Life Sciences Contribution 1 q 8 Royal Ontario Museum The Non-Passerine Pleistocene Avifauna of the Talara Tar Seeps, Northwestern Peru Kenneth E. Campbell, Jr. ROM ROYAL ONTARIO MUSEUM LIFE SCIENCES PUBLICATIONS INSTRUCTIONS TO AUTHORS Authors are to prepare their manuscripts carefully according to the following instructions. Failure to do so will result in the manuscript’s being returned to the author for revision. All manuscripts are considered on the understanding that if accepted they will not be offered for publication elsewhere. I; GENERAL Papers for publication are accepted from ROM staff members, Research Associates, or from researchers reporting on work done with ROM collections. In exceptional cases,monographic works on the flora and/or fauna of Ontario will be considered for publication by authors not affiliated with the ROM. Authors are expected to write clearly and concisely, and to omit all material not essential for an understanding of the main theme of the paper. . FORMAT Manuscripts are to be typed double-spaced (including captions, synonomies, literature cited, and tables) on 11’’ X 8%”’ paper with a 1%”’ margin on all sides. Three xerox copies are to be submitted to the Chairman of the Editorial Board, and the original retained by the author(s). A separate sheet is to be submitted giving author(s) names, affiliation, title of publication, series in which it is to appear, number of typed pages, number of tables, and number of figures. Manuscripts should normally be organized in the following order: Table of Contents, Abstract, Introduction, Materials and Methods, Results, Discussion, Conclusions, Summary (if paper is long), Acknowledgements, Literature Cited, and Appendices. Authors are encouraged to include foreign language translations of the Summary where appropriate. Headings of sections are to be left-justified to the text margin. The first line of the first paragraph in each new section should not be indented. Text-figures are referred to as ‘‘Fig. 1’. Literature cited in the text is in the form “*Jones (1972)’’ or ‘‘(Jones, 1972)’’ or ‘*(Smith, 1960:71-79, fig. 17)’’. . STANDARD SOURCES The primary source for decisions on format and style is A Guide for Contributors and Editors of ROM Life Sciences Publications, available from the Chairman of the Editorial Board. Otherwise, consult CBE (AIBS) Style Manual (3rd Edition). Other standard sources are as follows: for English spelling (Concise Oxford Dictionary), for Canadian place names and coordinates (Gazetteer of Canada), and for spelling of geographic names (Times [London] Atlas). . ABSTRACT All papers are preceded by a short and factual abstract, about 3 per cent as long as the text, but not longer than 400 words. The abstract is to be followed by four to six keywords enclosed in brackets. . TAXONOMY The name of a taxon is given in full in headings, where it appears for the first time, or when the name begins a paragraph. Use authority and date if appropriate, with first mention of each taxon and not thereafter. Taxonomic papers follow the layout in Life Sciences Contribution 99, particularly the synonomies. . LITERATURE CITED References in the text cite author and date and are enclosed in parentheses (Smith, 1978). Complete references are listed in alphabetical order by author at the end of the paper. When there are two or more citations for an author, the works are listed chronologically. Names of journals are not abbreviated. Consult Life Sciences Contributions beginning with 117 for correct bibliographic form. . TABLES All tables are numbered consecutively in arabic numerals in numerical order of their first mention in the text. Mark the appropriate text location of each table with a marginal notation. Each table is typed on a separate sheet. Avoid footnotes etc., to tables by building them into the title. . FIGURES All figures are numbered consecutively in arabic numerals. Component photographs or drawings are labelled sequentially in upper case letters. Mark the appropriate text location of each figure with a marginal notation. The intended reduction for figures is ideally one and a half to two times. All labelling on figures is in blue pencil and not inked or letraset. Halftones must be photographic prints of high contrast on glossy paper. Authors are to submit 10°’ x 8°’ copies with the MS and retain originals until they are requested. Figure captions are to appear grouped together on a separate page at the end of the MS. LIFE SCIENCES CONTRIBUTIONS ROYAL ONTARIO MUSEUM NUMBER 118 KENNETH E. CAMPBELL, JR. The Non-Passerine Pleistocene Avifauna of the Talara Tar Seeps, Northwestern Peru ROM ROYAL ONTARIO MUSEUM PUBLICATIONS IN LIFE SCIENCES The Royal Ontario Museum publishes three series in the Life Sciences: LIFE SCIENCES CONTRIBUTIONS, a numbered series of original scientific publications including monographic works. LIFE SCIENCES OCCASIONAL PAPERS, a numbered series of original scientific publications, primarily short and usually of taxonomic significance. LIFE SCIENCES MISCELLANEOUS PUBLICATIONS, an unnumbered series of publications of varied subject matter and format. All manuscripts considered for publication are subject to the scrutiny and editorial policies of the Life Sciences Editorial Board, and to review by persons outside the Museum staff who are authorities in the particular field involved. LIFE SCIENCES EDITORIAL BOARD Senior Editor: A. G. EDMUND Editor: J. H. McANDREWS Editor: R. D. JAMES KENNETH E. CAMPBELL, JR. is a Curator at the George C. Page Museum, Natural History Museum of Los Angeles County, California. Canadian Cataloguing in Publication Data Campbell, Kenneth E., 1943- The non-passerine Pleistocene avifauna of the Talara Tar Seeps, northwestern Peru (Life sciences contributions ; no. 118 ISSN 0384- 8159) Abstracts in English, French and Spanish. Bibliography: p. ISBN 0-88854-230-5 pa. 1. Birds, Fossil. 2. Paleontology - Pleistocene. 3. Paleontology - Peru. I. Royal Ontario Museum. If. Title: I. Series. QE871.C35 568.2 C79-094428-6 Publication date: 22 June 1979 © The Royal Ontario Museum, 1979 100 Queen’s Park, Toronto, Canada M5S 2C6 PRINTED AND BOUND IN CANADA BY THE ALGER PRESS ——— —— ogee ee Contents Abstracts 1 Introduction 5 Materials and Methods 8 Geology 9 Site Description 9 Processes of Entrapment 9 Systematic List 16 Discussion and Conclusions 138 Ecology 138 Recent Avifauna 138 Palaeoecology 139 Dating Some Extinct Species 141 Comparison of Avifaunas of Talara Tar Seeps and Rancho La Brea_ 141 Correlation of Talara Tar Seeps with La Carolina, Ecuador 142 Acknowledgements 143 Literature Cited 144 To Pierce Brodkorb The Non-Passerine Pleistocene Avifauna of the Talara Tar Seeps, Northwestern Peru Abstract The Talara Tar Seeps are a localized series of fossil-bearing deposits of late Pleistocene age that are located in northwestern Peru, approximately 04°33’'S, 81°7'W. The deposits lie on a marine terrace believed to be of Sangamon age. They have been dated at ca 13,900 Bee A discussion of processes of entrapment in tar pits suggests that oil seeping onto the ground surface may act as an animal trap in different ways under varying conditions and, for any given oil seep, through time. The effectiveness of any specific oil seep as a trap may also vary for different types of animals at any given time. The non-passerine portion of the palaeoavifauna from the Talara Tar Seeps, i.e., that portion reported on here, consists of over 6,200 specimens representing a minimum of 744 individuals of 13 orders, 24 families, 67 genera, and 89 species. Six new genera and 21 new species are described. The following families are represented, with the number of identified species for each (numbers of new species are in parenthesis): Tinamidae, 1; Podicipedidae, 2; Phalacrocoracidae, 2; Ardeidae, 6(1); Plataleidae, 4(2); Ciconiidae, 2; Anatidae, 9(4); Vulturidae, 6(3); Accipitridae, 7(2); Falconidae, 6(1); Cracidae, 1; Rallidae, 1; Charadriidae, 7(2); Scolopacidae, 14(3); Phalaropodidae, 3(1); Jacanidae, 1; Burhinidae, 1; Thinocoridae, 3(1); Laridae, 2; Columbidae, 4; Psittacidae, 1; Tytonidae, 1; Strigidae, 3; Caprimul- gidae, 2(1). The taxonomic positions of two previously recognized palaeospecies are altered; Archeoquerquedula lambrechti Spillman is synonymized with Anas bahamensis Linnaeus, and Morphnus woodwardi Miller becomes Amplibuteo woodwardi (Miller). Diagnos- tic osteological characters are given in all cases where necessary to justify species identification. The coast of northwestern Peru where the Talara Tar Seeps are located is desert today. Very rare, unpredictable rains may occur during the summer, generally following the occurrence of an oceanic phenomenon known as E/ Nino. During the winter a continuous cloud cover with no precipitation, a condition referred to as a garua, prevails. The non-passerine portion of the palaeoavifauna suggests that during the Wisconsin glaciation northwestern Peru was a savanna woodland, or a savanna with extensive riparian forests. At the end of the Wisconsin, climatic shifts brought about the current desert | conditions in the area, and resulted in the extinction of 23 per cent of the non-passerine species of the avifauna. The palaeoavifauna from the asphalt deposits of Rancho La Brea, California, which are similar in age, includes 98 non-passerine species, compared with the 89 non-passerine species found in the Talara deposits. Sixteen species are common to both localities. Twenty-three per cent of the non-passerine species known from Talara are extinct, in contrast with 17 per cent of the non-passerine species known from Rancho La Brea. The high degree of similarity between the palaeoavifaunas of the Talara Tar Seeps and the asphalt deposits of La Carolina, Ecuador, suggests that deposition may have occurred simultaneously at the two sites. The similarity also suggests that climatic conditions at the two sites were comparable during the Wisconsin, just as they are today; and that the two sites underwent the same desertification process at the end of the Wisconsin. (Pleistocene; avifauna; Talara Tar Seeps; tar pits; Peru; avian systematics) Resumen Los ‘‘Talara Tar Seeps’’ son una serie localizado de depositos fosiliferos de asfalto del Pleistoceno tardio que estan en el noroeste del Peru, aproximadamente 04°33’S, 81°7'W. Los depositos estan sobre una terraza maritima que piensan data de la Sangamon. Los depositos de asfaltos datan de aproximadamente 13,900 anos ante de presente. Una discusion de los procedimientos de entrampamiento en lagunajos de la brea sugiere que petroleo que mana sobre la superficie de la tierra puede ser una trampa de animales en modos diferentes con condiciones variantes y, por una lagunajo particular, durante tiempo. La efectividad de un lagunajo de la brea particular puede variar tambien segun animales de diferentes tipos a un tiempo especificado. La porcion de la avifauna de los ‘‘Talara Tar Seeps’’ que esta discutido aqui, es decir los no paserinos, contiene mas de 6,200 especimenes representando a minimo 744 individuos de 13 ordenes, 24 familias, 67 generos, y 89 especies. Seis géneros nuevos y 21 especies nuevas estan descritos. Las familias siguientes estan representadas, con el numero de especies identificado por cada uno (el numero de especies nuevos por familia esta en parentesis): Tinamidae, 1; Podicipedidae, 2; Phalacrocoracidae, 2; Ardeidae, 6(1); Plataleidae, 4(2); Ciconiidae, 2; Anatidae, 9(4); Vulturidae, 6(3); Accipitridae, 7(2); Falconidae, 6(1); Cracidae, 1; Rallidae, 1; Charadriidae, 7(2); Scolopacidae, 14(3); Phalaropodidae, 3(1); Jacanidae, 1; Burhinidae, 1; Thinocoridae, 3(1); Laridae, 2; Columbidae, 4; Psittacidae, 1; Tytonidae, 1; Strigidae, 3; Caprimulgidae, 2(1). La_ posicion taxonomica de dos paleoespecies reconocidas previamente esta cambiada. Archeoquerquedula lambrechti Spillman esta en sinonimia con Anas bahamensis Linnaeus y Morphnus woodwardi L. Miller esta cambiado a Amplibueto woodwardi (L. Miller). Caracteristicas osteologicas diagnosticas estan presentados en todos casos donde es necesario para justificar la identificacion de una especie. La costa del noroeste del Peru donde estan los ‘‘Talara Tar Seeps”’ es un desierto en el presente. Lluvias muy raras y no predicables pueden ocurrir durante el verano, generalmente sequiente la ocurren- cia de un fenomeno maritimo que se llama El Nino. Durante el invierno una capa ininterrumpida de nubes y neblina, pero sin precipitacion, prevalece. Esta se llama garua. La porcion no paserina de la paleoavifauna sugiere que durante la helamiento Wisconsin el noroeste del Peru fue una sabana arbolada, o una sabana con bastante ribereno. Al fin del Wisconsin cambios climaticos causaron las condiciones de un desierto en la region y dieron por resultado la extincion del 23 porciento de las especies no paserinas de la avifauna. La paleoavifauna de los depositos de asfaltos de Rancho la Brea, California, que son similar en edad incluye 98 especies no paserinas en comparacion con las 89 especies no paserinas de los depositos de Talara. Hay 16 especies que se encuentran en ambas localidades. Ventitres porciento de las especies no paserinas que se conocen de Talara estan extintos, en comparacion con 17 porciento de las especies no paserinas que se conocen de Rancho La Brea. La semejanza de alto grado entre las paleoavifaunas de los ‘‘Talara Tar Seeps’’ y los depositos de asfaltos de La Carolina, Ecuador, sugiere que la deposicion fue simultanea en las dos localidades. La semejanza sugiere tambien que las condiciones climaticas en las dos localidades fueron comparables durante el Wisconsin como en el presente y que las dos localidades sobrellevaron el mismo pro- cedimiento de desierto desarrollado al fin del Wisconsin. Resumé Les “‘Talara Tar Seeps’’ sont des séries localisées de dépots fossiliferes de la derniere partie de la période pleistocéne et se trouvent au nord-ouest de Pérou, approximativement a 04°33’S, 81°7’W. Les depots sont situés sur une terrasse marine, datant sans doute de la période sangamon. Ces dépots datent d’environ 13,900 ans. Une étude sur la fagon dont les animaux ftrent pris au piege suggere que |’infiltrations du goudron sur la surface de la terre petit étre un piege d’animaux de plusieurs manieres, selon de différentes conditions pendant une longue période de temps. L’efficaciteé de chaque infiltration, en tant que piege, varie selon les differents types d’animaux et le moment de |’enlisement. La paléoavifauna des ‘‘Talara Tar Seeps,’’ ne comprenant pas les passereaux est celle que nous eéetudions ici. Elle consiste de plus de 6,200 specimens représentant un minimum de 744 oiseaux de 13 ordres, 24 familles, 67 genres, et 89 especes. Six nouveaux genres et 21 nouvelles especes sont décrits ici. Les familles suivantes sont representées, avec le nombre d’especes, identifi¢es pour chaque famille (nombres des nouvelles espéces entre parentheses): ’ 3 Tinamidae, 1; Podicipedidae, 2; Phalacrocoracidae, 2; Ardeidae, 6(1); Plataleidae, 4(2); Ciconiidae, 2; Anatidae, 9(4); Vulturidae, 6(3); Accipitridae, 7(2); Falconidae, 6(1); Cracidae, 1; Rallidae, 1; Charadriidae, 7(2); Scolopacidae, 14(3); Phalaropodidae, 3(1); Jacanidae, 1; Burhinidae, 1; Thinocoridae, 3(1); Laridae, 2; Colum- bidae, 4; Psittacidae, 1; Tytonidae, 1; Strigidae, 3; Caprimulgidae, 2(1). La position taxonomique de deux paléoespeéces reconnues sont changées: Archeoquerquedula lambrechti Spillmann devient un synonyme de Anas bahamensis Linnaeus, et Morphnus woodwardi L. Miller, devient Amplibuteo woodwardi (L. Miller). Les caracteres ostéologiques sont décrits dans les cas ou il est nécessaire de justifier identification des espeéces. La cote nord-ouest de Pérou ou se trouvent les ‘‘Talara Tar Seeps”’ est une région désertique a notre epoque. Pendant |’été, les pluies sont rares et imprévisibles, faisant suite genéralement a un phénomene océanique connu sous le nom de ‘‘El Nino.’’ Pendant I’hiver, it ne pleut pas, mais le temps est toujours nuageux. Cette condition est connue sous le nom de “‘garua.’’ La portion de la paléoavifauna, (excepte les passereaux) fait penser que pendant la periode glaciaire du Wisconsin, la partie nord-ouest de Pérou était une savanne boisée, avec meme des foréts le long des rivieres. A la fin de la période glaciaire du Wisconsin, le changement de climat amena les conditions désertiques de cette region, et provoqua |’extinction de 23 pour cent des especes dont il est question ici. La paléoavifauna des dépots d’asphalt de Rancho La Brea, California, datant approximativement du méme age, comprend 98 especes (sans compter les passereaux), comparé a 89 dans les depots de Talara. Seize especes se trouvent aux deux endroits. Vingt-trois pour cent des especes connues de Talara sont éteintes, compare a dix-sept pour cent a Rancho La Brea. L’important similarité entre les ‘‘Talara Tar Seeps’’ et des dépots d’asphalt de la Carolina, en Equateur, indique que ces depots datent de la méme période, que les conditions climatiques étaient semblables pendant la periode glaciaire du Wisconsin et qu’elles le sont encore aujourd’hui; elles devinrent désertiques a la fin de la période glaciaire du Wisconsin. Introduction When one considers the large number of publications on fossil mammals from South America it is clear that palaeontologists have long been active and involved with the study of fossils from that continent. Studies on fossil birds from South America, however, have been few and far between. Most of the published works are very old, centred in the decades surrounding the turn of the century. There are at present 77 palaeospecies of birds recognized from South America, with 70 coming from Argentina, three from Ecuador, two from Brazil, and one each from Colombia and Chile (Brodkorb, 1963, 1964, 1967, 1971; Campbell, 1976; Simpson, 1972). While most of the palaeospecies that were described around 1900 were valid species according to the standards considered adequate at that time, a re-evaluation of these taxa using the much more rigorous standards of today would probably result in a substantial reduction in their number. This re-evaluation has recently been accomplished with the fossil penguins by Simpson (1972). The oldest avian palaeospecies known from South America is the Upper Cretaceous diving bird Neogaeornis wetzeli Lambrecht (1929), recorded from two localities in Chile. An additional, as yet undescribed, Cretaceous bird has recently been discovered in Argentina (pers. comm., Pierce Brodkorb). The remaining recognized palaeospecies are distributed temporally as follows: Eocene, 3; Oligocene, 15; Miocene, 29; Pliocene, 12; and Pleistocene, 17. The major published works describing avian palaeospecies from South America are those of Ameghino (1891, 1895, 1899, 1901, 1905), Moreno and Mercerat (1891), Mercerat (1897), Rovereto (1914), Kraglievich (1931), and Simpson (1972). All but 16 of the recognized extinct taxa are described in those papers, the remainder being scattered through the following 14 publications: Ameghino (1882, 1887), Lambrecht (1929), Ameghino and Rusconi (1932), Patterson (1941), Rusconi (1946), Spillmann (1942), A.H. Miller (1953), Howard (1955), Cattoi (1957), Rusconi (1958), Patterson and Kraglievich (1960), Craycraft (1970), Campbell (1976). Approximately 155 neospecies have been recorded as fossils from South America, with at least 25 species being recorded from more than one locality. The largest number (102) were recorded from the late Pleistocene to Recent cave deposits of Minas Gerais, Brazil, by Winge (1887), while the second largest collection described (43 neospecies) was that from the late Pleistocene deposits of La Carolina, Ecuador (Campbell, 1976). The other major papers in which neospecies are recorded are those of Ameghino (1891), Lambrecht (1933), and Wetmore (1935). In addition to those from Brazil and Ecuador, neospecies have been reported from Argentina (14), Bolivia (1), Chile (1), Peru (7), and Venezuela (14). Additional publications reporting the occurrence of neospecies are as follows: Clarke (1882), Moreno and Mercerat (1891), Stresemann (1924), and Kraglievich and Rusconi (1931). The palaeospecies described by Lonnberg (1902) is synonomized herein with a neospecies. The present work grew out of the collections of fossil material from numerous surface tar seeps in northwestern Peru which were made by Dr. A. Gordon Edmund, Dr. Roy Lemon, and R.R. Hornell of the Royal Ontario Museum, Toronto, Canada, during January and February, 1958. The materials collected included extensive 5 samples of plant, invertebrate, and vertebrate remains. The collections came from tar seeps located in the La Brea-Parinas oil field, near Talara, Peru (Fig. 1). These deposits are referred to as the ‘‘Talara Tar Seeps’’ to avoid confusion with the similar deposits at Rancho La Brea, California. This report is the first on the avian remains of the deposit and is concerned with the non-passerine portion of the avifauna. Additional work needs to be completed before the passerine species are satisfactorily identified. The Talara Tar Seeps are located in Departamento de Piura, just west of the Pan American Highway, km 1125, approximately 10 km west of the southern end of the LA CAROLINA Gulf of Guayaquil Punta Parinas* Fig. 1 Map of northwestern Peru showing location of the Talara Tar Seeps. Amotape Mountains, 37 km due east of the Pacific coast at Punta Parinas, the westernmost point of South America, and 18 km southeast of Talara. The site is marked on maps as the town of La Brea, but only a cemetery remains. The tar seeps were first worked by pre-Columbian Indians, who used the natural asphalt in their road construction programme. Spaniards made extensive use of the pitch later, developing it into a considerable economic resource. The first oil well in South America was drilled at the site in 1862, although at the present time there are no active wells nearby. Unfortunately, fossiliferous material is presently being removed and used as patching material for nearby roads. The non-passerine avifauna described herein is quite diverse, containing representatives of 13 orders, 24 families, 67 genera, and 89 species. Of the species recorded, 53 are represented by 5 or more specimens, 43 by 10 or more specimens, 17 by 15 or more specimens, and 10 species are represented by over 100 specimens. The species represented by the most specimens is Anas bahamensis, of which there are 2,864 bones from a minimum of 243 individuals. Unfortunately, 20 of the species are represented by only one specimen. While this does not bear on the accuracy of the identification of these species, a larger sample size is always preferred and is more convincing for a large number of readers. Honest scepticism of conclusions drawn from small sample sizes is quite understandable. However, the identification of each species recorded herein is based on clearly diagnostic material. The habitat requirements of the neospecies represented range from the semi-arid conditions found at the site today (Forpus coelestis, Thinocorus rumicivorus) to year-round standing water surrounded by some type of forest (Dendrocygna autumnalis, Cairina moschata). Many species that require very wet conditions for feeding and/or reproduction, such as the grebes (two species), herons (six species), ibises (three species), and ducks (nine species) are represented. Also requiring similar very wet conditions are the 17 charadriiform species that are winter migrants from North America. Nineteen species of diurnal raptors are represented, most of them probably drawn to the site by the attraction of dead or dying animals trapped in the pools of oil or tar. The size of these predatory birds ranged from the giant condors (three species) to the small kestrel, Falco sparverius. Of this group, two species of the condors have become extinct since the time of deposition of the fossils, as well as one very large vulture, two giant eagles, and one caracara. Three other raptors are recorded that probably represent extinct species, but which must remain unnamed until additional comparative material of living species is available. The owls are represented by four neospecies, all of which are at present widely distributed in North and South America, and elsewhere in some cases. In terms of specimens and individuals the Great Horned Owl, Bubo virginianus, is the most abundant species of owl (110 specimens), but the Barn Owl, Tyto alba, is not far behind (83 specimens). The remains of these two species are more abundant than those of any other species of bird of prey, with the exception of the two species of caracaras which are, or were, diurnal carrion feeders. The doves are also represented by four species, two of which are very common. The Croaking Ground-Dove, Columbina cruziana, is represented by 578 specimens from at least 68 individuals, which makes it the most abundant species of dove present in the collection. This dove is at present found in large flocks in areas with small to moderate vegetation cover throughout northwestern Peru. The remaining species represented are from several families and make up a very diverse group. They are from such groups as the tinamous, cormorants, curassows, rails, thick-knees, gulls, and nighthawks. Again, they represent a group of species with habitat requirements ranging’ from standing water. (cormorants) to forest (curassows) to semi-arid desert (thick-knees). The first palaeospecies of the families Thinocoridae and Caprimulgidae are described herein. Also described as new are six genera and 19 additional species. As indicated by the samples from the Talara Tar Seeps, approximately 25 per cent of the avifauna of northwestern Peru has become extinct during the short time since the fossils were deposited about 14,000 years ago. This paper is intended primarily as a systematic description of the avian species found at the Talara Tar Seeps. An attempt has been made to state clearly how closely related species differ osteologically, so as to facilitate the labour of later workers in the field. In a paper in preparation I will discuss in detail the palaeoecological implications of the palaeoavifauna of the Talara Tar Seeps for northwestern Peru, and northwestern South America. Materials and Methods The fossil specimens from the Talara Tar Seeps were loaned by the Royal Ontario Museum (designated as ROM), Toronto, Canada, to which most will be returned at the termination of this study. A representative sample will be placed in the Collection of Pierce Brodkorb, Department of Zoology, University of Florida. The type of one species of Anas described herein was taken from the collection of fossil birds from the Late Pleistocene of La Carolina, Ecuador. While that collection has been described earlier (Campbell, 1976) I thought it best to delay formally describing this species until it could be done in conjunction with the other new species of Anas. As the material from La Carolina is on loan to me, the holotype is uncatalogued pending its ultimate disposition. Nearly all modern skeletons used for comparative purposes were prepared by maceration. Skeletons that have been prepared by dermestid beetles are often not free of all tendinal materials, and the grease or fats found on the surface may obscure the fine detail of a bone. Osteological terminology was taken primarily from Howard (1929). Fisher (1946), Fisher and Goodman (1955), and Owre (1967) were also consulted. All measurements were taken with vernier dial calipers accurate to 0.05 mm, and rounded off to the nearest 0.1 mm. The following abbreviations are used in the tables and text: M, mean; n, number of specimens measured; OR, observed range. All measurements given in the text are in mm. Geology Site Description The geology of northwestern Peru has been discussed by Bosworth (1922), Iddings and Olsson (1928), Travis (1953), and Lemon and Churcher (1961). Additional information is presented by Sheppard (1937) in his discussion of the geology of Ecuador. The following is a brief summary of the geology of the Talara region based on these works. The current major oil-producing area of Peru is known as the La Brea-Parinas oil field, which straddles the 4th parallel. The oil field occupies the westermost part of the South American continent, and lies between the Pacific Ocean on the west and the Amotape Mountains, a small outlier of the Andes Mountains, to the east. The Amotape Mountains and basement rocks of the coastal area consist of Pennsylvanian quartzites, graywackes, argillites, and shales. Directly upon these lie a series of Cretaceous rocks in excess of 2700 M in thickness. The Cretaceous sediments consist of a lower series of limestones, and an upper series of shales, sandstones, and conglomerates. The main oil-producing rocks are of Eocene age, and consist of approximately 4,600 M of marine clastic sediments. Shales, sands, and a few conglomerates are present in these sediments that originated during the orogenic uplift of the Andes Mountains directly to the east. These sediments are highly faulted and of complex structure as a result of post-depositional orogenic movement. Early, or perhaps Middle, Pleistocene to Recent marine deposits overlie the Eocene rocks and consist of well stratified marls, limestones, calcareous sands, coquinas, and pebble beds. These deposits occupy three wave-cut terraces, named the Mancora Tablazo (highest and oldest), the Talara Tablazo (middle terrace, not much younger than the Mancora Tablazo), and the Lobitos Tablazo (lowest and youngest). The terraces represent periods of marine encroachment followed by orogenic uplift, which was greater in the north than in the south, giving the terraces a southward tilt. The deposits are thickest in the north and thin southward, with a maximum thickness of 16 to 25 M. The Talara Tar Seeps rest on the Mancora Tablazo (Fig. 2). Their age, as determined by radiocarbon dating, has been given as ca 13,900 B.P. (Churcher, 1966). Recent deposits consist of large alluvial breccia fans from the Amotape Mountains, valley terraces, sand dunes, and Recent marine deposits. Processes of Entrapment Of all the different ways that organisms can be fossilized, perhaps the one most likely to preserve the remains of an animal with little or no alteration is for the living animal to become mired in a surface flow of oil, or its distillate, tar. The term ‘‘tar seep’’ is a misnomer because tar itself does not seep from the ground. Rather, it is liquid petroleum, or crude oil, that escapes to the surface through cracks in the rocks that overlie the petroleum deposit. Once the crude oil is exposed to the atmosphere, the lighter, more volatile hydrocarbon compounds evaporate, leaving behind the heavier hydrocarbon compounds that make up what is referred to as tar, and after further distillation, asphalt. Since oil seeps can act as traps in different ways at different times, depending on the condition of the oil or tar and the type of organism 9 ‘¢ ‘Sy ur dn-asojo ut uses ysodap jo adA} Sursusy ‘podeys-o8pom oy) sayeorpul MOLY YoOYs sy], ‘daes [10 Jo adA} [eyUOZLOY ay) Aq powioy syisodap jo aanequasaidos are MOU BuO] ey) Aq payeoIpul eIeNS oY, “YWOU SUTyOO] ‘ni9g WaysamYYOU ‘sdaeg Ie] eIE[e] BY) JO MAIAIOAQ Z ‘BI es RSs eT re . “ W . ~ ~ : SS OO S 10 involved, a short consideration of the ways that oil seeps may serve as traps is appropriate. First, a simple pool of fresh crude oil may be expected to be an effective, although selective, trap. Mammals should be able to distinguish between pools of oil and water before entering them because they usually move slowly on the ground and possess well-developed olfactory organs that allow them to smell the normally odiferous crude oil. If the oil should be floating on water, and under such circumstances perhaps being less detectable, it is reasonable to expect a few mammals to wade into the water and become covered with oil. The size of the animal may then determine if it is able to escape successfully, or be trapped. A pool of oil is a much deadlier trap for birds, however, because from the air a pool of oil looks exactly like a pool of water. Because they move rapidly through the air and generally lack well-developed olfactory organs, birds would perceive the oil as water and alight. A bird could land either in the pool, as would a duck, or on the shore, after which it might wade into the pool, as would a sandpiper. If the pool were pure oil it might be reasonable to expect a wading bird to detect this and stay out after landing on the shore. Detection might be more difficult if there is a thin layer of oil on the surface of the water. In either case, once a bird moves into a pool of oil or oil-covered water it is in serious trouble, primarily because of the oil coating the feathers. This makes the feathers useless for, and even a hindrance to, locomotion. Unlike a mammal that might blunder into a pool of oil and still walk out, a bird cannot fly with oil-covered feathers, as anyone experienced with modern oil spills can attest. Struggling simply ensures that more oil will cover the feathers and hasten the end result, which can only be entombment. If the pool is of water with only a light coating of oil a bird might escape, only to perish elsewhere as a result of poisoning from oil inadvertently swallowed while being preened from the feathers. A second type of trap may be formed after the oil has been exposed at the surface for a period of time and, as a result, has lost the most volatile hydrocarbons. At this stage it forms a more sticky substance, more properly referred to as tar, which, depending on its depth, can be expected to trap any size of mammal or bird. The ‘‘stickiness’’ of the tar would tend to hold the animal, much like flypaper is used to trap flies. Struggling only tends to make an animal mired in the tar sink deeper. An added factor with oil as it approaches the consistency of tar is the degree to which its effectiveness as a trap increases as the air temperature rises. From late evening through the early morning the cool air temperature will tend to solidify the tar, making it quite safe to move around on. If an animal fails to leave the tar before it warms up, or if it should attempt to traverse during mid-day the same area it did during the night, it could very well become trapped. During this stage, the oil, or tar, may act to entrap birds and mammals in a more similar manner than in the first type of trap, and not select between the two groups. Animals trapped by tar at this stage probably act as an attractant that lure carnivores, raptors, and both mammalian and avian scavengers onto the tar. Because tar is much denser than crude oil, it would take a longer time for trapped animals to sink, or to be completely covered if the tar is deep enough, than in a pool of oil. Diurnal predators and scavengers would be attracted to the trapped animals at just those times when they themselves would be likely to be trapped, while nocturnal predators and scavengers would feed at times when the tar would be least ‘‘sticky’’. A third type of trap can be formed when debris covers the tar, making it Il undetectable. This occurs when enough of the volatiles have evaporated, leaving behind residual hydrocarbons dense enough to support sand, dust, leaves, or other light material that may fall on the surface of the tar. Under these conditions the tar is undetectable, and when the air temperature is low, the tar with its camouflaged surface may be able to support almost any size of organism. When warm, however, the debris forms a thin crust which may give way, allowing the animal to sink into the sticky tar under the crust from which it cannot escape. This type of trap is probably of less significance for trapping birds than it is for larger mammals. However, if a large mammal floundered enough to break up the covering crust surrounding it before dying, scavenging birds could then be trapped in the freshly exposed tar when they came to feed on the carcass. An example of a mammal trapped in a deep horizontal flow that had been covered with a crust of sand is shown in Fig. 3. Illustrated here is a goat that had walked out onto the sand and dust covered tar, only to have the crust give way under it. As the goat is only partially buried in the tar, it can act as an attractant for carnivores or scavengers, and only that portion of the skeleton actually buried will be preserved. There are two ways that oil may collect after seeping onto the surface to form a pool that acts as a trap. One way is for the oil to spread over a large horizontal area, forming either a shallow pool or thin coat on the ground surface. A deep pool may form if there is abundant oil. If these types of pools are very shallow they will not be able to effectively trap large mammals, but they will still be very effective in trapping any birds that might land on them believing them to be water. If this type of flow is too shallow to allow the animal to be completely covered, only a portion of the skeleton will be preserved. An example of this partial preservation is illustrated in Fig. 4, which shows an Olivaceous Cormorant (Phalacrocorax olivaceus) that landed in a shallow layer of oil. The bird and the oil surrounding it were covered with some of the sand, silt, or dust, that is blown about a great deal in the semi-arid environment after it was trapped. The skull of this cormorant is already badly weathered and it will not become fossilized. Other parts of the skeleton as well may not become fossilized, since it is only half buried in the tar. An illustration of a fossil deposit that formed from a horizontal flow is shown in Fig. 2. A second way that oil may collect on the surface is for it to fill a natural depression. The commonest depressions to be found on the flat, wave-cut terraces in this region would be the channel bottoms of streams, or gullies. In order for a pool of oil to gather in a stream channel it would have to be dry, or at least carrying only a small amount of water. This condition would be expected if rainfall in the vicinity was limited to a few months a year. If a slow movement of water existed in a channel filled with oil or tar, the effectiveness of the pool as a trap would be increased because mammals that would otherwise stay away might be drawn in by the water. The pools that formed in the channels were probably deeper than the horizontal flows, and were therefore, able to trap larger mammals. It is also likely that there were, at a minimum, riparian forests that followed the streams out of the Amotape Mountains, where they most likely originated. The deepest pools of oil can be expected to form where a stream is undercutting its channel bank, and some of the animals moving about in the vegetation above such a steep slope can be expected to fali into the pools below. An example of the lense-shaped type of fossil deposit that would form in a stream channel is shown in Figs. 5. It is evident from the fossil deposits illustrated in Figs. 2 and 5 that both types of 4 ~~ ST met fg 4 : Se & ESS 13 VJ . . . A I . x SAA HUQ << S\\ ] [eLiaj}eul [Isso JO soURepUNge puke UOTISOdap Jo adA) SuIsua] ZuUTMOYsS 6 yisodap [Isso} oy) JO Mota dn sola ¢ ‘“SI4 | 8) deposits were formed at the Talara Tar Seeps. Although there are no active oil seeps in the area now, oil and water currently escaping from abandoned oil wells attract both vertebrate and invertebrate animals and act as active traps. Windblown sand and dust usually form a crust over most of the oil seeping onto the horizontal surfaces. Persistent large pools of oil occur only in areas where man has excavated the natural asphalt for road material or museum collections. The preservation of the fossils is very good, and in numerous cases broken specimens could be pieced back together, even though the pieces were sometimes separated at the time of collection. The avian specimens range in size from complete humeri of condors to those of swallows. Small specimens are not lacking, as there are an estimated 2,500 to 3,000 specimens of passerines in the collection that are being worked on. Numerous skulls of various species are also present in the collection. The excellent condition of these specimens indicates that there has been little or no post-depositional alteration of the fossils. Since tar becomes very difficult to move once it has passed to the asphalt stage it can be assumed that the present deposits were formed in situ, and that there has been no selective sorting or alteration of the faunal composition once the bones were immersed in the tar. Any process of selection operating to produce more specimens of one species than another would be a direct result of the habitat preferences and/or behavior patterns of the respective species, and which of the three types of traps described earlier was involved. The possibility of predator or scavenger preference for a given species must also be taken into consideration as a selective force working to reduce the number of specimens of a species fossilized. For additional information on entrapment in pools of oil or tar see Stock (1956). Systematic List The following systematic list details the non-passerine species known from the Talara Tar Seeps, as summarized in Table 1. Nomenclature used above the species level is primarily that of Brodkorb (1963, 1964, 1967, 1971). Except where otherwise noted, species distributions are taken from Meyer de Schauensee (1966). The following osteological diagnoses are based on as many comparative specimens of each species that I considered reasonable, or necessary, to obtain. In many cases the series available is admittedly too small, and this should be kept in mind when using osteological characters given for each species. This problem was greatest when comparative skeletons of South American species were involved, as many species are simply not to be found in North American collections. The lack of comparative material from the South American avifauna has been particularly significant in working with certain taxonomic groups, such as the accipitriform species, and it has permitted the referral of three fossil species to genera only because of a lack of available specimens of all species of certain genera. It is hoped the diagnoses will stand the test of time, but I will not be surprised if future work leads to some alterations. 