary slabs (x3; specimen coated with ammonium chloride). 212 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 2.?Elements of shoulder girdle and forelimb (A-P) ofArgornis cauca sicus, new genus, new species (holotype, PIN 4425-18), and humerus (R) of Jungornis tesselatus. A-C, right coracoid in ventral (A), dorsal (B), and sternal (C) views; D, right clavicle with sternal part of left bone; E-G, right humerus in cranial (E), caudal (F), and dorsal (G) views; 1,H, proximal part of left humerus in caudal (H) and proximal (i) views; J,K, right ulna in cranial (J) and caudal (K) views; L,M, right radius in dorsal (L) and ventral (M) views; N, right car pometacarpus and phalanx of alular digit, dorsal view; O, right proximal pha lanx of major digit and phalanx of minor digit, dorsal view; P, right distal pha lanx of major digit, dorsal view; R, left humerus of Jungornis tesselatus (holotype, PIN 1413-208), caudal view (x5; specimens coated with ammonium chloride). NUMBER 89 213 AM AL AM AL AM AL AM AM AL AM FIGURE 3.?Comparison of coracoid of Apodiformes: A, Aegialomis gallicus; B, Hemiprocne mysta- cea; C, Argornis caucasicus, new genus, new species; D, Jungornis tesselatus; E, Palescyvus escampen- sis; F, Apus apus; G, Archilochus colubris. AM=angulus medialis, AL=angulus lateralis. Arrows indi cate proc. lateralis. Left side, dorsal view, standardized for comparison. (Scale-1 mm.) lis is not pronounced. The sulcus of M. humerotricipitis is wide and shallow. Ulna: The shaft is slender and straight. Both ends are rela tively narrow. The olecranon is tapered, with a narrow base. The cotyla ventralis is shallow. The ridge separating the cotyla dorsalis and cotyla ventralis is low and smooth. The tubercu lum bicipitale is located close to the distal border of the cotyla ventralis. There is a deep fossa for M. ulnometacarpalis ventra lis on the caudoventral side of the proximal end. Radius: The shaft is thin and slightly bowed craniad. Both ends are relatively narrow. The tuberculum bicipitale adjoins the ventral border of the cotyla humeralis. Carpi ulnare: The proximal margin of the corpus is notice ably beveled, so that the proximal border of the facies articu laris ulnaris lies much more distally than the base of the proc. muscularis. Carpometacarpus: The metacarpale majus is slender, slightly bowed craniad, and beveled ventrocaudally. The protu- berantia metacarpalis projects dorsally, is proximodistally elongated, and is placed somewhat distad to the middle of the craniodorsal surface of the major metacarpal. A well-devel oped sulcus tendinosus begins just proximocaudally of the pro- tuberantia metacarpalis and passes distad and somewhat caudad, where it widens into a clearly outlined and relatively deep depression on the dorsal side of the distal end of the major metacarpal. There is a pronounced tuberculum of M. extensor metacarpi ulnaris on the caudal margin of the major metacarpal just opposite the distal border of the proximal symphysis. The cranial margin of the labrum dorsale of the trochlea carpalis bears the impression of M. ulnometacarpalis ventralis. The caudal margin of the facies articularis digitalis major is deeply concave. The distal part of metacarpale minus is absent in the holotype; however, the facies articulares metacarpales of the proximal phalanx of the major digit and of the minor digit are on the same level in the holotype, which suggests that corre sponding facies articulares digitales major and minor also should be level relative to the long axis of the carpometacarpus. Ossa digorum manus: The phalanx digiti alulae is dam aged too badly to recognize any structural peculiarities. Two large oval fenestrae of the phalanx proximalis digiti ma- joris are separated by a dorsally pronounced, obliquely oriented pila transversa. There is a well-developed proc. distalis on the caudal margin. The caudal edge of the facies articularis metac arpalis forms a deep middle concavity. The phalanx distalis digiti majoris has a caudally enlarged proximal articulation and a caudally widened tip. The phalanx digiti minoris is narrow and tapering. It curves along the caudal margin of the proximal phalanx of the major digit and extends to the level of the distal margin of the proxi mal fenestra. The facies articularis metacarpalis of phalanx dig iti minoris is placed relative to the long axis of the manus on the same level as the metacarpal facet of the proximal phalanx of the major digit. Forelimb Muscles: Tendencies of Transformation Among the living Apodiformes, the most generalized config uration of the forelimb muscles is found in the Hemiprocnidae. The completeness of their set of forelimb muscles reveals a rel atively low level of specialization. Many of their muscles dem onstrate a comparatively simple inner structure. In particular, unlike the true swifts and hummingbirds, the crested swifts 214 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY possess well-developed Mm. extensor longus alulae and ul nometacarpalis dorsalis. Both of these muscles usually sustain a reduction when the automatic conjunction of movements in the elbow and carpal joints becomes more efficient (Stegmann, 1970). The Hemiprocnidae also are characterized by relatively simple inner differentiation of such important flight muscles as M. pectoralis, M. flexor carpi ulnaris, and M. extensor metacar- pi radialis. In the crested swifts, comparatively weak develop ment of M. extensor digitorum communis, M. flexor digitorum profundus, and M. ulnometacarpalis ventralis is obviously as sociated with their limited role of resisting aerodynamic forces, whereas in the true swifts and hummingbirds these muscles also participate in active rotation of the manus and its major digit. Certain flight muscles are developed in constant proportion to body size in all representatives of the families compared: M. rhomboideus superficialis, the group of Mm. serrati (with ex ception of M. serratus superficialis pars metapatagialis, it being absent in the hummingbirds), M. coracobrachialis caudalis, M. tensor propatagialis pars brevis, M. scapulotriceps, M. brachia- lis, M. expansor secundariorum, M. ectepicondylo-ulnaris, M. abductor alulae, and M. flexor digiti minoris. All three families are characterized by the weakness of both M. deltoideus minor and M. scapulohumeralis cranialis. Relative development of the following flight muscles in creases from the Hemiprocnidae to the Apodidae to the Tro chilidae: M. subcoracoideus caput ventrale, M. pectoralis, M. supracoracoideus, M. humerotriceps, M. flexor digitorum pro fundus caput humerale, M. flexor carpi ulnaris, M. extensor metacarpi radialis, M. extensor digitorum communis, M. exten sor longus digiti majoris, M. supinator, M. ulnometacarpalis ventralis, and M. abductor digiti majoris. Thus, in addition to an obvious and quite understandable hypertrophy of M. pecto ralis and M. supracoracoideus, the reinforcement of the flight muscles in the true swifts and hummingbirds involves those that supinate the humerus and forearm, extend the elbow, ex tend and flex the wrist, rotate the manus, supinate the major digit of the manus, and flex the major digit and pronate its dis tal phalanx. In the same sequence, the following muscles become rela tively less developed: M. scapulohumeralis caudalis, M. rhom boideus profundus, M. deltoideus major, M. latissimus dorsi pars cranialis, M. biceps brachii, M. extensor longus alulae, M. ulnometacarpalis dorsalis, and M. flexor alulae. In the true swifts and hummingbirds, a relative weakness of the muscles that elevate and retract the humerus without causing rotation (M. latissimus dorsi pars caudalis, M. scapulohumeralis cauda lis, M. deltoideus major) obviously results from a hypertrophy of both M. pectoralis and M. supracoracoideus. These two muscles provide mainly rotational mobility of the humerus rel ative to its long axis in the true swifts and hummingbirds, which correlates with the caudal orientation of the caput hu meri and the shortening of the humeral shaft in these families (Karhu, 1992b). The retracting action of M. pectoralis grows as its sternal attachment widens caudally, increasing the amount of muscular fibers oriented in a craniodorsolateral direction. In contrast to the large number of muscles with similar ten dencies of specialization in both the Apodidae and the Trochil idae, there are few muscles in which specific reinforcement or reduction is unique either to the Apodidae or to the Trochil idae. The Apodidae exceed the Trochilidae in relative develop ment of M. coracobrachilais cranialis, M. subscapularis caput laterale, M. flexor digitorum superficialis, M. extensor metac arpi ulnaris, and M. interosseus ventralis. In the Apodidae, en largement of the muscles listed above provides more efficient maintenance of the spread wing and prevents passive extension of the wrist and passive dorsal flexure of the major digit. These peculiar transformations of the flight muscles in the Apodidae correspond to their greater ability in gliding and fast, forward- flapping flight in comparison with the Trochilidae. In comparison with the Apodidae, the Trochilidae have much better developed Mm. pronator superficialis and pronator profundus but less developed M. subscapularis, M. latissimus dorsi pars caudalis, M. tensor propatagialis pars longa, and M. interosseus dorsalis. In addition, the hummingbirds lack both propatagial parts of M. pectoralis, and M. flexor digitorum su perficialis remains only as a short, stout tendon, attaching on the proximal part of the lig. humeroulnare. They have only two of the four alular muscles, namely, M. abductor alulae and M. adductor alulae, the latter being greatly reduced. Reinforcement of Mm. pronator superficialis and pronator profundus in hummingbirds indicates an extensive rotational mobility of the forearm relative to the humerus. This conclu sion conforms with the structure of the elbow joint in the hum mingbirds, which allows significant rotational movements, un like the more restricted mobility in the true swifts. A conspicuous example of divergence between the Apodidae and Trochilidae is provided by M. biceps brachii. In living Apodiformes, only the crested swifts have the M. biceps brachii ending on both the proximal end of the ulna and the proximal end of the radius. In hummingbirds there is a single insertion on the ulna, whereas in the true swifts the insertion is on the radius. Taking into account that the double insertion of M. biceps brachii is typical for most birds, it is obviously more generalized, and a single insertion, either on the ulna or on the radius, represents a morphological specialization. Discussion The following features show the general level of specializa tion to be lower in Argornis than in Jungornis: coracoid with facies articularis sternalis saddle-shaped; sternal facet of cora coid relatively narrow dorsoventrally with only the medial part ventrally protruded; sternal facet of coracoid with the angulus lateralis projecting beyond the level of the angulus medialis; both proximal and distal ends of humerus relatively narrow; humeral shaft more slender; only the smaller ventral part of the NUMBER 89 215 caput humeri perpendicular to the long axis of the bone; the middle of the caput humeri without a distal protrusion on the caudal side; distal part of humerus without a dorsal crest; hu merus with tuberculum M. pronator superficialis detached from tuberculum supracondylare ventrale; distal end of humerus with tuberculum supracondylare ventrale adjacent to the condylus ventralis; proc. flexorius of humerus projecting weak ly distad; ulna with relatively small tuberculum lig. collateralis ventralis; and double insertion of M. biceps brachii on both the radius and ulna. The Jungornithidae resemble the Apodidae in having a high, robust, tapering, and proximally placed deltopectoral crest. In the Apodidae the structure and position of the deltopectoral crest correlates with reinforcement of M. coracobrachialis cra nialis and M. pectoralis, pars cranialis. All these features are among the characters that provide a highly developed ability for gliding flight in the true swifts (Karhu, 1992a). The similar ity of structure and placement of the deltopectoral crest in the true swifts and jungornithids suggests that the latter could be well adapted for gliding flight, too. At the same time, Jungornis and the Trochilidae share some essential characters that distinguish them from the Apodidae. The structure of the humeral head in Jungornis clearly demon strates a trochilid-like specialization: presence of the distal en largement on the caudal surface. In the Trochilidae this modifi cation of the humeral head is associated with high specialization of the shoulder joint, which allows extreme supi nation of the adducted humerus during hovering flight (Cohn, 1968; Karhu, 1992a). In the Hemiprocnidae, Apodidae, and all known fossil genera, with the exception of Jungornis, the hu meral head lacks any distal enlargement on the caudal surface. Although Argornis possesses a double insertion of M. biceps brachii both on the ulna and on the radius, there is no sign of the radial insertion in Jungornis. This fact implies the presence of a single insertion on the ulna, although it cannot be deter mined directly because of poor preservation in the holotype of Jungornis. If so, it would represent another trochilid-like spe cialization within the Jungornithidae. Taking into account that the trochilid-like characters under discussion are absent in the relatively more generalized jungor- nithid genus Argornis, their occurrence in Jungornis should be considered a result of intrafamilial evolution parallel to that in the Trochilidae. Because two jungornithid genera, Argornis and Jungornis, show obvious similarities to the Apodidae, the origin of trochilid-like features in the Jungornithidae demon strates the feasibility of developing trochilid-like specializa tions from apodid-like adaptations. It suggests that the Trochil idae could have arisen from an apodid-like ancestor as well. This inference is supported by the analysis of evolutionary transformation of the forelimb muscles in the Apodiformes, which shows the widespread progression of certain specializa tions in the Trochilidae relative to the Apodidae. Agreement in the conclusions based on paleontological and myological data clearly contradicts the opinion of Cohn (1968) that similarities between the Apodidae and Trochilidae are convergent. Both Argornis and Jungornis are similar to the Trochilidae and differ from the Apodidae in having the cotyla ventralis ul nae with a weakly pronounced ventroproximal edge. Owing to this peculiarity, the condylus ventralis of the humerus can slide ventroproximally relative to the ulna during the supination of the forearm. In the Apodidae the ventroproximal edge of the cotyla ventralis is prominent and strongly restricts the possible rotational movements of the elbow joint in the spread wing (Karhu, 1992a). The position of the proc. supracondylaris dorsalis of the hu merus is among the number of especially important characters in Apodiformes. The tendency for proximal displacement of this process is obviously conditioned by the proximal enlarge ment of the places of origin of M. extensor digitorum commu nis and the ventral head of M. extensor metacarpi radialis on the craniodorsal side of the distal part of the humerus. Very in teresting indirect evidence of such a correlation is provided by Jungornis. Its humerus resembles the true swifts and humming birds in overall configuration, but it has the proc. supracondy laris dorsalis placed approximately on the same level as in Ae gialomis, the most generalized genus of Apodiformes known (Karhu, 1992b). The relatively distal position of the proc. su pracondylaris dorsalis in Jungornis may be explained by the presence of the high crest distally adjacent to the process. This crest allows enlargement of the place of origin of both exten sors without proximal displacement of the supracondylar pro cess. Hummingbirds, which have the best developed M. exten sor digitorum communis and ventral head of M. extensor metacarpi radialis, demonstrate both conditions: proximal dis placement of the proc. supracondylaris dorsalis, and muscle or igin from the crest. In Argornis the supracondylar process is lo cated distally, and the dorsal crest is absent, suggesting relatively weak development of both these muscles. The well-developed distal process of the caudal margin of the proximal phalanx of the major digit indicates the presence of long first primaries in Argornis (Stegmann, 1965). Evidence for long first primaries in other Eocene apodiforms (e.g., Lyde kker, 1891; Peters, 1985) suggests that the elongation of the distal part of the wing occurred in the early stages of apodiform evolution. In particular, this elongation might precede the di vergence of humeral structure that distinguishes the hemiproc- nid and apodid-trochilid directions of specialization of the flight apparatus within Apodiformes (Karhu, 1992a, 1992b). 216 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Literature Cited Bannikov, Alexandre F. 1993. The Succession of the Tethys Fish Assemblages Exemplified by the Eocene Localities of the Southern Part of the former USSR. Kaupia: Darmstddter Beitrage zur Naturgeschichte, 2:241?246, 3 figures. Baumel, J.J., A.S. King, J.E. Breazile, H.E. Evans, and J.C. Vanden Berge, editors 1993. Handbook of Avian Anatomy: Nomina Anatomica Avium. Publica tions of the Nuttall Ornithological Club, 23: second edition, xxiv+779 pages. Cambridge, Massachusetts: Nutall Ornithological Club. Bock, Walter J., and Ch.R. Shear 1972. A Staining Method for Gross Dissection of Vertebrate Muscles. Anatomischer Anzeiger, 130:222-227. Cohn, Jean M.W. 1968. The Convergent Flight Mechanism of Swifts (Apodi) and Hum ming-birds (Trochili) (Aves). 201 pages, 21 figures, 13 tables. Doc toral dissertation, University of Michigan, Ann Arbor, Michigan. Harrison, C.J.O. 1984. A Revision of the Fossil Swifts (Vertebrata, Aves, Suborder Apodi), with Description of Three New Genera and Two New Species. Med- edelingen van de Werkgroep voor Tertiaire en Kwartaire Geologie, 21(4):157-177, 8 figures, 2 tables. Karhu, Alexandr A. 1988. [A New Apodiform Family from the Paleogene of Europe.] Paleon tological Journal, 3:78-88, 6 figures, 2 plates. [In Russian.] 1992a. [Phylogenetic Relationships within the Order Apodiformes.] 285 pages, 33 figures, 2 plates. Candidate dissertation, Paleontological Institute of the Russian Academy of Sciences, Moscow. [In Rus sian.] 1992b. Morphological Divergence within the Order Apodiformes as Re vealed by the Structure of the Humerus. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36: 379-384, 6 figures. Lowe, Percy R. 1939. On the Systematic Position of the Swifts (Suborder Cypseli) and Humming-birds (Suborder Trochili), with Special Reference to Their Relation to the Order Passeriformes. Transactions of the Zoo logical Society of London, 24:307-348. Lucas, Frederic A. 1893. Swifts and Hummingbirds. Ibis, 5:365-371. Lydekker, Richard 1891. Catalogue of the Fossil Birds in the British Museum (Natural His tory). 368 pages, 75 figures. London: Taylor and Fransis. Mourer-Chauvire\ Cecile 1978. La poche a phosphate de Saint-Neboule (Lot) et sa faune de Verte- bres du Ludien superieur; Oiseaux. Paleovertebrata, 8(2-4): 217-229, 1 figure, 2 plates. Peters, D.S. 1985. Ein neuer Segler aus der Grube Messel und seine Bedeutung fur den Status der Aegialornithidae (Aves: Apodiformes). Senckenbergiana Lethaea, 66(1/2): 143-164, 8 figures, 4 tables. Stegmann, B.K. 1965. [On Morphology of the Distal Parts of Avian Wing.] Zoological Journal, 44(3):423^*32. [In Russian.] 1970. [On Reduction of the Wing Musculature in the Process of Evolution in Aves.] Transactions of the Zoological Institute of the USSR Acad emy of Sciences, 47:249-261, 8 figures. [In Russian.] Wetmore, Alexander 1960. A Classification for the Birds of the World. Smithsonian Miscella neous Collections, 139(11): 37 pages. Zusi, Richard L., and Gregory Dean Bentz 1982. Variation of a Muscle in Hummingbirds and Swifts and Its System atic Implications. Proceedings of the Biological Society of Washing ton, 95(2):412^120, 2 figures. 1984. Myology of the Purple-throated Carib (Eulampis jugularis) and Other Hummingbirds (Aves: Trochilidae). Smithsonian Contribu tions to Zoology, 385: 70 pages, 20 figures. Selmes absurdipes, New Genus, New Species, a Sandcoleiform Bird from the Oil Shale of Messel (Germany, Middle Eocene) D. Stefan Peters ABSTRACT Selmes absurdipes, new genus, new species, is established for two fossil specimens from Messel. The pamprodactyl foot, with unusually short toes and comparatively long tarsometatarsus and tibiotarsus, is the most characteristic feature of the new genus. Sandcoleiformes were not confined to North America and had a considerable morphological radiation. It might be appropriate to combine Sandcoleiformes and Coliiformes. Introduction Surprisingly, the great majority of fossil birds from the lake deposits of Grube Messel are land birds. Many of them are dif ficult to classify with extant taxa even on the ordinal level (Pe ters, 1991, 1992). Houde and Olson (1992) established the or der Sandcoleiformes for a variety of species from the Eocene of North America. Some of these birds were previously as signed to various other higher taxa. Subsequently, an examina tion of several Messel birds revealed that the new order was not confined to North America. Two specimens of a supposed spe cies of Sandcoleiformes from Messel are described herein. The anatomical terminology used is after Baumel et al. (1993) un less otherwise indicated. ACKNOWLEDGMENTS.?I am indebted to S. Rietschel, Lan- dessammlungen fur Naturkunde, Karlsruhe, who kindly lent the paratype of Selmes absurdipes. I wish to thank Storrs Ol son, Peter Houde, and an anonymous reviewer for their critical review of the manuscript. D. Stefan Peters, Forschungsinstitut Senckenberg, Senckenberganlage 25, D-60325 Frankfurt am Main, Germany. Systematics SANDCOLEIFORMES SANDCOLEIDAE REMARKS.?Placed within order and family by the following characters: apertura nasi ossea large and holorhinal; rostrum maxillae distal to the nostrils rather short; mandibula curved (ventrally concave, dorsally convex) and with a short symphys is; fenestrae mandibulae absent; furcula thin and lacking a hy- pocleideum; olecranon short and blunt; papillae remigiales ab sent; processus intermetacarpalis absent; os metacarpale majus and o. m. minus subequal in distal extent; the three proximal phalanges of digit IV very short; phalanges unguales large and with strong flexor tubercles. Selmes, new genus FIGURES 1-3 TYPE SPECIES.?Selmes absurdipes, new species, the only known species of the genus. ETYMOLOGY.?Anagram of Messel. Selmes should be treat ed as masculine in gender. DIAGNOSIS.?The new genus differs from all known sand coleiform genera by the unique morphology of its hind limbs. The tarsometatarsus is rather slender and is markedly longer than the longest toe; its distal end is only slightly broadened, approaching the condition of Coliiformes. Not only the proxi mal phalanges of toe IV but also the proximal two phalanges of toe III and the proximal phalanx of toe II are extremely short. Toes III and IV are of almost equal length. The foot is pampro dactyl and possibly was facultatively anisodactyl. In addition, Selmes differs from Sandcoleus Houde and Ol son, 1992, by having a comparatively shorter and thicker bill; it differs from Chascacocolius Houde and Olson, 1992, by having neither long processus retroarticulares mandibulae nor a marked epicondylus of humerus nor processus intermetacarpales. 217 218 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?Selmes absurdipes, n. gen., n. sp. Holotype, SFM-ME 2375. Scale=2 cm. Photo, Forschungsinstitut Senckenberg, S. Trankner. Selmes absurdipes, new species TYPE LOCALITY.?Olschiefergrube Messel, Hessen, Ger many. HOLOTYPE.?Slab with partly deformed skeleton, sternum HORIZON.?Lower middle Eocene, Lower Geiseltalium, and pelvis lacking, Forschungsinstitut Senckenberg SFM-ME Messel Formation. 2375. PARATYPE.?Slab with postcranial skeleton, lacking most of NUMBER 89 219 FIGURE 2.?Selmes absurdipes, n. gen., n. sp. Holotype, SFM-ME 2375. Coated with ammonium chloride. Scale=2 cm. Photo, Forschungsinstitut Senckenberg, S. Trankner. sternum and pelvis, Landessammlungen fur Naturkunde Karlsruhe, ME 313. ETYMOLOGY.?From the Latin absurdus, absurd, senseless, crazy; and pes, foot; noun used in apposition. DESCRIPTION.?Both skeletons are fixed in slabs and are heavily crushed. In addition, the bones of the holotype are plas tically deformed, a condition frequently encountered in fossils from Messel. For this reason, only approximate measurements can be given. Skull: The calvaria is crushed and compressed rostrocau- dally so that the rim of the orbita is bent off and is not inspect- able; therefore, little can be said about the various processus and fossae. They apparently were not very prominent, howev er, because no indications of these structures can be traced on the preserved parts of the calvaria. In the orbita most of the annulus ossicularis sclerae is preserved and contains a black "pupil" of some organic matter (Figure 1). The bill is nearly conical. The apertura nasi ossea is similar to that of Sandco- 220 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 3.?Selmes absurdipes, n. gen., n. sp. Paratype, ME 313. Coated with ammonium chloride. Scale=2 cm. Photo, Forschungsinstitut Senckenberg, S. Trankner. leus copiosus Houde and Olson (1992, fig. 2). Most probably there was a septum nasi, otherwise it would be difficult to in terpret the amorphous bony matter filling the nostril; how ever, a fragment of a strap-like structure might be a part of the palatinum. The dorsal outline of the rostrum maxillae is only moderately curved. The mandibula seems to be slightly more robust than the specimens figured by Houde and Olson (1992). There are no fenestrae mandibulae; some organic matter and shadows (Figures 1, 2) simulate apertures, but there are none. Measurements are as follows: maximal skull length 37.5 mm; nostril length 7.5 mm; mandibula length 26.5 mm. Apparatus hyobranchialis: A considerable number of frag ments of this structure are preserved. The paraglossum and the rostral end of the basihyale cannot be seen. The basihyale is rather broad and is not fused with the urohyale. The latter is a slender bone, tapering caudally, about 3.5 mm long. The right ceratobranchiale is completely preserved (6 mm long, rostral- end diameter 1 mm), whereas the caudal ends of both epibran- chialia are lost (length of preserved part of right epibranchiale 6 mm, rostral-end diameter 0.5 mm). Vertebrae: Although many vertebrae are preserved, they are very deformed and are almost useless for diagnostic pur poses. Even their exact number cannot be ascertained. In the paratype the last free thoracic can be measured ventrally (length 3.5 mm). The synsacrum consists very probably of 11 vertebrae. There are at least five free caudal vertebrae. The py- gostyle is not preserved. NUMBER 89 221 Costae: Deformed fragments of four left ribs can be seen in the paratype. No processus uncinati can be detected. Sternum: Pieces of the sternum are preserved only in the paratype. They are badly crushed, covering as a thin layer parts of the ribs and femur. Furcula: The furcula is thin and U-shaped. There is no hy- pocleideum. In both specimens the dorsal ends of this bone are insufficiently preserved. The thickness of the scapus is continu ously about 1 mm. Coracoideum: In the holotype, fragments of the coracoids are deformed past recognition. In the paratype they are in better condition. They have a slender shaft and a rather small extrem- itas omalis. The presence of a processus procoracoideus is un certain. Unfortunately, the lateral and medial parts of the ex- tremitates sternales are only partly preserved. What can be seen, especially from the left coracoid of the paratype, suggests abroad sternal extremity. Maximal length of the coracoid is ap proximately 16 mm. Scapula: Parts of the left scapula of the holotype and of both scapulae of the paratype are preserved. All three bones have an even, ribbon-like shape, without any terminal enlarge ment. They are 1.6-1.7 mm broad. The cranial extremities are hidden by other bones. Humerus: All humeri have the cranial surfaces exposed. They are robust, slightly curved, and approach in shape the hu merus of Anneavis Houde and Olson, 1992, but have the tuber culum dorsale and epicondylus dorsalis less prominent (fide Houde and Olson, 1992, fig. 8). The length is 21 mm and the midshaft width is 3 mm. Ulna: The ulna is stout, only slightly curved, and of the same length as the humerus or slightly shorter. Radius: The radius is straight and robust. The extremities are not well preserved. Os carpi radiale and o. c. ulnare: In the holotype a de formed ossicle attached to the distal end of the left ulna might be the o. c. ulnare, and both carpalia are preserved in the right hand of the holotype and in the left hand of the paratype. Their condition is so bad, however, that no useful details can be de tected. Carpometacarpus: This is a robust bone. Apparently, the processus extensorius was only moderately protuberant. No processus intermetacarpalis is present. The proximal end of the os metacarpale minus has a rectangular flange projecting ven trally. A blurred structure in the paratype suggests that this flange might have been perforated, as in Coracias garrulus Linnaeus. The length is 14-15 mm, and the distal-end width is 4.5 mm. Digitus alulae: In the holotype a small ossicle at the tip of the digit of the left hand very probably represents the second phalanx of this digit. The length of the first phalanx is 6 mm. Digitus major: The proximal phalanx is not fenestrated and approaches the shape of that of Passeriformes. Measurements are as follows: phalanx proximalis length 6 mm, distal-end width 3 mm; phalanx distalis length 6 mm. Digitus minor: The only phalanx is robust; it has a triangu lar outline and is about 3 mm long. Pelvis: The small fragments in the paratype suggest that the pelvis was rather wide. Inferring from the position of the femo ra, the distance between the acetabula was about 11-12 mm. Femur: No precise morphological details of this bone can be imparted. Its length is about 20 mm. Tibiotarsus: This is a slender, straight bone. In the holotype the cranial aspect of the right tibiotarsus is exposed; in the paratype the lateral side of the left tibiotarsus can be seen. The latter one is broken near the distal end, and both fragments have slipped together. In the holotype, part of the proximal end can be seen. Apparently, the cristae cnemiales were not very prominent, similar to the condition in Coracias. The distal end in both specimens is very deformed. The length is 30 mm, and the midshaft width is 3 mm. Fibula: Not preserved. Tarsometatarsus: The tarsometatarsus also is comparative ly slender and straight. In the holotype both tarsometatarsi have the dorsal aspect exposed, the left one in a reversed position. In the paratype the left tarsometatarsus is exposed lateroplantarly. Although the bones are plastically deformed, it can be seen that they had a small hypotarsus and an unusually small trochleae. Both the distal and the proximal ends of the bone are only slightly broader than the shaft. The length is 20-21.5 mm. Ossa digitorum pedis: In the holotype the entire set of toes is preserved; in the paratype only the toes of the left foot can be seen in their plantar and partly lateral aspect. In the right foot of the holotype and the left one of the paratype, the toes are in a pamprodactyl position. In the left foot of the holotype, the con figuration seems to be anisodactyl, but this might be an artifact because the toes are disarticulated from the tarsometatarsus. The most striking feature is in the proportions of the phalanges (p), as shown by the following measurements (in mm). digit Lpl, 4.5-5.5; p 2, 4.0-4.5 digit II: p 1, 2.0; p 2, 6.0; p 3, 4.5-5.5 digit III: p 1, 2.5; p 2, 2.5-3.0; p 3, 6.0; p 4, 5.0-6.0 digit IV: p 1,2.0; p 2, 1.5-2.0; p 3, 1.5-2.0; p 4,6.0; p 5,5.5-6.0 Feathers: There are some small remnants of feathers in the holotype (Figure 1), suggesting that the birds had quite long remiges or rectrices. Contents of the Digestive System: In the holotype at least 25 densely packed seeds are preserved. Surely, this was the bird's last meal. It is difficult, however, to decide whether the seeds were in the stomach or in the crop. The seeds obviously belong to a dicotyledonous plant, but their identity is as yet unknown. Discussion Selmes absurdipes shows that not only the distribution but also the morphological range of the Sandcoleiformes is wider than initially presumed. The birds of this order are real taxo nomic mosaics, having similarities with many other groups. For this reason it might be worthwhile to reflect on the defini- 222 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY tion of the Sandcoleiformes. It is my impression that they prob ably could be combined with the Coliiformes. The taxonomic definition of the latter is founded on the features of a small group of very similar species, and the resultant narrowness of the taxonomic definition makes it unduly difficult to include new taxa with the Coliiformes. There are, however, substantial similarities between sandcoleids and colies, and Selmes adds to these similarities (intemasal septum, distal end of tarsometar- sus). But because I have not seen the original material of most sandcoleiform species, these considerations should be taken only as suggestions. The new species undoubtedly was a highly specialized bird. Houde and Olson (1992:143) emphasized that the pedal pha langes in Sandcoleidae are "extremely short" and they figured as an example the foot of Anneavis anneae Houde and Olson (1992). In Selmes the phalanges are even more shortened. Whereas in Anneavis all phalanges are longer than broad, in Selmes the proximal phalanges of toes II, III, and IV are broader than long. In the latter the tarsometatarsus is marked ly longer than the longest toe, and toes III and IV are of al most equal length. In Anneavis, toe III is the longest by far and equals the tarsometatarsus in length. It is hard to imagine how Selmes used its feet. Inferring from the construction of the toes, which is closest to that of swifts, it could only cling to more or less sloping surfaces. Perching would seem to have been nearly impossible; but, why are the toes so short and the rest of the leg comparatively long? As yet, there is no answer. Maybe the identification of the seeds will shed some light on the behavior of this remarkable bird. KURZFASSUNG Selmes absurdipes, n. gen., n. sp., wird auf der Beschreibung zweier fossiler Vogel von Messel begriindet. Das wichtigste Kenn- zeichen der neuen Gattung ist der pamprodactyle FuB mit ungewohnlich kurzen Zehen, aber verhaltnismaBig langem Tar sometatarsus und Tibiotarsus. Die Sandcoleiformes waren nicht auf Nordamerika beschrankt und entwickelten eine beachtliche morphologische Radiation. Moglicherweise sollten Sandcolei formes und Coliiformes vereinigt werden. Literature Cited Baumel, J.J., A.S. King, J.E. Breazile, H.E. Evans, and J.C. Vanden Berge, editors 1993. Handbook of Avian Anatomy: Nomina Anatomica Avium. Publica tions of the Nuttall Ornithological Club, 23: second edition, xxiv+779 pages. Cambridge, Massachusetts: Nuttall Ornithological Club. Houde, P., and S.L. Olson 1992. A Radiation of Coly-like Birds from the Eocene of North America (Aves: Sandcoleiformes New Order). In K.E. Campbell, Jr., editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36: 137-160, 21 figures, 2 plates. Peters, D.S. 1991. Zoogeographical Relationships of the Eocene Avifauna from Messel (Germany). In Acta XX Congressus Internationalis Ornithologici, 1:572-577. Wellington, New Zealand: Ornithological Congress Trust Board. 1992. Messel Birds: A Land-Based Assemblage. In S. Schaal and W. Zie- gler, editors, Messel: An Insight into the History of Life and of the Earth, pages 135-151, figures 198-218. Oxford: Clarendon Press. A Fossil Screamer (Anseriformes: Anhimidae) from the Middle Tertiary of Southeastern Brazil Herculano M.F. Alvarenga ABSTRACT A new genus and species of anhimid is described based on 14 fragmentary, partially associated bones collected from the shales of the Tremembe Formation of the Taubate basin in southeastern Brazil. The age of this formation is either upper Oligocene or lower Miocene. An isolated, almost complete left coracoid was chosen as the holotype. This bird, the first paleospecies described for the family Anhimidae, was smaller and more gracile than Chauna chavaria (Linnaeus), the smallest species of living screamers. Introduction The screamers (Anseriformes: Anhimidae) are distinguished from the more widespread family Anatidae by the narrow, downwardly hooked bill lacking filtering fringes, long legs, and large, unwebbed feet with strong hind toes. The skeleton of anhimids is very noticeably pneumatized. There are three mod em species of screamers in two genera, Anhima cornuta (Lin naeus), Chauna chavaria (Linnaeus), and C. torquata (Oken), all of which are endemic to South America. There are no fossil species described for this family (Olson, 1985), but unde scribed fossils of possible screamers are known from the early Eocene of Wyoming and England (S.L. Olson, pers. comm., 1995). In the present study, a new genus and species of anhim id is described, the first record for this family in the Tertiary of South America. The Tremembe Formation, in the town of the same name in the Taubate basin of southeastern Brazil (Figure 1), is a lacus trine deposit of small extent. The sediments are composed of alternate layers of thinly foliated bituminous shales, 6 to 10 meters thick, and an almost homogeneous montmorillonitic Herculano M.F. Alvarenga, Departamento de Zoologia, Instituto de Biociencias, Universidade de Sao Paulo, Caixa Postal 11294, Sao Paulo-SP, 05422-970 Brazil. clay of about the same thickness (Figure 2). Both of these lay ers have produced a large diversity of fossil vertebrates, includ ing representatives of six families of birds (Alvarenga, 1982, 1985, 1988, 1990, 1995). The age of the Tremembe Formation is either upper Oligocene or lower Miocene, as discussed by Soria and Alvarenga (1989), Alvarenga (1990), and Vucetich etal. (1993). The presence of flamingos, such as Palaelodus and Ag- nopterus (Alvarenga, 1990), a cathartid vulture (Alvarenga, 1985), and a large number of small fossil fishes and Crustacea suggest the interpretation of the site as an old lake of shallow, alkaline water. Some large and small mammals (Soria and Al varenga, 1989; Vucetich et al., 1993) and a large phorusrhacid (Alvarenga, 1982) also are known from these sediments. MATERIALS AND METHODS.?Among the birds described from the Tremembe Formation, some specimens were found articulated in the shales, whereas others, in the layer of mont morillonitic clay, often occurred as fragmented and dissociated bones, which caused some difficulties in relegating them to a particular taxon. The fossils described herein were collected by the author from the montmorillonitic clay on different occa sions from 1978 to 1993. They belong to at least three individ uals, all identified as anhimids of a size compatible with that of a single species. The bones were compared with those of skeletons of almost all families of birds, especially Anhimidae {Chauna chavaria, C. torquata, Anhima cornuta), Anatidae {Anseranas semipal- mata (Latham)), and some Gruidae, including Grus, Balearica, and Anthropoides. The single skeleton of Chauna chavaria used in this study was obtained on loan from the National Mu seum of Natural History (USNM; collections of the former United States National Museum), Smithsonian Institution, and all others are from the author's collection (HA). Anatomical terminology follows that of Howard (1929) and Baumel et al. (1979), although I also follow Olson (1987) in using the term "procoracoid foramen" in the description of the coracoid. The fossil specimens herein described are housed in the Vertebrate Paleontology collection of Museu Nacional da Universidade Federal do Rio de Janeiro (MNRJ), Brazil. 223 224 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY SAO PAULO STATE 220 30'S Taubate Tropic of Capricorn*^ _ jjn Sao Paulo Rio de Janeiro Taubate Basin Ocean FIGURE 1.?Map of Taubate basin showing the town of Tremembe, the type locality of Chaunoides antiquus, n. gen. n. sp. <- Quaternary cover layer of pyrobituminous foliations (shales) rich in small fossil vetebrates and invertebrates (mainly fishes) <- layer of homogenious montmorillonitic clay where large and small bones have been found, and the position of bones studied herein FIGURE 2.?A simplified section of the Tremembe Formation showing the stratigraphic position of the layers of shales, clay, and fossils. ACKNOWLEDGMENTS.?I thank the National Geographic Society for grants 3183-85 and 3699-87 that supported my field work. I am greatly indebted to Storrs L. Olson and Helen James (National Museum of Natural History, Smithsonian In stitution, Washington, D.C.) for loan of comparative material for this study and for helpful suggestions and comments on this paper; to Kenneth E. Campbell (Natural History Museum of Los Angeles County) for discussions, suggestions, and English language assistance on this paper; and to Elizabeth Hoffling (Departamento de Zoologia, Instituto de Biociencias, Univer sidade de Sao Paulo) for suggestions and comments on the original manuscript. Finally, my thanks to Coordenadoria de Apoio a Pesquisa e Ensino Superior, Brazil, for financial sup port for my doctoral research. Order ANSERIFORMES Family ANHIMIDAE Chaunoides, new genus TYPE SPECIES.?Chaunoides antiquus, new species. HORIZON.?Middle Tertiary, late Oligocene or early Mi ocene, Tremembe Formation. GEOGRAPHIC DISTRIBUTION.?Taubate basin, southeastern Brazil. ETYMOLOGY.?From the Latinized suffix -oideus, meaning resembling, like; referring to a bird that was similar to the ge nus Chauna. DIAGNOSIS.?Coracoid of similar size and form to that of the living species of Anhima and Chauna (Figures 3-5), including NUMBER 89 225 FIGURE 3.?Chaunoides antiquus, n. gen. n. sp., holotype, left coracoid (MNRJ 4619-V), coated with ammonium chloride: A, ventral view; B, dorsal view. (Scale bar= 1 cm.) a large pneumatic foramen in the sterno-coracoidal fossa, a large procoracoid foramen, and a large transverse procoracoid process. The coracoid differs from that of Anhima by having a (1) narrower head; (2) glenoid facet with the transverse width greater than the length, also easily observed in ventral view (Figures 3A, 5) because of its lateral expansion (similar to Chauna); (3) larger and deeper scapular facet; and (4) larger pneumatic foramen, with rounded borders, in the sterno-cora coidal fossa. The coracoid differs from that of Chauna torquata and C. chavaria by having (1) a smaller acrocoracoid process; (2) the scapular facet more rounded and deeper; (3) the dorsal surface less swollen; and (4) the large pneumatic foramen in the sterno-coracoidal fossa wider and dorsoventrally more compressed, with a small process in the dorsal border. Chaunoides antiquus, new species HOLOTYPE.?Nearly complete left coracoid, lacking only the tip of the head, tip of the procoracoid process, and the sterno- coracoidal process (Figure 3). Vertebrate paleontology collec tions of Museu Nacional de Historia Natural da Universidade Federal do Rio de Janeiro, MNRJ 4619-V. TYPE LOCALITY.?Santa Fe Farm, 2 km north of Tremem be, Sao Paulo State, Brazil (22?30'S, 45?32'W) (Figure 1). Montmorillonitic clay, about 4 m below the most superficial level of shales. HORIZON AND AGE.?Tremembe Formation, Taubate basin, upper Oligocene or lower Miocene. MEASUREMENTS OF HOLOTYPE.?See Table 1. PARATYPES.?Another left coracoid lacking the sternal end (MNRJ 4620-V), identical in morphology to the holotype, as sociated with an almost complete left femur (MNRJ 4621-V); the distal end of a left ulna (MNRJ 4622-V); the distal end of a right radius (MNRJ 4623-V); a left radius lacking the proximal end (MNRJ 4624-V); a left ulna lacking the proximal end (MNRJ 4632-V), associated with a segment of distal shaft of a left tibiotarsus (MNRJ 4631-V); and two segments of the distal end of left tibiotarsi (MNRJ 4625-V, MNRJ 4629-V), the last being associated with the proximal end of a left tarsometatarsus TABLE 1.?Measurements (mm) of Chaunoides antiquus, n. gen. n. sp., compared with other Anhimidae. Measurement Coracoid midshaft width acrocoracoid apex to internal angle Femur top of head to internal condyle least width midshaft Radius greatest width, distal end Tarsometatarsus width of proximal articular surface Tibiotarsus width immediately above tendinal bridge Ulna greatest width of external condyle Chaunoides antiquus, n. gen. n. sp. Holotype, MNRJ4619-V 12.3 66.1 - - -- - - Paratypes MNRJ-4620-V 11.7 - MNRJ 4621-V 89.7 10.0 MNRJ 4623-V, 4624-V 12.2, 12.5 MNRJ 4630-V 21.05 MNRJ 4625-4629-V, 4631-V 14,2, 14.0, 14.5, 15.0, 13.8, 14.1 MNRJ 4622-V, 4632-V 14.4,16.1 Chauna chavaria USNM 347738 (sex indet.) 11.3 65.8 92.5 11.5 15.5 23.1 15.8 17.6 Chauna torquata HA 41 HA 389 HA 702 (female) ( 15.4 71.1 97.1 12.9 17.0 23.8 17.8 19.6 sex indet.) 11.6 62.6 88.6 10.6 14.0 22.0 15.0 17.2 (male) 15.7 68.0 101.0 13.7 16.1 26.1 17.9 20.1 Anhima HA 40 (female) 13.9 64.8 93.2 12.0 15.4 21.7 16.1 18.3 cornuta HA 902 (female) 13.6 69.5 94.3 14.4 16.4 23.0 20.3 18.1 226 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 4.?Dorsal view of the left coracoid of Chaunoides antiquus, n. gen. n. sp., compared with those of liv ing anhimids: A, Chauna torquata (HA 41); B, Chaunoides antiquus, n. gen. n. sp., holotype (MNRJ 4619-V); C, Chaunoides antiquus, n. gen. n. sp., paratype (MNRJ 4620-V); D, Anhima cornuta (HA 40). The fossil bones (B and c) are coated with ammonium chloride. (Scale bar= 1 cm.) FIGURE 5.?Coracoids of anhimids, in ventral view, showing variation: A, Chauna chavaria (USNM 347738); B, Chauna torquata (HA 389); C, Chauna torquata (HA 702); D, Chauna torquata (HA 41); E, Anhima cornuta (HA 902); F, Anhima cornuta (HA 40). Note the absence of a procoracoid foramen in A and B. (Scale bar= 1 cm.) (MNRJ 4630-V). There also are three unassociated segments of DESCRIPTION AND COMPARISONS.?The coracoid of Chau- distal, right tibiotarsi shafts (MNRJ 4626-V, MNRJ 4627-V, MNRJ4628-V). MEASUREMENTS OF PARATYPES.?See Table 1. ETYMOLOGY.?From the Latin antiquus, antique, old, an cient. DIAGNOSIS.?As for the monotypic genus. noides is proportionally more slender than in living anhimids, and the procoracoid process is located slightly more toward the shoulder than in living species of anhimids (see Figures 3-5). The distal end of the ulna has the carpal tuberosity less pro nounced than in living anhimids, somewhat similar to Chauna chavaria (Figure 6). The distal end of the radius of Chaunoides NUMBER 89 227 B C D FIGURE 6.?The distal end of left ulna, in distal view (left) and in anconal view (right), of Chaunoides antiquus, n. gen. n. sp., compared with other Anhimidae: A, Anhima cornuta (HA 902); B, Chauna torquata (HA 702); C, Chauna chavaria (USNM 347738); D, Chaunoides antiquus, n. gen. n. sp., (paratype, MNRJ 4632-V); E, Chaunoides antiquus, n. gen. n. sp. (paratype, MNRJ 4622-V). (Scale bar=l cm.) has a similar morphology to that in other anhimids, but, in contrast, there is no pneumatic foramen (Figure 7). The femur (Figure 8) is less robust than in Anhima and Chauna, with its neck longer and more similar to that in Anseranas and other anatids. Its distal end is similar to that of Anhima and Chauna, with a shallow patellar groove and a flat articular surface of the internal condyle. The prominent crista supracondylaris medialis, being in a posteromedial position, gives a square shape to the distal end of the femur in medial view. The tibiotarsus of Chaunoides has the distal end of the shaft slightly bowed toward the midline, and the internal condyle extends far mediad, as is generally the case in Anser iformes (Figure 9). The tendinal groove is not pneumatized, in contrast to that of extant anhimids, and its medial border is very pronounced and extends proximad to the middle of the FIGURE 7 (right).?The distal end of (A) right radius (paratype, MNRJ 4623-V), and (B) left radius (paratype, MNRJ 4624-V), of Chaunoides antiquus, n. gen. n. sp., in palmar view, compared with the left radius, also in palmar view, of other Anhimidae and Anatidae: C, Chauna chavaria (USNM 347738); D, Anhima cornuta (HA 902); E, Chauna torquata (HA 41); F, Anseranas semipalmata (HA 1201). (Scale bar=l cm.) 228 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 8.?The left femur of Chaunoides antiquus, n. gen. n. sp. (paratype MNRJ 4621 -V), coated with ammo nium chloride (right) compared with Chauna chavaria (USNM 347738) (left): A, anterior view; B, posterior view; c, distal view. (Scale bar= 1 cm.) shaft. In the proximal end of the tarsometatarsus (Figure 10), the intercotylar prominence is higher and sharper than in other Anhimidae; the morphology of the hypotarsus is very similar to that in Chauna, especially in C. chavaria, but the crista media lis is shorter and higher. Also, there is no pneumatic foramen in the proximal end of the tarsometatarsus, whereas in extant an himids this region is very pneumatized. Discussion The procoracoid foramen in Chaunoides is large and perfect ly formed, as it is in the available specimens of Anhima and some specimens of Chauna; however, in certain specimens, such as Chauna chavaria (USNM 347738) and C. torquata (HA 389), this foramen is not formed (Figure 5), a condition possibly due to immaturity or to intraspecific variation. Olson (1987) commented on this same variation in several genera of Accipitridae as possible intraspecific or intrageneric variation. The pneumatic foramen in the sterno-coracoidal fossa of Chaunoides and Chauna is a very large opening with rounded borders, being the extreme form of a condition that also is found in Anhima, Anseranas, Opisthocomus, gruids, some gal- liforms, and cathartids. In the three species of living anhimids, NUMBER 89 229 FIGURE 9.?Right tibiotarsus of (A) Chauna chavaria (USNM 347738) in anterior view, compared with paratyp- ical tibiotarsi of Chaunoides antiquus, n. gen. n. sp.: B, MNRJ 4628-V; C, MNRJ 4626-V; D, MNRJ 4627-V; E, MNRJ 4625-V; F, MNRJ 4631-V; G, MNRJ 4629-V. (Scale bar=l cm.) the bones are extremely pneumatized, especially the distal end of the radius, the tendinal groove of the tibiotarsus, and the proximal end of tarsometatarsus. This condition is not observed in Chaunoides. The distal end of the femur of Chaunoides is very similar morphologically to that of the other anhimids, but the long neck and the thin shaft give this bone an appearance more like that of anatids, including Anseranas. In the tarsometatarsus of Chaunoides, the higher intercotylar prominence is quite differ ent from that in extant screamers, but the hypotarsus, with only two calcaneal ridges, is typical of the Anhimidae and also is found in Paranyroca magna Miller and Compton, 1939, from the lower Miocene of South Dakota, in contrast to that of the remainder of the Anatidae, which have four calcaneal ridges. The proportions of the femur, tibiotarsus, and tarsometatarsus of Chaunoides suggest that it was slightly smaller than Chauna chavaria, the smallest living screamer, and also more gracile, with the leg bones more slender and with the skeleton less pneumatized than in the living anhimids. Conclusion Chaunoides antiquus is the first fossil species of the family Anhimidae to be recognized. At the Natural History Museum, London (formerly the British Museum (Natural History)), in 1992, I examined the holotype of Loxornis clivus Ameghino, 1895 (placed in incertae sedis by Tonni (1980)), which is from the Oligocene (Deseadean) of Argentina. I concluded that this bird may be a representative of the Anhimidae, but because of its poor preservation (only the distal end of a left tibiotarsus, with the medial condyle incomplete), additional material would be required to substantiate this hypothesis. 230 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 10.?Left tarsometatarsus of Chaunoides antiquus, n. gen. n. sp. (paratype, MNRJ 4630-V), coated with ammonium chloride (right) compared with Chauna chavaria (USNM 347738) (left): A, anterior view; B, poste rior view; c, lateral view; D, proximal view. (Scale bar= 1 cm.) Literature Cited Alvarenga, Herculano 1982. Uma gigantesca ave fossil do Cenoz6ico brasileiro: Physornis bra- siliensis sp. n. Anais da Academia Brasileira de Ciencias, 54(4):697-712. 1985. Notas sobre os Cathartidae (Aves) e descricao de um novo genero do Cenoz6ico brasileiro. Anais da Academia Brasileira de Ciencias, 57(3):349-357. 1988. Ave fossil (Gruiformes: Rallidae) dos folhelhos da Bacia de Taubate, Estado de Sao Paulo, Brasil. Anais da Academia Brasileira de Ciencias, 60(3):321-328. 1990. Flamingos f6sseis da Bacia de Taubate, Estado de Sao Paulo, Brasil: Descricao de nova especie. Anais da Academia Brasileira de Cien cias, 62(4):335-345. 1995. Um primitivo membro da Ordem Galliformes (Aves) do Terciario Medio da Bacia de Taubate, Estado de Sao Paulo, Brasil. Anais da Academia Brasileira de Ciencias, 67(l):33-44. Ameghino, Florentino 1895. Sur les oiseaux fossiles de Patagonie. Boletin del Instituto Geogrd- fico Argentino, 15:501 -602. Baumel, Julian J., A.S. King, A.M. Lucas, J.E. Breazile, and H.E. Evans, editors 1979. Nomina Anatomica Avium: An Annotated Anatomical Dictionary of Birds, xxv+637 pages. London: Academic Press. Howard, Hildegarde 1929. The Avifauna of the Emeryville Shellmound. University of Califor nia Publications in Zoology, 32(2):301-394. Miller, Alden H., and L.V. Compton 1939. Two Fossil Birds from the Lower Miocene of South Dakota. Con dor, 41:153-156. Olson, Storrs L. 1985. The Fossil Record of Birds. In Donald S. Famer, James R. King, and Kenneth C. Parkes, editors, Avian Biology, 8:79-238, 11 figures. New York: Academic Press. 1987. Variation in the Procoracoid Foramen in the Accipitridae. Rivista Italianadi Omitologia, 57(3?4): 161?164. Soria, Miguel, and H. Alvarenga 1989. Nuevos restos de mamiferos de la Cuenca de Taubate, Estado de Sao Paulo, Brasil. Anais da Academia Brasileira de Ciencias, 61(2):157-175. Tonni, Eduardo P. 1980. The Present State of Knowledge of the Cenozoic Birds of Argentina. