Fossil Vertebrates from the Bahamas STORRS L. OLSON, EDITOR SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION Emphasis upon publication as a means of "diffusing knowledge" was expressed by the first Secretary of the Smithsonian. In his formal plan for the Institution, Joseph Henry outlined a program that included the following statement: "It is proposed to publish a series of reports, giving an account of the new discoveries in science, and of the changes made from year to year in all branches of knowledge." This theme of basic research has been adhered to through the years by thousands of titles issued in series publications under the Smithsonian imprint, commencing with Smithsonian Contributions to Knowledge in 1848 and continuing with the following active series: Smithsonian Contributions to Anthropology Smithsonian Contributions to Astrophysics Smithsonian Contributions to Botany Smithsonian Contributions to the Earth Sciences Smithsonian Contributions to the Marine Sciences Smithsonian Contributions to Paleobiology Smithsonian Contributions to Zoology Smithsonian Studies in Air and Space Smithsonian Studies in History and Technology In these series, the Institution publishes small papers and full-scale monographs that report the research and collections of its various museums and bureaux or of professional colleagues in the world of science and scholarship. The publications are distributed by mailing lists to libraries, universities, and similar institutions throughout the world. Papers or monographs submitted for series publication are received by the Smithsonian Institution Press, subject to its own review for format and style, only through departments of the various Smithsonian museums or bureaux, where the manuscripts are given substantive review. Press requirements for manuscript and art preparation are outlined on the inside back cover. S. Dillon Ripley Secretary Smithsonian Institution SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY • NUMBER 48 Fossil Vertebrates from the Bahamas Storrs L. Olson EDITOR SMITHSONIAN INSTITUTION PRESS City of Washington 1982 ABSTRACT Olson, Storrs L., editor. Fossil Vertebrates from the Bahamas. Smithsonian Contributions to Paleobiology, number 48, 65 pages, 12 figures, 1982.—The three papers in this volume summarize the previous literature on fossil vertebrates from the Bahamas, provide revisions of the previously described fossil speci- mens, include identifications of newly collected material, and discuss changes in the late Pleistocene environment of the Bahaman archipelago. Olson and Pregill review the history of fossil exploration in the Bahamas, describe the known fossil localities, and briefly discuss the depauperate mammalian fauna. Pregill reviews the Pleistocene herpetofauna of New Providence Island, which is similar to that found on the island today, the only extinct taxa being a tortoise (Geochelone), a crocodile (Crocodylus), an iguana (Cyclura), and a gecko of the genus Aristelliger (previously misidentified as Tarentola). Taphonomy of the New Providence deposits and the zoogeographical patterns of the herpe- tofauna are discussed in relation to arid climatic conditions of the Wisconsinan glacial period. It is suggested that the establishment of a north-south rainfall gradient within the Bahamas has caused more extinctions in the wetter northern islands, whereas a more diverse herpetofauna persists in the drier southern islands. Olson and Hilgartner review the fossil record of birds from the Bahamas and propose the following changes in nomenclature: Calohierax quadratus = Buteo sp., Burhinus nanus = Burhinus bistriatus nanus, Glaucidium dickinsoni = Athene cunicularia, Otus providentiae = Athene cunicularia, Bathoceleus hyphalus = Melanerpes superciliaris, Corvus wetmorei = Corvus nasicus. About 50% of the fossil avifauna of New Providence no longer occurs there and 40% is extinct in the Bahamas. Species composition indicates that the Bahamas in the late Pleistocene were drier and had more open savanna-like and broadleaf scrub habitats. Subsequent increases in rainfall caused habitat changes that resulted in extinction. The implications of this for modern ecological theories are discussed. OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recorded in the Institution s annual report, Smithsonian Year. SERIES COVER DESICN: The trilobite Phacobs rana Green. ^ Library of Congress Cataloging in Publication Data Main entry under title: Fossil vertebrates from the Bahamas. (Smithsonian contributions to paleobiology ; no. 48) Bibliography: p. Includes index. Contents: Introduction to the paleontology of Bahaman vertebrates / by Storrs L Olson and Gregory K. Pregill-Fossil amphibians and reptiles from New Providence Island, Bahamas ' b,yG^gory K. Pregill-Fossil and subfossil birds from the Bahamas / by Storrs L. Olson and William B. Hilgartner. 