16 In some instances specimens are thought to represent immature individuals. This conclusion is based on the pitted appearance of the surface of the bone and incomplete ossification of the articular facets. No characters are listed for a species if the fossil specimens do not vary significantly from living representatives of the species, and if the species is sufficiently distinct from related species to minimize the possibility of error in identification. Order Tinamiformes (Huxley) Family Tinamidae Gray Crypturellus Brabourne & Chubb Crypturellus cf. transfasciatus (Sclater & Salvin) Pale-browed Tinamou Material Proximal end of one right humerus. ROM 13033. Remarks At the present time C. transfasciatus occurs from the equator in western Ecuador south to Tumbes and Piura in northwestern Peru. It is the only tinamou found in the dry deciduous forests that surround the semi-arid and arid regions of southwestern Ecuador and coastal Peru. The fossil specimen corresponds to the estimated size of the humerus of C. transfasciatus, but as no skeletal material was available for comparison the specimen is only tentatively referred to that species. Order Podicipediformes (Furbringer) Family Podicipedidae (Bonaparte) Podiceps Latham Podiceps dominicus (Linnaeus) Least Grebe Material One complete right coracoid, 1 nearly complete left femur, proximal end of | left femur, proximal end of 1 left tarsometatarsus. ROM 13034-13036, 13100. The 4 specimens represent a minimum of 2 individuals. Remarks The present range of P. dominicus includes northwestern Peru. 17 6] a8ed uo ponunuog Te Ol QS SiO sie si a SoS Gol Oc ™ — ert 10¢ nO -_ pu] weaq snjouloiun oanqodidd Zz ‘ds oaing | ‘ds oaing pwosodjod oaingy, SNINAJOUD]AW SNJADOUDAIDH)y. ‘aou ‘ds 39 ‘uad Ipapgqiy oainqydury,, “AOU ‘ds 39 ‘Udd S1jSasia] DJINbDAL eepundisoy DAND SAJADYID Dx SNIDAID “JD SdKSDIODy ‘aou ‘ds raysif ¢snyduvs0o4v¢ snyddus anqjn, ‘aou ‘ds appapmoy sdks0umdy ‘Aou ‘ds 30 ‘ua3d snnbijad sdk80u0say seplLinynA SOULIOJLNIC IDV snoiuiMop XKUOWON SISUIWDYDG | Vx ‘AOU ‘ds apuajayavdjOUDS * fy ‘aou ‘ds adpjown ‘py ‘AOU ‘dS aDADID] SDUPy, DIDYISOU DUIAIDI paajdouvjau vsvydaojy) “aou ‘ds 39 ‘uad DIDJISIAUI DJJ2UOUUDNy SIJDULININD DUSKIOAPUIG x sepneuy SOULIOJLIOSUY puy cel + Wo] DUDIIAIUD DIAAJIA DIAAJIAU NAIGDS sepITUOsT) vi{olv vivly SNq]D “4A ‘AOU ‘ds snupianiad snuiz0pnq “AOU “ds 1a40UJ9M SNINSIAIY J seplo[eyed XDAODNIAU XDAOINIKN y “AOU “ds IMijuDUoUDS DUS1AKS Dajnsavd “Fz DINY) DIJAIBT snq]D sniposqusvy 10909 vaply seploply SOULIOJIOPIY 1ypiauipsnog dq SNIIDAYO XDAOIOAIDIVDY seplovloooOIOR[eyd SOULOJIULIIIOg sdasipod snquapipog snoiuiuop sdaJ1pog epIpedioipog SOULIOJIPSdIdIpog snvlgsvf{suv.] “Jd SnpjaanjdKgD oeprmwmeul yl SOULIOJIWIEUI J, (OL61 ‘[aqdueD) 10opensy ‘euljores eT] WAY pap.1090.1 OSTe 319M (,,) YSII9}SB UB YIM payseUl saidadg *pajzeoipul (puy) S[ENprIAIpUI JO JaquINU WNUTUTUL puke (Wayq) S}UIUWIITa JO JaquiNnuU yQIM ‘Sdaag JB] BIE yay) WoIy UMOUY Satdads aulJassed-uoU 6g AY} JO ISIT =: FIQU 18 CbL 89 N 8SC9 £6 ‘aou ‘ds sisuasnid snéjnuiidvg 9 smuadynov sapiapsoy) oepisinuuded SouLojIs nuded ‘all snaumunvpf oisp 9 piapjnoiund okjoads cOl snupiuis41a oqng SePISINS €8 DqID OI I epIuojA |, SOULIOJIZINS I $118a]a09 snd10 4 seploeyIsg SOULIOJIOBI ISG L6S DUDIZNAID * De € Noovd]y] vuiqunjog L DINDISD “Z ZOI DIDINI1AND DPIDUIZ » seprquin[o) SOULIOJIQUINTOD 6 uvoxidid “J, I€ DIJ1214JD SNADT sepue’y] I ‘aou ‘ds apayodaoy * I SNAOAIINUNA SNAOIOULY J x I ‘yopur ‘ds 39 ‘uas oepliooOUTy sepuosoulY | 67 S1UDIpIIAIANS SNUIYANG » seprurying I psouids puvove oeprueoer I ‘aou ‘ds inpis “SG, 99 4010914] SNdouvs8aIS x 76 snjpqgo] SadiqoTx oeprpodoseyeud (0.0 cn cn I Boe = N OWNnwo-« AaAamNManmn me OM ‘Aou ‘ds Jo ‘Udd SniuDINOS snIUNNy $1]Da10q "JD Sniuauny Saddsaqul DIADUAAP y sndojuvuy ‘Wy ‘aou ‘ds 1upudnys nupvjodosriy LUNDU SAJAUNIAT y SOJOUDJAU "Fy DIPUNUIM D1]O47 snyoujpodiumas snsoydosjdojv Dy, DIAD]NIDU SIN IP SNINIJOUDIAU * [x sadiavl{ SnudjO J» DIADIJOS * Lx ‘aou ‘ds 1miysawp Dsus aeprloedojoos SNIDUDdIWAS * Dy SIUD]JOI “dD SNABJIIOA SNIAPDADY Dx pjoavjonbs * dx DIIUNMOP SI]DIAN] dx “aou ‘ds ipunupa snsajdouojag ‘aou ‘ds 19 ‘uas s121d 401014 oeprupereyo SOULIOJTLIPeIEYS DUIJOADI DUDZAOg oepiley SOULIOF EY suaosvandind ‘Jd adojauad oeploela SOULIOJTED “aou “ds 1quoypoig O8vapiWx snouvjd sn10qgx]odx ‘ds oojD4 snidaadvds * J syvsowaf * Tx snuiisaiad OJ]DA% oepluogey 19 Podilymbus Lesson Podilymbus podiceps (Linnaeus) Pied-billed Grebe Material Proximal end of 1 right humerus, 1 nearly complete right femur, proximal end of 1 right femur, 1 complete left tarsometatarsus. ROM 13101-13104. The 4 specimens represent a minimum of 2 individuals. Remarks The present range of P. podiceps includes northwestern Peru, and I would expect to find this grebe on almost any body of fresh water in coastal Peru, although perhaps not in large numbers. Order Pelecaniformes Sharpe Suborder Sulae Sharpe Family Phalacrocoracidae (Bonaparte) Phalacrocorax Brisson Phalacrocorax olivaceus (Humboldt) Olivaceous Cormorant Material Two maxillae, distal halves of 2 right and 2 left mandibles, 5 right and 2 left scapulae, 5 complete right and 11 complete left coracoids, humeral ends of | right and 3 left coracoids, sternal ends of 2 right and 4 left coracoids, 2 complete left humeri, proximal ends of 2 right and 4 left humeri, distal ends of 7 right and 6 left humeri, proximal ends of 3 right and 2 left ulnae, distal ends of 5 right and 2 left ulnae, proximal ends of 5 right and 2 left radii, distal ends of 4 right and 1 left radius, 3 complete right and | complete left carpometacarpus, proximal ends of | right and 1 left carpometacarpus, 4 complete right and 4 complete left femora, proximal end of 1 left femur, distal ends of 1 right and 1 left femur, 2 complete right and 1 complete left tibiotarsus, proximal ends of 1 right and 2 left tibiotarsi, distal ends of 1 right and 3 left tibiotarsi, 8 complete right and 8 complete left tarsometatarsi, proximal ends of 3 left tarsometatarsi, distal ends of 1 right and 2 left tarsometatarsi. ROM 13105-13237. The 133 specimens represent a minimum of 15 individuals. Characters Phalacrocorax olivaceus differs from P. bougainvillii and P. gaimardi by having a larger premaxillary and mandible. Postcranial elements of P. olivaceus are generally smaller than those of P. gaimardi, although size ranges of the two species overlap. They are much smaller, and especially more slender, than those of P. bougainvillii. Compared with P. bougainvillii and P. gaimardi, P. olivaceus has scapula with (1) glenoid facet distinctly triangular (more rounded in P. bougainvillii and P. gaimardi); (2) acromion wide and moderately long (wide and long in P. bougainvillii, narrow and moderately 20 long in P. gaimardi); (3) protrusion on lateral surface of shaft distal to glenoid facet round and prominent (elongated and only slightly raised, if any, in P. bougainvillii and P. gaimardi). Coracoid with (1) shaft very slender (very robust in P. bougainvillii, slender in P. gaimardi); (2) bicipital attachment narrow and deeply concave (wide and moderately to slightly concave in P. bougainvillii and P. gaimardi); (3) procoracoid short and rounded (moderate to long, stouter, and more angular in P. bougainvillii and P. gaimardi); (4) attachment of Lig. humero-coracoideum anterius superius deep (superficial in P. bougainvillii and P. gaimardi); (5) humeral end narrow (wide in P. bougainvillii and P. gaimardi). Humerus with (1) pneumatic fossa moderately excavated (deeply excavated in P. bougainvillii and P. gaimardi); (2) deltoid crest small, smoothly rimmed (larger and locally angular in P. bougainvillii and P. gaimardi); (3) attachment of M. extensor metacarpi radialis inset in shallow depression (located on raised platform in P. bougainvillii; inset in depression in P. gaimardi, but bordered distally by large ectepicondylar prominence); (4) attachment of anterior articular ligament with moderate slope (lesser slope in P. bougainvillii, greater slope in P. gaimardi). Ulna with (1) internal cotyla narrow (wider in P. bougainvillii and P. gaimardi); (2) external cotyla wide and moderately long (long and narrow in P. bougainvillii, short and narrow in P. gaimardi); (3) external condyle of moderate width and length (narrow and long in P. bougainvillii, moderately wide and short in P. gaimardi); (4) carpal tuberosity moderately large (large in P. bougainvillii and P. gaimardi). Radius with (1) insertion of M. biceps moderately deep (very deep in P. bougainvillii and P. gaimardi); (2) projection extending from lateral end of carpal facet sharp and pointed (only slightly developed in P. bougainvillii and P. gaimardi). Carpometacarpus with (1) external rim of carpal trochlea small (much wider in P. bougainvillii, slightly wider in P. gaimardi); (2) process of metacarpal I narrow (broad in P. bougainvillii and P. gaimardi); (3) pollical facet wide (narrower in P. bougainvillii and P. gaimardi). Femur with (1) iliac facet small (larger in P. bougainvillii and P. gaimardi, even though the femur is shorter in P. gaimardi); (2) internal condyle slightly angular distally (more rounded, and rotated posteriad, in P. bougainvillii and P. gaimardi); (4) shaft slender (much heavier in P. bougainvillii and P. gaimardi). Tibiotarsus with (1) outer cnemial crest wide and short (wide and long in P. bougainvillii, narrow and short in P. gaimardi); (2) shaft slightly to moderately concave distal to internal articular surface (deeply concave in P. bougainvillii, slightly to moderately concave in P. gaimardi); (3) internal condyle short and moderately wide (long and wide in P. bougainvillii, long and narrow in P. gaimardi); (4) size very small. Tarsometatarsus with (1) internal cotyla short, narrow, and moderately concave (longer, wider, and more concave in P. bougainvillii and P. gaimardi); (2) inner extensor groove wide (narrow in P. bougainvillii and P. gaimardi); (3) external trochlea narrow (wide in P. bougainvillii and P. gaimardi); (4) distal foramen of moderate size (smaller in P. bougainvillii, much larger in P. gaimardi). Remarks Phalacrocorax olivaceus is generally distributed throughout South America, and is common in coastal lagoons, marshes, lakes, and rivers. The osteological comparisons 21 were made between four specimens of P. olivaceus, 16 specimens of P. bougainvillii, and three specimens of P. gaimardi. Phalacrocorax bougainvillii (Lesson) Guanay Cormorant Material Distal end of one right humerus. ROM 13238 Remarks Phalacrocorax bougainvillii is a marine species that breeds on offshore islands and comes ashore only rarely, and then usually only during times of distress (Murphy, 1936). The population of this species numbers in the millions, and it undergoes wide fluctuations. At times, when the population crashes, the beaches are literally covered with corpses. This happens whenever the warm current known as El Nino appears. The single specimen in the collection suggests that this species can be regarded as a rare straggler inland in the Talara region. Order Ardeiformes (Wagler) Suborder Ardeae Wagler Family Ardeidae Vigors Ardea Linnaeus Ardea cocoi Linnaeus White-necked Heron Material One right scapula, proximal ends of | right and | left carpometacarpus, distal ends of 2 left carpometacarpi, distal end of 1 right tarsometatarsus. ROM 13239-13244. The 6 specimens represent a minimum of 2 individuals. Remarks The largest living South American heron, A. cocoi is generally distributed over the entire continent. It is found in coastal marshes and lagoons in Peru, although only in small numbers. Casmerodius Gloger Casmerodius albus (Linnaeus) Common Egret Material Two complete right and 2 complete left coracoids, humeral ends of 2 right coracoids, 1 right scapula, 1 complete left carpometacarpus, proximal end of 1 left 22 carpometacarpus, distal end of 1 right tibiotarsus, distal end of 1 might tarsometatarsus. ROM 13245-13255. The 11 elements represent a minimum of 4 individuals. Remarks Generally distributed over all of South America, C. albus is common in coastal ~marshes and lagoons of Peru. Egretta T. Forster Egretta thula (Molina) Snowy Egret Material Two right scapulae, 2 complete right coracoids, proximal end of | left humerus, 1 complete right femur, proximal end of | right femur. ROM 13256-13262. The 7 specimens represent a minimum of 2 individuals. Additionally, the following specimens are identified to genus only: one complete right and 2 complete left carpometacarpi, distal ends of 3 left tibiotarsi, proximal end of 1 right tarsometatarsus, distal ends of 3 right and 5 left tarsometatarsi. ROM 13263-13277. Characters The three species of Egretta, E. thula, E. (Hydranassa) tricolor, and E. (Florida) caerulea, are extremely difficult to separate osteologically. Characters used to identify the fossil specimens of E. thula are given below, but intraspecific variation is so great that even these characters may prove to be invalid if sufficient series are examined. Compared with E. caerulea and E. tricolor, E. thula has scapula with (1) thickness between acromion and coracoidal articulation moderate, but variable (similar in E. caerulea, slightly thicker in E. tricolor); (2) length of acromion base moderate (less in E. caerulea, greater in E. tricolor). Coracoid with (1) curvature of coracohumeral surface slight to moderate as traced by central groove (similar in FE. caerulea, very slight in E. tricolor); (2) angle between internal edge of furcular facet and triosseal canal slightly more than 90° (much larger in EF. caerulea, and the area is well rounded; slightly larger in E. tricolor). Humerus with (1) distal end of attachment of M. proscapulohumeralis surrounded by attachment of M. triceps, external head (not surrounded in E. caerulea, roughly similar in E. tricolor); (2) external attachment of M. supracoracoideus short and deep (short and shallow in E. caerulea, long and shallow in E. tricolor). Carpometacarpus with (1) intermuscular line between pollical facet and pisiform process lying near pollical facet (similar in E. caerulea, positioned posteriorly in E. tricolor). No characters were found that reliably separated the carpometacarpi of EF. thula from those of E. caerulea. The three fossil specimens could thus be separated from E. tricolor, but assignment to either E. thula or E. caerulea was not possible. 25 Femur with (1) iliac facet narrow anteroposteriorly (narrow to moderate width in E. caerulea, broader in E. tricolor); (2) external condyle flat dorsally (similar in E. caerulea, extending farther dorsad in E. tricolor); (3) distal end broad (narrow in E. caerulea, broad in E. tricolor). Tibiotarsus with (1) rotular crest low (low in E. caerulea, higher in E. tricolor); (2) external condyle slightly undercut dorsoproximally (not undercut in EF. caerulea; slightly, or not, undercut in E. tricolor). As no reliable characters were found to separate the tibiotarsi of FE. thula and E. tricolor, the specimens are referred to genus only. No reliable characters for separating the tarsometatarsi of the three species of Egretta were found. For this reason the specimens are referred to genus only. Remarks Generally distributed over almost all of South America, EF. thula is common in coastal lagoons and marshes of Peru. I have chosen to lump Egretta, Hydranassa, and Florida because of the very close similarity in their osteology. If the three cannot be reliably distinguished at the species level they should not be considered distinct genera. The comparisons were made using more than 10 specimens of each species. Egretta caerulea (Linnaeus) Little Blue Heron Material One complete left coracoid. ROM 13278. Remarks Generally distributed over almost all of South America, E. caerulea is found in coastal lagoons and marshes of Peru, but not in such large numbers as E. thula. Syrigma Ridgway Syrigma sanctimartini sp. nov. Fig. 6A Holotype Complete left coracoid. ROM 12892. Diagnosis Coracoid agrees with that of Syrigma and differs from that of all other genera of South American herons by having (1) head broad and slightly peaked in medial view (subangular to subrounded); (2) brachial tuberosity prominent and angular; (3) triosseal canal very deep adjacent to glenoid facet; (4) attachments of M. coracobrachialis anterior and Lig. humero-coracoideum anterius superius deep, elongated and notched, with center of latter lying parallel to, and roughly in line with, 24 anterior edge of shaft; (5) protrusion internal to glenoid facet large, bordering ventral side of triosseal canal; (6) external side of shaft with slight ridge running ventrad from middle of head. Coracoid differs from that of S. sibilatrix (Temminck) by having (1) coracohum- eral surface short, displacing dorsal-most point of head posteriad; (2) furcular facet larger; (3) brachial tuberosity larger; (4) humeral end with medial side flatter (locally more convex and concave in S. sibilatrix); (5) protrusion internal to glenoid facet smaller; (6) head thicker between attachment of Lig. humero-coracoideum anterius superius and anterior furcular facet; (7) glenoid facet rotated posteriad, lengthening distance from external edge of glenoid facet to anterior edge of shaft; (8) external sternal facet much higher. Measurements of the holotype, with those of one male specimen of S. sibilatrix in parentheses, are as follows: length, 42.3 (41.6); head to scapular facet, 13.3 (13.5); proximal depth, 9.5 (9.2); least depth of shaft, 4.2 (4.3). Etymology This species is named for José de San Martin, who proclaimed Peruvian independence at Lima on July 28, 1821. Remarks The genus Syrigma is represented by one living species in South America, S. sibilatrix, which occurs only east of the Andes. The presence of a second, extinct species west of the Andes suggests that perhaps the climate west of the Peruvian Andes was more favourable during the late Pleistocene than it is today. Syrigma sanctimartini is the first palaeospecies of the genus to be described. Legitimate concern may be expressed about the validity of this new species, and some to be described later, because of the small number of specimens of living species used for comparisons. Intraspecific variability can be a very serious problem in identifying a species from a single bone, but it is much more of a problem in certain groups than in others. I believe that in general the osteological variation within a group is correlated with how specialized it is; i.e., the narrower the niche, the less the osteological variation, and the broader the niche, the greater the osteological variation. The importance attached to any given character, i.e., its reliability, is dependent on the species involved. After examining very carefully thousands of bones of over two hundred species of birds, I believe I can determine what degree of variation can be expected in any given species by looking at a few specimens, and by knowing how specialized it is. Exceptions, of course, can always be found. Certain types of osteological characters, such as size, shape, and position of muscle scars and ligamental attachments, may be very reliable characters for some species and of no value whatsoever for other species. Other types of osteological characters, however, such as the spatial orientation of the proximal end of a bone compared with the distal end, or the size, shape, and position of articular facets, are much less likely to vary greatly between individuals. Size as a distinguishing osteological character at the species level is no more or no less reliable than many other characters; it all depends on the group in question. I have described species herein as new only when I believed the characters being used were not subject to intraspecific variation. While I used such characters to 2 support my conclusions, I have also included in the diagnoses how perhaps less reliable characters differed between the living species and the species being described. I can only hope that my conclusions will be verified when sufficient series of specimens of the living species can be assembled and the actual degree of variation is determined. Caution has been the by-word in naming new species, as evidenced, I believe, by the number of species herein which have not been named. Nycticorax T. Forster Nycticorax nycticorax (Linnaeus) Black-crowned Night Heron Material Middle third of 1 left mandible, 1 right and 3 left quadrates, 4 right and 3 left scapulae, | complete right coracoid, humeral ends of 9 right and 4 left coracoids, proximal end of | left humerus, proximal end of | right ulna, distal ends of 2 right and 3 left ulnae, distal end of | left radius, proximal end of | left carpometacarpus, 1 complete right femur, proximal ends of | right and | left femur, distal ends of 3 right and 5 left tibiotarsi, 1 complete right and 1 complete left tarsometatarsus, proximal end of | right tarsometatarsus, distal ends of 2 right and 3 left tarsometatarsi. ROM 13279-13331, 19532. The 54 specimens represent a minimum of 10 individuals. Characters All elements of Nycticorax nycticorax examined are separable from elements of all species of Egretta by size, those of Nycticorax being either much longer, shorter, wider, or stouter than corresponding elements of Egretta. Compared with Nyctanassa Stejneger and Tigrisoma Swainson, Nycticorax has scapula with (1) coracoidal articulation large (large and subrounded in Tigrisoma, smaller in Nyctanassa); (2) glenoid facet round (broad oval in Tigrisoma, narrow oval in Nyctanassa); (3) acromion moderately deflected externally (less in Tigrisoma, more in Nyctanassa). Coracoid with (1) triosseal canal moderately inset (slightly inset, if any, in Tigrisoma and Nyctanassa); (2) glenoid facet large (much smaller in Jigrisoma, of similar size in Nyctanassa); (3) dorsal surface of head notched by deep concavity in medial view (rounded in 7igrisoma, notched and more angular in Nyctanassa); (4) anterior end of head moderately rotated externally (rotated less in Tigrisoma, much more in Nyctanassa). Humerus with (1) attachment of M. proscapulohumeralis elongated and moderately inset (oval and inset only slightly in Tigrisoma and Nyctanassa); (2) pneumatic fossa very shallow (moderately deep in Tigrisoma and Nyctanassa); (3) dorsal edge of attachment of external head of triceps curving away from capital shaft ridge distal to attachment of M. latissimus dorsi posterioris (extends farther proximad on capital shaft ridge in Tigrisoma and Nyctanassa). Ulna with (1) olecranon narrow (broad in Tigrisoma, similar in Nyctanassa); (2) external cotyla large, with small ridge leading distad (smaller, without ridge in Tigrisoma and Nyctanassa), (3) external condyle long (short in Tigrisoma, moderate in Nyctanassa); (4) carpal tuberosity moderately long (long in Tigrisoma, short in Nyctanassa). 26 Radius with (1) palmar surface of distal end moderately excavated (slightly excavated in Jigrisoma, greatly excavated in Nyctanassa); (2) tendinal groove broad and deep (narrower and not as deep in Jigrisoma and Nyctanassa); (3) carpal facet broad (narrow in Jigrisoma and Nyctanassa). Carpometacarpus with (1) internal ligamental attachment of pisiform process and ligamental attachment of ulnare confluent (separated in Tigrisoma and Nyctanassa); (2) corner on distal end of internal rim of carpal trochlea small (large in Tigrisoma, absent in Nyctanassa); (3) carpal trochlea moderately wide (wider in Tigrisoma, similar in Nyctanassa), (4) process of metacarpal I slender (stouter in Tigrisoma and Nyctanassa). Femur with (1) proximal lateral face curving mediad (curvature slightly less in Tigrisoma, much less in Nyctanassa); (2) concavity distal to posterior edge of iliac facet deep (moderate concavity in Tigrisoma, slight concavity in Nyctanassa); (3) posterior intermuscular line curving internally to base of internal condyle (lies more in center of shaft and does not curve in Jigrisoma, does not curve as far internally in Nyctanassa), (4) gap between external condyle and origin of M. gastrocnemius, pars externa, wide (ridge present between the two in Jigrisoma, slight gap in Nyctanassa). Tibiotarsus with (1) internal articular surface large (similar in Tigrisoma, smaller in Nyctanassa); (2) inner cnemial crest short and thick (long and of moderate width, tapering distad, in Tigrisoma and Nyctanassa); (3) ridge leading to posteromost comer of internal articular surface absent (present in Jigrisoma, absent in Nyctanassa); (4) internal condyle undercut dorsally (not undercut in Tigrisoma and Nyctanassa), (5) internal condyle flaring extensively internally (flaring much less in Tigrisoma and Nyctanassa). Tarsometatarsus with (1) shaft large (more slender in Tigrisoma, much more slender in Nyctanassa); (2) cotylae and trochleae wide (narrow in Tigrisoma and Nyctanassa); (3) internal edge of shaft angular (rounded in Tigrisoma and Nyctanassa). Remarks Nycticorax nycticorax is generally distributed throughout South America, including coastal rivers, swamps, and lagoons of Peru. On the basis of osteological comparisons listed above I believe Nycticorax and Nyctanassa are unquestionably distinct genera. Comparative material included ten specimens of Nycticorax nycticorax, eight of Nyctanassa violacea, one of Tigrisoma lineatum, and two of T. fasciatum. Suborder Plataleae Newton Family Plataleidae Bonaparte Subfamily Threskiornithinae (Richmond) Theristicus Wagler Theristicus wetmorei sp. nov. Fig. 6B, C Holotype Complete left coracoid. ROM 12890. Referred Material Fragmentary proximal end of 1 right humerus. ROM 12891. The 2 specimens represent a minimum of | individual. Diagnosis Coracoid agrees with that of Theristicus and differs from that of all other genera of South American ibises (with possible exception of Cercibis Wagler which was unavailable for comparison), by having (1) head long and subangular in medial view; (2) attachment of M. coracobrachialis anterior moderately deep, long, with narrow posterior section and a much wider anterior section; (3) attachment of Lig. humero-coracoideum anterius superius shallow and confluent with anterior section of attachment of M. coracobrachialis anterior; (4) coracohumeral surface broad, of almost uniform width; (5) head very wide, with external and internal sides roughly parallel; (6) brachial tuberosity large, well rounded; (7) glenoid facet large, with widest area immediately dorsal to scapular facet; (8) shaft with anterior edge immediately dorsal to sternal facet flattened and rotated slightly internally; (9) anterior external sternal facet with large posterior protrusion. Coracoid differs from that of Theristicus caudatus (Boddaert) by having (1) coracohumeral surface much shorter, giving a more rounded head with lower profile in medial view; (2) attachments of M. coracobrachialis anterior and Lig. humero-coracoideum anterius superius shallower; (3) procoracoid deeply concave internally; (4) shaft with anterior edge immediately dorsal to sternal facet more flattened and less rotated internally; (5) external sternal facet and external distal angle much shorter; (6) concavity between external sternal facet and flattened anterior edge of shaft shallow. Measurements of the holotype, with those of three male specimens of T. caudatus in parentheses, are as follows: length, 50.5 (52.0-54.3; mean 52.8); head to scapular facet 19.9 (20.8-21.1; mean, 20.9); proximal depth, 12.9 (13.7—14.1; mean, 13.8); least depth of shaft, 8.5 (9.0—9.1; mean, 9.1). Characters Humerus agrees with that of Theristicus and differs from that of all other genera of South American ibises examined by having (1) pneumatic fossa shallow; (2) internal tuberosity short; (3) median crest very prominent. Humerus differs from that of 7. caudatus by having (1) internal tuberosity shorter; (2) attachment of M. coracobrachialis posterior rotated toward axis of shaft; (3) median crest more prominent; (4) attachment of Lig. humero-coracoideum anterius inferius positioned more anconally. The specimen is too broken to yield accurate measurements. Etymology This species is named for the late Dr. Alexander Wetmore of the United States National Museum in recognition of his lifetime dedication and service to ornithology. Remarks Theristicus is represented by one living species in South America, T. caudatus, which occurs over almost the entire continent, including cultivated areas of coastal Peru 28 (Koepcke, 1970). No significant differences were seen between specimens of T. c. caudatus and T. c. branickii examined. Although no specimens of T. c. melanopis were available, I doubt that it would vary significantly from T. c. caudatus or T. c. branickii. Theristicus wetmorei is the first described palaeospecies of the genus. Eudocimus Wagler Eudocimus peruvianus sp. nov. Fig. 7 A-F Holotype Complete left tarsometatarsus. ROM 12917. Paratype Distal end of 1 left tarsometatarsus. ROM 12918. Referred Material One right scapula, distal end of 1 left humerus, proximal ends of 2 radii, 1 complete left femur, proximal end of | right femur, 1 complete right tibiotarsus (immature), distal end of 1 right tibiotarsus. ROM 12919-12926. The 10 specimens, including holotype and paratype, represent a minimum of 2 individuals. Diagnosis Tarsometatarsus agrees with that of Eudocimus and differs from that of all other genera of South American ibises (with possible exception of Cercibis which was unavailable for comparison) by having (1) shaft long, moderately stout; (2) cotylae, trochleae, and hypotarsus moderately compressed. Tarsometatarsus differs from that of Eudocimus albus Wagler by having (1) shaft with channelling of anterior face very shallow, deepest on external side, with proximal ligamental attachment raised to level of sides of shaft (shaft with anterior face deeply channelled, deepest on internal side, with proximal ligamental attachment lying below level of sides of shaft in E. albus); (2) intercotylar prominence larger, more rounded, and rotated dorsad; (3) proximal end wider and distal end narrower (when compared with specimen of E. albus of same, or greater, length); (4) cotylae slightly longer and wider; (5) groove between intercotylar prominence and external cotylae distinct; (6) hypotarsal ridges extend farther distad and form a ‘‘V’’; (7) internal side of hypotarsus, and shaft anterior to internal side of hypotarsus, more excavated; (8) attachment of external ligament prominent, positioned posteriorly; (9) convexity at anterior end of internal trochlea slightly developed; (10) depth of middle trochlea less; (11) external trochlea with external face large, extending less posteriad. For measurements, see Table 2. Measurements of the holotype are as follows: length, 100.8; proximal width, 13.7; distal width, 12.7; least width of shaft, 5.3; width of middle trochlea, 4.9. Table 2 Measurements of the scapula, humerus, radius, femur, tibiotarsus, and tarsometatarsus of Eudocimus peruvianus sp. nov. and Recent E. albus; and of the scapula, tibiotarsus, and tarsometatarsus of E. albus from the Talara Tar Seeps (in mm). Eudocimus peruvianus Eudocimus albus sp. nov. Talara Tar Seeps Recent Scapula' Proximal OR 11.9 11.0 9.3-11.8 Height M = ~- 10.9 n l 1 5 Proximal OR SS) Sha 3 41=—401 Width M — — si) n l 1 5 Humerus Distal OR 172 14.6-17.8 Width M — 16.3 n 1 5 Depth of OR 8.9 7.9-9.9 External M — 8.9 Condyle n 1 5 Radius Proximal OR 5.6-5.9 4.3-5.4 Width M 5.8 4.9 n 2 5) Femur Length OR 59.0 56. 1=67.1 M as 62.9 n l 5 Proximal OR 11.9-12.2 10.8-—13.5 Width M 12.1 13-2 n 2 5 Distal OR 12.6 11.1-13.8 Width M os 12.4 n 1 =) Least OR SES) 5.0-6.5 Width of M = 5.8 Shaft n 1 5 Tibiotarsus Length OR 141.0 102.3-141.8 M = 12733 n 1 Re) 30 Distal Width Least Width of Shaft Depth of Internal Condyle Tarsometatarsus Length Proximal Width Distal Width Least Width of Shaft Width of Middle Trochlea Eudocimus peruvianus sp. nov. 10.6—11.0 10.8 11.3-11.7 PHS Eudocimus aibus Talara Tar Seeps 9.5-10.4 wey) 2 10.0-11.8 1OF9 2 4.2 Recent 8.5-10.6 9.8 73.0-103.8 9277 5 10.2-12.8 11.8 5 10.4-13.3 2g " Scapular measurements that can be used consistently with all avian species are difficult to find because the highly irregular shape of the proximal end varies considerably between taxa. Measurements used here are as follows: Proximal height—distance from tip of acromiom process to most ventral point of glenoid facet. Proximal Width—perpendicular distance from a line drawn between most internal points of acromiom and glenoid facet, to the most external point of acromiom or glenoid facet, whichever is greater. Si Characters Eudocimus peruvianus differs from E. albus by having scapula with (1) distal end of glenoid facet slightly elevated above surface of shaft, and rotated ventrad only slightly (elevated and rotated ventrad more in E. albus). Humerus with (1) attachment of anterior articular ligament broader, extending less distad; (2) attachment of M. pronator brevis located farther distally; (3) internal condyle projecting less distad, more palmarly; (4) intercondylar groove shallower. Radius with (1) capital tuberosity larger; (2) ligamental papillae smaller; (3) bicipital attachment a very deep pit; (4) ulnar facet larger. Femur with (1) anterior intermuscular line straight, lying near external side of shaft (curved and positioned medially in FE. albus); (2) iliac facet shorter; (3) concavity distal to posterior edge of iliac facet shallower; (4) trochanter projecting less anteriad; (5) internal flexure of shaft proximal to internal condyle greater; (6) ridge leading to posterior corner of internal condyle much smaller, with more curving edge; (7) external condyle smaller, less rounded. Etymology This species is named for Peru, the country in which it was first discovered. Remarks Eudocimus ruber differs more from E. albus and E. peruvianus than the latter two do from each other, thus a detailed comparison with E. ruber is not considered necessary. Eleven specimens of FE. albus from Florida were available for comparison. Eudocimus peruvianus is the first palaeospecies of the genus described from the New World. Its presence in the same deposit as E. albus, whose current range does not quite reach so far south as the fossil site, poses the interesting question of whether the two species were sympatric or had seasonally overlapping ranges. Neither of the two species occurs in very large numbers in the collection. Eudocimus albus (Linnaeus) White Ibis Material One left scapula, 1 complete left coracoid, humeral ends of 2 left coracoids, distal end of 1 right ulna, 1 complete right carpometacarpus, distal ends of 2 left tibiotarsi, proximal end of | right tarsometatarsus, distal end of 1 right tarsometatarsus. ROM 13332-13340; 19520. The 10 specimens represent a minimum of 3 individuals. Remarks Eudocimus albus occurs today in the northernmost part of coastal Peru, western Ecuador, Colombia and Venezuela. The specimens listed above did not differ noticeably from the comparative material from Florida. 32 Subfamily Plataleinae Bonaparte Ajaia Reichenbach Ajaia ajaja (Linnaeus) Roseate Spoonbill Material Proximal end of | right carpometacarpus. ROM 13341. Characters Carpometacarpus with (1) metacarpal I long and slender; (2) posterior carpal fossa elongated and very deep; (3) pollical facet moderately wide; (4) size large. Remarks Ajaia ajaja is widespread in South America, and has been recorded from northwestern Peru. Suborder Ciconiae Bonaparte Family Ciconiidae (Gray) Subfamily Ciconiinae Gray Jabiru Hellmayr Jabiru mycteria (Lichtenstein) Jabiru Material One cranium, | sternal fragment, proximal ends of 2 left scapulae, symphyseal area of 1 furculum, portions of 3 right and 4 left coracoids, distal ends of 1 right and 1 left humerus, proximal ends of 2 left ulnae, distal ends of 2 right and 2 left ulnae, distal ends of 2 right and 1 left radius, proximal end of | left carpometacarpus, distal ends of 2 right carpometacarpi, 5 right and 1 left carpal digit, proximal end of 1 left tibiotarsus, distal ends of 3 right and 4 left tibiotarsi, proximal ends of | right and 2 left tarsometatarsi, distal ends of 3 right and 1 left tarsometatarsus. ROM 13342-13388. The 47 elements represent a minimum of 5 individuals. Remarks Only isolated individuals of Jabiru mycteria have been recorded on the coast of southern Peru, where the species is regarded as a trans-Andean vagrant. The specimens represented here may also represent trans-Andean vagrants or they may represent individuals that moved southward from the moister coastal regions of southern Ecuador. Although there are no recent records of Jabiru in Ecuador, I see no reason why it may not have been there in the past, or why it may not be there now. 33 Subfamily Mycteriinae American Ornithologists’ Union Mycteria Linnaeus Mycteria americana Linnaeus Wood Stork Material Proximal end of | left carpometacarpus. ROM 13389. Remarks Mycteria americana has been recorded from northwestern Peru, but not from the remainder of coastal Peru. The scarcity of fossil specimens of M. americana may indicate a lack of the type of wooded swamps they prefer for nesting and roosting. The single specimen probably represents a rare vagrant. Order Anseriformes Wagler Suborder Anseres Wagler Family Anatidae Vigors Subfamily Dendrocygninae Reichenback Dendrocygna Swainson Dendrocygna autumnalis (Linnaeus) Black-bellied Tree Duck Material Portions of 2 crania, | frontal, 4 right and 7 left quadrates, 3 left articulars, 1 right and 5 left mandibular fragments, 15 right and 14 left scapulae, 13 complete right and 21 complete left coracoids, humeral ends of 5 right and 4 left coracoids, sternal ends of 2 left coracoids, 2 complete right and 2 complete left humeri, proximal ends of 14 right and 12 left humeri, distal ends of 17 right and 12 left humeri, 2 complete right and 1 complete left ulna, proximal ends of 6 right and 6 left ulnae, distal ends of 8 right and 7 left ulnae, 1 complete left radius, proximal ends of 5 right and 4 left radii, distal ends of 9 right and 8 left radii, 12 complete right and 10 complete left carpometacarpi, proximal ends of 4 right and 10 left carpometacarpi, 7 complete night and 4 complete left femora, proximal ends of 4 right and 5 left femora, distal ends of 1 right and 1 left femora, distal ends of 14 right and 4 left tibiotarsi, 1 complete right tarsometatarsus, proximal ends of | right and 1 left tarsometatarsus, distal ends of 4 right and 5 left tarsometatarsi. ROM 13390-13688, 19521-19522. The 301 specimens represent a minimum of 33 individuals. Characters The larger size of its elements distinguishes D. autumnalis from D. viduata (Linnaeus). Although D. bicolor (Vieillot) is usually smaller than D. autumnalis, the 34 two species overlap in size. Characters of D. autumnalis used to separate it from D. bicolor are listed below. Generic characters are also given for the scapula, radius, and ulna, elements not studied by Woolfenden (1961). Three specimens of each species were available for the comparisons. Scapula of Dendrocygna with (1) glenoid facet long and narrow, lacking an anteroventral projection. Scapula of D. autumnalis with (1) acromion longer than in D. bicolor; (2) glenoid facet longer and more elevated. Coracoid with (1) angle between furcular facet and edge of triosseal canal much smaller; (2) brachial tuberosity less rounded; (3) head more rounded in medial view; (4) attachment of Lig. humero-coracoideum anterius superius larger. Humerus with (1) proximal and distal ends much wider; (2) deltoid crest longer and more flaring; (3) internal tuberosity much larger; (4) ectepicondylar prominence more rounded; (5) attachment of anterior articular ligament higher; (6) entepicondyle much larger. Ulna of Dendrocygna with (1) external cotyla large; (2) bicipital attachment large and prominent; (3) carpal tuberosity very large. Ulna of D. autumnalis with (1) olecranon much thicker than in D. bicolor; (2) internal cotyla larger; (3) distal end of external cotyla more rounded; (4) external condyle longer; (5) carpal tuberosity larger. Radius of Dendrocygna with (1) shaft immediately posterior to head very narrow; (2) ulna facet narrow and U-shaped; (3) capital tubersoity large; (4) carpal facet narrow, with distinct anconal projection centrally located. Radius of D. autumnalis with (1) capital tuberosity smaller than in D. bicolor, but with more prominent shaft flaring more abruptly to meet a wider distal end. Carpometacarpus with (1) process of metacarpal I larger; (2) lobe at end of external rim of carpal trochlea more prominent; (3) area of fusion between metacarpal II and metacarpal III longer; (4) internal rim of carpal trochlea larger. Femur with (1) head rotated dorsad, and concavity between iliac facet and head greater; (2) iliac facet larger and more irregularly shaped; (3) trochanter larger and more prominent; (4) distal end wider; (5) ligamental attachment at base of fibular condyle more prominent; (6) condyles larger; (7) rotular groove deeper. Tibiotarsus with (1) shaft reasonably straight and stout (in D. viduata and D. bicolor there is a greater internal concavity and distal third of shaft has greater convexity directed posteriad); (2) inner cnemial crest larger; (3) condyles not reflected anteriad. Tarsometatarsus with (1) intercotylar prominence less prominent; (2) cotylae large; (3) external edge of hypotarsus longer, positioned proximally; (4) external trochlea slightly narrower, but middle and internal trochleae wider; (5) internal trochlea projecting less distad. Remarks Dendrocygna autumnalis has been recorded in western Ecuador south to the southernmost coastal province, El Oro, and occasionally on the coast of Peru (Koepcke, 1970). However, it was not recorded by Marchant (1958) on the Santa Elena Peninsula of Ecuador. The presence of this species may indicate an abundance of vegetation surrounding the fossil site at the time of deposition because of the species preference for quiet waters surrounded by forests (Delacour, 1954). a5 Subfamily Tadorninae Reichenbach ‘‘Humerus with (1) fairly prominent capital shaft ridge directed toward external tuberosity; (2) area of origin of external head of triceps relatively narrow; (3) deltoid crest relatively large and flaring, and in lateral view more rounded, or if an abrupt bend occurs, it lies more posteriad; (4) deltoid crest extending farther distad; (5) head rotated so that external tuberosity is higher.’’ (Woolfenden, 1961: 108). Nannonetta gen. nov. Diagnosis Humerus agrees with that of all living species of the subfamily Tadorninae and differs from that of all other species of the family Anatidae by having the diagnostic subfamilial characters listed by Woolfenden (1961). Humerus resembles that of Lophonetta Riley and differs from Chloephaga Eyton and Neochen Salvadori by having (1) head more rounded in anconal view; (2) external tuberosity reduced and elongated; (3) external head of triceps only slightly, if any, undercutting head; (4) deltoid crest shorter and less flaring; (5) capital groove rotated internally, reducing size of internal tuberosity and increasing size of head; (6) attachment of anterior articular ligament less elevated; (7) shaft flaring less to meet ectepicondyle; (8) internal condyle elongated, less bulbous. Humerus differs from that of Lophonetta by having (1) attachment of M. supracoracoideus on external tuberosity proportionately more elongated; (2) capital groove narrower and straighter; (3) deltoid crest shorter and less flaring, with sharp angle at its midpoint; (4) bicipital crest shorter and more rounded; (5) pneumatic fossa completely rimmed by heavy bone, such that distal edge is elevated above floor of bicipital crest; (6) pneumatic fossa with highest point located under tip of internal tuberosity (under median crest in Lophonetta); (7) capital shaft ridge reduced; (8) shaft with greater palmar flexure in distal quarter, such that a line drawn along anterior edge of shaft would almost bisect distalmost point of ectepicondyle (bisects anterior edge of ectepicondyle in Lophonetta); (9) ectepicondylar prominence smaller, with narrow ridge along proximal portion; (10) ectepicondyle less rounded and more V-shaped, extending farther distad; (11) attachment of anterior articular ligament wider than it is long (longer than wide in Lophonetta), less elevated, and located farther distad; (12) entepicondyle more prominent and rounded, not as large, with larger entepicondylar prominence and external curvature to anconal side; (13) condyles more rounded, giving deeper and narrower intercondylar groove; (14) external condyle shorter, retaining internal flexure found in Lophonetta, but lacking distinct ridge leading internally from tip of condyle; (15) shaft widening more gradually toward distal end. Type species Nannonetta invisitata sp. nov. Etymology From Greek, nannus, dwarf, and netta, feminine, duck. 36 Nannonetta invisitata sp. nov. Fig. oO AP Es SAGE: IQA, DiG, PATE Holotype Complete right humerus. ROM 13689. Paratypes One complete right and 3 complete left humeri, proximal ends of 6 right and 5 left humeri, distal ends of 2 right and 7 left humeri. ROM 13690-13713. Referred Material Three right and 4 left scapulae, 8 complete right and 7 complete left coracoids, humeral ends of 2 right and 1 left coracoid, 4 complete right and 3 complete left ulnae, proximal ends of 5 right and 3 left ulnae, distal ends of 5 right and 3 left ulnae, 2 complete left radii, proximal ends of 3 left radii, distal end of 1 right radius, 3 complete right and 2 complete left carpometacarpi, proximal ends of 2 right and 6 left carpometacarpi, 2 complete right and 2 complete left femora, proximal end of | right femur, 1 complete left tibiotarsus, distal ends of 4 right tibiotarsi, 4 complete right and 5 complete left tarsometatarsi, proximal end of 1 left tarsometatarsus. ROM 13714-13800. The 112 specimens, including holotype and paratypes, represent a minimum of 10 individuals. Diagnosis As for genus. For measurements see Table 3. Measurements of the holotype are as follows: length, 68.6; proximal width, 14.7; distal width, 11.0; least width of shaft, 2 5 Characters All elements of Nannonetta invisitata agree with the humerus in resembling Lophonetta more than Chloephaga or Neochen. N. invisitata differs from Lophonetta specularioides by having scapula with (1) coracoidal articulation proportionately more prominent, confluent with glenoid facet; (2) intermuscular line between attachments of M. rhomboidius superficialis and M. subscapularis better developed; (3) glenoid facet rotated such that it faces more anteriad and distal end is displaced dorsad. Coracoid with (1) head projecting less anteriad; (2) width of head much greater; (3) glenoid facet flaring less laterad; (4) scapular facet larger. Ulna with (1) attachment of anterior articular ligament more oval and less elongated; (2) external cotyla shorter but wider; (3) external condyle longer but less prominent; (4) carpal tuberosity more prominent. Radius with (1) capital tuberosity more prominent, and shaft more constricted immediately distal to it; (2) carpal facet less curved; (3) shaft with internal flexure at distal end. Carpometacarpus with (1) attachment of M. flexor carpi ulnaris brevis deeper; (2) external scapho-lunar ligamental attachment small, lying proximal to ridge extending or Table 3 Measurements of the scapula, coracoid, humerus, radius, ulna, carpometacarpus, femur, tibiotarsus, and tarsometatarsus of Nannonetta invisitata gen. et sp. nov. and Lophonetta specularioides (in mm). Scapula Proximal Height Proximal Width Coracoid Length Head to Scapular Facet Proximal Depth Least Depth of Shaft Length of Sternal Facet Humerus Length Proximal Width Distal Width Least Width of Shaft 38 Nannonetta invisitata gen. et sp. nov. 34.3-36.4 35.4 14 11.8-13.0 1253 13.4-14.9 14.2 4 61.9-68.6 65.0 4 13.9-14.7 14.5 Lophonetta specularioides 11.9-12.9 i235 45.4-51.0 49.1 B 16.7-19.2 18.2 17.7-20.3 19.4 3 95.3-101.8 PEND 3 19.0-22.4 2) Ua 3 14.5-15.7 5:3. Radius Length Proximal Width Distal Width Ulna Length Proximal Width Distal Depth Least Width of Shaft Carpometacarpus Length Height Through Metacarpal I’ Proximal Width Continued on page 40. Nannonetta invisitata gen. et sp. nov. d15=92:9 52:2 54.7-58.0 56.3 38.4-41.2 40.3 Lophonetta specularioides 82.6-86.8 85.1 89.3-94.7 a 3 10.0-11.8 10.9 57.1-60.7 59e1 39 Least Width of Shaft Length of Distal Fornix Femur Length Proximal Width Distal Width Least Width of Shaft Tibiotarsus Length Proximal Width Distal Width Least Width of Shaft Depth of Internal Condyle 40 Nannonetta invisitata gen. et sp. nov. 3.3-4.6 3.6 36.9-37.3 37-1 Lophonetta specularioides 3.9-4.8 4.5 52.6—59.1 56.1 11.7-12.9 W455) 85.7-95.4 9231 3 9.6-10.4 10.0 3 9.4-10.4 10.0 10.5-11.0 10.8 3 Nannonetta invisitata gen. et sp. nov. Lophonetta specularioides Tarsometatarsus Length OR 33,6-38.5 31 0=56.9 M 3523 54.3 n i! 3 Proximal OR 6.8=7.4 t0-3—=1059 Width M Ta OEY. n 9 3 Distal OR PTs 1O°4=120 Width M Tes 10.8 n yi 3 Least OR Soe 4.0-4.5 Width of M 3.6 4.3 Shaft n 10 3 Width OR 269-3 4.6—5.0 Middle M 3a 4.8 Trochlea n 9 3 " Height through Metacarpal I refers to the anteroposterior distance perpendicular to the long axis of the carpometacarpus as measured through metacarpal I. from external cuneiform ligamental attachment to proximal fornix, and lying very close to external cuneiform ligamental attachment [the large size and posterior position of this attachment in other genera of the subfamily Tadorninae is listed as a subfamilial character by Woolfenden (1961)]; (3) lobe at end of external rim of carpal trochlea, beginning at notch in rim, very long and prominent; (4) internal rim of carpal trochlea more prominent, giving more rounded proximal end in internal view; (5) internal ligamental fossa deeper; (6) area of fusion of metacarpal I and metacarpal If much shorter both proximally and distally. Femur with (1) neck more constricted; (2) area between anterior corner of trochanter and head more concave but similarly curved, not angular; (3) ridge on posterior surface of shaft leading to external condyle more prominent; (4) fibular condyle similar in shape, but less produced proximad; (5) external condyle well rounded where joined by fibular condyle (pointed in L. specularioides); (6) pit for M. tibialis anticus positioned more distad; (7) ridge on shaft leading distad to internal condyle, as traced by intermuscular line, less developed; (8) lip of internal condyle larger. Tibiotarsus with (1) rotular crest forming straight line (curved in L. specularioides); (2) outer cnemial crest thickened on distal side; (3) intermuscular line curving internally proximally to approach attachment of M. flexor digitorum longus (curves less internally in L. specularioides), (4) external ligamental attachment elliptical (similar in L. specularioides), but positioned more proximad. 4] Tarsometatarsus with (1) anterior lip on internal cotyla more undercut; (2) intercotylar prominence with greater anterior, but smaller dorsal component; (3) posterior border of internal cotyla with more distinct lip; (4) trochleae very spread (similar in L. specularioides, but internal and external trochleae turned slightly more mediad); (5) external rim of external trochlea less elongated; (6) internal side of middle trochlea lacking prominent internally directed ridge. Etymology Latin, invisitata, not seen before, unknown. Remarks Two extinct anseriform genera from North America have been assigned to the subfamily Tadorninae. These are Anabernicula Ross (1935) (A. minuscula Wetmore, A. gracilenta Ross, A. oregonensis Howard and A. robusta Short) and Brantadorna Howard (1963) (B. downsi Howard). From the descriptions given by Ross (1935), Wetmore (1924), Howard (1963, 1964), and Short (1970) it is readily apparent that Nannonetta is distinct from Anabernicula and Brantadorna, and that Anabernicula and Brantadorna appear to resemble Tadorna much more than Nannonetta or Lophonetta. Nannonetta differs from Anabernicula and Brantadorna by having humerus with (1) head more rounded in anconal view; (2) capital shaft mndge much reduced; (3) external tuberosity reduced and elongated. Coracoid with (1) angle between furcular facet and line drawn along the procoracoid smaller. Carpometacarpus with (1) length of metacarpal I much greater. The capital shaft ridge and external tuberosity of the humerus are much less prominent in Lophonetta than in Tadorna. Similarly, the capital shaft ridge and external tuberosity are much less prominent in Nannonetta than in Lophonetta. Nanonnetta, then, appears to represent one extreme of the variation found within the subfamily Tadorninae, with Tadorna representing the opposite extreme and Lophonetta occupying the central position. The intermediate position of Lophonetta can be seen in many of its elements, not just the humerus. Despite the large differences between Nannonetta and Tadorna, and the variance with a few subfamilial characters listed by Woolfenden (1961), the majority of the subfamilial characters displayed by Nannonetta agree with those of the subfamily Tadorninae. Therefore, I believe Nannonetta belongs within that subfamily. N. invisitata then becomes the smallest representative of the subfamily known, being approximately two-thirds the size of Tadorna and Lophonetta. Three specimens of Lophonetta specularioides were used for the osteological diagnoses. Chloephaga Eyton Chloephaga melanoptera (Eyton) Andean Goose Material One right quadrate, | right and 2 left scapulae, 4 complete right and 2 complete left 42 coracoids, humeral ends of | right and 3 left coracoids, 1 complete left humerus, proximal ends of | right and 4 left humeri, distal ends of | right and 5 left humeri, shaft of 1 left humerus, | complete right ulna, proximal end of | right ulna, distal ends of 5 right and 4 left ulnae, proximal ends of | right and | left radius, distal ends of 3 left radii, 4 complete right and 1 complete left carpometacarpus, proximal ends of 4 right and 2 left carpometacarpi, 1 complete pelvis, 2 complete right and 1 complete left femur, proximal ends of 2 right and 4 left femora, distal ends of 4 right and 4 left femora, proximal ends of 2 left tibiotarsi, distal ends of 3 right and 6 left tibiotarsi, | complete left tarsometatarsus, proximal ends of 2 right and 2 left tarsometatarsi, distal ends of 2 right and | left tarsometatarsus. ROM 13801-13891. The 91 specimens represent a minimum of 7 individuals. Remarks Chloephaga melanoptera is primarily resident in the high Andes, living up to the snow line. Individuals do, however, move to lower elevations during the non-breeding season. It has not been recorded in the Andes closer than 100 miles south of the fossil site (Koepcke, 1970). Since C. melanoptera feeds almost exclusively on grass and is seldom seen on water (Delacour, 1954), the presence of an abundant ground cover of vegetation near the fossil site would seem to be more likely to attract this species than a body of water surrounded by sparse vegetation. A number of the specimens appear to represent immature individuals. Subfamily Anatinae (Vigors) “Humerus with (1) capital shaft ridge obsolete; (2) pneumatic fossa opening, containing bony struts; (3) ectepicondyle typically equal or subequal in height to entepicondyle.’’ (Woolfenden, 1961: 109). Cairina Fleming Cairina moschata (Linnaeus) Muscovy Duck Material One right articular, 4 left scapulae, 2 complete right coracoids, humeral end of | right coracoid, proximal ends of 2 right humeri, distal ends of 2 right and | left humerus, 2 complete right ulnae, proximal end of 1 left ulna, distal ends of | right and 3 left ulnae, proximal ends of | right and 1 left radius, distal ends of 1 right and 1 left radius, 1 complete left carpometacarpus, proximal ends of 1 right and 1 left carpometacarpus, distal ends of 2 right tibiotarsi, 1 complete left tarsometatarsus, proximal end of | left tarsometatarsus, distal ends of 2 left tarsometatarsi. ROM 13892-13924. The 33 specimens represent a minimum of 4 individuals. Remarks Cairina moschata, a forest duck, occurs on the arid coast of Peru only by accident. It has been recorded only once, near Lima. Marchant (1958) did not report it from the arid Santa Elena Peninsula of Ecuador. Chapman (1926) found this species near the 43 coast only 35 km north of the Peru-Ecuador border, and less than 200 km from the fossil site, but in a completely different type of habitat. The sedentary habits of this species would seem to indicate either that its range extended much farther south than today, increasing the number of stragglers, or that it was a resident. If Cairina was a resident at the site its preference for water surrounded by forests would indicate a considerable amount of vegetation in the area. Anas Linnaeus Anas talarae sp. nov. Fig 86.8!) 9B) EG, 10B, F, 11C,G Holotype Complete left humerus (internal tuberosity and bicipital crest broken). ROM 12904. Referred Material One right scapula, 2 complete right coracoids, 1 complete right and 1 complete left ulna, distal end of 1 right ulna, proximal end of 1 left radius, distal end of 1 right radius, proximal end of | right carpometacarpus, distal ends of | right and 1 left tibiotarsus, 1 complete left tarsometatarsus. ROM 12905-12916. The 13 elements, including holotype, represent a minimum of 2 individuals. Diagnosis Humerus agrees with that of Anas and differs from that of all other genera of subfamily Anatinae by having those characters of the genus as described by Woolfenden (1961). Compared with the humerus of living species of Anas, the holotype is characterized by having (1) head of humerus small, profile rounded in palmar view, and projecting more anconally than distad (similar in Anas (Callonetta) leucophrys Vieillot; large, subrounded, and projecting more distad than anconally in Anas (Amazonetta) brasiliensis Gmelin, A. crecca Linnaeus, and A. cyanoptera (Vieillot); (2) head without inflection (similar in A. leucophrys, A. crecca, and A. cyanoptera; head inflected in A. brasiliensis); (3) internal tuberosity broken, but appearing prominent (similar in A. brasiliensis, A. crecca, and A. cyanoptera; reduced in A. leucophrys); (4) attachment of M. coracobrachialis posterior forming short, broad shelf, not extending down median crest (moderate to long, broad, usually not forming shelf, and usually extending down median crest in A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera); (5) median crest prominent (similar in A. /eucophrys and A. brasiliensis; slender to moderately prominent in A. crecca and A. cyanoptera); (6) pneumatic fossa elongated, not entering into internal tuberosity (moderately elongated, not entering into internal tuberosity in A. brasiliensis; slightly elongated and entering into internal tuberosity in A. leucophrys, A. crecca, and A. cyanoptera); (7) deltoid crest not flaring, with depression at distal end (similar in A. crecca, flaring, without depression in A. leucophrys and A. brasiliensis; not flaring, without 44 depression, and reduced in size in A. cyanoptera); (8) attachment of M. pectoralis superficialis rotated palmarly (similar in A. brasiliensis; not rotated in A. leucophrys, A. crecca, and A. cyanoptera); (9) shaft with moderate internal curvature immediately distal to bicipital crest, giving moderately sigmoid shaft (similar in A. leucophrys; sharp internal curvature in A. brasiliensis, giving strongly sigmoid shaft; similar length and curvature in A. leucophrys; short and wide, with very slight anterior shaft); (10) impression of M. brachialis anticus relatively deep, oriented at small angle to axis of shaft (deep in A. brasiliensis; shallow in A. leucophrys, A. crecca, and A. cyanoptera; oriented at large angle to axis of shaft in all); (11) attachment of anterior articular ligament longer than wide (width roughly equal to length in A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera); (12) entepicondyle rotated so internal side faces anterodistad (rotated strongly anteriad in A. leucophrys and A. brasiliensis; rotated slightly to moderately anteriad in A. cyanoptera; not rotated in A. crecca); (13) entepicondyle rounded distally, not produced (subrounded distally and very produced in A. leucophrys; subangular distally and moderately produced in A. brasiliensis; angular to subangular distally and moderately produced in A. crecca and A. cyanoptera); (14) internal tip of external condyle broken (undercut in A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera); (15) olecranol fossa long (similar in A. leucophrys and A. cyanoptera; short in A. brasiliensis and A. crecca). For measurement see Table 4. Measurements of the holotype are as follows: length, 56.4; proximal width, 13.0; distal width, 9.2; least width of shaft, 4.3. Characters Compared with living species of Anas, A. talarae is characterized by having scapula with (1) acromion of moderate length, narrow with strong anterior curvature (wide, of similar length and curvature in A. Jeucophrys; short and wide, with very slight anterior curvature in A. crecca; long and wide, with moderate anterior curvature in A. cyanoptera),; (2) glenoid facet of moderate length and narrow, lacking strong projection to anteroventral corner (similar, but long, in A. leucophrys; long and wide, with prominent anteroventral corner in A. brasiliensis, A. crecca, and A. cyanoptera); (3) attachment of Lig. furculo-scapulare dorsale oval, located dorsal to midline of shaft (similar in A. leucophrys, A. brasiliensis, and A. crecca; elongated and situated ventral to midline of shaft in A. cyanoptera). Coracoid with (1) coracohumeral surface long, relatively wide, not widening markedly anteriad, with moderate curvature, and positioned internally close to edge of triosseal canal (short, narrow, not widening markedly anteriad, of moderate curvature, positioned slightly internally in A. leucophrys; long, wide, widening markedly anteriad, of slight curvature, and positioned internally in A. brasiliensis; long, narrow, moderate to marked widening anteriad, of slight to moderate curvature, not positioned internally in A. crecca and A. cyanoptera); (2) attachment of Lig. humero-coracoideum anterius superius slightly elevated (similar in A. leucophrys and A. brasiliensis; not elevated in A. crecca and A. cyanoptera); (3) scapular facet small and elongated (similar in A. leucophrys and A. brasiliensis; large oval in A. crecca and A. cyanoptera). UlIna with (1) distal lip of external cotyla broad, extending moderately distad (similar in A. brasiliensis; internal side of external cotyla extends distad much farther than external side in A. leucophrys, A. crecca, and A. cyanoptera); (2) ridge leading 45 to shaft from carpal tuberosity large (similar in A. crecca; slight to moderate ridge in A. leucophrys, A. brasiliensis, and A. cyanoptera). Radius with no distinguishing characters other than size. Carpometacarpus with (1) process of metacarpal I low (i.e., short antero- posteriorly) and of moderate length (i.e., proximodistally) (low, with distal edge sloping sharply proximad in A. leucophrys; high and short in A. brasiliensis; of moderate height and length in A. crecca and A. cyanoptera); (2) proximal edge of metacarpal I sloping sharply proximad (similar in A. leucophrys and A. cyanoptera; moderate slope in A. crecca; slight slope in A. brasiliensis); (3) internal face of metacarpal I slightly excavated (similar in A. crecca; moderately to deeply excavated in A. leucophrys, A. brasiliensis, and A. cyanoptera); (4) proximal area of fusion of metacarpal II and metacarpal III very long (long in A. leucophrys and A. brasiliensis; short in A. crecca and A. cyanoptera); (5) external rim of carpal trochlea rounded proximally (subangular to angular in A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera); (6) area covered by external cuneiform ligament markedly convex (similar in A. leucophrys; slightly convex in A. brasiliensis; flat in A. crecca and A. cyanoptera); (7) posterior carpal fossa of essentially uniform width (similar in A. leucophrys and A. brasiliensis; wide distally, narrowing gradually proximad in A. crecca; wide proximally and distally with central constriction in A. cyanoptera). Tibiotarsus with (1) external face of distal end flat (similar in A. leucophrys, A. brasiliensis, and A. crecca; flat or slightly concave in A. cyanoptera); (2) internal condyle flaring mediad (not flaring in A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera). Tarsometatarsus with (1) intercotylar prominence broad (similar in A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera); (2) shaft narrowing sharply immediately distal to internal cotyla (similar in A. Jeucophrys, A. brasiliensis, and A. crecca; narrows gradually in A. cyanoptera); (3) hypotarsus short (similar in A. leucophrys and A. crecca; long in A. brasiliensis and A. cyanoptera); (4) anterior face of shaft straight in lateral view (similar in A. cyanoptera; slight concavity in A. brasiliensis and A. crecca; moderate concavity in A. leucophrys); (5) internal trochlea projecting distad without turning externally (similar in A. leucophrys, A. brasiliensis, and A. cyanoptera; turned slightly externally in A. crecca). Etymology This species is named for the town of Talara, Peru, from which the fossil site also takes its name. Remarks Anas talarae and the following two new species of Anas are compared with A. leucophrys, A. brasiliensis, A. crecca, and A. cyanoptera because the fossil species are all small, differing only slightly in size among themselves, and the above four living species represent the smallest forms of the genus. Although A. crecca does not reach South America it was used rather than A. discors, which does reach South America, because the two species are extremely difficult to separate osteologically, and of the two the former has smaller individuals. A. talarae is the Anas species 1 reported from La Carolina, Ecuador (Campbell, 1976). The numbers under each element correspond in the comparisons of all three new species. 46 Anas amotape sp. nov. Rise 8B, GC) IC aH, lOG BSH, 118, F Holotype Complete left humerus. ROM 12927. Paratypes One complete right and 1 complete left humerus, proximal ends of | right and 1 left humerus, distal ends of 2 right and | left humerus. ROM 12928-12934. Referred Material One left scapula, 2 complete right and 4 complete left coracoids, 1 complete right ulna, proximal ends of 2 left radii, 1 complete left carpometacarpus, proximal ends of 1 right and 1| left carpometacarpus, proximal end and shaft of 1 right femur, distal ends of 5 right and 1 left tibiotarsus, 1 complete left tarsometatarsus. ROM 12935-12955. The 29 specimens, including holotype and paratypes, represent a minimum of 5 individuals. Diagnosis Humerus agrees with that of Anas and differs from that of all other genera of the subfamily Anatinae by having those characters of the genus as described by Woolfenden (1961). Compared with the humerus of other species of Anas the holotype is characterized by having (1) head large, profile subrounded in palmar view, and projecting more anconally than distad; (2) head not inflected; (3) internal tuberosity reduced; (4) attachment of M. coracobrachialis posterior moderate to long, broad, forming slight shelf, extending down median crest; (5) median crest prominent; (6) pneumatic fossa round, entering moderately into internal tuberosity; (7) deltoid crest long, very flaring, without depression at distal end; (8) attachment of M. pectoralis superficialis not rotated palmarly; (9) shaft with sharp internal curvature immediately distal to bicipital crest, giving strongly sigmoid shaft; (10) impression of M. brachialis anticus extremely shallow, oriented at large angle to axis of shaft; (11) attachment of anterior articular ligament high and very small; (12) entepicondyle rotated moderately anteriad; (13) entepicondyle rounded distally, not produced; (14) internal lip of external condyle not undercut, but with slight ridge leading internally; (15) olecranol fossa short. For measurements see Table 4. Measurements of the holotype are as follows: length, 65.2; proximal width, 14.1; distal width, 9.7; least width of shaft, 4.5. Characters Compared with other species of Anas, A. amotape is characterized by having scapula with (1) acromion long and wide, with moderate anterior curvature; (2) glenoid facet long and wide, with anteroventral corner not projecting; (3) attachment of Lig. furculo-scapulare dorsale round, located in midline of shaft. 47 er ee ee Seer ee ae mS ti~C eee ees o EO EE EE EEE EE oOo 8 S I I r4 u }00R4 Qt Lol = = 6°Cl W [BUII}S Py yI-8°TI O:91-E-Cl v9! vl Seta Cnc re) jo w3sus7 8 Si I I I 9 Cc U yeys oe Le = rr = Ore Cre W jo yidaq OF-GeE PVauee ay me Be Onis te Lec te YO Le | 8 iS I I I v I U €9 L’‘9 = — = eo aed W yidaq O29 S=L—=G9 €'8 99 gS C96 v9 YO [BWIXOIdg 8 g I I I 1 I u eR) [Pegi 8 Il a = il Cat aa W Jepndeos cE s0ll SS eH 8 el (GI! €'Ol STS cs00 601 c[@) 0} peoH 8 S I I I € I u ONCE ESS a a > VAce a W PBEHS Le SiG6tro-Ce Deby, Bice LOE EC emOnce few |ls 6) ysuoy] proor10g 8 S I | I I uU EAE cc - = = — W YIPIM Ole SC Cava Or Okc SG Vac re [@) Ee erd 8 ¢ I I | I U iba €'8 a = = - W W319 i rciemte $16 Oi Om Sal o'6 lige v8 VL (6) [EUNTXOIg eindess DIIAAI piajdouvao SISUINISDAG sduydoona] “aou ‘ds “Aou ‘ds “aou ‘ds SPU sSpUYy spupy sDUPp IDUA]AYIDIIUDS adnjoun IDADIDI SDUp SDUPY SDUP “(WW UT) S2suau10409 DII2L9 “YY PUR ‘D4éajlOUDAD “WY ‘sisualisvig “Y ‘sduydoona] *y ‘(A[UO Sns.ieje}9UI0SIL} PUB ‘snsie}OIgH ‘snaawWINy ‘p1ode109) *aou “ds avuajayanjouvs spuy “cou ‘ds adpjown spuy ‘(anuiay SuIpnyjoxo) ‘AOU “ds apsnjD} spUY JO SNSAe}eJIUIOS.IB} PUB ‘sNsIeIOIQH ‘Inway ‘sndiesejauIOdAed ‘eUN ‘snipes ‘snsdWNY ‘prose.s0> ‘e[NdeIS ay) JO SjUaUIOANSKayYy Ff AIGE? 4§ 8 GOs. PUSses8 8 SL Ove Cv 8°Cl Cretaercl 8 6°LS ¢'6S-6'SS ‘OS aded uo ponunuod S vrs [ 199-08 gOS E SSee OV, 9c! Las Vee. ¢ 9°79 STO) £9 £99 g'19 8 PL Lvs LOS 86S er LOS eo) LOLs O°9¢ LSCn Ore 86 OI =ON(6) v cel LVipc cl (6 [seo e591 19 6 € €S ASS AWE Ce y9¢ uo ado rs (@) yisuay] Pu[f) UIPIAA [risiq YIPIM [PUIXOlg yisuay] snipey yeys JO UPI, Sea] WPI [eisiq YIPIM [PUITXO1g yisua7] sniswiny 49 50 8 S I I I u XIUIOJ Is Le) Sat 7 ro W [eIsIq OSS v 89-67 (Gye 67 Lev uO jo wsus7] 8 ¢ I I G I u eys gs 3 te a — Lee a W JO WIPIM Gtea Cc LCA BiG Be ce E-0'e CE rc(@) }SeaT] 8 ¢ I I (6 I u [ [edivorjapy L’8 G9) a < 6°8 aes W ysnoly 06-78 Clits Oot €'8 6°8 9°8 (6) Ysa 8 ¢ I I € | u Ore It eed - cv a W WIPIM eveRce LV20-€ 61 6€ 4 O'r uO [BWITXO1g 8 ¢ I I I uU COL 6 8e —— a = W TLE-S Pe GPO 9 0-9 ES Ge re (@) ysuaT] sndiesejawodieg 8 ¢ I I I ¢ u yeys Ve Ce 3 = tay ve W JO WPI aE ae Otaes vv Wee Be Smet ado SEO] 8 S I I I € u 8°9 ee oa = 3 9°9 W yidaq OL) (Gitsyml eS) 9°8 6°9 Y) i LoS!) rs (0) [BIST 8 S I I I I U sy) ye = = - aa W UPI (OG) £359 9 68 89 Ie 89 (6) [BWTXOlg DIIAAI piajdouvao SISUAI]ISDAG sduydoona] “aou ‘ds “aou ‘ds ‘aou ‘ds SDUY spup spuy SDUY IDUI] AYADJIUDS adpvjouw IDAD]DI spuy SDUP spuy ee ee ee ere see 7 ee ee ee Re 8 SO ee ee eS ‘7G ased uo ponunuog vs SLISee GS 6°CE 2) tea) fe re) 6 9nd Ss ¢ LLS CS P9-l’es 8 Ve CS CaS abe LL) veV [ss Once v9 0-9 C950 9 BE One S SACO ado re (@) ajApuoa [BuId}U] Jo yndaq YeYS JO PIM, }SPIT] UPL [easiq YIPIM [PUIIXOIg ysuaT] SNSIVIOIQL, Yyeys JO YIPIM SPIT] PIM. easiq WPI [BWIIXO1g ysuay Inwdj ay I U BITYOOL], LC Lc a aa = ies om W 9IPPIIN OrGas CG Ee Le 6c BC GE 6°C ado JO WPI 8 ¢ I | I I I u yeys BC Pee. a = =e oa = W JO WPIM Ont Sec She Sam bint Gre Ore ONC Sue Ore (8) Ag | 8 S I I I I Lf u 8S c9 = “i =o ie — W YIPIM Ose < LDeSrS [eal 59 (GE) Gy Si) re (0) [e1siq 8 g I I | I I u 6S £9 + - ae ae a W UPI OSALS Orga 9 8 oz Te) eS) ee Oo re[@) [BWIXOld 8 ¢ I I ! I I u 9°6C BCE = = i. = . 3 W 8 Te-I 8c SStac UC py 8t 6 6¢ ¢ 6¢ SIS e 8 0€ (0) yisucT SNSIe}LJOWOSIL |, DIIAAI viajdouvKo S1SUA1]1SDAQG skaydoona] “aou ‘ds “aou ‘ds “aou ‘ds spup spup spuy spuy DUA] AYaDJIUDS advjowv IDADIDI sSDUup sDUPp spuy a2 Coracoid with (1) coracohumeral surface wide and of moderate length, with moderate to marked widening anteriad, moderate to marked curvature, and positioned slightly internally; (2) attachment of Lig. humero-coracoideum anterius superius elevated; (3) scapular facet large oval. Ulna with (1) internal side of external cotyla extending much farther distad than external side; (2) ridge leading from carpal tuberosity to shaft very small. Radius with no distinguishing characters other than size. Carpometacarpus with (1) process of metacarpal I long and low; (2) proximal edge of metacarpal I sloping slightly proximad; (3) proximal area of fusion of metacarpal II and metacarpal III long, and distal area of fusion short (distal area short in A. leucophrys; long in A. brasiliensis and A. crecca; very long in A. cyanoptera); (4) internal face of metacarpal I slightly excavated; (5) external rim of carpal trochlea subangular proximally; (6) area covered by external cuneiform ligament markedly convex; (7) posterior carpal fossa wide distally, narrowing abruptly proximad. Femur with (1) head large, projecting mediad (small to moderate sized head projecting mediad in A. leucophrys, A. crecca, and A. cyanoptera; very large head projecting equally dorsad and mediad in A. brasiliensis); (2) attachment of M. ilio-trochantericus medius very long (long in A. leucophrys; of moderate length in A. brasiliensis and A. crecca; short in A. cyanoptera); (3) iliac facet curving sharply mediad posteriorly (moderate curvature in A. leucophrys and A. brasiliensis; slight if any curvature in A. crecca and A. cyanoptera). Tibiotarsus with (1) external face of distal end markedly concave; (2) internal condyle not flaring mediad. Tarsometatarsus with (1) intercotylar prominence broad; (2) shaft narrowing sharply immediately posterior to internal cotyla; (3) hypotarsus long; (4) anterior face of shaft straight in lateral view; (5) internal trochlea turned moderately externally. Etymology This species is named for the Amotape Mountains, an outlier of the Andes Mountains, that overlook the fossil site. Anas sanctaehelenae sp. nov. Figs ob. 9D, LID. i Holotype Complete left humerus. Uncatalogued. Type Locality La Carolina, - Ecuador (Campbell, 1976). Paratypes Proximal end of 1 right humerus, distal end of | right humerus. ROM 12899-12900. Referred Material One complete left coracoid, distal end of 1 left tibiotarsus, 1 complete right ei tarsometatarsus. ROM 12901-12903. The 5 specimens, excluding holotype but including paratypes, represent a minimum of | individual. Diagnosis Humerus agrees with that of Anas and differs from that of all other genera of the subfamily Anatinae by having those characters of the genus as described by Woolfenden (1961). When compared with the humerus of other species of Anas, the holotype is characterized by having (1) head large, profile rounded in palmar view, and projecting more distad than anconally; (2) head not inflected; (3) internal tuberosity reduced; (4) attachment of M. coracobrachialis posterior short broad shelf, not extending down median crest; (5) median crest moderately prominent; (6) pneumatic fossa slightly elongated, entering only slightly into internal tuberosity; (7) deltoid crest long, moderately flaring, without depression at distal end; (8) attachment of M. pectoralis superficialis not rotated palmarly; (9) shaft with moderate internal curvature immediately distal to bicipital crest, giving moderately sigmoid shaft; (10) impression of M. brachialis anticus shallow, oriented at large angle to axis of shaft; (11) attachment of anterior articular ligament as wide as it is long; (12) entepicondyle rotated moderately anteriad; (13) entepicondyle rounded distally, not produced, very small; (14) internal tip of external condyle not undercut, but with slight ridge present leading internally; (15) olecranol fossa short. For measurements see Table 4. Measurements of the holotype are as follows: length, 57.7; proximal width, 12.3; distal width, 9.1; least width of shaft, 4.4. Characters Compared with other species of Anas, A. sanctaehelenae is characterized by having coracoid with (1) coracohumeral surface short, narrow, widening moderately anteriad, of moderate curvature, and not positioned internally; (2) attachment of Lig. humero-coracoideum anterius superius elevated; (3) scapular facet large oval. Tibiotarsus with (1) external face of distal end flat; (2) internal condyle flaring mediad. Tarsometatarsus with (1) intercotylar prominence slender; (2) shaft narrowing sharply immediately posterior to internal cotyla; (3) hypotarsus short; (4) anterior face of shaft markedly convex in lateral view; (5) internal trochlea strongly turned externally. Etymology This species is named for the Santa Elena Peninsula of Ecuador, where the La Carolina site is located. Remarks The holotype of Anas sanctaehelenae is from La Carolina, Ecuador, not the Talara Tar Seeps. The Ecuadorian specimen was chosen for the holotype because of its completeness, and the lack of a complete specimen in the Talara material. The paratypes and all referred material are from the Talara Tar Seeps. A. sanctaehelenae is the Anas species 3 reported from La Carolina (Campbell, 1976). 54 Anas bahamensis Linneaus Bahama Pintail Material Portions of 14 crania, 5 frontals, 9 right and 11 left quadrates, 3 right and 3 left articulars, 5 right and 7 left mandibular fragments, 122 right and 106 left scapulae, 172 complete right and 128 complete left coracoids, humeral ends of 71 right and 87 left coracoids, 35 complete right and 34 complete left humeri, proximal ends of 86 right and 104 left humeri, distal ends of 103 right and 110 left humeri, 71 complete right and 64 complete left ulnae, proximal ends of 69 right and 61 left ulnae, distal ends of 87 right and 78 left ulnae, 14 complete right and 19 complete left radii, proximal ends of 83 right and 61 left radii, distal ends of 52 right and 41 left radii, 135 complete right and 103 complete left carpometacarpi, proximal ends of 73 right and 73 left carpometacarpi, 29 complete right and 21 complete left femora, proximal ends of 22 right and 25 left femora, distal ends of 11 right and 7 left femora, 5 complete right and 4 complete left tibiotarsi, proximal ends of 26 right and 15 left tibiotarsi, distal ends of 102 right and 91 left tibiotarsi, 60 complete right and 65 complete left tarsometatarsi, proximal ends of 17 right and 19 left tarsometatarsi, distal ends of 24 right and 21 left tarsometatarsi. ROM 13925-16788. The 2,864 specimens represent a minimum of 243 individuals. Characters Only two species (A. bahamensis and A. discors Linnaeus) of the subfamily Anatinae have been previously recorded from northwestern Peru, although I have seen A. cyanoptera (subspecies not determined) near Ciudad de Piura, Departamento de Piura (4 October 1971). A. bahamensis is easily separated on size from these two species, including the larger subspecies of A. cyanoptera, A. c. orinomus. Anas flavirostris Vieillot, A. georgica Gmelin, A. puna Tshudi, and A. platalea Vieillot also occur in Peru today, but not in the northwestern section. A. bahamensis is readily separable from A. puna and A. platalea by numerous osteological characters, and from the smaller A. flavirostris by size. Elements of A. bahamensis are usually smaller than those of A. georgica, the most closely related species, but the two species overlap in size. A. bahamensis differs from A. georgica by having scapula with (1) concavity dorsal to proximal end of glenoid facet deep; (2) acromion longer. Coracoid with (1) coracohumeral surface narrower and shorter; (2) head more rounded in medial view; (3) procoracoid projecting farther dorsad. Humerus with (1) capital groove narrower; (2) internal tuberosity projecting less anconally; (3) attachment of Lig. humero-coracoideum anterius inferius smaller and positioned anconally on internal tuberosity; (4) concavity immediately external to attachment of M. proscapulohumeralis much narrower; (5) ectepicondylar promi- nence larger; (6) entepicondyle smaller; (7) external condyle narrower. Ulna with (1) olecranon longer; (2) external cotyla smaller, with distal projection smaller; (3) internal cotyla more concave. Radius with (1) capital tuberosity smaller; (2) ligamental prominence narrower, more distinct, and situated more laterad. ys) Carpometacarpus with (1) process of metacarpal I slightly longer and more rectangular, lacking sharp anterior curvature; (2) posterior carpal fossa larger; (3) internal scapholunar ligamental attachment on proximal end of metacarpal III usually positioned proximally; (4) proximal area of fusion of metacarpal II and metacarpal III shorter. Femur with (1) head smaller; (2) iliac facet narrower; (3) internal condyle not as deep; (4) external condyle not as deep, more rounded, and extending farther proximad. Tibiotarsus with (1) concavity between internal articular surface and rotular crest shallower; (2) area between proximal internal ligamental attachment and flexor attachment slightly convex (very convex in A. georgica); (3) condyles less deep. Tarsometatarsus with (1) cotylae smaller; (2) external ligamental attachment positioned proximally; (3) trochleae less compressed; (4) middle trochlea extending less proximad. Remarks Anas bahamensis 1s widespread throughout tropical and subtropical South America, and is capable of rapidly colonizing temporary bodies of water in large numbers, even in very arid areas (Marchant, 1958). It is very characteristic of coastal lagoons and where found is present in large numbers. The large number of fossil specimens of this species represents almost one-half of the collection of fossil non-passerine birds from this site. Spillman (1942) described Archeoquerquedula lambrechti from late Pleistocene deposits of the Santa Elena Peninsula of Ecuador. Although I have not seen his material, I do have a number of topotypes and these agree with Anas bahamensis. Spillman did not describe any of the elements referred to A. lambrechti, but he did figure a skull. From his figure it is possible to determine that the length of the orbit is greater than the distance from the posterior edge of the orbit to the parietal. Of the species of Anas found in South America this condition occurs only in A. bahamensis. In all other species the distance from the posterior edge of the orbit to the parietal is greater than the length of the orbit. Therefore, I suggest that Archeoquerquedula lambrechti be synonomized with Anas bahamensis. Subfamily Oxyurinae Phillips Nomonyx Ridgway Nomonyx dominicus (Linnaeus) Masked Duck Material One complete left humerus. ROM 16789. Remarks Nomonyx dominicus has been recorded as far south in western Peru as Lambeyeque, 56 and Marchant (1958) reported it as scarce on the arid Santa Elena Peninusla of Ecuador. It is sedentary in habits, but strays great distances. It would appear from the single specimen that N. dominicus was not a regular resident at or near the fossil site. Discussion of Family Anatidae The nine species of waterfowl found in the Talara Tar Seeps avifauna represent slightly more than 10 per cent of the total number of non-passerine species known from the avifauna and give a good indication of what the habitat was like during the period of entrapment. Although some of the species are well-known colonizers of temporary ponds (Anas bahamensis), or widely wandering species (Nomonyx dominicus), others indicate either the presence of grasses in abundance (Chleophaga melanoptera) or forests (Dendrocygna autumnalis and Cairina moschata). The most interesting species are the four that became extinct. Special note should be made that they are all of very small size. The large to moderate number of elements representing each species, and the recognition of three of the four species in late Pleistocene deposits of southwestern Ecuador, is sufficient to remove any doubt that they are indeed separate species. This is particularly fortunate when considering the possibility of confusion among the three new small species referred to the genus Anas. It is possible that the four extinct species were derived from similarly sized, 1.e., small, ancestral species. However, the fact that the four species are all small also suggests the possibility that there may have been strong competition for available food resources. Schoener (1969, 1970) has constructed a model, using species of West Indies Anolis, demonstrating how species of similar size may converge toward a smaller size when in competition for food. Additional selective pressures contributing to the evolution of small sized individuals within a species would result from regular periods of extreme food stress because, other things being equal, small individuals require less food than large individuals. In this geographic region it is easy to visualize extreme food stress resulting from drought conditions. On the basis of the large number of specimens of ducks present in the deposit, however, I believe that there was sufficient water present to maintain resident populations of each species for a long period of time. The late Pleistocene age of the fossil deposits suggests that the four extinct species were endemic to northwestern Peru and, in all but one case, to southwestern Ecuador (Campbell, 1976). It also suggests that their extinctions were related in some way to the disappearance of suitable habitats. It is difficult to understand why not one species was able to survive elsewhere for such a short period of time, in the same way the five recorded extant species of widespread anatids have survived, unless the four species were endemic to the indicated region. Water is one of the most important factors in the ecology of ducks, and it must have been present when the fossil specimens were deposited. It is not present at the site now. The removal of the water supply would result in local extirpation of wide-spread species and the extinction of endemic species dependent upon water. If the four species were endemics, there must have been a sufficient period of habitat stability for their speciation. However, the process of speciation resulting in the three extinct species of Anas may have happened as rapidly as it now appears to be occurring in A. cyanoptera, where one finds three allopatric subspecies of quite different size classes in northern South America (Snyder and Lumsden, 1951). 