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Hildegarde Howard. Contributions in Science. Natural History Mu seum of Los Angeles County, 330:105-114. Vucetich, M., F. Souza-Cunha, and H. Alvarenga 1993. Un roedor Caviomorpha de la Formaci6n Tremembe (Cuenca de Taubat6), Estado de Sao Paulo, Brasil. Anais da Academia Brasileira de Ciencias, 65(3):247-251. The Anseriform Relationships of Anatalavis Olson and Parris (Anseranatidae), with a New Species from the Lower Eocene London Clay Storrs L. Olson ABSTRACT An associated partial skeleton, including the skull but lacking legs, from the lower Eocene London Clay of Essex, England, pos sesses derived characters of the coracoid and furcula that show it to belong to the Anseranatidae, which previously had no fossil record. Except for its much larger size, the humerus of this speci men is identical to that of Anatalavis rex (Shufeldt) from the late Cretaceous or early Paleocene of New Jersey. The Eocene speci men is described as a new species, Anatalavis oxfordi, and the genus Anatalavis is transferred from the form-family Graculavidae to a new subfamily, Anatalavinae, of the Anseranatidae. Anatala vis is characterized by a very broad duck-like bill, a proportion ately very short and robust humerus, and an anterior portion of the pelvis resembling that of ibises and other wading birds more than that of any known anseriform. Other features of its osteology are unique within the order. Introduction Waterfowl of the order Anseriformes are among the best known and most distinctive groups of modern birds. Although waterfowl are abundantly represented by Neogene fossils, much of their early evolutionary history has remained obscure. The most informative fossils until now have been the Pale ocene and early Eocene remains of Presbyornis and its close relatives, which were nearly cosmopolitan in the early Paleo gene. Presbyornis was shown to have a duck-like skull on the body of a long-legged wading bird and was interpreted as showing a derivation of the Anseriformes from a charadrii- form-like ancestor (Olson and Feduccia, 1980b) rather than from the Galliformes, as had been postulated previously. The Storrs L. Olson, Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560- 0116, United States. more recent studies of Ericson (1996, 1997) confirmed the lack of relationship between the Anseriformes and Galliformes, but the ancestry of the Anseriformes was unresolved beyond a complex of various groups of wading birds, including Charadriiformes. Presbyornis, however, was determined to have branched off within the order and constitutes the sister group of the Anatidae proper, with the Anhimidae and Anser anatidae being the primitive outliers of the Presbyornithidae/ Anatidae clade. The giant Paleocene and Eocene groundbirds of the genus Diatryma, once thought to have been predatory descendants of crane-like birds, also may be part of the anseriform radiation (Andors, 1988, 1992). The dietary habits of Diatryma, how ever, have been equivocated (Andors, 1992; Witmer and Rose, 1991). Although no fossils of screamers (Anhimidae) had hitherto been reported, a somewhat more primitive genus is now known from the middle Tertiary of Brazil (Alvarenga, this volume), and I have examined excellent fossils, as yet undescribed, from the lower Eocene Willwood Formation of Wyoming and from the contemporaneous London Clay of England. Thus, of the three major lineages of living Anseriformes, the only one with no early Tertiary (or later) fossil representative is the Anser anatidae, with its sole member being the Magpie Goose, Anser anas semipalmata, of Australia. Two bones from the Hornerstown Formation of New Jersey, a deposit of debated Late Cretaceous or early Paleocene age (Olson, 1994; Hope, this volume), were thought to show some similarities to the Anatidae, but in the absence of associated material they were assigned to the form-family Graculavidae, which contains various taxa resembling primitive Charadrii formes (Olson and Parris, 1987). A new fossil from the London Clay consisting of much of a skeleton, although lacking legs, permits positive identification of the New Jersey fossils as not only belonging to the Anseriformes but belonging to the family Anseranatidae. Thus, the New Jersey and London Clay fossils 231 232 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY provide the first record of this family, as well as the earliest certain occurrence of the entire order. The present paper is intended to provide a name for the new fossil from the London Clay and to place on record its more sa lient osteological features. Full analysis of anatomical details and phylogenetic significance will have to await the appear ance of information not presently available for certain extant but sequestered fossils, especially those of Eocene Anhimidae. Nomenclature for species' binomials and English names of modern birds herein follows Sibley and Monroe (1990). ACKNOWLEDGMENTS.?I must begin by crediting the singu lar cooperativeness of Andrew Oxford, who collected the spec imen that is the primary object of the present study and who, at my first suggestion, donated it to the Natural History Museum, London (formerly the British Museum (Natural History)) (BMNH). In connection with this transaction, I cannot fail to mention that the hospitality shown to me and my family by Mr. and Mrs. Oxford at their domicile at Great Mongeham, Deal, Kent, afforded us some of our most pleasant memories generat ed during a pleasant year in England. Mr. Oxford and I each owe a debt of gratitude to that veteran collector of fossil birds of the London Clay, Michael Daniels, who separately intro duced us to the incredibly productive mudflats at Wal- ton-on-the-Naze, who led me to contact Mr. Oxford about his fossil bird, and whose comments on the manuscript inspired me to delete much that was equivocal. The repeated hospitality of Mr. and Mrs. Daniels at their home in Holland-on-Sea, Clac- ton, Essex, has been of inestimable benefit to me for my knowledge of early Eocene birds. At the BMNH, Angela Milner greatly facilitated the present study by making arrangements for the incorporation of the specimen into the collections and having it prepared and photo graphed. In this connection, there are hardly sufficient words of praise for William Lindsay, whose painstaking removal of an exceedingly fragile and difficult specimen from its envelope of clay and pyrite was undertaken in such a spirit of collaboration and sensitivity that his contribution must rank equal to that of any scientist who studies these bones. Sandra Chapman has re peatedly been of assistance during my study of fossil bird col lections at the BMNH. Robert Prys-Jones and Cyril Walker, of the Bird Group at Tring, were instrumental in lending compar ative modern skeletons. The photographs are by Phil Crabb, Natural History Museum, London, Photo Unit, except Figure 8c,D, which is by the Smithsonian Photographic Services (SPS). Carl Hansen (SPS) was instrumental in assisting with the electronic composition of the figures. Order ANSERIFORMES The primary adaptation of the order Anseriformes is the modification of the bill so that the upper jaw houses an en larged tongue, which functions as a double-piston pump used in filter-feeding (Olson and Feduccia, 1980a), giving rise to a characteristic "duck-billed" shape that may be modified sec ondarily for other feeding functions. The following features also are characteristic of the Anseriformes, some being conver gent with Galliformes (Olson and Feduccia, 1980b; Ericson, 1996): the configuration of the quadrato-mandibular articula tion; the enlarged, deep, curved, blade-like retroarticular pro cess of the mandible; and the enlarged rounded or ovoid "ba- sipterygoid process" on the parasphenoid rostrum, with a corresponding enlarged facet on the pterygoid. These charac ters are practically all that exist to demonstrate the anseriform relationships of both living and fossil screamers (Anhimidae), in which the bill has either lost or never had the adaptations for filter-feeding. The holotype of the new species described below has a skull with all of the features typical of the order Anseri formes, to which it clearly belongs. Family ANSERANATIDAE Recognition of the living Australian Magpie Goose, Anser anas semipalmata, as a monotypic family of Anseriformes has been supported by several anatomical studies (e.g., Verheyen, 1953; Woolfenden, 1961; Livezey, 1986), and the distinctive ness of this species is confirmed by DNA sequencing as well (Michael Sorensen, University of Michigan, pers. comm., 1996). Two presumably derived characters group the fossil ge nus Anatalavis in the same family as Anseranas. The first is the unique V-shaped furcula with a large, deep symphysis. In other members of the order the furcula is an unelaborated U-shaped structure with a symphysis that is scarcely, if at all, larger than the rami. I interpret the former condition to be a derived char acter within the Anseriformes. If it is primitive, the only out groups that show any similarity (and this only in a general way) are storks (Ciconiidae) and, to an even lesser extent, herons (Ardeidae). The second character is the presence of a distinct, large pneumatic foramen in the dorsal surface of the sternal end of the coracoid. Although a similar condition exists in modem Anhimidae, which have one of the most pneumatized skeletons in any group of birds, this foramen is absent in Eocene Anhim idae, which are evidently completely nonpneumatic (pers. obs.). Therefore, the condition in modern screamers is obvious ly independently derived and is not an indication of relation ship with the Anseranatidae. Subfamily ANSERANATINAE INCLUDED GENUS.?Anseranas Lesson. DIAGNOSIS.?Subfamilial characters are intended as diag nostic only within the Anseranatidae. Rostrum strong, deep, and hooked; frontal area with a large bony casque; attachment for M. depressor mandibulae not greatly developed. Humerus of normal anseriform proportions. Pelvis with anterior portions of the ilia narrow. NUMBER 89 233 Subfamily ANATALAVINAE, new subfamily INCLUDED GENUS.?Anatalavis Olson and Parris, 1987. DIAGNOSIS.?Rostrum very broad and shallow, not hooked; frontal area without bony casque; attachment for M. depressor mandibulae greatly enlarged. Humerus proportionately very short and extremely robust. Pelvis with anterior portions of ilia markedly expanded and rounded. Anatalavis Olson and Parris, 1987 TYPE SPECIES.?Telmatomis rex Shufeldt, 1915. INCLUDED SPECIES.?Anatalavis rex (Shufeldt), Anatalavis oxfordi, new species. The genus Anatalavis was proposed for the species Telma tomis rex from the Hornerstown Formation in New Jersey, which differed from the type species Telmatomis priscus Marsh in the proportionately much shorter, more robust, and curved shaft of the humerus (Olson and Parris, 1987). Anatala vis rex is known so far only from two humeri lacking the prox imal ends and was assigned to the form-family Graculavidae, which was used to include various fragmentary postcranial fos sils showing similarities to the Presbyornithidae, Burhinidae, and other families. At the time, it was recognized that if cranial material could ever be associated with any of the genera of Graculavidae, it would probably prove possible to refer them to various other families or orders (Olson and Parris, 1987). The age of the fossil birds from the Hornerstown Formation re mains controversial, being either latest Cretaceous or early Pa leocene (Olson and Parris, 1987; Olson, 1994). The fossil from the London Clay is herein assigned to Anata lavis because the humerus is identical in proportions and has the same distinctive curvature and robustness as that of A. rex (Figure 8), which it matches in all details except size. At the time the genus Anatalavis was proposed, it was thought that the humerus appeared somewhat duck-like, hence the name, and this is fully borne out by the associated fossil from the London Clay. Anatalavis oxfordi, new species FIGURES 1-9 HOLOTYPE.?Partial, associated skeleton, BMNH Depart ment of Palaeontology registry number A5922. Collected 12 October 1991 by Andrew Oxford and Michael Daniels. TYPE LOCALITY.?Tidal mudflats and basal cliffs at Wal- ton-on-the-Naze, Essex, England. HORIZON.?London Clay (Ypresian), lower Eocene. MEASUREMENTS OF HOLOTYPE (in mm).?Skull (measure ments taken from ventral aspect): Total length from posterior of cranium to tip of bill as preserved, 100 (measurement longer than it should be due to separation of cranium and rostrum); length from posterior of cranium to apparent nasofrontal hinge, 46.5; length and width of right nostril, 8.6 x 5; length from an terior margin of nostril to bill tip, 34; maximum width of bill as preserved, 27.5. Skull (measurements taken from dorsal aspect): Width of interorbital bridge, 21.5; width of frontals at nasofrontal hinge, 12.7; dorsal length and width of narial opening, 12.0 x 6.5; width of internarial bar, -2.5; width of cranium across squamo sal protuberances, 25.3; depth (including occipital condyle) and width (across occipital condyle) of area of cervical muscle attachment, 18.5 x 19.3. Mandible: Length of retroarticular process, 11.0; depth of retroarticular process at midpoint, 6.9. Pterygoid: Total length, 11.4; greatest diameter of basip- terygoid facet, 4.6. Atlas: Depth, 9.8; width, 8.5. Axis: Length of centrum, 12.2; depth, 10.8. Thoracic Vertebra (19th?): Length of centrum, 10.9. Caudal Vertebra: Width (double the distance from tip of transverse process to midline), 17.8; length of centrum, 6.7. Furcula: Length from apex of right ramus to farthest extent of symphysis, 53.7; depth of symphysis, 14.4; width and depth of ramus at broadest point, 7.0 x 2.1. Coracoid: Length from head to internal angle, 48.0; width and depth of shaft at approximate midpoint (narrowest point below procoracoid process), 8.7 x 4.5; depth through head, 9.3; distance from distal margin of procoracoid foramen to internal angle, 29.0; width of sternal articulation, 22.3. Scapula: Length from acromion to posterior tip, -79; width of articulation including acromion, 12.5; depth of articu lar end, 4.4; greatest width of shaft, 7.0. Sternum: Anterior depth through carina, 39.3; estimated width through third costal facet, 42.5; width of anterior base of carina, 5.6. Pelvis: Width across antitrochanters (estimated by dou bling distance to midline), 46; length from anterior margin of iliac shield to posterior margin of antitrochanter, 52; anterior depth of synsacrum, 24. Humerus: Length, 119.3; length from head to distal margin of pectoral crest, 41.7; width and depth of shaft at midpoint, 9.7 x 7.8; depth through internal tuberosity, 20; distal width and depth, 22.3 x 11.7; length of radial condyle, 11.6. Ulna: Width and depth of shaft at approximate midpoint, 6.6 x 6.8; distal depth -13; distal width, 10.7+. Ulnare: Greatest diameter, 13.5. Carpometacarpus: Length, 69.5; length from proximal symphysis to distal end, 43.8; length of intermetacarpal space, 32.8; length of distal symphysis, 11.0; proximal depth, 18.5; width of trochlea, 9.7; distal width, 10.0; greatest width of ma jor metacarpal, 6.9. A lular Digit: Length, 21.3. Major Digit, Phalanx 1: Length, 30.2; proximal width, 8.0; distal depth, 7.8. Major Digit, Phalanx 2: Length, 22.7. Minor Digit: Length, 15.2. 234 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ETYMOLOGY.?Dedicated to the collector and donor, An drew Oxford, of Great Mongeham, Kent. DIAGNOSIS.?Much larger than Anatalavis rex (Table 1). PRESERVATION OF THE HOLOTYPE.?The holotype is essen tially an associated, partially articulated skeleton lacking the posterior half of the pelvis, the tail (except one caudal verte bra), and both hindlimbs. The skull, vertebral column, pectoral girdle, and wings are present, although in various states of preservation and with some elements missing (e.g., the right wing is missing except the proximal two-thirds of the humerus and one phalanx). Cranium and Bill (Figures 1, 2): The skull is considerably distorted through compression and by having been pushed into other bones. Although the rostrum and cranium appear to be continuous, they are actually no longer articulated and have bone, mostly of the palatal region, in the intervening space. The cranium is abraded along the left (dorsal) margin, the pa latines are crushed and distorted and are pushed over to the right of the midline. The right quadratojugal is present and more or less in place. What may be most of the left quadratoju gal was broken off and is present as a separated bone. The left pterygoid is well preserved as a separate bone (Figure 3). The TABLE 1.?Comparative measurements (mm) of the humerus of the two species of Anatalavis (A. rex, holotype and paratype, from Olson and Parris, 1987). Measurement Length from distal end of pectoral crest to ulnar condyle Shaft width at midpoint Width of shaft at proximal extent of brachial depression Depth through radial condyle Distal width A. rex 49.1,50.7 5.4, 5.6 7.2, 7.5 7.3, 7.5 3.6, 13.2 A. oxfordi 80.0 9.7 13.2 12.1 22.3 rostrum has portions of the left margin abraded. The entire bill is turned upward through compression, having been pressed into an underlying portion of humerus that has made a great de pression in the left dorsal surface, which is seen as a large tu mescence in ventral view. As seen in dorsal aspect, crushing has produced a large, somewhat triangular pit in the cranium just anterior to the parietals. The nasal part of the nasofrontal articulation is very badly crushed and distorted on the left side but is somewhat better preserved on the right. The left nostril is crushed and almost obliterated, whereas the right one is almost undistorted. Lacrimal?: What is possibly a portion of the right lacri- FlGURE 1.?Skull with rostrum of Anatalavis oxfordi (holotype, BMNH A5922): A, dorsolateral view; B, ventro lateral view; c, rostrum in direct dorsal view. Scales in mm. NUMBER 89 235 mal was found under the right orbit, but it is so fragmentary Mandible (Figure 4): The mandible appears to have that no interpretation can be made of it. Regardless, it is clear slipped forward during burial, and most of it has been eroded from the cranium that the lacrimal was not fused. off the original block of matrix. Only the right articular por- FlGURE 2.?Cranium of Anatalavis oxfordi (holotype, BMNH A5922): A, left lateral view; B, right lateral view; C, posterior view. Scale in mm. 236 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 3.?Left pterygoid of Anatalavis oxfordi (holotype, BMNH A5922): A, medial (internal) view; B, dorsal view; c, ventral view. Scale in mm. FIGURE 4.?Right articular portion of the mandible of Anatalavis oxfordi (holotype, BMNH A5922): A, lateral (external) view; B, medial (internal) view; c, dorsal view. Scale in mm. NUMBER 89 237 tion, with retroarticular process and an unidentified adherent piece of bone, was preserved with the specimen, lying under the rostrum. Vertebrae: Vertebrae 1 through 4 are present, as are at least four thoracics, fragments of other vertebrae, and a single caudal lacking most of the right transverse process and neural crest. There also are various fragments of ribs and other pieces of unidentified bone. Furcula (Figure 5): Complete except lacking the very tip of the left ramus and showing some lateromedial distortion. Coracoids (Figure 6A-F): The right coracoid lacks the ex ternal angle and the tip of the procoracoid process. The left lacks the head and the tip of the external angle. Scapulae (Figure 6G): Both are present and complete ex cept the left lacks much of the coracoidal articulation. Sternum (Figure 7): This lacks the posterior one-third or so, with much of the left side being badly damaged. The anteri or part of the carina is well preserved. The stemocoracoidal processes are variously damaged or obscured, and the dorsal surface of the anterior portion is obscured by the anterior part of the pelvis and matrix. Pelvis (Figure 7): This consists of the anterior half or more. The right side is lacking posterior to the anterior iliac shield, and the left side is obscured anteriorly by the sternum but pos teriorly preserves the dorsal part of the acetabulum, antitro- chanter, and part of the posterior portion of the ilium. FIGURE 5.?Furcula of Anatalavis oxfordi (holotype, BMNH A5922): A, anterior view; B, posterior view; c, right lateral view (rotated slightly clockwise). Scale in mm. 238 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 6.?Coracoids and scapula of Anatalavis oxfordi (holotype BMNH A5922): A-D, right coracoid in (A) ventral view, (B) dorsal view, (c) lateral (external) view, (D) medial (internal) view; E,F, left coracoid in (E) dorsal view, (F) ventral view; G, right scapula in ventral view. Scales in mm. Humeri (Figure 8): The left is complete, lacking a bit of the pectoral crest and has the proximal end somewhat crushed and distorted. The right lacks the distal one-third and is much crushed and obscured by adhering matrix and bone. Radius and Ulna: These are represented only by the distal two-thirds or more of the left radius and ulna. The radius lacks much of the articular end, and the ulna has the internal condyle broken and the distal end obscured by an adhering piece of bone. Carpal Bones: Only the left ulnare (Figure 9) was located and identified. Carpometacarpus (Figure 9): The right is lacking. The left is complete, having been broken and repaired, with a small piece of the major metacarpal missing and the whole el ement showing some compressional distortion. Alar Phalanges (Figure 9): All phalanges of the major and minor digits of the left wing are present, as is phalanx 1 of the right major digit. The single alular digit is presumably that of the left side as well. NUMBER 89 239 FIGURE 7.?Sternum and pelvis of Anatalavis oxfordi (holotype, BMNH A5922): A, left lateral view; B, dorsal view. Scales in mm. FIGURE 8.?Left humeri of Anatalavis. A-C, A. oxfordi (cast of holotype, BMNH A5922) in (A) anconal view, (B) palmar view, and (c) palmar view, at lesser magnification. D, A. rex (paratype Yale Peabody Museum 948), enlarged for comparison with c. The slight differences are mainly due to slightly different rotation of the speci mens. Scale bars=l cm. 240 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 9.?Left manus (ulnare, carpometacarpus, and phalanges) of Anatala vis oxfordi (holotype, BMNH A5922) in (A) dorsal view and (B) ventral view (bottom). Scale in mm. Description and Comparisons Direct comparisons were made with the three major groups of living Anseriformes. Anhimidae: Chauna chavaria male, BMNH S/1954.3.3. Anseranatidae: Anseranas semipalmata male (by skull morphology), BMNH 1891.7.20.110; female (by skull), BMNH 1862.7.6.6 (Vellum catalog 441c). Anatidae: Dendrocygna bicolor male, BMNH S/1952.1.163; Anserfabalis, BMNH 1930.3.24.204. Using the descriptions and photographs, further comparisons were made with various other taxa of Anseriformes and with Eocene fossils assigned to the Anhimidae and Presbyornithidae, using collections of the National Museum of Natural History, Smithsonian Institution, Washington, D.C, and of Michael Daniels. CRANIAL ELEMENTS SKULL (Figures 1, 2).?There is so much crushing and dis tortion of the skull that interpretation of many of its aspects is often very difficult. The bill is obviously very different from that of Anseranas in being short and very wide, with the bone being quite thin. How concave the ventral surface may have been is now difficult to say, but it cannot have been as deeply excavated as in Anseranas. The tip is extremely broad and rounded, quite unlike the pointed, nail-like tip of Anseranas. The nostrils are very short, broad, and rounded compared with any other anseriform. The internarial bar is very narrow. As preserved, the nostrils in ventral view are almost completely exposed, there apparently being no roofing over of bone (sec ondary palate) by the maxillopalatines. If this is not the product of breakage, then the condition is unlike that of Anseranas or the Anatidae and is more like that in the Anhimidae. The poste rior flange of the rostrum below the anterior articulation of the quadratojugal bar is reasonably well developed, about as in Dendrocygna, and is not as large, elongated, and pointed as in Anseranas. The interorbital bridge is much wider than in Anseranas and bears no hint of the bony casque of that species. From the ap pearance of the left side, the lacrimals must have been unfused, as in Anseranas, which is the primitive condition shared with Presbyornis and Anhimidae, as opposed to the Anatidae, in which the lacrimals are fused. The postorbital process is quite short and blunt, unlike any of the anseriforms compared (smaller in Chauna, but pointed). The posterior portion of the temporal fossa bears extremely broad, nearly rectangular scars indicating great development of M. depressor mandibulae, perhaps more so than in any extant waterfowl. In Anseranas these scars are narrow and are much less distinct. The anterior temporal fossa is hardly distinguish able in the fossil, however, indicating lesser development of the mandibular adductors. The combination of the deep scars of the depressor mandibulae and the well-developed area of at tachment of the cervical musculature produces very distinct nuchal crests on the occiput (Figure 2C). The occipital area may have been crushed lateromedially, thus reducing the size of the foramen magnum, which seems comparatively small, al though the occipital condyle is large relative to that of Anser anas. There are two vertically oriented, narrow, elongate fo ramina situated where the large, oval occipital fontanelles occur in most Anseriformes, the two presumably being homol ogous. It is not clear whether these are evolutionarily incipient fontanelles or whether the fontanelles have become mostly closed by bone, as occurs in certain modern waterfowl. PTERYGOID (Figure 3).?By virtue of the distinctive facet for articulation with the basipterygoid process of the parasphenoid rostrum, the pterygoid of Anatalavis is recognizably anseri form, yet it differs markedly from modern members of the or- NUMBER 89 241 der. The bone is very short and robust, with the basipterygoid facet proportionately very large but nearly round in shape rath er than being an elongate oval as in Anseranas and Anser. The quadrate articulation is much larger and more expanded, and the palatine articulation is completely offset laterally from the main (long) axis of the bone, rather than being in a line with it as in other anseriforms. MANDIBLE (Figure 4).?Only the right articular is preserved, and this has a large, blade-like retroarticular process that curves upward at the tip. It is very anseriform in appearance but is shorter, deeper, and thicker than typical forms, looking more like that in Chauna, although it is relatively larger. The medial process, although broken, is very small, unlike any modern anatid, and the lateral process is likewise not nearly as well de veloped. Although partly obscured by an adhering piece of bone, the articular surface looks typically anseriform, perhaps most similar to that in Chauna. The ramus seems to rise imme diately to a deep coronoid process. The recessus conicalis is absent. This is a deep conical hol low extending on the medial side from the retroarticular pro cess anteriorly under the articulation. It represents a derived character uniting the Presbyornithidae and Anatidae, but it is lacking in Anseranas and the Anhimidae (Ericson, 1997). AXIAL POSTCRANIAL SKELETON THORACIC VERTEBRA.?The best-preserved thoracic verte bra appears to be equivalent to the 19th of Anseranas (which is the sixth in front of the sacrum and the first with a full thoracic rib, but no sternal attachment, so that technically this would be the last cervical). This is very similar to that in Anseranas ex cept that the sides of the centrum are concave, with a small pneumatic foramen that is lacking in Anseranas. Evidence from various Eocene waterbirds suggests that the condition of having concavities on the centrum, which is characteristic of the Charadriiformes, for example, may be primitive. FURCULA (Figure 5).?The furcula is absolutely distinctive in being V-shaped and having a long, broad symphysis, thus re sembling only Anseranas among the Anseriformes, which oth erwise have a simple U-shaped furcula. The fossil differs from Anseranas in that most of the symphysis is a thin, nearly trans lucent sheet of bone with a sharp, low crest running down the midline on the posterior face, whereas in Anseranas the sym physis has become thickened and pneumatized, with a pneu matic foramen on the dorsal surface and the posterior crest much less distinct. In lateral view the fossil is wide throughout but is thin and flat lateromedially, whereas in Anseranas the rami become narrower toward the symphysis but are much thicker lateromedially than in the fossil. The ramus in the fossil comes to a very sharp point dorsally but is not expanded into an angular flange on the anterodorsal edge as in Anseranas, which gives the ramus in the latter a more curved appearance. The furcula in both Anatalavis and Anseranas differs from that in the Anatidae in being less curved, with the portion pos terior to the articulation not forming an angle and extending posteriorly. The modern Anhimidae are utterly different from any of these in having a furcula that is very broad, flat, and pneumatic posteriorly on both rami. CORACOID (Figure 6A-F).?The fossil has a narrow, pointed procoracoid process, whereas in Anseranas this is much broad er, blunter, and extends farther sternally. The procoracoid pro cess in A. oxfordi has a distinct circular foramen. This is the primitive condition that in Anseriformes is retained in Anser anas, Eocene and modern Anhimidae (in the latter it may sometimes be absent; see Alvarenga, this volume), and the Presbyornithidae. The procoracoid foramen has been lost in all Anatidae. Although a similar structure appears in the New Zealand fossil genus Cnemiornis and in some individuals of Cereopsis (Livezey, 1989), this probably evolved secondarily through ossification of ligaments. On the internal dorsal surface of the sternal end of the cora coid of the fossil there is a large, sharply delimited, ovoid pneumatic foramen. This also is found in Anseranas, where it may vary from a larger, although less distinctly edged, fora men, to a depression with only a pinhole foramen. A very large pneumatic foramen occurs in about the same place on the cora coid of modern Anhimidae, but this is entirely absent in the Eocene members of the family and must therefore have evolved independently as a result of the extreme pneumatiza- tion of the skeleton in modem screamers. The whole sternal end of the bone in the fossil is more ex panded, with the external angle a longer, sweeping wing. There also is a distinct projecting angle on the medial edge just above the internal angle that is not found in Anseranas. In sternal view the medial part of the sternal articulation is not nearly as deep and expanded as in Anseranas. SCAPULA (Figure 6G).?This bone appears to be relatively longer than in the Anatidae but is not as narrow as in Anser anas and apparently has the acromion narrower and more pointed. STERNUM (Figure 7).?In lateral view the sternum differs from that of Anseranas in having the apex of the carina more rounded and undercut by a broad, rounded notch. The distinctly projecting, blunt, triangular manubrial spine in the fossil is lacking in Anseranas. In these respects the sternum in the fossil is more similar to that in Chauna; however, in Chauna the manubrium is shorter and blunter. PELVIS (Figure 7).?The remaining portion of the pelvis is very different from that in any known member of the Anseri formes because the preacetabular portion is relatively short and the anterior iliac shield is very broad, rounded, and deeply ex cavated for the iliotrochantericus muscles, leaving a broad, well-defined dorsal ridge. In overall appearance, the pelvis in the fossil is more like that in certain wading birds, such as ibis es (Plataleidae) and other Ciconiiformes, or Charadriiformes, than in any anseriform. The acetabulum in the fossil is relative ly larger than in Anseranas. 242 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY The ilia appear to be fully fused to the sacral vertebrae, as in all living Anseriformes, which makes the unfused innominates of Presbyornis seem all the more anomalous. PECTORAL APPENDAGE HUMERUS (Figure 8).?The humems is characterized by the extremely short, robust shaft with markedly sigmoid curvatures both anteroposteriorly and lateromedially. The total length of the bone is 33% shorter than in Anseranas, yet the shaft is slightly wider, indicating very different proportions and pre sumably a much different flight pattern. The pectoral crest is very broad and rounded, unlike Anseranas or the Anatidae, but is somewhat similar to Chauna. The head and internal tuberos ity are massive, and the capital groove is extremely wide and deep, unlike other anseriforms. The orientation of the internal tuberosity may have been distorted by compression and per haps pushed distally; regardless, it overhangs the bicipital fos sa, which is small and may not have been pneumatic (obscured by pyrite). The capital ridge is very well defined. The distal end is expanded and flattened compared with most Anatidae but is generally similar to that of Anseranas except that the brachial depression is longer and narrower, the radial condyle is longer, and the olecranal and tricipital fossae are slightly deeper. The twisting of the shaft is remarkable because it appears as though the distal end has been rotated clockwise about the long axis of the bone by perhaps 15? or more. RADIUS AND ULNA.?In their incomplete condition there is little to be said about these elements. Compared with Anser, the fossil radius agrees with Anseranas in the more expanded distal end and more slender shaft, which is more sharply angular in cross section, with more flattened surfaces than in Anseranas. CARPOMETACARPUS (Figure 9).?This element agrees with that of Anseranas in being short and stout, although it is more robust even than that in Anseranas. It is intermediate in length between the male and female specimens compared. The alular metacarpal is more vertically oriented than in Anseranas and is more like that in the Anatidae. It also is much blunter, with a larger digital facet. ALAR PHALANGES (Figure 9).?Phalanx 1 of the major digit is relatively short and stout, as in Anseranas, but the proximal articulation is wider and not as deep. Phalanx 2 of the major digit and the minor digit are each relatively shorter, the former much more so, than in Anseranas. Discussion The two species of Anatalavis provide the only recognized occurrence of the family Anseranatidae in the fossil record. Anatalavis rex, from the Hornerstown Formation of New Jer sey, whether Late Cretaceous or early Paleocene in age, also provides the earliest certain record of the entire order Anseri formes. The material of Anatalavis oxfordi, from the lower Eocene London Clay, is more complete than that of any early Tertiary anseriform yet described, apart from Presbyornis, and provides us with a new set of clues regarding early evolution in waterfowl. The skull in Anatalavis oxfordi indicates that it was most likely an obligate filter feeder. The bill is very broad yet is thin and weak. The retroarticular process of the mandible is quite well developed, although not nearly to the extent observable in the more extreme members of the Anatidae, yet the massive de velopment of M. depressor mandibulae shows it to have been more adapted for straining, as opposed to grasping, in which the mandibular adductors play a greater role (Goodman and Fisher, 1962). This is practically the opposite of its nearest pre sumed relative, the Australian Magpie Goose {Anseranas semi- palmata), in which the bill is strong, deep, and hooked and is used in digging out tubers and other plant material (Frith, 1967). Although the palate is very poorly preserved in the holotype of A. oxfordi, the fundamentally different morphology of the pterygoid compared with modem Anseriformes suggests that aspects of the organization and function of the skull in Anatala vis may have differed considerably from that in living water fowl. In the shoulder girdle, the peculiar structure of the furcula and the pneumatic foramen in the dorsal surface of the sternal end of the coracoid are considered to be derived characters uniting Anatalavis and Anseranas in the family Anseranatidae. The two otherwise have very little else in common that is not generally present in most of the rest of the order. The proportionately short and very robust, twisted humems of Anatalavis is unique in the order and bespeaks a different mode of flight that probably was very strong and rapid. Many extant waterfowl are strong, fast fliers without having such a robust humems, however. The overall proportions of the hu mems are more like those of a falcon {Falco, Falconidae), al though why a filter feeder would need such a wing is not easily envisioned. The pelvis of Anatalavis is likewise peculiar for an anseri form in the short, expanded anterior portions of the ilia, where as in all other waterfowl, including the Anhimidae, the preace- tabular part of the pelvis is longer and narrower. From its resemblance to such wading birds as ibises (Plataleidae) and other Ciconiiformes, the pelvis of Anatalavis is presumably primitive within the order. The innominate bones, however, are fused to the sacrum, which is a more derived condition also found in Ciconiiformes and other Anseriformes except Presby ornis. In the lack of fusion of the innominates, Presbyornis more nearly resembles the Charadriiformes. With the recognition of Anatalavis as a member of the Anseranatidae, we can trace each of the three major modem lineages of Anseriformes back to the early Eocene, or earlier in the case of Anatalavis. According to the phylogeny developed by Ericson (1997), the Presbyornithidae, which were probably world-wide in distribution in the early Tertiary, are on the lin eage leading to the Anatidae proper. As yet unpublished early NUMBER 89 243 Eocene records of the Anhimidae from Wyoming and England establish that this group was in existence at the same time and occurred outside of South America, the modem members of the family evidently being highly derived relicts. Anseranas in Australia likewise now appears to be a rather specialized relict of a once more diverse family Anseranatidae. Although the Anatidae proper probably existed in the Paleo gene, they do not appear with any certainty or regularity in the Northern Hemisphere until the Miocene, from which it has been presumed that the family probably originated in the Southern Hemisphere (Olson, 1989). Possibly the Anseranatidae was the more diverse family in the Northern Hemisphere in the Paleo gene, and the possible affinities of such taxa as the late Eocene Romainvillia Lebedinsky (1927), from France, and Cygnopter- us Lambrecht (1931), from the early Oligocene of Belgium, with the Anseranatidae should be investigated. Literature Cited Andors, Allison Victor 1988. Giant Groundbirds of North America (Aves, Diatrymidae). xvi+ 577 pages. Doctoral dissertation, Columbia University, New York. [Available from University Microfilms International, Ann Arbor, Michigan, order no. DA8815650.] 1992. Reappraisal of the Eocene Groundbird Diatryma (Aves: Anserimor- phae). In Kenneth E. Campbell, Jr., editor, Papers in Avian Paleon tology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:109-125. Ericson, Per 1996. The Skeletal Evidence for a Sister-Group Relationship of Anseri form and Galliform Birds?A Critical Evaluation. Journal of Avian Biology, 27:195-202. 1997. Systematic Relationships of the Paleogene Family Presbyornithidae (Aves: Anseriformes). Zoological Journal of the Linnean Society, London,\2\A29-=facies articularis clav icularis.) the Presbyornithidae); the scar for M. scapulohumeralis crania lis forming a shallow to moderately deep, elliptical depression (a narrow, deep scar in the Presbyornithidae); the fossa pneu- motricipitalis ventralis small and perhaps pneumatic (a large, deep, nonpneumatic excavation in the Presbyornithidae); the scars for Mm. latissimus dorsi posterioris and anterioris are di rected toward a point well distal of where the deltoid crest sets off from the shaft (directed toward where the crest sets off from the shaft in the Presbyornithidae) (Figure 4). In the distal end of the humems, Juncitarsus differs from the Presbyornithidae in having the attachment area of the anterior articular ligament of Howard (1929) much more elevated and lateromedially narrow; the processus flexorius not as ventrally Juncitarsus Presbyornis FIGURE 3.?Left scapula in dorsal view. Juncitarsus gracillimus: USNM 468466. Presbyornis pervetus: UCMP 126193 (in mirror image). (a-craniodorsal margin.) protmding; probably a well-developed sulcus for M. scapulo humeralis, which is lacking in the Presbyornithidae; and the two scars for M. flexor carpi ulnaris on the processus flexorius distinctly unequal in size (the posterior is considerably larger), whereas these scars are of about equal size in the Presbyorni thidae, or the anterior is the larger (Figure 5). ULNA.?I have found no distinctive differences in the proxi mal end, maybe due to the poor preservation of the only speci men of Juncitarsus available. In the distal end, the tuberculum carpale is short and directed cranially, not clearly craniodistally as in Presbyornis (proximodistally much longer and blunter in Telmabates). RADIALE.?Although generally very similar, Juncitarsus differs from Presbyornis (the only presbyornithid genus in which this bone is known) in lacking the deep excavation of the dorsal side and in having the incisure at the cranial side deeply cut and not as wide (Figure 6). CARPOMETACARPUS.?The only proximal end known of Juncitarsus has the cranial margin of trochlea carpalis (border ing the anterior carpal fossa) convex, not deeply concave as in the Presbyornithidae (Figure 7). No difference has been found in the distal end of the carpometacarpus. FEMUR.?The femur of Juncitarsus seems to be more robust than that in the Presbyornithidae. Furthermore, the femoral neck in Juncitarsus is broader in proximal view; the impres- siones iliotrochanterici show a very different pattern than in the Presbyornithidae; and the trochlea fibularis is concave, or flat, NUMBER 89 249 FIGURE 4.?Proximal end of left humerus in anconal view. Juncitarsus gracillimus: proximal end USNM 468466. Presbyornis pervetus: UCMP 126205; the shape of crista deltoidea, however, is based on USNM 492550 (in mirror image). (a=area below caput humeri, 6=scar for M. scapulohumeralis cranialis, c=fossa pneumotricipitalis ventralis, t/=scars for Mm. latissimus dorsi posterioris and anterioris.) Presbyornis in lateral view in Juncitarsus (most often convex in the Presby ornithidae) (Figure 8). TIBIOTARSUS.?A part of the distal, lateral condyle is the only adult specimen of Juncitarsus available for comparison. It differs from the Presbyornithidae in having the distal margin distinctly concave, not smoothly rounded. TARSOMETATARSUS.?The tarsometatarsus (Figure 9) is considerably more elongated and slender in Juncitarsus than in the Presbyornithidae; the intercotylar knob is much larger; the cotylae are deeper and narrower; the hypotarsus is lateromedi ally narrower and protmdes more caudally, with the crista lat eralis as large as the crista medialis (the crista medialis is very low in the Presbyornithidae); and the shaft is rectangular in Presbyornis FIGURE 5.?Distal end of right humerus of Presbyornis pervetus, UCMP 119399 (in mirror image), in anconal (left) and palmar (right) views. (a=attachment for the anterior articular ligament, 6=processus flexorius, c= sulcus M. scapulohumeralis, t/=scars for M. flexor carpi ulnaris.) both taxa but is laterally very compressed in Juncitarsus, not craniocaudally compressed as in Presbyornis (insufficiently known in Telmabates). The distal end essentially is very simi lar to the Presbyornithidae but is more laterally compressed; the foramen vasculare is as large as in the Presbyornithidae but is narrower mediolaterally, with its caudal opening situated slightly more proximad; and the scar for the hallux is situated more distally on the shaft. Presbyornis Juncitarsus FIGURE 6.?Right radiale in dorsal view (left) and in ventral view (right). Jun citarsus gracillimus: USNM 244333 (in mirror image). Presbyornis pervetus: USNM 492552. (a=excavation of dorsal side, 6=incisure.) 250 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Presbyornis Presbyornis FIGURE 7.?Left carpometacarpus of Presbyor nis pervetus, UCMP 126228, in dorsal view. (a=cranial margin of trochlea carpalis.) Juncitarsus o- FIGURE 8.?Right femur in external view. Juncitarsus gracillimus: USNM 468466 (in mirror image). Pres byornis pervetus: UCMP 126244. (a=trochlea fibularis.) Discussion Despite their close match in general skeletal morphologies, Juncitarsus does differ from the Presbyornithidae in several as pects that can be useful in the identification of fragmentary ma terial. Given that Juncitarsus is correctly referred to the Phoeni copteridae, it may not be particularly closely related to the Prebyomithidae (Ericson, 1997). Rather, the similarities be tween the two taxa should be interpreted as symplesiomorphies. Surely, this also is the explanation of the observed morphologi cal similarities between the Presbyornithidae and certain other early Tertiary birds, such as the Graculavidae (Olson and Parris, 1987). Many aspects of the skeletal morphology might thus be less useful in the reconstmction of the phylogenetic relation ships between these taxa. The similarities do show, however, that these long-legged forms perhaps share a not much older common ancestor, and that they are parts of a radiation of late Mesozoic "shorebirds," as postulated by Martin (1983). Martin (1983:320) suggested that the supposed radiation of shorebird-like forms could possibly include "the progenitors of the entire Tertiary radiation" of birds. This seems less likely, however, given that many landbirds, such as galliforms, strigi- forms, caprimulgiforms, and cuculiforms, had evolved by the early Tertiary, some already by the Paleocene (Olson, 1985). These birds exhibit skeletal morphologies quite different from the group of "shorebirds" to which Juncitarsus and the Presby ornithidae belong. If we allow for the time necessary to evolve these various adaptations, it seems justified to assume that a few different phylogenetic lineages leading to subclades of modern neognaths were established by the late Mesozoic. It would not be very surprising if these lineages eventually prove to correspond to Olson's (1985:84) tentative division of the modem birds into the palaeognathous birds, "basal land bird assemblage," "higher land bird assemblage," and "water bird assemblage." Literature Cited Ericson, Per G.P. 1997. Systematic Position of the Paleogene Family Presbyornithidae (Aves: Anseriformes). Zoological Journal of the Linnean Society (London), 121:429-^483. In prep. The Early Tertiary Presbyornithidae (Anseriformes)?Paleoecol ogy, Anatomy, and Systematic Revision. Howard, Hildegarde 1929. The Avifauna of Emeryville Shellmound. University of California Publications in Zoology, 32:301-394. Martin, Larry D. 1983. The Origin and Early Radiation of Birds. In A.H. Brush and G.A. Clark, Jr., editors, Perspectives in Ornithology, pages 291-338. Cambridge: Cambridge University Press. Olson, Storrs L. 1985. The Fossil Record of Birds. In D.S. Farner, J.R. King, and KC. Parkes, editors, Avian Biology, 8:79-238. New York: Academic Press. Olson, Storrs L., and Alan Feduccia 1980. Relationships and Evolution of Flamingos (Aves: Phoenicop teridae). Smithsonian Contributions to Zoology, 316: 73 pages. Olson, Storrs L., and David C. Parris 1987. The Cretaceous Birds of New Jersey. Smithsonian Contributions to Paleobiology, 63: 22 pages. Peters, D. Stefan 1987. Juncitarsus merkeli n. sp. stiitz die Ableitung der Flamingos von Regenpfeifervogeln (Aves: Charadriiformes: Phoenicopteridae). Courier Forschungsinstitut Senckenberg, 97:141-155. NUMBER 89 251 m Presbyornis FIGURE 9.?Left tarsometatarsus. Different aspects of Juncitarsus gracillimus (left, based on USNM 244318 (holotype), except top left figure, which is the mirror image of USNM 244322) shown with the corresponding aspects of Presbyornis pervetus (right, based on USNM 492551, in mirror image, for complete bone; UCMP 126173, in mirror image, for proximal view; UCMP 126177, in mirror image, for anterior view of proximal end; UCMP 126178 for anterior view of distal end; UCMP 126182 for distal view). (a=hypotarsus, 6=intercotylar knob, c=cotyla, Across section of shaft, e=foramen vasculare.) Presbyornis isoni and Other Late Paleocene Birds from North Dakota Richard D. Benson ABSTRACT Paleocene fossil birds from North Dakota in the collections of the Science Museum of Minnesota range in age from Tiffanian 3 to Tiffanian 4 and seem to represent five taxa. A humems is referred to the anseriform Presbyornis isoni Olson, previously known from a less complete humerus from the late Paleocene of Maryland. All known specimens of Dakotornis cooperi Erickson, referable to the extinct charadriiform form-family Graculavidae, are reviewed. A cervical vertebra and a tarsometatarsal fragment, both within the probable body-size range of Dakotornis cooperi but probably representing different taxa, are referred to the Gracu lavidae. Another distal end of a tarsometatarsus, from perhaps the smallest currently known Paleocene bird, also is referred to the Graculavidae. These two tarsometatarsi exhibit a mosaic of charadriiform characters. Together with the tarsometatarsus of Tel matomis priscus Marsh, three size classes of North American Paleocene graculavid tarsometatarsi are now known. Introduction A new Paleocene species of the fossil anseriform genus Presbyornis, P. isoni, was established by Olson (1994) on the basis of two bones, an incomplete humems and a manual pha lanx 1 of the major digit, discovered in Maryland. This is the largest known presbyornithid. The purpose of this paper is to describe an additional, more complete humems of P. isoni as well as other avian fossils from three late Paleocene sites in North Dakota: the Wannagan Creek Quarry, the Judson Locali ty, and the Brisbane Locality. The Wannagan Creek Quarry of western North Dakota oc curs in the Bullion Creek (formerly "Tongue River") Forma tion, consisting of riverine and lacustrine deposits. The Wanna gan Creek fossil flora and fauna indicate a subtropical swamp environment: the most abundant large vertebrate is the 15-ft (4.5-m) crocodile Leidyosuchus formidabilis Erickson; the flo- RichardD. Benson, J.F. Bell Museum, and 100 Ecology Building, Uni versity of Minnesota, St. Paul, Minnesota 55108, United States. ra is dominated by bald cypress {Taxodium olriki (Heer) Brown), fig {Ficus spp.), magnolia {Magnolia spp.), and sy camore {Platanus spp.) (Erickson, 1991). About 110 mi (175 km) east of the Wannagan Creek Quarry, the Judson Locality also occurs in the Bullion Creek Formation, in deltaic sedi ments (Holtzman, 1978). The Brisbane Locality occurs in the underlying Slope Formation (Kihm, 1993) near the contempo rary marine Cannonball Formation. The paleontology of these two near-shore localities indicates warm-temperate cedar swamps, the faunas of which also included crocodilians (Holtz man, 1978). Fossil footprints of probable shorebirds have been reported from another late Paleocene site (Locality L6421) near Wannagan Creek Quarry (Kihm and Hartman, 1995). Nomenclature for species' binomials and English names of modem birds follows Sibley and Monroe (1990). AGE AND CORRELATION.?The age of the avian fossils from Maryland reported by Olson (1994:429) is "near the base of the Upper Paleocene (Landenian), Aquia Formation, Piscataway Member... probably upper nannoplankton zone NP5, but pos sibly lower NP6. ...On the scale of Berggren et al. (1985), the age would be somewhere between 61 and 62 million years." This falls within the Tiffanian North American Land Mammal Age, the Landenian of Europe being more or less coterminous with the Tiffanian of North America (Berggren et al., 1985). The Wannagan Creek Quarry dates to early Tiffanian 4 {=Plesiadapis churchilli zone), within the earlier half of paleo- magnetic chron 25 Reversed (Sloan, 1987), whereas the Mary land locality that yielded the type specimens of P. isoni dates to the middle of chron 26 Reversed (Berggren et al., 1985), seem ingly to Tiffanian 1. The Wannagan Creek beds of North Dako ta would correlate, in the notation used in Olson (1994), to nan noplankton zone NP7 or NP8, with an age of about 60 million years according to the scales of both Berggren et al. (1985) and Meehan and Martin (1994). The Judson Locality is within the later half of paleomagnetic chron 26 Normal, in early Tiffanian 4 (Kihm, 1993, and pers. comm., 1996), probably less than 100,000 years earlier than Wannagan Creek. The Brisbane Lo cality dates to early Tiffanian 3A {=Plesiadapis rex zone: Neo- 253 254 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY plagiaulax hunteri subzone) at the late part of chron 26 Re versed (Sloan, 1987), with an age of about 61 million years. ACKNOWLEDGMENTS.?I thank B.R. Erickson, Science Mu seum of Minnesota (SMM), St. Paul, for the loan of the Pale ocene avian specimens reported in this paper and for discus sions concerning Wannagan Creek Quarry; P.G.P. Ericson and S.L. Olson for sharing their knowledge of Presbyornithidae and Paleocene birds; and Julie Martinez, Science Museum of Min nesota, for drawing the figures. Comparative material used in this study was that of the avian skeletal collection of the Bell Museum of Natural History, University of Minnesota, and the National Museum of Natural History (USNM, which houses the collections of the former United States National Museum), Smithsonian Institution. This study was partially supported by funds (for illustrations) from the Department of Paleontology of the Science Museum of Minnesota. Systematic Paleontology Class AVES Order ANSERIFORMES Family PRESBYORNITHIDAE Presbyornis isoni Olson, 1994 FIGURE 1 REFERRED MATERIAL.?A badly crushed but mostly com plete right humems, lacking the external tuberosity, the proxi mal half of the pectoral crest, and the central portion of the bi cipital area; SMM P96.9.2; collected by Michael P. Ryan, 28 June 1989. For measurements, see Table 1. LOCALITY.?North Dakota, Billings County, -10 mi (16 km) NW of Medora; Wannagan Creek Quarry, field map quadrant P-6. HORIZON AND AGE.?Late Paleocene, early Tiffanian 4, Bullion Creek Formation, Wannagan Creek Quarry, Bed 2 (lig- nitic shale); absolute age, -60 Ma. COMPARISONS.?Specimen SMM P96.9.2 agrees in all char acters with the holotype of Presbyornis isoni (USNM 294116), the distal two-thirds of a humems from the Aquia Formation of Maryland, and confirms that P. isoni is an equally good Pres byornis at both ends of its humerus by comparison with the TABLE 1.?Measurements (in mm) of the two known humeri of Presbyornis isoni Measurement Length from head to internal condyle Distal width Depth through external (radial) condyle Greatest diameter of brachial depression SMM P96.9.2 194.9 -27 13.5 10.0 USNM 294116 (from Olson, 1994) 23.3 12.9 8.8 type species Presbyornis pervetus Wetmore of the early Eocene. The proximal end of the humems of P. isoni agrees with that of the smaller-bodied P. pervetus in the following characters: the head is undercut by a deep, arc-shaped exten sion of the capital groove; the pectoral crest is gently curved and long, about twice the length of the widely curved bicipital crest; the attachment of M. scapulohumeralis posterior is a wide, but elongated, kidney-shaped pit in the distal rim of the bicipital crest; the median crest is continuous with the extreme ly prominent central ridge of the shaft; and the attachment of M. latissimus dorsi posterior is a prominent oval structure very close to the central ridge. The proximal end of the humems of P. isoni differs from that of P. pervetus in lacking a prominent muscle-scar line distal to the attachment of M. latissimus dorsi posterior. In the South American presbyornithid Telmabates antiquus Howard, the humeral head and internal tuberosity are both undercut by an extension of the capital groove (P.G.P. Ericson, pers. comm., 1996), whereas in the North Dakota specimen of Presbyornis isoni, only the head is undercut, as is the case in Presbyornis pervetus; this is further indication that P. isoni was correctly assigned to genus. As in P. pervetus, the humems of P. isoni lacks the stoutness typical of most of the modem anseriforms, although without being very slender. The length/distal-width ratio of the humer us of P. isoni is about 7.2, in contrast to 7.6 for the slightly more gracile P. pervetus, 6.5 for the stout-winged Snow Goose {Anser caerulescens), and 9.4 for the slender-winged Northern Gannet {Morus bassanus). The ratios for Presbyornis spp. are in general agreement with those of charadriiforms: American Oystercatcher {Haematopus palliatus), 6.7; Double-striped Thick-knee {Burhinus bistriatus), 6.8; American Avocet {Re- curvirostra americana), 7.0; Marbled Godwit {Limosa fedoa), 7.1; Black Skimmer {Rynchops niger), 7.2; South Polar Skua {Catharacta maccormicki), 7.3; Franklin's Gull {Larus pipix- can), 7.5; and Common Murre {Uria aalge), 7.7. Similar values obtain for primitive ducks, such as the Fulvous Whistling-duck {Dendocygna bicolor), 7.2. DISCUSSION.?The Presbyornithidae are primitive anseri forms that share numerous character states with charadriiform birds. For example, the proximal end of the humems of Presby ornis is shallow, as in most charadriiforms and unlike modem anseriforms. Postcranially, presbyomithids resemble the "tran sitional charadriiform" Graculavidae, a group known from the Late Cretaceous and the Paleocene of North America, from the Paleocene of Sweden ("Scaniornithidae"; Olson and Parris, 1987) and France (Mourer-Chauvire, 1994), and apparently from the earliest Eocene of Australia (Boles et al., 1994). Re cently, one genus formerly classified as graculavid, Anatalavis, has been shown to be anseriform (Olson, this volume). Similar ly, Wetmore (1926), in describing Presbyornis pervetus, placed the new family Presbyornithidae in the same suborder as the Recurvirostridae. Feduccia and McGrew (1974) repeatedly called P. pervetus "the Green River flamingo," although they noted its duck-like appearance. The Recurvirostridae (avocets), NUMBER 89 255 FIGURE 1.?Referred right humerus of Presbyornis isoni, SMM P96.9.2 (xl): a, anconal view; b, palmar view. which are flamingo-like charadriiforms (Olson, 1985), have the (Eocene of Europe; Olson, 1985; Unwin, 1993). Presbyornis is oldest known fossil record of the extant charadriiform families the only presbyomithid genus yet recognized in North Ameri- 256 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ca, although Olson (1994) noted that the large-bodied P. isoni might well be assignable to a different genus were more of its skeleton known. The individual represented by SMM P96.9.2 was slightly larger than the one represented by the holotype of Presbyornis isoni (see Table 1). Their relative difference in size is not greater than that often seen within avian species (pers. obs.), including Presbyornis pervetus (P.G.P. Ericson, pers. comm., 1996). Order CHARADRIIFORMES Form-Family GRACULAVIDAE Dakotornis cooperi Erickson, 1975 FIGURE 2 MATERIAL.?Holotype: A complete right humems, SMM P74.24.106 (figured in Erickson, 1975), collected by Bmce R. Erickson and field crew, July 1974. Referred Material: A complete right humems, SMM P75.22.7 (Figure 2), collected by Bmce R. Erickson, summer 1975; a mostly complete left ti biotarsus, SMM P75.22.25 (not figured), collected by Tim Mc- Cutcheon, 6 July 1975. LOCALITY.?North Dakota, Billings County, -10 mi (16 km) NW of Medora, Wannagan Creek Quarry. Field map quad rants: holotype humems, G-5; other humems, G-2; tibiotarsus, K-6 (quadrants are 5 ft (1.5 m) on a side). HORIZON AND AGE.?Late Paleocene, early Tiffanian 4, Bullion Creek Formation, Wannagan Creek Quarry, Bed 2 (lig- nitic shale); absolute age, -60 Ma. DISCUSSION.?Dakotornis cooperi Erickson is the only bird from Wannagan Creek to have been previously published. It was originally described as a "primitive ibis-like bird" repre senting the extinct family Dakotomithidae within the suborder Plataleae, although its resemblance to thick-knees was men tioned (Erickson, 1975). Although the humems of Dakotornis shares numerous characters with the Plataleidae (ibises), espe cially in its stoutness and general outline, Dakotornis more closely resembles the Graculavidae, as Olson and Parris (1987) pointed out, and even some of the modem charadriiforms such as Burhinidae (thick-knees), Haematopodidae (oystercatchers), Recurvirostridae (avocets), Laridae (gulls), and Scolopacidae (godwits, phalaropes, etc.). These charadriiforms resemble Da kotornis in having a shallow proximal end of the humems that is not pneumatic, a median crest at a nearly right angle to the shaft, and a well-developed central ridge. In all of these charac ters these modem charadriiforms and Dakotornis differ from ibises. In two other characters of the humems?the proximally produced external tuberosity and the lack of a typically charadriiform ectepicondylar spur?Dakotornis is as similar to thick-knees as to ibises. The rounded ectepicondylar promi nence of Dakotornis closely resembles that of Burhinus, whereas in other charadriiforms a spur extends from the proxi mal rim of this prominence. Dakotornis would thus seem to be -^? :'/;?? 