7 1. Vertebrates, Fossil-Addresses essays lectures. 2. Paleontology-Bahamas-Addresses, es- Series CS' * g ' Gre%oryK' ni-Hilgartner, William B. IV. QE701.S56 no. 48 [QE841] 560s [566'.097296] 81-13543 AACR2 * Contents Page Introduction to the Paleontology of Bahaman Vertebrates, by Storrs L. Olson and Gregory K. Pregill 1 Fossil Amphibians and Reptiles from New Providence Island, Bahamas, by Gregory K. Pregill 8 Fossil and Subfossil Birds from the Bahamas, by Storrs L. Olson and William B. Hilgartner 22 Literature Cited 57 Index 61 in Fossil Vertebrates from the Bahamas Introduction to the Paleontology of Bahaman Vertebrates Storrs L. Olson and Gregory K. Pregill In relation to their accessibility and their bio- logical interest, the Bahamas have been relatively poorly studied. The present collection of papers attempts to bring together all that is as yet known of one aspect of the natural history of the archi- pelago, namely, vertebrate paleontology. This is largely an outgrowth of the more extensive and detailed collection of fossils obtained by Smith- sonian personnel in 1978 from the only truly productive vertebrate fossil locality yet discovered in the Bahamas—a sinkhole on the island of New Providence. In addition to analyzing this collec- tion, we have reviewed the literature on Bahaman fossil vertebrates and, with the exception of mam- mals, we have re-examined and re-evaluated all of the previously reported fossil specimens. Par- ticularly in the case of birds, this has resulted in a number of taxonomic and nomenclatural changes. Furthermore, we have attempted to in- terpret our findings in the context of the more general picture of Pleistocene biogeography and paleoecology that has emerged from our recent studies in the Antilles. Yet this does not alter the fact that the fossil vertebrates of the Bahamas are Storrs L. Olson, Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, D. C 20560. Gregory K. Pregill, Natural History Museum, P.O. Box 1390, San Diego, California 92112. still poorly known. What we hope to provide is an updated baseline, a consolidated starting point, from which our knowledge can be ex- panded by future exploration. We would like to think that we have shown that the little details of anatomy, systematics, geology, and distribution that are the building blocks of studies such as those included here, eventually lead to significant generalizations of wider application. We will have succeeded if this volume stimulates additional research of this nature. The Bahamas form an extensive archipelago of low (maximum elevation 67 m), limestone islands and cays that are spread out over a distance of nearly 1000 km. The larger islands are almost all in the northern part of the archipelago, where the predominant vegetation type is pine forest (Pinus caribaea). Rainfall (data from Buden, 1979; Young and Cant, 1977) is greatest at the northern end of the archipelago, averaging 1300 to 1550 mm per annum on Grand Bahama and Great Abaco, respectively. On the islands of the Great Bahama Bank, rainfall ranges from 1295 mm per annum on New Providence, in the north, to 920 mm per annum on Long Island, in the south. The south- ern islands are quite dry (860 mm per annum on Mayaguana; 738 per annum on Great Inagua) and vegetation here consists mainly of xeric SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY broadleaf scrub. This rainfall gradient has impor- tant zoogeographical ramifications (see other pa- pers in this volume). The most important factor in any consideration of the faunal history of the Bahamas is the effect that eustatic fluctuations in sea level in the Pleis- tocene would have had on the archipelago. Al- though there is not full agreement on the extent of Pleistocene changes in sea levels, a maximum lowering of 120 m during the height of the Wis- consinan glaciation, 17,000 years ago, is a reason- able estimate (Gascoyne et al., 1979). This would have consolidated the 29 islands and 661 cays of the present-day Bahamas into five major islands and several smaller ones (Figure 1). In addition, a large island would have been created from the Cay Sal Bank, where there is now almost no dry land, and several islands would have formed from the Mouchoir, Silver, and Navidad banks south- east of Turks and Caicos, where there is no land at all. Land area in the Bahamas was increased by more than an order of magnitude, from 11,406 km2 at present, to about 124,716 km in the Wisconsinan (Buden, 1979). What is now the Great Bahama Bank constituted the bulk of the land area and was separated from the Cuban mainland by the 15-km-wide Old Bahama Chan- nel (Buden, 1979), thus presumably greatly facil- itating the dispersal of organisms from Cuba. At the close of the Wisconsinan glaciation, 12,000 to 13,000 years ago, sea levels began rising and much land was submerged, fragmenting the for- merly continuous large islands into many smaller ones. Today, most of the Great Bahama Bank is under 3 to 30 m of water. During the Wisconsinan glaciation, average temperatures in the Bahamas would have been cooler than at present by about 4° C (Lynts and Judd, 1971), and the fossil record provides evi- dence that climate and vegetation in general were altered as well (see discussions on pp. 18, 52). This evidence corroborates that from other is- FIGURE 1.