7) ba | There is, of course, the possibility that future discoveries of fossil sites will show that one or more of the four extinct species were not endemic to the area as believed. I do not think that this in itself would alter the conclusions reached concerning the habitat at the Talara Tar Seeps during the time of deposition of the fossils, however, primarily because of the additional supportive evidence from other members of the known avifauna. The three new species referred to the genus Anas tend to fill in some of the larger gaps in osteological characters between typical Anas and Anas (Callonetta) leucophrys and Anas (Amazonetta) braziliensis noted by Woolfenden (1961). For this reason I have considered the latter two species and the three new species of the subfamily Anatinae as members of a larger, more inclusive genus Anas rather than proposing new genera for the new intermediate forms. Order Accipitriformes (Vieillot) Suborder Sarcoramphi (Ridgway) Family Vulturidae (Illiger) Geronogyps gen. nov. Diagnosis Tarsometatarsus agrees with that of Vu/tur Linnaeus and Gymnogyps Lesson by having (1) intercotylar prominence large; (2) anterior metatarsal groove very deep with large attachment of M. tibialis anticus located immediately distal to proximal foramina; (3) hypotarsus flattened, roughly triangular in shape, with central ridge leading distad; (4) shaft robust, rectangular, widening proximad and distad to meet articular surfaces; (5) external and internal trochleae much smaller than middle trochlea and turned slightly toward midline of shaft; (6) size large. Tarsometatarsus differs from that of Vultur and Gymnogyps by having (1) intercotylar prominence very high and narrow (low and broad in Vultur, high and of moderate width in Gymnogyps); (2) internal cotyla very deep, narrow, with very high, almost straight internal edge (deep, wide, internal edge moderately curved in Vultur; deep, of moderate width, with internal edge rounded in Gymnogyps); (3) external cotyla round, with articular surface extending onto external anterior metatarsal ridge (very elongated with articular surface restricted in Vultur, oval with articular surface restricted in Gymnogyps); (4) hypotarsus narrow, deeply grooved, moderately notched externally, with central ridge leading distad narrow (broad, very deeply grooved, deeply notched externally, with central ridge broad in Vultur; of moderate width, moderately grooved, moderately notched externally, with central ridge broad in Gymnogyps); (5) anterior metatarsal groove of moderate width and appears slightly pneumatic proximal to attachment of M. tibialis anticus (wide and highly pneumatic in Vultur, narrow and moderately pneumatic in Gymnogyps); (6) attachment of M. tibialis anticus very large, with length equal to or exceeding length of anterior metatarsal groove proximal to it (of moderate size, wider than long, much shorter than proximal portion of: anterior metatarsal groove in Vultur; of moderate size, length greater than width, much shorter than proximal portion of anterior 58 metatarsal groove in Gymnogyps); (7) attachment of M. extensor hallucis longus located approximately halfway between base of anterior metatarsal groove and top of internal metatarsal ridge (located near top of internal metatarsal ridge in Vultur and Gymnogyps); (8) external anterior metatarsal ridge broad and rounded and internal anterior metatarsal ridge of moderate width but with sharp proximal edge (broad, rounded external, and narrow, moderately rounded internal anterior metatarsal ridge in Vultur; both internal and external anterior metatarsal ridges broad and rounded in Gymnogyps); (9) shaft moderately depressed external to hypotarsus (slightly, if any, depressed in Vultur, moderately to deeply depressed in Gymnogyps); (10) shaft transversely concave posteriorly at midpoint of shaft (flat to rounded in Vultur; flat to slightly concave in Gymnogyps); (11) shaft widens slightly to moderately to meet internal cotyla, widens abruptly to meet external cotyla (widens abruptly to meet both external and internal cotylae in Vultur, widens moderately to abruptly to meet internal cotyla and abruptly to meet external cotyla in Gymnogyps); (12) shaft widens much more to meet internal trochlea than it does to meet external trochlea, placing middle trochlea off centre (widens equally to both internal and external trochleae, placing middle trochlea near midline of shaft in Vultur; similar in Gymnogyps); (13) internal trochlear surface small (large in Vultur; moderate to large in Gymnogyps; (14) internal trochlea extending distad to almost reach distalmost point of middle trochlea (does not project as far distad in Vultur and Gymnogyps); (15) internal trochlea with rounded posterior edge (posterior edge arrow shaped, directed posteriad in Vultur; arrow shaped, directed posterioproximad in Gymnogyps); (16) middle trochlea short (moderate to long in Vultur and Gymnogyps); (17) external trochlear surface large, not delimited sharply anteroproximally (moderate, sharply delimited anteroproxi- mally in Vultur; moderate to large, moderately delimited anteroproximally in Gymnogyps); (18) external trochlea with distinct posteroproximal corner to posterior edge (rounded in Vultur; rounded to prominent proximal projection in Gymnogyps); (19) external trochlear notch moderately wide (very wide in Vultur; moderately wide in Gymnogyps). Type Species Geronogyps reliquus sp. nov. Etymology From Greek, gerontos, genitive of geron, old man, and gyps, masculine, vulture. Geronogyps reliquus sp. nov. Bigel2Ae B, MAAS LSA, Biol 7As C,:E; iS Bev lGA Bi CyDeZ0A Holotype Complete right tarsometatarsus. ROM 12986. Paratypes Complete right tarsometatarsus and distal end of 1 right tarsometatarsus. ROM 12987-12988. 59 Referred Material Portions of 2 sterna, 1 complete left coracoid, 2 left coracoids lacking heads, humeral end of 1 right coracoid, sternal end of 1 left coracoid, shaft and distal end of 1 left humerus, proximal end of | right humerus, distal end of | right humerus, proximal end of | left ulna, distal end of 1 left ulna, proximal ends of | right and 1 left radius, distal ends of 1 right and 2 left radii, proximal end of | right carpometacarpus, distal end of | left tibiotarsus, proximal ends of 4 right femora, distal ends of 2 right femora. ROM 12989-13013. The 28 specimens, including holotype and paratypes, represent a minimum of 4 individuals. In addition, the following elements are tentatively referred to this species but cannot be assigned with certainty: 2 left articulars, 1 left quadrate, portion of 1 cranium. ROM 13014-13017. Diagnosis As for genus. For measurements see Table 5. Measurements of holotype are as follows: length, 121.8; proximal width, 25.4; distal width, 29.0; least width of shaft, 13.8; width of middle trochlea, 10.9. Characters Sternum with (1) ventral lip of coracoidal sulcus markedly concave along ventral edge, matching short external anterior sternal facet of coracoid (not concave in Vultur and Gymnogyps); (2) carinal apex situated very far distally (situated short distance distally in Vultur, moderate distance in Gymnogyps); (3) carina large (of moderate size in Vultur and Gymnogyps), (4) ventral manubrial spine very broad (broad in Vultur, narrow in Gymnogyps). Coracoid with (1) coracoidal fenestra lacking (present in Vultur and Gymnogyps); (2) shaft leading distad from external corner of glenoid facet forming large ridge (slight to moderate in Vultur, very slight in Gymnogyps); (3) shaft with internal contour very concave between procoracoid and anterior sternal facet (not as concave in Vultur and Gymnogyps); (4) external portion of anterior sternal facet short, distinctly set off from base of shaft distally (long, not distinctly set off in Vultur and Gymnogyps); (5) procoracoid directed internally along plane of glenoid facet (directed more anteriad in Vultur and Gymnogyps). Humerus with (1) head narrow (wide in Vultur, narrow in Gymnogyps); (2) capital shaft ridge dropping off sharply on both sides (shaft more rounded in Vultur and Gymnogyps); (3) attachment of M. proscapulohumeralis brevis elevated (not elevated in Vultur, elevated or not elevated in Gymnogyps); (4) pectoral attachment very prominent (less prominent in Vultur, more similar in Gymnogyps); (5) deltoid crest flaring distally (not flaring as much in Vultur and Gymnogyps); (6) shaft ventral to M. pectoralis superficialis not depressed (depressed in Vultur, not depressed in Gymnogyps); (7) external condyle wide and short (narrow and long in Vultur and Gymnogyps); (8) internal condyle short (very long in Vultur, moderately long in Gymnogyps); (9) ectepicondylar prominence long, moderately produced proximad (of moderate length, not produced distad, greatly produced and angular proximally in Vultur; short, greatly produced and rounded in Gymnogyps); (10) internal condyle moderately rotated anteriad, giving moderately flexed distal end (greatly rotated in Vultur, moderately rotated in Gymnogyps); (11) impression of M. brachialis anticus moderately deep (very deep in Vultur and Gymnogyps). 60 Ulna with (1) olecranon of moderate size (very large in Vultur, small to moderate in Gymnogyps); (2) attachment of anterior articular ligament small, not extending far distad (large, extending very far distad in Vultur; large, extending far distad in Gymnogyps); (3) area between external and internal condyles flat in external view (internal condyle projects slightly farther distad in Vultur and Gymnogyps); (4) carpal tuberosity large, joins shaft abruptly (large in Vultur, small in Gymnogyps, merges more gently with shaft in both). Radius with (1) bicipital attachment large (small in Vultur and Gymnogyps); (2) bicipital tubercle large (large in Vultur, very small in Gymnogyps); (3) capital tuberosity bordered internally by deep notch, pierced by large nutrient foramen (shallow in Vultur, moderate in Gymnogyps, nutrient foramen small to moderate in both); (4) external corner of distal end rounded (flattened so angle is formed at external end of carpal facet in Vultur, more rounded in Gymnogyps); (5) tendinal groove bordered externally by high ridge (very slight ridge in Vu/tur, moderate ridge in Gymnogyps); (6) distal end with external edge curving sharply palmarly (almost straight in Vultur and Gymnogyps). Carpometacarpus with (1) process of metacarpal I low (i.e., short anteropos- teriorly), curving sharply proximad (high, with moderate proximal curvature in Vultur; moderately high, with slight proximal curvature in Gymnogyps); (2) external ligamental attachment a small deep pit bordered posteriorly by raised area (of moderate size, lying in a pit, with no raised area posteriorly in Vultur and Gymnogyps). Femur with (1) trochanter of moderate size, subrounded (large, well rounded in Vultur; of moderate size, subrounded in Gymnogyps); (2) attachment of M. obturator externis deep (larger, not as deep, positioned similarly in Vultur; similar size, not as deep, located more posteriorly in Gymnogyps); (3) attachments of M. ilio- trochantericus medius and M. iliacus confluent and with surface ribbed (separate and not ribbed in Vultur and Gymnogyps); (4) concavity immediately distal to posterior edge of iliac facet deep, bordered externally by large protuberance (shallow, very small protuberance in Vultur; very deep, moderate sized protuberance in Gymnogyps); (5) distal end of internal condyle subrounded in medial view (more rounded in Vultur; less rounded in Gymnogyps); (6) rotular groove moderately undercut distally (deeply undercut in Vultur, moderately undercut in Gymnogyps). Tibiotarsus with (1) internal condyle short anteroposteriorly (long in Vultur and Gymnogyps); (2) distal internal ligamental attachment very deep (deep in Vultur, moderately deep in Gymnogyps). Etymology Latin, reliquus, remaining, left over. Remarks The tarsometatarsus is traditionally made the holotype of fossil accipitriform species when available, and the sum of many tarsometatarsal characters of Geronogyps reliquus establishes it as a distinctive genus and species. Additional very apparent large differences found in the other available elements strengthen this belief; i.e., the shape of the sternum and the manner of articulation with the coracoid, the absence of a coracoidal fenestra, and the confluent ribbed attachments of M. ilio-trochantericus 6] I 4 I u }00R4 a Lacy. - = W [BUIDIS Scr I'py-t tp 9°8E YO jo yysus7] € I Cc I G u yeys vv = CLT = y9I W jo uidaq GO Fl-o' Sl 691 081-9 91 v9 Jes ioees ll re (@) Seva] C I c I I u 6 9¢ — 1'0¢ = Ta W yidoq C.8CaE SG 0°0e © 0£-8 6¢ I°8¢ © SC (0) [BWITXOld 6 I (6 I l u 008] 6 LV an 6° OS = = W Jetndess 6 1S-0' br OCS € 1S-S OS 6 9P (Oe (0) 0} peoH | I ¢ I I u 7 S86 —; "as W b68 1 €0l 8 86—C 86 9°COl c 96 ado yysuoT] proorel0d 4 | | U g°8 = . W UWIPIM VO VL 18 C'8 (6) [BWIIXO1g C I I u S'6¢ = - W IYSsIIH C 0-9 87 GLC SLE ado [BWIXOIg eindess ——— ass ea ‘aou ‘ds jad ‘uas sdoog ie] eIele JUd99y SNUDIUAOLIDI “aou ‘ds snnbijaa sdXksouosay snydkzs anqyn, sdksouwmdky aDpADMOY sdXsouumay "(wu ut) (e[ndeds SuIpnpsxa) *aou ‘ds ja ‘uasd snnbiyas sdé3ouosay pure ‘(anwiay pue sndsesejawiodsed Surpnydxa) ‘aou ‘ds aDpsDMOY sdidoumdy ‘(sndseseyawiodse) pue ‘snisuiny ‘ejndeds surpnypoxe) sdaag sey eseyey, ayy wos snyddis +4 ‘snyd&u8 AnjnA yIdIdY ‘Ssnumiusofiyvs sdiJoumdy JO snszeyejouiosse} puwe ‘snsiejoIqH ‘inulay ‘sndaesejauIodsed ‘euN ‘snipes ‘snsouNY ‘plosV109 ‘etndeds ay) Jo sJUsWIIINSeayL «SFR 62 ‘pg ased uo ponunuod € Vee EVC L.Ce G (eel a ClaGre! LSs oT 6 OLE ONES On ele fe Sh Ele) S| Cc 0 ele 0° 8c£—-0 867 Cc Sle S lad ic ¢ pss 9°SS-I'SS (6 Oned. VACI=DCD G ¢8L7 0°C6C-0'$97 O'6le O' POL OmeLe [ETC e l6l 9'6I-C 81 ¢ OLY 0'6v-P OY Cc EES € ys-S CS 6 ¢ 097 ‘0'S9C-0 9ST dO rs (@) ado yysuoq Bull) UIPLAA eisiq YIPIM [PUNIXO1g yisue7] snipey yeys JO UIPIAA Sed] UPL [easiq YIPIM [BUIXO1g yisue7] snIouiny 63 c I u XIUJOJ t 6l a W [esiq S0c-1 81 vl YO jo yysueq C I u yeys 8'€1 = W JO UIPIAN Ser Orel Cel ps6) SBI I Cc ! u | [edieorjapy me oe) ES P= W ysnoy | O'CE OrBE=o9E Bese ao IYsIOH ! G I u = a7 a = W UIPIM ysl VileC El g 91 6) [BWITXO1g G ! uU Cc 9El rs W e OVisGcel 0'6¢1 re L@) yisucT sndieorjawiodies G I U yeys 6 tI Be W JO YIPIM Vole. vil Sreil ro (@) PSIG I Cc I G u = LC = Sr W tindaq msc cj Gale, ONE L YC SESE re(@) [eistq I G Cc I Cc iu a cohen OSE ee 9°0£ W WIPIM 10¢ OVEsS CE uC SmE NS Siete MWe OG re[@) [PWIXO1g ‘aou ‘ds ja ‘uas sdaag Ie] Pree JUD.0y snuDd1Usofi]DI “aou ‘ds snnbyjaa sdksouosay snydaus anqjn, sd&souwmay aDpADMOY sdksouwmAy ee ee ee ee ee ee eee eee 64 ‘99 ased uo ponunuod SEE AG CEE oh Chem Git3 v GSC OV. Cu® CC Tvl O'STc INS GaGa G 9°0€ I EO Oke ON Ee G OSC CIC OVC 6 vl t Sls Sl (6 I) Qe WEal Se G 9°6C DOE-9°3C G 0°6C7C 0.cve-O0 OT? c L 81 ¢ 6I-0'81 G C6£ i lite Le C O'8e £ Stoo Le (4 6 OSI © 8Seo Cri OIC JEG G5| 8 CC Cc 8 VC eSCIG VC r= (6) ado a, Apuod [ePuId}Uy jo yidaq yeys JO YIPIMA Sea] UIPLAA [Pasi YIPIA [PUNIXOIg ysudy SnsIeOIQL] yeus JO YIPIM Sea] UPI [Pasi WIPIM [PWIXO1g ysua7] nwa 65 c (6 I 9 u BaTYoOIy OFT Gl = ely W 2IPPIN Oso Ol 6:1 Pa Ges ).(0)) uo JO YIPIM G @ I C U yeys 8 el Cc vl = Orel W JO WPI 8 el |e i ba 3 | Ov Sela el rs (6) Svea] Cc 6 I v u S°6C LeGe roy, ere W UPI 0°0t—-0°6C CFSE OF 8 6C O'VE-S '6¢ uo [Bisiq 6 I c I iL u 8ST 7 L 8? a (OTEG W YIPIMA CIV SC 0'0e S6G Ss ic ERSNG ONGC SSC ado [BWwTxO1g C C I Z u sali EXC a © 0c! W C ycl-8 ICI LVel-8 sil Seal O'ccI-S 811 rc (@) yisueT SNSIEJVIOWIOSIV J, ‘aou ‘ds ja ‘ua3 sdoag Ivy eleye }Ud99y snupludofijDdo ‘aou ‘ds snnbyjaa Sdk80u0s1aH snyd&as anyjn/ SdXsOUWAD aDpADMOY sdX80UWAD 66 medius and M. iliacus of the femur. From the above diagnosis of the tarsometatarsus it is evident that Geronogyps reliquus resembles Gymnogyps more than Vultur. Two specimens of Vultur gryphus and one of Gymnogyps californianus were available for comparison of characters. Gymnogyps Lesson Gymnogyps howardae sp. nov. Fic clS Awl 48), OAGHB., 17B, Dy 18A, 19E; 21A Holotype Complete right tarsometatarsus. ROM 12956. Paratypes One complete left tarsometatarsus, proximal ends of 2 right and 3 left tarsometatarsi, distal ends of | right and 3 left tarsometatarsi. ROM 12957-12966. Referred Material Portion of 1 sternum, | left scapula, 1 complete right coracoid, 1 right coracoid lacking head, 1 complete right and 1 complete left humerus, shaft and distal end of 1 right humerus, proximal end of 1 left humerus, distal ends of | right and 2 left humeri, proximal ends of | right and 1 left ulna, distal ends of 1 right and 2 left ulnae, distal end of | right radius, distal ends of 2 left tibiotarsi. ROM 12967-12985. The 30 specimens, including holotype and paratypes, represent a minimum of 4 individuals. Diagnosis Tarsometatarsus agrees with that of Gymnogyps and differs from that of Vultur and Geronogyps by having those characters of Gymnogyps listed in description of Geronogyps reliquus. Tarsometatarsus differs from that of Gymnogyps californianus (Shaw) and G. amplus L. Miller (Collection of Pierce Brodkorb No. PB498, Reddick, Florida) by having (1) intercotylar prominence with prominent concavity on external side (slight to deep groove present in G. californianus and G. amplus); (2) internal cotyla with moderately curved posterior rim (moderately to sharply curved in G. californianus and G. amplus); (3) hypotarsus extending internally to at least middle of internal cotyla (does not extend as far internally in G. californianus and G. amplus); (4) posterior projection of external cotyla large (small to large in G. californianus and G. amplus); (5) ridge leading distad from hypotarsus very prominent, extending distad for more than third of length of bone, with deep depression of shaft on both sides (moderately prominent, extending distad for less than third of length of shaft, with depression internally greater than that externally in G. californianus and G. amplus); (6) anterior metatarsal groove with marked concavity distal to attachment of M. tibialis anticus (no noticeable concavity to G. californianus and G. amplus); (7) anterior metatarsal ridges well rounded and approximately equal in size (external anterior metatarsal ridge larger and more rounded than internal anterior metatarsal 67 ridge in G. californianus and G. amplus); (8) internal anterior metatarsal ridge continues distad and forms a marked corner to shaft that ends at base of internal rim of middle trochlea (is not traceable as a distinct corner in G. californianus and G. amplus); (9) internal trochlea long, large, rotated posteriad, with longest part of posterointernal edge directed posteroproximad (short, small to large, not rotated as much posteriad, with sharp to rounded posterointernal edge directed more posteriad than proximad in G. californianus and G. amplus); (10) middle trochlea very long, lying parallel to main axis of shaft, rims straight, and ending posteriorly in gradual taper (short, distal end turned slightly inward from main axis of shaft, rims flaring and terminating posteriorly more abruptly in G. californianus and G. amplus); (11) external trochlea short, not extending as far distad as internal trochlea, rotated slightly posteriad, trochlear surface very long anteriorly, with posterior edge well rounded and directed posteriad (long, extending as far distad as internal trochlea, rotated slightly, if any, posteriad, trochlear surface short, with posterior edge directed posteriad or proximad in G. californianus and G. amplus). For measurements see Table 5. Measurements of the holotype are as follows: length, 122.0; proximal width, 26.9; distal width, 30.2; least width of shaft, 13.5; width of middle trochlea, 11.1. Characters Sternum with (1) ventral manubrial spine narrow (wider in G. californianus); (2) ventral lip of coracoidal sulcus curved along ventral edge (less curvature in G. californianus, straight in Vultur). Scapula with (1) internal surface very concave anteriorly (less in G. californianus, similar in Vultur); (2) distal end of dorsal acromion edge greatly elevated above shaft (similar, but not quite as sharply set off in G. californianus; moderately elevated, merging gradually with shaft in Vultur). Coracoid with (1) anterior sternal facet with small dorsal slope (moderate slope in G. californianus , sharp slope in Vultur); (2) scapular facet moderately elevated above shaft (slightly elevated in G. californianus, greatly elevated in Vultur); (3) head narrow (similar in G. californianus, wide in Vultur). Humerus with (1) head narrow (slightly wider in G. californianus); (2) attachment of M. proscapulohumeralis brevis elevated (not elevated in G. californianus); (3) capital groove dividing base of median crest and capital shaft ridge (does not extend as far distad in G. californianus); (4) internal tuberosity moderately prominent (very prominent in G. californianus); (5) deltoid crest small, with anconal side very concave (larger, with anconal side less concave in G. californianus); (6) shaft moderately robust (very robust in G. californianus); (7) internal condyle small (large in G. californianus), (8) external condyle moderately produced distad (very produced in G. californianus); (9) ectepicondylar prominence short, distinctly set off from shaft distally (long, not distinctly set off from shaft in G. californianus); (10) entepicondyle moderately produced (more produced in G. californianus); (11) impression of M. brachialis anticus deep (very deep in G. californianus). Ulna with (1) olecranon of moderate size (short, narrow, but deep in G. californianus); (2) attachment of anterior articular ligament long, elevated, and terminating abruptly (long, not elevated, terminating gradually in G. californianus), (3) internal condyle projecting slightly distal to external condyle (projects moderately 68 distal to external condyle in G. californianus); (4) carpal tuberosity long (shorter in G. californianus). Radius with (1) ulnar depression moderately deep (shallow in G. californianus), (2) shaft widening very abruptly internally for carpal facet (widens more gradually in G. californianus). Tibiotarsus with (1) distal internal ligamental attachment moderately deep (moderately deep in G. californianus, deep in Vultur); (2) internal condyle not flaring internally at anteroproximal end (similar in G. californianus, flares internally in Vultur). Etymology This species is named for Dr. Hildegarde Howard in recognition of her many outstanding contributions to avian palaeontology. Remarks The appearance in the fossil record of a South American species of Gymnogyps poses the question of its relationships with the G. amplus—G. californianus line of North America. The continent of origin of the genus is then drawn into question, but must remain uncertain until specimens of an earlier period are found. Vultur Linnaeus Vultur gryphus Linnaeus Andean Condor Material One complete right coracoid, distal end of 1 right humerus, proximal ends of 1 right and 1 left ulna, distal end of 1 right ulna, proximal end of 1 left radius, distal end of 1 left radius, proximal end of 1 left femur, distal ends of 3 left femora, 1 complete right tibiotarsus, proximal end of 1 right tibiotarsus, distal ends of 2 right and 1 left tibiotarsus, proximal end of 1 left tarsometatarsus. ROM 16790-16806. The 17 specimens represent a minimum of 3 individuals. Remarks Vultur gryphus is found almost throughout the Andes of South America and often on the Pacific coast. Discussion In addition to the two species of living condors, eight large extinct vulturid species have been described from the Pliocene and Pleistocene of North and South America, one each in the two living genera and the other six in extinct genera. The possibility of one or more of these extinct North American species being identical to an extinct form from the Talara Tar Seeps requires a discussion of their characters. Three extinct monotypic genera, Pliogyps Tordoff, Breagyps Miller and Howard, and Antillovultur Arredondo are known in the subfamily Vulturinae (Illiger). 69 Pliogyps fisheri Tordoff is much smaller than any of the Talara specimens, and additional characters listed by Tordoff (1959), such as a very large internal trochlea and a broad, rounded hypotarsal ridge, indicates that this species is not similar to any of the Talara species. The distinctions among Breagyps, Vultur, and Gymnogyps have been described (L. Miller, 1910; Miller and Howard, 1938; Howard, 1974) and the characters of Gymnogyps listed hold for G. howardae. An inspection of the illustrations of Breagyps clarki (L. Miller, 1910: 9; Howard, 1974) immediately reveals large differences between that species and Geronogyps reliquus. No specimens of Breagyps were available for comparison. The following characters of the tarsometatarsus of Breagyps are most significant in distinguishing it from Geronogyps: (1) intercotylar prominence of similar height, but more rounded and broader; (2) shaft widening equally internally and externally for proximal articular surfaces, giving symmetrical appearance; (3) internal trochlea not extending nearly so far distad as middle trochlea; (4) internal intertrochlear notch wide. Antillovultur was described from the late Pleistocene of Cuba (Arredondo, 1971; see also, Arredondo, 1976) on the basis of a fragmentary tarsometatarsus and the distal end of a humerus. The original description distinguishes this genus from Vultur and Gymnogyps. The following characters of Antillovultur, as determined from illustrations of the holotype, distinguish it from Geronogyps: (1) tubercle for tibialis anticus smaller, positioned more toward midline of anterior metatarsal groove, and with distal portion prominently elevated, constricting anterior metatarsal groove; (2) anterior metatarsal groove narrower and more clearly defined; (3) internal anterior metatarsal ridge very thick. The proximal and distal ends are missing from the holotype of Antillovultur. Gymnogyps amplus is considered to be either the direct ancestor but specifically distinct from (Fisher, 1944), or a temporal subspecies (Brodkorb, 1964) of, G. californianus. The few specimens of G. amplus available indicate a closer relationship to G. californianus than to G. howardae. Lonnberg (1902) named Vultur patruus from a tarsometatarsus and femur from the Tarija Valley of Bolivia, listing primarily familial characters for identification of the material. The primary character used for the separation of V. patruus from V. gryphus was the small size of the fossil specimens in relation to one Recent specimen available to Lonnberg. Although V. patruus is probably near the lower limit of the size range of V. gryphus, it is not significantly smaller than V. gryphus. In addition, the following characters of the tarsometatarsus of V. patruus, all taken from Lonnberg’s photograph of the specimen, are similar to those of V. gryphus, namely; (1) intercotylar prominence broad; (2) the bone appears symmetrical, with proximal and distal widening of shaft approximately equal; (3) presence of a prominent point of inflection near midpoint of internal side of shaft, distal to which a long curve leads to the edge of the internal condyle, positioning the narrowest point of the shaft distal to point of inflection; (4) middle trochlea short; (5) internal and external intertrochlear notches approximately equal. That V. patruus does not represent Geronogyps reliquus is easily seen by two characters of the latter, i.e., (1) shaft widening much more externally than internally for proximal articular surfaces, giving an asymmetrical bone; (2) internal trochlea extending almost as far distad as middle trochlea (internal trochlea does not extend nearly so far distad in Vultur). 70 V. patruus superficially resembles G. californianus and G. amplus more than G. howardae, but G. howardae is easily distinguished by having (1) intercotylar prominence narrow; (2) shaft widening more gradually for proximal articular surfaces, more abruptly for distal articular surfaces; (3) no distinct point of inflection on internal side; (4) middle trochlea much longer; (5) trochlear surfaces asymmetrical. Although the associated material was considered Pliocene by Lonnberg, later studies by Boule (1920) and Oppenheim (1943) have placed the age of the Tarija beds as Pleistocene. From the above information it appears that V. patruus is a specimen of V. gryphus and should be placed in synonymy with the latter. This belief was expressed earlier by Fisher (1944: 294). The three species known from the subfamily Teratornithinae (L. Miller) are so distinctive they are readily seen to bear no resemblance to the Talara specimens. Monotypic genera are perhaps undesirable from certain viewpoints, but the discovery of another species of Gymnogyps indicates again that the taxonomy of living species is at best incomplete without greater knowledge of the fossil record, which is often slow in coming. The ‘‘lumping’’ of Vultur, Gymnogyps, and Pliogyps into one genus, Vultur, as suggested by Mayr and Short (1970) appears to be an example of eliminating monotypic genera solely for the purpose of reducing the number of taxonomic names and not for valid systematic, or phylogenetic, reasons. Sarcoramphus Dumeril Sarcoramphus? fisheri sp. nov. Fig eZ 2ApB iC Holotype Distal end of right tibiotarsus. ROM 12887. Referred Material Distal end of 1 right ulna, proximal end of | left radius. ROM 12888-12889. The 3 specimens, including holotype, represent a minimum of | individual. Diagnosis Compared with Vultur, Gymnogyps, Coragyps, and Cathartes, Sarcoramphus has tibiotarsus with (1) external condyle of greater depth than length (similar in Cathartes; more rounded in Gymnogyps, Vultur, and Coragyps); (2) tendinal groove passing under supratendinal bridge along straight line (similar in Cathartes; curving internally before reaching supratendinal bridge in Gymnogyps, Vultur, and Coragyps); (3) external condyle broad (similar in Gymnogyps, Vultur, and Coragyps; narrow in Cathartes). Tibiotarsus differs from that of Sarcoramphus papa (Linnaeus) by having (1) shaft very flattened posteroproximally to internal condyle (less flattened in S$. papa); (2) external condyle with rather even slope to medial surface (well-rounded in S. papa); (3) size larger, width and depth 40 mm proximal to base of intercondylar groove, 11.8 and 9.6, respectively (in S. papa, width, 10.2—-10.3, mean, 10.2; depth, 7.7-8.1, mean, 7.9; n = 3). 71 Characters Ulna with (1) external condylar surface passing medially into broad, shallow depression in shaft (passes onto shaft, but no depression in S. papa; does not pass onto shaft in Gymnogyps, Vultur, Cathartes, or Coragyps); (2) pneumaticity lacking in palmar surface (similar in §. papa, Gymnogyps, and Vultur; present in Cathartes and Coragyps). Distal width, 21.5 (in S. papa, 14.2-15.9; mean, 14.8; n = 3). Radius with (1) ligamental attachment of internal side of capital tuberosity divided into two distinct sections (similar in S$. papa; not divided into two distinct sections in Gymnogyps, Vultur, Cathartes, or Coragyps); (2) bicipital tubercle extending much farther distad than long attachment for M. biceps (similar in S. papa; does not extend as far distad in Gymnogyps, Vultur, Cathartes, or Coragyps, and attachment of M. biceps much shorter in all but Gymnogyps). Proximal width, 12.6 (in S. papa, 8.5—9.2; mean, 8.9; n = 3). Etymology This species is named for Dr. Harvy I. Fisher in recognition of his many contributions to our knowledge of New World vultures. Remarks Unfortunately, Sarcoramphus? fisheri is represented by such poor material that its true generic status must remain questionable until further collections of fossil material are made. That the material represents a large species of Sarcoramphus is reasonably certain, or at least it is much more similar to that genus than any other living genus of vulture. That the elements are intermediate in size between S. papa and Vultur indicates that they represent a new species as there is no living vulturid of comparable SIZe. One additional extinct species of Sarcoramphus, S. kernense (L. Miller) from the Middle Pliocene of California (L. Miller, 1931), has been described, and it is also larger than S. papa. Although the holotype of S. kernense, which is the distal third of a humerus, is not comparable with the specimens referred to Sarcoramphus? fisheri, the great difference in time and space between the occurrence of the two species makes it highly unlikely that the specimens of Sarcoramphus? fisheri actually represent S. kernense. Coragyps Geoffrey Coragyps cf. atratus (Bechstein) Black Vulture Material Four premaxillaries, 2 left quadrates, 1 right articular, 4 right and 3 left scapulae, 4 complete right and 3 complete left coracoids, humeral ends of 2 right coracoids, Ly sternal end of 1 left coracoid, proximal ends of 3 right and 2 left humeri, distal ends of 4 right and 2 left humeri, 2 complete right and 1 complete left ulna, proximal ends of 2 right and 1 left ulna, distal ends of 2 right and 2 left ulnae, proximal ends of 2 right and 5 left radii, distal ends of 3 nght and 2 left radii, proximal ends of 2 left carpometacarpi, | complete left femur, proximal end of | nght femur, proximal ends of 1 right and 2 left tibiotarsi, distal ends of 5 right and 4 left tibiotarsi, 1 complete right and 1 complete left tarsometatarsus, proximal ends of 3 left tarsometatarsi, distal ends of 2 right and 2 left tarsometatarsi. ROM 16807-16878, 19536-19545. The 82 specimens represent a minimum of six individuals. Characters Coragyps atratus and Cathartes aura (Linnaeus) are easily separated by many characters. Although the material listed above closely resembles Coragyps atratus, including C. a. brasiliensis, it differs in numerous small, but consistent, ways from that species. Measurements of the fossil material all fall within the range of living C. atratus and are much smaller than corresponding measurements for C. occidentalis (Howard, 1968). Remarks Despite strong resemblance to C. atratus, the fossil material is definitely distinctive, and may represent a new species, or at least a temporal subspecies. This being the case, earlier suggestions (Fisher, 1944; Howard, 1968) that C. occidentalis may be the direct ancestor to C. atratus must be re-examined. Cathartes Mlliger Cathartes aura (Linnaeus) Turkey Vulture Material One premaxillary, 2 left quadrates, 3 left scapulae, 2 complete right and 1 complete left coracoid, humeral end of 1 left coracoid, proximal end of | right humerus, distal ends of 2 left humeri, proximal ends of 2 left radii, distal end of 1 right radius, distal end of | left carpometacarpus, distal ends of 2 right and 3 left tibiotarsi, 1 complete right and 1 complete left tarsometatarsus, proximal ends of 1 right and 1 left tarsometatarsus. ROM 16879-16904. The 26 specimens represent a minimum of three individuals. Remarks Cathartes aura is generally distributed throughout South America. 73 Suborder Accipitres Vieillot Family Accipitridae (Vieillot) Subfamily Buteoninae (Vigor) Miraquila gen. nov. Diagnosis Tarsometatarsus characterized by having (1) inner calcaneal ridge of hypotarsus projecting posteriad at 90° to vertical axis of shaft (projects posteriad at approximately 80° to vertical axis of shaft in Heterospizias Sharp and Buteogallus Lesson; approximately 60°—-70° in Buteo Lacepede, Geranoaetus Kaup, Harpyhaliaetus Lafresnaye, Spizaetus Vieillot, and Aquila Brison); (2) intercotylar area distinctive ridge (low elevated area in Buteo, Geranoaetus, Harpyhaliaetus, Spizaetus, and Aquila; moderate ridge in Heterospizias and Buteogallus); (3) edge of shaft external to outer calcaneal ridge of hypotarsus distinct long ridge flaring externally (long ridge but not as large in Buteogallus; short, protuding externally only slightly in Buteo, Geranoaetus, Harpyhaliaetus, Heterospizias, Spizaetus, and Aquila); (4) anterior metatarsal groove very deep (shallow in Harpyhaliaetus; moderately deep in Heterospizias, Buteogallus, Geranoaetus and Buteo; deep in Spizaetus and Aquila); (5) internal anterior metatarsal ridge well developed, extending distad for more than third of shaft (moderately developed in Buteo; slightly developed in Geranoaetus and Buteogallus; poorly developed in Harpyhaliaetus, Heterospizias, Spizaetus and Aquila); (6) attachment of tibialis anticus separated from external anterior metatarsal ridge by deep groove (groove slightly developed in Buteo and Buteogallus; groove not distinct in Geranoaetus, Harpyhaliaetus, Heterospizias, Spizaetus, or Aquila); (7) middle trochlea long anteriorly, turned only slightly externally in anterior view (long anteriorly, turned moderately externally in Geranoaetus and Buteogallus, and very strongly externally in Aquila; long anteriorly, turned slightly externally in Heterospizias and Spizaetus; short, turned significantly externally in Buteo and Harpyhaliaetus); (8) internal trochlea rotated posteriad such that posteromedial edge is at same level as posterolateral edge of middle trochlea (similar in Buteo, Geranoaetus, Heterospizias, Buteogallus, Spizaetus, and Aquila; not rotated posteriad in Harpyhaliaetus); (9) internal trochlea not projecting distad beyond middle trochlea and not rotated externally in anterior view (similar in Harpyhaliaetus and Aquila; projecting distad beyond middle trochlea and not rotated in Heterospizias, Aquila, and Geranoaetus; projecting or not projecting distad beyond middle trochlea and not rotated in Buteogallus; projecting distad beyond middle trochlea and rotated significantly externally in Buteo); (10) internal trochlea without distinct angular proximolateral projection (similar in Harpyhaliaetus, Heterospizias, Aquila, and Spizaetus; small projection in Buteogallus; prominent in Buteo and Geranoaetus). Type Species Miraquila terrestris sp. nov. 74 Etymology From Latin, mirus, amazing or astonishing, and aquila, feminine, eagle. Miraquila terrestris sp. nov. Fig. 23A, 24C; 27C; D, E Holotype Proximal end of left tarsometatarsus. ROM 13025. Paratypes Proximal end of | right tarsometatarsus, distal ends of 2 right tarsometatarsi. ROM 13026-13028. Referred Material One complete left coracoid, 1 complete left carpometacarpus, proximal ends of 2 left tibiotarsi. ROM 13029-13032. The 8 specimens, including holotype and paratypes, represent a minimum of 2 individuals. Diagnosis As for genus. Tarsometatarsus agrees with that of Geranoaetus melanoleucus (Vieillot) and Harpyhaliaetus (Urubitornis) solitaria (Tschudi) in general form and proportions, and is similar in width but significantly longer than that of the latter. For measurements see Table 6. The only measurements available from the holotype is the proximal width, 20.8. Characters Coracoid with (1) coracohumeral surface twice as broad anteriorly as distal to bicipital tuberosity, sloping moderately externally (approximately same _ size throughout, sloping steeply externally in Geranoaetus, Buteo, Heterospizias, and Buteogallus; slightly broader anteriorly, sloping moderately steeply externally in Harpyhaliaetus, Spizaetus, and Aquila); (2) attachment of M. coracobrachialis anterior and Lig. humero-coracoideum anterius superius elevated (former not elevated and latter elevated only along posterior margin, if at all, in Geranoaetus, Buteo, Heterospizias, Buteogallus, Harpyhaliaetus, and Aquila; former not elevated but latter elevated in Spizaetus); (3) angle formed between internal side of shaft and anterior edge of glenoid facet approximately 155° (approximately 160° in Geranoaetus and Buteo; approximately 165° in Heterospizias, Buteogallus, Harpyhaliaetus, Spizaetus, and Aquila); (4) shaft very thick in distal view ventral to pneumatic foramen (slender in Geranoaetus, Buteo, Heterospizias, Buteogallus, Harpyhaliaetus, and Spizaetus; moderately thick in Aquila). Carpometacarpus with (1) process of metacarpal I low (i.e., short anteropos- teriorly) not turned proximad, and of approximately equal length (i.e., proximodis- tally) anteriorly and posteriorly (high, directed proximad, much longer anteriorly than posteriorly in Geranoaetus, Harpyhaliaetus, and Aquila; varying from high to low, 75 directed proximad, of equal length anteriorly and posteriorly in Buteo; high, directed slightly proximad, and of equal length anteriorly and posteriorly in Heterospizias and Buteogallus; low, directed proximad, and of equal length anteriorly and posteriorly in Spizaetus); (2) length of proximal fusion of metacarpal II and metacarpal III long (similar in Harpyhaliaetus, Heterospizias, and Buteogallus; of moderate length in Buteo, Geranoaetus, Spizaetus, and Aquila); (3) external edge of metacarpal III extending as slight ridge to base of external rim of carpal trochlea (similar in Harpyhaliaetus; extends external to midline of metacarpal II in Heterospizias and Buteogallus; extends to approximately midline of metacarpal II in Buteo, Spizaetus, and Aquila; lies directly in line with internal rim of carpal trochlea in Geranoaetus); (4) surface of carpal trochlea much wider anteriorly than posteriorly (similar in Buteo, Geranoaetus, Heterospizias, and Buteogallus; of approximately equal width throughout in Harpyhaliaetus, Spizaetus, and Aquila); (5) tendinal groove restricted to external side of metacarpal II (proximal half of groove lies on anterior face and distal half of groove on external side of metacarpal II in Buteo, Geranoaetus, Harpyhaliaetus, Heterospizias, Buteogallus, and Aquila; proximal half lies on corner of shaft in Spizaetus). Tibiotarsus with (1) rotular crest low for length (internal end high, remainder low in Geranoaetus and Buteo; moderately high for length in Heterospizias and Buteogallus; high for length in Harpyhaliaetus, Spizaetus, and Aquila); (2) internal edge of proximal end sloping slightly posteriad in internal view (moderate slope in Geranoaetus, Buteo, Buteogallus, Harpyhaliaetus, and Spizaetus; steep slope in Heterospizias and Aquila). Etymology Latin, terrestris, of the earth. Remarks Miraquila is a very distinctive genus, and perhaps should be placed in a separate subfamily. In such characters as the sharp angle of the coracoid shaft, the tendinal groove of carpometacarpus restricted to external side of metacarpal II, and the middle trochlea of tarsometatarsus lying almost parallel to midline of shaft, this genus resembles no known living accipitrid genus. Although Oroaetus isidori was not available for comparison its shorter tarsus and apparent relationship to the Spizaetus group (Brown and Amadon, 1968) would seem to indicate that Miraquila is different from that genus. Harpia is so distinctive that it is not necessary to draw a detailed comparison with that genus. The tarsometatarsus of M. terrestris differs from that of Titanohierax gloveralleni Wetmore (1937) from the late Pleistocene of the Bahamas by having the middle trochlea turned externally only slightly. It differs from that of Aquila borrasi Arredondo (1970) from the late Pleistocene of Cuba by having the tubercle for tibialis anticus positioned farther distally and more externally. Both of these characters are readily seen in published illustrations of the types of these two large extinct eagles, and there are undoubtedly many other characters that cannot be detected from the illustrations. The tarsometatarsus of Wetmoregyps daggetti (L. Miller) (1915) from the late Pleistocene of California and Nuevo Leon, Mexico, is much longer than that of M. 76 terrestris, and according to the original description has reduced trochleae which appear from the published photograph to lie parallel to the long axis of the shaft. In M. terrestris the trochleae are not reduced and the middle trochlea lies at a very slight angle to the long axis of the shaft. The diagnostic character of Calohierax quadratus Wetmore (1937) from the late Pleistocene of the Bahamas is the quadrangular shape of the internal trochlea, resulting from the lack of a large posteromedially projecting wing. A _ large posteromedially projecting wing is present in M. terrestris. Lagopterus minutus Moreno and Mercerat (1891) was described from the late Pleistocene of Argentina on the basis of a humerus. The large size difference between L. minutus and M. terrestris indicates that they are definitely not the same species. Although no humerus of M. terrestris is available for direct comparison with that of L. minutus, 1 do not believe the two species are congeneric because from the published illustration (Moreno and Mercerat, 1891) it appears that Lagopterus resembles Buteo very much, while the distinctive characters of the available specimens of Miraquila would lead me to expect a humerus with similarly distinctive characters. Although there is still some question as to the generic position of Harpyhaliaetus (Urubitornis) solitarius (Brown and Amadon, 1968; Wetmore, 1965), the species is used here as representative of Harpyhaliaetus until an osteological comparison can be made between H. solitarius and H. coronatus. Skeletons available for comparison included four of Geranoaetus, seven of Heterospizias , one of Harpyhaliaetus, five of Buteogallus, two of Aquila, two of Spizaetus, and over 50 of Buteo. Amplibuteo gen. nov. Diagnosis Tarsometatarsus is characterized by having (1) inner calcaneal ridge of hypotarsus projecting posteriad at approximately 60°—70° to vertical axis of shaft; (2) intercotylar area elevated slightly; (3) edge of shaft external to outer calcaneal ridge of hypotarsus without ridge; (4) anterior metatarsal groove moderately deep; (5) internal anterior metatarsal ridge only slightly developed; (6) attachment of tibialis anticus not separated from external anterior metatarsal ridge by groove; (7) middle trochlea short, turned significantly externally in anterior view; (8) internal trochlea rotated posteriad such that posteromedial edge is at same level as posterolateral edge of middle trochlea; (9) internal trochlea projecting distad beyond middle trochlea and rotated significantly externally in anterior view; (10) internal trochlea with distinct angular proximolateral projection. The numbers of the osteological characters correspond to those of the characters used in the diagnosis of Miraquila. Type Species Amplibuteo hibbardi sp. nov. Etymology From Latin, amplus, large, and buteo, masculine, buzzard 77 0°89 ‘aou ‘ds 39 ‘ua3 S1AJSAAAI] pjinbv.py ¢ 09 6°61 D1AD]1OS snjapijpyxdiv yy Be 9'el 0 Tcl € 8 VC VY SC SVC € t Oe 8 LEV Se G €°08 8 18-882 c i en EVSs Uh € 0'v~ O LCV CC ‘aou ‘ds }0 ‘uaa Ip4Dgqiy oainqydup V8 C OSI E Sle ceil (6 SEG le Com (CC Cc LAS Secs 11s sdoag Ie] Pleyel v 0'@ [ee C36 Ol 74 C8 TOT Dec. VA 074 Bec LwvCac. 