4 FIGURE 2.?Referred right humerus of Dakotornis cooperi, SMM P75.22.7 (x 1), anconal view. a good charadriiform. Peters (1983) and Olson (1985) noted the similarities between ibises and charadriiforms (and gruiforms). The two known humeri of Da kotornis cooperi, holotype and re ferred, discovered in different field seasons, have until now been separated from each other through loans to different persons; I am the first person to see both origi nal specimens side by side. The referred humerus (Figure 2) agrees in all characters with the holotype. The only apparent dif ference is that the external tuber osity of the holotype (as figured in Erickson, 1975) appears much narrower and sharper than the broad, rounded external tuberosity of the referred humems. The ap parent narrowness of the tuberosi ty in the holotype, however, is due only to breakage. These two right humeri of Dakotornis cooperi are of nearly identical size; the holo type is 87.1 mm long, and the oth er is 89.0 mm long. The tibiotarsus (SMM P75.22.25) was tentatively assigned to Dakotornis cooperi on the basis of size and provenance. Olson and Parris (1987) have noted the graculavid nature of this spec- GRACULAVIDAE, gen. et sp. probabiliter indescript. MATERIAL.?A mostly complete third cervical vertebra, 18.5 mm long, -15 mm wide (both anteriorly and posteriorly), SMM P77.7.159 (not figured); collected by Richard C. Holtz- man, summer 1977. LOCALITY.?North Dakota, Morton County, -8 mi (13 km) S of Judson; Judson Locality. HORIZON AND AGE.?Late Paleocene, early Tiffanian 4, Bullion Creek Formation, Judson Locality; absolute age, -60 Ma. COMPARISONS.?This specimen is referred to the Gracu lavidae due to its Paleocene age and its close resemblance to modem charadriiform cervical vertebrae. This cervical vertebra is identified as the third on the grounds of its caudally oriented, long-bottomed cariniform hypapophysis, gracile neural spine, and reduced pleurapophyses. The third cervical is typically a short vertebra, unlike this specimen, which is 1.2 times longer than its width, although this degree of elongation is common among charadriiforms. In its general outline and proportions between its parts, the specimen most closely resembles the cer- NUMBER 89 257 vical vertebrae of the Charadriiformes. The shapes of the facets of the zygapophyses, of the articulations of the centmm, of the ventral pit anterior to the keel, and of other features are like those of most Charadriiformes but are more like recurvirostrids than like burhinids. The unusual feature of its relatively narrow posterior width (which is not appreciably greater than its ante rior width) also is found in oystercatchers (Haematopodidae). Another striking feature is the transverse perforation of the centrum. This condition is observed in some of the more caudad cervical vertebrae (but not necessarily in the craniad vertebrae, such as the third) in avocets (Recurvirostridae), gulls (Laridae), skimmers (Rynchopidae), and sandpipers (Scolo- pacidae). Transverse perforation of the centmm is fairly com mon in Anseriformes, and in Dendrocygna bicolor even the third cervical is perforated; however, the specimen does not otherwise appear anseriform. DISCUSSION.?Although this specimen most likely repre sents an unknown species, the size of the vertebra would seem to fall within the upper limits of probable size for Dakotornis cooperi, even if it is somewhat large for a charadriiform body plan. A possible model for the body proportions of the very stout-winged Dakotornis might be the Canvasback {Aythya valisineria), in which the absolute and relative sizes of the hu mems, vertebrae, tibiotarsus, and tarsometatarsus are very sim ilar to those of the humems of Dakotornis, the present vertebra, the Wannagan Creek tibiotarsus, and the tarsometatarsus de scribed below. FIGURE 3.?Distal end of left tarsometatarsus of Graculavidae gen. et sp. prob. indescript., SMM P77.8.210 (x3): a, anterior view; b, medial view; c, posterior view; d, distal view. GRACULAVIDAE, gen. et sp. probabiliter indescript. FIGURE 3 MATERIAL.?Distal 20 mm of a left tarsometatarsus, lacking most of the outer trochlea, SMM P77.8.210 (Figure 3); collect ed by Richard C. Holtzman, summer 1977. LOCALITY.?North Dakota, Grant County, -5 mi (8 km) W of Raleigh; Brisbane Locality. HORIZON AND AGE.?Late Paleocene, early Tiffanian 3A, Slope Formation, Brisbane Locality; absolute age, -61 Ma. COMPARISONS.?This specimen is referred to the Gracu lavidae on the basis of its general resemblance to the tar sometatarsus of charadriiforms, including that of Telmatomis priscus, the only North American graculavid for which the tar sometatarsus was previously known (although T. priscus was a much smaller bird). The posterior surface of the specimen dis plays what might be called a typical "charadriiform basin," i.e., a deep, fairly symmetrical depression bounded by the longitu dinal ridges of the inner and outer trochleae and by the proxi mal rim of the middle trochlea. The specimen resembles Tel matomis in the following: the distal foramen is moderately large and oval, the metatarsal facet is well developed, the inner trochlea is oriented distomedially, and the inner trochlea is ele vated so that its distalmost extent is just proximal to the half- height of the middle trochlea's digital groove. The inner and outer trochleae, however, are considerably more posteriorly re tracted than in Telmatomis, so that the tarsometatarsus has a greater arch in distal view. The present specimen differs from the Presbyornithidae in having the metatarsal facet oval and distinct rather than long and weak, a relatively smaller distal foramen, and a lesser trochlear arch in distal view. Characters of SMM P77.8.210 resemble those of a number of extant charadriiform families in a mosaic manner. Olson and Parris (1987) noted skua-like (and presbyomithid-like) features in the tibiotarsus of another Paleocene graculavid, Laornis ed- vardsianus Marsh. SMM P77.8.210 combines tarsometatarsal characters especially of gulls (Laridae), skuas (Stercorariidae), and oystercatchers (Haematopodidae), among others. The gen eral outline of the bone is oystercatcher-like, except that the orientation of the inner trochlea is gull-like, being neither as distally oriented as in oystercatchers nor as medially oriented as in skuas. Its moderately large distal foramen matches that of oystercatchers, unlike the larger foramen in gulls and skuas. The metatarsal facet is skua-like (as seen in Catharacta mac cormicki) in both configuration and position (rather than lack ing a hind toe as in oystercatchers). The degree of trochlear arch in distal view matches that of gulls or skuas, unlike the very highly arched tarsometatarsus of oystercatchers. The low er elevation of the inner trochlea, unlike that in most of the modem charadriiforms, is similar to that seen in Telmatomis, thick-knees (Burhinidae), or skimmers (Rynchopidae). 258 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY DISCUSSION.?Considering that the distal end of this tar sometatarsus seems a strange mosaic, had the proximal end been preserved, it might well have resembled that of some oth er bird altogether, not to mention what the rest of the skeleton may have resembled. Because the preserved part of this bone is, however, a mosaic of exclusively charadriiform characters, the Graculavidae is the best present assignment for this speci men. This fragment would seem, at least in the absence of other Paleocene fossils of more modem aspect, to represent a bird older than the divergence time of any of the extant charadrii form families, as does Dakotornis. On body size, the present specimen could be referable to either of the monotypic genera Dakotornis or Graculavus. The specimens of Dakotornis coo peri are about one million years younger than this specimen, and those of Graculavus velox Marsh are about four million years older (see Olson, 1994, for the probable early Paleocene age of G. velox and the other New Jersey graculavids). GRACULAVIDAE, gen. et sp. indescript. FIGURE 4 MATERIAL.?Distal 6 mm of a right tarsometatarsus, with the outer trochlea broken away, SMM P96.9.3; collected by Bmce R. Erickson and field crew, summer 1976. LOCALITY.?North Dakota, Billings County, -10 mi (16 km) NW of Medora; Wannagan Creek Quarry, field map quadrant 0-19. HORIZON AND AGE.?Late Paleocene, early Tiffanian 4, Bullion Creek Formation, Wannagan Creek Quarry, Bed 2 (lig- nitic shale); absolute age, -60 Ma. DESCRIPTION.?This tarsometatarsal fragment is from a bird about the size of a Wilson's Plover {Charadrius wilsonia) and has the following characters: (1) the distal foramen is moder ately large and oval and occupies the distal end of a shallow an terior tendinal groove; (2) in posterior view, two tiny foramina occur just proximal to the distal foramen; (3) the metatarsal facet is shallow and oval; (4) the trochlear arch in distal view is relatively low, with the outer trochlea not posteriorly retracted, but with the inner trochlea retracted so that its central point is posterior to the posteriormost extent of the middle trochlea; the trochlear arch in posterior view (as is common in Charadrii formes) forms a subsymmetrical basin between the three tro chleae; (5) the inner trochlea is bulbous (as in most Charadrii formes) and is oriented distomedially; (6) the inner trochlea in distal view bears a very slight, posteriorly oriented wing on its medial side (a smaller wing than in modem Charadriiformes); (7) the inner trochlea is elevated so that its distal extent is about level with the proximal extent of the middle trochlea's digital groove; (8) the middle trochlea extends considerably farther distally than the other two; its lateral rim is slightly greater in distal extent and in anterior extent than is the medial rim; and (9) the outer trochlea would seem to have a slightly greater dis tal extent than the inner trochlea (the broken-off outer trochlea is preserved, but its contact with the adjacent part of the bone at the breakage is lost). FIGURE 4.?Distal end of right tarsometatarsus of Graculavidae gen. et sp. indescript., SMM P96.9.3 (x8): a, anterior view; b, medial view; c, posterior view; d, distal view. COMPARISONS.?Fossil tarsometatarsi of small, spurless gal- liforms can be easily mistaken for those of charadriiforms (Ol son and Farrand, 1974). Indeed, none of the nine characters listed in the previous paragraph would necessarily be inconsis tent with an assignment of the specimen to Galliformes. Olson and Farrand (1974), however, list 10 other tarsometatarsal characters, all relating to the trochleae, that distinguish galli- forms from charadriiforms. The present specimen differs from the Galliformes in all 10 of these trochlear characters. Specimen SMM P96.9.3 shares seven of the nine characters listed above (1-5, 8, 9) with the plover family, Charadriidae, a family not known in the fossil record earlier than the early Mi ocene (Unwin, 1993). The specimen differs from Charadriidae in two characters of the inner trochlea (6, 7), in which it more closely resembles the Paleocene Telmatomis and modem av- ocets (Recurvirostridae), respectively. It also resembles the Re curvirostridae in characters 3 and 9. DISCUSSION.?There is no evidence that the families Charadriidae and Recurvirostridae had diverged prior to the date of this fossil (Unwin, 1993), which may represent part of the "graculavid" stock prior to the divergence of these families. Of course, the presence of primitive charadriiform characters in this specimen (and in the Brisbane tarsometatarsus described above), perhaps by retention from much earlier forms, can tell us nothing certain about the divergence times of the extant charadriiform families. The very small body size of the bird represented by specimen P96.9.3, perhaps the smallest currently known Paleocene bird, need not exclude it from the Graculavidae (which are otherwise much larger birds), as this is only a form-family in any case NUMBER 89 259 (Olson and Parris, 1987). The relative difference in size be tween Graculavus velox and the species represented by P96.9.3 is comparable to that between the largest and smallest modem members of the sandpiper family, Scolopacidae. Literature Cited Berggren, William A., Dennis V. Kent, John J. Flynn, and John A. Van Cou- vering 1985. Cenozoic Geochronology. Bulletin of the Geological Society of America, 96:1407-1418. Boles, Walter E., Henk Godthelp, Suzanne Hand, and Michael Archer 1994. Earliest Australian Non-marine Bird Assemblage from the Early Eocene Tingamurra Local Fauna, Murgon, Southeastern Queens land. Alcheringa. 18:70. Erickson, Bruce R. 1975. Dakotornis cooperi, a New Paleocene Bird from North Dakota. Sci entific Publications of the Science Museum of Minnesota, 3(1): 7 pages. 1991. Flora and Fauna of the Wannagan Creek Quarry: Late Paleocene of North America. Scientific Publications of the Science Museum of Minnesota, 7(3): 19 pages. Feduccia, Alan, and Paul O. McGrew 1974. A Flamingolike Wader from the Eocene of Wyoming. Contributions to Geology of the University of Wyoming, 13:49-61. Holtzman, Richard C. 1978. Late Paleocene Mammals of the Tongue River Formation, Western North Dakota. Report of Investigations, North Dakota Geological Survey, 65: ix+88 pages. Kihm, Allen J. 1993. Late Paleocene Mammalian Biochronology of the Fort Union Group in North Dakota: Plesiadapis (Plesiadapiformes) from the Brisbane, Judson, and Wannagan Creek Quarry Local Faunas. In A.J. Kihm and J.H. Hartman, editors, The Marshall Lambert Symposium, pages 26-27. Bismarck: North Dakota Geological Society. Kihm, Allen J., and Joseph H. Hartman 1995. Bird Tracks from the Late Paleocene of North Dakota. Proceedings of the North Dakota Academy of Science, 49:63. Meehan, T.J., and Larry D. Martin 1994. Iterative Community Evolution in the North American Cenozoic. Journal of Vertebrate Paleontology, 14:3 8 A. Mourer-Chauvire, Cecile 1994. A Large Owl from the Paleocene of France. Palaeontology, 37:339-348. Olson, Storrs L. 1985. The Fossil Record of Birds. In Donald S. Farmer and James R. King, editors, Avian Biology, 8:79-238. New York, Academic Press. 1994. A Giant Presbyornis (Aves: Anseriformes) and Other Birds from the Paleocene Aquia Formation of Maryland and Virginia. Proceedings of the Biological Society of Washington, 107:429-435. Olson, Storrs L., and John Farrand, Jr. 1974. Rhegminornis Restudied: A Tiny Miocene Turkey. Wilson Bulletin, 86:114-120. Olson, Storrs L., and David C. Parris 1987. The Cretaceous Birds of New Jersey. Smithsonian Contributions to Paleobiology, 63: iii+22 pages. Peters, Dieter Stefan 1983. Die "Schnepfenralle" Rhynchaeites messelensis Wittich 1898 ist ein Ibis. Journal fur Ornithologie, 124:1-27. Sibley, Charles G., and Burt L. Monroe, Jr. 1990. Distribution and Taxonomy of Birds of the World. 1111 pages. New Haven: Yale University Press. Sloan, Robert E. 1987. Paleocene and Latest Cretaceous Mammal Ages, Biozones, Magne- tozones, Rates of Sedimentation, and Evolution. Geological Society of America Special Paper, 209:165-200, 4 charts. Unwin, David M. 1993. Aves. In Michael J. Benton, editor, The Fossil Record 2, pages 717-737. London: Chapman and Hall. Wetmore, Alexander 1926. Fossil Birds from the Green River Deposits of Eastern Utah. Annals of the Carnegie Museum, 16:391-402, plates 36, 37. A New Species of Graculavus from the Cretaceous of Wyoming (Aves: Neornithes) Sylvia Hope ABSTRACT A new species of Graculavus from the Lance Formation, Wyo ming, extends the range of the genus from the Atlantic paleocoast- line to the near-shore of the Cretaceous inland sea. The type and referred species were nearly contemporaneous in the late Maas trichtian. The new species was a very large flying bird with the proximal end of the humems in the size range of the largest mod em gulls or geese. The systematic and biogeographic significance of Graculavus-like birds is discussed. Introduction Graculavus velox Marsh, 1872, was described as a cormo rant, but since then it has been diagnosed as a shorebird (Shufeldt, 1915) and characterized as "transitional," referring to intermediacy between charadriiforms and gruiforms (Olson, 1985:171). The genus has been monotypic since Olson and Parris (1987) synonymized G. pumilus Marsh, 1872, with Tel matomis priscus Marsh, 1870. The changing taxonomic treat ments reflect the difficulty of identifying isolated fragments of unknown birds. The discovery of a very large new species of Graculavus provides an opportunity to review the significance of this genus of early neomithine birds. The two known species of Gracula vus are approximately contemporaneous representatives of widely separated, near-shore environments in the Late Creta ceous. Both species are known only from the proximal end of the humems, which, interestingly, is very similar to that of the early anseriform Presbyornis. Because the phylogenetic posi tion of the Anseriformes remains a key problem in avian taxon omy (Ericson, 1996), the resemblance of Graculavus to both Presbyornis and the Charadriiformes takes on systematic inter est. This paper describes the new species and provides a brief Sylvia Hope, Department of Ornithology and Mammalogy, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118, United States. perspective on the characters and biogeography of Graculavus- like birds. METHODS CHARACTER ANALYSIS.?Polarity of osteologie characters was judged from comparison of a broad selection of neomi thine birds with other Ornithurae and with Enantiornithes (Hope, unpublished data). Because uncertainties surround the basal phylogeny of Neornithes, the definition and phylogenetic diagnosis of higher groupings within it is uncertain. Principal comparisons herein are to "waterbirds" as defined below. Char acters cited for waterbirds occur in most of them, and some characters occur elsewhere among Neornithes as isolated in stances but not as a concerted complex. At the genus and spe cies levels, diagnosis of fragments by synapomorphy is rarely possible, so identification at these levels is based on a unique combination of attributes. NAMES.?Higher taxon names are used in the sense of Wet more (1960), except that these names are used herein in the node-based sense (de Queiroz and Gauthier, 1992) to include fossil forms sharing a most recent common ancestor with the extant crown taxon. The name "Ciconiiformes" is avoided be cause of gross disparity in usage, both historical and recent. English vernacular names are used for groups of birds with equivocal systematic status. Such names do not necessarily im ply monophyly and are as follows: "waterbirds," refers to all "seabirds," "shorebirds," and the Anseriformes; "seabirds" re fers to the Procellariiformes, Pelecaniformes, penguins, loons, and grebes; "shorebirds" refers to the Charadriiformes, ibises and flamingos, storks, and herons. ACKNOWLEDGMENTS.?For the opportunity to collect in the field and to describe this and other specimens, I thank Malcolm C. McKenna, American Museum of Natural History, New York (AMNH). Charlotte Holton (AMNH) arranged the loan of fossil material. Storrs L. Olson, National Museum of Natural History, Smithsonian Institution, Washington, D.C, gave ac cess to Hornerstown specimens and other fossil material in his care. David C. Parris and William B. Gallagher, New Jersey 261 262 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY State Museum, Trenton; Richard K. Olsson, Rutgers Universi ty; and Laurel M. Bybell, United States Geological Survey, provided stratigraphical information about the Hornerstown Formation. I thank Luis Baptista, California Academy of Sci ences, San Francisco, for providing work space and extended access to specimens. I also thank Ned K. Johnson, Museum of Vertebrate Zoology, Berkeley, and George C. Barrowclough, AMNH, for use of comparative material of recent species. Luis M. Chiappe, AMNH, and John H. Ostrom, Peabody Museum, Yale University, New Haven, made it possible for me to study comparative non-neornithine material in their care. I thank Olga Titelbaum, San Francisco, for her invaluable translations from the Russian. Comments by Storrs L. Olson, David C. Par ris, and Per G.P. Ericson have greatly improved the manu script. Fossil Localities and Stratigraphy Graculavus velox was recovered from greensand marl of the Navesink or overlying Hornerstown Formation of New Jersey during the last century in the course of commercial mining for the marl. The greensands were formed along the quiet margin of what was then the Atlantic coastline during the middle Maastrichtian through the earliest Paleocene. The phosphatic sediments provided ideal conditions for preservation. Birds from the Hornerstown and Navesink were reported earlier by Marsh (1870, 1872), Shufeldt (1915), and by Olson and Parris (1987). The new specimen represents a larger species of Grac ulavus from the late Maastrichtian Lance Formation, Wyo ming. The first bird discoveries in this area resulted from early exploration of the dinosaur fields in the North American West. The University of California began newer expeditions about 1955, and the American Museum of Natural History has been collecting in the Lance since 1960. Birds from the Lance were reported by Marsh (1889:83, footnote; 1892) and by Brodkorb (1963). The depositional setting and history of exploration in the New Jersey marls is summarized by Olson and Parris (1987). The Navesink Formation is entirely Maastrichtian. The age of the birds from the overlying Hornerstown Formation, whether Cretaceous or early Tertiary, was long debated because of the complexity of the sedimentation patterns and the enigmatic composition of the basal Main Fossiliferous Layer (MFL). This very narrow, densely fossiliferous zone lies directly over the Navesink Formation at the Inversand marl pit in Gloucester County, New Jersey. The MFL in this area includes Maastrich tian macrofossils, with ammonites, mosasaurs, and Enchodus. Typically Paleocene foraminifera occur in lower Hornerstown Formation levels at other localities. Vertebrates have not been recovered in the Hornerstown immediately above the MFL, but a Danian (Paleocene) fauna is found approximately 3 m higher (Gallagher and Parris, 1985). Very recent studies interpret the MFL as a Cretaceous lag deposit infilled with a Paleocene ma trix, possibly by burrowing arthropods (Gallagher, 1993; Kennedy and Cobban, 1996; other studies cited in both). There is iridium elevation within the MFL but not in a sharply de fined high peak. Thus, the Cretaceous/Tertiary boundary prob ably is within the MFL but is blurred and attenuated by rework ing of the sediments (Gallagher, 1992). The Lance Formation in the southeastern comer of the Pow der River basin, Wyoming, consists of massive, loosely consol idated sandstones with lenses of lignites and lignitic shales throughout. The Lance lies conformably directly under the en tirely Paleocene Fort Union Formation. The Lance sediments were deposited near the western margin of the North American interior seaway during its final retreat in the latest Cretaceous (Maastrichtian). Plant remains indicate a humid, subtropical environment. An abundant vertebrate fauna including sharks, lizards, mammals, and birds is preserved in channel fill of the ancient, meandering, near-shore streams. Faunal correlation shows a late Maastrichtian age for the fossiliferous sediments. The indurated streambeds have survived erosion better than the surrounding terrain has, and they are exposed now as "blow outs," or elevated sandstone outcrops (Dorf, 1942; Estes, 1964; Clemens, 1960, 1963; Lillegraven and McKenna, 1986). Systematic Paleontology NEORNITHES GRACULAVIDAE TYPE GENUS.?Graculavus Marsh, 1872:363. REMARKS.?The name Graculavidae is used herein in the sense of Olson and Parris (1987), except that Olson (this vol ume) has since referred Anatalavis to Anseriformes. Graculavus Marsh, 1872 Limosavis Shufeldt, 1915:19. TYPE SPECIES.?Graculavus velox Marsh, 1872. INCLUDED SPECIES.?Graculavus velox, Graculavus augus- tus, new species. Graculavus augustus, new species FIGURE 1 HOLOTYPE.?AMNH 25223; proximal end of left humems. TYPE LOCALITY.?From near Lance Creek, Niobrara Coun ty, Wyoming, University of California Museum of Paleontolo gy Locality V-5711 (Bushy Tailed Blowout), on the southern rim of a large valley that empties into Lance Creek, grid coor dinates 23,860-23,350 on reconnaissance map of Clemens (1963). Collected by Malcolm C. McKenna and party, August, 1985. HORIZON.?Upper part of the Lance Formation (late Maas trichtian). MEASUREMENTS.?Maximum depth of articular head, crani al to caudal, 6.8 mm; width of shaft through dorsal tubercle and NUMBER 89 263 10mm FIGURE 1.?Proximal end of the left humerus of Graculavus augustus, new species (holotype, AMNH 25223): a, cranial view; b, caudal view; c, medial view. base of ventral tubercle, 28.5 mm; internal width of tricipital fossa through base of ventral tubercle and distal border of im pression for M. scapulohumeralis caudalis, 9.5 mm; distance from capital incisure to dorsal tubercle, 17.3 mm ETYMOLOGY.?From the Latin augustus, majestic, for the large size of the bird, as well as the month of collection. DIAGNOSIS.?Derived characters of the Neornithes: moder ate enlargement of the articular head of the humems. Derived characters within the Neornithes: very large bicipital crest and prominence; thin, erect ventral tubercle; large dorsal tubercle; and well-defined caudal margin of the humems. The diagnosis of Graculavus is based on the differential di agnosis for Graculavus velox (Shufeldt, 1915; Olson and Par ris, 1987). Graculavus augustus is very similar to G. velox but is about one-third larger, and the area between the ventral and dorsal tubercles is relatively wider and flatter. DESCRIPTION.?The new specimen comes from a very large bird with the proximal end of the humems in the size range of the largest gulls or geese. Surface preservation is excellent. The tips of the dorsal and ventral tubercles are missing, but the shape of the remaining base of each is consistent with the mor phology of Graculavus velox. The pectoral crest is missing (as it is in G. velox). The bicipital crest is broken off just distal to the impression for M. scapulohumeralis caudalis. The shaft is broken off slightly distal to the tricipital fossa. The bone is delicately sculpted. The proximal end of the hu mems is very flat and broad. The articular head is small. On the cranial surface of the humerus, the bicipital prominence is large, slightly raised, and rounded. The sulcus for the trans verse ligament is deep and well defined but short, extending from the border of the bicipital crest only as far as the ventral tubercle. The impression for M. coracobrachialis cranialis is shallow and indistinct. In caudal view, the preserved base of the dorsal tubercle shows that it was moderately large and strongly protmdent from the shaft and was very far from the ar ticular head. The head of the humems does not overhang the capital incisure, which is deep and well defined. The caudal end of the incisure is excavated into a sulcus continuous with the deep sulcus undercutting the articular head. Distally this sulcus is bordered by a large, well-defined transverse scar ex tending from the base of the ventral tubercle diagonally toward the articular head. Evidently the dorsal tubercle also was deep ly undercut, but breakage obscures detail. The bicipital crest is broad and appears to have been rounded rather than sharply angular. The impression for M. scapulo humeralis caudalis is extremely large and well defined. The ventral tubercle is slender and erect, but breakage prevents see ing its total length. The tricipital fossa is very large and wide and is without a pneumatic foramen. There is a central tumes cence in the fossa, separating it into proximal and distal basins. The tumescence is the obverse side of the deep sulcus for the transverse ligament, visible because the cranial wall of the fos sa is very thin, and the fossa lacks the bony stmts and velum usually associated with pneumaticity. The more proximal of the two resulting basins is small in Graculavus. The surface of the fossa distal to the tumescence shows a se ries of narrow, shallow transverse ridges and sulci that appear to be impressions of parallel muscle fibers. The striations ter- 264 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY minate on a prominent, raised linear scar 7 mm long, extending from within the tricipital fossa near the base of the ventral tu bercle distally in the axis of the shaft. The scar may be the cen tral vane of a partly pinnate M. humerotriceps. Similar, al though shallower, striations occur in the tricipital fossa of the Western Gull, Larus occidentalis. A similar long scar is present in a slightly different position in several modem larids exam ined and in the Burhinidae. A short, lower crest or ridge ex tends from close to the proximal end of the long scar proximad and deeper into the triceps fossa. Dorsal to the longer scar there is a small, rough, irregular surface that may be the impression for M. scapulohumeralis cranialis. The caudal margin of the humems is very robust and distinct. It is far dorsal to the ventral tubercle. At the level of the base of the ventral tubercle the caudal margin bends abmptly dorsad to terminate at the proximal articular surface of the humerus about halfway between the apex and the dorsal tubercle. The area between the ventral tubercle and the caudal margin of the humems is broader than that in Graculavus velox. Discussion Cretaceous and Paleogene neornithines with the proximal end of the humems similar to that of Graculavus are Telmator- nis, Presbyornis, Telmabates, and Zhylgaia. Olson and Parris (1987) compared Graculavus most closely to the Burhinidae in the Charadriiformes. Telmatomis Marsh, 1870, has been re ferred to the Charadriiformes (Shufeldt, 1915; Cracraft, 1972; Olson and Parris, 1987) and is most similar to the Scolo- pacidae. Presbyornis Wetmore, 1926, is referable to the Anser iformes based on associated cranial material (Olson and Feduc cia, 1980). Telmabates Howard, 1955, evidently is referable to the Presbyornithidae (Feduccia and McGrew, 1974). Zhylgaia Nesov, 1988, was referred originally to the Presbyornithidae in Charadriiformes but later (Nesov, 1992) was referred, only ten tatively, to the Presbyornithidae. Zhylgaia has a very steeply angled ventral tubercle (almost 90? to the axis of the shaft), a distinct although shallow impression for M. coracobrachialis cranialis, the head not much undercut, the capital incisure broad and shallow, and the caudal margin of the humems well defined but not dorsal to the ventral tubercle. These conditions suggest that it does not belong with either the Anseriformes or the Charadriiformes. Thus, the humems in all of these early waterbirds is very similar, but the birds are not all referable to the same modem higher-level taxon. In general, the proximal end of the humems in these birds differs from that of most modem Charadriiformes as follows: 1. Among the articular head and associated structures, the head is smaller, and the dorsal tubercle is farther from the head and is smaller but more protmdent. 2. Impressions for ligaments and muscles on the cranial sur face are much less pronounced, including a shallow impression for M. coracobrachialis cranialis and a short sulcus for the transverse ligament. 3. Attachments of the muscles in and around the tricipital fossa are more robust, especially in Graculavus itself, but there is no dorsal (second) tricipital fossa. These differences suggest a distinctive flight mechanism. Graculavus, Telmatomis, Presbyornis, Telmabates, and Zhylgaia from the Late Cretaceous and the Paleogene are more similar to each other than to any modern bird. Many of the characters of Graculavus are not unique to the Charadriiformes but occur also among extant waterbirds in a mosaic pattern. The double-basined tricipital fossa occurs in virtually identical form in the Charadriiformes and the Oceanitidae (Procellarii- formes). Extreme elongation of the ventral tubercle, very deep ly undercut articular head, and very prominent caudal margin of the humerus are present in many Charadriiformes and in most Procellariiformes. A small articular head occurs in many Procellariiformes and in the Pelecaniformes. A small articular head, wide distance between the dorsal and ventral tubercles, and strong protmsion of the dorsal tubercle occur in some pe trels, especially Calonectris (Procellariiformes), in Phaethon, and in some other pelecaniforms (see comparisons in Olson, 1977, figs. 18-20). Judged on varied morphological and behav ioral grounds, Phaethon is highly plesiomorphic among Pele caniformes (Cracraft, 1985; Elzanowski, 1995). This mosaic pattern of sharing characters with Graculavus suggests that the various extant groups of waterbirds have re tained different suites of characters that were present in a com mon ancestor. The occurrence of many characters of Gracula vus in the overall very plesiomorphic Phaethon supports this suggestion. Conclusions The new, very large species extends the Cretaceous range of Graculavus from eastern to western North America. Presbyor- nithids are now reported from the Late Cretaceous through the early Eocene in North and South America, Antarctica, and Mongolia (Howard, 1955; Feduccia and McGrew, 1974; Ol son, 1994; Noriega and Tambussi, 1995). Zhylgaia comes from an estuarine habitat in the late Paleocene of Kazakhstan (Nes ov, 1988). Graculavus, Telmatomis, Zhylgaia, and the Presby ornithidae show that similar graculavid-like birds were wide spread in the Late Cretaceous through the early Tertiary. Reports of graculavid-like birds represented only by other parts of the skeleton are harder to evaluate in this context (e.g., Ol son and Parris, 1987; Kurochkin, 1988; Nesov and Jarkov, 1989; Elzanowski and Brett-Surman, 1995). Graculavus augustus was a very large bird and a strong fly er. The bones were delicately sculpted and were not highly pneumatic, resembling contours in flying swimmers and divers; they lacked the extreme inflation of soaring birds. Im pressions for tendons and muscles differ sufficiently from those of modem Charadriiformes to suggest distinctive flight mechanics. Graculavus is most similar to the Charadriiformes, but it shows a high proportion of characters that may be plesio- NUMBER 89 265 morphic among them. Whatever their subsidiary affiliations, the similarity of the humems among these Late Cretaceous and Paleogene waterbirds probably is due to recent common ances try. Literature Cited Brodkorb, P. 1963. Birds from the Upper Cretaceous of Wyoming. In Charles G. Sibley, editor, Proceedings of the XIII International Ornithological Con gress, Ithaca, 17-24 June, 1962, 1:55-70, 10 figures. Clemens, W.A. 1960. Stratigraphy of the Type Lance Formation. Report of the Twenty- First Session of the International Geological Congress, Norden, 1960, part 5:7-13. 1963. Fossil Mammals of the Type Lance Formation, Wyoming, Part I: In troduction and Multituberculata. University of California Publica tions in Geological Sciences, 48: 105 pages, 51 figures. Cracraft, J. 1972. A New Cretaceous Charadriiform Family. Auk, 89:36-46, 3 figures. 1985. Monophyly and Phylogenetic Relationships of the Pelecaniformes: A Numerical Cladistic Analysis. Auk, 102:834-853. de Queiroz, K., and J.A. Gauthier 1992. Phylogenetic Taxonomy. Annual Review of Ecology and Systemat ics, 23:449^180. Dorf, E. 1942. Flora of the Lance Formation at Its Type Locality, Niobrara County, Wyoming, 2: Upper Cretaceous Floras of the Rocky Mountain Re gion. Publications, Carnegie Institution of Washington, 508:83-144. Elzanowski, A. 1995. Cretaceous Birds and Avian Phylogeny. In D.S. Peters, editor, Acta Palaeomithologica: 3 Symposium SAPE: 5 Internationale Sencken- berg-Konferenz, 22-26 Juni 1992. Courier Forschungsinstitut Senckenberg, 181:37-53. Elzanowski, A., and M.K. Brett-Surman 1995. Avian Premaxilla and Tarsometatarsus from the Uppermost Creta ceous of Montana. Auk, 112:762-766. Ericson, P.G.P. 1996. The Skeletal Evidence for a Sister-Group Relationship of Anseri form and Galliform Birds?A Critical Evaluation. Journal of Avian Biology, 27(3): 195-202. Estes, Richard 1964. Fossil Vertebrates from the Late Cretaceous Lance Formation, East ern Wyoming. University of California Publications in Geological Sciences, 49: 180 pages, 73 figures, 5 plates. Feduccia, A., and P.O. McGrew 1974. A Flamingolike Wader from the Eocene of Wyoming. Contributions to Geology, University of Wyoming, 13(2):49-61, 13 figures. Gallagher, W.B. 1992. Geochemical Investigations of the Cretaceous/Tertiary Boundary in the Inversand Pit, Gloucester County, New Jersey. Bulletin of the New Jersey Academy of Science, 37:19-24. 1993. The Cretaceous/Tertiary Mass Extinction Event in the Northern At lantic Coastal Plain. The Mosasaur, 5:75-154. Philadelphia: Dela ware Valley Palaeontological Society. Gallagher, W.B., and D.C. Parris 1985. Biostratigraphic Succession Across the Cretaceous-Tertiary Bound ary at the Inversand Company Pit, Sewell, N.J. In R. Talkington, ed itor, Geological Investigations of the Coastal Plain of Southern New Jersey: Second Annual Meeting of the Geological Society of New Jersey, Pomona, New Jersey, 1985, pages C1-C16. Howard, H. 1955. A New Wading Bird from the Eocene of Patogonia. American Mu seum Novitates, 1710: 25 pages, 8 figures. Kennedy, W.J., and W.A. Cobban 1996. Maastrichtian Ammonites from the Hornerstown Formation in New Jersey. Journal of Paleontology, 70(5):798-804. Kurochkin, E.N. 1988. [Cretaceous Birds of Mongolia and Their Significance for Study of Phylogeny of Class Aves.] In E.N. Kurochkin, editor, Fossil Rep tiles and Birds of Mongolia. Transactions of the Joint Soviet-Mon golian Palaeontological Expedition, 34:33-41. [In Russian.] Lillegraven, J.A., and M.C. McKenna 1986. Fossil Mammals from the "Mesaverde" Formation (Late Creta ceous, Judithian) of the Bighorn and Wind River Basins, Wyo ming, with Definitions of Late Cretaceous North American Land- Mammal "Ages." American Museum Novitates, 2840:1-68, 13 fig ures. Marsh, O.C. 1870. Notice of Some Fossil Birds from the Cretaceous and Tertiary For mations of the United States. American Journal of Science, series 2, 49:205-217. 1872. Preliminary Description of Hesperornis regalis, with Notices of Four Other New Species of Cretaceous Birds. American Journal of Science, series 3, 3:360-365. 1889. Discovery of Cretaceous Mammalia. American Journal of Science, series 3, 38:81-92, 5 plates. 1892. Notes on Mesozoic Vertebrate Fossils. American Journal of Sci ence, series 3, 55:171-175, 3 plates. Nesov, L.A. 1988. [New Cretaceous and Paleogene Birds of Soviet Middle Asia and Kazakhstan and Their Environments]. Proceedings of the Zoologi cal Institute of Leningrad, 182:116?123, 2 plates. [In Russian, with English summary.] 1992. Mesozoic and Paleogene Birds of the USSR and Their Paleoenviro- ments. In K.E. Campbell, editor, Papers in Avian Paleontology Honoring Pierce Brodkorb; Proceedings of the II International Sym posium of the Society of Avian Paleontology and Evolution. Sci ence Series. Natural History Museum of Los Angeles County, 36:465^178. Nesov, L.A., and A.A. Jarkov 1989. [New Birds from the Mesozoic and Paleogene of the USSR and Some Remarks on the Phylogeny of Aves and Origin of Flight.] Proceedings of the Zoological Institute of Leningrad, 197:78-97, 2 plates. [In Russian, with English summary.] Noriega, J.I., and CP Tambussi 1995. A Late Cretaceous Presbyornithidae (Aves: Anseriformes) from Vega Island, Antarctic Peninsula: Paleobiogeographic Implications. Ameghiniana, 32(1):57?61. Olson, S.L. 1977. A Lower Eocene Frigate Bird from the Green River Formation of Wyoming (Pelecaniformes: Fregatidae). Smithsonian Contributions to Paleobiology, 35: 33 pages, 31 figures. 1985. The Fossil Record of Birds. In D.S. Farner, J.R. King, and K.C. Parkes, Avian Biology, 8:79-256, 11 figures. New York: Academic Press. 1994. A Giant Presbyornis (Aves: Anseriformes) and Other Birds from the 266 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Paleocene Aquia Formation of Maryland and Virginia. Proceedings of the Biological Society of Washington, 107(3):429-435. Olson, S.L., and A. Feduccia 1980. Presbyornis and the Origin of the Anseriformes (Aves: Charadrio- morphae). Smithsonian Contributions to Zoology, 323: 24 pages, 15 figures. Olson, S.L., and D.C. Parris 1987. The Cretaceous Birds of New Jersey. Smithsonian Contributions to Paleobiology, 63: 22 pages, 11 figures. Shufeldt, R.W. 1915. Fossil Birds in the Marsh Collection of Yale University. Transac tions of the Connecticut Academy of Arts and Sciences, 19:1-110, 15 plates. Wetmore, A. 1926. Fossil Birds from the Green River Deposits of Eastern Utah. Annals of the Carnegie Museum, 16(3^1):391^102, plates 36, 37. 1960. A Classification for the Birds of the World. Smithsonian Miscella neous Collections, 139( 11): 1 -37. Implications of the Cranial Morphology of Paleognaths for Avian Evolution Felix Y. Dzerzhinsky ABSTRACT In the early evolution of birds, bill formation produced a prob lem for muscular control of the thin, elongated upper jaw. In par ticular, it required a relatively high retracting force. Three sources of this force evolved. (1) A powerful M. retractor palatini (espe cially in Tinamiformes and Apteryx), originating primarily on the vomer and pterygoid, developed to provide direct muscular con nection between the dermal palate and the cranial base. It appar ently evolved due to a joining of the medial portions of the pterygoid and mandibular depressor muscles, which were aligned by development of the proc. mandibulae medialis (a character unique to birds). (2) The ancestral pseudotemporalis muscle devel oped into two portions, a large postorbital portion and an almost horizontally oriented intramandibular portion. Each portion seves to increase the retraction ability of the muscle as a whole. (3) The external mandibular adductor muscle developed, which, in neo- gnaths, is larger than either muscle previously mentioned. Its evo lutionary development was temporarily retarded by reduction of one of its places of origin?the upper temporal arch. Introduction For more than a century, paleognaths have been subjected to morphological studies in order to ascertain their apparently primitive nature and to discover their position in avian phylog eny (W.K. Parker, 1866; T.J. Parker, 1891; Pycraft, 1900; Mc Dowell, 1948; Hofer, 1945, 1950, 1955; de Beer, 1956; Webb, 1957; Muller, 1963; Bock, 1963; Cracraft, 1974; Yudin, 1970, 1978). I shall try to extract information on avian ancestry from the comparative and functional morphology of the feeding ap paratus in paleognaths. Nomenclature for species' binomials and English names of modem birds follows Sibley and Monroe (1990). ACKNOWLEDGMENTS.?In the process of this work I re ceived valuable assistance from the following individuals and Felix Y. Dzerzhinsky, Faculty of Biology, Moscow State University, Moscow 119899, Russia. institutions. F. Vuilleumier and A.V. Andors (American Muse um of Natural History, New York, New York) and E.G. Kordi- cova (Institute of Zoology, Kazakh Academy of Sciences, Al- maty, Kazakhstan) granted me access to the alcoholic specimen of Apteryx sp.; K.A. Yudin and V.M. Loskot (Zoological Insti tute, Russian Academy of Sciences, St. Petersburg, Russia) provided me with specimens of Eudromia and Casuarius and with several skulls of paleognaths, respectively. M.V. Bevol- skaya (Institute of Cattle-Breeding Askania-Nova, Ukraine) permitted me to dissect the alcoholic heads of Dromaius no- vaehollandiae (Latham), Rhea americana (Linnaeus), and Struthio camelus (Linnaeus). A. Elzanowski (National Muse um of Natural History, Smithsonian Institution, Washington, D.C.) and E.N. Kurochkin kindly assisted me in obtaining liter ature. A.N. Kuznetsov helped me in editing the manuscript and in translating it into English. S.C. Bennett, L.D. Martin, and S.L. Olson read the English version with fruitful criticism and helped me in editing it. I am sincerely grateful to all these per sons. This work was supported by The Cultural Initiative Foun dation, Moscow, and by The Russian Foundation of Basic Re searches (RFBR, grant N 96-04-50822). Skeleto-muscular Consequences of Bill Formation The adductory force of the mandible is transferred to the up per jaw through a food object. Resistance of the upper jaw to this force is produced (Figure 1) by a combination of tension on the ventral stalk (premaxillary and maxillary bones with palate caudally) and longitudinal compression of the dorsal stalk (frontal projection of premaxillary and premaxillary pro cesses of the nasal bones). The longer the jaw grew, the greater the forces became, and, due to jaw lightening, the stresses be came ever greater. The active forces necessary for normal grasping of food items must be supplied by muscles. The muscular force that creates tension in the palate and upper-bill floor also can ac complish ventral movement of the upper jaw by means of re- 267 268 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY compression FIGURE 1.?Lateral view depicting the combination of forces in the upper bill of a bird (Struthio sp.) that produce resistance against the push from the mandi ble when grasping food. Lateral view. F'=useful force applied to object, R=rectractory (active muscular) force, N=push transferred to the braincase via dorsal upper-bill stalk. traction (backward shift) of the palate. Therefore, it is called the retracting or retractory force. The ancestors of birds appar ently had no obvious source of such a force. Birds, however, have evolved the following three sources of retracting force. 1. M. retractor palatini (Figure 2) is the ventromedial part of the pterygoid muscle (of Moller, 1930, 1931), which is rather large in tinamous (Figure 2 A,B; Dzerzhinsky, 1983; Elzanows ki, 1987), ostriches (Figure 2 D,E), and especially in Apteryx (Figure 2 C). In paleognaths (sensu Pycraft, 1900) this muscle usually originates on the pterygoid and on the rear end of the vomer. Its caudal attachment is not situated at the midline on the base of the braincase, as in many neognaths (sensu Pycraft, 1900), but more laterally, near the caudal attachment of the oc- cipito-mandibular ligament, i.e., on the medial part of ala tym- FlGURE 2.?Retractor palatini muscle in ventral view: A,B, tinamou, Eudromia elegans; c, Kiwi, Apteryx sp. (scale=5 mm); D,E, Ostrich, Struthio camelus. aca=aponeurosis of insertion of M. pterygoideus caudalis; ad'=aponeurose of insertion of M. depressor mandibulae that is related to M. retractor palatini; art=aponeurosis of insertion of M. retractor palatini; Pt=pterygoid; Rtp=M. retractor palatini; Vbm=vomer (A,C,D, superficial layer; B,E, deeper layers) (A,B, after Dzerzhinsky, 1983.) NUMBER 89 269 C Psp Lep FIGURE 3.?Interrelationship of parts of the pseudotemporalis muscle and epipterygoid or its apparent remainder (lig. epipterygoideum): A,B, Tuatara, Sphenodon punctatus (Gray); C,D, tinamou, Eudromia elegans. Ept=epipterygoid; Lep=ligamentum epipterygoideum; Ls=laterosphenoid; Ps=undivided M. pseudotemporalis; Psp=M. pseudotemporalis profundus; Pss=M. pseudotemporalis superficialis; Q=quadrate. (A,B, after Dzerzhinsky and Yudin, 1979; C,D, after Dzerzhinsky, 1983.) panica. In adult Tinamiformes (e.g., Rhynchotus) these rela tions are obscured by later ossification, but they are quite clear in young Eudromia elegans Geoffroy Saint-Hilaire (Dzerzhin sky, 1983). Due to the occipito-mandibular ligament, the ptery goid muscle as a whole can act similarly to the retractor, but in contrast to it, via the mandible. 2. M. pseudotemporalis (part of the internal mandibular ad ductor) applies a retractory force to the mandible, and the force is transferred to the palate via the pterygoid muscle. In Spheno don and lizards (Figure 3A,B), M. pseudotemporalis is undivid ed and originates mainly from the epipterygoid. In birds, the epipterygoid would limit the mobility of the quadrate, so it either has been replaced by a flexible ligament, as in tinamous (Figure 3 C,D; Dzerzhinsky, 1983), or has been completely reduced. Consequently, the pseudot emporalis muscle has been divided into two portions, FIGURE 4 (right).?Contraction effect of the pseudotemporalis pro fundus muscle in the skull of the Common Raven, Corvus corax Linnaeus. APsp=immediate force; A=final force transferred to the braincase via the quadrate bone; Q=quadrate; Sq=squamosum. originating from two ends of the former epipterygoid. M. pseudotemporalis profundus originates on the tip of proc. or- bitalis quadrati, and M. pseudotemporalis superficialis origi nates on the front wall of the braincase. M. pseudotemporalis profundus produces a retracting force rather effectively, irre spective of the particular direction of its fibers, because the re sulting force is transferred to the braincase very caudally, through the quadrato-cranial joint (Figure 4). The main part of M. pseudotemporalis superficialis is well developed in paleog naths and occupies a considerable area of the temporal surface of the braincase (Figure 6A). But even here it passes rather 270 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY steeply, i.e., at a great angle to the jugal bar, and so produces a small retractory effect. It may, however, include (e.g., in Rhea) a very inclined portion, the so-called intramandibularis muscle (Figure 5B). In some birds another inclined portion of this muscle has evolved. The so-called caput absconditum (Hofer, 1950) appar ently is a derivative of the main part of M. pseudotemporalis superficialis that is situated in the posterior temporal fossa of a typically diapsid ancestor (Figure 5A; Dzerzhinsky and Yudin, 1979). It is found in Sphenisciformes, Procellariiformes (Fig ure 5B), and Pelecaniformes but never in recent paleognathous birds. 3. M. adductor mandibulae extemus (Figure 6) acting on the upper jaw via the pterygoid muscle is a very important retractor in most birds (e.g., cranes). In paleognaths, however, it shows a rather modest development and, except Apteryx, does not spread up over the temporal wall of the braincase. Thus, its ori gin is limited to the zygomatic process of the squamosal. This restriction might have resulted from a reduction of the main an cestral origin of the muscle, the upper jugal arch. Otherwise, it might be a result of the change of functional requirements in the muscle during the course of development of the long bill and the cranial kinesis. I presume that the immediate ancestors of birds had an aki netic skull that possessed some prerequisites of cranial kinesis, such as a loose basipterygoid articulation (Yudin, 1970). It seems likely that kinetic mobility appeared first in the most loaded zone, i.e., within the slender upper jaw (Figure 1), and thus resulted in an archaic rhynchokinesis (Yudin, 1970, 1978). One of the questions about the functional morphology of the avian skull is the influence of sharp strokes, such as are associ ated with pecking or with accidental strokes against hard sub strates while gathering grain or catching small, agile prey. In tinamous, the loose articulation of the frontal bone with the ad jacent parietal and laterosphenoid (Figure 7) is equivalent to the so-called "articulating frontoparietal joint" described by Houde (1981) in early Tertiary North American carinates. It does not allow significant rotary movements of any cranial part, so in my opinion it is not associated with ancient mesoki- netic mobility. Rather it is for damping shocks received along the dorsal bill stalk while pecking. The ventral stalk of the upper jaw was initially compliant, and it had to be supported by some solid framework able to transfer to the braincase large, but not dangerous, forces. This framework is formed by the bony palate, and among recent pa leognaths it is strongest in Apteryx (Figure 8A), doubtless due to its specialization for probing. In tinamous, Rhea (Figure 8B), and, apparently, recent Ca- suariiformes, the main trajectory of compression stresses mns from the palatine process of the premaxillary bone to the vomer, then to the pterygoid, the quadrate, and finally via the quadrate's otic process to the braincase (Dzerzhinsky, 1983). In ostriches (Figure 8C), where the palatal processes of the pre maxillary are missing and the vomer is partly reduced, com- Pss Psp aim aps aca FIGURE 5.?Comparison of the pseudotemporalis muscle in lateral view: A, lizard, Cyclura nubilis Gray; B, pro- cellariiform bird, Northern Fulmar, Fulmarus glacialis (Linnaeus) (mandible and side wall of braincase partly destroyed and removed). aca=aponeurotic insertion lobe of the caput absconditum of the pseudotemporalis superficialis muscle; Aex=M. adductor mandibulae extemus; aim= aponeurotical lobe of origin of the M. intra mandibularis; apm=aponeurosis of insertion of the pseudotemporalis profundus muscle; aps=aponeurosis of insertion of the pseudotemporalis superficialis muscle; Ca=caput absconditum of the pseudotemporalis superfi cialis muscle; Im=intramandibularis muscle, part of the pseudotemporalis superficialis muscle; Psp=pseudotemporalis profundus muscle; Pss=pseudotemporalis superficialis muscle; *=especially inclined part of the pseudotemporalis muscle. (A, after Iordansky, 1990; B, after Dzerzhinsky and Yudin, 1979.) NUMBER 89 271 FIGURE 6.?Lateral surface of jaw adductors in lateral view: A, Rhea, Rhea americana; B, White-naped Crane, Grus vipio. Aex=M. adductor mandibulae extemus; Ap=M. adductor mandibulae posterior; Dm=depressor mandibulae muscle; Pss=pseudotem-poralis superficialis muscle; Pvl=ventrolateral por tion of the pterygoid muscle. (B, after Kuular and Dzerzhinsky, 1994.) pression stresses mn from the bill tip through the premaxillary and maxillary bones to the palatine and then almost directly to the apex of the basipterygoid process. In the roof of the mouth, the ostrich possesses a broad gap that is closed only by skin. This patch of skin gains a gliding support from the parasphe- noidal rostmm via a long, thin, anterior extension of the vomer. Source of the Medial Mandibular Process The processus mandibulae medialis is highly specific for Aves. For example, in Gobipteryx, a fossil Mongolian bird, El zanowski (1974) regarded this process as a distinctly avian character. The functional properties of the muscular portion (ventromedial portion of the pterygoid muscle) inserting on its tip are influenced significantly by the particular position of the tip. It is placed extremely high in the sagittal plane, so corre sponding muscular forces pass almost through the pivot of the quadrato-mandibular joint (Figure 9A) and therefore apply a negligible adductory component to the mandible as compared to the retractory one. In the frontal plane, the tip of the process is extremely close to the midline, and therefore those muscular forces tend to rotate the caudal part of mandibular branch and so expand the lower jaw as a whole (Figure 9B; Yudin, 1961). The functional conditions discussed above, however, do not seem to account for the first steps in the evolution of the medial mandibular process. There is a peculiarity in the paleognath jaw musculature that is more useful in this respect: the above- mentioned M. retractor palatini. I suggest that this muscle may have arisen by a joining of two muscular units?the ventrome dial portion of the pterygoid muscle and the depressor mandib ulae muscle. Thus, their primitive interconnection via the cau dal portion of the mandible formed a two-link chain that foreshadowed the recent M. retractor palatini (Figure 10). The cmcial event in their further evolution has been an optimization of their ability to exert a single force, which has been ensured by alignment of both muscular links, due to the displacement Ls Fr FIGURE 7.?Lateral view of tinamou skull (Tinamiformes), showing the loose articulation of the frontal with adjacent bones. Fr=frontal; Ls=laterosphenoid; Pa=parietal. 272 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 8.?Comparison of palate structures in ventral view: A, Kiwi, Apteryx sp.; B, Rhea, Rhea americana; C, Ostrich, Struthio camelus. Probable paths for transferring longitudinal compression forces stippled. Bpt=basipterygoid process; Mx=maxillary; Pal=palatine; Pmx=premaxillary; Pt=pterygoid; Q=quadrate; Vbm=vomer. Scale=20 mm. of their interconnecting point by means of the formation and gradual elongation of the medial mandibular process. After the alignment, this hypothetical digastric muscular complex must have separated from the mandible. Apparently, the recent oc- cipito-mandibular ligament represents the reduced caudal belly of the digastric complex. Conclusion I would like to comment on the reinterpretation by McDow ell (1978) of the homologies in the avian upper jaw and palate. It is, of course, tempting to use the kinetic mobility in the skull as a cause of fragmentation of a huge, ancient pterygoid bone into two; however, many traits in the general arrangement of the bones (primarily palatine position relative to the choana, premaxillary, etc.) seem to be consistent with the traditional in terpretation that these two bones represent the reptilian palatine and pterygoid. The skull in ancient birds almost certainly had less internal mobility than it does in recent paleognaths, and such characters as the shape of the lateral rim of the palate or the pattern of epidermal papillae can hardly be valid. Finally, it is too difficult to accept McDowell's proposed loss of the max illary bone in birds and his consequent thesis that the maxillo- palatine of birds is equivalent to the reptilian palatine. Literature Cited Bock, W.J. 1963. The Cranial Evidence for Ratite Affinities. In Charles G. Sibley, ed itor, Proceedings of the XIII International Ornithological Congress, Ithaca 17-24 1962, 1:39-54. Buhler, P. 1985. On the Morphology of the Skull of Archaeopteryx. In M.K. Hecht, editor, The Beginnings of Birds: Proceedings of the International Archaeopteryx Conference, Eichstdtt, 1984, pages 135-140. Eich- statt: Freunde des Jura-Museums, Eichstatt. Cracraft, J. 1974. Phylogeny and Evolution of the Ratite Birds. Ibis, 116:494-521. de Beer, G.R. 1956. The Evolution of Ratites. Bulletin of the British Museum (Natural History), Zoology, 4(2):59-70. NUMBER 89 273 Dm Dm Pmm Pmm FIGURE 9.?Some functional effects of the medial mandibular process: A=skull of the Carrion Crow, Corvus corone comix Linnaeus, in sagittal sec tion, seen from left; B=skull of the Herring Gull, Larus argentatus Pontoppi- dan, with mandible broadened by contraction of the ventromedial portion of the pterygoid muscle, ventral view. F/>vm=force of the ventromedial portion of the pterygoid muscle; ^pvi+pdi=force of its lateral portions; Pmm=processus mandibulae medialis; Pvm=ventromedial portion of the pterygoid muscle. (B, after Yudin, 1961). Dzerzhinsky, F.Ya. 1983. On the Feeding Apparatus of Eudromia elegans; On the Question of Morphological Specificity of the Feeding Apparatus in Paleognaths. Trudy Zoologicheskogo Instituta Akademii Nauk SSSR [Proceedings of the Zoological Institute of the Academy of Sciences of USSR], 116:63-118, 4 figures. [In Russian.] Dzerzhinsky, F.Ya., and K.A. Yudin 1979. On the Homology of Jaw Muscles in Tuatara and Birds. Orni- tologiya (Moscow), 14:14-34, 7 figures. [In Russian; English edi- Pmm FIGURE 10.?Hypothetical stages in the formation of the proc. mandibulae medialis in birds, drawn on the model of the skull of Archeopteryx (Buhler, 1985). Dm=depressor mandibulae muscle; Pmm=processus mandibulae medi alis; Pvm=ventromedial portion of the pterygoid muscle; Rtp=M. retractor palatini. tion appeared in 1982 in Ornithological Studies in the USSR (Moscow), 2:408^136.] Elzanowski, A. 1974. Preliminary Note on the Palaeognathous Bird from the Upper Creta- cious of Mongolia. Palaeontologia Polonica, 30(5):103-109, 2 plates. [Results of the Polish-Mongolian Palaeontological Expedi tions.] 1987. Cranial and Eyelid Muscles and Ligaments of the Tinamous (Aves: Tinamiformes). Zoologische Jahrbucher, Abteilung fur die Anat omie und Ontogenie der Tiere, 116(1):63-118. Hofer, H. 1945. 1950. 1955. Houde, P. 1981. Untersuchungen iiber den Bau des Vogelschadels. Zoologische Jahrbucher, Abteilung fur die Anatomie und Ontogenie der Tiere, 69(1):1-158. Zur Morphologie der Kiefermuskulatur der Vogel. Zoologische Jahrbucher, Abteilung fur die Anatomie und Ontogenie der Tiere, 70:427-556. Neuere Untersuchungen zur Kopfmorphologie der Vogel. In A. Portmann and E. Sutter, editors, Acta XI Congressus Internationalis Ornithologici, Basel, 1954, pages 104-137. Basel: Birkhauser. Paleognathous Carinate Birds from the Early Tertiary of North 274 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY America. Science, 214(4526): 1236-1237. Iordanski, N.N. 1990. Evolution of Complex Adaptations: Feeding Apparatus in Amphibi ans and Reptiles. 