—Map of the Bahamas showing the extent of the present banks and thus the approximate configuration of the islands during the last glaciation. NUMBER 48 lands, showing that in Wisconsinan times, the West Indies as a whole were more arid, and open savanna and scrubland habitats were much more prevalent than today (Pregill and Olson, 1981). Although the Bahamas are riddled with sink- holes and caves, such as usually provide ideal sites for the accumulation of fossil vertebrates, the islands have been relatively little explored paleon- tologically. Barbour (in Lawrence, 1934) and Hecht (1955) lamented that many deposits of cave earth have been obliterated for use as fertil- izer, but there must still be many unexploited sources of fossils awaiting discovery on the various islands of the Bahamas. Among the first fossil or subfossil vertebrates collected in the Bahamas were those obtained by Froelich Rainey in 1933-34 for Yale University. He obtained bones from Great Abaco, Eleuthera, Long, and Crooked islands. The material from Crooked Island came from one of the Gordon Hill caves along the northeast shore, where bones were obtained from a hard stratum under a layer of sand associated with charcoal, food remains, and artifacts indicating human occupancy. The specimens were "considered pre-Columbian though there is no definite criterion to establish the actual age" (Wetmore, 1938:51). Apart from introduced mammals, only the rodent Geocapromys ingrahami and remains of birds were recorded from this site (Lawrence, 1934; Wetmore, 1938). Only Geocapromys was reported from Great Abaco (Im- perial Lighthouse Cave, Hole in the Wall), Eleuthera, and Long Island (no precise localities given). Although Lawrence's (1934) account im- plies that these specimens were all derived from archeological deposits, Barbour's comments (in Lawrence, 1934) imply that Rainey may have been sampling genuine fossil deposits as well. Shortly after Rainey's discoveries, vertebrate fossils were forwarded to the Museum of Com- parative Zoology at Harvard University, suppos- edly from Great Exuma. Glover Allen (1937) discussed the remains of Geocapromys from this collection and Wetmore (1937b) identified the birds, among which were three species of raptors that he described as new. Bones of "reptiles, and frogs" were also included in this sample (G. M. Allen, 1937:370), but these were never studied and they cannot now be located in the MCZ collections (Ernest Williams, pers. comm.). This may have been the "mass of small miscellaneous bones" from which Williams forwarded speci- mens to Koopman (1951), who identified two mandibles of bats, one of which belonged to a species (Mormoops blainvillii) that no longer occurs in the Bahamas. Very little was originally recorded about the Exuma site other than that fossils "were sorted out by Mrs. Vivienne Knowles from cave earth excavated for use as fertilizer . . . during the early part of 1937" (Wetmore, 1937b;427). Although Wetmore referred to the fossils as having come from Great Exuma, G. Allen (1937) made refer- ence only to "Exuma Island." It was later deter- mined that these fossils were actually taken on Little Exuma, from a cave known as "Upper Pasture Cave" located "about fifty feet above sea level facing the western side of Little Exuma" (Hecht, 1955:134). Hecht and party found in 1953 that this and all similar sites in the vicinity "had been completely cleared of the fossiliferous cave soils." The presence of extinct species, such as the gigantic hawk Titanohierax, combined with the heavy mineralization of the specimens from the Little Exuma deposits, probably indicates a Pleistocene age, but as with most Antillean cave deposits, there is little direct evidence for dating. Hecht and Koopman's expedition to the Exu- mas in 1953 resulted in the collection of a few additional fossils (Hecht, 1955; Koopman et al., 1957). Remains of Geocapromys ingrahami were re- ported from "a small cave on the west side of Pigeon Cay just off Mosstown, Great Exuma" and in "Robertson cave near Williamstown, Lit- tle Exuma" (Hecht, 1955:134-135). Koopman et al. (1957) reported five bones of four species of bats from a cave near Forest Settlement, Great Exuma, and a mandible of an additional species from the cave near