1G v O's LASSaer SV, yO)! 8 LI-0 91 JUDN0Y snonajouvjaul SnjapouvdsayH (6) ado 10084 [BUuIaS jo y3uaT eys jo uidoq SBI] yidoq [PWIIXOlg 1008] Jetndesg 0} peoy usu] proovlog UIPIM [BUIIXOIg 1YysIOH [PWIxXO1g eindess ‘(uu ut) ‘AOU ‘ds ja ‘Uad S2.4jSa413} DJINbD.NPF JO SNS.1V}eJIWIOS.1e} PUB ‘SNsIeJOIGH ‘sndzedej}IUIOd.Ied ‘pl0dB.109 IY} JO pue ‘$s1upjy0s snjanyvytduvyy ‘(e{ndeds Suipnydxe) sdaeg ey, Viele] IY) WOIy sNInajounjau *D ‘snonajounjawm Snjavounsayy yuIddyY ‘*Aou ‘dS 4a “UIs Ipsvqgqiy oajnqyduy Jo sns.iejeyawi0s.1e} pues ‘snsiejOIqH ‘ANwWay ‘sndsvdejaWIOd sed ‘eUlN ‘snipe’ ‘sNAVUINY ‘p10ode.109 ‘epndeds ay} JO SyUIWIaAINSeIy, «9 FIqQUL 78 ‘0g osed uo panunuod 8 681 0'6L1 Vicor FS) Cisne (6 etl CS rI-I vl (6 6SE Estar ce O'S6l LS SUE L87¢ COG V8 v vy LOI O VETS cSt 0'8SI € S9I-L trl v 86 COIS 6G v C ve ESC-U CC v 6 LC Vi6c1 SC v 6 cl LiOS Isl OC rc (@) ado ado ysuay eu) UIPIAA [easiq UIPIM [PUITXO1g ysuay snipey YeYS JO YIPIM Sea] UIPIAN [easiq WPI [PWIXO1g ysuay snisuiny 79 I I v u XIWIO4 = ae Pe W [RIsIq 1°8 ¢'6 8'L-$'9 YO Jo yysueT I I Cc I 14 u yeus = ae v8 = G W JO UPI Rae Ee 88-08 v9 S9OFLS re (@) |SvoT I I Cc t v u = = Ore! L’8 Bye W UPIM 0°OI 16 O°CI E Ones Con0e (6) [BWITXOld I I I € v u [ [edievsrja/y — —~ = L'61 707 W ysnory | OT 617 C 87 S.0Gec- oI VIC9 8I rs (@) 1YysI9H I | I 74 u aa = a 8 08 W 1°98 8° ¢8 8 Ol 898-I PL rs (6) wsuay sndieorjawiodies I v U yes a Sol W JO YIPIM v8 2 re (6) $SEO'] I 4 v v u — r'si Sel O'rI W yidaq eS 1 6I-9'L1 Or CleS Cl ORs el rs (6) [BISIq I I v U = = jee en W UIPIM (GLOXG GSC OSlss Oil rs (@) [EwIxord ‘aou ‘ds ja ‘ua3s DIADIIJOS ‘aou ‘ds ja ‘uaa sdoag Ie] Pree }UIDNY S1AJS IAAI] snjapypyAdiv yy 1pavqqiy snonajouv] au SnjapouvsaH pjinbowpy oainqyudup 80 ‘7g oded uo ponunuog Cc I € 0c ae VGC (OG 0°0¢ Ik Go) y 901 ¢ cav Savi Ono L MELE Orcas 1 OReEC VEG 8 91 Cc 88 CiOnG 8 c L6l Gao) s)| v 06! 60c-S LI G Je I SOI-€ 06 574 8 Ol c II-001 v C8 Wy om 0) fs v OVE (0) /Elligte “SI V 6°SI [EL ONAE 4! 74 Ze 6 ChI—-8 87l v 16 vols v © ALG SC Cmeaoll v 96! SiGe ss) v £66 6 COl=2 6 ado rs [@) a,Apuod [PUIO}U] jo uWdaq YeYs JO YIPIM Sea] UIPIAA [eisiq YIPIM [BUITXOIg yysuoT SNSIPIOIQL], yes JO WPIM Seaq UPI [BISIG YIPIMA [BUIIXOIg yysuaq] Inuld,j &] Cc I 8 (S v u BITYOOI 9 — tL Eg GES W S[PPIIN o:97t 9 9 18-69 LeSetes 9°S-6'P ao JO YIPIM I I G v v U yeus — = 6 Ol SIG 89 W JO UPI 88 6°6 EC biel Ol Cosa 0 5 ERS) ao 1SeaT] I 9 (6 v u — ECE 0°0¢ 6°81 W WIPIM OEEC CiGGeC SC 9'0C-€ GI COC+8eE1 rs (@) [eisiq C I is S$ v u 6 07 = [ERIE vst Vy fel W UPI 60-8 0¢ LOC OPSGaG CC 9'6I-€ 91 8° 8I-c 91 rs [@) [eWITxXO1g I 3 I v u (poyeuns| ) er evel — 6 COI W Srl 8 Pl SONS NS Ge € 001 LALO Tee 86 ro (8) ysueT SNSIEILJOWIOSIV |, ‘aou ‘ds ja ‘uas DIADII]OS ‘aou ‘ds ja ‘ua3 sdaag Ie] eleye JUd99y SIS AAAI] snjapiyjoyAdivy ipavqqiy snonajouvjau Snjavouv41dayH) ppinboalpy oainq dup 82 Amplibuteo hibbardi sp. nov. Pig: 22D,,23B. 244A; (BD) E, 25A-E, 260A, 27A, B, 28A Holotype Complete left tarsometatarsus (inner calcaneal ridge and external trochlea broken). ROM 16905. Paratypes One complete right and 1 complete left tarsometatarsus, shaft of 1 left tarsometatarsus, proximal ends of | right and 1 left tarsometatarsus, distal ends of 6 right and 2 left tarsometatarsi. ROM 16906-16918. Referred Material Two rostra, 1 right ramus, | right and 2 left scapulae, 2 complete left coracoids, humeral end of 1 right coracoid, 1 complete right humerus, proximal end and shaft of 1 right humerus, proximal end of 1 left humerus, distal end of 1 left humerus, proximal end of | right ulna, distal ends of 2 right and 2 left ulnae, proximal ends of 2 right and 2 left radii, distal end of 1 left radius, 1 almost complete right carpometacarpus, proximal end of 1 right carpometacarpus, shaft of 1 right carpometacarpus, distal end of | right carpometacarpus, | complete right femur, proximal ends of 2 left tibiotarsi, distal ends of 2 right and 6 left tibiotarsi. ROM 16919-16956. The 52 elements, including holotype and paratypes, represent a minimum of 7 individuals. Diagnosis As for genus. For measurements see Table 6. Measurements of the holotype are as follows: length, 134.1; proximal width, 25.0; distal width, 27.1; least width of shaft, Met: Characters From the diagnosis of the tarsometatarsus it is apparent that Amplibuteo is very closely related to Buteo and Geranoaetus , and is quite distinct from other living South American accipitrid genera, Miraquila, Titanohierax, Wetmoregyps, Calohierax, and Lagopterus. Accordingly, the referred material is compared with B. jamaicensis and G. melanoleucus to facilitate future comparison. Because of the large size of Amplibuteo, in comparison with B. jamaicensis and G. melanoleucus, all comments as to size must be considered relative. Amplibuteo hibbardi is characterized by having rostrum with (1) external nares long and narrow (shorter and wider in B. jamaicensis and G. melanoleucus). Ramus with (1) very slight curvature in lateral view (much greater curvature in B. jamaicensis and G. melanoleucus); (2) post-articular process small (more prominent in B. jamaicensis and G. melanoleucus); (3) ventral corner of articular approximately 90° angle (less acute an angle in B. jamaicensis and G. melanoleucus), (4) medial third of internal articular process projecting anteriad (projects little, if any, anteriad in B. jamaicensis, and G. melanoleucus). §3 Scapula with (1) dorsointernal corner of acromion process well developed, giving straight dorsal edge (angular mediad projection present, giving sharp internal slope in B. jamaicensis; corner well rounded, giving moderate internal slope in G. melanoleucus); (2) dorsoexternal corner of acromion process prominent (similar in B. jamaicensis; less developed in G. melanoleucus); (3) glenoid facet wide anteriorly and short, with anteroventral corner well developed (similar in shape with moderately developed anteroventral corner in B. jamaicensis, more elliptical with poorly developed anteroventral corner in G. melanoleucus). Coracoid with (1) coracohumeral surface twice as wide dorsal as ventral to bicipital tuberosity, short, with moderate external slope (narrow throughout, longer, with steep external slope in B. jamaicensis and G. melanoleucus); (2) attachments of M. coracobrachialis anterior and Lig. humero-coracoideum anterius superius promi- nently elevated (former not elevated, and latter elevated only at posterior edge, if at all, in B. jamaicensis and G. melanoleucus); (3) ventral extension of glenoid facet large (similar in B. jamaicensis, small in G. melanoleucus); (4) depression distal to intermuscular line and dorsal to external posterior sternal facet deep (slight to moderate depression in B. jamaicensis, slight to no depression in G. melanoleucus); (5) distal half of internal edge of sternal facet projecting significantly internally, giving deep sterno-coracoidal impression (similar in B. jamaicensis; projects internally only slightly, giving shallow or no sterno-coracoidal impression in G. melanoleucus); (6) prominent swelling dorsal to proximal end of internal sternal facet (slightly developed in B. jamaicensis; prominent in G. melanoleucus). Humerus with (1) internal tuberosity moderately produced and rounded (similar in B. jamaicensis, more produced and less rounded in G. melanoleucus); (2) edge of bicipital crest merging gradually with shaft posteriorly (end of bicipital crest well marked in B. jamaicensis and G. melanoleucus); (3) base of deltoid crest very thick distally (thin in B. jamaicensis and G. melanoleucus); (4) area between shaft and distal edge of deltoid crest essentially flat (moderately concave in B. jamaicensis; very concave in G. melanoleucus); (5) impression of M. brachialis anticus deep, giving high narrow ridge leading proximad from attachment of anterior articular ligament (shallow, with only slight ridge in B. jamaicensis; moderately deep, with moderate ridge in G. melanoleucus; (6) entepicondyle well rounded in medial view, not projecting distad beyond internal condyle (rounded in medial view and projecting distad significantly beyond internal condyle in B. jamaicensis; subangular in medial view, not projecting distad beyond internal condyle in G. melanoleucus). Ulna with (1) olecranon very rounded proximaily on anconal side (not rounded in B. jamaicensis or G. melanoleucus); (2) internal cotyla sloping steeply externally (sloping moderately in B. jamaicensis, sloping slightly in G. melanoleucus); (3) external cotyla with moderate palmar projection (much larger projection in B. jJamaicensis and G. melanoleucus); (4) external condyle short and well rounded (longer and less rounded in B. jamaicensis and G. melancleucus); (5) carpal tuberosity short, projecting sharply internally (much longer and projecting less internally in B. jamaicensis and G. melanoleucus). Radius with (1) ligamental papilla short, very prominent (long, not prominent in B. jamaicensis; very short, not prominent in G. melanoleucus); (2) carpal facet well rounded internally (only slightly rounded in B. jamaicensis and G. melanoleucus); (3) external corner not extending distad beyond carpal facet (extending significantly beyond carpal facet in B. jamaicensis and G. melanoleucus). 84 Carpometacarpus with (1) end of process of metacarpal I subangular to rounded in external view (almost straight in B. jamaicensis and G. melanoleucus); (2) internal ligamental fossa shallow (deep in B. jamaicensis, shallow to moderate depth in G. melanoleucus); (3) internal side of shaft of metacarpal II narrowing distally to form ridge leading to base of distal metacarpal symphysis (internal side of shaft narrowing slightly to form small ridge in B. jamaicensis, no narrowing of shaft and only slight indication of ridge in G. melanoleucus). Femur with (1) abundant pneumaticity proximally (similar in B. jamaicensis, limited pneumaticity in G. melanoleucus); (2) proximal edge of trochanter moderately rounded in lateral view (similar in B. jamaicensis, less rounded in G. melanoleucus); (3) neck short and head projecting mediad in B. jamaicensis; neck long and head projecting dorsomediad in G. melanoleucus); (4) iliac facet with well developed posteroexternal corner (corner moderately developed in B. jamaicensis, area well rounded in G. melano!leucus); (5) attachment of M. gluteus profundus short (similar in B. jamaicensis, long in G. melanoleucus); (6) proximal end of lateral surface curving sharply internally in anterior view (curves moderately internally in B. jamaicensis , curves only slightly internally in G. melanoleucus); (7) proximal end of fibular condyle small, directed externally (larger and directed more proximad in B. Jamaicensis and G. melanoleucus); (8) attachment of M. gastrocnemius, pars externa, shallow (of moderate depth in B. jamaicensis, deep in G. melanoleucus); (9) lateral surface of external condyle at large angle to vertical axis of shaft (at small angle to vertical axis of shaft in B. jamaicensis and G. melanoleucus); (10) shaft moderately convex anteriorly in lateral view (more convex in B. jamaicensis and G. melanoleucus). Tibiotarsus with (1) rotular crest slightly elevated internally (elevated much more in B. jamaicensis and G. melanoleucus); (2) internal articular surface slightly convex and sloping slightly posteriad (quite convex and sloping steeply posteriad in B. jJamaicensis, of moderate convexity and sloping steeply posteriad in G. melanoleucus); (3) anterior notch between internal articular surface and extemal articular surface well marked (only slight indentation in B. jamaicensis and G. melanoleucus); (4) external condyle short proximodistad and long anteroposteriad, with small anterior projection at anteroproximal end (more rounded with moderate projection in B. jamaicensis, rounded with very prominent projection in G. melanoleucus); (5) anterior portion of internal condyle long (similar in B. jJamaicensis, short in G. melanoleucus); (6) opening in tendinal groove distal to supratendinal bridge does not extend distal to proximal edge of internal condyle (similar in B. jamaicensis, does extend distal to internal condyle in G. melanoleucus); (7) internal condyle only slightly undercut proximally (similar in B. jamaicensis, moderately undercut in G. melanoleucus). Etymology The species is named for the late Dr. Claude W. Hibbard in recognition of his many outstanding contributions to vertebrate palaeontology. Position of Morphnus woodwardi Through the courtesy of Dr. Hildegarde Howard and the Natural History Museum of Los Angeles County, I have been able to compare a representative coracoid, 8&5 humerus, femur, and tarsometatarsus of Morphnus woodwardi L. Miller (1911) from Rancho la Brea with corresponding elements of Amplibuteo hibbardi. The two species are essentially identical in size, and woodwardi agrees with A. hibbardi in having those characters of Amplibuteo listed for the tarsometatarsus that serve to distinguish it from the tarsometatarsi of other accipitrid genera. A body skeleton of Morphnus guianensis has recently become available from the United States National Museum and a comparison of the coracoid, humerus, and femur of that specimen with woodwardi quickly reveals that woodwardi is not of that genus. From the specimen available it appears, osteologically, Morphnus is more similar to Harpia than any other genus. The very close resemblance of woodwardi to A. hibbardi in all elements examined indicates that the two species should be considered congeneric, and that woodwardi be removed from Morphnus and placed in Amplibuteo. Amplibuteo hibbardi differs from A. woodwardi by having coracoid with (1) coracohumeral surface longer; (2) posterior edge of glenoid facet projecting much farther from shaft and more rounded; (3) attachment of M. coracobrachialis anterior and Lig. humero-coracoideum anterius superius more elevated and distinctly marked; (4) ventral end of furcular facet projecting more externally; (5) head more rounded; (6) anterior internal sternal facet longer, ending less abruptly; (7) area dorsal to distal external sternal facet more concave. Humerus with (1) internal tuberosity more prominent; (2) concavity leading distad from base of median crest deep, not shallow; (3) capital shaft ridge more prominent, lying more internally; (4) external tuberosity more prominent, set off from head by slight groove; (5) bicipital crest longer; (6) attachment of Lig. humero-coracoideum anterius inferius larger and deeper; (6) shaft more curved in anconal view; (7) distal end flexed less palmarly; (8) ectepicondylar process slightly more produced; (9) intercondylar furrow deeper, particularly in palmar view. Femur with (1) neck longer; (2) shaft less stout, more curved anteroposteriad; (3) internal condyle less rounded in medial view, and proximal ridge elevated more above shaft and joining shaft more abruptly; (4) rotular groove narrower and deeper; (5) proximal end of external condyle more rounded, and turned more externally; (6) fibular condyle much more prominent, projecting more sharply externally; (7) attachment of M. gastrocnemius, pars externa, larger in area, but shallower; (8) proximal end of fibular condyle larger. Tarsometatarsus with (1) inner calcaneal ridge longer through base; (2) intercalcaneal furrow narrower; (3) tubercle for tibialis anticus larger, positioned more externally; (4) shaft curving externally less at level of metatarsal facet, giving larger external angle between line drawn along most distal point of trochlea and perpendicular axis of shaft; (5) internal trochlea narrower through base, thus having more distinct angular proximolateral projection; (6) external trochlea rotated less posteriad, extending farther distad. Although only one specimen of each element of A. woodwardi was available I believe most of the characters will hold if a larger series is studied. It should be stressed that A. woodwardi differs from B. jamaicensis and G. melanoleucus as much as, if not more than, A. hibbardi in those characters listed in the diagnoses of A. hibbardi. Although Buteo and Geranoaetus are closely related as revealed in the preceding diagnoses, I believe Geranoaetus is a valid genus and should be maintained (see also Amadon, 1963; Brown and Amadon, 1968). 86 Geranoaetus Kaup Geranoaetus melanoleucus (Vieillot) Black-chested Buzzard-Eagle Material Two complete right coracoids, proximal ends of 2 right and 2 left humeri, distal end of 1 right humerus, distal ends of 3 right and 1 left ulna, proximal end of 1 left radius, 1 complete right carpometacarpus, proximal ends of 1 right and 1 left carpometacarpus, 1 complete right and | complete left femur, proximal ends of 2 right and 1 left femur, distal end of 1 right femur, proximal end of 1 left tibiotarsus, distal ends of | right and 1 left tibiotarsus, 2 complete left tarsometatarsi, proximal ends of 1 right and 2 left tarsometatarsi, distal ends of 2 right tarsometatarsi, shaft of 1 right tarsometatarsus. ROM 16957-16987, 19548-19549. The 32 specimens represent a minimum of 4 individuals. Remarks Widely distributed in western South America, G. melanoleucus is common along coastal hills and western slopes of the Peruvian Andes. Buteo Lacepede Buteo polyosoma (Quoy & Gaimard) Red-backed Hawk Material Two left articulars, | right scapula, 1 complete right and 4 complete left coracoids, humeral end of | right coracoid, proximal end of | right humerus, distal end of 1 right humerus, distal ends of 2 right ulnae, proximal end of 1 right radius, distal end of 1 left radius, proximal ends of 1 right and 1 left carpometacarpus, proximal ends of 2 left femora, proximal ends of 1 right and 1 left tibiotarsus, distal end of 1 right tibiotarsus, 2 complete right tarsometatarsi, proximal ends of 1 right and 1 left tarsometatarsus, distal ends of 2 right and 3 left tarsometatarsi. ROM 16988-17017, 19523. The 31 specimens represent a minimum of 4 individuals. Remarks Buteo polyosoma occurs all along the west coast of South America, and is very common in northwestern Peru. Buteo sp. 1 Material Complete right ulna, proximal ends of 2 left ulnae, 1 complete left carpometacarpus. ROM 17018-17021. The 4 specimens represent a minimum of 2 individuals. 87 Characters Size larger than Buteo nitidus (Latham) and overlapping that of small B. polyosoma. Agrees with Buteo and differs from related genera by having those characters of carpometacarpus listed for Buteo in description of Miraquila terrestris. Carpometacarpus with (1) internal side of metacarpal II deeply excavated with ridge separating internal ligamental fossa from anterior portion (slightly excavated in B. polyosoma; moderately excavated in B. swainsoni Bonaparte; similar but lacking ridge in B. platypterus (Vieillot); excavated more anteriorly in B. nitidus and Parabuteo unicinctus (Temminck)); (2) external rim of carpal trochlea parallel with shaft and internal rim of carpal trochlea almost straight (external rim at angle to shaft and internal rim moderately to strongly curved in B. polyosoma; similar in B. swainsoni; external similar but internal at greater angle to shaft in B. platypterus and P. unicinctus; external rim similar, internal rim curved in B. nitidus); (3) process of metacarpal I very short, rounded on proximal corner (long, not rounded in B. polyosoma, B. nitidus, and P. unicinctus; moderately long, not rounded in B. swainsont, slightly rounded in B. platypterus); (4) external edge of pollical facet long (short in B. polyosoma, B. swainsoni, B. platypterus, and Parabuteo unicinctus; moderately long in B. nitidus). Length, 51.4; proximal width, 5.9; height through metarcarpal I, 13.5; least width of shaft, 4.4. UlIna with (1) anconal side of olecranon continuing curve of shaft (reflected anconally, interrupting curve of shaft in B. polyosoma, B. swainsoni, B. platypterus, B. nitidus, and P.. unicinctus); (2) olecranon long and broad in anconal view (long and narrow in B. polyosoma, B. swainsoni, B. nitidus, and P. unicinctus; short and broad in B. platypterus); (3) external cotyla greatly produced (similar in B. polyosoma and B. nitidus; less produced in B. swainsoni and B. platypterus; more produced in P. unicinctus ); (4) carpal turberosity short, well set off from shaft (long, not well set off from shaft in B. polyosoma, B. platypterus, B. nitidus; moderately long and moderately set off from shaft in B. swainsoni and P. unicinctus). Length, 110.1 (n = 1); proximal width, 11.5-12.4; mean, 11.9 (n = 3); distal depth, 9.7 (n = 1); least shaft width, 5.4 (n = 1). Remarks Although distinct from the species on hand, the material described above is referred only to genus because of the lack of comparative material of B. brachyurus and B. leucorrhous. Although the latter species is much smaller than the fossil material, B. brachyurus is in the right size range and until the fossil specimens can be compared with specimens of that species the question of whether the fossil material represents a living or extinct species must remain unanswered. Buteo sp. 2 Material Proximal end of 1 left ulna, distal end of 1 left ulna, proximal ends of 2 right carpometacarpi, distal ends of 4 right tibiotarsi, proximal end of 1 might tarsometatarsus, distal end of 1 left tarsometatarsus. ROM 17022-17031. The 10 specimens represent a minimum of 4 individuals. 88 Characters Agrees with Buteo and differs from related genera by having those characters for carpometacarpus, tibiotarsus, and tarsometatarsus listed for Buteo in description of Miraquila terrestris. Buteo sp. 2 is similar in size to B. albicaudatus, smaller than B. peocilochrous, and larger than B. polyosoma, B. swainsoni, and B. nitidus. Buteo sp. 2 differs from B. albicaudatus by having ulna with (1) olecranon narrower, More rounded in internal view; (2) internal cotyla more concave; (3) protuberance between internal and external cotylae prominent (not present in B. albicaudatus); (4) external condyle smaller; (5) carpal tuberosity positioned farther proximad. Proximal width, 12.6 (n = 1); distal depth, 10.3 (n = 1). Carpometacarpus with.(1) process of metacarpal I higher, but shorter; (2) internal rim of carpal trochlea much shorter; (3) carpal trochlea narrowing distad more sharply, giving more curved external rim of carpal trochlea; (4) protuberance at distal end of external carpal trochlea much smaller; (5) external rim of carpal trochlea angular at point of attachment of M. flexor carpi ulnaris brevis, in external view (rounded in B. albicaudatus). Proximal width, 6.0 (n = 2); height through metacarpal I, 14.5-14.8 (n = 2); least width of shaft, 4.5 (n = 1). Tibiotarsus with (1) tendinal groove passing under tendinal bridge more proximad; (2) internal condyle longer anteroposteriad, shorter proximodistad, with deeper depression at posteroproximal end; (3) external condyle much smaller; (4) shaft proximal to tendinal bridge and medial to tendinal groove depressed, and lateral to tendinal groove elevated. Distal width, 12.0-12.8 (n = 4); least width of shaft, 6.3—6.4 (n = 2); depth of internal condyle, 8.2—8.4; mean, 8.5 (n = 2). Tarsometatarsus with (1) anteroproximal edge of shaft straighter; (2) proximal portion of anterior metatarsal groove narrower and shallower; (3) area between hypotarsi less excavated; (4) posterolateral corner of proximal end rounded, lacking protuberance; (5) external condyle projecting distad more than middle trochlea (projects distad less than middle trochlea in B. albicaudatus); (6) external trochlea projects less posteriad; (7) protuberance on posteroproximal corner of internal trochlea larger; (8) middie trochlea shorter, with shallower trochlear groove. Proximal width, 13.3 (n = 1); distal width, 14.5 (n = 1); depth of middle trochlea, 4.1 (n = 1). Remarks Buteo sp. 2 probably represents a new species, but until specimens of B. ventralis, B. albonotatus, and B. brachyurus are available for comparison its exact status must remain unknown. Parabuteo Ridgway Parabuteo unicinctus (Temminck) Bay-winged Hawk Material Two right and 1 left scapula, humeral end of 1 right coracoid, distal ends of | right and | left humerus, distal ends of 1 right and 2 left ulnae, proximal ends of 2 right and 8&9 2 left radii, 1 complete right carpometacarpus, proximal end of 1 left carpometacar- pus, proximal end of | left femur, distal ends of 1 right and 2 left femora, proximal ends of 3 right tibiotarsi, distal ends of 4 right and 1 left tibiotarsus, 2 complete left tarsometarsi, distal ends of | right and 1 left tarsometatarsus. ROM 17032-17062. The 31 specimens represents a minimum of 4 individuals. Remarks The proximal end of the left carpometacarpus differs from the other fossil carpometacarpus and the one Recent carpometacarpus available by having internal rim of carpal trochlea longer, extending to level of pollical facet. Without a greater series to determine variation of this character, the specimen is placed with P. unicinctus , as it is essentially identical otherwise. P. unicinctus occurs along the west coast of South America south to Aysen, Chile. Family Falconidae Vigors Subfamily Falconinae (Vigors) Falco Linnaeus Falco peregrinus Tunstall Peregrine Falcon Material One left quadrate, 1 left articular, 3 mandibular symphyses, 2 right scapulae, 1 complete left coracoid, humeral ends of 2 right and 1 left coracoid, 1 complete right humerus, proximal ends of 2 right humeri, 1 complete left ulna, distal ends of 2 left radii, 1 complete right and 2 complete left femora, proximal ends of | right and 1 left femur, | complete right tibiotarsus, proximal ends of 2 right and 1 left tibiotarsus, distal ends of 2 right and 1 left tibiotarsus, 1 complete right and 3 complete left tarsometatarsi, proximal end of 1 right tarsometatarsus. ROM 17063-17093, 19533-19535. The 34 specimens represent a minimum of 3 individuals. Remarks Falco peregrinus occurs on the Peruvian coast as a migrant from both North and southern South America. Falco femoralis Temminck Aplomado Falcon Material Two complete right and 1 complete left coracoid, 1 complete right ulna, distal end of 1 left ulna, distal end of 1 right radius, 1 complete right carpometacarpus. ROM 17094-17100. The 7 specimens represent a minimum of 2 individuals. 90 Remarks Falco femoralis is generally distributed throughout Ecuador and Peru. Falco sparverius Linnaeus American Kestrel Material One premaxillary, 1 right and 2 left scapulae, 2 complete left coracoids, sternal end of 1 left coracoid, 3 complete right and 1 complete left humerus, distal ends of 3 left humeri, 3 complete right and 4 complete left carpometacarpi, 1 proximal and 2 distal ends of left carpometacarpi, 1 complete right and 2 complete left ulnae, proximal ends of 2 right ulnae, distal ends of 2 right ulnae, 2 complete right and 1 complete left femur, distal ends of 1 right and 1 left femur, 1 complete right tibiotarsus, distal ends of 2 right and 3 left tibiotarsi, 3 complete right and 2 complete left tarsometatarsi, proximal ends of 2 right and 1 left tarsometatarsus, distal ends of 2 right and 3 left tarsometatarsi. ROM 17101-17147, 19550-19557. The 55 specimens represent a minimum of 6 individuals. Remarks Falco sparverius is generally and commonly distributed throughout Peru. Falco sp. Material One complete right coracoid, humeral end of 1 left coracoid, distal end of 1 left humerus, proximal end of | right ulna, distal end of | right ulna, proximal end of 1 right femur, distal end of 1 right femur. ROM 17148-17153, 19524. The seven specimens represent a minimum of one individual. Characters Falco sp. is intermediate in size between F.. peregrinus and F. femoralis, and differs from them by having coracoid with (1) head long, with slight curvature to dorsal edge in external view (long, well rounded in F. peregrinus; short, subrounded in F. femoralis); (2) attachment of Lig. humero-coracoideum anterius superius large, with depression lying posterior to it (large, no depression in F. peregrinus; small, no depression in F. femoralis); (3) attachment of M. coracobrachialis anterior small oval (elongated in F. peregrinus, small oval in F. femoralis); (4) coracohumeral surface moderately long (long in F. peregrinus, short in F.. femoralis); (5) brachial tuberosity large, extending farther internally than head (small, flush with head in F. peregrinus, of moderate size, extending slightly farther internally than head in F’. femoralis); (6) glenoid facet wide, widest near ventral end (wide, widest at dorsal end in F. peregrinus; narrow in F. femoralis). Length, 36.8 (n = 1); head to scapular facet, 13.5—13.9 (n = 2); proximal depth, 8.0-8.1 (n = 2); least depth of shaft, 4.1 (n = 1). 9] Humerus fragmentary, with (1) attachment of anterior articular ligament low, wider than long (high, longer than wide in F. peregrinus; high, width about equal to length in F. femoralis); (2) attachment of M. pronater brevis very deep pit (similar in F.. peregrinus; not as deep, longer in F. femoralis); (3) attachment of M. extensor metacarpi radialis large and slightly elevated (large and very elevated in F. peregrinus, small and moderately elevated in F. femoralis). Distal width, 12.9; least width of shaft, 5.9. Ulna with (1) external cotyla large (large in F’. peregrinus, small in F. femoralis); (2) pronounced concavity between tricipital attachment and external cotyla (slight concavity in F’. peregrinus and F. femoralis); (3) carpal tuberosity short and stout (long and more slender in F.. peregrinus and F. femoralis). Proximal width, 10.0; distal depth, 8.4. Femur with (1) attachment of M. gluteus profundus and M. obturator externus very deep (deep in F. peregrinus, moderately deep in F. femoralis); (2) protrusion between neck and trochanter broad (small in F. peregrinus and F. femoralis); (3) anterior intermuscular ridge prominent (similar in F. peregrinus, interrupted proximally in F. femoralis); (4) attachment of M. gastrocnemius, pars externa, large, very shallow (large, deep in F’. peregrinus; small, shallow in F.. femoralis); (5) ridge at base of internal condyle high with internal surface rotated to face distad (moderately high, not rotated in F. peregrinus; high, not rotated in F.. femoralis); (6) internal condyle angular distally in internal view (rounded in F. peregrinus, subrounded in F’. femoralis). Proximal width, 9.0; distal width, 8.8. Remarks The material listed above belongs without question in the genus Falco, and it probably represents a new species. It is not named at this time, however, on the slight chance that it represents F. deiroleucus, specimens of which were not available for comparison. Subfamily Polyborinae Lafresnaye Polyborus Vieillot Polyborus plancus (Miller) Crested Caracara Material One cranium, 4 right and 6 left quadrates, 4 left articulars, 6 mandibular symphyses, 2 premaxillaries, 4 right and 10 left scapulae, 9 complete right and 15 complete left coracoids, humeral ends of 6 right and 4 left coracoids, sternal ends of 4 right and 2 left coracoids, 1 complete right and 1 complete left humerus, proximal ends of 4 right and 3 left humeri, distal ends of 9 right and 2 left humeri, 5 complete right ulnae, proximal ends of 7 right and 2 left ulnae, distal ends of 15 right and 6 left ulnae, proximal ends of 3 right and 1 left radius, distal ends of 7 right and 6 left radii, 6 complete right and 6 complete left carpometacarpi, proximal ends of 6 right and 3 left carpometacarpi, distal ends of 4 right and 4 left carpometacarpi, 2 complete right and 3 complete left femora, proximal ends of 2 right and 4 left femora, distal ends of 3 92 right femora, 1 complete right tibiotarsus, proximal ends of 2 right and 3 left tibiotarsi, distal ends of 16 right and 14 left tibiotarsi, 8 complete right and 5 complete left tarsometatarsi, proximal ends of 12 right and 11 left tarsometatarsi, distal ends of 19 right and 11 left tarsometatarsi. ROM 17154-17446, 19525. The 294 specimens represent a minimum of 27 individuals. Remarks Although the subspecies of P. plancus living in northwestern Peru today, P. p. cheriway, is much larger than that represented by the fossil material, the latter is still larger than specimens of P. p. plancus from Bolivia. Despite numerous minor differences from living forms, I do not feel there is sufficient basis for placing the fossil material in a new species, especially when considering the great range of variation found in living forms. Instead, this material should be regarded as a temporal subspecies. As I do not believe there is any value in naming subspecies I have not put a separate name on this material. Howard (1938) named the Rancho La Brea caracara as a new species and stated that she believed P. prelutosus represented an ancestral form of the living caracara. At the time she indicated disbelief in temporal subspecies. After examining specimens of ‘‘P. prelutosus’’ from Florida, and comparing her descriptions with the Peruvian material and Recent forms available I do not believe P. prelutosus is a valid species, but rather represents another temporal subspecies, if it is in fact separable from the highly variable living forms. Unfortunately Howard did not have available for comparison any representatives of the smallest members of P. p. plancus which are smaller than measurements given for P. prelutosus. Olson (1976: 363), in a discussion of the fossil species of Polyborus stated, ‘‘P. prelutosus in almost surely the temporal equivalent of, and on a direct genetic line with, P. plancus.’’ The problem of P. prelutosus may need additional study before it can be settled. Milvago brodkorbi sp. nov. Figs 268), C, 29426, 30A, 8 Holotype Complete left tarsometatarsus. ROM 17447. Paratypes One complete right and 4 complete left tarsometatarsi, proximal ends of 2 right and 3 left tarsometatarsi, distal ends of 7 right and 6 left tarsometatarsi. ROM 17448-17470. Referred Material One right and 1 left quadrate, 4 right and 4 left scapulae, 9 complete right and 6 complete left coracoids, humeral ends of 4 right coracoids, proximal ends of 3 right humeri, distal ends of 4 right and 2 left humeri, 3 complete right and 3 complete left ulnae, proximal ends of 7 right and 3 left ulnae, distal ends of 5 right and 2 left ulnae, 2 complete right radii, proximal ends of 4 right and 4 left radii, distal ends of | right and 3 left radii, 12 complete right and 5 complete left carpometacarpi, proximal ends 93 of 5 right and 1 left carpometacarpus, distal ends of 3 right and 1 left carpometacarpus, | complete right femur, proximal ends of | right and 1 left femur, distal ends of | right and 3 left femora, 2 complete right tibiotarsi, proximal ends of 4 right tibiotarsi, distal ends of 10 right and 13 left tibiotarsi. ROM 17471-17606, 19526, 19527. The 162 specimens, including holotype and paratypes, represent a minimum of 17 individuals. Diagnosis Tarsometatarsus agrees with that of Milvago and differs from that of all other South American falconid genera by having (1) internal calcaneal ridge short; (2) external calcaneal ridge directed proximad; (3) shaft with prominent ridges bordering both sides of anterior metatarsal groove; (4) trochleae compressed; (5) area internal to proximal end of internal calcaneal ridge depressed. Compared with the tarsometatarsus of M. chimachima and M. chimango, the holotype is characterized by having (1) intercotylar prominence high and broad (low and broad in M. chimachima, high and narrow in M. chimango); (2) external calcaneal ridge large, with external side long (slender, external side short in M. chimachima and M. chimango); (3) tubercle for M. tibialis anticus very prominent (moderately prominent in M. chimachima, similar in M. chimango); (4) trochlear groove of middle trochlea very deep (moderately deep in M. chimachima, similar in M. chimango); (5) distal projection of internal trochlea long, broad, and only slightly depressed on internal surface (long, narrow, deeply notched on internal surface in M. chimachima; short, broad, deeply notched on internal surface in M. chimango); (6) internal intertrochlear notch deep and wide (moderately deep and narrow in M. chimachima, shallow and narrow in M. chimango); (7) external trochlea of moderate depth (lesser in M. chimachima, greater in M. chimango); (8) trochleae very slightly compressed in M. chimachima and M. chimango); (9) shaft large (small in M. chimachima and M. chimango). For measurements see Table 7. Measurements of the holotype are as follows: length, 61.5; proximal width, 9.8; distal width, 10.4; least width of shaft, 4.1; width of middle trochlea, 4.0. Characters Scapula with (1) acromion broad and rounded proximally in distal view (narrow and straight in M. chimachima and M. chimango); (2) groove dorsal to proximal end of glenoid facet shallow to moderately deep (deep in M. chimachima and M. chimango); (3) glenoid facet broad (narrow in M. chimachima and M. chimango). Coracoid with (1) external side with pronounced internal curvature at dorsal end (slight curvature in M. chimachima and M. chimango),; (2) coracohumeral surface long (short in M. chimachima and M. chimango); (3) attachment of Lig. humero-coracoideum anterius superius greatly elevated anteriorly (slightly elevated in M. chimachima, moderately elevated in M. chimango); (4) external intermuscular line lying on pronounced ridge (slight ridge in M. chimachima, similar in M. chimango), (5) external face in distal view very concave (moderately concave in M. chimachima, similar in M. chimango); (6) concavity dorsal to sternal facet and anterior to external intermuscular line pronounced (slightly developed in M. chimachima, similar in M. chimango). 