310 pages, 65 figures. Moscow: Nauca. [In Rus sian.] Kuular, U.S., and F.Ya. Dzerzhinsky 1994. Trophical Adaptations of Cranes (Gruidae) as Seen from the Com parative and Functional Morphology of the Jaw Apparatus. 62 pages, 11 figures. Deposited Paper in VINITI, Russian Academy of Sciences, N 2904, Moscow State University, Moscow. [In Russian.] McDowell, S. 1948. The Bony Palate of Birds, Part 1: The Palaeognathae. Auk, 65(4):520-549. 1978. Homology Mapping of the Primitive Archosaurian Reptile Palate on the Palate of Birds. Evolutionary Theory, 4:81-94. Moller, W. 1930. Uber die Schnabel?und Zungenmechanik blutenbesuchender Vo gel, Teil I. Biologia Generalis, 6:651-726. 1931. Uber die Schnabel?und Zungenmechanik blutenbesuchender Vo gel, Teil II. Biologia Generalis, 7:99-154. Muller, H.J. 1963. Die Morphologie und Entwicklung des Craniums von Rhea ameri cana Linne, II: Viszeralskelett, Mittelohr und Osteocranium. Zeitschrift fur die Wissenschaftliche Zoologie, 168(l/2):35-l 18. Parker, T.J. 1891. Observations on the Anatomy and Development of Apteryx. Philo sophical Transactions of the Royal Society of London, series B, 182:25-134. Parker, W.K. 1866. On the Structure and Development of the Skull in the Ostrich Tribe. Philosophical Transactions of the Royal Society of London, 156:113-183, 15 plates. Pycraft, W.P. 1900. On the Morphology and Phylogeny of the Palaeognathae (Ratitae and Crypturi) and Neognathae (Carinatae). Transactions of The Zoological Society, London, 15(5, 6): 149-290. Sibley, C.G., and B.L. Monroe 1990. Distribution and Taxonomy of Birds of the World. 1111 pages. New Haven: Yale University Press. Webb, M. 1957. The Ontogeny of the Cranial Bones, Cranial Peripheral and Cranial Parasympathetic Nerves, Together with a Study of the Visceral Muscles of Struthio. Acta Zoologica (Stockholm), 38:81-203. Yudin, K.A. 1961. On the Mechanismus of Mandible in Charadriiformes, Procellarii formes, and Some Other Birds. Trudy Zoologicheskogo Instituta Ak ademii Nauk SSSR [Proceedings of the Zoological Institute of the Academy of Sciences of USSR], 29:257-302, 30 figures. [In Rus sian.] 1970. Biological Significance and Evolution of the Cranial Kinetics in Birds. Trudy Zoologicheskogo Instituta Akademii Nauk SSSR [Pro ceedings of the Zoological Institute of the Academy of Sciences of USSR], 47:32-66. [In Russian.] 1978. Classical Morphological Characters and Recent Avian Systematics. Trudy Zoologicheskogo Instituta Akademii Nauk SSSR [Proceedings of the Zoological Institute of the Academy of Sciences of USSR], 76:3-8. [In Russian.] Zusi, R.L. 1984. A Functional and Evolutionary Analysis of Rhynchokinesis in Birds. Smithsonian Contributions to Zoology, 395: 40 pages, 20 figures. The Relationships of the Early Cretaceous Ambiortus and Otogornis (Aves: Ambiortiformes) Evgeny N. Kurochkin ABSTRACT Ambiortus from the Khurilt beds (Neocomian) of central Mon golia shows a combination of characters that confirms the assign ment of this fossil to a separate order, Ambiortiformes. Otogornis genghisi Hou, 1994, of the Yijinhouluo Formation (earliest Creta ceous or latest Jurassic) of Ordos, China, was first described as Aves incertae sedis. Ambiortus and Otogornis share specialized characters, such as a thickened, three-edged acrocoracoid with an acute top; a flat, wide humeral articular facet of the scapula; ven tral position of a small, oval humeral articular head; and a thin, long intermediate phalanx of the major wing digit. The generic sta tus of Otogornis is supported by some other diagnostic characters. Several advanced characters demonstrate the assignment of Ambiortus and Otogornis to the Palaeognathae. These two forms show the occurrence of paleognathous birds in the Early Creta ceous of Central Asia. Introduction The Early Cretaceous Ambiortus dementjevi Kurochkin, 1982, was described as a member of the new family Am- biortidae and order Ambiortiformes, which was assigned to the infraclass Carinatae (Kurochkin, 1982). Ambiortus was based on an associated portion of the skeleton, including the verte brae, the shoulder girdle, and some wing bones preserved on the main slab and counterslab, and also on the distal portion of the wing bones and feather imprints, which are preserved on an associated slab. Two papers containing more extensive descrip tion and comparison of this fossil were published later (Kuro chkin, 1985a, 1985b). The surprising appearance of this Early Cretaceous tme bird in the paleontological record made com parison with other birds very difficult. In these first papers I at tempted to compare Ambiortus with members of the living Gmidae, Rallidae, Strigidae, Alcedinidae, Momotidae, and, su- Evgeny N. Kurochkin, Paleontological Institute of the Russian Acad emy of Sciences, 123 Profsouznaya Street, 117868 Moscow, GSP-7, Russia. perficially, with the Archaeopterygiformes, Ichthyornithi- formes, Paleogene paleognaths (later described as Lithornithi- formes Houde, 1988), and Wyleyia Harrison and Walker, 1973. However, further study and a possibility of a direct comparison with the Enantiornithes, Archaeopteryx, Ichthyornis, Wyleyia, Paracathartes, Lithomis, Palaeotis, and living Palaeognathae, and also additional preparation of the holotype of Ambiortus, showed the published anatomical descriptions and morphologi cal comparison of this fossil to be incomplete and partly erro neous. Evidence for the relationships of Ambiortus with the Palaeognathae was published beginning in 1985 (Kurochkin, 1985a, 1988, 1995a, 1995b). This mainly was based on com parison with the Paleogene Lithornithiformes studied by Houde and Olson (1981) and Houde (1988). Olson (1985) em phasized that Ambiortus shows some decided similarities with the Paleogene paleognathous birds and also may share some common characters of the humems with Ichthyornis. Martin (1987, 1991) united Ambiortus with Apatomis and the Ichthy- omithiformes. Cracraft (1986) concluded that Ambiortus can be assigned to the Carinatae, in which he included the Palaeog nathae, Neognathae, and Ichthyornis. Sereno and Rao (1992) have placed Ambiortus outside the Ornithurae without charac ter evidence. Chiappe (1995) considered Ambiortus to be the oldest probable ornithurine, yet one of unclear relationships. Elzanowski (1995) assigned Ambiortus to the Carinatae and primitive Neornithes, close to Ichthyornis; however, his cladis tic analysis of the skeletal characters also placed the Enantior nithes, Cathayomis, and Concomis among the Carinatae. Ob viously, Ambiortus has nothing in common with the Enantiornithes. In that paper, Elzanowski (1995) made mainly mistaken interpretations of skeletal characters in Ambiortus that are discussed below. Feduccia (1995, 1996) placed Am biortus in the basal Ornithurae together with Gansus, the Hes- perornithiformes, and Ichthyornithiformes. Thus, controversial phylogenetic conclusions exist concern ing Ambiortus, and it has remained isolated in the avian phylo genetic tree as the result of incomplete and questionable de scriptions (Kurochkin, 1982, 1985a, 1985b), especially those 275 276 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY concerning the amphicoelous structure of the vertebrae and the bones of the shoulder girdle. New preparation and observation of the specimen provide a complete and corrected anatomical description of this bird. It was a great surprise to discover that Ambiortus dementjevi is similar to Otogornis genghisi, which was described by Hou (1994) from the Ordos Basin at the Chabu Sumu locality, Otog Qi, Yikezhao-meng, Inner Mongolia, China. The specimen was collected in the thin, grey green mudstones of the Yijinhuoluo Formation of the Zidan Group in the Lycoptera-bearing depos its and represents the earliest Cretaceous or even a Late Juras sic avian fossil from China (Hou, 1994). Otogornis is based on associated elements of the forelimb and shoulder girdle (VP- 9607, holotype, Institute of Vertebrate Paleontology and Pale- anthropology (IVPP), Beijing). The slab also displays some feather impressions. Otogornis was described as Aves incertae sedis and was compared with Archaeopteryx, Chaoyangia, and the enantiornithines Sinornis and Cathayornis (Hou, 1994). Earlier, the same specimen was assigned to the indeterminate Enantiornithes (Dong, 1993). ACKNOWLEDGMENTS.?I thank very much L. Hou and Z. Zhou for permitting investigation of the holotype of O. genghi si in the IVPP collection. L. Martin prompted me to use for the investigation the mold from specimen 3790-272, which led to the discovery of a contact with the counterslab specimen 3790-271- (specimens in the Paleontological Institute, Russian Academy of Sciences (PIN)). Comparison with the lifhorni- thids was made possible by the courtesy of R. Emry in the De partment of Vertebrate Paleontology, National Museum of Nat ural History, Smithsonian Institution, Washington, D.C. For comments on the manuscript and English corrections I am thankful to S. Lucas, R. Zusi, S. Olson, and two anonymous re viewers. The stereophotographs were made by S. Morton in the Faculty of Sciences of Monash University, Clayton, Australia. The x-radiograph of Ambiortus dementjevi was made by L. Martin and J. Chom in the Museum of Natural History, Univer sity of Kansas. All drawings, including Otogornis, were made by the author. This study was supported by a travel grant and grant 96-04-50822 of the Russian Fund for Fundamental Re search, as well as by funds from the IVPP, Academy of Scienc es of China. Age of the Khurilt Beds The geological age of Ambiortus is problematical. It was found in Mesozoic rocks of the Gobian Altai in central Mongo lia in the Khurilt-Ulan Bulak locality. This discovery caused some paleontologists to doubt the correct definition of the fos sil and the age of the deposits. At present, no doubts exist about the advanced avian condition of this specimen. The geologic age of the Upper Cretaceous lacustrine shales and sandstones of the Khurilt locality, however, is discussed in contradictory terms by different experts in biostratigraphy, paleobotany, and paleoentomology, with the dates ranging from the latest Juras sic to the Aptian. Numerous and various insects were collected in these beds (Zherikhin, 1978; Sinitsa, 1993). The insect fauna is very con stant in a number of localities in central Mongolia (Khurilt, up per members of Kholbotoo, Bon Tsagan). This is known as the Bon Tsagan assemblage, the youngest among three Lower Cre taceous assemblages in central Mongolia (Ponomarenko, 1990). Dmitriev and Zherikhin (1988) supported an Aptian age of the deposits in these localities. The plant associations of the Khurilt, neighboring deposits of the Kholbotoo, and of the middle levels of the Bon Tsagan lo calities include, following Krassilov (1982), four phytostrati- graphic units in the Upper Cretaceous of Mongolia. The third unit is the Baierella hastata (Bennettitales; and its cones, Karkenia mongolica)/'Araucaria mongolica zone, which in cludes localities of Shin Khooduk-Anda Khooduk level, and most of the paper shales of the Bon Tsagan, Kholbotoo Gol, Khurilt, Erdeni Ula, Shin Khooduk, and Modon Usoo locali ties. The sediments of the Khurilt and Kholbotoo localities were assigned by Krassilov (1982) to the Anda Khooduk For mation. The plant communities from these localities he corre lated with the Aptian flora of the Russian Far East (Primorye). Thus, phytostratigraphic data suggest an Aptian age for the Khurilt beds (Krassilov, 1982). Based on geological data, Martinson (1973) and Shuvalov (1975, 1982, 1993) referred the Khurilt and Kholbotoo beds to the Anda Khooduk Formation, which they correlated with the Hauterivian-Barremian. Sinitsa (1993), based on the lithofacial data and ostracod as semblages, did not provide a definite age for the Khurilt beds and defined it as a task for future exploration. She specified that the Khurilt beds belong to the Khurilt Section of the Bon Tsagan Series. The Khurilt beds are underlaid by the Dund Ar- galant Series of the latest Jurassic, which includes the Anda Khooduk Formation (Tithonian), and are overlapped nearby by the sediments of the Kholbotoo Section (younger beds of the Bon Tsagan Series). In more western areas of Mongolia, the Bon Tsagan Series are overlapped by the Khoolsyn Gol Forma tion, which is correlated with the Aptian-Albian. Perhaps the Mongolian and Chinese Lower Cretaceous shale sediments were deposited simultaneously. The problem of age determination of the lacustrine Lower Cretaceous beds in Mon golia is the same as for the Jehol Group in China (Matsukawa and Obata, 1994). The Khurilt outcrop and the upper members of the neighboring Kholbotoo outcrop are very similar to the grey green, thin-bedded sandy and oil shales and siltstones of the Jiufotang Formation in Liaoning Province of China. The Ji- ufotang Formation is a member of the Jehol Group, which is subdivided into four lithostratigraphic units: Yixian, Jiufotang, Shahai, and Fuxin formations (Smith et al., 1995). The Jiufo tang Formation is correlated with the Berriasian-Valanginian (Li and Liu, 1994), or Tithonian-Valanginian (Lin, 1994), or even with the Tithonian (Chen and Chang, 1994). On the basis NUMBER 89 277 of fossil fishes, the Yixian and Jiufotang formations are corre lated with the Late Jurassic and Neocomian, in agreement with the fish faunas of Japan, Kazakhstan, and western Europe (Fan, 1996). Published radiometric ages for the base of the Yixian Forma tion include 137?7 Ma using K/Ar and 142.5 Ma using Rb/Sr (Wang, 1983; Wang and Diao, 1984), which corresponds with a Berriassian age, using the time scale of Harland et al. (1989). For the Fuxin Formation, K/Ar ages range from 100 Ma to 137 Ma (Mao et al., 1990), corresponding with an Aptian-Valang- inian age. New age dates were reported for the Yixian Forma tion, however, that are based on a laser 40Ar/39Ar study of sin gle mineral grains (Smith et al., 1995). This study estimated the age of the lower Yixian Formation as 121.2?0.3 Ma and 121.3?2.3 and 121.4?0.7 Ma for the upper Yixian Formation. Smith et al. (1995) also tested the absolute ages using 40Ar/ 39Ar in very fine crystallites of glaucony from the white lacus trine Ershilipu sediments that occupy a stratigraphic position between the upper and lower parts of the Yixian Formation. The resulting ages of 122.1 ?0.2 Ma and 122.5?0.3 Ma agree with the strict chronostratigraphic constraints imposed by the 40Ar/39Ar ages of the upper and lower Yixian Formation. Thus, Smith et al. (1995) provided an integrated age range of 121.1-122.9 Ma for the Yixian Formation, which corresponds to the Barremian using the time scale of Harland et al. (1989), and these dates are much younger than the K/Ar and Rb/Sr dates of other authors. Thus, insects and plants suggest an Aptian age for the Khu rilt deposits, but the geological data and ostracods indicate a Neocomian age. The probably contemporaneous Yixian and Ji ufotang formations in northeastern China are assigned to the Neocomian on faunistic, plant, and radiometric data, although the laser 40Ar/39Ar study gives a Barremian age. I am inclined to accept a Neocomian age for the Khurilt deposits. Comprehensive Description of Ambiortus Ambiortus dementjevi is represented on three slabs. The main slab (PIN 3790-271+) bears the cervical and thoracic ver tebrae, furcula, left scapula and coracoid, a portion of the ster num, some thoracic ribs, the proximal portion of the left hu merus, distal portions of the radius and ulna, ulnare, the proximal portion of the left carpometacarpus, and phalanges of the major wing digit (Figures 1, 3-5). This slab also shows an isolated impression of a feather vane about 12 mm long and the probable impression of the soft body of the specimen, with small contour feathers that surround the body impression. The feathers and body impression are better represented on the counterslab (PIN 3790-271-). The counterslab bears small frag ments of the vertebrae and a portion of the major metacarpal, and it has a good mold of the humems, coracoid, clavicle, and carpometacarpus. A small associated slab (PIN 3790-272) shows just the mold of three phalanges of the major left wing digit, the mold of the ulna and major metacarpal, and a frag- FlGURE 1.?Main slab with Ambiortus dementjevi Kurochkin, 1982, holotype PIN 3790-271+; Khurilt Ulan Bulak locality, central Mongolia, Neocomian. Stereopairs. (Scale bar= 1 cm.) 278 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ment of the radius that are continued from the main slab (Fig ures 2, 3). This slab also bears several impressions of the wing feathers. The counterslab and associated slab form a broken contact with each other (Figure 3). VERTEBRAE.?The preserved 15 or 16 cervical and thoracic vertebrae (3-17(18)) are ventrally exposed on the main slab (Figures 1, 3, 5). The first (atlas) and second (axis) vertebrae are absent, although a probable fragment of the latter is pre served in front of the first preserved vertebra, which I consider to be the third cervical. The first two preserved vertebrae each have a short, wide, flat ventral body surface; thus, I consider them to be the third and fourth cervicals. Most of the portions of the probable fifth and sixth cervicals are disarticulated from the other vertebrae. The seventh and eighth vertebrae each show a long, flattened body ventrally. They have well-devel oped cranial zygapophyses with costal processes and caudal zygapophyses (Figures 1, 3, 5). The carotid processes form a shallow open carotid canal. The strong, paired, caudal, trans verse processes are developed on the ventral side lateral to the caudal articular surface. Such processes are present in Lithornis and in Rhea; they are different from the postlateral processes in grebes (Zusi and Storer, 1969). The ninth cervical was extracted in order to investigate the articular surfaces of the vertebral bodies because in earlier pa pers the vertebrae of Ambiortus were regarded as probably amphicoelous. Although this was widely used to establish the relationships of this bird, it appears to have been a misinter pretation. As a result of this new preparation of the vertebrae, it has become evident that the caudal articular surface of the eighth cervical and the cranial surface of the tenth cervical were certainly heterocoelous. The tenth and eleventh vertebrae have a longer centrum with a narrower ventral side than do those of the third and fourth vertebrae and those of the thir teenth through sixteenth vertebrae, but they have shorter cen tra than in the seventh and eighth vertebrae. The twelfth, thir teenth, and fourteenth cervical vertebrae show shortened bodies and enlarged costal processes represented only by the sturdy basal portions. The basal portions of two well-ex pressed ribs are preserved near the right side of the fifteenth vertebra. These ribs are small and represent the floating ribs. The sixteenth vertebra is very compressed. The seventeenth vertebra has a wide and nearly flat ventral body surface and a very narrow caudal articular surface. On the left side of this vertebra the dorsal portion of a large rib is present; it has two articular facets, although the dorsal one is not preserved in this sample. A probable portion of the eighteenth vertebra is pre served caudal to the seventeenth vertebra in the angle between the coracoid and sternum. The thirteenth through seventeenth vertebrae have shortened bodies. The ventral crests (hypapo- physes) are not present in either the cranial cervical or the tho racic vertebrae. I think that this specimen preserves at least 15 cervical vertebrae, which mainly have wide and short centra. Thus, Ambiortus was a short-necked bird. FIGURE 2.?Associated slab with Ambiortus dementjevi, PIN 3790-272; Khu rilt Ulan Bulak locality, central Mongolia, Neocomian. (Scale bar= 1 cm.) SHOULDER GIRDLE.?The cranial portion and left costal margin of the sternum preserves the base of a thick pillar of the sternal keel and a damaged sternal costal margin (Figures 1, 3). The cranial surface of the pillar is directed down and somewhat caudally. The furcula is represented by the dorsal portions of both clavicles, which terminate at slightly thinned ends that are neither enlarged nor flattened. The small articular facet for the coracoid is directed caudomedially. The cross section of the clavicle in its middle portion is nearly circular. The medial side of the clavicle is slightly flattened. Originally, the holotype preserved the mold of the interclavicular symphysis, which was placed superficially on the level of the fifteenth vertebra. This was destroyed during preparation. The mold of the ventral fur cula showed that the ventral portion of both clavicles have the same diameter as dorsal ones. In the interclavicular area, a thin, slightly projecting eminence around the symphysis was present, without any vestige of the hypocleideum. In the left scapula, the caudal end of the scapular body is ab sent. The humeral articular facet is flat and wide and faces late- ro-cranially but is not nearly perpendicular to the scapular blade as noted by Elzanowski (1995). In its cranial area, the facet converges into a slightly convex, ellipsoidal coracoidal tubercle. A strongly projecting and sturdy acromion is very dis tinctive for this scapula (Figure 4). The acromion is dorsoven trally compressed, and its cranial ending is blunt and attenuated cranially. The dorsal surface of the acromion possesses a con spicuous dorsal tubercle (not crest) that is very similar to that in Lithornis celetius Houde, 1988. The acromion is not turned mediad, contrary to Elzanowski (1995). The scapular body is blade-like, strongly flattened lateromedially, and slightly NUMBER 89 279 FIGURE 3.?Composition of the main slab (PIN 3790-271+) and of the mold of the associated slab (PIN 3790-272), with the holotypical partial skeleton of Ambiortus dementjevi. APR=acrocoracoid process, CL=clavicle, CPR=costal process, CRP=cranial pit, CTP=caudal transverse process, CTR=carpal trochlea, DCR=deltopectoral crest, DPG=deltopectoral groove, EPR=extensor process, FIM=feather imprints, HAF=humeral articular facet, IPH=inter-mediate phalanx of major wing digit, ITF=infratrochlear fossa, LFO=liga-mentaI fossa, LHM=left humerus, LPC=lateral process of coracoid, LSC=left scapula, MAM=major metacarpal, MIM=minor metacarpal, PPH=proximal phalanx of major wing digit, PPR=procoracoid process, R=radius, RB=rib, RCM=right carpometacarpus, SBI=soft-body imprint, ST=sternum, V3-V15=3rd-15th ver tebra, UL=ulna, UPH=ungual phalanx of major wing digit, UR=ulnare. (Scale bar=l cm.) curved dorsally. The scapula has a moderately long body bear ing an elongated tuberosity on the dorsolateral edge in its crani al half and has a well-expressed depression along the lateral surface of its caudal half. 280 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY The left coracoid is represented by the shoulder end and shaft. The sternal end is covered by matrix and sternal bone, but it is well seen in the x-radiograph (Figure 5). The sternal end is wide and flat, with a long-pointed medial angle and with a rectangular lateral process. Such a structure of the medial an gle is very similar to that in Lithornisplebius Houde, 1988. The acrocoracoid is sturdy, relatively short, and its dorsal top is three-edged and bluntly acute. The craniomedial side of the ac rocoracoid bears an elongate depression that probably repre sents an articulation for the clavicle. The lateral side of the ac rocoracoid possesses a wide, slightly concave depression of the acrocoracohumeral tendon. Ventral to this depression, a rela tively small humeral articular facet is located, the facet being exposed laterally. An ellipsoidal scapular cotyla is exposed caudolaterally and is located on an enlarged base of a wide, flat, long procoracoid process. The sternal portion of the cora coidal shaft is strongly broadened. None of the elements of the shoulder girdle are compressed, and they all preserve the tme configurations of the bones. WING BONES.?The proximal end of the left humerus was strongly compressed in its plane during preservation. The hu meral articular head is small, bean-shaped, and located in the ventral position of the proximal end. The humerus has a well-developed deltopectoral crest beginning very close to the humeral articular head in the most proximal position of the proximal end; it is similar to that in Lithornis plebius. The del topectoral crest is flat but is rather deflected dorsally, contrary to Elzanowski (1995), who described it as projecting laterally. FIGURE 4.?Shoulder articulation in Ambiortus dementjevi on the opposite view of the main slab, PIN 3790-271+. ACR=acromion, APR=acrocoracoid process, ATB=acromial dorsal tubercle, CMR=caudal margin, DPT=dorsal pit, HAH=humeral artiular head, TRF=tricipital fossa, VTB=ventral tubercle. (Scale bar= 1 cm.) FIGURE 5.?X-radiograph of the main slab with Ambiortus dementjevi. The sternal end of the coracoid with the medial angle and structure of the vertebrae can be clearly seen. CTP=caudal transverse process, LPC=lateral process of coracoid, MAN=medial angle of stemal end of coracoid, RB=rib, V10=10th vertebra. Along the dorsal margin, a shallow groove appears in the prox imal half on the cranial side. The bicipital crest and pneumotri- cipital fossa are absent; the latter is expressed only as a tricipi tal depression (Figure 4). The ventral edge of the proximal end of the humerus is remarkably projected ventrally. Its distal edge is like a boss. The cranial surface of this boss possesses a slightly pronounced cranial tubercle with a pit in the center. Cranially from this tubercle is a noticeable ligamental fossa (not a groove). Lithornis plebius also has a similar tubercle possessing a pit and has a ligamental fossa instead of a furrow. Such a ligamental fossa is probably the homolog of the trans verse ligamental furrow. On the caudal surface of the proximal end, a small dorsal pit is developed in the usual place of the dorsal tubercle. A small ventral tubercle is represented on the caudal surface of a projecting ventral edge. The capital groove is not developed. A slightly elevated caudal margin mns along the middle of the shaft and is directed toward the middle of the humeral head. The ulna is badly damaged. Only its distal end and a mold of a portion of the shaft are preserved on the main and associated NUMBER 89 281 slabs, respectively. The radius also is represented by portions of the shaft on both these slabs (Figure 3). It shows a circular cross section in the midshaft. The ulnare lies near the proximal end of the carpometacarpus. This curved element exhibits just its oval-shaped proximal portion, in which a small fossa is de veloped in the place of attachment of the humerocarpal liga ment. The left carpometacarpus displays its ventral side and the proximal articular surface (Figures 1, 3). It has a well-devel oped ulnocarpal trochlea and a deep infratrochlear fossa. The carpal trochlea appears to be small and narrow, and the exten sor process is poorly developed, in accord with Elzanowski (1995). The pisiform process is either not preserved or is not developed. The major metacarpal is represented by bone frag ments on the main slab and by a mold on the associated slab. The minor metacarpal is represented by the most proximal part of the base and by a mold of a small portion of the shaft on the main slab. The metacarpals are completely fused at their proxi mal ends. The proximal shafts of both metacarpals are similar in size. The molds of all three phalanges of the major wing dig it are displayed on the associated slab, with the ventral sides exposed. The proximal phalanx has a typically avian morphol ogy, with a flat cranial surface and a thin, flat caudal plate, with two divided depressions on the ventral side. The intermediate phalanx is long and thin, and it does not show a vestigial condi tion. The ungual phalanx is flat, short, and slightly bowed. The intermediate and ungual phalanges form a good articular joint with each other. SOME FEATURES IN THE MORPHOLOGY OF Ambiortus AND Otogornis One of the most characteristic properties of Ambiortus de mentjevi was supposed to be the amphicoelous cervical verte brae, as I had proposed in earlier publications on this fossil (Kurochkin, 1982, 1985a, 1985b). As emphasized above, how ever, the eighth and tenth cervical vertebrae are now known to have heterocoelous centra. New observations also revealed a contact between the broken edges of the counterslab and asso ciated slab. Thus, the major metacarpal, radius, and ulna in the main slab show extension on the associated slab with specimen PIN 3790-272 that provides certain confirmation of belonging to the same specimen. I have not attempted a detailed description of Otogornis genghisi, but I mention just some corrections to the original pa per and the characters important for comparison with Ambior tus dementjevi. Most characters of Otogornis genghisi that are used in this paper were published in the original description by Hou (1994). In contrast to Hou's observations, however, I discovered that the deltopectoral crest is present, the transverse ligamental fur row is only expressed as a distinctive fossa, the dorsal cotyla of the proximal end of the ulna is well preserved, and the metacar pals are fused at their proximal base, although this area is very cmshed (Figures 6, 7). HAH RSC LRD FIM LRD DCT FIGURE 6.?Otogornis genghisi Hou, 1994, from the Chabu Sumu locality, Ordos Basin, China, Yijinhuoluo Formation. Cranial view; drawing made from a slide. CGR=capital groove, CRP=cranial pit, DCD=dorsal condyle, DCR=deltopectoral crest, DCT=dorsal cotyla, FIM=feather imprints, HAH= humeral articular head, LCR=left coracoid, LFU=ligamental furrow, LHM= left humerus, LRD=left radius, LUL=left ulna, OL=olecranon, OLF= olecranal fossa, RHM=right humerus, RRD=right radius, RSC=right scapula, SGR=scapular groove, TRF=tricipital fossa, VCD=ventral condyle. VTB= ventral tubercle. (Scale bar=l cm.) A wide, flat humeral articular facet in the scapula of Otogor nis faces latero-cranially. The projecting ventral edge of the proximal end of the humerus in Otogornis possesses a small cranial tubercle with a pit in the center that is very similar to Ambiortus. Perhaps Otogornis is similar to Ambiortus and the Lifhornifhiformes in the specialized morphology of having a dorsoventrally compressed scapular acromion with a tubercle on its dorsal side. Despite being broken, the acromion of Otogornis shows some dorsoventral flattening with a promi nence on the dorsal surface. The cranial portion of the lateral surface of the scapula bears a distinctive scapular groove. The specimen of Otogornis genghisi exhibits the imprints of two wing feathers. Hou (1994) pointed out an important char acteristic of these feathers, which is that they are not tightly ar ranged, i.e., there is no bonding of the barbs by barbules. Am biortus dementjevi also preserves some feather imprints in the area of the wing feathers, although these show bonded feather vanes. COMPARISON The bones of the shoulder girdle and forelimb in Otogornis genghisi are somewhat longer than those in Ambiortus de- 282 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY LCR CTR PPR LRD LUL DMB FIGURE 7.?Otogornis genghisi Hou, 1994. Caudal view. ACR=acromion, APR=acrocoracoid process, CTR=carpal trochlea, DMB=damaged bones, HAF= humeral articular facet, IPH=intermediate phalanx of major wing digit, LCM=left carpometacarpus, LCR=left coracoid, LRD=left radius, LUL=left ulna, MAM=major metacarpal, MIM=minor metacarpal, PPR=procoracoid process, RHM=right humerus, RRD=right radius, RSC=right scapula, SGR=scapular groove, TRF=tricipital fossa, UPH=ungual phalanx of major wing digit, VTB= ventral tubercle. (Scale bar=lcm.) mentjevi, and they also look more robust. Comparison of the separate morphological characters, however, shows that these two Early Cretaceous birds have a close relationship. The char acters and their conditions in Ambiortus, Otogornis, and some higher avian taxa are shown in Table 1. The Ichthyomithes and Neognathae are accepted as outgroups for determining polarity (advanced or primitive condition). The Enantiornithes are used only for general comparison because they have no close rela tionship with the Ornithurae. Ambiortus and Otogornis share the combination of the fol lowing characters: a thickened, three-edged acrocoracoid with an acute top (character 8); a flat, wide humeral articular facet of the scapula; ventral position of a small, short, and oval humeral articular head (character 10); and a long, thin intermediate pha lanx of the major wing digit (character 16). These characters provide evidence for a close relationship between Ambiortus and Otogornis, and for the assignment of Otogornis to the Am biortiformes. A convex coracoidal cotyla in the scapula (character 7) and concave scapular cotyla in the coracoid (character 9) unite Ambiortus and Otogornis with the Ornithurae. Two advanced characters support the assignment of Ambiortus to the Neorni thes. These are the heterocoelous cervical vertebrae (character 1) and the U-shaped furcula (character 3). With the Palaeog nathae, Ambiortus and Otogornis share the advanced condi tion of a projecting, dorsoventrally compressed scapular acro mion (character 5) with a tubercle or prominence on its dorsal side (character 6); a projecting ventral edge of the humeral proximal end (character 13); and a remarkable cranial tuber cle with a pit in the center of the cranial surface of this pro jecting edge (character 14). Ambiortus also shares with the Palaeognathae the strong, ventral, caudal transverse processes of the cervical vertebrae (character 2; unknown for Otogor nis). Ambiortus and Otogornis also have a number of generalized characters that are common to the Palaeognathae and/or Lithor- nithiformes and are primitive in respect to the Neognathae. TABLE 1.?Distribution of some characters among Ambiortus, Otogornis, and other birds. Character 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Cervical vertebrae Ventral caudal trans verse process Furcula Hypocleideum Acromion Acromial dorsal tubercle Coracoidal cotyla Acrocoracoid Scapular cotyla Humeral articular head Bicipital crest Pneumotricipital fossa Ventral edge, proximal end of humerus Cranial tubercle Transverse ligamental depression Intermediate phalanx of major wing digit Ambiortus heterocelous present U-shaped absent projecting, blunt present convex three-edged, acute concave small absent depression strongly project ing present fossa long, thin Otogornis ? 7 7 ? projecting present? convex three-edged, acute concave small absent depression strongly project ing present fossa long, thin Ichthyomithes amphicelous absent U-shaped absent attenuated absent convex rounded concave large absent tricipital fossa projecting absent absent short Neognathae heterocelous absent U, V-shaped absent or pesent attenuated or short absent convex rounded or elongate concave or flat large present fossa or foramen rounded or project ing absent furrow long, flat Palaeognathae heterocelous present U-shaped absent projecting present convex short, rounded concave small absent fossa or depression strongly projecting present? shallow furrow long, three-edged Lithomithiformes Enantiornithes heterocelous present U-shaped absent projecting, acute present convex rounded concave small absent fossa strongly project ing present fossa ? opisthocelous absent V-shaped present short absent concave stick-like boss small absent fossa rounded absent present long, flat NUMBER 89 283 These are the absence of the bicipital crest and intumescence (character 11); the absence of the pneumatic foramen in the pneumotricipital fossa, being expressed only as an undivided tricipital depression (character 12); a small ventral tubercle on the proximal end of the humems; the presence of a ligamental fossa instead of a transversal ligamental furrow (character 15); a rounded cross-section of the shaft of the radius; and the pres ence of an ungual phalanx on the major wing digit. These char acters demonstrate that the Ambiortiformes have a common or igin with other orders of paleognathous birds. Otogornis differs from Ambiortus in having a smaller proco racoid process; a deep groove on the lateral side of the shoulder end of the scapula; a wide scapular blade (narrow in Ambior tus); a flat, elongated excavation along the cranial side of the deltopectoral crest; and the presence of a capital groove, which is divided into two furrows (Figures 6, 7). Differences in the detailed morphology of the scapula and humems support their separate generic status, although it could be argued that they are only two species of a single genus. Ambiortus dementjevi is smaller than Otogornis genghisi. The maximum width of the proximal end of the humems of A. dementjevi is 13.0 mm, and the maximum width across the most projecting edge of the deltopectoral crest is 11.2 mm. The same measurements in O. genghisi are 15.8 mm and 12.2 mm, respectively. Conclusions Ambiortus from central Mongolia and Otogornis from the Or- dos Basin, China, show a close relationship based on the shared, specialized characters 8, 10, and 16 in the structure of the fore limb and shoulder girdle (Table 1). At the same time, Ambiortus and Otogornis show some differences in the shoulder girdle and the forelimb that support their separate generic status. The relationships of Ambiortus and Otogornis with other birds are determined by comparison with the Ichthyomithes, Neornithes, Palaeognathae, and Neognathae. Ambiortus and Otogornis share an advanced condition of characters 7 and 9, which are common to the Ornithurae. Ambiortus shares with the Neornithes an advanced condition of characters 1 and 3, which are unknown for Otogornis. At the same time, the Am biortiformes share with the Palaeognathae (including Lithomi- thiformes) such specialized characters as 5, 6, 13, and 14, which suggests their assignment to the parvclass Palaeog nathae, sensu Kurochkin (1995b). No common advanced char acters were found for the Ambiortiformes, Ichthyomithes, and Enantiornithes. This study confirms that the Ambiortiformes are not closely related to the Ichthyomithes or the Neognathae and are totally unrelated to the Enantiornithes. The Early Cretaceous Ambiortiformes were flying palaeog- nathous birds. Thereby, they document an early diversification of ornithurine birds into two main evolutionary branches: Palaeognathae and Neognathae. Literature Cited Chen, Peiji, and Zhenlu Chang 1994. Nonmarine Cretaceous Stratigraphy of Eastern China. Cretaceous Research, 15:245-257, 4 figures. Chiappe, Luis M. 1995. First 85 Millions Years of Avian Evolution. Nature. 378:349-355, 6 figures. Cracraft, J. 1986. The Origin and Early Diversification of Birds. Paleobiology, 12:383-399, 1 table, 4 figures. Dmitriev, V.Yu., and V.V. Zherikhin 1988. [The Changes of Diversity in the Insect Families on the Data by the Method of Accumulated Occurrences.] In A.G. Ponomarenko, edi tor, Cretaceous Biocenotic Crisis and the Evolution of Insects, pages 208-215, 1 table, 1 figure. Moscow: Nauka. [In Russian.] Dong, Zhi-Ming 1993. A Lower Cretaceous Enantiornithine Bird from the Ordos Basin of Inner Mongolia, People's Republic of China. Canadian Journal of the Earth Sciences, 30:2177-2179, 1 table, 2 figures. Elzanowski, A. 1995. Cretaceous Birds and Avian Phylogeny. Courier Forschungsinstitut Senckenberg, 181:37-53, 4 figures. Fan, Jin 1996. New Advances in the Late Mesozoic Stratigraphic Research of Western Liaoning, China. Vertebrata Palasiatica, 34(2): 102-122, 3 tables. [In Chinese, with English summary.] Feduccia, Alan 1995. Explosive Evolution in Tertiary Birds and Mammals. Science, 267:637-638, 2 figures. 1996. Origin and Evolution of Birds, x+420 pages. New Haven and Lon don: Yale University Press. Harland, W.B., R.L. Armstrong, A.V. Cox, L.E. Craig, A.G. Smith, and D.G. Smith 1989. A Geologic Time Scale 1989. xvi+263 pages. Cambridge: Cam bridge University Press. Harrison, C.J.D., and CA. Walker 1973. Wyleyia: A New Bird Humerus from the Lower Cretaceous of Eng land. Palaeontology, 16:721-728, 2 figures. Hou, Lianhai 1994. A Late Mesozoic Bird from Inner Mongolia. Vertebrata Palasiatica, 32(4):258-266, 3 figures, 1 plate. Houde, P. 1988. Paleognathous Birds from the Early Tertiary of the Northern Hemi sphere. Publications of the Nuttall Ornithological Club, 22: vii+148 pages, 27 tables, 41 figures. Houde, P., and S.L. Olson 1981. Paleognathous Carinate Birds from the Early Tertiary of North America. Science, 214:1236-1237, 2 figures. Krassilov, V.A. 1982. Early Cretaceous Flora of Mongolia. Palaeontographica, series B, Paldophytologie, 181:1-43, 1 table, 11 figures, 20 plates. 284 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Kurochkin, E.N. 1982. [New Order of Birds from the Lower Cretaceous of Mongolia.] Doklady Academii Nauk SSSR, 262:452-455, 2 figures. [In Russian.] 1985a. Lower Cretaceous Birds from Mongolia and Their Evolutionary Sig nificance. In V.D. Ilyichev and V.M. Gavrilov, editors, Acta XVIII Congressus Internationalis Ornithologici, 1:191-199, 4 figures. Moscow: Nauka. 1985b. A True Carinate Bird from Lower Cretaceous Deposits in Mongolia and Other Evidence of Early Cretaceous Birds in Asia. Cretaceous Research, 6:271-278,4 figures. 1988. [Cretaceous Birds of Mongolia and Their Significance for Study of Phylogeny of Class Aves.] Trudy, Sovmestnaya Sovetsko-Mon- gol'skaya Paleontologicheskaya Ekspeditsiya, 34:33?42, 1 table, 2 figures, 1 plate. [In Russian.] 1995a. Morphological Differentiation of the Palaeognathous and the Neog- nathous Birds. Courier Forschungsinstitut Senckenberg, 181:79-88, 2 tables, 1 figure. 1995b. Synopsis of Mesozoic Birds and Early Evolution of Class Aves. Ar chaeopteryx, 13:47-66, 1 figure. Li, Wenben, and Zhaosheng Liu 1994. The Cretaceous Palynofloras and Their Bearing on Stratigraphic Correlation in China. Cretaceous Research, 15:333-365, 18 figures. Lin, Qi-bin 1994. Cretaceous Insects of China. Cretaceous Research, 15:305-316, 6 figures. Mao, S., J. Yu, and J.K. Lentin 1990. Palynological Interpretation of Early Cretaceous Non-marine Strata of Northeast China. Reviews of Paleobotany and Palynology, 65:115-118. Martin, L.D. 1987. The Beginning of the Modem Avian Radiation. Documents des Lab- oratoires de Geologie Lyon, 99:9?19, 3 figures. 1991. Mesozoic Birds and the Origin of Birds. In H.P. Schultze and L. Trueb, editors, Origins of the Higher Groups ofTetrapods: Contro versy and Consensus, pages 485-540, 51 figures. Ithaca: Cornell University Press. Martinson, G.G. 1973. [On Stratigraphy of Jurassic and Cretaceous Deposits of Mongolia.] Izvestiya Akademii Nauk SSSR, Geologicheskaya Seriya, 12:89-95, 1 table. [In Russian.] Matsukawa, Masaki, and Ikuwo Obata 1994. Dinosaurs and Sedimentary Environments in the Japanese Creta ceous: A Contribution to Dinosaur Facies in Asia Based on Mollus can Palaeontology and Stratigraphy. Cretaceous Research, 15:101-125, 12 figures. Olson, S.L. 1985. The Fossil Record of Birds. In D.S. Farner, J.R. King, and K.C. Parkes, editors, Avian Biology, 8:79-252, 11 figures. New York: Academic Press. Ponomarenko, A.G. 1990. [Insects and the Lower Cretaceous Stratigraphy of Mongolia.] In V.A. Krassilov, editor, Continental Cretaceous of the USSR, pages 103-108. Vladivostok: Far Eastern Branch of the USSR Academy of Sciences. [In Russian.] Sereno, P.C, and Chenggang Rao 1992. Early Evolution of Avian Flight and Perching: New Evidence from the Lower Cretaceous of China. Science, 255:845-848, 5 figures. Shuvalov, V.F. 1975. [Mesozoic Stratigraphy in the Central Mongolia.] Trudy, Sovmest naya Sovetsko-Mongol 'skaya Geologicheskaya Ekspeditsiya, 13:50-112, 3 tables, 21 figures. [In Russian.] 1982. [Paleogeography and the History of Development of the Mongolian Lake Basins in the Jurassic and Cretaceous.] In G.G. Martinson, ed itor, Mesozoiskie Ozernie Basseiny Mongolii, pages 18-80, 1 table. Leningrad: Nauka. [In Russian.] 1993. [The Late Jurassic and Neocomian (?) Deposits in Trans-Altai Gobi (Mongolia).] Stratigraphiya, Geologicheskaya Korreliatziya, 1(3):76-81, 1 figure. [In Russian.] Sinitsa, S.M. 1993. [Jurassic and Lower Cretaceous of Central Mongolia.] Trudy, Sovmestnaya Rossiisko-Mongol 'skaya Paleontologicheskaya Eks peditsiya, 42:1-238, 4 tables, 71 figures, 16 plates. [In Russian.] Smith, Patrick E., Norman M. Evensen, Derek York, Mee-mann Chang, Fan Jin, Jin-ling Li, Stephen Cumbaa, and Dale Russell 1995. Dates and Rates in Ancient Lakes: 40Ar-39Ar Evidence for an Early Cretaceous Age for the Jehol Group, Northeast China. Canadian Journal of Earth Sciences, 32:1426-1431, 1 table, 4 figures. Wang, D. 1983. On the Age of the Rehe Group in Western Liaoning Province, China. Bulletin of the Chinese Academy of Geological Sciences, 1: 37-64, 2 tables. Wang, D.F, and N.C Diao 1984. Geochronology of Jura-Cretaceous Volcanics in West Liaoning, China. In Scientific Papers on Geology for International Exchange, Prepared for the 27th International Geological Congress, 1:1-12, 2 tables. Beijing: Geological Publishing House. [In Chinese, with En glish summary.] Zherikhin, V.V. 1978. [Development and Changing of the Cretaceous and Cenozoic Fau- nistic Complexes.] In Trudy, Paleontologicheskogo Instituta Aka demii Nauk SSSR, 165:1-198, 3 tables, 20 figures. Moscow: Nauka. [In Russian.] Zusi, Richard L., and Robert W. Storer 1969. Osteology and Myology of the Head and Neck of the Pied-Billed Grebes (Podilymbus). Miscellaneous Publications, Museum of Zool ogy, University of Michigan, 139:1-49, 4 tables, 19 figures. Enantiornithes: Earlier Birds than Archaeopteryx? Zygmunt Bocheriski ABSTRACT The oldest known remains of the Enantiornithes come from the Early Cretaceous of Spain and northeast China. They represent birds capable of flight, although it was not efficient enough to enable them to fly over the Turgai Strait, which at that time sepa rated the eastern and western parts of the present-day Palaearctic. A comparison of the coastlines of the continents in consecutive epochs of the Jurassic and Cretaceous suggests that in order to spread by land over all of Eurasia, both Americas, and Australia, the Enantiornithes would have had to differentiate at the latest by the Middle Jurassic (Bajocian), or about 25 million years before the period from which Archaeopteryx is known (Tithonian). Introduction Remains of Cretaceous birds of the subclass Enantiornithes, as described by Walker (1981), are known from many locali ties in North and South America, Europe, Asia, and Australia (Bocheriski, 1997) (Figure 1). The earliest remains may be the European representative, Nogueromis gonzalezi Lacasa from Spain, which has been referred to a period between the upper Berriasian and lower Valanginian (Lacasa, 1989). The remains of Sinornis santensis Sereno and Rao, Cathayomis yandica Zhou, Jin, and Zhang, and Boluochia zhengi Zhou from northeastern China come from the Valanginian (Sereno and Rao, 1992; Zhou, 1995a, 1995b). Three other Spanish species, Iberomesomis romerali Sanz and Bonaparte (1992), Concomis lacustris Sanz and Buscalioni (1992), and Eoalu- lavis hoyasi Sanz et al. (1996), are younger by several million years. The remaining Enantiornithes, from Asia (Mongolia and Uzbekistan), North and South America, and Australia have been obtained from deposits representing a period be tween the Albian and the Maastrichtian (Molnar, 1986; Chi- Zygmunt Bocheriski, Institute of Systematics and Evolution of Ani mals, Polish Academy of Sciences, Slawkowska 17, 31-016 Krakow, Poland. appe, 1991, 1993; Dong, 1993; Lamb et al., 1993; Kurochkin, 1995b, 1996; Martin, 1995). When describing particular forms included in this subclass, authors have paid attention to the characters indicative of ac tive flying. Martin (1995) emphasized that, unlike the state ob served in Archaeopteryx, the stmcture of the shoulder girdle and the possession of a keeled sternum indicate that these birds were able to rise from flat surfaces. On the other hand, the ster num and, especially, the carina stemi were proportionally very small, pointing to restricted flight efficiency that did not permit the birds to cover long distances. Poor powers of flight charac terized all Sauriurae, in contrast to contemporary Ornithurae (Zhou, 1995c). The geographic situation of the earliest Early Cretaceous localities (Figure 2) as seen against a background of the paleocoastlines at that time (Smith et al., 1995) shows that Enantiornithes occurred on both sides of the Turgai Strait, which then was at least several hundred kilometers wide and constituted a substantial obstacle for terrestrial vertebrates, as pointed out earlier by Kurten (1967-1970). Naturally, some cases of passive crossing of this barrier cannot be excluded, al though they seem unlikely. The Turgai Strait existed uninter ruptedly for many millions of years, from the Callovian to the Aptian (Smith et al., 1995). Protoavis texensis Chatterjee is considered to be a bird by some authors (e.g., Chaterjee, 1991, 1994; Kurochkin, 1995b). Its detection in the Upper Triassic layers (Chatterjee, 1991, 1994) and the presence of tme avian forms by the latest Juras sic (Hou, 1995) indicate that, despite the lack of direct evi dence, the differentiation of birds occurred in the Jurassic. The differentiation of the European and East-Asiatic Early Creta ceous Enantiornithes on both sides of the Turgai Strait into rather high systematic units (i.e., orders and families according to Martin, 1995), or into various genera (Kurochkin, 1996), oc curred independently under relatively stabilized biotopic con ditions, and so they must have been the result of a long-lasting evolutionary process. Thus, it seems plausible that the Enantio rnithes separated and spread in the Bajocian, 166-171 Ma BP (Haq and Van Eysinga, 1987), when it would still have been possible for them to spread over all the continents by land (Smith et al., 1995). European sea straits at that time were nar row and so could have been crossed much more easily than the 285 286 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?Distribution of localities, or groups of localities if situated close to each other, of enantiornithine birds (solid circles). later Turgai Strait (Figure 3). Later on, there was no land con nection between Laurasia and Gondwana until the Tithonian, when the present Gibraltar Strait was already very narrow. It seems, however, that the present-day Iberian Peninsula was colonized from other parts of Europe in the Valanginian-Berri- asian because previously it had been an island surrounded by ^> FIGURE 2.?Coastlines in the Valanginian-Berriasian, 138 Ma BP (modified from Smith et al., 1995, reprinted with the permission of Cambridge University Press). Squares indicate Early Cretaceous enantiornithine localities in Spain and northeastern China, which at that time were situated on continents divided by the Turgai Strait. (E=Europe, IP=Indian Peninsula, TS=Turgai Strait.) more or less wide seas. After the Tithonian, the part of the Tethys dividing Laurasia and Gondwana was wide again until the Tertiary. Probably the enantiornithine birds inhabiting these parts of the earth evolved independently during that time. The colonization of Gondwana took place in the Bajocian via the eastern part of North America and Africa (although we do not have any evidence for the occurrence of the Enantiornithes in Africa), in view of its land connection with South America C=> FIGURE 3.?Coastlines in the Bajocian, 170 Ma BP (modified from Smith et al., 1995, reprinted with the permission of Cambridge University Press). Arrows indicate hypothetical directions of dispersal of Enantiornithes. NUMBER 89 287 and Australia by way of the Antarctic (Figure 4). A second wave of colonization of North America was possible later, in dependent of the colonization of Gondwana, in the Early Creta ceous (Valanginian). In Chiappe's (1991) opinion, the Enantiornithes evolved in Gondwana, whereas Zhou (1995a) claims that Laurasia was their cradle. In Chiappe's (1991) conception, the whole process of colonization of the earth ran in the opposite direction, which, from the viewpoint of the history of continents, also is possible. If the Enantiornithes separated in the Bajocian, they theoretical ly could have originated anywhere on the earth. Vorona berivot- rensis Forster et al. (1996), discovered in Madagascar and con sidered to be a sister group of the Enantiornithes, speaks in favor of Chiappe's conception. On the other hand, the age of the remains and the differentiation of the forms from Laurasia seem to support Zhou's (1995a) opinion, and this is the reason for adopting my present course of reasoning. No matter which of these two theories is right, the history of continents indicates that the subclass Enantiornithes evolved in the Middle Jurassic, more than 25 million years before Archaeopteryx. The genera Nanantius and Enantiornis were first described from Gondwana in the Albian of Australia (Molnar, 1986) and the Maastrichtian of South America (Walker, 1981), respec tively. The acceptance of land dispersal for the Enantiornithes against a background of the history of continents raises doubts that the Late Cretaceous remains mentioned from Uzbekistan and the Gobi Desert (Nesov and Panteleev, 1993; Kurochkin, 1995a, 1996) could actually belong to these genera. Even if their flight abilities were considerably greater than in the Early Cretaceous Enantiornithes, at that time the oceans between all FIGURE 4.?Coastlines in the Albian, 105 Ma BP (modified from Smith et al., 1995, reprinted with the permission of Cambridge University Press). The star indicates the Albian locality of Nanantius in Queensland. Arrow indicates the latest possibility of colonization of Australia, assuming that the Antarctic was colonized earlier (not later than in the Tithonian). (IP=Indian Peninsula.) the places mentioned above were too wide to permit crossing (see Figure 4 and Rich, 1976). It also is doubtful that the genus Nanantius would have survived for 25 million years (i.e., from the Albian to the Campanian) or even longer. Literature Cited Bocheriski, Z. 1997("1996"). Enantiornithes?dominuj^ca grupa ladowych ptakow kredow- ych [Enantiornithes?A Dominant Group of the Cretaceaous Terrestrial Birds]. Przegl^d Zoologiczny, 40(3^1): 175-184, 4 figures. [In Polish, with English summary. Date on title pages is 1996; actually published in 1997.] Chatterjee, S. 1991. Cranial Anatomy and Relationships of a New Triassic Bird from Texas. Philosophical Transactions of the Royal Society of London, Biological Sciences, 332(1265):277-342, 40 figures. London: The Royal Society. 1994. Protoavis from the Triassic of Texas: The Oldest Bird. [Abstract.] Journal fur Ornithologie, 135(3):330. Chiappe, L.M. 1991. Cretaceous Birds of Latin America. Cretaceous Research, 12:55-63. 1993. Enantiornithine (Aves) Tarsometatarsi from the Cretaceous Lecho Formation of Northwestern Argentina. American Museum Novi tates, 3083: 27 pages, 13 figures. Dong, Zhi-Ming 1993. A Lower Cretaceous Enantiornithine Bird from the Ordos Basin of Inner Mongolia, People's Republic of China. Canadian Journal of Earth Sciences, 30:2177-2179. Forster, C.A., L.M. Chiappe, D.W. Krause, and S.D. Sampson 1996. The First Cretaceous Bird from Madagascar. Nature, 382:532-534, 4 figures. Haq, B.W., and F.W.B. Van Eysinga 1987. Geological Time Table. Fourth edition. Amsterdam: Elsevier Sci ence Publishers. [Wall plate.] Hou, L. 1995. Morphological Comparisons between Confuciusomis and Archae opteryx. In A. Sun and Y. Wang, editors, Sixth Symposium on Meso zoic Terrestrial Ecosystems and Biota; Short Papers, pages 193-202. Beijing: China Ocean Press. Kurochkin, E.N. 1995a. The Assemblage of the Cretaceous Birds in Asia. In A. Sun and Y. Wang, editors, Sixth Symposium on Mesozoic Terrestrial Ecosys tems and Biota; Short Papers, pages 203-208, 3 figures. Beijing: China Ocean Press. 1995b. Synopsis of Mesozoic Birds and Early Evolution of Class Aves. Ar chaeopteryx, 13:47-66. 1996. A New Enantiornithid of the Mongolian Late Cretaceous, and a General Appraisal of the Infraclass Enantiornithes (Aves). 50 pages, 13 figures, 3 plates. Moscow: Russian Academy of Sciences, 288 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Palaeontological Institute. [Special issue.] Kurten, Bjom 1967-1970. Continental Drift and the Palaeogeography of Reptiles and Mammals. Commentationes Biologicae, Societas Scientiarum Fen- nica, 31(1): 1-8, 2 figures. [Volume 31 number 1 published in 1967.] Lacasa Ruiz, A. 1989. Nuevo genero de ave fosil del yacimiento neocomiense del Montsec (provincia de Lerida, Espana). Estudios Geologicos, Instituto de Geologia (Madrid), 45(5-6):417-425, 5 figures. Lamb, J.P, L.M. Chiappe, and P.G.P. Erickson 1993. A Marine Enantiornithine from the Cretaceous of Alabama. [Ab stract.] Journal of Vertebrate Paleontology, 13(3):45. Martin, L. 1995. The Enantiornithes: Terrestrial Birds of the Cretaceous. In D.S. Pe ters, editor, Acta Palaeomithologica, 3 Symposium SAPE; 5 Inter nationale Senckenberg-Konferenz 22-26 Juni 1992. Courier Forschungsinstitut Senckenberg, 181:23-36, 10 figures. Molnar, R.E. 1986. An Enantiornithine Bird from the Lower Cretaceous of Queensland, Australia. Nature, 322:736-738, 4 figures. Nesov, L.A., and A.V. Panteleev 1993. [On the Similarity of the Cretaceous Ornithofaunas between South America and Western Asia.] Russian Academy of Sciences, Pro ceedings of the Zoological Institute, St. Petersburg, 252:84-94. [In Russian, with English summary.] Rich, P.V. 1976. The History of Birds on the Island Continent Australia. Proceedings of the 16th International Ornithological Congress, Canberra, pages 53-65, 6 figures. Sanz, J.L., and J.F. Bonaparte 1992. A New Order of Birds (Class Aves) from the Lower Cretaceous of Spain. In K.E. Campbell, Jr., editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:39-49, 8 figures. Sanz, J.L., and A.D. Buscalioni 1992. A New Bird from the Early Cretaceous of Las Hoyas, Spain, and the Early Radiation of Birds. Palaeontology, 35(4):829-845, 11 figures. Sanz, J.L., L.M. Chiappe, P.B. Perez-Moreno, A.D. Buscalioni, J.J. Moratalla, F. Ortega, and F.J. Poyato-Ariza 1996. An Early Cretaceous Bird from Spain and Its Implications for the Evolution of Avian Flight. Nature, 382:442-445. Sereno, PC, and C.G. Rao 1992. Early Evolution of Avian Flight and Perching: New Evidence from the Lower Cretaceous of China. Science, 255:845-848, 5 figures. Smith, A.G., D.G. Smith, and B.M. Funnell 1995. Atlas of Mesozoic and Cenozoic Coastlines, ix+99 pages, 2 figures, 31 numbered maps, 1 unnumbered map. Cambridge: Cambridge University Press. [First published in 1994, reprinted in 1995.] Walker, CA. 1981. New Subclass of Birds from the Cretaceous of South America. Na ture, 292:51-53. Zhou, Z. 1995a. The Discovery of Early Cretaceous Birds in China. In D.S. Peters, editor, Acta palaeomithologica, 3 Symposium SAPE; 5 Internation ale Senckenberg-Konferenz 22-26 Juni 1992. Courier Forschungs institut Senckenberg, 181:9-22, 10 figures. 