Williamstown, Little Exuma. For the sake of completeness, we shall mention two other papers that are of minimal paleonto- logical interest. The first of these is a study of SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY vertebrates from an Amerindian midden on San Salvador (Watling's) Island, which contained, in addition to fish and introduced mammals, the remains of a large iguanid lizard (Cyclura), a booby (Sula), and "a few individuals" oi Geocapro- mys ingrahami (Wing, 1969:26). Conklin (1971) reported 19 modern species of birds from a "Post- Columbian" cave deposit from "Abaco" contain- ing the bones of Rattus, as well as birds. It is probably misleading to refer to these specimens as fossils and the identifications of some of the specimens, particularly of small passerines, are very much in doubt in any case (p. 22). The most important fossil vertebrate locality in the Bahamas, and the only one yielding a true faunal assemblage, is Banana Hole, situated on the west end of New Providence Island. The entire fossil herpetofauna thus far known from the Bahamas, as well as the vast majority of specimens of fossil birds, were obtained from this site. New Providence is a low limestone island with an area of about 150 km2 that lies in the northwest central part of the Bahamas and is separated from the adjacent, much larger island of Andros by a deep trench, the Tongue of the Ocean. Although called Banana Hole by the original collectors for a "large banana tree growing in the sink" (Brodkorb, 1959:351), this designation is nowise distinctive. It is common practice in the arid, rocky Bahamas to cultivate banana trees in the rich, moist earth of sinkholes, and large sinks are therefore generally referred to as "banana holes." The Banana Hole, is a limestone sink about 6 m deep, with an opening nearly 12 m across. The hole is undercut on the western and northern sides for approximately another 12 m and it was in this sheltered portion that fossils occurred. Bones were first collected here in 1958 for the Florida State Museum by J. C. Dickinson, Jr., and Walter Auffenberg, who evidently sam- pled but a small portion of what was available. Brodkorb (1959) analyzed the fossil birds that they collected. The tortoise material and some of the fossil lizards from this sample were reported on by Auffenberg (1967) and Etheridge (1966), respectively. Additional collections were made in 1959-1960 by William Hanon, the birds from which are in the collection of Pierce Brodkorb. From 11 through 15 August 1978, Olson visited Banana Hole with Helen F. James, Frederick V. Grady, and Charles A. Meister, and collected still more material. At this time, a ladder was present that allowed easy access to the bottom of the sink. There were two large areas in the sheltered por- tion from which a layer of about 20 cm of soil had been removed, possibly for agricultural pur- poses, as the soil had not been dumped elsewhere in the sink. Olson's party was most successful in finding fossils by excavating pockets and crevices between large blocks of fallen rock, which were filled with loose, unstratified, powdery red soil. All soil was screened through 1/8 inch mesh and most of it through window screen as well. Large bones were removed on the spot. Later, the re- maining concentrate was washed, dried, and sorted. The fossils are buff to dark brown in color, some having a leached appearance, and a few showing signs of permineralization. The greatest number of vertebrate fossils were bones and teeth of Geocapromys. The concentra- tions of this relatively large mammal were almost certainly the result of predation by the giant extinct barn owl Tyto pollens. Smaller prey remains are presumably attributable in part to the mod- ern Barn Owl, Tyto alba. Fossils of both these predators are represented. The age of the Banana Hole deposits is gauged only by inference. Brodkorb (1959:368) reasoned that the "richer Pleistocene avifauna would seem to require a larger land mass" and therefore that the deposits must have been formed "during the early portion of the Wisconsin (fourth) glacial stage and preceding the mid-Wisconsin (Pamlico) temporary rise in sea level." The avifauna con- tains a number of species characteristic of arid or savanna-type habitats, such as no longer exist in the Bahamas. Evidence from fossil reptiles (Pre- gill, 1981) and birds (Pregill and Olson, 1981) indicate that similar conditions prevailed in the late Pleistocene of Puerto Rico. One of the older Puerto Rican fossil sites, that containing the most NUMBER 48 species of xerophilic vertebrates, has been tenta- tively radiocarbon dated at about 18,000 to 20,000 years before present (Pregill, 1981), which is at or just before the maximum extent of the Wisconsinan glaciation. While the fossil avifauna from Banana Hole indicates a