94 Humerus with (1) profile of head low in palmar view (high in M. chimachima and M. chimango); (2) anconal surface of shaft immediately distal to external tuberosity very concave (moderately concave in M. chimachima and M. chimango; (3) attachment of M. proscapulohumeralis brevis located midway between head and base of median crest (located near base of median crest in M. chimachima, similar in M. chimango); (4) internal tuberosity large (small in M. chimachima, moderate in M. chimango); (5) impression of M. brachialis anticus long and wide (moderately long and wide in M. chimachima and M. chimango); (6) attachment of anterior articular ligament rotated distad (rotated only slightly in M. chimachima, similar in M. chimango); (7) entepicondyle very produced internally (moderately produced in M. chimachima and M. chimango); (8) attachment of M. pronator brevis a large shelf (pit in M. chimachima, slight shelf in M. chimango). Ulna with (1) internal cotyla with deep concavity (moderate in M. chimachima, similar in M. chimango); (2) external corner of external condyle projecting greatly palmarly (projecting only slightly in M. chimachima, projecting moderately in M. chimango); (3) impression of M. brachialis anticus shallow to moderately deep (very deep in M. chimachima, similar in M. chimango)); (4) ligamental attachment at base of internal side of internal condyle a deep pit (shallow in M. chimachima, moderately deep in M. chimango). Radius with (1) bicipital tubercle large (small in M. chimachima and M. chimango); (2) capital tuberosity large (small in M. chimachima and M. chimango); (3) tendinal groove deep (shallow in M. chimachima, moderate in M. chimango); (4) external end of carpal facet extending well onto ligamental prominence (extending only slightly onto prominence in M. chimachima, similar in M. chimango). Carpometacarpus with (1) external rim of carpal trochlea moderately produced proximad, giving a moderate slope to carpal trochlea (greatly produced in M. chimachima, giving a steep slope to carpal trochlea; similar in M. chimango); (2) process of metacarpal I large, greatly produced internally (small, moderately produced in M. chimachima; small, slightly to moderately produced in M. chimango); (3) internal ligamental fossa deep (shallow in M. chimachima, moderate in M. chimango); (4) internal rim of carpal trochlea very large (small in M. chimachima, moderate in M. chimango); (5) depression distal to pisiform process deep (similar in M. chimachima, moderate in M. chimango), (6) attachment of M. flexor digitalis III greatly produced internally (moderately produced in M. chimachima and M. chimango). Femur with (1) attachments of M. gluteus profundus, M. flexor ischiofemoralis, and M. iliacus long (all short in M. chimachima and M. chimango); (2) attachment of M. flexor perforans et perforatus digiti II large, but shallow (small and deep in M. chimachima, of moderate size and depth in M. chimango); (3) rotular groove broad (narrow in M. chimachima and M. chimango),; (4) internal condyle angular in internal view (subangular in M. chimachima and M. chimango); (5) external condyle rounded, with moderate posterior projection (subrounded, with large posterior projection in M. chimachima; less rounded, with moderate posterior projection in M. chimango); (6) ridge dorsal to internal condyle long and very prominent (short and moderately prominent in M. chimachima and M. chimango); (7) shaft with strong curvature in external view (slight curvature in M. chimachima and M. chimango). Tibiotarsus with (1) area distal to internal articular surface deeply excavated, giving a prominent lip to posterior side of head (moderately excavated, with small lip 95 Table 7 Measurements of the scapula, coracoid, humerus, radius, ulna, carpometacarpus, femur, tibiotarsus, and tarsometatarsus of Milvago brodkorbi sp. nov., M. chimachima, and M. chimango (in mm). Scapula Proximal Height Proximal Width Coracoid Length Head to Scapular Facet Proximal Depth Least Depth of Shaft Length of Sternal Facet Humerus Length Proximal Width Distal Width Least Width of Shaft 96 Milvago brodkorbi sp. nov. 35.0-36.9 36.0 9 12.3-13.7 12.9 14.1-15.6 14.9 3 13.6-14.6 14.0 Milvago chimachima Milvago chimango 30.9-31.0 309 2 11.5-15.1 13.3 12.8-13.5 h52 A: 67.9-69.3 68.6 2 15.1-15.3 [5.2 2 12.0-12.2 Milvago brodkorbi sp. nov. Milvago chimachima Milvago chimango Radius Length Proximal Width Distal Width Ulna Length Proximal Width Distal Depth Least Width of Shaft Carpometacarpus Length Height Through Metacarpal I Proximal Width Least Width of Shaft Continued on page 98. 74.7-75.5 I-A 77.3-82.9 80.6 41.7-46.7 43.8 15 11.0-12.2 Gee) 18 4.85.5 =| 20 4.14.6 4.4 23 75.1 39.6 69.4-72.8 Teal 74.6-77.2 TSS 40.7-42.2 41.8 2 10.1—10.5 10.3 97 Length of Distal Fornix Femur Length Proximal Width Distal Width Least Width of Shaft Tibiotarsus Length Proximal Width Distal Width Least Width of Shaft Depth of Internal Condyle Tarsometatarsus Length 98 OR OR Milvago brodkorbi sp. nov. 4.1-5.4 4.8 16 56.2 10.5-11.2 10.8 3 10.1-10.8 10.4 5) 3:3 I 78.2-80.6 79.4 57.9-61.8 59.8 5 Milvago chimachima 4.1 50.6 70.4 D257 Milvago chimango 4.3-4.9 4.6 2 47.6-49.3 48.4 69.0-72.4 107. 7.9-8.0 58.2-60.8 5903 Z Milvago brodkorbi Milvago Milvago sp. nov. chimachima chimango Proximal OR 9.4-10.9 8.9 8.6-8.9 Width M 10.0 8.8 n S) l j3 Distal OR 9.1-10.4 8.8 8.4-8.9 Width M on7 — 8.7 n 14 Z Least OR 3.64.4 3.6 3.0-3.1 Width of M 4.0 — Sal Shaft n 15 | 2 Width of OR 3.44.1 333 3.2-3.3 Middle M 327 — Bee) Trochlea n 19 l 2! in M. chimachima and M. chimango); (2) external condyle with very deep concavity in side (deep in M. chimachima, moderate in M. chimango); (3) internal condyle with moderate concavity in side (deep in M. chimachima, shallow in M. chimango). Etymology This species is named for Dr. Pierce Brodkorb in recognition of his many outstanding contributions to avian palaeontology. Remarks It is apparent from the above comparisons that M. brodkorbi, although quite distinct from M. chimango, is more similar to that species than to M. chimachima. As M. chimango is found today west of the Andes in Chile, and M. chimachima only east of the Andes, a closer relationship between M. chimango and M. brodkorbi is perhaps to be expected. Milvago brodkorbi is the second palaeospecies of Milvago to be described, the first being M. alexandri Olson (1976) from Pleistocene cave deposits in Haiti. The holotype of M. alexandri is a tarsometatarsus (length, 56.2; proximal width, 8.2; distal width, 9.1; least width of shaft, 3.4; width of middle trochlea, 3.4 (Olson, 1976: 358) which is significantly smaller than M. brodkorbi, and which is described as being morphologically more similar to M. chimachima than to M. chimango. In addition, from the published photograph of the type of M. alexandri it is possible to determine that the tarsometatarsus of M. brodkorbi differs from that of M. alexandri by having (1) intercotylar prominence higher and possibly narrower; (2) trochlear groove of middle trochlea deeper; (3) distal projection of internal trochlea larger; (4) anterior metatarsal groove not as deeply excavated. 99 The two living species of Milvago are residents of savannas and lightly wooded areas, and often feed as scavengers. If the fossil material represented a living species its presence in a currently unsuitable habitat could be attributed to its feeding habits and the long-distance attraction of the tar pits for scavengers, but as the material represents an extinct species a postdepositional change in the habitat at the site is implied. Order Galliformes (Temminck) Family Cracidae Vigors Subfamily Penelopinae Bonaparte Penelope Merrem Penelope cf. purpurascens Wagler Crested Guan Material One right scapula, proximal end of 1 left humerus, distal end of 1 right humerus, distal end of 1 right tibiotarsus. ROM 17607-17610. The 4 specimens represent a minimum of | individual. Characters The specimens listed above are readily recognizable as belonging to the genus Penelope by their size and many distinctive characters. They compare favorably with Recent P. purpurascens from Central America (none was available from South America), but are slightly smaller, corresponding to the current trend toward smaller size in South American forms. The fossil material appears to resemble Central American P. purpusascens more than one available Bolivian P. jacquacu Spix, but the great intraspecific variation makes a definitive identification impossible without a larger series. Remarks The known range of P. purpurascens extends along the western slopes of the Andes only as far south as Naranjo (Naranjito), Ecuador, slightly over 125 km from the fossil site. However, suitable habitat does exist today south of its known range, and it may simply not have been found there yet. The presence in the fossil collection of a tropical forest bird such as P. purpurascens suggests that a heavy stand of forest, or at least a good riparian forest, was very near. The fragmentary nature of the fossil material, a lack of comparative material from South America, and the availability of only one P. jacquacu prevents any comment on whether P. purpurascens and P. jacquacu are conspecific or distinct species (Vuilleumier, 1965; Hellmayr and Conover, 1942). 100 Order Ralliformes (Reichenbach) Suborder Ralli (Reichenbach) Family Rallidae Vigors Subfamily Rallinae (Vigors) Porzana Vieillot Porzana carolina (Linnaeus) Sora Material Distal end of 1 left humerus, distal end of 1 left tibiotarsus, distal end of 1 left tarsometatarsus. ROM 17611-17613. The 3 specimens represent a minimum of 1 individual. Remarks A North American migrant, P. carolina occurs as far south as central Peru. Order Charadriiformes (Huxley) Suborder Charadrii (Huxley) Family Charadriidae (Vigors) Coracoid with procoracoid perforated by coracoidal fenestra. Subfamily Vanellinae Blyth Viator gen. nov. Diagnosis Coracoid agrees with that of other genera of Vanellinae, and differs from those of genera of Charadriinae, by having (1) attachment of Lig. humero-coracoideum anterius superius long, lying close to internal side of head; (2) procoracoid long; (3) sterno-coracoidal impression deep. Coracoid differs from that of Belonopterus chilensis (Molina), Vanellus vanellus (Linnaeus) and Hoploxypterus cayanus (Latham), by having (1) head turned approximately 85° from line drawn along scapular facet and procoracoid (turned approximately 65° in B. chilensis, 70° in V. vanellus, 55° in H. cayanus); (2) furcular facet with anterior portion without large dorsal and ventral projections (present in B. chilensis, V. vanellus, and H. cayanus); (3) head very deeply excavated internally (deeply excavated in B. chilensis and H. cayanus, moderately excavated in V. vanellus); (4) shaft with side between glenoid facet and head markedly concave (essentially flat in B. chilensis and V. vanellus , excavated so that large shelf is formed extending from distal end of glenoid facet to middle of head in H. cayanus); (5) attachment of M. coracobrachialis anterior large (smaller in B. chilensis, V. vanellus , and H. cayanus); (6) glenoid facet rotated internally (not rotated in B. chilensis, V. vanellus, and H. cayanus); (7) scapular facet large (smaller in B. chilensis, V. 101 vanellus, and H. cayanus); (8) depression immediately dorsal to sternal facet and anterior to external intermuscular line pronounced, with deeper internal pits (much less pronounced, without pits in B. chilensis, V. vanellus, and H. cayanus); (9) internal sternal shelf very large (smaller in B. chilensis, V. vanellus, and H. cayanus); (10) external sternal angle very long, projecting sharply ventrodistad so that attachment of M. coracobrachialis posterior does not reach sternal facet (of moderate length, not projecting sharply distad in B. chilensis, V. vanellus, and H. cayanus). Coracoid differs from that of Belonopterus (Ptiloscelys) resplendens (Tschudi) in all ways that it differs from that of B. chilensis, and additionally by having (1) scapular facet positioned much lower on shaft; (2) glenoid facet much larger. Type Species Viator picis sp. nov. Etymology Latin, viator, masculine, wayfarer. Viator picis sp. nov. Fig. 30C Holotype Complete right coracoid, procoracoid broken. ROM 13021. Diagnosis As for genus. For measurements see Table 8. Etymology Latin, picis, genitive of pix, tar. Remarks An exhaustive comparative study of the osteology of the genera of lapwings is beyond the scope of this paper. However, osteological evidence cited above and in the following description of Belonopterus sp. nov., and examination of other skeletal elements, indicates that the inclusion of the genera Belonopterus, Ptiloscelys, and Hoploxypterus in the genus Vanellus as proposed by Bock (1958) is unwarranted and serves only to confuse relationships of the several species of lapwings rather than clarify. them. The suggestion by Bock (1958) that B. chilensis and P.. resplendens be placed in a single genus is supported by their similar osteological characters as listed above, but these characters require that these two species be placed in the genus Belonopterus, not combined with Vanellus. The osteological differences between Belonopterus and Vanellus are more than sufficient to separate them at the generic level. Viator picis and Belonopterus sp. nov. (described below) represent the second and third palaeospecies of lapwings to be described from the Pleistocene of the New World, Dorypaltus prosphatus Brodkorb (1959) being recorded from two localities in 102 Florida. The discovery of three palaeospecies of lapwings from the late Pleistocene suggests that perhaps even more palaeospecies will be found in earlier deposits. It also indicates the hazards of describing the origins of current worldwide distributions of the members of a group (Bock, 1958) without knowledge of their fossil record. Belonopterus Reichenbach Belonopterus edmundi sp. nov. Fig. 31h, 8 Holotype Complete left humerus. ROM 13022. Paratype Distal end of one left humerus. ROM 13023. Referred Material One complete left femur. ROM 13024. The 3 specimens, including holotype and paratype, represent a minimum of 2 individuals. Diagnosis Humerus agrees with Belonopterus (including Ptiloscelys) and differs from Vanellus (sensu stricto) and Hoploxypterus by having (1) ectepicondyle reduced, not forming prominent spur (very prominent in Vanellus; moderately prominent in Hoploxyp- terus); (2) attachment of M. extensor metacarpi radialis, pars anconalis, with proximal end forming prominent projection equal to, or slightly lower than, proximal end of ectepicondylar prominence (not projecting, much lower than proximal end of ectepicondylar prominence in Vanellus; not projecting, but only slightly lower than proximal end of ectepicondylar prominence in Hoploxypterus); (3) olecranal fossa very poorly developed (moderate to deep, undercutting proximal end of internal condyle in Vanellus and Hoploxypterus); (4) brachial depression deep (shallow to moderately deep in Vanellus and Hoploxypterus). Humerus is characterized by having (1) head more rounded anconally than palmarly in external view; (2) attachment of M. triceps, external head, very deep, excavated below attachment of M. proscapulohumeralis and undercutting head; (3) bicipital crest very wide; (4) entepicondyle with entire side very concave; (5) entepicondyle oblique, moderately produced; (6) olecranal fossa poorly developed; (7) concavity extending to internal side of external tricipital groove immediately proximal to ectepicondyle deep, giving ectepicondyle a very prominent external projection; (8) ectepicondylar prominence with external side concave; (9) attachment of M. flexor metacarpi radialis, pars anconalis, with proximal end very prominent, equal in height and separated by concavity from proximal end of ectepicondylar prominence; (10) external condyle wide, undercut externally and proximally by narrow groove, and flat internally; (11) brachial depression deep, extending less than 25 per cent of length proximal to proximal end of ectepicondylar prominence; (12) 103 I I ¢ € u = sae c el Cel W WPI 0°6 S'6 Slav cl Tyl-9 TI uo [BWITXO1g I I S € u -- aa Leg 9°89 W 10) ctr ¢°S9-8'6S T'IL-0'L9 uo ygue7] sniouiny I I ¢ € U 10087 ae m2 68 9°8 W [euiai¢ v's 16 76-18 06-78 r(@) jo ysus7T I I S € uU yeys = = TE oe W jo yrdaq CC ve tetas ¢ VCE uo 1seo'] I I S e u a a v9 L‘9 W yidaq 0O'V lee |) bass eS 69-99 uo [BUTXOIdg I I ¢ € u yo0e == = 0°6 9°6 W endeos 09 €Ol €'6- 8 0'OI-1'6 uo 0} peoH I I g € U << = EST L vc W 7h vy S7 8 S7-8°EC T9C-9 tc uo wsueT plooe1od snuvkv9 ‘aou ‘ds jo ‘ua3 “aou ‘ds suapuajdsas sisuaq1yo snaajdkxojdoy s1did 1punupa sniajdouojag snsajdouojag 4OID1A snsajdouojag ‘(WU UT) ‘AOU ‘ds 3a *Uad S191d 10J01,4 JO P10dB10D 9y) JO pue {-AouU “ds 1punwpa snsajdouojag jo INuIaj PUB SNAIUINY 9Y} JO ‘snuvdvI snsajdX{xojdoy pue ‘suapuazdsas *g ‘sisuapiys snsajdouojag JO ANUWII} PUB ‘P10IE.109 ‘sNAVUINY IY} JO SJUIWIAINSEay, «-g AIGE TL, 104 6ST 6 °C |lewe 2) ha) ¢ L Ov 6 Cy-6 LE 86-76 Val Stim ol e 16¢ LOV-LiLe yeys jo yrdoq Seay WyeYS JO PIAA SPIT] UPI [eIsiq WPI [BWIXOlg y3sua7] Inw3,j afApuog yeuia}xq yidoq yes JO WIPIAA Seay (PIAA [easiq 105 attachment of anterior articular ligament with proximal third rotated to face internal condyle; (13) shaft strongly sigmoid in external view. Humerus resembles B. chilensis more than B. (Ptiloscelys) resplendens, but is less than two-thirds the size of either of those species. For measurements see Table 8. Measurements of the holotype are as follows: length, 43.5; proximal width, 9.5; distal width, 6.7; least width of shaft, 2.8; depth of external condyle, 4.2. Characters Femur with (1) head facing approximately equally dorsad and internally; (2) concavity distal to posterior edge of iliac facet very deep; (3) concavity undercutting anterior edge of iliac facet and proximal end of anterior intermuscular line, which remains parallel to edge of trochanter to its end; (4) shaft flattened, with deep depression on internal side of shaft at point of flexure distally; (5) internal condyle slightly rounded, with distalmost point centrally located; (6) external condyle moderately rounded; (7) attachment of M. gastrocnemius, pars externa, very large and elevated; (8) shaft relatively straight, with only slight anteroposteriad curvature. The femur agrees with the humerus in resembling B. chilensis more than B. (Ptiloscelys) resplendens. Etymology This species is named for Dr. A. Gordon Edmund, the individual most responsible for the palaeontological collections from the Talara Tar Seeps. Remarks Belonopterus edmundi is clearly related to B. chilensis and B. (Ptiloscelys) resplendens, and is quite distinct from Vanellus (sensu stricto) and Hoploxypterus, providing additional evidence for the congeneric status of B. chilensis and B. (Ptiloscelys) resplendens. Subfamily Charadriinae (Vigors) Pluvialis Brisson Pluvialis dominica (Muller) Golden Plover Material One left scapula, 4 complete left coracoids, 3 complete right humeri, shaft and distal ends of 2 right and 1 left humerus, proximal ends of 2 left humeri, distal ends of 3 right and 5 left humeri, 1 complete left carpometacarpus, distal ends of 1 right and 1 left tibiotarsus, 1 complete left tarsometatarsus, proximal ends of 2 right and 3 left tarsometatarsi, distal ends of 1 right and 1 left tarsometatarsus. ROM 17614-17645. The 32 specimens represent a minimum of 6 individuals. 106 Characters The smaller size of P. dominica is generally sufficient to separate its elements from those of P. squatarola (Linnaeus). For those elements of the two species that are of similar length, P. squatarola is much stouter or broader. Additional characters of P. dominica separating the two species are coracoid with (1) angle between furcular facet and edge of triosseal canal approximately 135° (approximately 125° in P. squatarola). Humerus with (1) base of median crest slightly nearer capital shaft ridge; (2) internal condyle narrow; (3) olecranal fossa narrow, and deepest on internal side. Carpometacarpus with (1) process of metacarpal I lower; (2) carpal trochlea with internal rim projecting less distad. Femur with (1) head much smaller; (2) greater trochanter much smaller; ridge at internal corner of internal condyle turned externally at greater angle. Tibiotarsus with (1) external articular surface narrower; (2) posterior rim of internal condyle extending less proximad. Tarsometatarsus with (1) intercotylar prominence more undercut externally; (2) proximal foramina situated more proximally; (3) middle trochlea extending distad less in relation to the internal and external trochleae. Remarks A North American migrant, P. dominica occurs on the west coast of South American as far south as Santiago, Chile. Pluvialis squatarola (Linnaeus) Black-bellied Plover Material One complete left femur, distal end of 1 right tibiotarsus, distal end of 1 left tarsometatarsus. ROM 17646-17648. The 3 specimens represent a minimum of 1 individual. Remarks The differences between P. dominica and P. squatarola described above are not considered to be significant enough to place the two species in separate genera. This is in agreement with the recent allocation to congeneric status (A.O.U., 1973). A North American migrant, P. squatarola occurs along the west coast of South America as far south as Concepcion, Chile. Charadrius Linnaeus Charadrius vociferus Linnaeus Killdeer Material Two complete left coracoids, 1 complete right humerus, distal end of | right 107 humerus, proximal end and shaft of 1 right femur, 1 complete right and 1 complete left tibiotarsus. ROM 17649-17655. The 7 specimens represent a minimum of 2 individuals. Remarks Charadrius vociferus is resident on the coast of Peru today. Charadrius collaris Vieillot Collared Plover Material Proximal end of 1 left humerus, shaft and distal end of 1 right humerus, 1 complete left tarsometatarsus. ROM 17656-17658. The 3 specimens represent a minimum of 1 individual. Characters Charadrius collaris is separated from C. alexandrinus Linnaeus, the only other plover of similar size, by having humerus with (1) proximal end narrow; (2) deltoid crest smaller; (3) head less undercut; (4) ectepicondyle and entepicondyle less produced; (5) external condyle wider and shorter. Tarsometatarsus with (1) internal trochlea placed farther distally. Remarks Charadrius collaris occurs all along the west coast of South America. Charadrius semipalmatus Bonaparte Semipalmated Plover Material Complete left tarsometatarsus. ROM 17659. Characters Although the specimen represents an immature bird, the small, but stout, limb elements of C. semipalmatus are sufficient to distinguish it from all other species of the genus. Remarks A North American migrant, C. semipalmatus winters as far south as Llanquihue, Chile. Family Scolopacidae Vigors Coracoid with procoracoid not perforated by coracoidal fenestra. 108 Tringa Linnaeus Tringa ameghini sp. nov. Fig. 30E, F Holotype Complete left coracoid. ROM 12897. Referred Material Proximal end of 1 right tarsometatarsus. ROM 12898. Diagnosis Coracoid agrees with that of Tringa and differs from that of all other scolopacid genera, including Totanus, by having (1) head of moderate height, rectangular, with anterodorsal corner extending farthest anteriad; (2) external side of shaft between glenoid facet and attachment of Lig. humero-coracoideum anterius superius flat; (3) length of procoracoid through base short; (4) procoracoid set off sharply from shaft distally; (5) internal ridge of shaft leading dorsad from internal distal angle similarly developed and at similar angle as in 7. solitaria; (6) sterno-coracoidal impression deep, extending more than halfway up shaft. Coracoid differs from that of T. solitaria by having (1) shaft ventral to coracohumeral surface very flat, excavated from glenoid facet to base of attachment of Lig. humero-coracoideum anterius superius so that external edge of coracohumeral surface forms a lip and a large ridge extends from end of coracohumeral surface to attachment of Lig. humero-coracoideum anterius superius; (2) angle between line drawn along furcular facet and plane of procoracoid approximately 10°—15° less; (3) internal distal angle lying at greater angle to horizontal in internal view; (4) scapular facet directed more posteriad. For measurements see Table 9. Characters Tarsometatarsus differs from that of 7. solitaria by having (1) proximal end narrower, with shaft flaring more gradually toward cotylae; (2) external cotyla positioned more distally; (3) proximal foramina and tubercle for tibialis anticus located more distally. Etymology This species is named for Florentino Ameghino, the great Argentine vertebrate paleontologist. Remarks Although there are not many differences between T. ameghini and T. solitaria (including fossil specimens) the few present are very good specific characters. Tringa ameghini is the first recorded palaeospecies of the genus. While it cannot be proven, it is probable that 7. ameghini was a northern migrant, as are all related species and genera today. 109 Table 9 Measurements of the coracoid and tarsometatarsus of Tringa ameghini sp. nov., Recent T. solitaria, and T. solitaria from the Talara Tar Seeps (in mm). Tringa ameghini Tringa solitaria sp. nov. Talara Tar Seeps Recent Coracoid Length OR 16.2 15.9-16.6 15.0-15.5 M — 16.3 LSE. n l 6 e Head to OR D8 5.8-6.2 5.1-5.5 Scapular M — 6.0 a3 Facet n 1 t | Proximal OR 4.8 4.8-5.1 4.14.5 Depth M — 5.0 4.4 n l 5 3 Least OR 1.9 1.9-2.2 1.7-1.8 Width of M — 2.0 1.8 Shaft n ] 7 3 Length of OR St 5.3-5.7 4.3-5.1 Sternal M — 5.4 4.7 Facet n ] 5 3 Tarsometatarsus Proximal OR 3.6 4.0 3.8 Width M — — i n l ] Tringa solitaria Wilson Solitary Sandpiper Material Two complete right and 5 complete left coracoids, 2 complete right and 3 complete left humeri, 1 complete right carpometacarpus, proximal end of 1 might carpometacarpus, proximal end of | right tibiotarsus, distal ends of 1 right and 2 left tibiotarsi, 1 complete right tarsometatarsus, distal ends of 1 right and 1 left tarsometatarsus. ROM 17660-17680. The 21 specimens represent a minimum of 5 individuals. Characters Tringa solitaria differs from other scolopacid species of similar size by having humerus with (1) proximal end small and bicipital surface small and narrow; (2) groove between capital shaft ridge and base of median crest very deep, deeply undercutting head; (3) line of M. latissimus dorsi anterioris prominent, lying near top 110 of capital shaft ridge; (4) impression of M. brachialis anticus ellipitical, ending distally in deep pointed pit; (5) area between ectepicondylar process and ectepicondyle only slightly concave; (6) entepicondyle produced. Carpometacarpus with (1) ligamental attachment of pisiform process well developed; (2) external side of metacarpal I sharply concave. Tibiotarsus with (1) internal articular surface small; (2) external condyle joining shaft abruptly distally; (3) tendinal groove positioned near internal condyle. Tarsometatarsus with (1) hypotarsus high and rectangular; (2) distal foramen long; (3) shaft long and narrow, and anterior face channeled for more than half its length; (4) shaft widening gradually to meet outer and inner trochleae; (5) trochleae compressed. Remarks For characters of the coracoid of Tringa that distinguish it from that of all other scolopacid genera see diagnosis of T. ameghini. Although the sample is admittedly very small, the measurements given in Table 9 for T. solitaria suggest there has been a significant reduction in the body size of the species through time. A North American migrant, 7. solitaria occurs all along the Peruvian coast. Totanus Bechstein Totanus flavipes (Gmelin) Lesser Yellowlegs Material Seven complete right and 9 complete left coracoids, humeral end of 1 left coracoid, 6 complete right and 7 complete left humeri, proximal ends of 1 right and 7 left humeri, distal ends of 4 right and 3 left humeri, 6 complete right and 2 complete left carpometacarpi, proximal ends of 1 right and 2 left carpometacarpi, distal ends of 2 right and 6 left tibiotarsi, proximal ends of 2 right and 3 left tarsometatarsi, distal ends of | right and 1 left tarsometatarsus. ROM 17681-17749, 19528, 19529. The 71 specimens represent a minimum of 17 individuals. Characters Totanus flavipes differs from other species of scolopacids of similar size by having coracoid with (1) anterior furcular facet with large dorsal extension and small ventral extension; (2) protrusion distal to attachment of Lig. humero-coracoideum anterius superius. Humerus with (1) external tuberosity facing anconally, forming a large shelf; (2) area between attachment of M. supracoracoideus and deltoid crest usually concave; (3) capital shaft ridge large, with anconal projection of head at its proximal end; (4) internal tuberosity prominent and slender; (5) ectepicondyle well developed; (6) attachment of M. anconeus, M. extensor digitorum communis, and M. flexor metacarpi radialis strongly developed; (7) brachial depression moderately deep, not extending to internal condyle. II] Carpometacarpus with (1) metacarpal I long; (2) internal rim of carpal trochlea slanting externally only slightly; (3) lobe at end of external rim of carpal trochlea well developed; (4) concavity at distal end of tendinal groove deep. Tibiotarsus with (1) distal end compressed; (2) condyles raised above shaft anteroproximally. Tarsometatarsus with (1) hypotarsus distinctive; (2) trochleae long and compres- sed; (3) shaft widening gradually distad. Remarks On the basis of the large number of significant osteological differences between Tringa (T. solitaria and T. ameghini) and Totanus (T. flavipes and T. melanoleucus), I have refrained from accepting the recent lumping of the two genera into the genus Tringa (A.O.U., 1973). A North American migrant, 7. flavipes occurs all along the Pacific coast of South America. Totanus melanoleucus (Gmelin) Greater Yellowlegs Material One complete right coracoid, humeral end of 1 left coracoid, 1 complete left humerus, proximal end of | right humerus, proximal end of | left carpometacarpus, distal end of 1 left tibiotarsus, proximal ends of 2 right and 1 left tarsometatarsus, distal ends of 1 right and 1 left tarsometatarsus. ROM 17750-17760. The 11 specimens represent a minimum of 2 individuals. Remarks A North American migrant, 7. melanoleucus occurs all along the Pacific coast of South America. Actitis Illiger Actitis macularia (Linnaeus) Spotted Sandpiper Material Distal end of 1 left humerus. ROM 17761. Characters Humerus with (1) attachment of M. flexor carpi ulnaris large and round; (2) attachment of anterior articular ligament short and high; (3) impression of M. brachialis anticus completely inset; (4) ectepicondylar process small. Remarks A North American migrant, A. macularia occurs all along the coast of Peru. 112 Catoptrophorus Bonaparte Catoptrophorus semipalmatus (Gmelin) Willet Material Humeral ends of 2 right coracoids, 1 complete right and 2 complete left humeri, proximal end of 1 left humerus, distal ends of 1 right and 2 left humeri, 2 complete right and 2 complete left carpometacarpi, distal ends of 2 left tibiotarsi, distal end of 1 left tarsometatarsus. ROM 17762-17777. The 16 specimens represent a minimum of 4 individuals. Remarks A North American migrant, C. semipalmatus occurs all along the coast of Peru. Erolia Vieillot Erolia minutilla (Vieillot) Least Sandpiper Material One complete left coracoid, 9 complete right and 9 complete left humeri, proximal ends of 3 right humeri, distal ends of 2 right and 1 left humerus, 1 complete left femur, | complete right tarsometatarsus, distal end of 1 left tarsometatarsus. ROM 17778-17805. The 28 specimens represent a minimum of 12 individuals. Characters Erolia minutilla is the smallest species of the genus and is readily separated on size. As this species is approximately the same size as Ereunetes mauri and E. pusillus, the following characters were used to distinguish it from the genus Ereunetes. Coracoid with (1) external rim of glenoid facet rotated more posteriad; (2) distal rim of scapular facet small. Additional characters listed by Brodkorb (1963a: 4-5) include (1) coracohumeral surface shorter, making head appear to rise at a lesser angle; (2) juncture of coraco-humeral surface and glenoid facet notched; (3) base of procoracoid process abruptly joining shaft; and (4) ligamental attachment at lower medial edge of shaft shortened. Humerus with (1) internal condyle/external condyle ratio high (Brodkorb, 1955); (2) bicipital crest usually not extending as far distad; (3) internal condyle less produced; (4) attachment of anterior articular ligament lower; (5) impression of M. brachialis anticus long and moderately deep. Carpometacarpus with (1) internal rim of carpal trochlea less rounded; (2) proximal area of fusion of metacarpal II and metacarpal III longer; (3) pisiform process deflected slightly more outward; (4) attachment of M. flexor carpi ulnaris brevis not positioned as far posteriorly. Femur with (1) rotular groove wide, and internal condyle appearing to taper and run off shaft proximally (rather than curving with edge of shaft as in Ereunetes). 113 Tibiotarsus with (1) length short; (2) outer cnemial crest directed more toward fibular crest; (3) distal rim of internal condyle much more developed. Tarsometatarsus with (1) length short; (2) trochleae less compressed; (3) external trochlea extending farther distad. Remarks In view of the numerous osteological differences between the genera Erolia, Ereunetes, and Calidris, I believe it is best to maintain each as a separate genus pending a detailed analysis of their osteological similarities and differences. This belief conflicts with the recent placement of Erolia and Ereunetes in the genus Calidris (A.O.U., 1973). A North American migrant, E. minutilla occurs all along the Peruvian coast. Erolia melanotos (Vieillot) Pectoral Sandpiper Material Five complete right and 8 complete left coracoids, humeral ends of 2 left coracoids, 19 complete right and 16 complete left humeri, proximal ends of 10 right and 6 left humeri, distal ends of 32 right and 41 left humeri, 22 complete right and 16 complete left carpometacarpi, proximal ends of 4 right and 2 left carpometacarpi, 3 complete left femora, distal end of 1 left femur, 1 complete right and 1 complete left tibiotarsus, distal ends of 6 right and 7 left tibiotarsi, 5 complete right and 2 complete left tarsometatarsi, proximal ends of 2 right and 2 left tarsometatarsi, distal ends of 4 right and 4 left tarsometatarsi. ROM 17806-18025, 19560. The 221 specimens represent a minimum of 57 individuals. Characters Erolia melanotos differs from other species of scolopacids by having coracoid with (1) brachial tuberosity large, and protrusion posterior to attachment of Lig. humero-coracoideum anterius superius large, combining to give very wide head; (2) coracohumeral surface short, wide, and sharply curved; (3) shaft large. Humerus with (1) proximal end wide, capital shaft ridge dropping off sharply on either side; (2) attachment of M. supracoracoideus long; (3) area between ectepicondyle and ectepicondylar process usually quite concave; (4) impression of M. brachialis anticus deep distally, undercutting attachment of anterior articular ligament, and becoming gradually shallower proximad; (5) entepicondyle rectangu- lar; (6) external condyle curving slightly toward anterior articular ligament. Carpometacarpus with (1) internal rim of carpal trochlea rounded posteriorly; (2) lobe at end of external rim of trochlea long and well developed, tapering sharply to end; (3) ligamental attachment of pisiform process well developed; (4) ligamental attachment of ulnare deep, separated from ligamental attachment of pisiform process. Femur with (1) trochanter high; (2) attachment of M. gluteus profundus wide; (3) attachment of M. obturator externus deep; (4) space between anterior intermuscular line and trochanteric ridge narrow; (5) internal condyle more oval than elliptical; (6) 114 external condyle quite produced; (7) attachment of M. gastrocnemius, pars externa, large, positioned near base of fibular condyle. Tibiotarsus with (1) internal articular surface undercut distally; (2) outer cnemial crest directed away from fibular crest; (3) shaft convex externally and concave internally between fibular crest and articular surface in distal view; (4) tendinal groove directed toward inner condyle; (5) posterior intercondylar surface wide, but short. Tarsometatarsus with (1) shaft moderately large, widening abruptly both proximad and distad; (2) intercotylar prominence produced; (3) hypotarsus roughly triangular, long on external side, very short on internal side; (4) trochleae not greatly compressed or widened. Remarks A North American migrant, E. melanotos occurs all along the Peruvian coast. Ereunetes Illiger Ereunetes mauri (Cabanis) Western Sandpiper Material One complete right and 2 complete left coracoids, humeral end of | left coracoid, 1 complete right and 2 complete left humeri, shaft and distal end of 1 left humerus, 2 complete right and 2 complete left carpometacarpi, 1 complete right tibiotarsus, distal ends of 2 right and 1 left tibiotarsus, 1 complete left tarsometatarsus. ROM 18026-18041, 19531. The 17 specimens represent a minimum of 3 individuals. Characters Ereunetes mauri and E. pusillus are very similar osteologically, but some elements of the two species can be separated from one another. Of the material listed above I was not able to find any reliable characters for separating carpometacarpi and tibiotarsi of the two species. However, as all other elements were identifiable as E. mauri and not E. pusillus, 1 have assigned the carpometacarpi and tibiotarsi to E. mauri. E. mauri differs from E. pusillus by having coracoid with (1) coracohumeral surface rising at steeper angle; (2) brachial tuberosity larger: the combined effect of (1) and (2) being a larger, thicker head. Humerus with (1) deltoid crest slightly larger; (2) attachment of anterior articular ligament slightly shorter and significantly higher; (3) impression of M. brachialis anticus much deeper. Tarsometatarsus with (1) external rims of internal and external cotylae higher, the cotylae more produced anteriad; (2) distal extension of internal trochlea more distinctly set off from rest of trochlea. Remarks A North American migrant, E. mauri occurs as far south as central Peru. 115 Micropalama Baird Micropalama chapmani sp. nov. Fig. 31G Holotype Complete right femur. ROM 13019. Diagnosis Femur agrees with that of Micropalama and differs from that of all other South American scolopacid genera by having (1) iliac facet slightly curved and extending over very short neck to head; (2) trochanter protruding moderately proximad; (3) iliac facet undercut posteriorly for entire width; (4) ridge leading to external rim of internal condyle from posterior intermuscular line; (5) attachment of M. gastrocnemius, pars externa, large, elevated, and extending distad in a long crescent to reach base of fibular condyle; (6) shaft with distal end turned sharply mediad posterior to anteroproximal termination of internal condyle, and very excavated on medial surface at that point. Femur differs from that of M. himantopus (Bonaparte) by having (1) attachment of M. iliacus much shorter; (2) shaft much stouter; (3) attachment of M. gastrocnemius, pars externa, positioned much more proximad; (4) internal condyle with posterior articular surface extending farther proximad; (5) rotular groove much narrower and deeper; (6) shaft with less excavation at point of medial flexure of distal end; (7) internal and external condyles more rounded ridges at their anteroproximal terminations. Measurements of the holotype are as follows, with those of the coracoids of six Recent M. himantopus and the one coracoid of M. himantopus from the Talara Tar Seeps in parentheses: length from iliac facet to external condyle, 22.2 (21.2-22.2, mean, 21.7; 21.2); width of proximal end, 4.0 (3.6-4.0, mean, 3.9; 4.1); depth of proximal end, 3.3 (2.7—3.0, mean, 2.9; 2.7 +); least width of shaft, 1.8 (1.7-1.8, mean, 1.8; 1.7); width of distal end, 4.3 (3.8-4.1, mean, 4.0; 4.1). Etymology This species is named for the late Dr. Frank M. Chapman in recognition of his many contributions to our knowledge of the ornithology of South America. Remarks Micropalama chapmani is the second fossil species of Micropalama described, M. hesternus Wetmore (1924) from the Upper Pliocene of Arizona being the first. Although based on different elements, the great difference in the ages of the two specimens would tend to preclude their being conspecific. 116 Micropalama himantopus (Bonaparte) Stilt Sandpiper Material One complete right femur. ROM 18042. Remarks A North American migrant, M. himantopus occurs all along the coast of Peru. Arenaria Brisson Arenaria interpres (Linnaeus) Ruddy Turnstone Material Proximal end of 1 rmght humerus, 1 complete left carpometacarpus. ROM 18043-18044. The 2 specimens represent a minimum of 1 individual. Characters Humerus with (1) proximal end wide, and head large; (2) capital groove wide; (3) groove between capital shaft ridge and base of median crest moderately deep; (4) pneumatic fossa wide; (5) shaft widening considerably to meet bicipital crest. Carpometacarpus with (1) trochlear surface wide, and internal rim of carpal trochlea very high; (2) pisiform process positioned near pollical facet; (3) shaft rather large. Remarks A North American migrant, A. interpres occurs all along the coast of Peru. Numenius Brisson Numenius cf. borealis (Foster) Eskimo Curlew Material Complete right coracoid. ROM 18045. Characters Agrees with Numenius and differs from all other scolopacid genera by having (1) head extending only short distance anterior to shaft, and with dorsoventral length very large; (2) attachment of Lig. humero-coracoideum anterius superius very long, shallow, with only slight convexity distal to it; (3) internal edge of coracohumeral surface elevated; (4) internal excavation of head external to furcular facet shallow. 