1995b. Discovery of a New Enantiornithine Bird from the Early Cretaceous of Liaoning, China. Vertebrata Palasiatica, 33(2):99-l 13,4 figures, 1 plate. 1995c. New Understanding of the Evolution of the Limb and Girdle Ele ments in Early Birds?Evidence from Chinese Fossils. In A. Sun and Y. Wang, editors, Sixth Symposium on Mesozoic Terrestrial Ec osystems and Biota; Short Papers, pages 209-214. Beijing: China Ocean Press. Feathered Dinosaur or Bird? A New Look at the Hand of Archaeopteryx Zhonghe Zhou and Larry D. Martin ABSTRACT A detailed examination of wrist and manus skeletons in Archae opteryx, and their comparison with those of modem birds, demon strates an overwhelmingly avian appearance, much more so than has been previously recognized. Many workers have considered feathers to be the only indisputable evidence for the avian identity of this early bird. Although only a few skeletal characters have been used to support its avian identity, we believe that this is due to a lack of detail in previous analyses. We offer a list of eight uniquely derived avian characters or character complexes in the wrist and manus of Archaeopteryx. This further indicates that Archaeopteryx is a bird, with wings used for flying rather than for predation, and provides some fundamental skeletal differences between the oldest birds and their immediate ancestors. We extend our comparisons to the only other bird with Archaeopteryx-like morphology in the manus, Confuciusornis, and show how the wrist and manus may provide useful clues for discerning poten tially older and unknown birds in the future. In addition, the large number of uniquely avian characters in the wrist and manus con trasts with a more primitive anatomy in other parts, providing another example of mosaic evolution, as the structure of the wing modernized at a more rapid rate than other anatomical units. Introduction Since the discovery of Archaeopteryx in 1861, extensive studies have been conducted on this genus, and the past two de cades marked a new era of study for Jurassic birds. One result has been the resurrection, mainly by Ostrom in the 1970s, of the theory of the dinosaur origin of birds. This hypothesis de rives most of its support from comparison between Archaeop teryx and a few theropod dinosaurs, primarily Deinonychus. Although strongly challenged by ornithologists and many pale- Zhonghe Zhou, Institute of Vertebrate Paleontology and Paleoanthro pology, Chinese Academy of Sciences, P.O. Box 643, Beijing 100044, China. Larry Martin, Natural History Museum and Department of Ecology and Evolutionary Biology, University of Kansas, Lawrence, Kansas 66045, United States. ornithologists, this theory has been broadly acclaimed among vertebrate paleontologists. Ostrom even went farther, stating that, were it not for the remarkable feather imprints, both of the early Archaeopteryx specimens (London and Berlin) would have been identified unquestionably as coelurosaurian thero- pods (Ostrom, 1976). This argument has been echoed in an ex tensive literature. Less attention has been paid to the significant similarity between Archaeopteryx and modem birds apart from the feathers and claws (Feduccia and Tordoff, 1979; Feduccia, 1993). The wrist and manus bones in Archaeopteryx, when submit ted to detailed analysis and comparison with modem birds, il lustrate many avian skeletal characters that are important to the flight of birds and that were subject to complex morphological change in early avian evolution. ACKNOWLEDGMENTS.?We thank Desui Miao for reading the manuscript, and we benefited from discussion with Christo pher Bennett. We are grateful to Peter Wellnhofer, Burkhard Stephan, Lawrence Witmer, and John Ostrom for their critical reviews and valuable advice and suggestions. We also are in debted to John Chorn for reading the abstract and preparing some of the figures. Character Analyses A total of eight uniquely derived avian characters or charac ter complexes from the wrist and manus skeletons of Archae opteryx are recognized. We have been able to examine the orig- inals or good casts of all seven known specimens of Archaeopteryx, especially the Berlin and Eichstatt specimens. Ostrom listed several theropod dinosaurs as having the closest wrist and manus structure to Archaeopteryx. Among the genera most frequently used in comparisons are Deinonychus, Velocir- aptor, Omitholestes, and Chirostenotes; therefore, our compar isons will focus on the similarities between Archaeopteryx and modem birds on one hand, and the difference between Archae opteryx and these dinosaurs on the other. The homologies of the digits of birds and dinosaurs is still controversial among pa leontologists and embryologists (Hinchliffe, 1985; Martin, 289 290 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 1991). In this paper we use a "1, 2, 3" numbering of the digits in birds, but we have no objection to the "2, 3, 4" identification of embryologists. If the latter scheme is accepted, almost all comparison with dinosaurs disappears. These characters are as follows. 1. The semilunate carpal (Figure 1) is centered on the sec ond metacarpal (see Martin, 1991). In modern birds, this is known to be a single distal carpal (II or III). A similar bone, supposedly homologous to the semilunate carpal in Archaeop teryx, is found in Deinonychus, Sinomithoides, and Velocirap- tor (Ostrom, 1995). In Archaeopteryx (Wellnhofer, 1974), the semilunate carpal is in contact with the first and second metac arpals, but the articulating surface of the second metacarpal is about 2.5 times as long as that of the first one. In contrast, in Deinonychus, Sinomithoides, and Velociraptor the semilunate carpal is articulated almost equally with each of the first two metacarpals. From embryological evidence (Holmgren, 1955), it is known that the semilunate carpal is centered on the second metacarpal in modem birds and that this is clearly an advanced avian character. 2. The third metacarpal slants ventrally toward the distal end as in modem birds (Figure 2), as clearly revealed in the Eich- B A \ 'lfc'%%,,,, FIGURE 2.?Comparisons of the carpometacarpus: A, Archaeopteryx (cast of Berlin specimen in the Natural History Museum of the University of Kansas) in dorsal and slightly posterior view; B, a modern bird, Bubo virginianus (Gmelin), in posterior view to show the similarly ventrally slanting profile of metacarpal III toward the distal end. (2=metacarpal II, 3=metacarpal III.) c D FIGURE 1.?Comparisons of the articulation between the semilunate carpal (black) and the metacarpals: A, Velociraptor mongoliensis Osborn (modified from Ostrom, 1976); B, Deinonychus antirrhopus Ostrom (modified from Ostrom, 1976); C, Archaeopteryx (modified from Wellnhofer, 1974); D, a 19- day-old Struthio camelus Linnaeus (modified from Holmgren, 1955). Draw ings not to scale. (sl=semilunate carpal, 1 =metacarpal I, 2=metacarpal II. start and Berlin specimens of Archaeopteryx. The phalanges of the outer digit also are lower and flatter than those of the mid dle digit. As a result of this, the Eichstatt, Berlin, and Soln- hofen specimens show the third digit (as preserved) crossed by the second digit. This relationship exists in part because the shafts of the feathers ride over the outer phalanges and insert in a fold of skin that forms the edge of the fleshy portion of the wing. The manus of birds is bound together in the postpatagial skin that bears the flight feathers. An impression of the postpat- agium appears to be present on the Berlin specimen and is indi cated in Heilmann's restoration of the wing (Heilmann, 1926, fig. 21). It is not clear whether his restoration of the patagium was based on the specimen or was inferred from modem birds. The fact that the feathers extend onto the digit and are enclosed NUMBER 89 291 in the skin of the patagium makes it impossible for the manus to actually grip objects or act as a prey-capture mechanism. The proximal portion of the third metacarpal is markedly an teroposteriorly compressed and is tightly attached to the poste rior side of the second metacarpal. This character is obviously present in modem birds but is absent in Deinonychus. 3. Four carpals are present in an avian arrangement (Martin, 1991). In the Berlin specimen (de Beer, 1954), there are four preserved carpals, and they are even better displayed in the Eichstatt specimen (Wellnhofer, 1974). Two of them are the ulnare and radiale, which serve to connect the manus with the forearm (Fisher, 1957), and the third (and largest) is the semi lunate carpal. The fourth carpal is relatively small and fuses to metacarpal III (IV?) in later birds (Figure 3). No dinosaurs have been described with these four carpals in an avian ar rangement. The semilunate has a proximal articulating facet for the ulna on the ulnare. The Eichstatt ulnare is better preserved and exposed than in the other specimens. Its tight articulation with a semicircular external condyle on the ulna facilitates the stabilization of the distal portion of the wing. In addition, the third metacarpal does not extend as far proximally as the other two. In Archaeopteryx, proximal to the third metacarpal, there is a small carpal ("x-bone" of Hinchliffe (1985)) that in modem birds fuses with the semilunate carpal to form the proximal end of the carpometacarpus (Figure 3). B LU FIGURE 3.?Comparisons of the wrist pattern: A, Archaeopteryx (a reconstruc tion based on Wellnhofer, 1974); B, a modem bird, Gallus gallus (Linnaeus) (modified from Hogg, 1980). Drawings not to scale. (r=radius, rd=radiale, sl= semilunate carpal, u=ulna, ul = ulnare, x="x-bone" of Hinchliffe (1985), 1 = metacarpal I, 2=metacarpal II, 3=metacarpal III.) 4. The distal metacarpals are simplified. The articulations between the metacarpals and the phalanges are as in modem birds and are different from dinosaurs. The distal end of the first metacarpal is markedly narrower than the proximal end, and the contact between the first and second metacarpals is straight and tightly appressed along its length. Modem birds all have a fused carpometacarpus, and this fusion is clearly a de rived character for birds. The first and second metacarpals di verge distally in Deinonychus (Ostrom, 1976). 5. The second metacarpal is more robust than the other two. This character in Archaeopteryx is related to the support of the feathers provided by this element. In Deinonychus, Sinomi thoides, and Velociraptor the first metacarpal is, on the con trary, more robust than the second one, indicating a totally dif ferent adaptation for the hand. The second digit in Archaeopteryx also is more robust than the other two (Welln hofer, 1988). In Deinonychus and Oviraptor the first digit is more robust than the second one. The first digit of Archaeop teryx is proportionally the same length as that of the juvenile Hoatzin {Opisthocomus) and is well suited to climbing (Heil mann, 1926). In modem birds the first digit is reduced and is never robust, whereas in Deinonychus and Velociraptor the first digit is relatively massive. 6. The proximal end of the first metacarpal is simple and round. This appears to be another avian character unknown in dinosaurs. 7. The first and second phalanges of the second digit form a high, sharp ridge on their dorsal surfaces. This ridge assists in the attachment of the primary feathers and is not known in Deinonychus or Velociraptor. 8. The distal end of the first phalanx of the second digit an teroposteriorly is as wide as, or slightly wider than, the proxi mal end. In theropod dinosaurs such as Deinonychus, Ovirap tor, and Omitholestes, the first phalanx of the second digit is wider proximally than distally. In Archaeopteryx the posterior margin of the distal portion of this phalanx is slightly convex in shape compared with the concave posterior margin in dino saurs. Both of these characters become progressively more ad vanced in Confuciusomis, Cathayomis (Zhou et al., 1992), and modem birds (Figure 4). In modem birds, the distal portion of the first phalanx, together with the proximal portion of the sec ond phalanx, forms a prominently expanded convex posterior margin of the main digit, which provides a combined, solid, bow-shaped support for the primary feathers. We should note that the above-mentioned characters are not functionally independent from each other. They are mostly a result of the morphological requirements of feathered flight. Confuciusomis is the only other bird known with an Archae- opteryx-like morphology in the manus. It also is probably the oldest bird known except Archaeopteryx. All of the above char acters that can be ascertained are present in Confuciusomis, the most notable being characters 1, 4, 6, and 8. Character 3 also appears to be recognizable in the holotype of Confuciusomis, although the wrist area is somewhat cmshed. Conclusions Ostrom (1976) argued that the chief difference between the hands of Archaeopteryx and those of theropods is one of size, all of the theropods being larger. Also, the fingers are relatively 292 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY B D FIGURE 4.?Comparisons of the first phalanx of the second manual digit in dorsal view: A, Ornitholestes her- manni Osborn (modified from Osborn, 1917); B, Deinonychus antirrhopus (modified from Ostrom, 1976); c, Archaeopteryx bavarica Wellnhofer (modified from Wellnhofer, 1993); D, Confuciusomis sanctus Hou et al. (from a cast of the holotype of Confuciusomis); E, Cathayornis yandica Zhou et al. (modified from Zhou et al., 1992); F, Meleagris gallapavo Linnaeus. Drawings not to scale. shorter in the theropods. In addition to the differences men tioned by Ostrom, Archaeopteryx has many advanced avian at tributes of the wrist and hand, with the complex and peculiar pattern of the avian carpometacarpus already formed at this early stage of avian evolution. Without these specializations, the attachment of the primary feathers to the hand would be hardly imaginable, let alone flight. Vazquez (1992) discussed the functional morphology of the avian wrist and stated that Archaeopteryx lacks many of the features of modem birds and was probably incapable of execut ing all the kinematics of modern avian powered flight. Al though we might agree that modem birds have a wrist better designed for powered flight, Archaeopteryx is not so deficit in those features as Vazquez (1992) supposed. As shown by Mar tin (1991), Archaeopteryx has a basically avian wrist, with all of the bones found in modern birds, including an L- or V- shaped cuneiform (ulnare) to glide on the articular ridge of the carpometacarpus. This fact has been missed by most workers, including Martin (1983), because the preservation in Archae opteryx usually shows only the dorsal or ventral side. Fortu nately, the Eichstatt specimen (Wellnhofer, 1974, figs. 8, 9) presents palmar and anconal (ventral and dorsal) views of the same specimen. In palmar view the ulnare is large and elon gate; in anconal view it is small and round. It could therefore be either pyramid-shaped, which would make it unlike the shape of any known relative, including other birds or dino saurs, or L-shaped as in modem birds but not in dinosaurs. We accept the L-shaped interpretation. The perching capability of Archaeopteryx has recently been argued with strong evidence (Feduccia, 1993). The wing claws in Archaeopteryx are long and curved. The first digit diverges from the others (Zhou, 1995). The manual digits are relatively slender. All these characters, in combination, seem to show an overwhelmingly avian pattern and show that the wings could not have been used for predation (Ostrom, 1974). It seems more reasonable to suggest that the oldest bird, although limit ed in flying power, lived an arboreal life just as do most mod em birds, with its wings used for both flight and climbing. The appearance of feathers was the critical point in avian evolution, and the modern appearance of the feathers in Ar chaeopteryx has often been noted (Feduccia and Tordoff, 1979; Norberg, 1995). The close match of the bones of the wrist and manus with modem birds suggests that flight played a vital role in the early evolution of birds. Furthermore, there was coevolu- tion of the skeleton and feathers as two inseparable parts of the flight mechanism. The recognition of many avian characters in Archaeopteryx is important not only for identifying more fragmentary fossils but also for recognizing potential protobirds from even older strata. Because more and more people believe that Archaeop- NUMBER 89 293 teryx is a side branch in avian evolution (Martin, 1983; Feduc cia, 1995), the oldest ancestor of birds might have existed in the Early or Middle Jurassic or even Late Triassic. The recent Chinese finding of a Late Jurassic-Early Cretaceous beaked bird (Hou et al, 1995) seems to lend further credibility to this proposal. Many of the characters discussed above may appear to be subtle, but their importance and evolutionary implication are probably no less than many superficially significant mor phological changes. Although many shared features have been suggested between theropod dinosaurs and Archaeopteryx, they often lack the detailed similarity we should expect in ho mologous characters. Ostrom (1985) recognized only two uniquely avian charac ters in Archaeopteryx: an ossified furcula and feathers. The dis closure of eight uniquely avian characters in the wrist and manus of Archaeopteryx provides further evidence for mosaic evolution in the vertebrate history, and encourages us to exam ine the anatomy of these unique fossils more closely. Literature Cited de Beer, G. 1954. Archaeopteryx lithographica: A Study Based on the British Museum Specimen. 68 pages. London: British Museum (Natural History). Feduccia, A. 1993. Evidence from Claw Geometry Indicating Arboreal Habits of Ar chaeopteryx. Science, 259:790-793. 1995. Explosive Evolution in Tertiary Birds and Mammals. Science, 267:637-638. Feduccia, A., and H.B. Tordoff 1979. Feathers of Archaeopteryx: Asymmetric Vanes Indicate Aerody namic Function. Science, 203:1021-1022. Fisher, H. 1957. Bony Mechanism of Automatic Flexion and Extension in the Pi geon's Wing. Science, 126:446. Heilmann, G. 1926. The Origin of Birds, iii+208 pages. London: H.F. and G. Whitherby. Hinchliffe, J.R. 1985. "One, Two, Three" or "Two, Three, Four": an Embryologist's View of the Homologies of the Digits and Carpus of Modern Birds. In M.K. Hecht, J. K. Ostrom, G. Viohl, and P. Wellnhofer, editors, The Beginning of Birds, pages 141-147. Eichstatt: Freunde des Jura-Mu seums Eichstatt, Willibaldsburg. Hogg, D.A. 1980. A Re-investigation of the Centers of Ossification in the Avian Skel eton at and After Hatching. Journal of Anatomy, 130(4):725-743. Holmgren, N. 1955. Studies on the Phylogeny of Birds. Acta Zoologica, 36:1085. Hou, L., Z. Zhou, L. Martin, and A. Feduccia 1995. A Beaked Bird from the Jurassic of China. Nature, 377:616-618. Martin, L.D. 1983. The Origin and Early Radiation of Birds. In A.H. Bush and G.A. Clark, editors, Perspectives in Ornithology: Essays Presented for the Centennial of the American Ornithologist's Union, pages 291-338. Cambridge: Cambridge University Press. 1991. Mesozoic Birds and the Origin of Birds. In H.-P. Schultze and L. Trueb, editors, Origins of the Higher Groups ofTetrapods, pages 485-540. Ithaca: Cornell University Press. Norberg, R.A. 1995. Feather Asymmetry in Archaeopteryx (Response to J.R. Speakman and S.C Thomson in Vol. 370, p. 514). Nature, 374(6519):221. Osborn, H.F. 1917. Skeletal Adaptations of Ornitholestes, Struthiomimus, Tyrannosau- rus. Bulletin of the American Museum of Natural History, 35: 733-771. Ostrom, J.H. 1974. Archaeopteryx and the Origin of Flight. Quarterly Review of Biol ogy, 49:21-41. 1976. Archaeopteryx and the Origin of Birds. Biological Journal of the Linnean Society, 8:91-182. 1985. The Meaning of Archaeopteryx. In M.K. Hecht, J.K. Ostrom, G. Viohl, and P. Wellnhofer, editors, The Beginning of Birds, pages 161-176. Eichstatt: Freunde des Jura-Museums Eichstatt, Willi baldsburg. Ostrom, J. 1995. Wing Biomechanics and the Origin of Bird Flight. Neues Jahrbuch fiir Geologie und Paldontologie Abhandlungen, 195(1?3):253?266. Vazquez, R.J. 1992. Functional Osteology of the Avian Wrist and the Evolution of Flap ping Flight. Journal of Morphology, 211:259-268. Wellnhofer, P. 1974. Das funfte Skelettexemplar von Archaeopteryx. Palaeontographica, A, 147:169-216. 1988. Ein neues Exemplar von Archaeopteryx. Archaeopteryx, 6:1-30. 1993. Das siebte Exemplar von Archaeopteryx aus den Solnhofener Schichten. Archaeopteryx, 11:1?47. Zhou, Z. 1995. The Discovery of Early Cretaceous Birds in China. In D.S. Peters, editor, Acta Palaeomithologica, 3 Symposium SAPE; 5 Internation ale Senckenberg-Konferenz 22-26 Juni 1992. Courier Forschungs institut Senckenberg, 181:9-22. Zhou, Z., F. Jin, and J. Zhang 1992. Preliminary Report on a Mesozoic Bird from Liaoning, China. Chi nese Science Bulletin, 37:1365-1368. Implantation and Replacement of Bird Teeth Larry D. Martin andJ.D. Stewart ABSTRACT Study of the teeth of the Mesozoic birds Hesperornis, Parahes- perornis, Ichthyornis, Cathayornis, and Archaeopteryx provides new evidence for avian tooth implantation and replacement. Birds share with crocodilians and, to a lesser extent, mammals, a com plex mode of tooth implantation, with deep sockets walled lin- gually by the dentary, maxilla, or premaxilla. These walls crowd the replacing teeth so mat early in ontogeny the teeth migrate labi- ally and continue their development under the crown of their pre decessor. They thus form a vertical tooth family, as opposed to the horizontal tooth family found in dinosaurs and most other tetra- pods. Birds, crocodilians, and mammals have root cementum on their teeth and presumably attach teeth to the socket with peri odontal ligaments. The sockets in mammals and presumably in birds are formed by the outside of the periodontal sac, whereas cementum is deposited by the inside of the sac. Bird teeth are ini tially formed in a groove, and ontogenetically the sockets (in socket-forming species) form first at the front of the jaw. Socket formation then proceeds posteriorly, as in crocodilians. Young dinosaurs have the lingual side of the jaw around the teeth open, so that the roots are exposed. The sockets form around dinosaur teeth as bone of attachment, which is probably the same periodontal bone that forms sockets in mammals, crocodilians, and birds. The sites of new tooth formation extend lingually within the so-called "special foramina" that separate the interdental plates. The inter dental plates represent the surrounding attachment bone and are similar to the attachment bone in pleurodont lizards. In fact, dino saurs might be characterized as having a superpleurodonty that results in sockets. Introduction In our previous paper on avian teeth (Martin et al., 1980), we called attention to numerous features shared by crocodilians and birds but not found in theropod dinosaurs. At that time, we were unaware of how fundamentally different the whole dental Larry D. Martin, Natural History Museum and Department of Ecol ogy and Evolutionary Biology, University of Kansas, Lawrence, Kan sas 66045, United States. J.D. Stewart, Los Angeles County Museum, 900 Exposition Boulevard, Los Angeles, California 90007, United States. system is in crocodilians and dinosaurs and how similar the dentition is in crocodilians and birds. The characteristic tooth morphology of crocodilians and birds includes a flattened, unserrated crown that becomes con stricted as it approaches the crown/root juncture. The tooth is narrow at this point and then expands into a cement-covered root at least as broad as the crown and usually broader. Resorp tion begins as circular to oval pits in the lingual side of the root, and the replacement tooth has most of its formative history be neath the tooth that it will replace (below or above depending on lower or upper dentition). This morphology is found in all of the Triassic and Early Jurassic crocodilians that we have been able to examine. For instance, this tooth form is very clearly shown in acid-prepared specimens from the Liassic (Early Jurassic) marine crocodilian Pelagosaums in The Natu ral History Museum, London, collections. ACKNOWLEDGMENTS.?We are grateful to Alan Charig, Cyril Walker, and Angela Milner of The Natural History Muse um, London (formerly the British Museum (Natural History); BMNH) for access to specimens; P. Wellnhofer (Bayerische Staatssamlung, Munich) generously shared access to speci mens and insights, as did P. Currie (Tyrrell Museum, Drum- heller) and G. Edmund (Royal Ontario Museum, Toronto). C. Bennett and John Chom read the manuscript, and we especially thank Zhonghe Zhou for helpful suggestions. The photography staff at BMNH made the excellent ultraviolet photographs, and the drawings are by Mary Tanner (Craneview Studio, North Platte, Nebraska) and A. Aase (University of Kansas Natural History Museum). Funding was provided by National Science Foundation grant DEB 7821432 and National Geographic So ciety grant 2228-80. Discussion Aside from the nature of the teeth themselves, their mode of implantation in vertebrates also has proven to be useful in working out relationships. The earliest reptiles had acrodont teeth, as are found in the labyrinthodont amphibians and the captorhinomorph reptiles (Figure lA). In the earliest diapsid reptile known (Petrolacosaums), this condition has been modi fied (Reisz, 1981) by the upward (in the lower dentition) exten- 295 296 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY FIGURE 1.?A-E, Cross-sectional diagrams showing relationships of the tooth-bearing bone (unshaded), attachment bone (black), adult tooth (diagonal hatching), and replacement teeth (stippled): A, stem-reptiles (acrodont); B, ear liest diapsid reptile (Petrolacosaurus); c, crocodilians and birds (thecodont with a groove); D, lizards (pleurodont); and E, "thecodonts" and dinosaurs (the codont through superpleurodonty). F, cross section of a dentary of Troodon showing the replacement tooth standing lingual to its predecessor (modified from Currie, 1987, fig. Id). G, medial view of the maxilla of the theropod dino saur Megalosaurus (modified from Charig, 1979:15), showing replacement teeth (rt), interdental groove (ig), and interdental plate (ip). sion of the labial margin of the tooth-bearing bones so that the teeth rest on a shelf (Figure IB). Replacement teeth form lin- gually to the teeth they replace, and this probably inhibits the formation of an inner bony wall. The teeth are attached to the bony shelf by a highly cancellous bone called "bone of attach ment" (Tomes, 1923). Attachment bone also surrounds the de veloping teeth, which are able to migrate labially through it. Edmund (1969:126) briefly discussed attachment bone, saying that it "appears to be embryologically and histologically identi cal with the alveolar bone of mammals and thecodont reptiles. It develops from the dentigenous bone near the base of the new tooth, fuses with it, and is, in turn, resorbed in the process of shedding of the old tooth." It also attaches the roots of the teeth of lizards to the labial side of the jaw, thus forming the pleuro dont condition (Carlson, 1990). In mammals the periodontal sac encloses the developing tooth and deposits bone (cemen tum) around the root using the mesenchymal tissues on its in ner surface. Imbedded in the cementum are collagenous fibers from the periodontal sac, called periodontal ligaments, that also become imbedded in the bone laid down by the external sur face of the sac (Carlson, 1990), which corresponds to bone of attachment. Mammalian teeth have a very limited replacement sequence and have dense, well-formed alveolar bone. Attach ment bone of animals with continuous replacement is constant ly being remodeled. It has the distinct porous (fibrous) mor phology of bone that is subject to rapid resorption and regrowth. The tooth when embedded in the jaw is within the dental sac and thus can be surrounded by attachment bone. The dental lamellae form tooth papillae on the lingual side in an ordered sequence. This is generally tme for all vertebrates, and primi tively much of the formation of a replacement tooth occurs lin gual to the tooth it will replace. The tooth migrates through the easily reconfigured attachment bone and is always internal to the dense bone of the jaw. Because the tooth family is lingual, it is not obstmcted by a labial wall supporting the root. An outer wall helps the teeth to resist outward strain, but stabilizing the teeth to inward pres sure is attained either by fusing the teeth (Figure ID) to the la bial wall (pleurodonty) or by building an inner wall lingual to the tooth row (thecodonty). In the latter case, space for the tooth family must still be provided. Archosaurs evolved two solutions to the problem of building an inner wall. One method was to bring up the inner edge of each tooth-bearing bone to form a groove (Figure lC). In its most primitive stages this groove may not have contained com plete septa, although it seems likely that the groove would have had some constrictions around the teeth. This is essentially the situation that we see in young crocodilians and in young birds. In these archosaurs, the required anteroposterior stabilization of the dentition is provided in part by expanding the roots of the teeth so that they nearly contact one another, and this also gives more surface for periodontal ligaments. The more ad vanced condition is seen in adult crocodilians. Here, lingual and labial projections meet to form septa, and teeth of adult crocodilians tend to have less bulbous roots than do the teeth in juveniles (Martin et al., 1980). An alternate solution is expan sion of the attachment bone until it forms the lingual wall, which is found in several archosaur groups, including dino- NUMBER 89 297 saurs (Figure 1E-G). This is not surprising when we consider that it is merely an elaboration of bone that was already in volved with fixing the teeth to the jaw. In camosaurs, this ex pansion forms structures that have been variously termed "in terdental rugosae" (Osborn, 1912), "interdental plates'' (Madsen, 1976), or "infradental plates" (Gardiner, 1982). In the mandible, these "plates" lie on top of the dentary and are slightly labially inset to the lingual wall. In the upper jaw, they lie beneath the maxilla and premaxilla, slightly labial to the lin gual walls (Figure lF,G). They are generally bounded anteriorly and posteriorly by vertical grooves leading into foramina at the base of the plates. The foramina also are connected by a hori zontal groove on the ledge at the base of the interdental plates. Each foramen is paired to one tooth site and commonly con tains a developing tooth (Figure lG). The grooves and foramina may mark the sites for the dental lamellae, an interpretation that is consistent with their termination at the location of newly deposited tooth crowns. Because the grooves are at the tooth sites of the jaw, the flat attachment bone between them is "in terdental." Interdental plates of this sort occur in most sauris- chians and in many thecodonts (Martin et al., 1980). The only significant variation we have seen in the morphology of inter dental plates is the occasional obliteration of the vertical grooves in presumably older individuals. That the interdental plates are continuous with the interdental septae and distinct from the tooth-bearing bones themselves was observed by Os born (1912) and Walker (1964). Each method of lingual wall formation is accompanied by a characteristic mode of tooth replacement. In fact, in the croco dilian mode of replacement, the new tooth has most of its for mation in the pulp cavity of its predecessor. This mode of re placement also is facilitated by the expanded root and was described by Edmund (1960:114-115) thus: "The crown of a replacement tooth develops within the body of the old tooth, mainly below the neck separating the wider base from the nar rower crown. In this way the diameter of the replacement crown can become greater than that of the crown of the tooth within which it lies." The signature feature of this type of re placement is a pit that completely surrounds the developing re placement tooth (Figure 2E,F), a feature that is absent in all of the many thousands of known dinosaur teeth. Edmund (1969:186) pointed out that saurischian dinosaurs differ from crocodilians in that the replacement tooth did not enter its pre decessor's pulp cavity at an early stage, but seems to have been associated with progressive lingual resorption, with the result ing appearance of having dissolved its way into the lingual wall. The new tooth does not become central in the alveolus until it is about half grown, and much of its predecessor has been resorbed. Frequently a replacement tooth can be seen in the alveolus lingual to its predecessor, the latter being still per fectly functional. From the discussions of Edmund (1960, 1969), and from examination of many saurischian specimens, it is clear that the replacement teeth of saurischians form and continue in an upright position to their maturity. In camosaurs FIGURE 2.?A-D, Teeth of theropod dinosaurs thought by various authors to be especially close to birds: A, Mononykus (modified from Perle et al., 1993); B, Troodon; C, Saurornitholestes; D, Dromaeosaurus (C-D modified from Currie et al., 1990). Teeth showing constricted crown, replacement tooth tip, and expanded base: E, bird, Parahesperornis alexi Martin; F,G, crocodilian, Alliga tor; G, lateral cross section showing the tilted replacement tooth resorbing the root of its predecessor (modified from Edmund, 1962). the replacement teeth form rows on the lingual side of the ma ture tooth, and we have seen as many as three generations of teeth ranked side by side. In crocodilians, however, the replace ment tooth prepares to enter the pulp chamber of its predeces sor by first tilting toward it (Figure 2G). The developing crown then passes in and upward through a circular resorption win dow in its predecessor (Figure 2F). The teeth of crocodilians are attached by periodontal ligaments running from the jaw bones to the root cementum on the expanded roots (Miller, 1968). This mode of attachment has not been recognized in other diapsid reptiles, which also may lack the necessary root cementum. 298 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Bird teeth do contain circular or oval replacement pits that are closed on the bottom, and the developing crown also must tilt as it enters the root of its predecessor (Figures 2E, 3A-D). They also have root cementum (Schmidt and Keil, 1958), and therefore their teeth are attached with periodontal ligaments. Peyer (1968) reported the presence of tooth cementum in crocodiles among modem reptiles and among ichthyosaurs in fossil reptiles, but in ichthyosaurs it is developed only in the geologically younger forms (Peyer, 1968:146). Ichthyosaurs also developed expanded roots, and Peyer (1968:146) suggest ed that this was "to offer the connective tissue fibers an ade quate surface for attachment." It seems likely that in crocodil ians, birds, and ichthyosaurs, the expanded roots are related to a combination of problems resulting from teeth set in a groove. We have not clearly identified the crocodilian pattern of thec- odonty in any other diapsid reptile, but we would expect that the peculiar mode of tooth replacement would accompany it, if it does occur elsewhere. Currie (1987) made an effort to identify bird-like characteris tics in the jaws and teeth of troodontid theropods but figured a cross section (reproduced herein as Figure IF) showing a typi cal dinosaurian lingual replacement pattern and interdental plates (Currie, 1987, figs. 1, 3). The crowns of the teeth are wider than the roots and are widest at the point that they join the roots (not waisted). The teeth are heavily serrated (Figure 2B). In other words, they do not show a single feature thought to characterize bird teeth (Martin et al., 1980). Currie and Zhao (1993) published a drawing of an undetermined dinosaur tooth (?dromaeosaurid) thought to show an oval replacement pit; however, this tooth was an isolated find, was poorly figured, and cannot be relocated. No other dinosaurian taxa with bird-like teeth have been identified in the 20 years since the unique features of bird teeth were first described (Martin et al., 1980). Currie and Zhao (1993:2245) also suggested that be cause bird teeth tend to drift out of the jaws after death, they could not have been attached by cementum. This must be based on a misunderstanding because the decay of the periodental lig aments would release the teeth of birds and young crocodilians, which would be found as relatively intact teeth with roots at tached, as noted by Currie and Zhao (1993). This happens much more rarely with dinosaurs, where the teeth are fixed by attachment bone. It also should be pointed out that theropod di nosaurs have relatively much more room for the lingual tooth family than is found in either birds or crocodilians (contrary to Currie and Zhao, 1993:2245). The ornithomimid dinosaur Mononykus, considered to be a bird by some (Perle et al., 1993), has teeth (Figure 2A) resembling those of the ornitho mimid Pelecanimimus, not those of birds. Elzanowski and Wellnhofer (1996) took the opposite tack by attempting to show that the jaws and teeth of Archaeopteryx are like dinosaurs rather than like other toothed birds. This FIGURE 3.?A-C, Lingual views of the premaxillary and maxillary teeth of Archaeopteryx lithographica von Meyer, London specimen (BMNH 37001): A, left premaxilla and right maxilla; B, maxilla and isolated tooth; c, isolated tooth (from right premaxilla?); D, Parahesperornis alexi, left lower tooth (from the holotype); E, drawing taken from photograph of a tooth of the sev enth specimen of Archaeopteryx (in Wellnhofer, 1993, pl. 6: fig. 3), showing similarity to sockets in the London maxillary; F, right, lingual view of an alli gator maxilla showing similarity of tooth and socket formation to A and B. NUMBER 89 299 would seem unlikely at the outset because the enantiornithine birds of China have Hesperornis-like teeth, almost certainly es tablishing this structure as primitive for birds. Examination of the London specimen of Archaeopteryx shows typical avian structure, with a waisted crown, expanded root, and oval re placement pits (Figure 3A-C). Elzanowski and Wellnhofer (1995:42) deny the presence of expanded roots in Archaeop teryx in spite of the fact that such roots are clearly shown in Wellnhofer (1988, pl. 8: figs. 2, 4; 1993, pl. 6: figs. 1, 3, 5). It also is ironic that the first mention of expanded roots in bird teeth dates back to the first description of teeth in Archaeop teryx by Evans (1865), and this was later confirmed by Ed mund (1960). There is an excellent oblique photograph (Welln hofer, 1993, pl. 5: fig. 9) of the seventh specimen of Archaeopteryx showing the waisted crown and expanded root typical of other birds and the so-called interdental plates cross ing as tooth septa. Although the features of bird teeth may be most easily seen in large isolated specimens from Hesperornis and Parahesperornis, it is clear that all known bird teeth are closely similar to each other and that the teeth of Archaeop teryx are not atypical or especially primitive. The claim for in terdental plates (Elzanowski and Wellnhofer, 1996) in the sev enth skeletal specimen is based on the labial margin of the jaw being higher than the lingual margin and exposing distinct al veolar septa (interdental plates?) between the tooth sites. In fact, this condition is better displayed in the London maxillary (Howgate, 1984), where it has been interpreted as a tooth sock et (Martin, 1991). Because the alveolar bone of the sockets and the attachment bone of the interdental plates are ultimately de rived from the same source, we must carefully describe what is meant by interdental plates versus sockets. When interdental plates are present, they generally expose the replacement teeth, and most of the length of the root is surrounded on the lingual side by the interdental plates. The jaws of the seventh specimen of Archaeopteryx are spread and compressed so that we get a slightly oblique view of the jaws on the slab (see Wellnhofer, 1993, pl. 4: fig. 1). We are indebted to Wellnhofer (1993) for excellent photographs that clearly show the crown-root junc ture on the teeth of the seventh example and show that the con striction at the base of the crown is almost at the lingual edge of the mandible (Figure 3E), so that the replacement teeth are mostly hidden by the side of the dentary as in other birds. When we look at typical interdental plates (Figure lG), we see that not only the socket but also part of lingual side of the jaw is produced by the interdental plates, and that there is a dis tinct groove separating the individual plates, terminating in the "special foramina" and the replacing teeth. The replacing teeth lie to the lingual side of the adult tooth as shown by the replac ing tooth in Troodon (Figure IF). These are not the relation ships shown in the London Archaeopteryx (Figure 3A-C) or in the seventh specimen (Figure 3E). The intersepta of the tooth sockets of the London maxilla closely resemble a similar view of an alligator maxilla (Figure 3F), as well as the sockets of Ar- chaeopteryx (Figure 3E) so well photographed by Wellnhofer (1993). A close examination of Wellnhofer's photographs also shows the intersepta widening again as they come to the labial edge, as expected in a dorsal view of the socket. If the view were entirely medial, we would not expect to see this widening, even if these were interdental plates (see Figure lG). The condition in coelurosaurs is not as clear as it is in camo saurs. Compsognathus is reported by Ostrom (1978) to have small interdental plates. Dromaeosaurs were not thought to have interdental plates (Colbert and Russell, 1969). According to Currie (1995), dromaeosaurs have fairly typical interdental plates forming much of the lingual side of the jaw below the sockets except that the grooves fuse across, forming a solid wall. This should indicate a modified tooth replacement, and indeed it appears that the tooth family may be thrown into diag onal lines so that replacing teeth are both lingual and posterior, as, for example, in the overlapping replacement tooth in the ra mus of Deinonychus illustrated by Ostrom (1969). This is not bird-like, nor is the covering of the interdental plates by special bones (supradentary) in Allosaurus and Tyrannosaurus (inter- coronoid of Brown and Schlaikjer, 1940). It is clear that coelu- rosaur teeth are very similar to the teeth of camosaurs and do not show the specialized type of tooth replacement found in birds (Figure 2A-D). Conclusion The argument that birds are related to dinosaurs is now most often restated that birds are dinosaurs. If this is the case, we would expect their anatomical structures to maintain similarity under a very rigorous analysis. We see that this is not tme for almost any aspect of tooth form, implantation, or replacement. The tooth structures identified as interdental plates in Archae opteryx by Wellnhofer (1993) do not agree in detail with those structures in dinosaurs and can be closely duplicated by croco dilians. We now know from the abundant Chinese enantiorni thine material that the tooth form of birds is similar in all known groups of birds and must have been established at least by the Jurassic. Crocodilians and birds form the inside walls of their tooth-bearing bones differently from dinosaurs and have a different mode of tooth replacement. Their common ancestor with dinosaurs may not have been "thecodont" in the descrip tive sense of that word. Crocodilians have derived features that prevent them from being ancestral to birds, but a sister-group relationship is still possible. 300 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Literature Cited Brown, B., and E.M. Schlaikjer 1940. A New Element in the Ceratopsian Jaw with Additional Notes on the Mandible. American Museum Novitates, 1092:1-13. Carlson, S.J. 1990. Vertebrate Dental Structures. In Joseph G. Carter, editor, Skeletal Biomineralization: Pattern, Processes and Evolutionary Trends, 1:531-556. New York: Van Nostrand Reinhold. Charig, A. 1979. A New Look at the Dinosaurs. 160 pages. London: Heinemann in as sociation with the British Museum (Natural History). Colbert, E.H., and D.A. Russell 1969. The Small Cretaceous Dinosaur Dromaeosaurus. American Museum Novitates, 2380:1^19. Currie, P.J. 1987. Bird-Like Characteristics of the Jaws and Teeth of Troodontid Theropods (Dinosauria, Saurischia). Journal of Vertebrate Paleon tology, 7(1): 72-81. 1995. New Information on the Anatomy and Relationships of Dromaeo saurus albertensis (Dinosauria: Theropoda). Journal of Vertebrate Paleontology, 15(3):574-591. Currie, P.J., J.K. Rigby, Jr., and R.E. Sloan 1990. Theropod Teeth from the Judith River Formation of Southern Alberta, Canada. In K. Carpenter and P.J. Currie, editors, Dinosaur System atics: Approaches and Perspectives. 8:107-125. Cambridge: Cam bridge University Press. Currie, P.J., and X. Zhao 1993. A New Troodontid (Dinosauria, Theropoda) Braincase from the Di nosaur Park Formation (Campanian) of Alberta. Canadian Journal of Earth Sciences, 30:2231-2247. Edmund, A.G. 1960. Tooth Replacement Phenomena in the Lower Vertebrates. Journal of Vertebrate Paleontology, 52:1-190. 1962. Sequence and Rate of Tooth Replacement in the Crocodilia. Journal of Vertebrate Paleontology, 56:1?42. 1969. Dentition. In C. Gans, A.d'A. Bellairs, and T.S. Parsons, editors, Bi ology of the Reptilia, pages 117-200. London: Academic Press. Elzanowski, A., and P. Wellnhofer 1995. The Skull of Archaeopteryx and the Origin of Birds. Archaeopteryx, 13:41^16. 1996. Cranial Morphology of Archaeopteryx: Evidence from the Seventh Skeleton. Journal of Vertebrate Paleontology, 16( 1 ):81-94. Evans, J. 1865. On Portions of a Cranium and of a Jaw, in a Slab Containing the Fossil Remains of the Archaeopteryx. Natural History Review, new series, 5:415-421. Gardiner, B. 1982. Tetrapod Classification. Journal of Vertebrate Paleontology, 74:207-232. Howgate, M.E. 1984. The Teeth of Archaeopteryx and a Reinterpretation of the Eichstatt Specimen. Journal of Vertebrate Paleontology, 82:152-175. Madsen, J.H., Jr. 1976. Allosaurus fragilis: A Revised Osteology. Utah Geological and Mineral Survey Bulletin, 109:1-163. Martin, L.D. 1991. Mesozoic Birds and the Origin of Birds. In H.P. Schultze and L. Trueb, editors, Origin of Higher Groups ofTetrapods, pages 480-540. Ithaca: Cornell Press. Martin, L.D., J.D. Stewart, and K. Whetstone 1980. The Origin of Birds: Structure of the Tarsus and Teeth. Auk, 97: 86-93. Miller, W.A. 1968. Periodontal Attachment Apparatus in the Young Caiman sclerops. Archives of Oral Biology, 13:735-748. Osborn, H.F. 1912. Crania of Tyrannosaurus and Allosaurus. Bulletin of the American Museum of Natural History, new series, 1:1-30. Ostrom, J.H. 1969. Osteology of Deinonychus antirrhopus, an Unusual Theropod from the Lower Cretaceous of Montana. Bulletin of the Peabody Museum of Natural History, 30:1-165. 1978. The Osteology of Compsognathus longipes Wagner. Zitteliana, 4: 73-118. Perle, A., M.A. Norell, L.M. Chiappe, and J.M. Clark 1993. Flightless Bird from the Cretaceous of Mongolia. Nature, 362: 623-626. Peyer, B. 1968. Comparative Odontology, xvi+348 pages. Translated and edited by R. Zangerl. Chicago: The University of Chicago Press. Reisz, R.R. 1981. A Diapsid Reptile from the Pennsylvanian of Kansas. Special Publi cation, University of Kansas Museum of Natural History, 7:72-74. Schmidt, W.J. and A. Keil 1958. Die Gesunden und die erkrankten Zahngewebe des Menschen und der Wirbeltiere in Polarisationsmikroskop. 386 pages. Miinchen: Carl Hanser Verlag. Tomes, CS. 1923. A Manual of Dental Anatomy, Human and Comaprative. Eighth edi tion, 616 pages. New York: Macmillan Co. Walker, A.D. 1964. Triassic Reptiles from the Elgin Area: Omithosuchus and the Origin of Camosaurs. Philosophical Transaction of the Royal Society of London, series B, 2478:53-134. Wellnhofer, P. 1988. Ein neues Exemplar von Archaeopteryx. Archaeopteryx, 6:1-30. 1993. Das siebte Exemplar von Archaeopteryx aus den Solnhofener Schichten. Archaeopteryx, 11:1-48. Humeral Rotation and Wrist Supination: Important Functional Complex for the Evolution of Powered Flight in Birds? John H. Ostrom, Samuel O. Poore, and G.E. Goslow, Jr. ABSTRACT To achieve a better understanding of the function of the M. supracoracoideus in extant birds, we measured the mechanical properties and actions of the supracoracoideus in the European Starling (Sturnus vulgaris Linnaeus) and a pigeon (Columba livia Gmelin). We performed three sets of acute in situ experiments by direct nerve stimulation. We measured length-active and length-passive tension, forces of humeral elevation and rotation (torque), and humeral excursion (elevation and rotation). The supracoracoideus is capable of generating a tetanic force 7-10 times the bird's body weight, imparts a torque about the longitudi nal axis that is greater than its force of humeral elevation, and, when tetanically stimulated, elevates the humems a limited 50?-60? above the horizontal but rotates it through 80?. We con clude that the primary role of the supracoracoideus is high-veloc ity rotation of the humerus, a movement critical to achieving the upstroke portion of the wingbeat cycle. In addition, we propose that high-velocity humeral rotation may also serve to augment supination of the wrist during upstroke. A morphologically derived supracoracoideus to produce rapid humeral rotation and the skeletal features associated with it, an acrocoracoid, triosseal canal, and tuberculum dorsale, are not evi dent in Archaeopteryx or Sinornis. These features also appear undeveloped in Iberomesornis and Concornis, dirt unknown in Cathayornis, and apparently are not preserved in the most recent find, Confuciusomis. Introduction In order for a flapping wing stroke to be effective, the wing surface must be converted from an aerofoil to a "nonaerofoil" surface as the wing changes from the powered downstroke to JohnH. Ostrom, Division of Vertebrate Paleontology, 170 Whitney Av enue, P.O. Box 6666, New Haven, Connecticut 06511, United States. Samuel O. Poore and G.E. Goslow, Jr., Brown University, Department of Ecology and Evolutionary Biology, Box G-BMC 204, Providence, Rhode Island 02912, United States. the recovery upstroke, either powered or unpowered. The wing of birds is pronated and extended during the downstroke to pro vide maximum surface area for both lift and thrust, but it is su- pinated during the recovery upstroke so as to minimize the sur face area and thereby reduce drag. The focus of this contribution centers around the M. supracoracoideus and its role in modem birds for augmenting supination at the wrist. These data provide a new perspective on the fossil record of birds. The statement written twenty years ago that "any consider ation of the evolution of flight must start with Archaeopteryx" (Ostrom, 1976a:3) is more important today than when it was written. This is because more is now known about the anatomi cal details of Archaeopteryx, especially after publication of the last three finds (Wellnhofer, 1974, 1988, 1992, 1993), and be cause of advances in our understanding of bird flight mechanics (Gauthier and Padian, 1985; Ostrom, 1986, 1994, 1995; Jenkins et al., 1988; Rayner, 1988a; Dial et al., 1991; Vazquez, 1992; Pennycuick, 1993). This fact is brought home in a most com pelling manner when the now seven nearly complete, articulat ed specimens of Archaeopteryx are compared with the often in complete, solitary Mesozoic bird specimens {Iberomesornis, Sinornis, Cathayornis, Concornis, Otogornis, Confuciusomis, and others) that have been reported since the Eichstatt specimen was recognized. Recognition in the Eichstatt specimen of the maniraptoran-like semilunate carpal (Ostrom, 1976a, 1976b) resulted in a careful reevaluation of the hypothesis of the thero pod origin of birds (Hecht et al., 1985; Schultze and Tmeb, 1991) and the origins of flight in birds (Padian, 1986; Gauthier and Padian, 1989; Bock and Buhler, 1995). These arguments aside, however, from a functional standpoint this same semilu nate carpal was central to the maintenance of pronation during downstroke and to the execution of supination during upstroke. After meticulous investigations of the morphology of the avian carpal-metacarpal complex and functional morphology of the pigeon carpometacarpus, Vazquez (1992, 1995) demon strated that the articular surface of the trochlea carpalis in mod em birds acts to automatically supinate the hand upon wrist 301 302 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY flexion. Ostrom's analysis also can be extended to address the presence of the semilunate carpal, the precursor of the trochlea carpalis, in Archaeopteryx; it too may have served for automat ic supination of the hand and metacarpus (Vazquez, 1992). Such supination undoubtedly served to streamline the distal portion of the wing during the upstroke, an action necessary to reduce profile drag (Rayner, 1988b). Our experimental evi dence concerning the action of the supracoracoideus in pow ered flight reveals this muscle's potential for increasing the rate, and perhaps extent, of supination during upstroke. These findings underscore the functional importance of a morpholog ically derived supracoracoideus with a dorsally directed tendon in modem birds. The implications of these findings have been reported (Poore et al., 1997). ACKNOWLEDGMENTS.?We gratefully acknowledge the as sistance of our good colleagues Peter Wellnhofer (Bayerische Staatssammlung fur Palaontologie und historische Geologie, Munich), Gunter Viohl (Jura-Museum, Willibaldsburg, Eich statt), Jose Sanz (Universidad Autonoma de Madrid), and Her man Jaeger (Humboldt Museum fur Naturkunde, Berlin) for providing access to the specimens in their charge and for all their hospitality and generosity. In fond memory, we dedicated this "exploration" to our dear colleague Herman, without whom most of this would not have been possible. For technical assistance with the experimental procedures, we thank Maki Morimoto and Amie Valore. We thank Kathy Brown-Wing (Brown University, Providence, Rhode Island) for preparation of the illustrations used in Figures 3 and 4. This work was sup ported in part by National Science Foundation grant IBN 9220097 to one of us (GEG). Wrist Anatomy THE MANIRAPTORAN WRIST Important to this discussion of supination of the hand in modern birds is a review of the evolution of the wrist in maniraptoran theropods and Archaeopteryx. Unknown before 1964, the semilunate carpal of theropods was first reported in 1969 (Ostrom, 1969a), and a detailed, functional interpretation followed a few months later (Ostrom, 1969b). That analysis has apparently been accepted. It was shown therein that the semilunate carpal element of Deinonychus (and later of Ve lociraptor and other maniraptoran taxa) articulated in a tight, rigid union with the first and second metacarpals distally. The opposite proximal surface was a well-finished trochoidal artic ular surface that permitted a high degree of flexion-extension with the ulna (Figure 1). Because of its highly canted or pro nounced asymmetrical shape on the proximal surface, the semi lunate carpal also forced the metacarpus to supinate (circum duct) up to 45? as the wrist was flexed. At the carpal joint, supination must have been just as important as flexion because the articular facet was well formed and highly finished in all of the specimens in which it was found. That particular kind of wrist motion was believed at the time (1969) to have been an FIGURE 1.?Key components of a maniraptoran theropod (Deinonychus anti- rrhopus Ostrom) forearm to illustrate hypothesized action imposed by the semilunate carpal (key carpal) during flexion-extension: A, metacarpals I and II and their relationship to the key carpal; B, the asymmetrical proximal gingly- mus of the key carpal causes supination (circumduction) of the closely adjoined metacarpus through approximately 45?; c, surface aspects of the key carpal: l=distal articular surface, 2=proximal articular surface, 3 = lateral sur face. (Modified from Ostrom, 1969b.) important part of the predator action of Deinonychus, the first taxon in which it had been found. Subsequently, when recog nized in other maniraptoran specimens {Velociraptor, Stenony- chosaurus, and Sinomithoides), it apparently was presumed to have served a similar raptorial role related to the function of the manus. Although the biomechanical actions that were produced by this particular wrist appear quite obvious, exactly what biologi cal role these movements played in maniraptoran life is not so NUMBER 89 303 apparent. It is now clear that this semilunate carpal is character istic of the clade, but the adaptive meaning of these clade-com- mon unique wrist movements that seem to have been typical of maniraptoran theropods remains unknown. THE WRIST OF Archaeopteryx Revelation of the remarkable details of the anatomy pre served in the Eichstatt specimen (the fifth) of Archaeopteryx (Wellnhofer, 1974) caused renewed interest in the question of the origin of birds, culminating in the 1984 International Ar chaeopteryx Conference (Dodson, 1985) in Eichstatt, Germany (Hecht et al., 1985). Although not unanimous, that conference reached a consensus that the ancestral stock from which birds arose was probably a primitive archosaurian, but the details of this origin soon dissolved into three distinctly different hypoth eses that persist to this day: the primitive thecodontian theory (Hecht and Tarsitano being the principal advocates), the cro- codylomorph theory (championed effectively by Walker, Mar tin, and Whetstone), and the theropod ancestral theory (argued by Ostrom, Padian, and Wellnhofer, sometimes by Gauthier, and occasionally by others). The most important evidence provided by the Eichstatt spec imen is the well-preserved semilunate carpal almost exactly as it is preserved in Velociraptor and Sinomithoides from Mongo lia and China, respectively (Figure 2). The semilunate carpal appears to have functioned in the same way in these forms, just as originally visualized in Deinonychus (Ostrom, 1969b); flex ion at the wrist forced a pronounced supination (circumduc tion) of the metacarpus-manus. If this interpretation also is cor rect for Archaeopteryx, as we believe, that carpal manipulation has profound implications regarding the flight capability of the Urvogel and has even stronger implications concerning the ear ly stages of flapping flight in birds. Because of this apparent wrist action in Archaeopteryx, supi nation was initially equated with the modem avian wrist (Os trom, 