more open, arid environment than at present, the herpetofauna suggests that extinction of the more xerophilous elements may already have taken place (p. 20), thus implying that the deposits formed after the Wisconsinan glacial maximum but before the close of the Pleistocene. By comparison with the reptiles and birds, the mammalian fauna of the Bahamas is quite de- pauperate. Because we do not otherwise deal with the fossil mammals, we shall use this opportunity to evaluate the status of the few taxa thus far known. Raccoons (Procyon lotor elucus) are known to have been introduced to Grand Bahama from Florida around 1932 or 1933 (Sherman, 1954), but a population found on New Providence was described from a single specimen as an endemic species, Procyon maynardi Bangs (1898). In his re- vision of Procyon, Goldman (1950) retained P. maynardi as a full species on the basis of three specimens, but noted that it was most similar to Procyon lotor incautus of the extreme southern Flor- ida keys. With seven additional specimens, Koop- man et al. (1957:164) found that the cranial and dental characters alleged by Goldman did not hold and they considered "maynardi as at best no more than a subspecies of P. lotor." Presumably the only distinction that can now be ascribed to maynardi is Goldman's assertion that it is darker than incautus. The fact that P. I. maynardi occurs only on the most densely populated island in the Bahamas but nowhere else on the Great Bahama Bank, and the fact that there are no fossils of Procyon in the rich material from Banana Hole, makes it very unlikely that Procyon can be re- garded as occurring naturally in the Bahamas. Until there is evidence to the contrary, we shall regard it as having been introduced. Thus, discounting Procyon and excepting bats, there is only one species of native land mammal known in the Bahamas, the endemic rodent Geo- capromys ingrahami J. A. Allen (1891). The genus Geocapromys is otherwise known only from the Cayman Islands (Patton, 1966; Morgan, 1977), Jamaica, Little Swan Island, and Cuba, whence the ancestors of G. ingrahami doubtlessly origi- nated. The nominate subspecies, G. i. ingrahami, is known only from tiny, arid East Plana Cay (near Acklins Island but not on the Crooked- Acklins Bank), where it survives today and is relatively abundant (Clough, 1972). On the basis of midden material from Crooked Island, Lawrence (1934) named a new subspecies, Geocapromys ingrahami irrectus, to which she referred specimens from Eleuthera and Long islands, al- though as we have mentioned, it is not certain whether these came from midden or fossil depos- its. G. M. Allen (1937) and Koopman et al. (1957) referred fossil specimens from Great and Little Exuma to G i. irrectus. Remains far in excess of 500 individuals of Geocapromys were collected from Banana Hole on New Providence, and Morgan (1977) mentions specimens at Florida State Mu- seum from Andros and Cat islands. Presumably these would be referable to the same taxon as the specimens from other islands on the Great Ba- hama Bank. The specimens of G. ingrahami re- ported by Wing (1969) from midden deposits on San Salvador were not identified to subspecies and were subsequently lost (G. S. Morgan, pers. comm.). San Salvador was never part of a larger island and was isolated from the island formed by the Great Bahama Bank during the Wiscon- sinan glaciation. Furthermore, Crooked Island, the type-locality of G. i. irrectus, is likewise part of a separate island system—the Crooked-Acklins Bank. A few specimens from Great Abaco, on the Little Bahama Bank, were designated by Law- rence (1934) as a third subspecies, Geocapromys ingrahami abaconis, but again it is not certain whether these were derived from midden or fossil deposits. The extent to which Amerindians trans- ported Geocapromys from one island to another is not known, so that in the absence of series of genuine fossils from each bank system, it would be difficult to evaluate the taxonomic status and SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY natural distribution of the various populations of G. ingrahami, even if consistent morphological dif- ferences could be demonstrated between popula- tions on different banks. What caused G. ingrahami to become extinct everywhere in the Bahamas except on East Plana Cay? To what extent was it affected by the late Pleistocene changes in habitats that caused the extinction of many other vertebrates in the Ba- hamas, and to what extent was it affected by human predation and human-wrought environ- mental changes? Again, the possibility of Amer- indian transport obfuscates our knowledge of events in the history of Geocapromys, because Amerindians could have carried the animals to islands where they had become extinct naturally. The fact that remains of Geocapromys occur in