117 Differs from N. phaeopus (Linnaeus) by having (1) much smaller size; (2) head proportionately longer dorsoventrad, and extending less anteriad; (3) dorsal and anterior edge of head straight in external view. Remarks Although the specimen listed above probably represents N. borealis, it can be only tentatively referred to that species until a comparative specimen is examined. If the specimen does represent N. borealis it indicates a change in the distribution of this species in South America, as the winter range of this almost extinct species along the west coast of South America is from Arica, Chile, southward, and the migratory routes are poorly known. Nuntius gen. nov. Diagnosis Coracoid agrees with that of all other genera of scolopacids by having (1) head elongated anteroposteriorly; (2) furcular facet long and wide, forming medial side of deep concavity underlying head; (3) procoracoid not perforated by coracoidal fenestra. Coracoid differs from that of all other scolopacid genera by having (1) head approximately two-thirds as high as long in external view, with anterodorsal corner protruding farther anteriad than anteroventral corner, and dorsal edge essentially straight [length-height ratio similar or greater, with anteroventral corner protruding farther anteriad than anterodorsal corner, and dorsal edge convex in Limnodromous griseus (Gmelin)]; (2) head turned internally at anterior end (turned externally in L. griseus); (3) internal excavation of head deep (shallower in L. griseus); (4) coracohumeral surface short and narrow (longer and wider in L. griseus); (5) area between furcular facet and coracohumeral surface slightly convex, sloping slightly (moderately to strongly convex along dorsal half and essentially vertical in L. griseus); (6) angle between line drawn along glenoid facet and from ventral edge of glenoid facet to tip of external distal angle approximately 160° (approximately 170° in L. griseus); (7) glenoid facet narrow (wider in L. griseus); (8) procoracoid with short base, large anterior end and broad dorsal edge that does not dip below level of anterior corner of scapular facet (short base, small anterior end, and narrow dorsal end that dips below level of anterior corner of scapular facet in L. griseus); (9) sternocoracoidal impression very deep (moderately deep in L. griseus); (10) shaft moderately convex at anterior end of ridge leading along internal side of internal distal angle in internal view (slightly convex in L. griseus); (11) internal distal angle directed anteriad in internal view (directed more ventrad in L. griseus); (12) hyposternal process directed sharply internally (not turned as much internally in L. griseus); (13) sternal facet narrow, with tips of both internal and external distal angles projecting internally well beyond internal edge (wide, with both distal angles projecting only slightly, if at all, internally toward internal edge in L. griseus). Type species Nuntius solitarius sp. nov. 118 Etymology Latin, nuntius, masculine, messenger. Nuntius solitarius sp. nov. Fig. 30D, G Holotype Complete right coracoid. ROM 13018. Referred Material One complete left tarsometatarsus. ROM 13020. The two specimens, including holotype, represent a minimum of one individual. Diagnosis As for genus. Measurements of the holotype are as follows: length, 16.7; head to scapular facet, 5.9; proximal depth, 4.6; least width of shaft,,.2.1; sternal facet length, 526. Characters Tarsometatarsus with (1) intercondylar prominence of moderate width and very prominent (broad and moderately prominent in L. griseus); (2) hypotarsus long and broad, proximal end lying almost at level of internal cotyla (shorter and narrower, proximal end placed well distal to internal cotyla in L. griseus); (3) shaft narrows abruptly distal to condyles, widens slightly immediately distal to proximal foramina, and narrows markedly proximal to trochleae (narrows abruptly distal to condyles and then narrows gradually until it begins widening for trochleae in L. griseus); (4) trochleae spread (similar in L. griseus); (5) middle trochlea of moderate size (larger in L. griseus in specimen of same length); (6) anterior edge of internal trochlea lies anterior to midline of middle trochlea in internal view (lies posterior to midline in L. griseus); (7) trochlear grooves of middle and external trochlea narrow, with well-marked edges (broader and more rounded in L. griseus); (8) middle trochlea ends posteriorly elevated above shaft (merges with shaft in L. griseus); (9) trochlear surfaces of moderate length (longer in L. griseus); (10) metatarsal facet well marked (similar in L. griseus); (11) posterior projection of internal trochlea broad dorsoventrad and set off from rest of condyle by groove (narrower and not set off from rest of condyle in L. griseus); (12) distal third of shaft flattened anteroposteriad (not flattened in L. griseus). Length, 35.0; proximal width, 4.4; distal width, 4.3; least width of shaft, 1.6; width of middle trochlea, 1.4. Etymology Latin, solitarius, solitary. Remarks From the specimens on hand it is not possible to determine which genus within the 119 Scolopacidae is most closely related to Nuntius. Nuntius does not bear any close resemblance to any other scolopacid genus except Limnodromus, which it superficially resembles in size and general form. This superficial resemblance of Nuntius to Limnodromus is the reason why L. griseus was chosen to provide a reference point for the osteological characters of this new genus and species presented in the diagnosis. No close taxonomic or evolutionary relationship between Nuntius and Limnodromus is implied, and indeed, the two genera are quite different in numerous critical areas. For example, in the coracoid the shape and orientation of the head, the orientation of the glenoid facet, the shape of the procoracoid, and the size and position of the sternal articular facets are all quite different. These are the types of osteological differences that are to be expected between unrelated genera of different habits, and evolutionary history. The tarsometatarsus is referred to N. solitarius because it also bears a close superficial resemblance to L. griseus and is of a similar size. It does not resemble Tringa solitaria sufficiently to warrant consideration as an element of T. ameghini. Palnumenius victima L. Miller (1942) was described on the basis of a complete tarsometatarsus from the Upper Pleistocene of San Josecito Cave, Nuevo Leon, Mexico. Palnumenius was described as resembling Numenius more than any other scolopacid genus, and the tarsometatarsus of P. victima is slightly more than twice as large as the tarsometatarsus referred to N. solitarius. Additional characters of the tarsometatarsus referred to N. solitarius that distinguish it from that of P. victima, such as the more proximal position of the internal cotyla relative to the external cotyla, the presence of four hypotarsal ridges instead of three, and a smaller internal trochlea indicate that the two genera are quite different. If the tarsometatarsus referred to N. solitarius should prove to belong to another species, the large size of Palnumenius and its resemblance to Numenius would still indicate that Palnumenius and Nuntius are not congeneric. Family Phalaropodidae Bonaparte Femur with attachment of M. gastrocnemius, pars externa, elevated, lying on corner of shaft, and connected by ridge to base of fibular condyle. Lobipes Cuvier Lobipes lobatus (Linnaeus) Northern Phalarope Material Two complete left humeri. ROM 18046-18047. Characters The humerus of L. /obatus is distinguished from humeri of other phalaropes by size, and from humeri of scolopacids of similar size by having (1) groove between capital shaft ridge and base of median crest very deep, undercutting head; (2) capital shaft ridge depressed, so that it curves anconally to meet base of head; (3) impression of M. n20 brachialis anticus deeply inset proximally and distally; (4) entepicondyle little produced. Remarks A North American migrant, L. Jobatus occurs south on the Pacific coast of South America to Santiago, Chile. Steganopus Vieillot Steganopus tricolor (Vieillot) Wilson’s Phararope Material Five complete right and 6 complete left coracoids, 8 complete right and 8 complete left humeri, proximal ends of 3 right and 2 left humeri, distal ends of 1 right and 3 left humeri, 3 complete right and 6 complete left carpometacarpi, proximal ends of 1 right and 2 left carpometacarpi, 1 complete left femur, distal ends of 6 right and 3 left tibiotarsi, 1 complete left tarsometatarsus, proximal ends of 3 right and 3 left tarsometatarsi, distal end of 1 left tarsometatarsus. ROM 18048-18112, 19559. The 66 specimens represent a minimum of 11 individuals. Characters Steganopus tricolor is the largest of the three species of phalaropes and can generally be easily separated on a size basis. In addition to size, S. tricolor is characterized by having coracoid with (1) shaft long and slender; (2) glenoid facet deeply concave; (3) anterior portion of head long and narrow; (4) coracohumeral surface long; (5) scapular facet elliptical. Humerus with (1) proximal end narrow, and bicipital surface small; (2) attachment of M. supracoracoideus moderately long; (3) impression of M. brachialis anticus deeply inset on all sides and undercutting attachment of anterior articular ligament; (4) external condyle straight; (5) entepicondyle merging smoothly with shaft proximally, and curving slightly palmarly; (6) area between ectepicondyle and ectepicondylar prominence moderately to slightly concave. Carpometacarpus with (1) external side of metacarpal I noticeably concave; (2) internal rim of carpal trochlea flattened posteriorly, distal end protruding farther posteriad than proximal end. Femur with (1) head prominently elevated above shaft on all sides but dorsally, where it extends slightly above level of iliac facet; (2) shaft convex anteriorly; (3) attachment of M. gastrocnemius, pars externa, elevated, facing dorsad, and connected to base of fibular condyle by marked ridge; (4) pit at base of fibular condyle deep; (5) external condyle deep and short. Tibiotarsus with (1) proximal end of external condyle joining shaft gradually; (2) tendinal groove curving slightly under internal condyle; (3) internal condyle of moderate size. Tarsometatarsus with (1) shaft long and narrow; (2) proximal and distal ends laterally compressed; (3) hypotarsus high, narrow, and rectangular; (4) anterior face of shaft flat. 121 Remarks A North American migrant, S. tricolor occurs all along the west coast of South America. Steganopus graui sp. nov. Fig. 31F Holotype Complete right femur. ROM 12896. Diagnosis Femur agrees with that of Steganopus and differs from that of Lobipes and Phalaropus Brisson by having (1) head small, with dorsal surface essentially flush with iliac facet (large, protruding well above iliac facet in Lobipes and Phalaropus), (2) shaft narrows rather abruptly distal to head in posterior view (narrows more gradually in Lobipes and Phalaropus); (3) trochanter narrow (slightly wider in Lobipes and Phalaropus); (4) shaft slender (slender in Lobipes, stout in Phalaropus); (5) attachment of M. gastrocnemius, pars externa, of moderate length, greatly elevated, with marked ridge leading to base of fibular condyle (long, moderately elevated, with moderate ridge, and located more on side of shaft in Lobipes and Phalaropus); (6) size large (small in Lobipes and Phalaropus). Femur differs from that of Steganopus tricolor by having (1) head smaller, not extending as far proximad; (2) trochanter narrower, more concave on internal side; (3) shaft much more concave distal to anterior edge of iliac facet; (4) shaft more compressed laterally, less compressed anteroposteriad, with greater internal flexure at distal end; (5) internal and external condyles deep, with anterior end markedly elevated above shaft (of moderate depth, with anterior end slightly, if any, elevated above shaft in S. tricolor); (6) attachment of M. gastrocnemius, pars externa, more elevated, positioned more distally, with a larger ridge leading to base of fibular condyle. For measurements see Table 10. Etymology This species is named for Admiral Miguel Grau, Peruvian patriot and hero of the War of the Pacific with Chile. Remarks Steganopus graui is the first recorded palaeospecies of the family Phalaropodidae. The greater development of the attachment of M. gastrocnemius, pars externa, the greater lateral compression of the shaft, and the greater elevation of the condyles at their anteroproximal ends suggest that perhaps S. graui was a much stronger swimmer than S. tricolor. 122 Table 10 Measurements of the femur of Steganopus graui sp. nov., Recent S. tricolor, and S. tricolor from the Talara Tar Seeps (in mm). Steganopus graui Steganopus tricolor sp. nov. Talara Tar Seeps Recent Femur Length OR NS) 23.4 20.7-23.5 M — — Migs) n 1 3. Proximal OR 3.5 4.1 3.2-3.9 Width M — _— 3.6 n 1 l 3 Distal OR 3.8 4.3 3.74.0 Width M = = 3.8 n | ] 3 Least OR [3 L.9 1.4-1.5 Width of M — — iTS) Shaft n 1 1 3 Least OR 1 2a0) 1.5-1.6 Depth of M — = 1.6 Shaft n | l 3) Family Jacanidae (Stejneger) Subfamily Jacaninae (Stejneger) Jacana Brisson Jacana spinosa (Linnaeus) Jacana Material Distal end of 1 left humerus. ROM 18113. Characters Humerus with (1) ectepicondylar prominence well developed but small compared with other charadriiform species; (2) impression of M. brachialis anticus deep, undercutting attachment of anterior articular ligament; (3) entepicondyle produced. Remarks Jacana spinosa is resident in areas with abundant floating aquatic vegetation, and was recorded by Chapman (1926) as abundant along tropical lowlands of southern coastal Ecuador. Although the single specimen may represent a vagrant, it is probable that the range of the species previously extended farther south. 123 Family Burhinidae Mathews Burhinus IMliger Burhinus superciliaris (Tschudi) Peruvian Thick-knee Material Three right and 2 left scapulae, 2 complete right and 2 complete left coracoids, humeral end of 1 left coracoid, proximal ends of 1 right and 4 left humeri, distal ends of 2 right humeri, distal ends of 1 right and 1 left radius, 2 complete right femora, distal ends of 3 right and 3 left tibiotarsi, proximal end of 1 right tarsometatarsus, distal end of 1 left tarsometatarsus. ROM 18114-18142. The 29 specimens represent a minimum of 4 individuals. Remarks Burhinus superciliaris is quite common all along the arid littoral of the Pacific coast of Peru, and it is also present on the Santa Elena Peninsula of Ecuador. A bird of dry areas with little ground cover, B. superciliaris is primarily nocturnal in habits, although I have seen them moving about during the day. Family Thinocoridae (Gray) Humerus with (1) attachment of external head of triceps very deep, undercutting head; (2) attachment of M. proscapulohumeralis brevis large, very prominent; (3) attachment of M. brachialis anticus extending under attachment of anterior articular ligament. Thinocoridae gen. et sp. indet. Material Complete right coracoid. ROM 18143. Characters Coracoid with (1) procoracoid non-perforate (similar in Thinocorus rumicivorus Eschscholtz); (2) head long, anterodorsal corner rounded, and area ventral to attachment of Lig. humero-coracoideum anterius superius subangular, extending far ventrad (head longer, anterodorsal corner subangular, and ventral projection rectangular and longer anteroposteriad in T. rumicivorus); (3) head moderately high (very high and angular in T. rumicivorus); (4) internal excavation of head deep (moderate in T. rumicivorus); (5) external side of head concave ventral to dorsal end of coracohumeral surface (similar in 7. rumicivorus); (6) internal edge of coracohumeral surface raised considerably above that surface (only slightly elevated in T. rumicivorus); (7) glenoid facet broad and rounded (broad and less rounded in T. rumicivorus); (8) posteroventral corner of scapular facet prominent (similar in T. 124 rumicivorus); (9) external surface of procoracoid markedly elevated above shaft (slightly elevated in T. rumicivorus); (10) sternal facet long and narrow (short and broad in 7. rumicivorus); (11) external surface of shaft dorsal to external distal angle slightly concave (deeply concave in 7. rumicivorus); (12) sterno-coracoidal impression moderately deep (deep in 7. rumicivorus); (13) anterior edge of attachment of M. supacoracoideus marked by prominent ridge (small ridge in T. rumicivorus). Length, 21.4; head to scapular facet, 7.5; proximal depth, 5.6; least width of shaft, 2.4; length of sternal facet, 6.7. Remarks The specimen listed above resembles Thinocorus more than any other charadriiform genus available for comparison, particularly in the ventral projection of the anterior end of the head and the concavity on its external side. There are too many differences, however, between the fossil specimen and the humerus of Thinocorus for the two to be considered congeneric. No specimens of Attagis were available for comparison. Neither of the two living species of Attagis occurs on the coast of Peru today, although A. gayi Lesson occurs in the Andes from Ecuador southward. Until the specimen can be compared with Afttagis its generic status must remain undetermined. Thinocorus Eschscholtz Thinocorus rumicivorus Eschscholtz Least Seedsnipe Material One complete left humerus. ROM 18144 Characters The single specimen of 7. rumicivorus does not differ in any way from Recent specimens of the species examined, except for having a smaller size. Remarks Thinocorus rumicivorus is the smallest of the four living species of the family Thinocoridae, and is the only one that occurs on the Pacific coast of South America today. T. rumicivorus is represented by five subspecies. The smallest, 7. r. pallidus, which occurs on the Santa Elena Peninsula of southwestern Ecuador, was not available for comparison. Thinocorus koepckeae sp. nov. Fig..3 iC Holotype Complete left humerus (internal tuberosity and bicipital crest broken). ROM 12893. 125 Diagnosis Humerus agrees with that of Thinocorus by having (1) capital groove extremely deep, very deeply undercutting head. Humerus differs from that of 7. rumicivorus by having (1) head shorter but wider; (2) capital groove smaller; (3) attachment of M. supracoracoideus shorter, more shelflike; (4) capital shaft ridge straighter; (5) ligamental furrow more distinct, with deep, narrow distal extension; (6) concavity internal to attachment of M. pectoralis superficialis deeper; (7) attachment of M. pectoralis superficialis rotated less palmarly, located more on deltoid crest; (8) impression of M. brachialis anticus longer, extending proximad beyond proximal edge of ectepicondyle; (9) attachment of anterior articular ligament much smaller, as wide as long (much longer than wide in T. rumicivorus); (10) internal condyle narrower, extending farther distad than external condyle; (11) external condyle narrower, with proximal end not directed toward attachment of anterior articular ligament; (12) internal face of entepicondyle at very small angle to long axis of shaft (rotated approximately 45° to long axis of shaft in T. rumicivorus ); (13) area proximal to entepicondylar prominence slightly concave (deeply concave in T. rumicivorus); (14) external side of ectepicondylar prominence slightly concave (deeply concave in 7. rumicivorus); (15) posterior edge of ectepicondylar prominence straight, not bowing externally, and with slight palmar convexity (slightly concave in T. rumicivorus); (16) shaft much less sigmoid in external view; (17) shaft more slender. Thinocorus koepckeae is smaller than T. rumicivorus (see Table 11), and much smaller than T. orbignyianus (Lesson). Table 11 Measurements of the humerus of Thinocorus koepckeae sp. nov., Recent T. rumicivorus , and 7. rumicivorus from the Talara Tar Seeps (in mm). Thinocorus koepckeae Thinocorus rumicivorus sp. nov. Talara Tar Seeps Recent Humerus Length OR 23:6 24.5 28.8-34.1 M = 30.6 n l l 3 Distal OR 338 4.3 4.8-5.6 Width M — -- Soi n 1 l 3 Least OR 1.6 2.0 2.1-2.3 Depth of M — —— Dad Shaft n l l 3 Depth of OR 2 2.8 3.1-3.5 External M —— a 3.2 Condyle n l 3 126 Etymology This species is named for the late Dr. Maria Koepcke, in recognition of her contributions to our knowledge of Peruvian ornithology. Remarks Although no specimens of Aftagis were available for comparison, no differences between T. koepckeae and T. rumicivorus were large enough to justify placing the two species in separate genera. Thinocorus koepckeae is the first recorded palaeospecies of the family. Suborder Lari Sharpe Family Laridae Vigors Subfamily Larinae (Vigors) Larus Linnaeus Larus atricilla Linnaeus Laughing Gull Material Two premaxillaries, | right scapula, 2 complete right and 1 complete left coracoid, humeral end of | right coracoid, proximal ends of 2 right and 3 left humeri, distal ends of 6 right and | left humerus, 1 complete left carpometacarpus, proximal ends of 2 right and | left carpometacarpus, distal ends of 1 right and 1 left tibiotarsus, 2 complete left tarsometatarsi, proximal end of 1| right tarsometatarsus, distal ends of 3 left tarsometatarsi. ROM 18145-18175. The 31 specimens represent a minimum of 6 individuals. Characters Although considerable intraspecific size variation occurs in gulls, four of the nine species of gulls found in Peru are readily separated from the fossil specimens on a size basis, either because they are too large (L. belcheri Vigors, L. dominicanus Lichenstein, and Creagrus furcatus (Neboux)) or too small (Xema sabini Sabine). Of the remaining five species, L. modestus Tschudi, L. cirrocephalus Vieillot, and L. serranus Tschudi are rather distinctive and easily separated from the fossil material, but L. atricilla and L. pipixcan Wagler are very similar osteologically, and the usual intraspecific variation found in gulls makes their identification very difficult. L. atricilla tends to be larger than L. pipixcan, but the two species overlap broadly in size. In comparison with L. pipixcan, L. modestus, L. cirrocephalus, and L. serranus, L. atricilla is characterized by having scapula with (1) coracoidal articulation of moderate size (similar in L. modestus and L. cirrocephalus; large in L. serranus; small in L. pipixcan); (2) glenoid facet long, narrow, and elliptical (short and triangular in L. modestus; long and triangular in L. cirrocephalus,; long, notched, and elliptical in L. serranus; long, narrow, and elliptical, with very abrupt distal termination in L. pipixcan). Coracoid with (1) anterior projection of head short (long in L. modestus and L. cirrocephalus; similar in L. serranus; short to long in L. pipixcan); (2) glenoid facet elliptical and of moderate width (wide and round in L. modestus and L. serranus; long, elliptical, and of moderate width in L. cirrocephalus; narrow elliptical to round in L. pipixcan); (3) concavity where M. biceps attaches moderate (deep in L. modestus and L. serranus; shallow in L. cirrocephalus , moderate in L. pipixcan); (4) anterior furcular facet elliptical and of moderate length (long and broad in L. serranus; of moderate length and rectangular in L. pipixcan); (5) attachment of Lig. humero-coracoideum anterius superius elliptical, deep, and narrow (large and round in L. modestus and L. serranus; short, deep, and narrow in L. cirrocephalus; small, round to long oval, and deep in L. pipixcan). Humerus with (1) head large (large in L. modestus, L. cirrocephalus, and L. serranus; small in L. pipixcan); (2) capital groove wide and shallow (wide and deep in L. modestus and L. cirrocephalus; narrow in L. serranus; moderately wide and deep in L. pipixcan); (3) distal end laterally compressed with condyles of moderate size (broad, with side condyles in L. modestus and L. cirrocephalus; laterally compressed with large condyles in L. serranus; laterally compressed with small condyles in L. pipixcan); (4) intercondylar groove of moderate depth (very deep in L. modestus; deep in L. cirrocephalus, L serranus, and L. pipixcan); (5) brachial depression deep (very deep in L. modestus, L. cirrocephalus, and L. serranus; of moderate depth in L. pipixcan); (6) ectepicondylar prominence large, projecting palmarly almost perpendicular to palmar surface (broader, but shorter, in L. modestus; larger in L. cirrocephalus and L. serranus; of moderate size and projecting palmarly at angle to palmar surface in L. pipixcan). Carpometarcarpus with (1) internal ligamental fossa deep (deep in L. modestus, L. serranus, and L. pipixcan; shallow in L. cirrocephalus); (2) metacarpal I long with moderately prominent process (short and thick, with prominent process in L. modestus;, short, with moderately prominent process in L. cirrocephalus and L. serranus; moderate, with moderately prominent process in L. pipixcan); (3) groove in carpal trochlea moderately deep (deep in L. modestus and L. serranus; moderately deep in L. cirrocephalus and L. pipixcan); (4) external slant to internal carpal trochlea moderate (similar in L. modestus; less in L. cirrocephalus; greater in L. serranus and L. pipixcan); (5) shaft depth immediately posterior to pollical facet great (great in L. modestus; moderate in L. cirrocephalus and L. serranus; small in L. pipixcan). Tibiotarsus with (1) condyles narrow and not well-rounded (wide external and narrow internal condyle in L. modestus; wide and rounded condyles in L. cirrocephalus; narrow external and wide internal condyle in L. serranus; narrow, well-rounded condyles in L. pipixcan); (2) anterior intercondylar fossa wide with internal condyle undercut internally (narrow in L. modestus; wide in L. cirrocephalus, L. serranus, and L. pipixcan); (3) external distal ligamental attachment shallow (deep, set in deep concavity in L. modestus; deep, not set in concavity in L. cirrocephalus; deep, set in poorly defined concavity in L. serranus, deep, set in moderate concavity in L. pipixcan); (4) notch in distal edge of internal condyle of moderate size (moderate size in L. modestus; deep in L. cirrocephalus, L. serranus, and L. pipixcan). Tarsometatarsus with (1) cotylae very concave (similar in L. modestus; small degree of concavity in outer cotyla in L. serranus and L. pipixcan); (2) hypotarsus wide (similar in L. modestus and L. serranus; narrow in L. pipixcan); (3) trochleae 126 proportionately narrow, with pointed posterior termination of innner trochlea and shallow groove in external trochlea (inner trochlea narrow, and articular surface of outer trochlea with prominent dorsal projection in L. modestus; trochleae thick, with posterior projection of medial edge of inner trochlea small in L. serranus; middle trochlea large in L. pipixcan, with inner and outer trochleae not projecting posteriad as much as in L. atricilla, and a greater posteromedial slant to inner and outer trochleae than is found in L. atricilla). The tarsometatarsus of L. cirrocephalus is readily identified by incomplete fusion of middle and internal trochleae in the adult. Remarks A migrant to the coast of Peru, L. atricilla occasionally reaches as far south as Lima, but usually it does not go beyond the Gulf of Guayaquil. The presence of warm water apparently determines the southward range of this species along the west coast of South America (Murphy, 1936). In years when the warm northern current known as El Nino moves farther south than usual, the range of L. atricilla extends correspondingly. Primarily a coastal bird, L. atricilla occasionally occurs inland around bodies of water. Considering the current ranges and habitat preferences of L. atricilla and L. pipixcan, it appears that there are more specimens of L. atricilla in the fossil collection than would be expected under prevailing conditions, a possible indication of climatic changes. Larus pipixcan Wagler Franklin’s Gull Material One premaxillary, 3 right and 1 left quadrate, 3 right and 4 left scapulae, 3 complete right and 14 complete left coracoids, humeral ends of 4 right and 3 left coracoids, 1 complete right and 1 complete left humerus, proximal ends of 3 right and 4 left humeri, distal ends of 6 right and 4 left humeri, 4 complete right and 3 complete left carpometacarpi, proximal ends of 7 right and 2 left carpometacarpi, distal ends of 7 right and 6 left tibiotarsi, 1 complete right and 2 complete left tarsometatarsi, proximal ends of 2 right tarsometatarsi, distal ends of 6 right tarsometatarsi. ROM 18176-18270. The 95 specimens represent a minimum of 17 individuals. Remarks A North American migrant, L. pipixcan occurs in great numbers along the Peruvian coast, and also inland wherever there is sufficient food. 129 Order Columbiformes (Latham) Suborder Columbidae Latham Family Columbidae (Illiger) Subfamily Columbinae (Illiger) Zenaida Bonaparte Zenaida auriculata (Des Murs) Eared Dove Material Three left scapulae, 3 complete right and 7 complete left coracoids, humeral ends of 1 right and | left coracoid, 4 complete right and 2 complete left humeri, proximal end of 1 left humerus, distal ends of 2 right and 1 left humerus, 9 complete right and 3 complete left ulnae, proximal ends of 3 right and 1 left ulna, distal ends of 4 right and 2 left ulnae, 4 complete right and 2 complete left radii, 8 complete right and 8 complete left carpometacarpi, proximal ends of | right and | left carpometacarpus, distal end of 1 left carpometacarpus, 2 complete right and 2 complete left femora, proximal ends of | right and 1 left femur, distal end of 1 left femur, 5 complete left tibiotarsi, proximal end of 1 left tibiotarsus, distal ends of 1 right and 1 left tibiotarsus, 1 complete right and 5 complete left tarsometatarsi, proximal ends of 1 right and 2 left tarsometatarsi, distai ends of 2 right and 4 left tarsometatarsi. ROM 18271-18372. The 102 specimens represent a minimum of 13 individuals. Characters Zenaida auriculata differs considerably in numerous osteological characters from Leptotila verreauxi (Bonaparte), which is of similar size. Size is sufficient to distinguish it from all other species of doves from coastal Peru and southern Ecuador. Remarks Generally distributed throughout South America, Z. auriculata is very common in coastal Peru. Zenaida asiatica (Linnaeus) White-winged Dove Material One right scapula, 1 complete left coracoid, humeral end of 1 right coracoid, 1 complete right humerus, distal end of 1 right humerus, 1 complete left ulna, proximal end of | left carpometacarpus. ROM 18373-18379. The 7 specimens represent a minimum of 2 individuals. Characters Zenaida is readily distinguished from the genera Columba Linnaeus and Geotrygon 130 Gosse by many characters. Z. asiatica differs considerably from Z. auriculata in size, being a much larger and heavier-bodied bird. Remarks Zenaida asiatica is relatively common in South America from southwestern Ecuador down the coast to northern Chile. Columbina Spix Columbina talpacoti (Temminck) Ruddy Gound Dove Material Two complete left coracoids, 1 complete right carpometacarpus. ROM 18380-18382. The 3 specimens represent a minimum of 2 individuals. Remarks The range of C. talpacoti in coastal northwestern Peru today is limited to those areas of the northernmost departments that have any amount of heavy vegetation. The habitat preferred by this species is apparently heavy brush, and the isolated specimens in the fossil material may indicate that heavy brush cover existed close to the fossil site. At the same time the small numbers may indicate that such cover was also not very close. The large number of specimens of C. cruziana (see below), which is the same size as C. talpacoti, would suggest that size was not a factor in preservation, but rather that the habitat in the vicinity of the site of deposition was not suitable to C. talpacoti. Columbina cruziana (Knip & Pervost) Croaking Ground Dove Material Three left and 1 right scapula, 49 complete right and 39 complete left coracoids, humeral ends of 6 right and 4 left coracoids, 45 complete right and 56 complete left humeri, proximal ends of 12 right and 12 left humeri, distal ends of 10 right and 6 left humeri, 29 complete right and 23 complete left ulnae, proximal ends of 7 right and 6 left ulnae, distal ends of 6 right and 5 left ulnae, 8 complete right and 13 complete left radii, proximal ends of 4 right and 4 left radii, distal ends of 2 right and 3 left radii, 27 complete right and 29 complete left carpometacarpi, proximal ends of 3 right carpometacarpi, 16 complete right and 26 complete left femora, proximal ends of 8 right and 1 left femur, distal ends of 11 right and 4 left femora, 8 complete right and 10 complete left tibiotarsi, proximal ends of 7 right and 5 left tibiotarsi, distal ends of 14 right and 11 left tibiotarsi, 25 complete right and 26 complete left tarsometatarsi, proximal ends of 4 right and 2 left tarsometatarsi, distal ends of 4 right and 6 left tarsometatarsi. ROM 18383-18960, 19561-19574, 20676-20680. The 597 specimens represent a minimum of 68 individuals. {37 Characters Columbina cruziana differs from Claravis pretiosa (Ferrari-Perez) (larger) and Columbina minuta (Linnaeus) (smaller) sufficiently in size so as to readily separate these three species. However, C. cruziana is essentially the same size as C. talpacoti, and they are very similar osteologically. No specimens of C. ¢. buckleyi were available, so the following comparisons with characters of C. cruziana are based on specimens of C. ¢t. rufipennis from Panama. Columbina cruziana differs from C. talpacoti by having scapula with (1) acromion smaller and more rounded; (2) glenoid facet smaller, with dorsal edge more set off from shaft; (3) anteromedial edge between glenoid facet and acromion straighter, less rounded. Coracoid with (1) procoracoid much smaller. Humerus with (1) head rotated more anconally and of less depth; (2) capital groove narrower, but deeper, cutting more into head; (3) impression of M. brachialis anticus deeper, undercutting a more elevated attachment of anterior articular ligament; (4) entepicondyle rotated slightly anconally. Ulna with (1) olecranon smaller; (2) external cotyla smaller; (3) intercotylar area lacking a high ridge. Radius with (1) bicipital tubercle larger; (2) capital tuberosity more pointed; (3) palmar surface of shaft at distal end more rounded. Carpometacarpus with (1) groove separating metacarpal II and metacarpal III under carpal trochlea deep; (2) internal face proximal to pisiform process much more excavated. Femur with (1) attachment of M. gastrocnemius, pars externa, large and elongated, positioned anteriorly (small, oval, situated on posterior side of shaft in C. talpacoti); (2) posterior end of external condyle extending less dorsad; (3) opening between base of internal condyle and ridge leading toward center of shaft posteriorly lacking. Tibiotarsus with (1) rotular crest lower; (2) outer cnemial crest rotated less proximad; (3) external condyle of greater depth. Tarsometatarsus with (1) external cotyla better developed; (2) proximal third of internal edge of shaft very angular, and middle third flat, the transition effected by passing of angular edge posteriad (in C. talpacoti the proximal third may not be as angular, and the edge does not pass posteriad, giving a more rounded edge to the middle third of shaft); (3) ridge leading from metatarsal facet to base of internal trochlea much larger. Remarks Columbina cruziana is very common from coastal Ecuador to the coast of northern Chile, often occurring in large flocks. 132 Order Psittaciformes (Wagler) Family Psittacidae (Illiger) Subfamily Psittacinae (Illiger) Forpus Boie Forpus coelestis (Lesson) Pacific Parrotlet Material One complete left humerus. ROM 18961. Remarks The smallest parrotlet in northwestern Peru, F. coelestis is relatively common throughout its range. As an arboreal species that lives throughout the year in areas with little or no water, its occurrence in the fossil material must be regarded as a fortunate accident. Order Strigiformes (Wagler) Family Tytonidae Ridgway Tyto Billberg Tyto alba (Scopoli) Barn Owl Material Five right and 8 left scapulae, 8 complete right and 1 complete left coracoid, distal end of 1 right humerus, | complete right ulna, proximal end of 1 left ulna, distal ends of 4 left ulnae, proximal ends of 2 right radii, distal ends of 2 right radii, 2 complete left carpometacarpi, proximal end of 1 right carpometacarpus, distal ends of 2 left carpometacarpi, 1 complete right and 1 complete left femur, proximal ends of 4 right and 2 left femora, proximal end of 1 right tibiotarsus, distal ends of 7 right and 8 left tibiotarsi, 2 complete right and 2 complete left tarsometatarsi, proximal ends of 5 right and 4 left tarsometatarsi, shaft of 1 left tarsometatarsus, distal ends of 4 right and 3 left tarsometatarsi. ROM 18962-19044. The 83 specimens represent a minimum of 8 individuals. Remarks Tyto alba is generally, but often locally, distributed throughout South America. (33 Family Strigidae (Vigors) Subfamily Buboninae (Vigors) Bubo Dumeril Bubo virginianus (Gemlin) Great Horned Owl Material One premaxillary, 1 mandibular symphysis, 1 right articular, 3 right and 3 left scapulae, 2 complete left coracoids, humeral ends of 2 right and 2 left coracoids, sternal ends of 1 right and | left coracoid, distal ends of 4 right and 3 left humeri, proximal end of | right ulna, distal ends of 2 right and 3 left ulnae, proximal ends of 3 right and 3 left radii, distal ends of 1 right and 4 left radii, 3 complete right and 1 complete left carpometacarpus, proximal ends of 3 right and | left carpometacarpus, distal ends of | right and 1 left carpometacarpus, 3 complete right and 4 complete left femora, proximal ends of 2 right and 4 left femora, distal ends of 1 right and 1 left femur, proximal ends of 2 left tibiotarsi, distal ends of 5 right and 4 left tibiotarsi, 3 complete right and 6 complete left tarsometatarsi, proximal ends of 4 right and 4 left tarsometatarsi, distal ends of 3 right and 5 left tarsometatarsi, shaft of 1 tarsometatarsus. ROM 19045-19099, 19450-19495, 19546-19547. The 103 speci- mens represent a minimum of 11 individuals. Remarks Generally, although often locally, distributed throughout South America, B. virginianus 1s uncommon in coastal Peru. This species does, however, increase in numbers on the western slopes of the Andes in northern Peru. Speotyto Gloger Speotyto cunicularia (Molina) Burrowing Owl Material One complete right and 1 complete left coracoid, proximal end of 1 left carpometacarpus, 1 complete left femur, proximal end of 1 left tibiotarsus, distal end of 1 left tibiotarsus. ROM 19496-19501. The 6 specimens represent a minimum of 1 individual. Remarks Generally distributed throughout South America, S. cunicularia is very common in coastal Peru today. 134 Subfamily Striginae (Vigors) Asio Brisson Asio flammeus (Pontoppidan) Short-eared Owl Material One mandibular symphysis, | right scapula, 2 complete right coracoids, 1 complete left femur, distal end of 1 right tibiotarsus, 2 complete left tarsometatarsi, proximal ends of 2 right tarsometatarsi, distal ends of 2 right tarsometatarsi. ROM 19502-19513. The 12 specimens represent a minimum of 2 individuals. Remarks Generally distributed in open areas throughout South America, A. flammeus occurs in northwestern Peru today, although not in large numbers. Order Caprimulgiformes (Ridgway) Suborder Caprimulgi Ridgway Family Caprimulgidae Vigors Chordeiles Swainson Chordeiles acutipennis (Hermann) Lesser Nighthawk Material One left scapula, 1 complete right and | complete left coracoid, distal end of 1 right humerus, 1 complete right and | complete left carpometacarpus. ROM 19514-19519. The 6 specimens represent a minimum of | individual. Characters In addition to its larger size, C. acutipennis differs from C. minor (Forster) by having scapula with (1) acromion smaller, projecting less anteriad; (2) concavity between distal end of glenoid facet and ventral edge of shaft slightly deeper. Coracoid with (1) head shorter, more rounded, and wider; (2) attachments for M. coracobrachialis anterior and Lig. humero-coracoideum anterius superius smaller and shallower. Humerus with (1) attachment of anterior articular ligament less elevated; (2) curvature of shaft in region of attachment of M. gastrocnemius, pars externa, greater, resulting in ectepicondylar prominence appearing more produced; (3) concavity in proximal side of ectepicondyle greater. Carpometacarpus with (1) process of metacarpal I shorter and slightly curved; (2) area immediately posterior to pisiform process more excavated; (3) external side of metacarpal I more concave; (4) internal ligamental fossa much deeper. 