1976a), where the shape of the modem carpometacarpus is so similar to that formed by the semilunate carpus-metacar pus complex of the Eichstatt specimen. In fact, the trochlea car- palis of the modem carpometacarpus forms the key articulation essential for modem flapping bird flight (Vazquez, 1992). It was proposed that the maniraptoran-like semilunate carpal, through time, fused with metacarpals I and II to form the mod em carpometacarpus with its distinctive trochlea carpalis and its unique action so characteristic of all flying birds (Ostrom, 1976a). As Vazquez (1992) described, flexion of the wrist of the modem avian wing forces the more distal wing segments to su pinate, streamlining those wing components for the ensuing up stroke. Flexion at the wrist displaces the cuneiform distally, causing it to slide along the trochlea carpalis, which results in supination, although there are no muscles that directly supinate the hand (Vazquez, 1995, and references therein). Thus, supi nation is dependent on the trochlea carpalis-cuneiform com plex and air resistance on the dorsal surface of the wing during upstroke. In addition to the wrist's osteology, we propose here in that a derived supracoracoideus contributes to supination by rapidly rotating the humems on its longitudinal axis. Below we report the experimental evidence to support this position. The Role of the M. Supracoracoideus in Flapping Flight Numerous derived features characterize the pectoral girdle and associated musculature of the Neornithes. The most strik ing of these, and the one that represents an extreme departure from a primitive tetrapod organization, is that of the M. supra coracoideus. The supracoracoideus in all birds possessing powered flapping flight lies deep to the pectoralis, arises from the carina, sternum, and coracoclavicular membrane, and pos sesses a bipinnate architectural organization of its fascicles. The most distinctive feature of the supracoracoideus, however, is the course of its tendon of insertion (Figure 3). The tendon passes dorsally through the triosseal canal (formed by the cora coid, scapula, and furcula) and attaches on the dorsal aspect of the humems above the glenohumeral joint. The seemingly ob vious function of this dorsally inserting tendon is that the su pracoracoideus is for wing elevation. The presence or absence of this anatomical arrangement has been a central question in debates concerning the evolution of flapping flight and has been given considerable attention in interpreting the flight ca pabilities of the Late Jurassic bird Archaeopteryx (Ostrom, 1976a, 1976b; Olson and Feduccia, 1979). We studied the in situ contractile properties of the supracora coideus to clarify its role during flapping flight in two species of extant birds, the European Starling {Sturnus vulgaris Lin naeus) and a pigeon {Columba livia Gmelin). Starlings and pi geons contrast in their wing loading (wing area/body weight) and flight styles. In both species, we measured the absolute force generated by the supracoracoideus, the humeral excur sion (elevation and rotation), and the forces of humeral eleva tion and humeral axial rotation. Electrical activity of the supracoracoideus of a starling flying in a wind tunnel (Dial et al., 1991) and pigeons in free flight (Dial et al., 1988) begins in late downstroke and ends prior to the upstroke-downstroke transition. The electrically active pe riod is not coincident in time with force. The electromechanical delay reported in the pectoralis during flight in starlings (Bie- wener et al., 1992) and pigeons (Dial and Biewener, 1993) sug gests electrical activity anticipates force at burst onset by sev eral milliseconds (ms). After electrical activity ceases, however, force continues for 20-25 ms, leading us to conclude the force produced by the supracoracoideus in both species is sustained through most of the upstroke. We used as a reference for our physiological measurements the wing kinematics for European Starlings reported in the cineradiographic study by Dial et al. (1991). Kinematic data of comparable precision are not available for the pigeon; we made estimates from Brown (1951) and Simpson (1983). The downstroke-upstroke transi- 304 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY :m& FIGURE 2.?Articular joints of the wrist: A, Archaeopteryx lithographica von Meyer (Eichstatt specimen), scale units=0.5 mm; B, Deinonychus antirrhopus, scale bar=30 mm; c, Velociraptor mongoliensis Osborn, image -half size. Arrows indicate the semilunate carpal in each. (After Ostrom, 1995.) tion in both species begins with the humems below the hori- eluding retraction of the humems, elevation, and rapid flexion zontal and is characterized by a rapid sequence of events, in- at the wrist and elbow. NUMBER 89 305 FIGURE 3.?Dorsal view of the right shoulder to illustrate the emergence of the supracoracoideus tendon (arrow) through the triosseal canal to the tuberculum dorsale (external tuberosity) and its favorable angle for humeral rotation in four birds with different flight characteristics. A, European Starling (Sturnus vulgaris), a representa tive passeriform. During upstroke, the angle of retraction becomes more acute, increasing the ability of the ten don of insertion to rotate the humerus on its longitudinal axis. A sesamoid bone, the Os humeroscapularis, serves to deflect the tendon to maintain optimality for humeral rotation. B, pigeon (Columba livia), a representative columbiform specialized for vertical ascent and descent. The ability of this species to take off from a flat surface is dependent on the supracoracoideus. c, Great Black-backed Gull (Larus marinus Linnaeus), a charadriiform that soars and uses relatively low amplitude wingbeats during flapping flight. Note the relatively large tubercu lum dorsale. A decreased angle of retraction coupled with a large tuberculum dorsale results in relatively high torque. D, Atlantic Puffin (Fratercula arctica (Linnaeus)), a charadiiform specialized for wing-propelled diving. The supracoracoideus in this and other wing-propelled divers is relatively large to rotate the wing under water. Each scale bar= 1 cm. All experiments in this study were performed in accordance with National Institutes of Health guidelines for animal re search and were approved by the Institutional Animal Care and Use Committee of Brown University. We measured the me- 306 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY chanical properties and actions of the supracoracoideus in eight acute in situ experiments for each species by direct nerve stim ulation. All experiments were performed following anesthesia with ketamine (60 mg/kg) and xylazine (6 mg/kg); supplemen tal ketamine was given as needed. We bisected the lattisimus dorsi and rhomboideus muscles to expose the brachial plexus and isolate the nerve to the supracoracoideus. We intubated the birds unidirectionally via the trachea (80% oxygen, 20% nitro gen) after opening the posterior air-sacs. We severed all com ponents of the brachial plexus except the nerve to the supraco racoideus to prevent stimulation of adjacent muscles. Following surgical preparation we clamped the sternum and coracoid to a rigid frame and maintained body temperature at 40? C with warmed avian ringers and a heat lamp. We mounted the supracoracoideus nerve on silver bipolar electrodes and es tablished a stimulation voltage (2x threshold) to elicit a twitch or a tetanus. For four birds of each species, we measured maxi mal tetanic tension by connecting the tendon of the supracora coideus directly to a force transducer. ROTATION AND ELEVATION.?We made independent mea surements of the rotational force (torque) about the longitudi nal axis of the humems and of the force of elevation on the hu mems during isometric contraction of the supracoracoideus for two birds of each species. To measure torque, we threaded a short piece of silver wire (0.38 mm diameter) through a small hole drilled in the deltopectoral crest, attached the wire to the force transducer, and measured isometric force at that point. We placed a 23-gauge pin in the shaft of the humems to pre vent elevation while still permitting "free" rotation about the bone's longitudinal axis. To measure elevational force, we se cured the humerus to the transducer with surgical silk. We stimulated the supracoracoideus nerve tetanically with the hu mems positioned at joint angles of elevation/depression and protraction/retraction coincident with the downstroke-upstroke transition and midupstroke of flight. EXCURSION OF THE HUMERUS.?We measured the total in situ elevation excursions of the humems during tetanus of the supracoracoideus for two birds of each species. During these measurements, the humems was not restricted in any way but was allowed to move during stimulation. We stimulated the nerve tetanically (60 hertz; 500 ms train duration) and mea sured elevation of the humems with a protractor. We made all elevational measurements relative to the dorsal border of the scapula in lateral view. Subsequent to the elevation measure ments, we measured rotation by placing a 23-gauge pin guided by a rack and pinion through a small hole drilled in the distal end of the humems. We threaded the needle into the long axis of the humeral shaft, which served as a pivot for rotation while restricting the elevational component of movement. We placed a 26-gauge pin perpendicular to the long axis of the humems, which served as a dial with which to measure the degree of ro tation. We made measurements at the two wing positions noted earlier; the downstroke-upstroke transition and midupstroke. Discussion The downstroke-upstroke transition in both species begins with the humems depressed below the horizontal (10? for star ling, estimated 10? for pigeon). The angle formed by the long axis of the humems and the vertebral column in dorsal view at the downstroke-upstroke transition is about 55?-60? in both species. Upstroke commences by retraction, rotation, and ele vation of the humems, flexion of the elbow, and flexion/supi- nation of the wrist. During upstroke, the right humems rotates counterclockwise about its longitudinal axis and elevates about 40? above the horizontal (Figure 4). During muscle shortening the potential for active force production decreases as the hu mems is rotated and the wing is elevated. Nevertheless, at hu meral angles corresponding to the downstroke-upstroke transi tion, we measured tetanic forces of 6.5 ?1.2 newtons (N) in the starling (H=3) and 39.4 ?6.2 N in the pigeon {n=6); forces 8 times or more the body weight of each species. The supracora coideus imparted an average isometric force for rotation mea sured at the deltopectoral crest for the starling of 4.9 N (down stroke-upstroke transition) and for the pigeon of 32.1 N. The forces at the midupstroke positions were about half of these values. Although we measured in situ humeral rotations of up to 80?, maximum elevations of the humems were only about 55? above the horizontal. From these data we conclude the pri mary action of the supracoracoideus to be high-velocity rota tion of the humems about its longitudinal axis during wing up stroke; active wing elevation may be of secondary importance. Further support for this conclusion comes from an analysis of the glenoid and the anatomical arrangement of the avian suprac oracoideus. The avian shoulder joint is structurally derived and functionally complex. The glenoid, best described as a hemisel- lar (half-saddle) joint, faces dorsolateral^ and articulates with a bulbous humeral head. Jenkins (1993) reviewed the structural/ functional evolution of this joint and provided an interpretation of its function based on a cineradiographic analysis of the wing- beat cycle. His study illustrated the articulation of the humeral head on a dorsally facing surface of the glenoid, the labmm cavitatis glenoidalis, which allows for full abduction of the wing into the parasagittal plane at the upstroke-downstroke transition. We believe full abduction is not so much by eleva tion of the humems but by rotation about its longitudinal axis. It bears emphasis that during the wingbeat cycle of European Starlings flying in a wind tunnel, where we have precise cinera diographic data, the angle formed by the long axis of the hu mems and the vertebral column is never greater than 55? (Jen kins et al., 1988, fig. 1; Dial et al., 1991, fig. 4). We have made in situ measurements of humeral protraction/retraction in an aesthetized, intact starlings and pigeons. The humems cannot be drawn forward to intersect the body axis at an angle greater than 60?-65? unless forced; its forward angle beyond these an gles is constrained by the ligaments and muscles surrounding the shoulder. The mechanics of the musculoskeletal organization of the su pracoracoideus also supports our conclusion. The supracoracoi deus in both pigeons and starlings, as well as in all other species NUMBER 89 307 we examined (see Figure 3), is a bipinnate muscle with relative ly short but numerous fascicles. This architecture is favorable for high forces and limited excursion. Additionally, the tendon of insertion inserts circumferentialy on the long axis of the hu mems, further contributing to the role of the supracoracoideus as a humeral rotator. The moment arm of the tendon of inser tion in both species is short; we estimate its maximum to be 2 mm in the starling and 4 mm in the pigeon. Although the me chanical advantage of the supracoracoideus is low, its high in put force, particularly at the downstroke-upstroke transition, is favorable for the production of high-velocity movements at the distal portion of the wing. We predict that during the upstroke, the distal portion of the wing experiences extremely high rota tion velocity. Although still to be determined, these rotary forc es may act to augment supination at the wrist in addition to su pination provided by the trochlea carpalis-cuneiform complex. FIGURE 4.?Wing of the European Starling (Sturnus vulgaris Linnaeus) dur ing upstroke in frontal view (after Dial et al., 1991) at the down stroke-upstroke transition (A), midupstroke (B), and upstroke position (c) at maximum humeral rotation and elevation. At the downstroke-upstroke posi tion, the humerus is depressed 10? below the horizontal and the hand is pronated. During upstroke, the humerus rotates 80? on its longitudinal axis and elevates 55? above the horizontal, and the hand is fully supinated. Scale bar=l cm. THE FOSSIL EVIDENCE In view of the evidence for the role of the supracoracoideus during the wingbeat cycle in modem birds, the obvious ques tion before us is, when did the supination/humeral rotation ac tion of the supracoracoideus come into play? Rotation of the humems by the supracoracoideus is enhanced by the leverage provided by the tuberculum dorsale (external tuberosity), the derived site of insertion of the supracoracoideus (Figure 3). Re orientation of the supracoracoideus to this insertion on the hu mems is accomplished by the passage of the tendon through the triosseal canal and around the acrocoracoid. At what point in the fossil record can we recognize any of these features? None of these features have been noted in any of the speci mens of Archaeopteryx (Figure 5). As reported by Sereno and ARCHAEOPTERYX Deltopectoral Crest -????^?? External Tuberosity Head Deltopectoral Ectepicondyle Bicipital Crest Internal Tuberosity CATHARTES FIGURE 5.?Comparison of the humeri of Archaeopteryx and a modem flying bird, the Turkey Vulture (Cathartes aura (Linnaeus)), in dorsal aspect. Humeri are drawn to unit length for easy comparison; each scale bar equals 3 cm (external tuberosity=tuberculum dorsale). The humerus of Archaeopteryx is devoid of most of the tubercles and crests that are well developed in most mod em birds. Most of these features are the attachment sites of muscles that retract and rotate the humerus. (After Ostrom, 1976a.) 308 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Rao (1992), the critical region of Sinornis is not preserved, so we have no evidence. Zhou and Zhang (1992) did not note ei ther feature in any of the material of Cathayornis, and the spec imens of Iberomesornis (Sanz et al., 1988; Sanz and Bonaparte, 1992) and Concornis (Sanz et al., 1988; Sanz et al., 1995) do not show them either (confirmed by JHO). The precise strati graphic source of the newly reported Confuciusomis is in doubt, but Hou (1995; see also Hou et al., 1995) made no men tion of either feature. EARLY STAGES OF FLIGHT Sy (1936) described humeral axial rotation as a mechanism for the execution of wing upstroke and downstroke in pigeons and generalized its importance for other relatively small birds possessing powered flight. His observations that pigeons with bilateral tenotomy of the supracoracoideus are capable of flight, but cannot take off from the ground, are often cited in discussions of the evolution of powered flight (Olson and Fe duccia, 1979; Ostrom, 1976b; Ruben, 1991). Perhaps less ap preciated was Sy's (1936) identical procedure on at least one adult crow, for which he reported not only normal takeoff but normal flight. Sokoloff et al. (1994), in a cinematographic/ electromyographic analysis of adult starlings, bilaterally dener- vated {n?4) or tenotomized (?=2) the supracoracoideus and re ported that all birds but one could take off, but not without dif ficulty. Of particular importance are their observations that takeoff and flight in these deprived birds is not normal. In our estimation, the extent of impairment incurred by loss of the su pracoracoideus for different species is a function of wing load ing (body weight/wing area), the mechanical organization of the supracoracoideus, or some combination thereof. We be lieve the impaired takeoff capability of birds deprived of a functional supracoracoideus relates to their inability to rapidly rotate the humems on its axis. The earliest unequivocal evidence pertaining to bird flight is that of Archaeopteryx. There is no debate concerning its strati graphic age. There is no question about its avian affinities. The famed feather impressions on most of the seven specimens es tablish that if Archaeopteryx flew, a feathered airfoil was avail able. The apparent presence of an ossified sternum in the most recently found specimen suggests that a skeletal origin for the pectoralis (and supracoracoideus?) existed in at least some of the specimens, as is also suggested by the pronounced deltoid crest on the humems for insertion of the pectoralis. The suprac oracoideus of Archaeopteryx was not diverted to a dorsal inser tion, however, and thus could neither rapidly rotate the humer us nor augment supination of the distal wing. Conclusions In summary, Archaeopteryx was apparently incapable of the high-velocity rotation of the humems about its longitudinal axis that would have been generated by a derived supracoracoi deus with a dorsally inserting tendon. The subsequent evolu tion in later forms of an acrocoracoid, dorsally inserting ten don, and tuberculum dorsale resulted in (1) rotation of the humerus on its longitudinal axis to position the forearm and hand so their extension orients the fully outstretched wing in the parasagittal plane (i.e., the wing's ventral surface faces lat erally, the position appropriate for the beginning of the subse quent downstroke), (2) increased speed of the upstroke, and (3) augmented supination of the hand to reduce drag. By relocating the site of insertion of the supracoracoideus to an elevated posi tion on the dorsal surface of the humems (the novel external tu berosity), wing supination was accelerated, and the range of movement perhaps increased. Such augmentation of supination and an increase in the velocity of upstroke must have provided high selective value, particularly for rapid takeoff and landing. Literature Cited Biewener, A.A., K.P. Dial, and G.E. Goslow, Jr. 1992. Pectoralis Muscle Force and Power Output during Flight in the Star ling. Journal of Experimental Biology, 164:1-18. Bock, W.J., and P. Buhler 1995. Origin of Birds: Feathers, Flight and Homoiothermy (Ursprung der Vogel: Feder, Flug und Homoiothermie). Archaeopteryx, 12:5-13. Brown, R.HJ. 1951. Flapping Flight. Ibis, 93:333-359. Dial, K.P, and A.A. Biewener 1993. Pectoralis Muscle Force and Power Output during Different Modes of Flight in Pigeons. Journal of Experimental Biology, 176:31-54. Dial, K.P, G.E. Goslow, Jr., and F.A. Jenkins, Jr. 1991. The Functional Anatomy of the Shoulder in the European Starling (Sturnus vulgaris). Journal of Morphology, 207:327-344. Dial, K.P, S.R. Kaplan, G.E. Goslow, Jr., and F.A. Jenkins, Jr. 1988. A Functional Analysis of the Primary Upstroke and Downstroke Muscles in the Domestic Pigeon (Columba livia) during Flight. Journal of Experimental Biology, 134:1-16. Dodson, P. 1985. International Archaeopteryx Conference. Journal of Vertebrate Pa leontology, 5:177-179. Gauthier, J.A., and K. Padian 1985. Phylogenetic, Functional, and Aerodynamic Analysis of the Origin of Birds and Their Flight. In M.K. Hecht, J.H. Ostrom, G. Viohl, and P. Wellnhofer, editors, The Beginnings of Birds, pages 185-197. Eichstatt: Freunde des Jura-Museums, Eichstatt. 1989. The Origin of Birds and the Evolution of Flight. In K. Padian and D.J. Chure, editors, The Age of Dinosaurs. Short Courses in Pale ontology, 2:121-133. Knoxville, Tennessee: The Paleontological Society. Hecht, M.K., J.H. Ostrom, G. Viohl, and P. Wellnhofer, editors 1985. The Beginnings of Birds. 382 pages. Eichstatt: Freunde des NUMBER 89 309 Jura-Museums, Eichstatt. Hou, L.-H. 1995a. Morphological Comparisons between Confuciusomis and Archae opteryx. In A. Sun and Y. Wang, editors, Sixth Symposium on Meso zoic Terrestrial Ecosystems and Biota, Short Papers, pages 193-201. Beijing: China Ocean Press. Hou, L.-H., Z. Zhou, L.D. Martin, and A. Feduccia 1995. A Beaked Bird from the Jurassic of China. Nature, 377:616-618. Jenkins, F.A., Jr. 1993. The Evolution of the Avian Shoulder Joint. American Journal of Science, 293A:253-267. Jenkins, F.A., Jr., K.P. Dial, and G.E. Goslow, Jr. 1988. A CineradiogTaphic Analysis of Bird Flight: The Wishbone in Star lings Is a Spring. Science, 241:1495-1498. Olson, S.L., and A. Feduccia 1979. Flight Capability and the Pectoral Girdle of Archaeopteryx. Nature, 278:247-248. Ostrom, J.H. 1969a. A New Theropod Dinosaur from the Lower Cretaceous of Montana. Post ilia, Yale Peabody Museum of Natural History, 128: 1-17. 1969b. Osteology of Deinonychus antirrhopus, an Unusual Theropod from the Lower Cretaceous of Montana. Bulletin of the Peabody Museum of Natural History, 30: 165 pages. 1976a. Some Hypothetical Anatomical Stages in the Evolution of Avian Flight. In S.L. Olson, editor, Collected Papers in Avian Paleontol ogy Honoring the 90th Birthday of Alexander Wetmore. Smithso nian Contributions to Paleobiology, 27:1-21. 1976b. Archaeopteryx and the Origin of Birds. Biological Journal of the Linnean Society, 8:91-182. 1986. The Cursorial Origin of Avian Flight. In K. Padian, editor, The Ori gin of Birds and the Evolution of Flight. California Academy of Sci ences, Memoirs, 8:73-81. 1994. On the Origin of Birds and of Avian Flight. In D.R. Prothero and R.M. Schoch, editors, Major Features of Vertebrate Evolution. Short Courses in Paleontology, 7:160-170. The Paleontological Society. 1995. Wing Biomechanics and the Origin of Bird Flight. In W.-E. Reif and F. Westphal, editors, Festschrift: Herm Professor Adolf Seilacher zur Vollendung des 70 Lebensjahres am 24 Februar 1995 gewidmet. Neues Jahrbuch fiir Geologie und Paldontologie, 195:253-266. Padian, K., editor 1986. The Origin of Birds and the Evolution of Flight. California Academy of Sciences, Memoirs, 8: viii+98 pages. Pennycuick, C.J. 1993. Mechanics of Flight. In D.S. Farner, R.J. King, and K.C. Parkes, ed itors, Avian Biology, 9:1-75 London: Academic Press. Poore, S.O., A. Sanchez-Haiman, and G.E. Goslow, Jr. 1997. Wing Upstroke and the Evolution of Flapping Flight: A "Twist" in the Tale. Nature, 387:799-802. Rayner, J.M.V. 1988a. The Evolution of Vertebrate Flight. Biological Journal of the Lin nean Society, 34:269-287. 1988b. Form and Function in Avian Flight. Current Ornithology, 5:1-77. Ruben, J. 1991. Reptilian Physiology and the Flight Capacity of Archaeopteryx. Evolution, 45:1-17. Sanz, J.L., and J.F. Bonaparte 1992. A New Order of Birds (Class Aves) from the Lower Cretaceous of Spain. In K.E. Campbell, Jr., editor, Papers in Avian Paleontology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:39-49. Sanz, J.L., J.F. Bonaparte, and A. Lacasa 1988. Unusual Early Cretaceous Birds from Spain. Nature, 331:433^135. Sanz, J.L., L.M. Chiappe, and A.D. Busalioni 1995. The Osteology of Concornis lacustris (Aves: Enantiornithes) from the Lower Cretaceous of Spain and a Reexamination of Its Phyloge netic Relationships. American Museum Novitates, 3133:1-23. Schultze, H.P, and L. Trueb, editors 1991. Origins of the Higher Groups of Tetrapods: Controversy and Con sensus. xii+724 pages. Ithaca: Cornell University Press. Sereno, P.C, and C. Rao 1992. Early Evolution of Avian Flight and Perching: New Evidence from the Lower Cretaceous of China. Science, 255:845-848. Simpson, S.F. 1983. The Flight Mechanism of the Pigeon Columbia livia during Take-Off. Journal of Zoology (London), 200:435^43. Sokoloff, A., J. Gray, and J. Harry 1994. Supracoracoideus Is Not Necessary for Take-Off in the Starling. American Zoologist, 311:64A. Sy, M. 1936. Funktionall-anatomische Untersuchungen an Vogelfliigel. Journal fur Ornithologie, 84:199-296. Vazquez, R.J. 1992. Functional Osteology of the Avian Wrist and the Evolution of Flap ping Flight. Journal of Morphology, 211:259-268. 1995. Functional Anatomy of the Pigeon Hand (Columba livia): A Muscle Stimulation Study. Journal of Morphology, 226:33?45. Wellnhofer, P. 1974. Das funfte Examplar von Archaeopteryx. Palaeontographica, sec tion A, Palaozoologie, 147:169-216. 1988. Ein neues Examplar von Archaeopteryx. Archaeopteryx, 6:1-30. 1992. A New Specimen of Archaeopteryx from the Solnhofen Limestone. In K.E. Campbell, Jr., editor, Papers in Avian Paleontology Honor ing Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:3-23. 1993. Das siebte Examplar von Archaeopteryx aus dem Solnhofener Schichten. Archaeopteryx, 11:1-48. Zhou, Z., F. Jin, and J.-Y Zhang 1992. Preliminary Report on a Mesozoic Bird from Liaoning, China. Chi nese Science Bulletin, 37:1365-1368. A Comparison of the Jaw Skeleton in Theropods and Birds, with a Description of the Palate in the Oviraptoridae Andrzej Elzanowski ABSTRACT Similarities to birds in the structure of the jaws and palate sug gest that oviraptorosaurs (oviraptorids and caenagnathids), therizi- nosauroids, and omithomimosaurs are the closest theropodan relatives of birds, which is in conflict with recent phylogenetic reconstructions based on postcranial evidence. No specific avian similarities could be found in the jaws and palate of dromaeosau- rids. The ectopterygoid of the oviraptorids connects the lacrimal to the palatine, as does the avian uncinate (lacrimopalatine). This and other cranial similarities between the oviraptorosaurs and ornithu- rine birds raise the possibility that oviraptorosaurs are the earliest known flightless birds. With Archaeopteryx and the theropods pro viding evidence of plesiomorphic conditions, similarities in the mandibles, teeth, and tooth implantation in the Ichthyomithidae and Hesperomithidae may be interpreted as synapomorphies sup porting monophyly of the Odontognathae. Introduction Until recently, evidence for the theropod relationships of birds was derived almost exclusively from the postcranial skel eton (Ostrom, 1976; Gauthier, 1986; Holtz, 1994, 1996). Crani al comparisons have been used primarily by the proponents of alternative hypotheses (Tarsitano, 1991), with the notable ex ceptions of Currie's (1985; see also Currie and Zhao, 1993) studies of Troodon and Raath's (1985) studies of Syntarsus. Unequivocal cranial evidence for the theropod relationships of birds has only recently been provided by the exceptionally well-preserved skull of the seventh skeleton of Archaeopteryx (Elzanowski and Wellnhofer, 1995, 1996). The skull of Archaeopteryx, however, turned out to be very different from that of any known theropod, and the relation- Andrzej Elzanowski, Institute of Zoology, University of Wroclaw, UI. Sienkiewicza 21, 50335 Wroclaw, Poland. ships of birds within the theropods remain unsettled, as do the relationships between the theropod taxa (Russell and Dong, 1993; Holtz, 1996). The major cladistic analyses based prima rily on postcranial characters (Gauthier, 1986; Holtz, 1994, 1996) singled out the dromaeosaurids as the closest relatives of birds and, thus, echoed Ostrom's (1976) comparisons of Ar chaeopteryx to Deinonychus. This is inconsistent with cranial evidence, at least from the palate and jaws, which does not sup port a dromaeosaurid relationship for birds. The present paper provides a detailed description of the ovi- raptorid palate and examines avian similarities in the bony jaws and palate of oviraptorosaurs, therizinosauroids, and omitho mimosaurs. The oviraptorosaurs include the families Ovirap toridae and Caenagnathidae (=Elmisauridae), and the latter in cludes Chirostenotes {=Caenagnathus) (Sues, 1997). ACKNOWLEDGMENTS.?I am grateful to Halszka Osmolska, Institute of Paleobiology, Warsaw (ZPAL), for discussion, lit erature, and permission to study the oviraptorid skull ZPAL MgD-I/95; Storrs Olson, National Museum of Natural History, Smithsonian Institution, Washington, D.C. (NMNH), and Lar ry Witmer, Ohio University, Athens, for excellent reviews; Hans-Dieter Sues, Royal Ontario Museum (ROM), Toronto, for discussion and access to comparative materials; and Kieran Shepherd, Canadian Museum of Nature, Ottawa (CMN), for fa cilitating my study of the mandible of Chirostenotes pergracil- is Gilmore. The figures were prepared by Karol Sabbath, Insti tute of Paleobiology, Warsaw (Figures 1, 2); Taina Litwak, Gaithersburg, Maryland (Figure 3); and Claudia Angle, NMNH (Figures 4, 5). The Oviraptorid Palate The holotype of Oviraptorphiloceratops Osborn, which was for half a century the only known oviraptorid specimen, has a badly crushed skull that provides virtually no information about the cranial interior (Osborn, 1924; Smith, 1992). Well-preserved oviraptorid specimens had been collected over 311 312 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 20 years ago by the Polish- and Soviet-Mongolian expeditions, but the first review of their osteology was published only re cently (Barsbold et al., 1990), and most of the long-known Mongolian cranial specimens still need to be thoroughly stud ied. New finds of oviraptorids, including an embryonic skele ton, have been reported from Mongolia (Norell et al., 1994; Dashzeveg et al., 1995). The following description is based on the skull of Oviraptor sp. (ZPAL MgD-I/95) in the collection of the Institute of Pale obiology, Polish Academy of Sciences, Warsaw, Poland. This specimen has been illustrated and interpreted in the context of broad, primarily functional, comparisons (Osmolska, 1976) but was never described in detail. Osmolska's (1976) reconstruc tion of the oviraptorid palate proved correct except for details of the median pterygoid contact. Barsbold (1983a, fig. 12) pub lished an illustration of the similar palate of Oviraptor philo- ceratops, which is not accompanied by a description and con tains a misinterpretation of the vomer as the rostral part of the pterygoid. Both the premaxilla and maxilla bear sharp tomial edges that suggest a cutting function of the jaws (Figure 1). The palatal shelves of the premaxillae and maxillae form an entirely closed "secondary" palate. The two shelves of each bone remain sepa rated by a median suture. Dorsal to the shelves, the body of the premaxilla encloses a spacious sinus. The ventral (palatal) sur face of the premaxillary shelves is overlapped by the maxillary shelves. Each maxillary shelf is made of two longitudinal bulges sep arated by a shallow groove. The shelf is separated from the to mial edge by a deep, probably neurovascular, groove that emp ties into an opening at the suture with the premaxilla. The tomial edge is continuous with the lateral wall of the bone, which overlaps the premaxilla rostrally. The medial wall is seen in part as a perpendicular stmt, visible in the antorbital fenestra (Figure 2 A), that rises in the midlength of the bone and leaves a large maxillary foramen in front. The medial and later al walls are both fenestrated and enclose a spacious maxillary sinus (a part of the antorbital fossa) that extends all the way to the caudal (jugal) end of the bone (Figure 2A). The maxilla of Oviraptorphiloceratops has a similar structure (Barsbold et al., 1990, fig. 10.1 A). The maxilla is forked caudally (Figure 1). Each palatal shelf has a prominent knob- or tooth-like caudomedial process. The two caudomedial processes brace the vomer. Caudolaterally, the maxilla continues as a palato-jugal wing that articulates with the palatine, jugal, ectopterygoid, and lacrimal. The vomer is tightly held between the maxilla and the ptery goid. The rostral end of the vomer is strongly expanded and composed of a median knob and lateral wings. The knob is braced and the wings are overlapped by the caudomedial pro cesses of the maxilla. The convoluted suture to the pterygoid suggests a deep interdigitation. The palatine is composed of the maxillary process, which is its only prominent rostral process, and the pterygoid (caudal) pm mp plm 1 v fp A^?p 1 c W-ec qj FIGURE 1.?Reconstruction of the oviraptorid bony palate based on the speci men ZPAL MgD-I/95. Arrow points to the palatine's ascending wing (invisible in this view, see Figure 2). Scale bar=10 mm. (ec=ectopterygoid, fp=postpalatine fenestra, in=intemal naris, ip=interpterygoid vacuity, j=jugal, l=lacrimal, m=maxilla, mp=tooth-like caudomedial process of maxilla, plc=choanal conch (pterygoid wing) of palatine, plm=maxillary process of palatine, pm=premaxilla, pt=pterygoid, ptb=basal wing of pterygoid, q=quadrate, qj=quadratojugal, t=tomial edges of premaxilla and maxilla, v= vomer.) wing, which encloses the choana caudally. The maxillary pro cess has a triangular ascending wing that forms the lateral wall of the choana and articulates with the lacrimal dorsally and the maxilla rostrally (Figure 2B). The pterygoid wing, which is pa per-thin and poorly preserved, is strongly convex-concave dor- 313 B Pic FIGURE 2.?Oviraptorid specimen ZPAL MgD-I/95: A, upper jaw in lateral view; B, lateral wall of the choana in medial view, with a schematic reconstruction of the pterygoid wing of the palatine (pic). Arrow points to a nar row pocket that opens laterally by the postpalatine fenestra. Each scale bar=5 mm. (ec=ectopterygoid, fp=postpalatine fenestra, in=intemal naris, j=jugal, 1= lacrimal, m=maxilla, mf=maxillary foramen, ml=lateral wall of caudal maxillary sinus, mm=medial wall of caudal maxillary sinus, mp=tooth-like caudomedial process of maxilla, nf=nasal foramen, pl=palatine, pla=ascending wing of palatine, pIc=choanal conch (pterygoid wing) of palatine, plm=maxillary process of palatine, pm=premaxilla, pt=pterygoid, s= openings to caudal maxillary sinus, v=vomer.) soventrally: its dorsal vault fits in the ventral trough of the pterygoid and appears to be partly fused with that bone. The rostromedial comer of the wing probably formed a diminutive counterpart of the vomeral process of other theropods (Elza nowski and Wellnhofer, 1993, fig. 4), as in Oviraptorphilocer- atops (Barsbold et al., 1990, fig. 10.ID). Laterally, the palatine 314 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY is separated from the ectopterygoid by a vestigial postpalatine ("palatine") fenestra (Figure 2A). The fenestra leads to a flat bony pocket, which opens medially by a long fissure between the ascending wing of the palatine and the ectopterygoid (Fig ure 2B). Although the palatine/pterygoid contact is not pre served, there is no indication of the presence of a pterygopa latine ("subsidiary palatine") fenestra between the pterygoid wing of the palatine and the pterygoid. The details of pterygoid structure remain unclear due to its tight contact with or fusion to the adjacent bones. The rostral end of the pterygoid cannot be precisely distinguished from the palatine and ectopterygoid but seems to be forked rostrally. The rostral part of the bone is strongly concave ventrally and receives the pterygoid wing of the palatine. The caudal part of the pterygoid has a short wing for the basipterygoid articulation and a quadrate wing that closely adheres to what has been iden tified as the pterygoid ramus of the quadrate (Barsbold et al., 1990, fig. 10.1 A). The pterygoquadrate articulation, at angles to the long axis of the pterygoid, is very tight and does not sug gest any mobility. The ectopterygoid provides a smooth rostral continuation of the thick lateral margin of the pterygoid. (A large bone in the type specimen of Oviraptor philoceratops that is attached to the caudal end of the pterygoid was identified as an ectoptery goid (Smith, 1992), which is clearly a mistake, the bone proba bly being the left quadrate.) The ectopterygoid is oriented verti cally, almost in a parasagittal plane, and its rostral end lies much more dorsally than does the caudal end (Figure 2A). The rostral end articulates primarily with the lacrimal and maxilla and marginally contacts the jugal and the ascending wing of the palatine (Figure 2). The ectopterygoid is well delimited from the pterygoid across the ridge, but the suture wanes more medi ally (in the trough). The medial margin of the bone is over lapped by the palatine. Rostrally, the two bones are separated by a small postpalatine fenestra, which is well exposed in later al view (Figure 2A) but is barely exposed in ventral view (Fig ure 1). The lacrimal is oriented transversely, and its cross section gradually expands from a flattened lateral ridge to a broad me dial base (Figure 2A). Its ventral extremity is irregularly crenate (Figure 2B). Comparisons Characters of the Oviraptoridae (Figures 1, 2; Barsbold et al., 1990), Caenagnathidae (Figure 5; Sternberg, 1940; Currie et al., 1993; Sues, 1997), Therizinosauroidea as represented by Erlikosaurus (Clark et al., 1994), and Omithomimosauria (Os- molska et al., 1972; Barsbold and Osmolska, 1990) that are specifically shared with Archaeopteryx (Elzanowski and Wellnhofer, 1995, 1996), Confuciusomis (Hou et al., 1995), Gobipteryx (Elzanowski, 1977, 1995), the Odontognathae (Marsh, 1880; Elzanowski, 1991), and other ornithurine birds (Jollie, 1957; Elzanowski, 1995) are analyzed below, and their distribution is summarized in Table 1. All these characters show the opposite states in the Dromaeosauridae (Colbert and Russell, 1969; Ostrom, 1969; Sues, 1977; Currie, 1995) and usually in Allosaurus (Madsen, 1976) and other tetanuran theropods. The subdivision of birds and, especially, the defini tion of Ornithurae, follow Elzanowski (1995). Unfortunately, very little is known about the jaws of tro- odontids, which show some avian similarities in their braincase (Currie, 1985; Currie and Zhao, 1993). Similar to the troodon- tids are jaw fragments of Archaeornithoides (Elzanowski and Wellnhofer, 1992, 1993), which may in fact represent a juve nile troodontid. One of the main reasons for describing it as a separate genus in a family of its own was its tooth stmcture, TABLE 1.?Potential cranial synapomorphies of the Omithomimosauria (Omim), Therizinosauroidea (Ther), Oviraptoridae (Ovir), Caenagnathidae (Caen), and birds as represented by Archaeopteryx (Arch), Gobipteryx (Gobi), and Hesperomis (Hesp). The opposite character states (0) are present in the Dromaeosauridae and the majority of known theropods. Parentheses indicate that the homology of noted similarities may be open to inter pretation. A=ambiguous character state. See text for complete definitions and discussion of the characters. Character 1. Palatine with long maxillary process 2. Coronoid absent 3. Inrraramal articulation absent 4. Maxilla with broad palatal shelf 5. Quadrate head bent backwards 6. Palatine with broad pterygoid wing 7. Pterygoid with basal process 8. Ectopterygoid in rostral position 9. Articular and surangular co-ossified 10. Articular with lateral process 11. Articular with medial process 12. Mandibular symphysis fused 13. Jugal bar rod-shaped 14. Ectopterygoid contacts lacrimal Omim 1 1 1 1 1 0 (1) A 0 0 0 0 0 7 Ther 0 0 0 0 0 Ovir Caen 1 1 1 1 1 1 1 1 1 ? 1 1 1 ? 1 ? 0 1 1 1 1 1 1 1 1 ? 1 ? Arch 1 1 1 ? 1 1 A 0 0 0 0 0 0 0 Gobi ? (1) ? Hesp 1 1 0 1 0 1 1 (1) 1 1 1 (1) 1 (1) NUMBER 89 315 which is unlike that in any theropod. Theropod teeth were once strongly believed not to vary with age (Currie et al., 1990), but this belief has been refuted by the discovery of subcorneal, un- serrated teeth in dromaeosaurid hatchlings (Norell et al., 1994). Similar teeth are present in Archaeomithoides and may have been present in the early juveniles of troodontids. MAXILLA.?The maxilla has broad palatal shelves that meet at, or at least approach, the midline in the Therizinosauroidea, Oviraptoridae, Caenagnathidae, Omithomimosauria, Archae omithoides, Hesperomis, and the paleognaths except Struthio. A fairly broad palatal shelf of the maxilla was probably present in Gobipteryx. Unfortunately, contrary to previous interpreta tions, the palatal aspect of the maxilla (as well as premaxilla) remains entirely unknown in Archaeopteryx. In Hesperomis each palatal shelf of the maxilla ends with two processes, the lateral and the medial palatine process. The medial process, known as the maxillopalatine, is the only por tion of the palatal shelf that remains in Struthio and the neo- gnaths. The peg-like process of the oviraptorids and caena- gnathids corresponds in position to the maxillopalatine of neomithine birds in being a caudomedial extension of the pal atal shelf that contacts the vomer. At least in the oviraptorosaurs, Archaeomithoides, and birds, the maxillary shelf provides a floor for the caudal maxillary si nus. Other theropods are believed to have only the rostral sinus (Witmer, 1990). Among Mesozoic birds, the caudal maxillary sinus is well documented in Hesperomis (Witmer, 1990) and may have been present in Archaeopteryx, although the evi dence of its presence in the fifth skeleton provided by Witmer (1990:360, fig. 14) is probably incorrect (Elzanowski and Wellnhofer, 1995). The upper jaw of the fifth skeleton, how ever, contains a vertical, median or paramedian element (Elza nowski and Wellnhofer, 1995, fig. 7X) that is similar in shape and location to the medial wall of the caudal sinus in the ovi raptorids. PALATINE.?The palatine of the omithomimosaurs, therizi- nosauroids, oviraptorids, caenagnathids (Sues, 1997, fig. 2), and birds has a maxillary process that is much longer than the rostromedial vomeral process and overlaps the maxillary pala tal shelf ventrally. In the neomithines, including Hesperomis, the maxillary process is known as the premaxillary process be cause it extends even further rostrally and reaches the premax illa. The palatine has a broad pterygoid (caudal) wing that over- laps the pterygoid ventrally in the therizinosauroids, oviraptorids, birds, and probably in caenagnathids. In the omithomimosaurs, the palatine has a dorsal, trans versely oriented process situated close to the lacrimal (Osmol- ska et al., 1972:116, 136). A prominent transverse crest is present in Archaeopteryx in a comparable location (Elzanows ki and Wellnhofer, 1996:89, fig. 9A), and three transverse crests are present in Chirostenotes (Sues, 1997, fig. 2). In con trast, the dorsal process of the palatine that ascends to the lac rimal in the oviraptorids is oriented in the parasagittal plane (Figure 2B). PTERYGOPALATINE FENESTRA.?Gauthier (1986) used the pterygopalatine (subsidiary palatine) fenestra as one of two di agnostic cranial characters of the newly defined Coelurosauria, although it is known to be present in only two of the originally included families, the omithomimosaurs and dromaeosaurids. The pterygopalatine fenestra has been subsequently identified in Archaeomithoides, the therizinosauroids, and tentatively in Chirostenotes and Gobipteryx. This fenestra is lacking in the oviraptorids (Figure 1). In Hesperomis and other neomithines there is no separate fenestra, although the situation in Gobip teryx suggests that it may have merged with the choana. The lack of an appropriate embayment in either the palatine or pterygoid of Archaeopteryx suggests either the absence of this fenestra or a configuration similar to that in the neomithines. The uncertain status of the pterygopalatine fenestra in birds makes it of little use in the search for avian relatives. ECTOPTERYGOID AND POSTPALATINE FENESTRA.?The ecto pterygoid is situated rostrally and, as a result, the postpalatine (palatine) fenestra is reduced in the oviraptorids (Figures 1, 2) and therizinosauroids (although it is unclear whether the fenes tra is present in the latter). The fenestra is much smaller in Gal- limimus (Osmolska et al., 1972:108) than it is in the dromaeo saurids and may be even smaller, if present at all, in Omithomimus edmontonicus Sternberg (ROM 851), where the ectopterygoid is preserved in contact with the lacrimal (pers. obs.) in a position very different from that reconstructed in Gallimimus. In the oviraptorids and Erlikosaurus, the ecto pterygoid is lateral to the palatine, which is due to its rostral position and the presence of the pterygoid wing of the palatine. This also may be tme of the Caenagnathidae (Sues, 1997:701). Although the ectopterygoid is positioned caudally in the fifth specimen of Archaeopteryx, it may have overlapped, at least in part, the long pterygoid wing of the palatine. The oviraptorid ectopterygoid differs from that of Archaeop teryx and from other theropods, including the therizinosau roids, in the lack of the jugal hook, its distal articulation prima rily with the lacrimal and maxilla rather than the jugal, and its strongly slanting position between the lacrimal and the palatine (Figure 2). In all these differences, the oviraptorid ectoptery goid agrees with the avian uncinate (uncinatum=lacrimopala- tinum), which in the neomithine birds articulates with the cau- doventral margin of the lacrimal and descends caudoventrally to the palate (Figure 3). Although in modem birds the uncinate tapers toward the ventral tip and either articulates with the pa latine or ends free above it, in oviraptorids the bone flares out ventrally and articulates with both the palatine and the ptery goid. This difference may be accounted for by the reduction of the rostral part of the pterygoid (which became separated and reduced as the hemipterygoid) in the neomithine birds. In Hes peromis the uncinate probably approached or marginally con tacted the large hemipterygoid (Elzanowski, 1991, fig. 3). The uncinate of extant birds is clearly a vestigial structure that is extremely variable in shape. The uncinate is widespread among the neomithines. It has been found in the Stmthionidae and Rheidae among the paleog- 316 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY B 5mm 5mm FIGURE 3.?Fregata magnificens Mathews (Neognathae). The antorbital area of the skull in A, lateral, and B, dor- solaterocaudal views. Each scale bar=5 mm. (j=jugal, l=lacrimal, m=maxilla, n=nasal, nf=nasal foramen, pl=palatine, u=uncinate (lacrimopalatine).) naths and in nine families representing six neognathous orders, including Fregatidae (Pelecaniformes), Procellariidae and Di- omedeidae (Procellariiformes), Laridae and Alcidae (Charadrii formes), Cariamidae (Gruiformes), Accipitridae (Falconi formes), and Cuculidae and Musophagidae (Cuculiformes) (Burton, 1970; pers. obs.). Monophyly of this set of taxa is out of the question, and four of the included orders (Pelecani formes, Procellariiformes, Charadriiformes, and Gmiformes) have been repeatedly considered to be among the earliest neog nathous branches. In addition, strong evidence of the presence of the uncinate has been found in Hesperomis (Elzanowski, 1991). This suggests that the presence of the uncinate is plesio morphic at least for the neomithines and that an ancestral bone was present in the primitive, pre-neomithine birds. The origin of the uncinate has remained obscure. Frank (1954:232) suggested that it is an avian neomorph. Burton (1970) raised the possibility of its derivation from a ligament (which is more widespread than the bone) but admitted that this begs the question of the origin of the ligament. McDowell (1978) mapped the uncinate to the archosaurian ectopterygoid because this bone appeared to be the only possible avian homo- logue for a reptilian element that lies lateral to the pterygoid and caudal to the maxilla. Thus far, no intermediate condition has been found. The oviraptorid ectopterygoid provides at least a stmctural intermediate between the reptilian ectopterygoid and the avian uncinate. Whether it is an evolutionary interme diate remains to be decided in a broader phylogenetic analysis. The similarities to the avian uncinate cannot be homologous unless the oviraptorids are more closely related to the ornithu- rine birds than is Archaeopteryx because the latter possesses a typical theropodan, hooked ectopterygoid, which is preserved in a caudal position in the fifth (Eichstatt) skeleton. BASIPTERYGOID ARTICULATION.?The pterygoid has a basal process for the articulation with the cranial base in oviraptorids (Figure 1), Hesperomis, and the majority of those neognaths that have a basipterygoid articulation (sensu stricto) as distin guished from the rostropterygoid articulation of the galliforms and anatids (Weber, 1993). The basal process is well developed in many Charadriiformes and some Caprimulgiformes (Ste- atomithidae, Caprimulgidae, Nyctibiidae) and in all the Tumi- NUMBER 89 317 cidae, Anhimidae, Columbidae, Sagittariidae, Strigiformes, and Trogonidae (Figure 4). The pterygoid of the Procellariidae and Pelecanoididae has a rostral wing that extends to the rostral end of the bone and articulates only caudally with the basip terygoid process. The basal articular surface is least set off from the shaft of the pterygoid in the Cathartidae. The basal process of the pterygoid is poorly developed in Ar chaeopteryx, and the structure of the basipterygoid articulation remains unknown in Gobipteryx, which makes the phylogenet ic significance of this character somewhat ambiguous. Because the basal process of the oviraptorids (and probably in the caen agnathids, the basipterygoid articulation of which is very simi lar to that of oviraptorids) is distinctly better developed than it is in Archaeopteryx, it is interpreted herein as another possible synapomorphy linking the oviraptorosaurs to the omithurines. The quadrate ramus of the pterygoid is closely appressed on the braincase in the oviraptorids, omithomimosaurs, and theriz inosauroids. The basipterygoid process is virtually absent in the oviraptorosaurs, which have the articular surface of the ptery goid developed directly on the cranial base (Sues, 1997, fig. 3), and the same seems to be tme of the therizinosauroids (Clark et al., 1994:20). The identification of the basipterygoid processes in the omithomimosaurs remains uncertain. Their quadrate wing of the pterygoid has a prominent medial process that seems to abut the cranial base at the level of mandibular articu lation (i.e., behind the sphenoid capsule). This is clearly visible in an undescribed skull (ROM 851) of Omithomimus edmon- tonicus (pers. obs.). The apparent basal process of the ptery goid was left uninterpreted in Gallimimus bullatus Osmolska et al. (Osmolska et al., 1972, fig. 2; Barsbold and Osmolska, 1990, fig. 8.1), and the basipterygoid articulation was identi fied farther rostrally on the sphenoid capsule, a contact that is unlike a theropodan basipterygoid articulation and may have arisen as a result of the transverse expansion of the capsule. FIGURE 4.?The pterygoid in various neognathous birds, medioventral view: A, Turnix varia (Latham); B, Pele canoides urinatrix (Gmelin); C, Procellaria aequinoctalis Linnaeus; D, Coragyps atratus (Bechstein); E, Sagit tarius serpentarius (Miller); F, Anhima cornuta (Linnaeus); G, Chauna torquata (Oken); H, Columba fasciata Say; I, Philomachus pugnax (Linnaeus); J, Vanellus vanellus (Linnaeus); K, Trogon rufus Gmelin; L, Strix aluco Linnaeus. In B-E, the length of the basal articular surface on the pterygoid (indicated by "b" in D,E) is much greater than the rostrocaudal width of the basipterygoid process. (b=basal process, p=palatine articular surface, q=quadrate articular surface, w=rostrobasal wing.) 318 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY In the troodontids (Currie, 1985; Currie and Zhao, 1993) the basipterygoid processes are strong but are directed caudolater ally rather than rostrolaterally as in other theropods. This pre cludes a typical theropod basipterygoid articulation, and their facets are strongly concave, which suggests a basal process on the pterygoid (the caudal part of which is unknown in the tro odontids). In the dromaeosaurids and most of the remaining theropods, the pterygoids have sockets that receive prominent basipterygoid processes of the sphenoid. QUADRATE.?In the dromaeosaurids and most of the remain ing theropods, the caudal profile of the quadrate is straight, and the quadrate socket in the squamosal opens more or less ven trally. In contrast, the quadrate head is bent backward in the oviraptorids (Maryariska and Osmolska, 1997), Erlikosaurus, omithomimosaurs, troodontids, Gobipteryx, and the majority of other omithurines except for Hesperomithidae, Ichthyornis, and some modem birds. In Archaeopteryx the caudal bend of the quadrate head may be partly obscured by the position in which the quadrate is preserved; it is exposed in caudal aspect in the seventh skeleton and in rostral aspect in the fifth skele ton. The oviraptoid quadrate alone agrees with that of the omi thurines in having an otic capitulum for the articulation with the braincase, the pterygoid articular surface approaching the medial mandibular condyle, and a distinct (but shallow) quadratojugal cotyla (Maryariska and Osmolska, 1997). CORONOID.?The coronoid bone is absent in the oviraptoro saurs (Figure 5), omithomimosaurs, Erlikosaurus, and all adult birds, including Archaeopteryx and Gobipteryx. It may have been incorporated into the prearticular (Nemeschkal, 1983a), and a tiny splint of bone just rostral to the tip of the prearticular in the hatchling Golden Eagle {Aquila chrysaetos (Linnaeus)) probably is a vestige of the coronoid (Jollie, 1957). ARTICULAR AND PREARTICULAR.?The articular of ornifhu- rine birds (including Gobipteryx), but not Archaeopteryx, co-ossifies with the prearticular, surangular, and angular at var ious, mostly posthatching, stages of development. The articular and surangular are co-ossified in the caenagnathids, Erlikosau rus, and all other adult omithurines. The articular is co-ossified with the angular and surangular in Avimimus, in which the prearticular is not preserved (Kurzanov, 1987). The articular is separated by sutures from the adjacent bones in the omithomi mosaurs, oviraptorids, and other theropods, including dro maeosaurids and Allosaurus. The articular surface for the quadrate is expanded both medi ally and laterally beyond the ramus in ornithurine birds. The lat eral expansion is coextensive with the small lateral process that bears it. By contrast, the medial expansion covers only the basal part of the prominent medial process, which provides an area of attachment for the pterygoideus muscle. In Gobipteryx, how ever, there is no connection between the medial expansion of the articular surface and the medial process, which may or may not be due to damage. There is no evidence of any distinctive projections of the articular surface in Archaeopteryx or in the omithomimosauria. The articular surface is strongly expanded laterally but not medially in Erlikosaurus and is expanded both medially and laterally in the oviraptorids and caenagnathids. The two projections in the oviraptorosaurs are similar to those in Gobipteryx (Elzanowski, 1974, fig. XXXIII/2). Aside from being reduced in a few neognaths (such as the phasianids), the prearticular shows two fairly distinctive mor phologies. It turns dorsally and expands into an ascending blade in Erlikosaurus, the majority of theropods (including the dromaeosaurids), Archaeopteryx, Hesperomis (pers. obs.), and some neognaths (e.g., Spheniscidae, Laridae). The prearticular continues far rostrally as a straight bony rod in the caenagna- FlGURE 5.?Mandible of Chirostenotes pergracilis (-Caenagnathus collinsi R.M. Sternberg), CMN 8776 in medial view. The medial aspect of this mandible has been hitherto described only by Sternberg (1940, fig. 2). Sternberg's illustration showed a well-defined caudal outline of the dorsomedial process of the dentary, whereas in reality this process seems to be fused to the surangular. Absent from the mandible is the splenial, which may have fallen off, as frequently happens in birds, but it also is possible that it was reduced because the mandible is unusual in having a rostral extension of the angular that reaches the symphysis. The prearticular extends to the tip of the retroarticular process and covers the articular, which is fused to the surangular but not to the prearticular. (an=angular, d=ventral ramus of dentary, dv=ventral process of dorsal ramus of dentary, pa=prearticular, pc=coronoid process, pm=medial process, pr=retroarticular process, rl=lateral ridge, rm=medial ridge, sa=surangular.) NUMBER 89 319 thids (Figure 5), oviraptorids (Barsbold, 1983b, fig. 13), at least in Garudimimus among the omithomimosaurs (Barsbold and Osmolska, 1990, fig. 8.3.D), the paleognaths (Muller, 1963; pers. obs.), the remaining neognaths, and probably Gobipteryx. MANDIBULAR SYMPHYSIS.?The symphysis is unfused in Archaeopteryx, although the rostral tips of the rami are con nected very tightly. Among the theropods, the ends of the man dibular rami are fused only in the oviraptorosaurs and are fused even in the smallest caenagnathid specimens (Currie et al., 1993), but they remain unfused in the oviraptorid embryo (Norell et al., 1994), suggesting that fusion occurred at or soon after hatching, as in Gobipteryx (Elzanowski, 1981). The sym physis is fused in Confuciusomis, which combines an Archae- opteryx-like postcranial skeleton with a Gobipteryx-like skull (Hou et al., 1995), Gobipteryx (Elzanowski, 1977), and other omithurines except for the odontognaths (Marsh, 1880), tera torns (Campbell and Tonni, 1981), and pseudodontoms (Odon- topterygia) (Zusi and Warheit, 1992). The lack of a symphysis in the odontognaths is most proba bly secondary because in the hesperomithids and possibly in Ichthyornis, the tips of the rami were connected by a preden- tary bone (Martin, 1987), which is definitely a derived charac ter because no other bird or theropod has it. The dentary of birds