middens on a particular island does not necessar- ily mean that the rodents occurred there naturally at the same time or were even liberated there, as they could well have been raised in captivity. For example, the rodent Isolobodon portoricensis is known as a fossil only from Hispaniola. Although it is abundant in middens on Puerto Rico, whence originally described, it does not occur there as a fossil, nor in any noncultural context (Olson, pers. observ.). Furthermore, the same species, along with the flightless rail Nesotrochis debooyi, which is known from fossils in Puerto Rico, is found in middens in the Virgin Islands, including St. Croix (Olson, 1974). Because St. Croix is not part of the Puerto Rican Bank, it is unlikely that it would have had the same species of flightless rail as found on Puerto Rico. Thus N. debooyi must have been brought there by Amerindians. Not only did Amerindians clearly transport native animals from place to place in the West Indies, but they also appear to have practiced some form of ani- mal husbandry as well. Why is there only a single native species of terrestrial mammal in the Bahamas? In Cuba, the most likely source of colonizers, there were several different genera and species of insectivores and edentates, as well as a considerable radiation of capromyid rodents (Varona, 1974). The fact that Geocapromys has differentiated at best only at the subspecific level within the Bahamas is prob- ably an indication of the relative recency of its arrival, but there is no apparent reason why it should have arrived alone. A review of the extant and fossil bats of the Bahamas provides more useful information than the non-chiropterans. The fossil bats from Ba- nana Hole, New Providence, have not yet been studied in detail, but Karl Koopman, to whom we continue to be indebted, has identified speci- mens representing three significant range exten- sions among this material. Disregarding two living species that are known only from single accidental or mislabeled speci- mens, eleven species of bats are extant in the Bahamas (Koopman et al., 1957; Buden, 1975, 1977; Baker and Genoways, 1978). Four of these, Monophyllus redmani, Artibeus jamaicensis, Lasiurus borealis, and Tadarida brasiliensis, have not been reported as fossils in the Bahamas. Macrotus water- housii, Erophylla sezekorni, Eptesicus fuscus, Natalus lepidus, and N. tumidifrons are known from fossils from the Exumas (Koopman et al., 1957). Of these, only N. tumidifrons does not occur in the Exumas today. Two species of bats are known from the Baha- mas only as fossils. Mormoops blainvillii, endemic to the Greater Antilles, was found in the Little Exuma deposits (Koopman, 1951) and in Banana Hole, New Providence (Koopman, pers. comm.). Pteronotus parnelli, which occurs in the Greater Antilles and much of Central and South America, but has not hitherto been recorded from the Bahamas, was also found in the Banana Hole deposits (Koopman, pers. comm.). Monophyllus redmani and Brachyphylla cavernarum have only recently been taken in the Bahamas, and then only on islands south of the Crooked Island Passage (Buden, 1975, 1977). B. cavernarum, however, was found as a fossil in Banana Hole (Koopman, pers. comm.), suggesting that this species, and perhaps Monophyllus redmani as well, was once more widely distributed in the Bahamas and subsequently withdrew to the south. Natalus tumidifrons is endemic to the Bahamas, where it is known only from San Salvador and NUMBER 48 Great Abaco. It is absent from the islands of the Great Bahama Bank, where, however, it was found as a fossil on Great Exuma. Similarly, the present Bahaman distribution of the West Indian Red-bellied Woodpecker, Melanerpes superciliaris, includes only San Salvador, Great Abaco, and also Grand Bahama (where, however, it may now be extinct). It, too, is absent from islands of the Great Bahama Bank, but fossils of it are known from Little Exuma and New Providence. Its patchy distribution appears to have been caused by environmental changes (p. 45), and it is prob- able that the same is true of Natalus tumidifrons. We doubt that competitive replacement by the smaller species N. lepidus, as suggested by Koop- man et al. (1957), is a likely explanation. No such competitor exists for the woodpecker, yet its pat- tern of distribution is virtually the same as that of N. tumidifrons. Although the published fossil record of bats in the Bahamas is scant, there is evidence of the same patterns of extinction and fragmentation of range that are so strongly manifested among the birds (p. 54), and also of contractions in range southward, as has apparently occurred with rep- tiles (p. 20) and to a lesser extent with birds. Fossil Amphibians and Reptiles from New Providence Island, Bahamas Gregory K. Pregill Introduction As noted in the