35 Remarks Chordeiles acutipennis is common in river valleys of coastal Peru. Caprimulgus Linnaeus Caprimulgus piurensis sp. nov. Fig. 31D, E Holotype Complete left coracoid. ROM 12894. Referred material Proximal end and shaft of one right carpometacarpus. ROM 12895. Diagnosis Coracoid agrees with that of Caprimulgus and differs from that of Chordeiles, Nyctidromus Gould, Nyctiphrynus Bonaparte, and Hydropsalis Wagler by having (1) head projecting moderately beyond anterior edge of shaft; (2) brachial tuberosity small. The remaining six genera of South American caprimulgids were unavailable for comparison. Coracoid differs from that of Caprimulgus longirostris Bonaparte and C. parvulus Gould, the only species of the genus occurring today in southwestern Ecuador and northwestern Peru, and from C. cayennensis Gmelin, the only other small South American species of the genus available, by having (1) head narrow, and anterior projection small (narrow with much larger anterior projection in C. /ongirostris , wide with much larger anterior projection in C. cayennensis, wide with larger anterior projection in C. parvulus); (2) angle between furcular facet and triosseal canal approximately 125° (approximately 125° in C. longirostris, 135° in C. cayennensis, and 115° in C. parvulus); (3) attachment of M. coracobrachialis of moderate depth, separated anteriorly from attachment of Lig. humero-coracoideum anterius superius by a high ridge (attachment shallow, ridge very small in C. /ongirostris and C. parvulus; attachment of moderate depth, ridge of moderate size in C. cayennensis); (4) attachment of Lig. humero-coracoideum anterius superius deep, consisting of deep dorsal and moderately deep ventral part separated from each other by high ridge (shallow and of one part in C. longirostris, deep and of one part in C. cayennensis , moderately deep and of one part in C. parvulus); (5) brachial tuberosity undercut by deep concavity (deep in C. longirostris and C. cayennensis, moderate in C. parvulus). For measurements see Table 12. Characters Carpometacarpus with (1) process of metacarpal I long through base (short in C. longirostris, C. cayennensis, and C. parvulus); (2) external side of metacarpal I deeply concave (similar in C. cayennensis, less concave in C. longirostris and C. parvulus); (3) external ligamental attachment long (short in C. longirostris, C. cayennensis, and C. parvulus); (4) fusion of metacarpal II and metacarpal III long and deeply excavated in internal view (similar in C. cayennensis, shorter and less 136 Table 12 Measurements of the coracoid and carpometacarpus of Caprimulgus piurensis sp. nov., C. cayennensis, C. longirostris, and C. parvulus (in mm). Caprimulgus Caprimulgus Caprimulgus Caprimulgus piurensis sp. nov. cayennensis longirostris parvulus Coracoid Length OR 14.1 14.3-14.8 15.0 14.5-14.9 M — 14.6 —— 14.6 n 4 l 4 Head to OR 4.3 4.7-5.0 4.6 4.8-5.] Scapular M = 4.8 — 4.9 Facet n l pe l 4 Proximal OR 2.8 2.9-3.0 Sol 2.9-3.2 Depth M = a0) — 3.0 n l 2 l 4 Least OR 1.4 1.3-1.4 1.4 1.3-1.5 Depth of M — 1.4 -- 1.4 Shaft n l 2 1 4 Carpometacarpus Proximal OR Di 2.3-2.5 2.4 2.32.6 Width M _ 2.4 = 2.4 n Z l 4 Height OR Sho 5.0-5.4 aE 5.2-5.4 Through M — 52 — 2 Metacarpal I n ] 2 1 + excavated in C. longirostris and C. parvulus); (5) proximal end of metacarpal III wide and flat (narrow and rounded in C. longirostris and C. cayennensis, narrow and flat in C. parvulus); (6) shaft very robust (slender in C. longirostris, C cayennensis , and C. parvulus); (7) tendinal groove wide (similar in C. cayennensis , narrow in C. longirostris and C. parvulus). Etymology This species is named for the Departamento de Piura, within the borders of which lie the Talara Tar Seeps. Remarks Caprimulgus piurensis is the first recorded palaeospecies of the family Caprimul- gidae. It appears more similar to C. cayennensis than to either C. longirostris or C. parvulus. Although five species of similar size (C. maculicaudus, C. maculosus, C. nigrescens, C. whiteleyi, and C. hirundinaceus) were not available for comparison, current distribution and habitat preferences of those five make it very unlikely that C. piurensis represents any of them. 137 Discussion and Conclusions Ecology From the northern limit of the Santa Elena Peninsula of southwestern Ecuador (approximately 2°S) to Coquimbo in central Chile (30°S) the western coast of South America is essentially an unbroken 3300-km strip of desert. The desert is delimited on the west by the Pacific Ocean and on the east by the Andes Mountains. The only interruption in the long desert between the two points mentioned above is the forested coastal area along the eastern side of the Gulf of Guayaquil. Variation in local orographic conditions and distances from the coast permit slight Variations in preciptation amounts and cloud cover, and a corresponding variation in vegetation. Following the Holdridge (1967) system of life zone ecology, Tosi (1960) divided the coastal desert into a western subtropical desert and an eastern tropical scrub desert, with local patches of differing vegetation found in both major zones. Zonation at the borders of the desert, both horizontally at the northern and southern ends and vertically up the western slopes of the Andes, is abrupt. For example, along the southern edge of the Gulf of Guayaquil mean annual rainfall at Zorritos is about 25 cm while only 100 km southwest at Lobitas it is less than 5 cm. The vegetation cover decreases correspondingly. La Brea, Peru, the site of the Talara Tar Seeps, lies approximately two-thirds the distance from the nearest point on the coast to the mountains (see Fig. 1), and on Tosi’s map (Tosi, 1960) it falls on the dividing line between the subtropical desert and the tropical scrub desert. While there is some vegetation present at the site it consists only of scattered thorn shrubs and epiphytes. In northern coastal Peru during the local summer, which lasts from September 21 to March 23, the skies are clear with few clouds and little precipitation, except in years when a climatic phenomenon known as E1 Nino appears. Heavy rains may then fall and temporary streams and ponds or small lakes may form. Winds are usually strong, hot, and very dry during this season. During the local winter, which lasts from March 23 to September 21, there exists a heavy cloud cover, or garua, near the coast and reaching variable distances inland, usually to the mountains when they are near the coast. At this time light mists may occasionally prevail near the coast, but inland no precipitation occurs except well up the western slopes of the mountains. Winds are strong, cooler than in summer, and have a higher relative humidity at this time. Recent Avifauna No thorough analysis of the avifauna of northwestern Peru has ever been done. The work of Marchant (1958, 1959, 1960) on the avifauna of the Santa Elena Peninsula of Ecuador gives the most comprehensive list of birds to be expected in northwestern Peru because the two areas show a very high correlation between their avifaunas. Although Chapman (1926) did not discuss this area of Peru, his comments on the avifauna of the arid region of Ecuador also pertain in part to the avifauna of northwestern Peru, particularly that area north of the Sechuran Desert. Basically the modern avifauna consists of two groups of species, specific examples of which are listed in the following section on palaeoecology. One group comprises species adapted to the arid: habitat, and most are representative of what Chapman 13S (1926) termed the Equatorial Arid Fauna, which is found in portions of both Ecuador and Peru. The second group comprises opportunistic species that are able to live for short periods of time in the arid regions when there has been a sufficient amount of precipitation. Within this group are those species living in habitats adjoining the arid regions that extend their range, especially along river courses, during those summers when a short heavy rain may cause an increase in the desert vegetation. If the rains are sufficiently heavy and of sufficient duration these species may be able to breed in an area from which they are normally excluded. They are, however, forced to contract their ranges when the desert resumes its normal arid state. Transient, or migrating, species and wandering individuals of wide-ranging species are also found in this second group. These species may temporarily stop to feed while the desert vegetation is abundant after rains, or to take advantage of temporary ponds or streams that may form. If the summer rains occur, they come during the northern hemisphere winter. North American migrant species are thus very likely to find and stay at these temporary bodies of water. Water being the powerful attractant that it is in a desert, most of the wandering individuals of local species will represent such groups as herons, ibises, ducks, shorebirds, cormorants and coots. Marchant (1958) listed numerous species attracted to the arid Santa Elena Peninsula after a man-made dam created a lake. Palaeoecology Earlier work on various faunal components of the Talara Tar Seeps (Churcher, 1959, 1962, 1965, 1966; Churcher and van Zyll de Jong, 1965; Lemon and Churcher, 1961) all indicate that at the time of deposition of the fossils the region was much more hospitable to a greater variety of animals than at present. Vertebrate remains other than of birds include those of a frog, crocodilians, tortoises, armadillos, ground sloths, bats, skunks, canids, felids, mastodonts, horses, deer, camelids, and rodents. The insect fauna contains several families of coleopterans, lepidopterans, and orthopterans, with aquatic forms a leading element. The molluscan fauna includes a freshwater pulmonate. The palaeoavifauna can be broken into several component groups with each group analyzed for its role in the interpretation of the site. First, a few species are well adapted to arid conditions. These include Burhinus superciliaris, Thinocorus rumicivorus, Columbina cruziana, and Forpus coelestis. Although these species suggest the presence of nearby arid or semi-arid habitats, almost all may be found in cultivated areas and oases within arid zones today. This suggests that while these species may be arid-adapted forms, they are not bound to that habitat but do wander into savanna habitats. This might be particularly true during dry seasons when vegetation cover decreases. These species would not be expected to occur in heavily forested areas as they are all open ground species, with the exception of F. coelestis. A second group consists of wide-ranging species that may colonize suitable habitats over long distances for short periods of time. Examples of such species would include all northern hemisphere migrants, Jabiru mycteria, Anas bahamensis, the condors and vultures, and Polyborus plancus. All of these species can cover large distances in short periods of time and could be expected in any suitable habitat, no matter how limited in size, anywhere they could range. 139 A third group consists of savanna or open woodland species that are also found in sparsely vegetated semi-arid habitats. Examples of these species include Geranoaetus melanoleucus, Buteo polyosoma, Parabuteo unicinctus, Zenaida auriculata, Speotyto cunicularia, and Asio flammeus. While individuals of these species may occasionally wander into arid zones, or even nest there, as does S. cunicularia, they prefer at least a small to moderate vegetation cover. Most of these species are quite common today in the transition zone that extends from the forested slopes of the Andes to the barren coastai desert that is without any vegetation. A fourth group consists of those species found only in forested or heavy scrub habitats. Examples of these species would include Crypturellus transfasciatus, Dendrocygna autumnalis, Cairina moschata, and Penelope purpurascens. These species require at least a heavy scrub or riparian forest within their range. D. autumnalis and C. moschata also require the presence of suitable amounts of water. A fifth group consists of those species that require substantial bodies of water, either as lakes, streams, or swamps. In addition to migrating or wide-ranging species that also fall into this group because of their attraction for water, there are included the grebes, a number of ardeids, and Porzana carolina. Many species are not easily assigned to a specific category and some of those listed above as examples of one category may also be considered in another. Nevertheless, when all groups are considered together the overwhelming evidence is that northwestern Peru was considerably wetter during the period of entrapment of the fauna, and the habitat was probably that of a savanna woodland, or savanna with extensive riparian forests. This is also the conclusion reached by considering the fossil mammals and insects from the deposit. Archeological work (Richardson, 1973) also supports the idea that there has been a considerable change in the climate of the region. Perhaps the best indicator that the habitat of northwestern Peru has changed considerably as a result of climatic change, rather than slightly or not at all, is the very high number of extinct species, comprising at least 23 per cent of the non-passerine portion of the avifauna. Some of the species probably became extinct, as did similar species at Rancho La Brea, California, because their primary food source of large land mammals also became extinct. Examples of such species are the condors and large eagles. More important, however, are the supposedly non-migratory extinct ardeiform (3), anseriform (4), and charadriiform (37?) species that comprise at least half of the known extinct portion of the avifauna. What this suggests is that these species adapted to a specific habitat during the Wisconsin glaciation, and once this habitat began to disappear there was no alternative place for the species to survive. Being non-migratory these species could not have moved elsewhere to survive, and once their habitat had deteriorated past a certain point their fate was sealed. To the south there existed only the desert conditions found today; to the north the humid forests of Ecuador would not provide suitable habitat for savanna or aquatic species, and neither would the mountains to the east. Although savanna or thorn scrub habitat does exist between the arid and humid habitats in this area today, the zonation is so abrupt that the total area is too small to support a varied avifauna, especially those species that require permanent or semipermanent bodies of water for feeding and breeding. 140 The causes of the present-day climate, as opposed to the wetter climate present during glacial periods, are the subject of another paper that is in preparation. Dating Some Extinct Species It was suggested earlier that a deteriorating habitat at the end of the Wisconsin brought about the extinction of certain species thought to be endemic to the area, i.e., the ducks, herons, ibises, and lapwings. If these species became extinct at the end of the Wisconsin when the climate changed, they could not have lived in the area prior to the onset of the glaciation and the formation of the savanna woodland habitat. During the Sangamon interglacial the area would have been desert, as it is now. The question then becomes whether or not the now extinct species came into the area as distinct species from another region, or whether they speciated in situ. Speciation in situ means that the species would have evolved and become extinct within a period of less than 50,000 years, if one considers the onset of the Wisconsin glaciation to be 60,000 to 70,000 years ago and the age of the deposit to be 14,000 years B.P. (Churcher, 1966). Knowing what we do of the rate of divergence within certain species common to southwestern Ecuador and northwestern Peru since the end of the Wisconsin, a figure of 50,000 years is not inconceivable as sufficient time for the formation of a distinct species. Dunn (1974) makes a similar observation by citing the post-Pleistocene speciation patterns within the Irenidae. Selander (1971) suggests that speciation can occur even more rapidly than suggested above, particularly when new habitats appear, as would be the case here. Additionally, the work on rates of speciation of Passer domesticus in North America (Johnston and Selander, 1964, 1971; Selander and Johnston, 1967); the work of Haffer (1967, 1974) on speciation of neotropical birds; and the work of Muller (1973) on neotropical dispersal centers all present evidence of subspeciation occurring over very short periods of time. Comparison of Avifaunas of Talara Tar Seeps and Rancho La Brea Howard (1962) summarized the known avifauna from Rancho La Brea, and from her list the following comparison can be drawn. At least 98 non-passerine and 35 passerine species are known from Rancho La Brea, of which 19 species, or approximately 14 per cent (17 per cent of the non-passerines) are extinct. From the Talara Tar Seeps, 89 non-passerine species are known, of which at least 21, or approximately 23 per cent, are extinct. When the passerine component of the Talara avifauna is completely studied, and the identity of the few questionable non-passerine species is resolved, the final total of species from that locality may exceed that known from Rancho La Brea, and the percentage of extinct species should remain higher. At least 16 species of non-passerines are common to the two faunas, most of which are wide-ranging accipitriform and ardeiform species, or northern hemisphere species that migrate to South America during the winter. The woodpeckers are the only group found in significant numbers at Rancho La Brea and not found at all in the Talara Tar Seeps. Whether this means there were no stands of forest in the immediate vicinity of the tar seeps, or that their remains simply have not been found yet is not known. Because remains of large tree limbs have been found in the matrix, the latter is probably the most likely explanation. 14] Correlation of Talara Tar Seeps with La Carolina, Ecuador Lemon and Churcher (1961) argued that the Talaran faunas are younger than those of the Carolinian of Ecuador, and that the two faunas cannot be of equal age, nor can the Talaran faunas be older than the Carolinian faunas. Edmund (1965) stated that the Carolinian deposits are certainly very late Pleistocene. Of the 89 non-passerine species recorded from the Talara Tar Seeps, 38 (43%) were also recorded from La Carolina; see Table 1 (Campbell, 1976). This includes seven extinct species described in this paper. In terms of specimens, the sample size of the La Carolina avifauna is approximately 25 per cent of that from the Talara Tar Seeps, and 44 per cent of the La Carolina avifauna is comprised of one species of parrot. If the sample size from La Carolina were greater it is probable that there would be an even greater correspondence between the two sites. As I stated earlier (Campbell, 1976), the high degree of similarity between the two localities strongly suggests that there was a great deal of interchange between the two localities. It is not unreasonable to assume that the two local avifaunas were only portions of what would be considered today to be a single avifauna, occupying a unique, but continuous, widespread habitat. The mammalian faunas also demonstrate the same high degree of similarity between the two sites (Lemon and Churcher, 1961). If the Carolinian is older that the Talaran, but still of approximately the same age as indicated by the similar faunas, it can be dated as late Sangamon or early Wisconsin. This would place it prior to the establishment of extensive forests in western Ecuador during the Wisconsin. However, the presence of three extinct species of ducks in the La Carolina avifauna that I believe speciated in situ would require me to argue that a late Sangamon or early Wisconsin age is too old for the site. If the Carolinian is post- Wisconsin the deposits may have been formed early, prior to the establishment of desert conditions and the extinction of the large land mammals. I base this on the presence of the extinct species, and also on the large number of parrots in the La Carolina deposits. The similarity between the two faunas leads one to believe that deposition at the two sites was contemporaneous, although deposition at La Carolina possibly began later and lasted longer into the post- Wisconsin than did deposition at the Talara Tar Seeps. It is possible, of course, that some of the Carolina deposits are pre- or early- Wisconsin and others are late- or post- Wisconsin. 142 Acknowledgements I wish to express my deep and sincere gratitude to Pierce Brodkorb for his abiding interest in, and critical evaluation of, my research and the preparation of this manuscript. His dedication to the field of avian palaeontology has been a source of continuing inspiration. I am deeply indebted to the staff members of the Department of Zoology and the Florida State Museum, both of the University of Florida, for their interest, encouragement, and financial support during my years as a graduate student, and later as a fellow staff member. David Johnston and S. David Webb have been particularly helpful. I wish to thank A. Gordon Edmund of the Royal Ontario Museum, Toronto, Canada, for the opportunity to study the Talara collections. Peruvian officials, particularly Dr. Marc Dourojeanni R., Ing. Eduardo Izquierdo C., Ing. Carlos Ponce del Prado, and Dr. Antonio Brack Egg, were most cooperative in providing me with the necessary permits required to collect comparative materials in Peru. Professor Gustav Orces V. of the Escuela Politecnica Nacional, Quito, Ecuador, kindly permitted me to study the fossil material from La Carolina, Ecuador, and extended many courtesies to me while I was in Ecuador. Arturo Arellano, of the Universidad Nacional Pedro Ruiz Gallo, Lambayeque; Jose Durand, of the Universidad Tecnica del Altiplano, Puno; Max Cerro of Piura; Manuel Plenge of Lima; and Robert Hughes of Mollendo all provided invaluable encouragement and assistance during my visits to Peru. The following individuals and institutions generously loaned many specimens for comparative purposes: Dean Amadon, American Museum of Natural History; N. Philip Ashmole, Yale Peabody Museum; Hildegarde Howard, Natural History Museum of Los Angeles County; Robert Mengel, Museum of Natural History, University of Kansas; Robert W. Storer, Museum of Zoology, University of Michigan; Glen Woolfenden, University of South Florida; and Richard Zusi, United States National Museum. Special acknowledgement is due the Organization for Tropical Studies, Inc., for providing field research funds under OTS Pilot Research Grant N69-23. The Frank M. Chapman Fund, American Museum of Natural History, and the Society of Sigma Xi also provided field research funds. Without their financial support this study would be very incomplete. To all of these people, and many others too numerous to mention, I am deeply grateful. 143 Literature Cited AMADON, D. 1963 Comparison of fossil and recent species: some difficulties. Condor 65:407-409. AMEGHINO, F. 1882 Catalogo explicativo de las colecciones de antropologia prehistorica y de paleontologia de Florentino Ameghino. Catalogo de seccion de la provincia de Buenos Aires, en la Exposicion continental de Sudamerica, 1882, pp. 35-42. 1887 Enumeracion sistematica de las especies de mamiferos fosiles coleccionadas por Carlos Ameghino en los terrenos eocenos de la Patagonia austral y depositados en el Museo La Plata. Boletin del Museo La Plata 1: 1-26. 1891 Enumeracion de las aves fosiles de la Republica Argentina. Revista Argentina de Historia Natural 1:441-453. 1895 Sur les oiseaux fossiles de la Patagonia. Boletin del Instituto Geografico Argentino 15 (11-12):1-104. 1898 Sinopsis geologico-paleontologica. Segundo Censo Nacional de la Republica Argentina lic P1255. 1899 Sinopsis geologico-paleontologica; suplemento. Adiciones y correcciones. La Plata, Imprenta y Encuaderacion ‘‘La Libertad’’. 13 pp. 1901 L’age des formations sédimentaires de Patagonie. Anales de la Sociedad Cientifica Argentina 51:20-39, 65-91. 1905 Enumeracion de los impennes fosiles de la Patagonia y de la isla Seymour. Anales dei Museo Nacional de Buenos Aires 13:97—167. AMEGHINO, F. and C. RUSCONI 1932 Nueva subespecie de avestruz fosil del pampeano inferior Rhea americana anchorenense subesp.n. Anales de la Sociedad Cientifica Argentina 114:38—-42. AMERICAN ORNITHOLOGISTS’ UNION 1973 Thirty-second supplement to the American Omnithologists’ Union check-list of North American birds. Auk 90:411-419. ARRENDODO, O. 1970 Nueva especie de ave pleistocénica de orden Accipitriformes (Accipitridae) y nuevo genero para las Antillas. Ciencias Biologicas Universidad Habana 8:1—-19. 1971 Nuevo género y especie de ave fosil (Accipitriformes: Vulturidae) del pleistoceno de Cuba. Memoria de la Sociedad de Ciencias Naturales La Salle 31(90):309-323. 1976 The great predatory birds of the Pleistocene of Cuba. Jn Olson, S. L., ed., Collected papers in avian paleontology honoring the 90th birthday of Alexander Wetmore. Smithsonian Contributions to Paleobiology 27: 169-187. BOCK, W. J. 1958 A generic review of the plovers (Charadriinae, Aves). Bulletin of the Museum of Comparative Zoology 118:27-97. BOSWORTH, T.O. 1922 Geology of the Tertiary and Quaternary periods in the northwest part of Peru. London, Macmillan. 434 pp. 144 BOULE, M. and A. THEVENIN 1920 Mammiferes fossiles de Tarija. Paris, Imprimerie Nationale. 256 pp. BRODKORB, P. 1955 The avifauna of the Bone Valley formation. Florida Geological Survey, Reports of Investigations 14: 1-57. 1959 The Pleistocene avifauna of Arredondo, Florida. Bulletin of the Florida State Museum, Biological Sciences 4:269-291. 1963 Catalogue of fossil birds. Part 1. Archaeopterygiformes through Ardeiformes. Bulletin of the Florida State Museum, Biological Sciences 7:179-293. 1963a Fossil birds from the Alachua Clay of Florida. Florida Geological Survey, Special Publications 2, Paper 4: 1-17. 1964 Catalogue of fossil birds. Part 2. Anseriformes through Galliformes. Bulletin of the Florida State Museum, Biological Sciences 8:195-335. 1967 Catalogue of fossil birds. Part 3. Ralliformes, Ichthyornithiformes, Charadriiformes. Bulle- tin of the Florida State Museum, Biological Sciences 11:99-220. 1971 Catalogue of fossil birds. Part 4. Columbiformes through Piciformes. Bulletin of the Florida State Museum, Biological Sciences 15: 163-266. BROWN, L. and D. AMADON 1968 Eagles, hawks and falcons of the world. New York, McGraw-Hill, 2 vols., 945 pp. CAMPBELL, K. 1976 The late Pleistocene avifauna of La Carolina, Ecuador. Jn Olson, S. L., ed., Collected papers in avian paleontology honoring the 90th birthday of Alexander Wetmore. Smithso- nian Contributions to Paleobiology 27: 155-168. CATTOI, N. 1957 Contribucion al estudio de la avifauna extinguida del pleistoceno de la Republica Argentina. Ameghiniana 1(1—2):17-24. CHAPMAN, F. M. 1926 The distribution of bird-life in Ecuador: a contribution to a study of the origin of Andean bird-life. Bulletin of the American Museum of Natural History 55: 1-784. CHURCHER, C. S. 1959 Fossil Canis from the tar pits of La Brea, Peru. Science 130:564—565. 1962 Odocoileus salinae and Mazama sp. from the Talara tar seeps, Peru. Royal Ontario Museum, Life Sciences Contributions 57: 1-27. 1965 Camelid material of the genus Palaeolama Gervais from the Talara tar-seeps, Peru, with a description of a new subgenus Astylolama. Proceedings of the Zoological Society of London 145: 161-205. 1966 ‘The insect fauna from the Talara tar-seeps, Peru. Canadian Journal of Zoology 44:985—993. CHURCHER, C. S. and C. G. van ZYLL de JONG 1965 Conepatus talarae n. sp. from the Talara tar-seeps, Peru. Royal Ontario Museum, Life Sciences Contributions 62:1—15. 145 CLARKE, W. 1882 The composition of some fossil eggs, from the Peruvian guano deposits. Proceedings Philosophical Society of Glasgow 13:573-575. CRAYCRAFT, J. 1970 A new species of Telmabates (Phoenicopteriformes) from the lower Eocene of Patago- nia. Condor 72:479-480. DELACOUR, J. 1954 The waterfowl of the world. Vol. 1. London, Country Life. 284 pp. DUNN, D. 1974 Zoogeography of the Irenidae (Aves:Passeres). Biotropica 6: 165-174. EDMUND, A. G. 1965 ___ A late Pleistocene fauna from the Santa Elena Peninsula, Ecuador. Royal Ontario Museum, Life Sciences Contributions 63: 1-21. FISHER, H. 1944 The skulls of the cathartid vultures. Condor 46:272-296. 1946 Adaptations and comparative anatomy of the locomotor apparatus of New World vultures. American Midland Naturalist 35:545-727. FISHER, H. I. and D. C. GOODMAN 1955 The myology of the whooping crane, Grus americana. Illinois Biological Monographs 24:1-127. HAFFER, J. 1967 Speciation in Colombian forest birds west of the Andes. American Museum Novitates 2294: 1-57. 1974 Avian speciation in tropical South America. Publications of the Nuttall Ornithological Club 14: 1-390. HELLMAYR, C. and B. CONOVER 1942 Catalogue of birds of the Americas. Part 1, no. 1. Field Museum of Natural History, Publication 514, Zoological Series 13: 1-636. HOLDRIDGE, L. 1967 Life zone ecology. San Jose, Costa Rica, Tropical Science Center. 206 pp. HOWARD, H. 1929 The avifauna of Emeryville Shellmound. University of California Publications in Zoology 32:301-394. 1938 The Rancho La Brea caracara: a new species. Publications of the Carnegie Institute of Washington 487:217—240. 1955 Anew wading bird from the Eocene of Patagonia. American Museum Novitates 1710: 1-25. 1962 A comparison of avian assemblages from individual pits at Rancho La Brea, California. Los Angeles County Museum, Contributions in Science 58:3—24. 1963 Fossil birds from the Anza-Borrego desert. Los Angeles County Museum, Contributions in Science 73: 1-33. 146 1964 A new species of the ‘‘Pygmy Goose,’’ Anabernicula, from the Oregon Pleistocene, with a discussion of the genus. American Museum Novitates 2200: 1-14. 1968 Limb measurements of the extinct vulture, Coragyps occidentalis, with a description of a new subspecies. Jn Schroeder, A. H., ed., Collected papers in honor of Lyndon Lane Hargrave. Santa Fe, Museum of New Mexico Press, pp. 115-127. 1974 Postcranial elements of the extinct condor, Breagyps clarki (Miller). Los Angeles County Museum of Natural History, Contributions in Science 256: 1—24. IDDINGS, A. and A. A. OLSSON 1928 Geology of northwest Peru. Bulletin of the American Association of Petroleum Geologists 12:1-39. JOHNSTON, R. and R. SELANDER 1964 House sparrows: rapid evolution of races in North America. Science 144:548-550. 1971 Evolution in the house sparrow. II. Adaptive differentiation in North American popula- tions. Evolution 25: 1-28. KOEPCKE, M. 1970 ‘The birds of the Department of Lima, Peru. Wynnewood, Pa., Livingston Pub. Co. 144 pp. KRAGLIEVICH, L. 1931 Contribucion al conocimiento de las aves fosiles de la ¢poca Arauco-Entrerriana. Physis 10:304-315. KRAGLIEVICH, L. and C. RUSCONI 1931 Restos de vertebrados vivientes y extinguidos hallados por los senores E. R. Wagner y hermano en tumulos precolombianos de Santiago del Estero. Physis 10:229-241. LAMBRECHT, K. 1929 Neogaeornis wetzeli n. g., n. sp., der erste Kreidevogel der sudlichen Hemisphare. Palaeontologische. Zeitschrift 11:121-129. 1933 Handbuch der palaeornithologie. Berlin, Gebruder Borntrager. 1024 pp. LEMON, R. and C. S. CHURCHER 1961 Pleistocene geology and paleontology of the Talara region, northwestern Peru. American Journal of Science 259:410-429. LONNBERG, E. 1902. On some fossil remains of a condor from Bolivia. Bulletin of the Geological Institution of the University of Upsala 6:1-11. MARCHANT, S. 1958 The birds of Santa Elena Peninsula, S. W. Ecuador. Ibis 100:349-387. 1959 The breeding season in S. W. Ecuador. Ibis 101:137—152. 1960 The breeding of some S. W. Ecuadorian birds. Ibis 102:349-382, 584-599. MAYR, E. and L. L. SHORT 1970 Species taxa of North American birds. Publications of the Nuttal Ornithological Club 9:1-127. 147 MERCERAT, A. 1897 Note sur les oiseaux fossiles de la République Argentine. Anales de la Sociedad Cientifica Argentina 43:222-—240. MEYER de SCHAUENSEE, R. 1966 ‘The species of birds of South America. Wynnewood, Pa., Livingston Pub. Co. 577 pp. MILLER, A. 1953 A fossil hoatzin from the Miocene of Colombia. Auk 70:484—489. MILLER, L. 1910 The condor-like vultures of Rancho La Brea. University of California Publications, Bulletin of the Department of Geology 6:1-19. 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. 1915. A walking eagle from Rancho La Brea. Condor 17:179-181. 1931 Bird remains from the Kern River Pliocene of California. Condor 33:70-72. 1942 Two new bird genera from the Pleistocene of Mexico. University of California Publications in Zoology 47:43-46. MILLER, L. and H. HOWARD 1938 The status of the extinct condor-like birds of the Rancho La Brea Pleistocene. Publications of the University of California at Los Angeles in Biological Sciences 1:169-175. MORENO, F. and A. MERCERAT 1891 Catalogo de los pajaros fosiles de la Republica Argentina conservados en el Museo de La Plata. Anales del Museo de La Plata, Paleontologia Argentina 1:1—71. MULLER, P. 1973 The dispersal centres of terrestrial vertebrates in the neotropical realm. The Hague, W. Junk. 244 pp. MURPHY, R. 1936 Oceanic birds of South America. New York, American Museum of Natural History. 2 vols., 1245 pp. OLSON, S. 1976 A new species of Milvago from Hispanola, with notes on other fossil caracaras from the West Indies (Aves: Falconidae). Proceedings of the Biological Society of Washington 88(33):355-366. OPPENHEIM, V. 1943 The fossiliferous basin of Tarija, Bolivia. Journal of Geology 51:548—-555. OWRE, O. 1967 Adaptations for locomotion and feeding in the anhinga and double-crested cormorant. Ornithological Monographs 6: 1-138. PATTERSON, B. 1941 A new phororhacoid bird from the Deseado Formation of Patagonia. Field Museum of Natural History, Geology Series 8:49-54. 148 PATTERSON, B. and J. L. KRAGLIEVICH 1960 Sistematica y nomenclatura de las aves fororracoideas del Plioceno Argentino. Pub- licaciones del Museo Municipal de Ciencias Naturales y Tradicionales de Mar del Plata 1:3-51. RICHARDSON, J. 1973 The Preceramic sequence and the Pleistocene and post-Pleistocene climate of northwestern Peru. /n Lathrop, D. W. and J. Douglas, eds., Variation in anthropology: essays in honour of John C. McGregor. Urbana, Illinois Archaeological Survey, pp. 73-89. ROSS, R. 1935 A new genus and species of pigmy goose from the McKittrick Pleistocene. Transactions of the San Diego Society of Natural History 8:107-114. ROVERETO, C. 1914 Los estratos araucanos y sus fosiles. Anales del Museo Nacional de Historia Natural de Buenos Aires 25: 1-249. RUSCONI, C. 1946 Ave y reptil oligocenos de Mendoza. Boletin Paleontologia de Buenos Aires 21:1-15. 1958 Ave fosil del plioceno de Buenos Aires. Revista del Museo del Historia Natural de Mendoza 11:157-160. SCHOENER, T. 1969 Size patterns in West Indian Anolis lizards: I. Size and species diversity. Systematic Zoology 18:386—401. 1970 Size patterns in West Indian Anolis lizards. I]. Correlations with the sizes of particular sympatric species — displacement and convergence. American Naturalist 104: 155-174. SELANDER, R. 1971 Systematics and speciation in birds. Jn Farner, D. S. and J. R. King, eds., Avian biology. New York, Academic Press, pp. 57-147. SELANDER, R. and R. JOHNSTON 1967 Evolution in the house sparrow. I. Intrapopulation variation in North America. Condor 69:217-258. SHEPPARD, G. 1937 The geology of south-western Ecuador. London, T. Murby. 275 pp. SHORT, L. 1970 Méid-Pleistocene birds from western Nebraska, including a new species of sheldgoose. Con- dor 72:147-152. SIMPSON, G. 1972 Conspectus of Patagonian fossil penguins (Argentina). American Museum Novitates 2488: 1-37. SNYDER, L. and H. LUMSDEN 1951 +‘ Variation in Anas cyanoptera. Occasional Papers of the Royal Ontario Museum of Zoology 10:1-18. SPILLMAN, F. 1942 Contribucion al conocimiento de fosiles nuevos de la avifauna ecuatoriana en el pleistoceno de Santa Elena. Proceedings of the Eighth American Science Congress, Washington 4:375-389. 149 STOCK, C. 1956 Rancho La Brea; a record of Pleistocene life in California. 6th ed. Los Angeles County Museum, Science Series 20, Paleontology 11:1-81. STRESEMANN, E. 1924 Oceanodroma hornbyi (Gray) aus einem chilenischen Salpeterfelde. Ornithologische Monatsberichte 32:61-—63. TORDOFF, H. 1959 A condor from the Upper Pliocene of Kansas. Condor 61:338-343. TOS: 1960 Zonas de vida natural en el Peru, memoria explicativa sobre el mapa ecologico de Peru. Boletin Técnico no. 5, Zona Andina, Proyecto 39. Lima, Instituto Interamericano de Ciencias Agricolas. 271 pp. TRAVIS, R. 1953 La Brea-Parinas oil field, northwestern Peru. Bulletin of the American Association of Petroleum Geologists 37:2093-2118. VUILLEUMIER, F. 1965 Relationships and evolution within the Cracidae (Aves, Galliformes). Bulletin of the Museum of Comparative Zoology 134: 1-27. WETMORE, A. 1924 Fossil birds from southeastern Arizona. Proceedings of the United States National Museum 64:1-18. 1935. Pre-Columbian bird remains from Venezuela. Auk 52:328-329. 1937 Bird remains from cave deposits on Great Exuma Island in the Bahamas. Bulletin of the Museum of Comparative Zoology 80:427-441. 1965 The birds of the Republic of Panama. Part I. Tinamidae (tinamous) to Rhynchopidae (skimmers). Smithsonian Miscellaneous Collections 150: 1-483. WINGE, O. 1887 Fugle fra Knoglehuler: Brasilien. .E Museo Lundii 1(2):1—54. WOOLFENDEN, G. 1961 Postcranial osteology of the waterfowl. Bulletin of the Florida State Museum, Biological Sciences 6:1—129. 150 < 1 Left tarsometatarsus of Anas sanctaehelenae sp. nov., in anterior and posterior view, ROM 12903, xX 1. C(i) (i B(i) (i Ali) fi D(i) (i E(i) (i ae H(i) (i 163 Fig. 12. A(i), (ii) Proximal end of right humerus of Geronogyps reliquus gen. et sp. nov., in anconal and palmar view, ROM 12997, x 2/3. B(i), (1i) Distal end of right humerus of Geronogyps reliquus gen. et sp. nov., in anconal and palmar view, ROM 12998, x 2/3. 164 165 Fig. 13. A(i), (ii) Left humerus of Gymnogyps howardae sp. nov., in anconal and palmar view, ROM 129725 «2/3: 166 A(i) (11) 167 Fig. 14 a(i)-(iii) Left coracoid of Geronogyps reliquus gen. et sp. nov., in anterior, lateral, and posterior view, ROM 12891, xX I. B(i)—(il1) Right coracoid of Gymnogyps howardae sp. nov., in anterior, lateral, and posterior view, ROM 12969, x 1. 168 2 ————————E———————E——E———————————————————~~—~ss~~m ESSERE aes I 169 Fig. 15 A(i), (ii) Proximal end of left ulna of Geronogyps reliquus gen. et sp. nov., in palmar and anconal view, ROM 12999, x 1. B(i)-(iii) Distal end of left ulna of Geronogyps reliquus gen. et sp. nov., in external, anconal, and internal view, ROM 13000, x 1. 170 ) B(i iy) ( A(i) Fig. 16 (i), (ii) Proximal end of right ulna of Gymnogyps howardae sp. nov., in palmar and anconal view, ROM 12978, x 1. B(i)—(iii) Distal end of right ulna of Gymnogyps howardae sp. nov., in external, anconal, and internal view, ROM 12980, x 1. 172 Fig. 17 174 A(i), (ii) B(i), (ii) C(i), (ii) D(i), (11) E(i), (11) Proximal end of left radius of Geronogyps reliquus gen. et sp. nov., in palmar and anconal view, ROM 13002, x 1. Left scapula of Gymnogyps howardae sp. nov., in external and internal view, ROM 12968, Xl. Distal end of right radius of Geronogyps reliquus gen. et sp. nov., in palmar and anconal view, ROM 13003, x 1. Distal end of right radius of Gymnogyps howardae sp. nov., in palmar and anconal view, ROM 12938, x 1. Proximal end of right carpometacarpus of Geronogyps reliquus gen. et sp. nov., in internal and external view, ROM 13006, x 1. ty D(i) L/S Fig. 18 (i), (ii) Proximal portion of sternum of Gymnogyps howardae sp. nov., in lateral and anterior view, ROM 12967, xX 1. B(i) Proximal portion of sternum of Geronogyps reliquus gen. et sp. nov., in anterior view, ROM 12990, x 1. c(i) Sternum of Geronogyps reliquus gen. sp. nov., in lateral view, ROM 12989, x 1. 176 a LIF. Fig. 19 178 A(i), (ii) B(i) C(i), (11) D(i), (ii) E(i), (ii) Proximal end of right femur of Geronogyps reliquus gen. et sp. nov., in anterior and posterior view, ROM 13008, x 1. Proximal end of right femur of Geronogyps reliquus gen. et. sp. nov., in external view, ROM 13009, x 1. Distal end of right femur of Geronogyps reliquus gen. et sp. nov., in anterior and posterior view, ROM 13012, xX 1. Distal end of left tibiotarsus of Geronogyps reliquus gen. et sp. nov., in anterior and posterior view, ROM 13007, x 1. Distal end of left -tibiotarsus of Gymnogyps howardae sp. nov., in anterior and posterior view, ROM 12984, x 1. B(i) 179 Fig. 20 A(i)-(v) Holotype tarsometatarsus of Geronogyps reliquus gen. et sp. nov., in anterior, external, posterior, proximal, and distal view, ROM 12986, xX 1. 180 Ali) 181 Fig. 21 A(i)-(v) Holotype right tarsometatarsus of Gymnogyps howardae sp. nov., in anterior, external, posterior, proximal, and distal view, ROM 12956, x 1. 182 (il) Ali 183 Fig. 22 a(i)-(iii) B(i), (il) C(i), (ii) D(i)-{il1) 184 Distal end of right ulna of Sarcoramphus? fisheri sp. nov., in external, anconal, and internal view, ROM 12888, x 1. Distal end of holotype right tibiotarsus of Sarcoramphus? fisheri sp. nov., in anterior and posterior view, ROM 12887, xX 1. Proximal end of left radius of Sarcoramphus? fisheri sp. nov., in palmar and anconal view, ROM 12889, x 1. Left scapula of Amplibuteo hibbardi gen. et sp. nov., in external, internal, and proximal view, ROM 16923, x 1. B(i) Vlas (i AQ: ‘ AWN “o RRS, C(i) (i 185 Fig. 23. A(i)-iv) Left coracoid of Miraquila terrestris gen. et sp. nov., in anterior, lateral, posterior, and medial view, ROM 13030, x 1. B(i)—(ili) Left coracoid of Amplibuteo hibbardi gen. et sp. nov., in anterior, lateral, and posterior view, ROM 16925, x 1. 186 A\i) 187 Fig. 24 188 A(i), (il) B(i) C(i), (il) D(i), (ii) E(i), (il) Right ramus of Amplibuteo hibbardi gen. et sp. nov., in dorsal and external view, ROM 16921, x 1. Rostrum of Amplibuteo hibbardi gen. et sp. nov., in external view, ROM 16919, x 1. Left carpometacarpus of Miraquila terrestris gen. et sp. nov., in internal and external view, ROM 13029, x 1. Proximal end of right carpometacarpus of Amplibuteo hibbardi gen. et sp. nov., in internal and external view, ROM 16943, x 1. Distal end of right carpometacarpus of Amplibuteo hibbardi gen. et sp. nov., in internal and external view, ROM 16945, x 1. Ali) D(i) ii) ( ) E( C(i) 189 Fig. 25 190 A(i), (ii) B(i), (ii) C(i), (i) D(i), (1i) E(i)—(iii) Right humerus of Amplibuteo hibbardi gen. et sp. nov., in anconal and palmar view, ROM 16927, x 2/3. Proximal end of left radius of Amplibuteo hibbardi gen. et sp. nov., in palmar and anconal view, ROM 16939, x 1. Distal end of left radius of Amplibuteo hibbardi gen. et sp. nov., in palmar and anconal view, ROM 16941, x 1. Proximal end of right ulna of Amplibuteo hibbardi gen. et sp. nov., in palmar and anconal view, ROM 16932, x 1. Distal end of right ulna of Amplibuteo hibbardi gen. et sp. nov., in external, anconal, and internal view, ROM 16933, x 1. 19] Fig. 26 A(i)—(iii) Right femur of Amplibuteo hibbardi gen. et sp. nov., in anterior, external, and posterior view, ROM 16946, x 1. 192 193 Fig. 27 194 A(i)—(iii) B(i), (il) C(i)-(ili) D(i)—(iii) E(1), (11) Proximal end of right tibiotarsus of Amplibuteo hibbardi gen. et sp. nov., in posterior, internal, and proximal view, ROM 16947, x 1. Distal end of left tibiotarsus of Amplibuteo hibbardi gen. et sp. nov., in anterior and posterior view, ROM 16951, x 1. Proximal end of left tibiotarsus of Miraquila terrestris gen. et sp. nov., in posterior, internal, and proximal view, ROM 13031, x 1. Proximal end of holotype right tarsometatarsus of Miraquila terrestris gen. et sp. nov., in anterior, external, and posterior view, ROM 13025, x 1. Distal end of right tarsometatarsus of Miraquila terrestris gen. et sp. nov., in anterior and posterior view, ROM 13028, x 1. ri LY yy, MEM ty E(i) (1i) 195 Fig. 28 A(i)—-(iii) Holotype left tarsometatarsus of Amplibuteo hibbardi gen. et sp. nov., in anterior, external, and posterior view, ROM 16905, xX 1. B(i), (ii) Right quadrate of Milvago brodkorbi sp. nov., in external and internal view, ROM L74T i x1 c(i), (ii) Right scapula of Milvago brodkorbi sp. nov., in external and internal view, ROM WAT3 5 xa. 196 Ali) (i “(i 197 Fig, 29 198 A(i), (ii) B(i), (ii) c(i), (ii) p(i), (ii) E(1), (ii) F(i), (ii) G(i), (11) Proximal end of right humerus of Milvago brodkorbi sp. nov., in anconal and palmar view, ROM 17500, x 1. Distal end of left humerus of Milvago brodkorbi sp. nov., in anconal and palmar view, ROM 17507, x 1. Left coracoid of Milvago brodkorbi sp. nov., in anterior and posterior view, ROM 17490, x 1. Left carpometacarpus of Milvago brodkorbi sp. nov., in internal and external view, ROM 17558, X 1. Right ulna of Milvago brodkorbi sp. nov., in palmar and anconal view, ROM 17509, x 1. Right radius of Milvago brodkorbi sp. nov., in palmar and anconal view, ROM IW5S2-.5¢ 1c Right tibiotarsus of Milvago brodkorbi sp. nov., in anterior and posterior view, ROM 175805.