arises from two centers of ossification (Nemeschkal, 1983b), and the predentary bone most probably ossified from the rostral (mentomandibular) center. Consequently, this bone probably represents the co-ossified symphyseal part that be came separated from the remainder of the mandible. The pseudodontoms are highly specialized, fish-eating neog- nathous birds related to the pelecaniforms, which makes the lack of mandibular symphysis in these birds unquestionably secondary. As of now, there is no evidence for multiple origins of the fused symphysis in birds. It may have evolved only once in the omithurines, and its presence could be another synapomorphy of the oviraptorosaurs and omithurines. INTRARAMAL JOINT.?This joint includes the articulations between the dentary and surangular and between the splenial and angular. It permits mediolateral mobility within each ra mus of the mandible. The majority of the theropods, including the dromaeosaurids, have an intraramal joint. It is absent in the omithomimosaurs, Erlikosaurus, oviraptorosaurs, and primi tive birds, including Archaeopteryx, Gobipteryx, and probably Confuciusomis. In more advanced birds, there are at least two types of intraramal joints: one in the pelecaniforms (including pseudodontoms) and another in the odontognaths. They are made of different components, which indicates that they evolved independently of each other (Zusi and Warheit, 1992). Gingerich (1973) proposed that the intraramal joints in Hes peromis and Ichthyornis have been inherited from the thero pods and thus represent yet another theropod/avian synapo morphy. Consequently, the similarity between Ichthyornis and hesperomithids would be symplesiomorphic. This, however, is inconsistent with the intraramal joint being absent in Archae opteryx, Confuciusomis, and Gobipteryx (see also Feduccia, 1996:155). In addition, the two kinds of intraramal joints in birds are more similar to each other (Gingerich, 1972, fig. 1; Zusi and Warheit, 1992, fig. 7) than either of them is to the joint in the dromaeosaurids (Currie, 1995, fig. 7). In the odon tognaths, the hinge is formed primarily by the splenial and an gular, and these bones articulate at an angle of-45? to the long axis of the ramus. In the dromaeosaurids the hinge, reinforced by bony knobs, is primarily between the dentary and surangu lar, whereas the splenial and angular seem to form a gliding ar ticulation that is oriented much more horizontally than the sple- nio-angular hinge of the odontognaths. In all probability, birds started their evolution with a rigid mandible, and detailed simi larity of the intraramal joints in Ichthyornis and hesperomithids is synapomorphic. JUGAL BAR.?Confuciusomis and probably all the ornithu- rine birds have a thin, rod-like jugal bar that is formed in sub stantial part, and in some neognaths exclusively, of the quadra- tojugal. In the oviraptorids and Avimimus (Kurzanov, 1987), the bar is thin, as it is in the omithurines. The slender jugal bar stands out in an otherwise massive oviraptorid skull adapted for durophagy. In Archaeopteryx and the remaining known theropods, the jugal bar is a robust slat. It is formed almost exclusively of the jugal in most of the theropods. In Archaeopteryx, the quadrato- jugal may have extended far rostrally on the medial side of the jugal (Elzanowski and Wellnhofer, 1996, fig. 7). The postorbital process of the jugal is in the terminal caudal position, and the infratemporal fossa is reduced in the omitho mimosaurs, the troodontids, and Archaeopteryx. Erlikosaurus is unique among the theropods in having two ascending pro cesses of the jugal: a postorbital process one-fourth the length from the caudal end, and a terminal process overlying the quad rate and approaching the squamosal. The latter corresponds in position to the postorbital process of the omithomimosaurs and Archaeopteryx. This unique jugal morphology raises the possi bility of a secondary enlargement of the infratemporal fossa, accompanied by the division of the initially terminal postorbit al process. Although the rostral position of the postorbital bar in the oviraptorids looks like a symplesiomorphy with the ma jority of theropods, it is conceivable that this position is sec ondary and that the postorbital process of the jugal corresponds to the rostral, possibly secondary, process in Erlikosaurus. INTERDENTAL PLATES AND TEETH.?The presence of inter dental plates is a primitive archosaurian character (Elzanowski and Wellnhofer, 1993). Among the theropods, inderdental plates are known to be absent in troodontids (Currie, 1987), Ar chaeomithoides (Elzanowski and Wellnhofer, 1993), Baryonyx and Spinosaurus (Buffetaut, 1992), and omithomimosaurs (Perez-Moreno et al., 1994). Separate interdental plates are present in juvenile dromaeosaurids (Norell et al., 1994) and seem to be replaced by porous interdental bone in adults (Cur rie, 1987). The discovery of interdental plates in Archaeopteryx (Welln hofer, 1993) makes it clear that their absence in Hesperomis and Ichthyornis is secondary, as probably are other aspects of 320 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY their tooth implantation, such as the expanded dental roots and the equal height of the lingual and buccal margins of the den tary (Elzanowski and Wellnhofer, 1995). The teeth of Ichthyor nis are usually anchored in separate alveoli, whereas those of the hesperomithids are in a common groove; however, Martin and Stewart (1977) described a mandible of Ichthyornis with teeth in a groove. Avian teeth do not show any close similarity to any known theropod teeth except for the early juvenile teeth of Archaeor- nithoides (Elzanowski and Wellnhofer, 1993). The teeth of Ar chaeopteryx and Ichthyornis are similar in having the crowns divided into a straight base and a recurved tip. On the other hand, the teeth of Ichthyornis were reported to have the roots distinctly expanded, as in the hesperomithids (Martin and Stewart, 1977). Unfortunately, the radiographs on which this observation is based have never been published. Dental roots are not expanded in Archaeopteryx (Howgate, 1984; Elzanows ki and Wellnhofer, 1996).1 Entirely toothless jaws evolved independently in the ovirap torosaurs, advanced omithomimosaurs, Confuciusomis and Gobipteryx (these two genera were included in the Enantiorni thes by Hou et al., 1995), and the advanced neomithines. An edentulous premaxilla coexists with a dentigerous maxilla and dentary in the therizinosauroids and hesperomithids. It is clear that the reduction of teeth occurred several times independently within the analyzed set of taxa. Discussion Some of the avian characters, such as the toothless beaks of the oviraptosaurs and advanced omithomimosaurs, are certain ly convergent and evolved more than once within the set of taxa under comparison. Reduction of teeth is a recurrent theme in vertebrate evolution. There is, however, no reason to expect a pervasive conver gence on birds in the cranial morphology of the omithomimo saurs, therizinosauroids, and oviraptorosaurs because their skulls are unlike any avian skull. This is what one could expect as a result of turning a primitive, perhaps Archaeopteryx-like cranial morphology into several highly specialized kinds of jaw apparatus of the gigantic descendants of the ancestors of birds. In terms of jaw function, the oviraptorids are comparable to the dicynodonts (Osmolska, 1976), and the therizinosauroids are convergent on the omithischians in having cheeks (Clark et al., 1994). It seems reasonable, therefore, to suspect that the unique cranial similarities of birds, oviraptorosaurs, therizinosauroids, and omithomimosaurs (Table 1, characters 1-6) are synapo- 1 Editor's Note: This is at variance with Martin and Stewart, elsewhere in this volume, who consider previously published photographs and descriptions to indicate that the teeth of Archaeopteryx do have expanded roots. morphic and thus indicative of the monophyly of a clade com posed of these taxa and probably troodontids (Figure 6). This would agree with the phylogeny reconstmcted from 20 cranial and 39 postcranial characters by Russell and Dong (1993), who subdivided the tetanurans into two groups: one that included the dromaeosaurs and camosaurs and one that included the or- nithomimosaurs, troodonts, therizinosauroids, and oviraptoro saurs. Most intriguing are four characters (Table 1, characters 10-13) that are shared by the oviraptorosaurs and the ornithu- rine birds but are absent in Archaeopteryx. These suggest that the oviraptorosaurs branched off after Archaeopteryx and thus represent the earliest known flightless birds. Except for the elongate forelimbs (which become shortened in all flightless forms), the postcranial skeleton of Archaeopteryx does not have any avian traits that would be absent in the oviraptorids (Barsbold 1983a, 1983b). Therefore, if flightlessness had evolved at a stage of avian evolution close to Archaeopteryx, this would be extremely difficult to distinguish from the prima ry flightlessness of the theropods. Relationships of the oviraptorosaurs have been enigmatic ever since their discovery (Osbom, 1924). The only consensus reached in recent phylogenetic reconstmctions is that the ovi raptorosaurs belong in the Coelurosauria, a major clade that gave rise to birds, but their placement within that clade varies considerably (Gauthier, 1986; Barsbold et al., 1990; Holtz, 1996). Evidence for their affinities comes almost exclusively from the postcranial skeleton because the oviraptorid skull is difficult to compare with the skull of any other group of thero pods. In contrast, even the highly specialized, edentulous skulls of omithomimosaurs are still clearly identifiable as theropodan. The two cranial characters used by Gauthier (1986) to define the Coelurosauria, namely, the pterygopalatine fenestra and ventral pocket in the ectopterygoid, are absent in the oviraptori ds (as well as in the omithurines). Several similarities of the jaws of Hesperomithidae and Ich- thyomithidae prove likely to be synapomorphies when ana lyzed against the plesiomorphic background provided by the theropods and Archaeopteryx. Wetmore (1930) combined the Hesperornifhiformes and Ichthyornithiformes in the superorder Odontognathae, but monophyly of this taxon has never been explicitly suggested because most of their similarities were thought to be primitive. A closer relationship between these two orders is now suggested by the lack of mandibular sym physis, probably due in both taxa to the separation of the tip of the mandible as a predentary bone; the presence and detailed similarities of the intraramal joint; the absence of interdental plates, which is probably correlated with the lingual alveolar margin being flush with the buccal margin; and probably the expansions of the dental roots (fide Martin and Stewart, 1977) and the straight quadrate. NUMBER 89 321 i' FIGURE 6. Table 1. -Unresolved polytomy of the avian and three theropod lineages, as suggested by characters 1-6 in Literature Cited Barsbold, Rinchen 1983a. O "ptich'ikh" chertakh v stroyenii khishchnykh dinozavrov [On the "Avian" Characters in the Structure of Predatory Dinosaurs]. Trudy, Sovmestnaya Sovetsko-Mongol 'skaya Paleontologicheskaya Eks peditsiya, 24:96-103. [In Russian.] 1983b. Khishchnyye dinozavry mela Mongolii [Predatory Dinosaurs of the Cretaceous of Mongolia]. Trudy, Sovmestnaya Sovetsko-Mon gol'skaya Paleontologicheskaya Ekspeditsiya, 19:1-117. [In Rus sian.] Barsbold, Rinchen, Teresa Maryanska, and Halszka Osmolska 1990. Oviraptorosauria. In D.B. Weishampel, P. Dodson, and H. Osmol ska, editors, The Dinosauria, pages 249-258. Berkeley: University of California Press. Barsbold, Rinchen, and Halszka Osmolska 1990. Omithomimosauria. In D.B. Weishampel, P. Dodson, and H. Os molska, editors, The Dinosauria, pages 225-244. Berkeley: Univer sity of California Press. Buffetaut, Eric 1992. Remarks on the Cretaceous Theropod Dinosaurs Spinosaurus and Baryonyx. Neues Jahrbuch fur Geologie und Paldontologie, Monat- shefte. (1992):88-96. Burton, P.J.K. 1970. Some Observations on the Os Uncinatum in the Musophagidae. Os trich, supplement, 8:7-13. Campbell, Kenneth E., and Eduardo P. Tonni 1981. Preliminary Observations on the Paleobiology and Evolution of Ter atorns (Aves: Teratornithidae). Journal of Vertebrate Paleontology, 1:265-272. Clark, James M., Altangerel Perle, and Mark Norell 1994. The Skull of Erlicosaurus andrewsi, a Late Cretaceous "Segnosaur" (Theropoda: Therizinosauridae) of Mongolia. American Museum Novitates, 3115: 35 pages. Colbert, Edwin H., and Dale A. Russell 1969. The Small Cretaceous Dinosaur Dromaeosaurus. American Museum Novitates, 2380: 49 pages. Currie, Philip J. 1985. Cranial Anatomy of Stenonychosaurus inequalis (Saurischia, Theropoda) and Its Bearing on the Origin of Birds. Canadian Jour nal of Earth Sciences, 22:1643-1658. 1987. Bird-like Characteristics of the Jaws and Teeth of Troodontid Theropods (Dinosauria, Saurischia). Journal of Vertebrate Paleon tology, 7:72-81. 1995. New Information on the Anatomy and Relationships of Dromaeo saurus albertensis (Dinosauria, Theropoda). Journal of Vertebrate Paleontology, 15:576-591. Currie, P.J., S.J. Godfrey, and Lev Nessov 1993. New Caenagnathid (Dinosauria: Theropoda) Specimens from the Upper Cretaceous of North America and Asia. Canadian Journal of Earth Sciences, 30:2255-2272. 322 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Currie, P.J., J.K. Rigby, Jr., and R.E. Sloan 1990. Theropod Teeth from the Judith River Formation of Southern Al berta, Canada. In Kenneth Carpenter and Philip J. Currie, editors, Dinosaur Systematics: Approaches and Perspectives, pages 107-125. Cambridge, Cambridge University Press. Currie, Philip J., and X. Zhao 1993. A New Troodontid (Dinosauria, Theropoda) Braincase from the Di nosaur Park Formation (Campanian) of Alberta. Canadian Journal of Earth Sciences, 30:2231-2247. Dashzeveg, D., M.J. Novacek, M.A. Norell, J.M.Clark, L.M. Chiappe, A. Davidson, M.C. McKenna, L. Dingus, C. Swisher, and A. Perle 1995. Extraordinary Preservation in a New Vertebrate Assemblage from the Late Cretaceous of Mongolia. Nature, 374:446-449. Elzanowski, Andrzej 1974. Preliminary Note on a Palaeognathous Bird from the Upper Creta ceous of Mongolia. Palaeontologia Polonica, 30:103-109. 1977. Skulls of Gobipteryx (Aves) from the Upper Cretaceous of Mongo lia. Palaeontologia Polonica, 37:153-165. 1981. Embryonic Bird Skeletons from the Late Cretaceous of Mongolia. Palaeontologia Polonica, 42:147-179. 1991. New Observations on the Skull of Hesperomis with Reconstructions of the Bony Palate and Otic Region. Postilla, 207: 20 pages. 1995. Cretaceous Birds and Avian Phylogeny. Courier Forschungsinstitut Senckenberg, 181:41-59. Elzanowski, Andrzej, and Peter Wellnhofer 1992. A New Link between Theropods and Birds from the Cretaceous of Mongolia. Nature, 359:821-823. 1993. Skull of Archaeomithoides from the Upper Cretaceous of Mongolia. American Journal of Science, 293-A:235?252. 1995. The Skull of Archaeopteryx and the Origin of Birds. Archaeopteryx, 13:41-46. 1996. The Cranial Morphology of Archaeopteryx: Evidence from the Sev enth Skeleton. Journal of Vertebrate Paleontology, 16:81-94. Feduccia, Alan 1996. The Origin and Evolution of Birds. 420 pages. New Haven and Lon don: Yale University Press. Frank, G.H. 1954. The Development of the Chondrocranium of the Ostrich. Annals of the University of Stellenbosch, 30:179-248. Gauthier, Jacques 1986. Saurischian Monophyly and the Origin of Birds. Memoirs of the California Academy of Sciences, 8:1?46. Gingerich, P.D. 1972. A New Partial Mandible of Ichthyornis. Condor, 74:471^173. 1973. Skull of Hesperomis and Early Evolution of Birds. Nature, 243:70-73. Holtz, T.R., Jr. 1994. The Phylogenetic Position of the Tyrannosauridae: Implications for Theropod Systematics. Journal of Paleontology, 68:1100-1117. 1996. Phylogenetic Taxonomy of the Coelurosauria (Dinosauria: Ther opoda). Journal of Paleontology, 70:536-538. Hou, Lianhai, Zhonghe Zhou, Larry D. Martin, and Alan Feduccia 1995. A Beaked Bird from the Jurassic of China. Nature, 337:616-618. Howgate, M.E. 1984. The Teeth of Archaeopteryx and a Reinterpretation of the Eichstatt Specimen. Zoological Journal of the Linnean Society, 82:159-175. Jollie, Malcolm 1957. The Head Skeleton of the Chicken and Remarks on the Anatomy of this Region in Other Birds. Journal of Morphology, 100:389-436. Kurzanov, S.M. 1987. Avimimidy i problema proiskhozhdenia ptits [Avimimids and the Question of the Origin of Birds]. Trudy, Sovmestnaya Sovetsko- MongoVskaya Paleontologicheskaya Ekspeditsiya, 31:5-95. Madsen, J.H., Jr. 1976. Allosaurus fragilis: A Revised Osteology. Utah Geological and Mineral Survey Bulletin, 109: 163 pages. Marsh, O.C. 1880. Odontornithes: A Monograph on the Extinct Toothed Birds of North America. Memoirs of the Peabody Museum of Yale College, 1: 201 pages. Martin, L.D. 1987. The Beginning of the Modem Avian Radiation. Documents des Lab- oratoires de Geologie Lyon, 99:9-19. Martin, L.D., and J.D. Stewart 1977. Teeth in Ichthyornis (Class: Aves). Science, 195:1331-1332. Martin, L.D., J.D. Stewart, and K.N. Whetstone 1980. The Origin of Birds: Structure of the Tarsus and Teeth. Auk, 97:86-93. Maryariska, T, and H. Osm61ska 1997. The Quadrate of Oviraptorid Dinosaurs. Acta Palaeontologica Po lonica, 42:377-387. McDowell, Samuel 1978. Homology Mapping of the Primitive Archosaurian Reptile Palate on the Palate of Birds. Evolutionary Theory, 4:81-94. Muller, Hans-Joachim 1963. Die Morphologie und Entwicklung des Craniums von Rhea ameri cana Linne, II: Viszeralskelett, Mittelohr und Osteocranium. Zeitschriftfiir Wissenschaftliche Zoologie, 168:35-118. Nemeschkal, H.L. 1983a. Zum Nachweis eines Os coronoideum bei Vogeln?ein Beitrag zur Morphologie des Sauropsiden-Unterkiefers. Zoologische Jahr- buecher, Abteilung fiir Anatomie, 109:117-151. 1983b. Zur Morphologie des Unterkieferskeletes: das Sauropsidenden- tale?ein aus 2 Deckknochen zusammengesetztes Element (Verte brata, Sauroposida). Zoologische Jahrbucher, Abteilung fiir Anatomie, 109:369-396. Norell, M.A., J.M. Clark, D. Dashzeveg, R. Barsbold, L.M. Chiappe, A.R. Davidson, M.C. McKenna, A. Perle, and M. Novacek 1994. A Theropod Dinosaur Embryo and the Affinities of the Flaming Cliffs Dinosaur Eggs. Science, 266:779-782. Osborn, H.F. 1924. Three New Theropoda, Protoceratops Zone, Central Mongolia. American Museum Novitates, 144: 12 pages. Osmolska, Halszka 1976. New Light on the Skull Anatomy and Systematic Position of Ovi raptor. Nature, 262:683-684. Osm61ska, Halszka, Ewa Roniewicz, and Rinchen Barsbold 1972. A New Dinosaur, Gallimimus bullatus n. gen., n. sp. (Ornithomimi- dae) from the Upper Cretaceous of Mongolia. Palaeontologia Po lonica, 27:103-143. Ostrom, J.H. 1969. Osteology of Deinonychus antirrhopus, an Unusual Theropod from the Lower Cretaceous of Montana. Bulletin of the Peabody Museum of Natural History, 30: 165 pages. 1976. Archaeopteryx and the Origin of Birds. Biological Journal of the Linnean Society, 8:91-182. Perez-Moreno, B.P., J.L. Sanz, A.D. Buscalioni, J.J. Moratalla, F. Ortega, and D. Rasskin-Gutman 1994. A Unique Multitoothed Ornithomimosaur Dinosaur from the Lower Cretaceous of Spain. Nature, 370:363-367. Raath, M.A. 1985. The Theropod Syntarsus and Its Bearing on the Origin of Birds. In M.K. Hecht, J.H. Ostrom, G. Viohl, and P. Wellnhofer, editors, The Beginnings of Birds, pages 219-227. Eichstatt: Freunde des Jura Museums Eichstatt. Russell, D.A., and Z. Dong 1993. The Affinities of a New Theropod from the Alxa Desert, Inner Mon golia, People's Republic of China. Canadian Journal of Earth Sci ences. 30:2107-2127. NUMBER 89 323 Smith, David 1992. The Type Specimen of Oviraptor philoceratops, a Theropod Dino saur from the Upper Cretaceous of Mongolia. Neues Jahrbuch fiir Geologie und Paldontologie, Abhandlungen, 186:365-388. Sternberg, R.M. 1940. A Toothless Bird from the Cretaceous of Alberta. Journal of Pale ontology, 14:81-85. Sues, Hans-Dieter 1977. The Skull of Velociraptor mongoliensis, a Small Cretaceous Thero pod Dinosaur from Mongolia. Palaeontologische Zeitschrift, 51:173-184. 1997. On Chirostenotes, a Late Cretaceous Oviraptorosaur (Dinosauria: Theropoda) from Western North America. Journal of Vertebrate Paleontology, 17:698-716. Tarsitano, Samuel 1991. Archaeopteryx: quo vadis? In H.P. Schultze and L. Trueb, editors, Origins of the Higher Groups ofTetrapods, pages 541-576. Ithaca: Comstock Publishing Associates. Weber, Erich 1993. Zur Evolution basicranialer Gelenke bei Vogeln, insbesondere bei Huhner?und Entenvogeln (Galloanseres). Zeitschrift fiir Zoolo gische Systematik undEvolutionsforschung, 31:300?317. Wellnhofer, Peter 1993. Das siebte Exemplar von Archaeopteryx aus den Solnhofener Schichten. Archaeopteryx, 11:1-47. Wetmore, A. 1930. A Systematic Classification for the Birds of the World. Proceedings of the United States National Museum, 76:1-8. Witmer, L.M. 1990. The Craniofacial Air Sac System of Mesozoic Birds (Aves). Zoolog ical Journal of the Linnean Society, 100:327-378. Zusi, R.L., and K.I. Warheit 1992. On the Evolution of Intraramal Mandibular Joints in Pseudodon toms (Aves: Odontopterygia). In K.E. Campbell, editor, Papers in Avian Biology Honoring Pierce Brodkorb. Science Series, Natural History Museum of Los Angeles County, 36:351-360. Early Evolution of Birds: Roundtable Discussion The final day of the Washington, D.C, meeting of the Society of Avian Paleontology and Evolution (7 June 1996) was organized by Peter Wellnhofer and was devoted to Me sozoic birds and avian origins. The morning session consisted of individual paper presen tations, some of which appear elsewhere in this volume. The afternoon session was given over to a roundtable discussion in which the participants were gathered in an atmosphere conducive to audience participation. The roundtable consisted of three parts, each with a discussion leader who outlined the problems to be considered and who invited comment. The entire roundtable was recorded on audio- and videotapes, which were later used by the discussion leaders to produce the following partial transcripts and summaries. There was broad participation in the discussion. Those who are cited directly in the tran scripts are listed below in alphabetical order by surname. Buhler, Paul, Universitat Hohenheim, Stuttgart, Germany Chatterjee, Sankar, Texas Tech University, Lubbock, Texas, United States Chiappe, Luis M., American Museum of Natural History, New York, New York, United States Elzanowski, Andrzej, University of Wroclaw, Poland Forster, Catherine, State University of New York, Stony Brook, New York, United States Gatesy, Steve M., Brown University, Providence, Rhode Island, United States Goslow, G.E., Brown University, Providence, Rhode Island, United States Kurochkin, Evgeny, Paleontological Institute, Moscow, Russia Martin, Larry D., University of Kansas, Lawrence, Kansas, United States Mayr, Gerald, Forschungsinstitut Senckenburg, Frankfurt am Main, Germany Naples, Virginia, Northern Illinois University, De Kalb, Illinois, United States Olson, Storrs L., Smithsonian Institution, Washington, D.C, United States Ostrom, John H., Yale University, New Haven, Connecticut, United States Parkes, Kenneth C, Carnegie Museum, Pittsburgh, Pennsylvania, United States Paul, Gregory S., Baltimore, Maryland, United States Peters, D. Stefan, Forschungsinstitut Senckenburg, Frankfurt am Main, Germany Sereno, Paul C, University of Chicago, Chicago, Illiniois, United States Wellnhofer, Peter, Bayerische Staatssammlung fur Palaontologie, Munich, Germany Witmer, Lawrence M., Ohio University, Athens, Ohio, United States Zhou, Zhonghe, University of Kansas, Lawrence, Kansas, United States 325 New Aspects of Avian Origins: Roundtable Report Lawrence M. Witmer Introduction The convening of the fourth International Meeting of the So ciety of Avian Paleontology and Evolution (SAPE) in the year 1996 is significant in that it coincides with the anniversary of several important dates in the history of the debate on the ori gin of birds. It marks the seventieth anniversary of the publica tion of Gerhard Heilmann's (1926) The Origin of Birds, a vol ume that established the orthodox view?that birds descended from basal archosaurs in the Triassic?for the next 50 years. Furthermore, 1996 marks the twentieth anniversary of the pub lication of John H. Ostrom's (1976) magnum opus "Archaeop teryx and the Origin of Birds," a comprehensive treatment ar guing cogently that Archaeopteryx and all other birds are derived from coelurosaurian theropod dinosaurs. Finally, 1996 is the tenth anniversary of the publication of Jacques A. Gauth ier's (1986) paper "Saurischian Monophyly and the Origin of Birds," a widely cited work that, among other things, offered critical cladistic support for the theropod affinities of birds. This paper is not intended as a review of avian origins but rather as a report of the proceedings of an SAPE roundtable discussion organized by Peter Wellnhofer and moderated by myself on 7 June 1996. I was charged by Dr. Wellnhofer to provide the roundtable discussants with a brief overview of current notions on the origin of birds and then present several topics for discussion. I will first expand somewhat on the over view of current opinion to enable readers with less background to follow the discussion. Then the discussion topics will be in troduced and their rationale presented. The relevant portion of the ensuing roundtable discussion will be reported after the in troduction of each topic. The discussion itself was fairly wide- ranging, and participants often commented on more than one discussion topic. As a result, I will not present the report in its strict chronological order, but rather in the order of the discus sion topics. Participants were aware that the proceedings were being recorded on audio- and videotape for subsequent report Lawrence M. Witmer, Department of Biomedical Sciences, College of Osteopathic Medicine, Ohio University, Athens, Ohio 45701, United States. in this volume. Quotes herein are direct transcriptions from the audiotape, with trivial editing (e.g., deletion of false starts or midstream rephrasing) to enhance flow. Paraphrasings also de rive from the audiotape. Overview of Current Opinion on the Origin of Birds As mentioned, it is beyond the scope of this report to review the history of the debate. I previously provided a summary up through the late 1980s (Witmer, 1991), and Feduccia (1996) brought the review up to the present. The modem debate is typ ically characterized as a trio of hypotheses?the "pseudosu- chian thecodont" hypothesis, the crocodylomorph hypothesis, and the theropod hypothesis?with theropod relationships holding sway and the other views decreasing somewhat in pop ularity. Several important developments have arisen in the in tervening years, however, suggesting that opinion has not fully consolidated around the conventional theropod hypothesis. It is not the intent herein to provide a critical evaluation of these hy potheses but rather simply to present a thumbnail sketch and provide references. 1. Relationships with basal archosauriforms ("pseudosu- chian thecodonts," to use the old paraphyletic taxonomy) were suggested originally by Broom (1913), and this was the idea popularized by Heilmann (1926). The basic premise is that Tri assic archosauriforms, such as Euparkeria, are "sufficiently primitive" to have been ancestral to birds (and to other groups of archosaurs, as well). Although revived by Tarsitano and Hecht (1980; see also Tarsitano, 1991), the idea was widely criticized, particularly by supporters of theropod relationships (e.g., Thulbom and Hamley, 1982; Gauthier and Padian, 1985), for being uninformative and for offering few or no supporting synapomorphies. It had seemed that this view had passed away?principally because it was so nonspecific?until a re cent paper by Welman (1995), who proposed numerous syna pomorphies from the basicranial region of the skull, suggesting that Euparkeria is closer to avian ancestry than anyone ever thought. 2. A close relationship with crocodylomorphs, such as the Triassic form Sphenosuchus, was originally proposed by Walk er (1972) and was supported by L.D. Martin (e.g., 1991) and 327 328 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY his students (see Witmer, 1991, for references). Supporting characters included aspects of tympanic pneumaticity, cranial circulation, and dental morphology and replacement. The cro- codylomorph hypothesis was challenged on a number of counts (Gauthier, 1986) and received an apparent deathblow when Walker (1985) himself apparently recanted. Interestingly, Walker (1990, pers. comm., 1995) essentially recanted his re cantation and offered renewed support for avian relationships with crocodylomorphs. 3. Nonarchosauriform archosauromorphs, such as the Trias sic form Megalancosaurus, have been suggested to be close to avian ancestry by a number of authors (e.g., Hecht and Tarsi tano, 1982; Martin, 1983; Tarsitano, 1985, 1991), most force fully by Feduccia (1996; see also Feduccia and Wild, 1993). In all formulations, the origin of birds is tightly linked with the or igin of flight, which is presumed to have required an initial ar boreal phase. Therefore, it is reasoned, because avian ancestors must have been small and quadmpedal, bird-like forms, such as Megalancosaurus (and also Longisquama, Cosesaurus, or Scleromochlus), make good models for avian ancestors (Feduc cia and Wild, 1993; Feduccia, 1996). 4. Theropod dinosaurs are certainly the group most com monly cited as being involved in the origin of birds (Witmer, 1991; Chiappe, 1995); however, the specific nature of the rela tionship, that is, which specific group of theropods is closest to birds, remains controversial. Ostrom (1976) proposed that dro- maeosaurid coelurosaurs, such as Deinonychus and Velocirap tor, were closest to birds based on a large suite of derived char acters, principally from the manus and pelvic limb. This hypothesis received cladistic support from Padian (1982), Gauthier (1986), and Holtz (1994) and is the most commonly encountered version of the theropod hypothesis. Alternate ver sions differ in which clades are hypothesized to be the sister group of Archaeopteryx and/or other birds (see Witmer, 1991, for additional references): coelophysoid ceratosaurians such as Syntarsus (Raath, 1985), troodontid coelurosaurs (Currie, 1985, Paul, 1988), bullatosaurs (troodontids+ornifhomimo- saurs) (Thulbom, 1984), or oviraptorosaurs (Elzanowski, 1995, this volume). Sorting out this confusion will require a compre hensive and up-to-date phylogenetic analysis of Coelurosauria, itself involving a very careful analysis of many characters. 5. Under the broad heading of "the theropod hypothesis" is G.S. Paul's unique formulation (Paul, 1984, 1988). In Paul's view, not only are birds phylogenetically nested within Theropoda, but in fact some forms traditionally interpreted as nonavian theropods are actually secondarily flightless "proto- birds." Paul (1988) envisioned a lineage of protobirds begin ning in the Jurassic with Archaeopteryx and becoming even more bird-like in the Cretaceous, culminating in true birds (Metomithes, to use Chiappe's (1995) terminology). Along the way, the protobird lineage repeatedly gave off clades of terrestrial, secondarily flightless forms, such as Dromaeosau ridae, Oviraptorosauria, Omithomimosauria, and Troodon- tidae. For support, Paul (1988) cited characters from the skull and pelvic limb and offered additional evidence at the 1996 SAPE conference. This hypothesis has received scant atten tion in the literature. 6. A related notion is G. Olshevsky's (1994) "Birds Came First" (BCF) theory. This hypothesis suggests that the avian lineage is a tmly ancient one. That is, archosaur phylogeny is characterized by a "central line" of persistently arboreal, qua dmpedal "dino-birds" that, beginning in the Permian, continu ously gave off branches of terrestrial archosaurs throughout the Mesozoic Era. These secondarily terrestrial clades went on to become the various clades of archosauriforms (e.g., proterosu- chians, aetosaurs, sauropodomorph dinosaurs, etc.). Forms like Megalancosaurus and Longisquama are very close to this cen tral line and never left the trees. This central line of arboreal dino-birds became progressively more bird-like through time and thus so did their terrestrial descendants. Theropods are on the central line, and thus, as in Paul's (1988) formulation, the Cretaceous bird-like theropods are deemed secondarily flight less forms. Also like Paul's hypothesis, Olshevsky's ideas have been virtually ignored in the literature. The Roundtable Discussion Six major topics were presented at the roundtable for discus sion. The topics were chosen to stimulate debate, to examine critical issues, and, it was hoped, to reach agreement on at least some points. Again, each topic is briefly outlined below to set up the ensuing discussion. 1. THE CENTRAL ROLE OF Archaeopteryx IN THE DEBATE The history of the debate on avian origins, almost since its inception, has been focused on Archaeopteryx. In fact, Archae opteryx has been the key player in not just the origin of birds but in virtually all ancillary debates: the origin of flight (Rayner, 1988; Feduccia, 1993, 1996; Herzog, 1993), the ori gin of feathers (Parkes, 1966; Dyck, 1985), the origin of endot- hermy (Ruben, 1995), and others. An entire conference and the resulting volume (Hecht et al., 1985) were devoted to Archae opteryx and its impact on these questions. Moreover, Archae opteryx has importance beyond its technical significance as a symbol of organic evolution. As Ostrom stated during the roundtable, "the Berlin specimen [of Archaeopteryx] is the most valuable and most famous specimen of anything." Given this historically central role, the discussion topic posed to the roundtable participants was whether or not this role is deserved. The first sentence of Ostrom's (1976:91) pa per states, "The question of the origin of birds can be equated with the origin of Archaeopteryx,'''' which clearly articulates the feeling that if we can understand Archaeopteryx, we will auto matically understand the origin of birds (and the origin of flight, etc.). The avian status of Archaeopteryx is an unstated assumption of most analyses. The worry is that if all argumen tation is founded on this assumption and this assumption is NUMBER 89 329 proved questionable or even invalid, then an enormous amount of scientific discourse will have to be called into question. The stakes are quite high. Interestingly, the history of the debate (Witmer, 1991) shows a persistent minority arguing that Ar chaeopteryx may not be part of the true avian clade but rather is a feathered dinosaur (e.g., Lowe, 1944; Thulbom, 1975, 1984; Thulborn and Hamley, 1982; Barsbold, 1983; Kurzanov, 1985). The intent of raising the issue about the central role of Archaeopteryx was to nurture healthy skepticism and to offer the opportunity to reinforce (or dispute) its avian status. The discussion opened with G.S. Paul taking up the issue he presented in his poster and abstract, namely, that Archaeop teryx is skeletally a small dromaeosaurid and perhaps not a tme bird at all. Paul began by doubting that Archaeopteryx had the features of avian craniofacial kinesis suggested by Elzanowski and Wellnhofer (1996), citing the presence of a complete pos torbital bar and a strong maxillary-lacrimal contact, both of which would prevent intracranial mobility; furthermore, Paul questioned their interpretation of bird-like features in the ptery goid. "Years ago when I saw the Eichstatt skull," Paul contin ued, "I thought that I saw an essentially theropod skull, and I believe that with the newest skull this is, in most ways, tmer than I ever thought.... I don't really see very much evidence of anything avian in the skull of Archaeopteryx. Except, as Elza nowski and Wellnhofer [1996] have pointed out, apparently the palatine is fairly avian [in being] triradiate and having a small palatine hook [i.e., the vomeropterygoid or choanal process]. But even there, some theropods get very close to that. For ex ample, dromaeosaurs have virtually no fourth process, the maxillary process of the palatine. Postcranially, again, Archae opteryx is very, very similar to dromaeosaurid theropods. The main features that are avian are in the forelimb and, as pointed out today [in Zhou and Martin's talk], particularly in the wrist and hand?and those are features associated with flight. I hadn't really realized until very recently how extremely similar Archaeopteryx is to dromaeosaurs in very detailed characters." To illustrate this point, Paul distributed handouts derived from his poster and led the participants through the intricacies of a single character, the twisting of the paroccipital process, which is very similar in Archaeopteryx and dromaeosaurids like Ve lociraptor and is unlike other archosaurs, with perhaps the ex ception of Mononykus. "This is what we're getting down to now," Paul continued. "We're getting down to little tiny details shared by dromaeosaurs and Archaeopteryx" L.M. Witmer suggested that Paul's comments primarily pro vided "further evidence, I think many of us would say, support ing that birds are related to small theropods, in particular dro maeosaurs. [But the issue is] not necessarily what are the features that Archaeopteryx shares with dromaeosaurs, but what are the features that Archaeopteryx shares with other birds?" A. Elzanowski responded that Archaeopteryx has "very well-defined avian characters in the skull," such as those asso ciated with the palatine and pterygoid. He went on to enumer ate features in Archaeopteryx that are unique and that set it apart from dromaeosaurids. For example, the pterygoid of Ar chaeopteryx is "so different from a typical theropod or dro maeosaurid pterygoid that we [he and Wellnhofer] had prob lems, I admit, in identifying what is the left and what is the right element. No one would have any problems of this sort with [theropods given] John Ostrom's excellent documentation of dromaeosaurids.... The [pterygoid] wing that Greg [Paul] wants to see as an ectopterygoid process is certainly not an ec topterygoid process. ...The quadrate part of the pterygoid is radically, dramatically different from the dromaeosaurids? The skull is in many characters dramatically different from any known theropod.... The nasal cavity has very peculiar struc tures that are very difficult to compare with anything known so far. The pterygoid has an absolutely peculiar longitudinal divi sion which is very hard to interpret and to compare with any thing else." Elzanowski argued that molecular systematics pro vides insight into the importance of weighing characters, such that "characters like bending of the paroccipital process are simply not comparable, and can never outweigh a radical, dra matic difference in, for example, the palatine bone, which is definitely avian in Archaeopteryx and is clearly theropodan in dromaeosaurids." Furthermore, he suggested that the presence of an avian palatine reflects significant transformation of the skull and evolution of an avian kinetic apparatus. J.H. Ostrom argued passionately for the significance of the specimens of Archaeopteryx, yet he also noted that "the magni tude of Earth's history is enormous. With a handful of speci mens, you think you're going to draw conclusions about who evolved from whom?" In a similar vein, K.C Parkes offered, "With Archaeopteryx we have a snapshot?a snapshot of a brief moment in time A hell of a lot of things must have happened between the time of our still arguable ancestral form [and Archaeopteryx}. ...We have absolutely no evidence of what happened up to the point of that snapshot in time, which means we have to take Archaeopteryx for what we have.... The argument back and forth?is it a bird or not?seems to me al most fruitless because we don't know what came [even] half a million years before Archaeopteryx. So that to some extent, all of the conjecture as to where Archaeopteryx came from is go ing to be very fruitless until we can find something that's a lot closer to Archaeopteryx in time than anything we have now." The role of Archaeopteryx continued to be debated by the participants but as part of other discussion topics, which appear in their appropriate contexts. 2. THE ROLE OF THE CRETACEOUS AVIAN RADIATIONS IN THE DEBATE Certainly part of the reason Archaeopteryx has been so im portant is that for very many years it had been almost the only relevant Mesozoic bird {Hesperomis was too aberrant and Ich thyornis was too "modem" to be pertinent). With the numerous new discoveries of Early Cretaceous (perhaps even Late Juras sic) birds in Spain and China, the database has changed dramat- 330 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ically. Because most analyses (e.g., Chiappe, 1995; Sanz et al., 1995) suggest that these birds are more closely related to mod em birds than is Archaeopteryx, what relevance do these new discoveries hold for the debate on avian origins? The discussion began with L.M. Chiappe, who suggested that "the role of the Cretaceous avian radiations, in my point of view, is very clear. Without disregarding the data that Ar chaeopteryx provides, I think that we actually don't need Ar chaeopteryx right now, for example, to support the idea that birds are descended from theropod dinosaurs. ...We have enormous support from this Cretaceous radiation." Likewise, P.C. Sereno argued for the critical role played by the new dis coveries of Cretaceous birds: "Recently, even aside from these possible Late Jurassic-earliest Cretaceous forms, new birds have presented other combinations of characters that include even more advanced avian characters while still retaining things like gastralia and things that we've never seen in bird like creatures before." Some participants saw the Cretaceous avian radiations as helping to refine and redefine the role that Archaeopteryx plays in the debate. For example, Elzanowski suggested that "barely anything points so much to the central role of Archaeopteryx? its central position in [the] evolution of birds?than this record of Sinornis and Confuciusomis. [Sinornis is] perfectly interme diate between Archaeopteryx and more modem birds.... So, if let's say Sinornis is intermediate between modem birds and Ar chaeopteryx, therefore, by purely logical reasoning, Archaeop teryx has to be central to the evolution of birds in morphologi cal terms." For Sereno, the combination of Archaeopteryx and the Early Cretaceous birds presents "really a nice phylogenetic situation. I think for a cladist to look at Archaeopteryx, it addi tionally presents a strong argument for the origin of birds be cause it has so few autapomorphies. When you put it up on a cladogram, you try to see what are the characters that are unique to itself and to help to map its phylogenetic informa tion. There are so few that we almost want to call it a metatax- on (something that you can't actually link the specimens to gether by [apomorphic] features). I think that's the important thing, to reiterate what Luis [Chiappe] is saying, that we've got confirmatory evidence from other animals." G.S. Paul respond ed that "it is very possible that morphologically Archaeopteryx basically is a theropod dinosaur with wings It is very possi ble thai Archaeopteryx maybe was allied with dromaeosaurs or was a completely independent development from birds. On the other hand, what Paul [Sereno] just said is also tme?it's so primitive that it could be at the base of the bird radiation In a way, we really don't know whether Archaeopteryx has a cen tral role or not?we do not have the information yet." 3. THE THEROPOD HYPOTHESIS AND THE "TIME PROBLEM" Although the theropod hypothesis has been the most popular one for more than twenty years, it has always faced what may be regarded as "the time problem" (Witmer, 1995; see also Fe duccia, 1996), namely, the most bird-like of the nonavian theropods (e.g., dromaeosaurids, troodontids, oviraptorosaurs) are younger in age than Archaeopteryx. If the conventional hy pothesis is correct, then birds and nonavian coelurosaurs di verged in at least the Jurassic. Where, for example, are the Ju rassic dromaeosaurs? How disturbed should we be by this discordance in the fossil record? Does it severely damage the theropod hypothesis, as has been suggested (Tarsitano, 1991; Feduccia, 1994, 1996)? The discussion of this topic was limited. Sereno, who has studied the temporal ranges of theropod clades in conjunction with the pattern of phylogenetic branching, acknowledged that "there is a time discordance between Archaeopteryx and its nearest sister group. But when you look at the overall phyloge ny of theropods, there are many time discordances?but also many missing lineages with much greater length than that actu ally. For example, if we look at the origin of coelurosaurs, we now have radiometrically dated allosaur-like animals for the Lower Jurassic. We know that there was a coelurosaur lineage at the base of that radiation for which we have no evidence for the Jurassic, essentially until the Late Jurassic. So, we're miss ing maybe 20-30 million years of early coelurosaur evolution before we get to the point where we were talking about Archae opteryx and these other things. So, it's not that unusual. It seems that small theropods in general are your worst case ex treme for taphonomists, because you don't have the option usu ally of lake beds or near-shore marine sedimentary localities, but neither do you have the size that will often carry you through in a fluvial environment. So, you fall in-between the cracks in a very poor record." 4. THE SIGNIFICANCE OF Protoavis FOR THE DEBATE From the time of its discovery, Archaeopteryx was regarded as both the oldest and the most primitive bird. Reports of Trias sic avian remains from Texas (Chatterjee, 1987, 1991, 1995, this volume) would appear to challenge one or both of these claims. According to Chatterjee's (1991) cladogram (see also Kurochkin, 1995), Protoavis is closer to the modem radiation than is Archaeopteryx. In other words, Archaeopteryx would remain the basal member of Aves, but not the oldest. Thus, what is the significance?even relevance?of Protoavis for the debate on avian origins? Obviously it would make the time problem of topic three, above, much worse, telescoping much of theropod cladogenesis into the Norian or even Camian. Oth erwise, Protoavis might behave phylogenetically much like the components of the Cretaceous avian radiation (topic two, above). The discussion began with S. Chatterjee, who saw the time problem as less of an issue, suggesting that "we're caught up in a stratophenetic approach We are very content with Archae opteryx?this is the primitive one. You can derive anything from it. When the new evidence comes, look at it. Look at the bones. I think what it tells us is that, like mammals, there was a NUMBER 89 331 very bush-like radiation of birds." Chatterjee argued that the real significance of Protoavis resides in its prospects for estab lishing a skeletal definition of birds: "Once you define it, then there is no problem. Vox Archaeopteryx or for birds it is really a circular argument: we are defining on feathers. Do we define mammals on hair? No. ...We need some practical, tangible ev idence preserved in the fossils so that we can call it 'bird.' Once you define it, then you can see whether Archaeopteryx falls under the definition or not The time has come: we have to give the osteological definition of birds. For that matter, I think Protoavis really has a much, much better chance. You can define birds on the basis of the quadrate. You can define birds on the basis of the cheek region If you can document that the orbit and the two temporal openings are confluent, it is a bird." Sereno stated that the significance of Protoavis cannot be ad equately assessed until the professional community takes a se rious approach to the fossils: "It seems that most people ignore Protoavis, and I think that this is a sad situation. I think there's a lot of very different opinions about what Protoavis is, and some of these have been aired. [But] if we're going to move on the significance of Protoavis, it probably would be in having some type of a consortium with the fossil material, with people actually commenting on what they think it is in a serious-sci ence forum." 5. THE ORIGIN OF BIRDS VERSUS THE ORIGIN OF FLIGHT In some formulations (Tarsitano, 1991; Feduccia, 1993, 1996; Feduccia and Wild, 1993), the origin of flight and the or igin of birds are inextricably united: flight "from the ground up" with the theropod hypothesis and flight "from the trees down" with more basal archosaurs. The protocol appears to be to develop a concept of the hypothetical proavis based on one's notion of the origin of flight and then survey the animal king dom for a match; that is, the functional inference precedes the phylogenetic inference. The intent of this topic is not to exam ine the origin of flight, but rather to discuss the necessity of coupling these two issues. In other words, what is the relation ship of phylogenetic inference to functional inference? The discussion began with Elzanowski, who proposed, "the strict coupling of theropod/'from-the-ground-up' and alterna- tive-hypotheses/'from-the-trees-down' is not really warranted. I think that the discussion of the taxonomic origin of birds should be decoupled from the mechanics?the evolutionary mechanism?of the origin of flight. As all of us probably agree, we really don't know, in a strict sense, the ancestor of birds?we can't agree which are the closest theropod relatives of birds. We have no idea [of their] size or what those ancestors looked like. We know that they certainly were smaller than ba sically all the dinosaurs we have fossils of." Elzanowski argued that, as observed in mammals, small theropod dinosaurs would have had much more "flexible ankles" than large dinosaurs. This is "a known generalization? There is no reason to ques tion that there were arboreal or slightly arboreal theropods that would just climb on the tree or mn on the tree trunk and [then] just jump and glide from the tree trunk." Chiappe agreed that the two should be decoupled, saying, "The kind of data that we have is completely different. For the origin of birds, [it] is exclusively phylogenetic. We have a lot of data. We have fossils we can measure, look at, and examine. The origin of flight is a totally different question?a very inter esting one, but the kind of data that we have is certainly ten times more speculative. ...First, we should come up with an idea, a notion, about the origin of birds,... and then try to see how we can explain the origin of flight within the framework of that particular idea." Sereno likewise argued "that the two are very separate, because when you start looking at the prob lem phylogenetically, only some of the characters that are link ing these animals together into an evolutionary sequence actu ally are related to flight. Some of the most interesting things are the characters that were co-opted but were not evolved for flight in the first place. We have the extraordinary opportunity, with the great functional work that's being done and a series of fossils, to go at this functional transformation like we cannot in the case of bats and pterosaurs. We can actually tease apart the functional sequence, but all of the characters are not related to that functional sequence, so the two are pretty separate." P. Wellnhofer provided some important cautionary remarks about, again, over-reliance on Archaeopteryx, commenting that "we have to be careful in our conclusions. I think it's not so im portant what lifestyle Archaeopteryx as an animal really had. Maybe [it] could even climb or sit on a tree or on a tree branch or something like that. I think what's more important is the general architecture of the skeleton. The lifestyle of Archaeop teryx [itself] can be quite different from what we suggest." 6. THE VALIDITY OF "NONSTANDARD" HYPOTHESES As in probably all areas of human endeavor, science tends to eschew the iconoclastic in favor of familiar things from famil iar sources. In the present case, the "nonstandard" views of Paul and Olshevsky seem to be examples of this phenomenon in that they reverse the typical ancestor-descendant relation ship, derive from individuals that are outside the "fold" of uni versity and museum professionals, and have not been published in the conventional outlets. As mentioned, these views have been almost totally ignored. Ironically, both views agree with the current orthodoxy that birds and theropods are very closely related and, moreover, present the advantage that the time problem disappears. The intent of this topic is to examine the status of these views in the current debate. The discussion was limited to a statement by L.D. Martin: "One of the things about this conference that I've found ex tremely interesting is how many of the papers that were pre sented today could be taken to support Gregory Paul's so- called 'nonstandard' hypothesis. I would say he's getting so much support that we can view it as a school?'the Paulian 332 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY School of Bird Origins.' The only thing I see that it lacks for a confirmation would be the discovery of a Cretaceous dinosaur with enlarged feathers...and I would really think that we would have very strong support for Paul's viewpoint." Recapitulation and Conclusions Perhaps the best quote from the roundtable came from S.L. Olson: "There is no hypothesis involving the origin and evolu tion of birds that's too ridiculous that somebody won't propose it." This sentiment was shared by many of the participants, al though?and this is the interesting part?there would not be much agreement as to which hypotheses were the ridiculous ones. The goal of the roundtable was not to establish winners and losers, or to be able to come away with a broad consensus on avian ancestry. The goal was to raise issues, discuss them openly, and establish some common ground, and in this the roundtable was very successful. The role of Archaeopteryx received a rare critical appraisal. There was general agreement that Archaeopteryx will continue to merit a cmcial role in not only this debate but in all the de bates associated with the early radiation of birds. In an impor tant departure from the past, however, Archaeopteryx may slowly be heading toward a more appropriate position as only one of a number of important fossil taxa. The rapidly growing number of Early Cretaceous (and perhaps even Late Jurassic) discoveries, some species of which are represented by dozens of complete skeletons with feathers, are tremendously helpful in reducing the weight of inferences that Archaeopteryx must bear. Furthermore, these Cretaceous fossils provide important corroborating information with regard to the origin of birds such that Archaeopteryx apparently could be dropped from many analyses with little resultant change in the phylogenetic pattern of avian ancestry. Several synapomorphies of Archae opteryx and "tme" birds were discussed. Nevertheless, the sta tus of Archaeopteryx as a tme bird was challenged by other participants, and, given the controversial status of a number of taxa discussed at the conference (e.g., "protobirds," Monon- ykus, oviraptorosaurs, new Malagasy fossils), perhaps it is in deed pmdent to exercise caution about all taxa positioned phy- logenetically near that transitional nexus. For many partici pants, it is likely that the roundtable ultimately did little to diminish either the avian status or the importance of Archae opteryx. For others, the phylogenetic position of at least Ar chaeopteryx remains somewhat more uncertain. As for myself, I continue to regard Archaeopteryx as the basal member of Aves, while at the same time recognizing that I have been wrong before. As mentioned above, the recent discoveries of indisputable Cretaceous birds were widely seen as contributing very impor tant new data for the origin of birds. They confirm findings previously based solely on Archaeopteryx and provide new in sights as well. The time problem facing the theropod hypothe sis was discussed, and it was pointed out that the fossil record is rife with similar (and even worse) time discordances and that a stratophenetic approach is inappropriate. Perhaps the broad est level of agreement was that the functional issue of the ori gin of flight needs to be clearly separated from the phylogenet ic issue of the origin of birds, although the discussion perhaps was hampered by the absence of several of the chief propo nents of the linkage of these issues. Nevertheless, several par ticipants voiced strong opinions that the issue of phylogenetic origin logically and methodologically precedes the exploration of models on the origin of flight. There was little focused dis cussion on what we should do with nonstandard hypotheses such as those of Paul and Olshevsky, although it was clear that Paul's ideas received an open hearing with perceptions ranging from receptive to skeptical. In general, there was virtually no discussion of any hypothe ses other than the theropod hypothesis, which received strong support from several participants. This situation probably gen uinely reflects the broad acceptance that this notion has, but it probably also reflects the fact that several key proponents of al ternative views were not in attendance, whereas most of the theropod principals were present. There was only passing dis cussion of the alternate versions of the theropod hypothesis with the only notable outcome being the survival (even thriv ing) of Paul's protobird hypothesis. Finally, it is worth reporting that the roundtable was remark ably noncontentious. Participants listened patiently to the views of others, offered their responses with often amiable good humor, and generally seemed receptive to different ideas. That is not to say that strong views were not expressed, but only suggests that the tenor of the debate has moderated and moved onto a more professional and thoughtful level. Literature Cited Barsbold, Rinchen 1983. [Carnivorous Dinosaurs from the Cretaceous of Mongolia.] Trudy, Sovmestnaya Sovetsko-Mongol 'skaya Paleontologicheskaya Eks peditsiya [Transactions of the Joint Soviet-Mongolian Paleontologi cal Expedition], 19:5-119. [In Russian, with English summary.] Broom, Robert 1913. On the South-African Pseudosuchian Euparkeria and Allied Genera. Proceedings of the Zoological Society of London, 1913:619-633. Chatterjee, Sankar 1987. Skull of Protoavis and Early Evolution of Birds. Journal of Verte brate Paleontology, supplement, 7(3): 14A. 1991. Cranial Anatomy and Relationships of a New Triassic Bird from Texas. Philosophical Transactions of the Royal Society of London, series B, 332:277-342. NUMBER 89 333 1995. The Triassic Bird Protoavis. Archaeopteryx, 13:15-31. Chiappe, Luis M. 1995. The First 85 Million Years of Avian Evolution. Nature, 368: 349-355. Currie, Philip J. 1985. Cranial Anatomy of Stenonychosaurus inequalis (Saurischia, Theropoda) and Its Bearing on the Origin of Birds. Canadian Jour nal of Earth Sciences, 22:1643-1658. Dyck, Jan 1985. The Evolution of Feathers. Zoologica Scripta, 14(2): 137-154. Elzanowski, Andrzej 1995. Cranial Evidence for the Avian Relationships of Oviraptorosauria. Journal of Vertebrate Paleontology, supplement, 15(3):27A. Elzanowski, Andrzej, and Peter Wellnhofer 1996. Cranial Morphology of Archaeopteryx: Evidence from the Seventh Skeleton. Journal of Vertebrate Paleontology, 16(l):81-94. Feduccia, Alan 1993. Evidence from Claw Geometry Indicating Arboreal Habits of Ar chaeopteryx. Science, 259:790-793. 1994. The Great Dinosaur Debate. Living Bird, 13(4):28-33. 1996. The Origin and Evolution of Birds. 420 pages. New Haven: Yale University Press. Feduccia, Alan, and Rupert Wild 1993. Birdlike Characters in the Triassic Archosaur Megalancosaurus. Natunvissenschaften, 80:564?566. Gauthier, Jacques A. 1986. Saurischian Monophyly and the Origin of Birds. In Kevin Padian, editor, The Origin of Birds and the Evolution of Flight, pages 1-55, 9 figures. San Francisco: California Academy of Sciences. Gauthier, Jacques A., and Kevin Padian 1985. Phylogenetic, Functional, and Aerodynamic Analyses of the Origin of Birds and Their Flight. In Max K. Hecht, John H. Ostrom, Gunter Viohl, and Peter Wellnhofer, editors, The Beginnings of Birds, Pro ceedings of the International Archaeopteryx Conference, Eichstatt, 1984, pages 185-197, 3 figures. Eichstatt: Freunde des Jura-Muse ums, Eichstatt. Hecht, Max K., John H. Ostrom, Gunter Viohl, and Peter Wellnhofer 1985. The Beginnings of Birds, Proceedings of the International Archae opteryx Conference, 1984. 382 pages. Eichstatt: Freunde des Jura- Museums, Eichstatt. Hecht, Max K., and Samuel F. Tarsitano 1982. The Paleobiology and Phylogenetic Position of Archaeopteryx. Geo bios, special report, 6:141-149. Heilmann, Gerhard 1926. The Origin of Birds. 210 pages, 142 figures. London: Witherby. Herzog, Karl 1993. Stammesgeschichtliche Entwicklung des Flugvermogens der Vogel. Archaeopteryx, 11:49-61. Holtz, Thomas R., Jr. 1994. The Phylogenetic Position of the Tyrannosauridae: Implications for Theropod Systematics. Journal of Paleontology, 68:1100-1117. Kurochkin, Evgeny N. 1995. Synopsis of Mesozoic Birds and Early Evolution of Class Aves. Ar chaeopteryx, 13:47-66. Kurzanov, Sergei M. 1985. The Skull Structure of the Dinosaur Avimimus. Paleontological Journal, 1985(4):92-99. Lowe, Percy R. 1944. An Analysis of the Characters of Archaeopteryx and Archaeomis; Were They Reptiles or Birds? Ibis, 86:517-543. Martin, Larry D. 1983. The Origin of Birds and of Avian Flight. In Richard F. Johnston, ed itor, Current Ornithology, 1:106-129, 5 figures. New York: Plenum 1991. Mesozoic Birds and the Origin of Birds. In Hans-Peter Schultze and Linda Trueb, editors, Origins of the Higher Groups ofTetrapods: Controversy and Consensus, pages 485-540, 51 figures. Ithaca: Cornell University Press. Olshevsky, George 1994. The Birds First? A Theory To Fit the Facts. Omni, 16(9):34-36, 38, 40, 42^13, 80, 82-84. Ostrom, John H. 1976. Archaeopteryx and the Origin of Birds. Biological Journal of the Linnean Society, 8:91-182. Padian, Kevin 1982. Macroevolution and the Origin of Major Adaptations: Vertebrate Flight as a Paradigm for the Analysis of Patterns. In Bernard Mamet and M.J. Copeland, editors, Third North American Paleontological Convention, Proceedings, 2: 387-392. Toronto, Canada. Parkes, Kenneth C. 1966. Speculations on the Origin of Feathers. Living Bird, 5:77-86. Paul, Gregory S. 1984. The Archosaurs: A Phylogenetic Study. In Wolf-Ernst Reif and Frank Westphal, editors, Third Symposium on Mesozoic Terrestrial Ecosystems, Tubingen, 1984: Short Papers, pages 175-180, 1 fig ure. Tubingen: Attempto Verlag. 1988. Predatory Dinosaurs of the World. 464 pages. New York: Simon and Schuster. Raath, Michael A. 1985. The Theropod Syntarsus and Its Bearing on the Origin of Birds. In Max K. Hecht, John H. Ostrom, Gunter Viohl, and Peter Welln hofer, editors, The Beginnings of Birds, Proceedings of the Interna tional Archaeopteryx Conference, Eichstatt, 1984, pages 219-227, 2 figures. Eichstatt: Freunde des Jura-Museums, Eichstatt. Rayner, Jeremy M.V. 1988. The Evolution of Vertebrate Flight. Biological Journal of the Lin nean Society, 34:269-287. Ruben, John A. 1995. The Evolution of Endothermy in Mammals and Birds: From Physi ology to Fossils. Annual Review of Physiology, 57:69-95. Sanz, Jose L., Luis M. Chiappe, and Angela D. Buscalioni 1995. The Osteology of Concornis lacustris (Aves: Enantiornithes) from the Lower Cretaceous of Spain and a Reexamination of Its Phyloge netic Relationships. American Museum Novitates, 3133:1-23. Tarsitano, Samuel F. 1985. The Morphological and Aerodynamic Constraints on the Origin of Avian Flight. In Max K. Hecht, John H. Ostrom, Gunter Viohl, and Peter Wellnhofer, editors, The Beginnings of Birds, Proceedings of the International Archaeopteryx Conference, Eichstatt, 1984, pages 319-332, 6 figures. Eichstatt: Freunde des Jura-Museums, Eichstatt. 1991. Archaeopteryx: Quo Vadis? In Hans-Peter Schultze and Linda Trueb, editors, Origins of the Higher Groups ofTetrapods: Contro versy and Consensus, pages 541-576, 16 figures. Ithaca: Cornell University Press. Tarsitano, Samuel F., and Max K. Hecht 1980. A Reconsideration of the Reptilian Relationships of Archaeopteryx. Zoological Journal of the Linnean Society, 69(2): 149?182. Thulbom, Richard A. 1975. Dinosaur Polyphyly and the Classification of Archosaurs and Birds. Australian Journal of Zoology, 23:249-270. 1984. The Avian Relationships of Archaeopteryx, and the Origin of Birds. Zoological Journal of the Linnean Society, 82(1-2): 119-158. Thulbom, Richard A., and Timothy L. Hamley 1982. The Reptilian Relationships of Archaeopteryx. Australian Journal of Zoology, 30:611-634. Walker, Alick D. 1972. New Light on the Origin of Birds and Crocodiles. Nature, 237: 257-263. 334 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 1985. The Braincase of Archaeopteryx. In Max K. Hecht, John H. Ostrom, Gunter Viohl, and Peter Wellnhofer, editors, The Beginnings of Birds, Proceedings of the International Archaeopteryx Conference, Eichstatt, 1984, pages 123-134, 7 figures. Eichstatt: Freunde des Jura-Museums, Eichstatt. 1990. A Revision of Sphenosuchus acutus Haughton, a Crocodylomorph Reptile from the Elliot Formation (Late Triassic or Early Jurassic) of South Africa. Philosophical Transactions of the Royal Society of London, series B, 330:1-120. Welman, Johann 1995. Euparkeria and the Origin of Birds. South African Journal of Sci ence, 91:533-537'. Witmer, Lawrence M. 1991. Perspectives on Avian Origins. In Hans-Peter Schultze and Linda Trueb, editors, Origins of the Higher Groups ofTetrapods: Contro versy and Consensus, pages 427-466, 17 figures. Ithaca: Cornell University Press. 1995. The Search for the Origin of Birds. 64 pages, 28 figures. New York: Franklin Watts. Early Avian Evolution: Roundtable Report Luis M. Chiappe Introduction Few areas of vertebrate paleontology have advanced more over the last few years than that of the early evolution of birds. Recent findings of primitive, non-neomithine birds have been so numerous that we have more than doubled the number of valid taxa described between 1861, when the first early bird, Archaeopteryx lithographica von Meyer, was reported, and 1990. Thus, to address the plethora of new ideas and discus sions that all these new findings have triggered, in the single hour of roundtable discussion that I had been assigned to mod erate, was a daunting, if not impossible, task. With this in mind, and after discussing possible topics of debate with other colleagues, I decided to center the discussion on only three top ics within this new profusion of evidence. The aim of this report is not to provide a review of the new data on early bird evolution, nor is it to defend my own views over those of others. Much of the new evidence has already been reviewed, and a variety of choices are available for the in terested reader. Wellnhofer (1994) and Feduccia (1996) pro vide reviews based on a traditional "evolutionary" approach, whereas I have reviewed the new data from a strict cladistic perspective (Chiappe, 1995a). Discussion Topics Before going into the actual debate at the roundtable, it would be helpful to provide a general overview of the three topics that were discussed. 1. PHYLOGENETIC RELATIONSHIPS AND SIGNIFICANCE OF Mononykus.?Mononykus olecranus was first reported by A. Perle, M. Norell, L. Chiappe, and J. Clark on the basis of a par tial specimen from the Late Cretaceous Nemegt Formation of southern Mongolia (Perle et al., 1993). This flightless, turkey- sized animal, with short, stout forelimbs instead of wings, was regarded as phylogenetically closer to modem birds than is Ar chaeopteryx, and it was thus interpreted as a bird (Perle et al., Luis M. Chiappe, Department of Vertebrate Paleontology, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007, United States. 1993, 1994; Norell et al., 1993; Chiappe, 1995a; Chiappe et al., 1996a). Placement of the bizarre Mononykus within birds raised significant debate. Opponents expressed their views both in scientific journals and books (e.g., Patterson, 1993; Ostrom, 1994; Wellnhofer, 1994; Martin, 1995a; Zhou, 1995a; Feduc cia, 1996) and in popular magazines and newspapers (e.g., Fe duccia, 1994; Martin and Rinaldi, 1994; see also Norell et al., 1993; Chiappe et al., 1995, 1996, 1997, for responses to these criticisms), but with the exception of L. Martin, who regarded Mononykus as a bizarre ornithomimid (Martin and Rinaldi, 1994; Martin, 1995a), critics of the avian hypothesis have not proposed an alternative, specific hypothesis of relationships. Moreover, proponents of the avian relationship of Mononykus found additional support for their views in enlarged cladistic analyses (Chiappe et al., 1996; Forster et al., 1996a) that in clude data on new specimens (some preserving nearly com plete skulls) from the Mongolian Djadokhta-like beds of Ukhaa Tolgod (Dashzeveg et al., 1995) and from close relatives of Mononykus found in southern South America (Novas, 1996). In addition, this hypothesis received support from the work done by colleagues performing independent cladistic analyses (e.g., Chatterjee, 1995; Novas, 1996). Another topic of discussion surrounding Mononykus con cerns its life style, namely, whether its short, robust forelimbs were used for digging (e.g., Ostrom, 1994; Zhou, 1995a) or for other activities (Norell et al., 1993; Chiappe, 1995b). Although this appears to be a more trivial issue, it has been used as an ar gument against the hypothesis of avian relationships. For ex ample, Z. Zhou interpreted several of the characters used to support the placement of Mononykus within birds as the result of digging adaptations (Zhou, 1995a), concluding that a dig ging animal cannot be a bird. 2. THE AGE OF Confuciusomis.?For more than a century, and with the only exception being a "feather" of controversial origin (see Bock, 1986), Archaeopteryx lithographica stood alone as the oldest and only known Jurassic bird {Protoavis is left outside this discussion because its avian nature still needs to be confirmed; see Ostrom, 1987, 1996; Chiappe, 1995a). In 1995, L. HOU, Z. Zhou, L. Martin, and A. Feduccia reported on a Chinese bird, Confuciusomis sanctus, from lacustrine depos its of the Yixian Formation in the northwestern Liaoning Prov- 335 336 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ince (Hou, Zhou, Martin et al., 1995; see also Hou, Zhou, Gu et al, 1995; Hou et al., 1996). Although these authors pointed out that the chronology of the Yixian Formation was far from be ing settled, their paper was entitled "A Beaked Bird from the Jurassic of China," and thus Archaeopteryx's "new partner" was heralded as such by the press. Doubtless, in combining a modem-looking, toothless snout with short wings bearing mas sive, large claws, Confuciusomis is of extreme relevance. Yet whether it compares in age with Archaeopteryx or not is an is sue that still needs to be analyzed, especially now that new ra diometric dates have placed the Yixian Formation in the Early Cretaceous, with dates of roughly 121 million years (Smith et al., 1995, 1996). 3. THE PHYLOGENETIC POSITION OF THE ENANTIORNITHES AND THE MONOPHYLY OF "SAURIURAE."?Many of the new early fossils show a number of derived features that were first reported in an array of mostly disarticulated elements from the Late Cretaceous of Argentina, which C. Walker named Enan tiornithes (Walker, 1981). The new cohort of fossils has shown not only that the Enantiornithes are tme birds (confirm ing the perceptiveness of Walker's early observations), but that they formed a large and diverse clade of Mesozoic fliers as well. In his original paper, Walker (1981) regarded the Enantior nithes as phylogenetically intermediate between Archaeop teryx and modem birds. In 1983, L. Martin proposed a basal avian dichotomy leading to modem birds on the one hand, and to Archaeopteryx and the Enantiornithes on the other (Martin, 1983). Martin's characters in support of the close relationship of Archaeopteryx and the Enantiornithes, a group for which he rescued Haeckel's term "Sauriurae," ranged from being re garded as "not one" (Steadman, 1983:342), to "cannot be shown to exist" (Olson, 1985:94), to "either plesiomorphic or uncertain" (Chiappe, 1995b:60). At the same time, the non- monophyletic status of the Sauriurae has been broadly disre garded in numerous cladistic analyses (e.g., Cracraft, 1986; Chiappe, 1991, 1995b, 1996; Sanz and Buscalioni, 1992; Chi appe and Calvo, 1994; Sanz et al., 1995, 1996; Forster et al., 1996a). Yet in recent years, Martin's hypothesis has been re newed with the addition of more characters and defenders (e.g., Hou, Zhou, Martin et al., 1995; Kurochkin, 1995; Mar tin, 1995b; Zhou, 1995b; Feduccia, 1996; Hou et al., 1996). What does not seem to emerge from the discussions of these authors is the realization of the fact that an enormous amount of convergence (and its corollaries) has to be explained for the hypothesis of the monophyly of the Sauriurae to be seriously entertained (see below). The Roundtable Discussion THE PHYLOGENETIC POSITION OF Mononykus.?The debate was opened by L. Martin asking A. Elzanowski "whether Mononykus can be embedded somewhere in the same scheme where we have Oviraptor and ornithomimids." At one of the regular presentations of the Symposium on Mesozoic Birds that morning, Elzanowski had presented a cladogram, based on cranial data, supporting the idea that Oviraptor was closer to modern birds than is Archaeopteryx. In other words, Ovi raptor was regarded as a flightless bird. In responding to Mar tin, Elzanowski posited that Mononykus "would be on an earli er branch than Archaeopteryx," but added, "I would easily agree that Mononykus is closer to birds than a typical theropod... [yet] I cannot provide evidence in support of Mononykus [having] been related to birds." Martin then asked, "Do you think it [Mononykus} is related to Oviraptor?" Elza nowski disregarded that alternative, saying he could not think of any potential synapomorphy between Mononykus and Ovi raptor. Martin asked, "Do you think [Mononykus] is a more advanced bird than Oviraptor?" Elzanowski replied, "I don't think so." This initial exchange between Martin and Elzanowski was followed by J. Ostrom who, with intense democratic spirit, in quired, "How many people here believe Mononykus is a bird, and why?" The almost palpable hesitation of the audience was broken by L. Witmer who, after acknowledging that he had re viewed some of the papers defending the hypothesis of avian relationships (and had seen the material as well), asserted that "they [A. Perle, L. Chiappe, M. Norell, and J. Clark] have ar gued appropriately with the data they have. ... I think they have scored the specimens honestly and put them into their analysis, and they [the specimens] fell out between Archaeop teryx and modem birds." Put another way, Witmer was taking up the issue that part of the disagreement, as L. Chiappe put it, "is more related to methodological issues." The atmosphere of hesitation evolved into one of critical, scientific evaluation of the available data when P. Sereno sur mised that there could be "crucial data from the skull of the excellent specimens, and perhaps that would be the decisive data." Chiappe then projected the slides of two new, nearly complete skulls of Mononykus from Ukhaa Tolgod (collected in 1994 and 1995, and still unpublished), pointing out that "the jugal bar is rod-like... there is not even a slightly ascending process for its contact to the postorbital [bone]... [yet] there is a postorbital like in Archaeopteryx." The audience followed up with numerous questions about specific anatomical fea tures. P. Biihler inquired about the relationships of the two heads of the quadrate to other bones, which articulated with both the braincase and the squamosal, and was puzzled by the fact that the external nares open at the tip of the snout, saying, "It looks like a kiwi." S. Chatterjee asked about the shape of the orbital process of the quadrate, which is broad as in other basal birds (e.g., Archaeopteryx, Enantiornithes, Patagop- teryx) and, as Chiappe remarked, "not a pointy, typical orni- thurine quadrate." Elzanowski and S. Olson followed with questions about the condition of the dentition, to which Chi appe responded, "there are teeth in the mandible and those are set in a groove...their crowns are not serrated...they look quite like those of birds.... [There] may be a few teeth [in the NUMBER 89 337 maxilla], but those would be in the very anterior tip." P. Wellnhofer commented that "Archaeopteryx's teeth are not serrated, but they have a sharp edge that mns to the tip," and he asked Chiappe, "can you see anything like this in Mononykus?" Chiappe agreed that in Mononykus, as in Ar chaeopteryx, "there is a carina going throughout the edge." Martin, however, disagreed with this, and showed slides of the teeth of both the Aktien-Verein and London Archaeopteryx specimens, pointing out that "the base of the [tooth] of Ar chaeopteryx is as broad as or broader than the crown itself. ...This is the antithesis of what we see in Mononykus. ...Mononykus' teeth are identical to the teeth of Pelecanimimus [see Perez-Moreno et al., 1994], the Spanish Lower Cretaceous ornithomimid." Although Martin was cor rect in that the dentition of Pelecanimimus is very bird-like, with teeth lacking serrations, his remarks on the teeth of Ar chaeopteryx satisfied neither Elzanowski, who pointed out that "most of the other teeth [of Archaeopteryx] don't show any indication of expansion of the roots," nov Archaeopteryx expert Wellnhofer, who concurred with Elzanowski's view point. The issue of the life style of Mononykus was not discussed, although after noting that its hyoids were very well devel oped, Olson pointed out, "so this [Mononykus] is a termite eater." This was an interesting observation because it matches the suggestion made by Norell et al. (1993) that Mononykus may have used its forelimbs to tear apart insect nests, and it also coincides with ideas suggesting digging activities for the forelimbs of this animal, although not implying that it was fossorial. Sereno wondered about the "other related material from Ar gentina" and its implications for the phylogenetic placement of Mononykus. Chiappe stated that "there are some relatives of Mononykus in the Upper Cretaceous of Argentina [see Novas, 1996],... [but] those are more primitive forms,... and these new findings have demonstrated that some of the characters [used by Perle et al., 1993] are independently derived between Mononykus and more advanced birds,...but even including those taxa, the results are exactly the same" (see Chiappe et al., 1996; Novas, 1996). Chiappe remarked that "[this] is the way we can set this issue, finding new primitive members of this very weird and peculiar lineage,... [but] unfortunately the Ar gentine forms are very incomplete." AGE OF Confuciusomis.?This section of the debate began with Martin discussing the accuracy of the absolute dates pro vided by Smith et al. (1995; see also Smith et al, 1996), stat ing, "We were familiar with the results of the Canadian team and received a copy of the dates before publication." Martin then read the dates from the "lower part of the section [Yixian Formation] Remember that there [are] over 1500 meters of section involved ... [the dates are] 119.5, 119.2, 121.8, 123.1, 120.8. ...These are argon-argon [dates], which means that they can have very high precision;... it does not mean they have very [high] accuracy Now I will read to you [the dates of] the top of the section, which is 1500 meters above: 120.2, 121.8, 122.7. ...There is more variability in any of the sets of dates they got than they have for the entire section. ...They may be very accurate dates, but they behave as the dates of one unit, not a section.... This could happen if you have an in trusive event; in other words, these are all of the same event." Later on he stated that "at least some of these [basalts] are in trusive." It must be said, though, that the 1500 meters of thick ness mentioned by Martin are not for the Yixian Formation but for the entire Jehol Group, which includes three additional for mations (see Smith et al., 1995:1427), and that glaucony dates from the sedimentary rocks in between these two basalt levels also provided comparable dates (Smith et al., 1995, 1996). Martin made a valid point by questioning the accuracy of the Ar-Ar dates provided by Smith et al. (1995, 1996), yet that did not address the main point, which was, as Chiappe argued, "what is the data supporting a Late Jurassic age?" Martin con tinued: "If you look at our paper very carefully, you will see that we said that we felt that the ages were controversial," al though this consideration was omitted in the title. Z. Zhou fol lowed up, saying, "I don't know if this is Late Jurassic or Early Cretaceous. ...The reason we thought it could be Late Jurassic is based on absolute dating, potassium-argon [dates] from a dif ferent area supposed to be the same formation... not from the same locality." Sereno rightly argued that "having argon-argon dates from basalts, you could not ask for anything more, that's the best I think that great attention should be paid to these basalts." The discussion branched off to the ages of Archaeopteryx and Sinornis (Sereno and Rao, 1992) relative to Confuciusor- nis. Wellnhofer stated that "the correlation [between the Soln- hofen limestones and the Yixian Formation] may not be possi ble on biostratigraphic evidence. ...We rely on the absolute dating that can be applied in China but cannot be applied in Solnhofen." Martin agreed and argued that "even if we were sure that we have dated the unit [Yixian Formation] correctly,... we may still not know what the relative age is to Archaeopteryx." There was, however, a clear agreement that Confuciusomis is younger than Archaeopteryx and, on strati graphic grounds, is older than Sinornis. THE MONOPHYLY OF "SAURIURAE."?The discussion start ed with Sereno, who inquired as to whether this hypothesis "has been seriously entertained... after decent skeletons of Enantiornithes [have been found]." The answer is yes. G. Paul cheered up the Sauriurae affair with his statement that "it is possible... though very remote... that dromaeosaurs, Archae opteryx, and troodontids formed a clade with the Enantiorni thes, separated from Mononykus and other birds. ...A few characters may suggest that that may be tme, but it requires massive convergence in the flight apparatus and also in the skull, [which] may be more serious." His remarks on conver gence were reiterated by Chiappe, who stated that "[if Sauri urae is going to be accepted,] there are a number of characters, certainly flight correlated, that have to be assumed to have 338 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY evolved independently twice." Martin expanded on his thesis that the structure of the metatarsals and distal tarsals was es sentially different in Enantiornithes and Ornithurae: "In Ar chaeopteryx and the enantiornithine birds, the proximal end of the tarsometatarsus fuses; the distal end, however, does not. This is tme even of Maastrichtian Enantiornithes." Chiappe pointed out that this was tme "except for Avisaurus gloriae [see Varricchio and Chiappe, 1995] from the [Campanian] Two Medicine Formation, which has some fusion [distally]." Martin continued, "In all modem birds without exception, the metatarsal bones begin to fuse distally, and this fusion then moves forward to the proximal articulation. ... Modem birds built an epiphysis; that epiphysis is created by one or more distal tarsals,...but it makes a cap. ...The proximal end of metatarsal III is wedge shaped.... In Archaeopteryx and Enan tiornithes the metatarsal bones are together in a row, and they don't build a tarsal cap;.. .you can literally follow the metatar sals up, look at the proximal end of the articulation and see the ending.... So my argument is that indeed Archaeopteryx and the enantiornithine birds and modern birds all have fused metatarsal bones, but the way they put together the ontogenet ic constraints are different." He then added with emphasis, "I would call this a fundamental way to discover convergence," and continued, "In all modern ornithurine birds...there is a single prominence from the ischium.... In Archaeopteryx and all the enantiornithine birds you get a double prominence,... a little square thing that comes up... and then there is a little tri angular process behind that. ... The triangular process is ho mologous with the stmcture in ornithurine birds, the other structure is not found in ornithurine birds, it is found in all enantiornithines." After Martin's arguments, E. Kurochkin followed in the same vein, reaffirming the metatarsal thesis as well as arguing that the articulation between the coracoid and scapula is dif ferent (reversed) in the Enantiornithes and Ornithurae. Unfor tunately, by that point time had mn out, and there was no pos sibility of rebutting on morphological grounds. The interested reader can find a specific analysis of the evidence in support of and against the monophyly of the Sauriurae in Chiappe (1995b), or can simply analyze the character distribution pro vided in various cladistic analyses (e.g., Cracraft, 1986; Chi appe and Calvo, 1994; Sanz et al., 1995; Chiappe et al., 1996). The discussion was closed by C. Forster and S. Peters. For ster presented a new, spectacular specimen from the Late Cre taceous of Madagascar that combines an ulna with quill knobs, a long tail, and a typical dromaeosaur/troodontid, sickle- clawed digit II of the foot (see Forster et al., 1996b). Peters showed a specimen of Confuciusomis (recently acquired by the Senckenberg Museum in Frankfurt) that proves that the tail of this early bird was not long (as reconstmcted by Hou, Zhou, Martin et al., 1995) but short, ending in a pygostyle (see Peters, 1996). Concluding Remarks Although the roundtable discussion was played out in an arena of cordiality, it was evident that the different method ological approaches of the participants (cladists versus noncla- dists) were clouding the debate on the interpretation of the ac tual evidence. The methodological miscommunication was more apparent when analyzing the phylogenetic position of Mononykus and the monophyletic status of the Sauriurae. A great many of these misunderstandings appeared to center around the criteria used toormulate and test homology and the way in which phyloge netic statements are justified. For example, the hesitation in ac cepting Mononykus as a bird appears to stem more from the fact that its overall aspect (most notably its forelimbs) and its presumed fossorial life style (e.g., Ostrom, 1994; Zhou, 1995a) are at odds with the stereotypical view of a bird than from the critical evaluation of the distribution of anatomical characteris tics among taxa. As has been shown by several researchers (e.g., Perle et al., 1993; Chatterjee, 1995; Chiappe et al., 1996; Forster et al., 1996a; Novas, 1996), cladistic analyses that have used complete data sets have concluded that, in contrast to any initial intuition, Mononykus is closer to modem birds than is Archaeopteryx. Clearly, those arguing against the hypothesis of avian affinities were understanding homology as being validat ed by overall similarity (both morphological and functional) rather than by congruence of derived characters (see Hall, 1994; Shubin, 1994, for a discussion of the homology con cepts). These different approaches to the concept of homology were best portrayed by Martin. After remarking upon the dif ferent ontogenetic pathways of Archaeopteryx and Enantiorni thes (proximal to distal metatarsal fusion) on the one hand and the ornithurine birds on the other (distal to proximal metatarsal fusion) in his defense of the monophyly of the Sauriurae, he emphatically declared, "I would call this a fundamental way to discover convergence." Again, the conflict between different homology concepts ("biological homology" versus "phyloge netic homology"; see Shubin, 1994) becomes apparent. Martin prefers to assume the convergent evolution of the flight appara tus (among other features) in the Sauriurae and Ornithurae over the equally possible alternative of similar developmental con straints evolving independently in Archaeopteryx and the Enantiornithes. We are living in an exceptional period of discovery. With several early birds being described every year, new ideas are being formulated at a pace that exceeds our ability to blend them into a theory structured over this burst of new evidence. Methodological miscommunication stands as another obstacle in this process of assimilation. Clearly, fmitful discussions such as this roundtable, along with a better understanding of the methodological differences between us, can put us one step closer to the most exciting goal of reaching a sound, com prehensive theory of the early evolution of birds. NUMBER 89 339 Literature Cited Bock, W. 1986. The Arboreal Origin of Avian Flight. In K. Padian, editor, The Ori gin of Birds and the Evolution of Flight. Memoirs of the California Academy of Sciences, 8:57-72. Chatterjee, S. 1995. The Triassic Bird Protoavis. Archaeopteryx, 13:15-31. Chiappe, L.M. 1991. Cretaceous Avian Remains from Patagonia Shed New Light on the Early Radiation of Birds. Alcheringa, 15(3-4):333-338. 1995a. The First 85 Million Years of Avian Evolution. Nature, 378: 349-355. 1995b. The Phylogenetic Position of the Cretaceous Birds of Argentina: Enantiornithes and Patagopteryx deferrariisi. In D. Stefan Peters, editor, Acta Palaeomithologica: 3 Symposium SAPE; 5 Internatio nale Senckenberg-Konferenz, 22-26 Juni 1992. Courier For- schungsinstitut-Senckenberg, 181:55-63. 1996. Late Cretaceous Birds of Southern South America: Anatomy and Systematics of Enantiornithes and Patagopteryx deferrariisi. In G. Arratia, editor, Contributions of Southern South America to Verte brate Paleontology. Miinchner Geowissenschaftliche Abhandlun gen, series A, 30:203-244. Miinchen: Verlag Dr. Friedrich Pfeil. Chiappe, L.M., and JO. Calvo 1994. Neuquenornis volans, a New Upper Cretaceous Bird (Enantiorni thes: Avisauridae) from Patagonia, Argentina. Journal of Vertebrate Paleontology, 14(2):230-246. Chiappe, L.M., M.A. Norell, and J.M. Clark 1995. Comment on Wellnhofer's New Data on the Origin and Early Evo lution of Birds. Compte Rendus de I'Academie des Sciences, Paris, series 2a, 320:1031-1032. 1996. Phylogenetic Position of Mononykus from the Upper Cretaceous of the Gobi Desert. Memoirs of the Queensland Museum, 39: 557-582. 1997. Mononykus and Birds: Methods and Evidence. Auk, 114(2): 300-302. Cracraft, J. 1986. The Origin and Early Diversification of Birds. Paleobiology, 12 (4):383-399. Dashzeveg, D., M.J. Novacek, M.A. Norell, J.M. Clark, L.M. Chiappe, A. Davidson, M.C. McKenna, L. Dingus, C. Swisher, and A. Perle 1995. Extraordinary Preservation in a New Vertebrate Assemblage from the Late Cretaceous of Mongolia. Nature, 374:446-449. Feduccia, A. 1994. The Great Dinosaur Debate. Living Bird, 13(4):28-33. 1996. The Origin and Evolution of Birds. 420 pages. New Haven: Yale University Press. Forster, C.A., L.M. Chiappe, D.W. Krause, and S.D. Sampson 1996a. The First Cretaceous Bird from Madagascar. Nature, 382:532-534. 1996b. The First Mesozoic Avifauna from Eastern Gondwana. Journal of Vertebrate Paleontology, 16(3):34A. Hall, B.K., editor 1994. Homology: The Hierarchical Basis of Comparative Biology. 483 pages. New York: Academic Press. Hou, L., L.D. Martin, Z. Zhou, and A. Feduccia 1996. Early Adaptive Radiation of Birds: Evidence from Fossils from Northeastern China. Science, 274:1164-1167. Hou, L., Z. Zhou, Y Gu, and H. Zhang 1995. Confuciusomis sanctus, a New Late Jurassic Sauriurine Bird from China. Chinese Science Bulletin, 40(18): 1545-1551. Hou, L., Z. Zhou, L.D. Martin, and A. Feduccia 1995. A Beaked Bird from the Jurassic of China. Nature, 377:616-618. Kurochkin, E. 1995. Synopsis of Mesozoic Birds and Early Evolution of Class Aves. Ar chaeopteryx, 13:47-66. Martin, L.D. 1983. The Origin and Early Radiation of Birds. In A.H. Brush and G.A. Clark, Jr., editors, Perspectives in Ornithology, pages 291-338. Cambridge: Cambridge University Press. 1995a. The Relationship of Mononykus to Omithomimid Dinosaurs. Jour nal of Vertebrate Paleontology, supplement, 15(3):43A. 1995b. The Enantiornithes: Terrestrial Birds of the Cretaceous. In D. Stefan Peters, editor, Acta Palaeomithologica: 3 Symposium SAPE; 5 In ternationale Senckenberg-Konferenz, 22-26 Juni 1992. Courier Forschungsinstitut Senckenberg, 181:23-36. Martin, L.D., and C. Rinaldi 1994. How to Tell a Bird from a Dinosaur. Maps Digest, 17(4): 190-196. Norell, M.A., L.M. Chiappe, and J. Clark 1993. New Limb on the Avian Family Tree. Natural History (New York), 102(9):38^t3. Novas, F.E. 1996. Alvarezsauridae, Cretaceous Maniraptorans from Patagonia and Mongolia. Memoirs of the Queensland Museum, 39(3):675-702. Olson, S.L. 1985. The Fossil Record of Birds. In D.S. Famer, J.R. King, and K.C. Parkes, editors, Avian Biology, 8:79-252. New York: Academic Press. Ostrom, J.H. 1987. Protoavis, a Triassic Bird? Archaeopteryx, 5:113-114. 1994. On the Origin of Birds and of Avian Flight. In D.R. Prothero and R.M. Schoch, conveners, Major Features of Vertebrate Evolution. Short Courses in Paleontology, 7:160-177. 1996. The Questionable Validity of Protoavis. Archaeopteryx, 14:39?42. Patterson, C. 1993. Bird or Dinosaur? Nature, 365:21-22. Perez-Moreno, B.P., J.L. Sanz, A.D. Buscalioni, J.J. Moratalla, F. Ortega, and D. Rasskin-Gutman 1994. A Unique Multitoothed Ornithomimosaur from the Lower Creta ceous of Spain. Nature, 370:363-367. Perle, A., L.M. Chiappe, R. Barsbold, J.M. Clark, and M.A. Norell 1994. Skeletal Morphology of Mononykus olecranus (Theropoda: Avi- alae) from the Late Cretaceous of Mongolia. American Museum No vitates, 3105:1-29. Perle, A., M.A. Norell, L.M. Chiappe, and J.M. Clark 1993. Flightless Bird from the Cretaceous of Mongolia. Nature, 362:623-626. Peters, D.S. 1996. Ein nahezu vollstandiges Skelett eines urtumlichen Vogels aus China. Natur und Museum, 126(9):298-302. Sanz, J.L., and A.D. Buscalioni 1992. A New Bird from the Early Cretaceous of Las Hoyas, Spain, and the Early Radiation of Birds. Paleontology, 35(4):829-845. Sanz, J.L., L.M. Chiappe, and A. Buscalioni 1995. The Osteology of Concornis lacustris (Aves: Enantiornithes) from the Lower Cretaceous of Spain and a Re-Examination of Its Phylo genetic Relationships. American Museum Novitates, 3133:1-23. Sanz, J.L., L.M. Chiappe, B.P. Perez-Moreno, A.D. Buscalioni, J. Moratalla, F. Ortega, and F.J. Poyato-Ariza 1996. A New Lower Cretaceous Bird from Spain: Implications for the Evolution of Flight. Nature, 382:442^445. Sereno, P., and C. Rao 1992. Early Evolution of Avian Flight and Perching: New Evidence from 340 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY the Lower Cretaceous of China. Science, 255:845-848. Shubin, N.H. 1994. The Phylogeny of Development and the Origin of Homology. In L. Grande and O. Rieppel, Interpreting the Hierarchy of Nature, pages 201-225. New York: Academic Press. Smith, P.E., N.M. Evensen, D. York, S. Cumbaa, and D. Russell 1996. 40Ar-39Ar Laser-Probe Dating of the Yixian Formation, Northeast China and the Early Evolution of Birds. Journal of Vertebrate Pale ontology, 16(3):67A. Smith, P.E., N.M. Evensen, D. York, C. Mee-Mann, J. Fan, L. Jin-Ling, S. Cumbaa, and D. Russell 1995. Dates and Rates in Ancient Lakes 40Ar-39Ar Evidence for an Early Cretaceous Age for the Jehol Group, Northeast China. Canadian Journal of Earth Sciences, 32:1426-1431. Steadman, D. 1983. Commentary on "The Origin and Early Radiation of Birds" (L. Martin). In A.H. Brush and G.A. Clark, Jr., editors, Perspectives in Ornithology, pages 338-344. Cambridge: Cambridge University Press. Varricchio, D.J., and L.M. Chiappe 1995. A New Bird from the Cretaceous Two Medicine Formation of Mon tana. Journal of Vertebrate Paleontology, 15( 1 ):201 -204. Walker, CA. 1981. New Subclass of Birds from the Cretaceous of South America. Na ture, 292:51-53. Wellnhofer, P. 1994. New Data on the Origin and Early Evolution of Birds. Compte Ren dus de I 'Academie des Sciences, Paris, 319(2):299-308. Zhou, Z. 1995a. Is Mononykus a Bird? Auk, 112(4):958-963. 1995b. The Discovery of Early Cretaceous Birds in China. In D. Stefan Pe ters, editor, Acta Palaeomithologica: 3 Symposium SAPE; 5 Inter nationale Senckenberg-Konferenze, 22-26 Juni 1992. Courier Forschungsinstitut Senckenberg, 181:9-22. The Origin of Bird Flight: Roundtable Report G.E. Goslow, Jr. Introduction Our charge during this roundtable disussion was to consider evidence, old and new, regarding the origin of flight in birds. This topic, one that many of us approach with what seems to be a genetically programmed fascination, had been broached nu merous times in both the morning contributed paper session and during the previous two roundtables. To initiate discussion, Sankar Chatterjee kindly agreed to share an illustration sum marizing several competing hypotheses regarding the transition from nonflying to flying forms. The ensuing hour focused on four general questions/themes: (1) were theropods capable of climbing; (2) what can claws tell us; (3) what are the limita tions of the cursorial theory; and (4) do we have the right per spective? My first priority in this record is to reproduce the ideas and thoughts expressed by the participants and, whenever possible, to do so by reporting the conversation verbatim (from audio- and videotapes made with the knowledge and cooperation of the participants). In some instances for clarity, I made editorial alterations that are not intended to change the meaning of what was said. I took the liberty to rearrange the order of some com ments to group them within logical topical headings. The Roundtable Discussion 1. WERE THEROPODS CAPABLE OF CLIMBING??Pondering the various scenarios for the origin of flight, Larry Martin asked Gregory Paul, "Do you think these (theropod) dinosaurs were good climbers or not? I would not have thought so from some of your reconstructions." Paul responded, "A good analo gy would be a jaguar. If I were being chased by a jaguar, the jag could catch me on the ground. If I ran to a tree, the jag could climb the tree and catch me there as well. Jags are about the same size as the dromaeosaurs. The jaguar scenario sug gests the situation for dromaeosaurs. I think they were very good mnners, but I also think they were good climbers, as is tme for many of these small theropods. My theory is that there G.E. Goslow, Jr., Department of Ecology and Evolutionary Biology, Brown University, Providence, Rhode Island 02912, United States. exists a group of small arboreal theropods from the Triassic or perhaps Jurassic that we have not found yet, because they will be very hard to find, that were good climbers. That's where you get Archaeopteryx from as well as some of the Cretaceous theropods. I agree with Sankar Chatterjee that overlapping fields of vision and large brains are not necessary for flight be cause pterosaurs do not have large brains, nor do insects. Nor do pterosaurs have overlapping binocular vision, but primates evolved these things in trees. [There] are other suggestions [that] these theropods were climbing; they had raptorial hands and three-toed feet with reversed hallux trackways, which sug gests they could wrap this stucrure around. Even Tyrannosau- rus has a reversed hallux trackway, so yes, I would agree that many of these small theropods could be semiarboreal forms." "So you do not have any problem with these forms being arbo real?" asked Martin. "No, I would agree, the arboreal hypothe sis is far superior," responded Paul. Steve Gatesy raised a cautionary note regarding tracks and a reversed hallux by adding, "We are finding in the Triassic Greenland forms what we are calling a tetradactyl trackway, where we have shown that a 'reversed-hallux' trackway can ac tually be made by a form without a reversed hallux by plunging the foot into the substrate in a certain way that the toe is not re ally reversed anatomically. We must be careful about looking for perching feet in Triassic forms from trackways." Paul expanded further about theropod design by commenting on their shoulder architecture and by referring to a set of recon structive drawings he provided for the participants. "There have been some misconceptions about the shoulder girdles of dinosaurs. Quadmpedal forms, of course, walked with fore limbs outstretched to the ground and the limbs under the body. In a lot of the theropods, for example Syntarsus and other Cre taceous forms, the shoulder glenoid faces laterally so that the humems can be brought out laterally. When I manipulate the humems in the shoulder of Syntarsus, I can extend the humems laterally and slightly dorsally as I have illustrated. The range of motion in these theropods is very similar to that of Archaeop teryx; there is very little difference. Not until later birds do we see the glenoid facing dorsally so that the wing can be brought higher up over the back. This ability to position the humems 341 342 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY laterally and slightly dorsally is evidence for raptorial adapta tions early in theropods." 2. WHAT CAN CLAWS TELL Us??Kenneth Parkes precipi tated some lively discussion by saying, "I'd like to get back to the previously suggested rejection of trees as a source of gravi ty. I was very impressed at the Archaeopteryx conference with a paper that John will remember, where direct comparisons were made of the claw morphology of bark-climbing and non- bark- climbing organisms?lizards, squirrels, and birds [Yalden, 1985]. He found that the morphology of the claws of the bark-climbing animals differed from the ground-dwelling forms but were exactly like those of Archaeopteryx. After that paper, I recall you John [Ostrom] saying, 'I am convinced.' So if Archaeopteryx did not go up trees, why did it have tree- climbing claws?" Paul quickly reminded us of the potential flaw in arguments that place Archaeopteryx in trees by saying, "There's an alternate hypothesis to trees; there is a scmb cover problem with that hypothesis in that there were no trees. Arid islands with a scmb cover?zero trees." Paul Buhler, however, pointed out, "The problem is that you can have desert and an inland sea and still have a nice forest not too far away. In the Solenhofen near Eichstatt they have found dragonflies and other aquatic insects. That means that not too far away there must have been a forest present which was denser than the one documented in the fossil record. So you cannot tmst the fossil record in the immediate vicinity of the specimen to reflect the ecological situation of the entire surrounding area." "The claws of Archaeopteryx are indeed, in superficial form, similar to something like a woodpecker," added Stefan Peters, "but they are not strong enough. There are claws in animals that do not climb at all; for instance, in some cuckoos you will find claws that look like the claws of climbers. It may be a climber has to have claws similar to this or that, but you cannot reverse this argument. You cannot say, 'If I find an animal with such claws, it must have been a climber.' Lions, for example, have very similar claws but do not climb very often. We pub lished a paper on this; I am not very convinced by this argu ment. As far as I can see, the claws [of Archaeopteryx] are the only argument which remains for the arboreal theory." 3. WHAT ARE THE LIMITATIONS OF THE CURSORIAL THEO RY??After discussing the above evidence, which lends some support for the notion that the flight of birds arose from an ar boreal ancestor, discussion of evidence in support of the "cur sorial theory" was inevitable. Buhler initially asked, "What benefit can you get by mnning along and getting away from the earth [using wings]? There is a problem?the most probable situation is that the animal is a prey which is mnning away. In that case, by jumping from the ground the animal will be giv ing up its energy transformation system. That means it will be getting slower by gliding or flying, and I cannot think of any possibility where mnning and jumping into the air will be ad vantageous." To address the question about possible advantages of leaving the ground through flight, Ted Goslow asked, "Our work on the organization/action of the supracoracoideus in birds has led us to wonder if just the act of getting out of the way, or of tak ing off quickly, could be reason enough. Could early flight have been an erratic behavior to evade predation by jumping from the ground or from a tree for that matter? Does this make any sense?" Storrs Olson responded by saying, "In the case of flying from trees where you are going into a different medium and you are experiencing an optical change it does, but not in a terrestrial situation." "Among students of mammalian locomotion," noted Gos low, "the question of why saltation [richochetal locomotion] as a form of locomotion would ever evolve is often asked. Is not one possible selective advantage thought to be predator evasion through erratic movements?" Virginia Naples indicated that two points need to be consid ered in any discussion of leaping and its relationship to early flight. "If an animal is mnning and intends to jump and remain in the air for any length of time, that animal must get high enough to complete a downstroke, an upstroke, and a second downstroke in order to stay in the air. Secondly, I am con cerned that if you are leaping into the air to escape a predator or startle a predator, you are only going to be successful if you also change direction. That requires a tremendous level of [neural] sophistication in terms of maneuverability in flight, and I do not know if these early forms were capable of this." "We have a modern analog," said Sankar Chatterjee, "the kangaroo. They jump, they leap, but they never use their fore limbs in any way suggesting flight. It doesn't matter how far or how long they jump?they never use their forelimbs." "This brings up another point," said Storrs Olson, "although I do not know how related the origin of flight is to the evolution of bipedality, but when you have bipedal, terrestrial animals, the tendency is always to shorten the forelimbs. You have kan garoos, kangaroo rats, and humans. When animals come down from the trees and assume a terrestrial position, the forelimbs are shortened. This is directly against everything that happens with birds where you have elongation of the forelimbs." John Ostrom recalled that in 1974, "I wondered why Archae opteryx had hands that were designed like 'flyswatters.' I thought maybe the primaries could be used as flyswatters [Os trom, 1974]. Critics at the time did not like this, and I do not blame them; now I have a better understanding of powered flight. A former student, Rick Vazquez, described how the hand of a bird is supinated upon the trochlea carpalis and how this supination acts to streamline the upstroke [Vazquez, 1992]. I illustrated this morning in my presentation how this ability was already present in dromeosaurs, such as Deinonychus and Velociraptor. There is something in the gene pool which allows for this. In modem birds, for example the starling, a wingbeat cycle occurs in just 70 milliseconds. In these small birds at least, the wing must supinate many times per second and it does so automatically. Archaeopteryx had this same ability. It NUMBER 89 343 had the big primaries, secondaries, modem wing feathers, and the automatic streamlining mechanism. Modem birds lose trac tion on the ground, but in air they build up the speed of the up stroke to get to the next power stroke?a dozen times a second. When this transition occurred in birds, I do not know. But Ar chaeopteryx had most of the required arm structures necessary for flight. It has been asked, 'Is gliding required or is it more primitive than powered flight?' This group keeps referring to flight by having the animal get up into the trees and gliding down. Why is the elevation in a tree required when the addi tional lifting power is available by increasing the rate of the wingbeat cycle? Why are those wrists built that way, to climb trees?" Martin asked, "John, do you think flight got started essen tially straight up from a standing start?" Ostrom responded, "No, it has some forward velocity motion by the hindlegs; Ar chaeopteryx is built as a cursorial biped." For clarification, Martin asked, "Do you think the motion of supination enabled the animal to get started right off or are you saying it gained forward velocity from mnning? My question is, is the supina tion motion you are describing adequate in and of itself, or does the animal need velocity from some other means? Are we talking once again about cursorial flight?" Ostrom re sponded, "Most birds can mn and take off too. Many birds walk or run into their flight. They do not all begin from a standing start. I am just saying that the modified carpus was doing something. What?" In response to this question, Olson asked another: " What was it doing in those dinosaurs? They were not flying." To which Ostrom responded, " I do not know. The theropod/bird plan?they all have that carpal plan. Why?" Gerald Mayr asked, "What was the selective advantage of the ability to supinate to an intermediate stage, i.e., to a crea ture with small feathers that was mnning?" "Birds are bipeds and have long forelimbs," offered Paul, "and early forms have raptorial hands like Archaeopteryx. Among archosaurs, the only other forms like that are theropod dinosaurs [Paul, 1988]. The arms of some giant theropods, such as Deinocheirus, were about 10 feet long. As far as the lunate carpal block, every single dinosaur that is a theropod has this lunate and some other avian features in other parts of the skele ton and skull that are not present in Archaeopteryx. This sug gests or implies, but I cannot prove it, that the reason they do have the system is that they are secondarily flightless. A new troodontid from China, Sinomithoides, a photo of which I re produce on my handout, can fold the manus over the radius and ulna well over 90?; it possesses a very good folding mecha nism. Sinomithoides was described by Russell and Dong in 1993." Paul Sereno quickly responded, "I disagree with that inter pretation. There are different interpretations of the carpi posi tions on that specimen; one is slightly higher than 90?, the other is about 90? The specimen is coming at you a little and the photo is deceiving. I have looked at that specimen, and I found evidence it could not retract the manus any more than Archae opteryx" Martin noted, "I was recently asked a question about the long forelimbs of dromeosaurs, so I measured the limbs of a few. When I took off the manus and compared the length of the arm to the hindlimb without the foot, I found that the forelimb bones were all significantly shorter than those of the hindlimb. In primitive animals, one expects to find the forelimbs and hindlimbs to be about equal length. In Archaeopteryx, if you take the manus off and compare the arm length to the hindlimb length, you will find the forelimb is considerably longer than the hindlimb. So in comparison to the primitive plesiomorphic condition, even the dromeosaurs are shortened and Archaeop teryx is elongated like a bird. It seems evolution is going in dif ferent directions." 4. Do WE HAVE THE RIGHT PERSPECTIVE??Through the course of the roundtable, questions regarding the group's per spective or orientation to the question of the origin of flight were expressed. These thoughtful comments do, of course, force us to evaluate our own perspectives and biases and serve to stimulate new lines of investigation. Paul noted, "There is another issue people have not really looked at. I have done cal culations on the number of insects a ground-dwelling, in sect-eating bird the size of Archaeopteryx would have to con sume. Flying insects are a very small package of energy so you have to get a lot of them; something like 100 of them per day if the protobird had an overall energy budget similar to modem Aves. And I did the figures to determine how far it would have to mn and so on to catch these things and the numbers did not work out very well. The foraging range, mnning an average of 10 miles per hour, would be far beyond that observed for ani mals living today. So there are real serious energetic problems with the historic insect-catching hypothesis. Most insectivores the size of Archaeopteryx or bigger tend to feed on insect colo nies so they can have a concentrated resource. Most insecti vores that feed on individual insects are small so they do not have to eat so many each day. The basic insect catching hy pothesis is energetically very implausible." A question was raised, "Where did you get the numbers for the insects?" "I just looked them up," Paul responded. He continued by saying, "There exists no animal today that mns around on the ground and gets the majority of its energy from flying insects. Proba bly energetically this is not a good idea. Plus you are fighting gravity, and the insects are far superior in agility; it is just not a good idea." Andrzej Elzanowski cautioned the group early in the discus sion to be careful about adaptive arguments that simply support either an arboreal theory or cursorial theory for the origin of flight by saying, "I find something in our approach to be very confusing, and I may be speaking for others as well. Today I have heard much about coupling morphology to environmental factors. For example, there is coupling of arboreal adaptations with gliding to support an arboreal origin of flight. I think this linking of one model of the origin of flight with the peculiar 344 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ecological conditions of say, an arboreal habitat, is not really justified. It has been pointed out that there are different elevat ed objects in the environment, not just trees, that can be used to provide gravitation as a source of acceleration for getting up speed and gaining lift. Therefore, I think this dichotomy, this contradistinction, between an arboreal and terrestrial origin of flight is not really justified. I think we should first discuss the gliding model then the cursorial model. These should be con sidered separately. Following this, if we come to the conclu sion, as many have, like Jeremy Rayner [1991], who knows a lot more about this than I do, that the gliding model is aerody- namically much more likely than the cursorial model, we can evaluate the data to see if this model is supported ecologically. But we should not start out with the arboreal versus the cursori al theory per se; this is simply misleading." Peter Wellnhofer concluded the afternoon's discussion with a plea for putting Archaeopteryx into the perspective of an evolved flyer, not as the first bird. His thoughtful comment was: "The focus of our discussion here has centered around the abilities of Archaeopteryx?as an animal could it climb, could it mn, how fast, et cetera? I think these issues are not so impor tant. The early origin of flight happened well before, much ear lier than the evolution of Archaeopteryx. We must not compare the life style of Archaeopteryx?i.e., the environment, the hab itat. It doesn't matter whether there were trees or not; these are absolutely unimportant in the present context. Even if Archae opteryx could climb trees, it doesn't change the general bau- plan of the skeleton of Archaeopteryx, which is a bauplan for bipedal mnning. What is displayed in Archaeopteryx is a later adaptation in the direction of climbing flight." Concluding Remarks While listening to the various points raised and discussed in this roundtable, one could not help but be reminded of the Eichstatt conference on Archaeopteryx held 12 years ago in 1984. Have our opinions about the origin of flight changed since then? If so, in what way? In my estimation, some progress has been made in that the participants seemed not only willing to consider views at odds with their own but were anx ious to entertain new information and approaches. I believe many of us are guardedly optimistic about the promise of new insights into the question of the origin of flight in birds, aided by the additional specimens of Archaeopteryx unavailable to us 12 years ago and by the recent additions to our database of a se ries of Cretaceous fossil birds (for a review, see Chiappe, 1995). Literature Cited Chiappe, L.M. 1995. The First 85 Million Years of Avian Evolution. Nature, 378: 349-355, 6 figures. Ostrom, J.H. 1974. Archaeopteryx and the Origin of Flight. Quarterly Review of Biol ogy, 49:27-47, 10 figures. Paul, G. 1988. Predatory Dinosaurs of the World. 464 pages. New York: Simon and Schuster. Rayner, J.M.V. 1991. Avian Flight Evolution and the Problem of Archaeopteryx. In J.M.V. Rayner and R.J. Wooton, editors, Biomechanics in Evolu tion, pages 183-212, 6 figures. Cambridge: Cambridge University Press. Russell, D.A., and Zhi-Ming Dong 1993. A Nearly Complete Skeleton of a New Troodontid Dinosaur from the Early Cretaceous of the Ordos Basin, Inner Mongolia, People's Republic of China. Canadian Journal of Earth Sciences, 30: 2163-2173, 3 figures. Vazquez, R.J. 1992. Functional Osteology of the Avian Wrist and the Evolution of Flap ping Flight. Journal of Morphology, 211:259-268, 5 figures. Yalden, D.W. 1985. Forelimb Function in Archaeopteryx. In Max K. Hecht, John H. Os trom, Gunther Viohl, and Peter Wellnhofer, editors, The Beginnings of Birds, Proceedings of the International Archaeopteryx Confer ence, Eichstatt, 1984, pages 91-97, 4 figures. Eichstatt: Freunde des Jura-Museums, Eichstatt.