introductory paper (p. 3), the remains of frogs and lizards collected on Little Exuma Island were never described and are now apparently lost. Therefore, the only fossils of am- phibians and reptiles identified from the Baha- mas are specimens from Banana Hole on New Providence Island, which makes a full account of this material of particular importance. Etheridge (1966) reported seven species of lizards represent- ing five genera from the Banana Hole collection made by the Florida State Museum. Of these, only Cyclura no longer occurs on New Providence. Auffenberg (1967) recognized the tortoise fossils from the same collection as belonging to an ex- tinct form of Geochelone. When the Smithsonian expedition visited Ba- nana Hole in 1978, fossils were collected using fine-mesh screen, a technique that substantially increases the rate of recovery of small vertebrate remains. In this manner, the number of species in the herpetofauna was more than doubled, in combination with re-identification of specimens in the Florida State Museum collections. With the exception of tortoises, I examined all the existing fossils of amphibians and reptiles from Banana Hole. Seventeen species are treated here, with comments on taxonomy and morphology, and analyses of taphonomy, paleoecology, and zoogeography. The expanded list of species now Gregory K. Pregill, Natural History Museum, P.O. Box 1390, San Diego, California 92112. includes two frogs, four snakes, and several more lizards, including Aristelliger, a genus of gecko that today does not occur in the Bahamas north of the Crooked Island Passage. A fragment of a vertebra from a crocodile was also recovered, marking the first fossil occurrence of these animals in the Bahamas. ACKNOWLEDGMENTS.—This research was car- ried out during my tenure as a Smithsonian Institution postdoctoral fellow. Throughout the study I was privileged to tap the expertise of Ronald Crombie, George Zug, and Gary Morgan. Storrs L. Olson lent stimulating discussions and considerable comment on the manuscript, which was also read by Richard Etheridge, Jacques Gauthier, Linda Trueb, and Charles Woods. For the loan of fossil material I am grateful to S. David Webb of the Florida State Museum (UF) and Charles Scharff of the Museum of Compar- ative Zoology. All uncataloged material is depos- ited in the Florida State Museum. Most of the comparative skeletal material used in this study came from the National Museum of Natural History, Smithsonian Institution (USNM collec- tions). I thank Ernest Williams for his effort in attempting to locate the missing fossils from Little Exuma Island. I also thank Marianne Scott, who typed several drafts of the manuscript. Family LEPTODACTYLIDAE Eleutherodactylus cf. planirostris Cope MATERIAL EXAMINED.—87 ilia (USNM 283410, UF 3002); 3 maxillae (USNM 283411); 1 pre- NUMBER 48 maxilla (USNM 283411); 1 squamosal (USNM 283411); 16 mandibles (USNM 283412); 14 scap- ulae (USNM 283413, UF 3002); 21 vertebrae (USNM 283414). REMARKS.—Among the other remains of a small eleutherodactyline frog are numerous small ilia bearing round, dorsal protuberances. These are indistinguishable from ilia of Eleutherodactylus p. planirostris, which inhabits New Providence to- day. The largest fossil ilium is 10.6 mm in total length and 2.9 mm at its greatest posterior width, and came from a frog with an estimated snout- vent length of 32 mm. Schwartz (1960:23) gives a maximum length of 28 mm for E. p. planirostris on New Providence. The procedure for estimating snout-vent length of fossil individuals by extrap- olating from ratios of measurements made on modern specimens is fully described elsewhere (Pregill, 1981). Because ilia and most other skel- etal elements of West Indian Eleutherodactylus are very similar between species, I have only tenta- tively assigned the fossils to E. planirostris. This species, which is also found on Cuba and the Cayman Islands, is the only member of the genus presently inhabiting the Bahamas. Elsewhere in the West Indies, fossils oi Eleuth- erodactylus are known from the Cayman Islands (E. planirostris; Morgan, 1977), Barbuda (E. bar- budensis; Lynch, 1966) and Puerto Rico (Eleuther- odactylus sou.; Pregill, 1981). Family HYLIDAE Osteopilus septentrionalis Dumeril and Bibron MATERIAL EXAMINED.—76 ilia (USNM 283415, UF 2996, and 29 uncataloged); 25 maxillae (USNM 283416, UF 2996, and 4 uncataloged); 18 premaxillae (USNM 283417); 27 squamosals (USNM 283418, UF 2996); 15 sphenethmoids (USNM 283418, UF 2996, and 10 uncataloged); 1 parasphenoid (USNM 283418); 3 basicrania (UF 2996); 73 mandibles (USNM 283419, UF 2996); 60 scapulae (USNM 283420, UF 2996, and 25 uncataloged); 118 vertebrae (USNM 283421, UF 2996, and 62 uncataloged). REMARKS.—The osteology of Osteopilus septen- trionalis has been described in detail by Trueb and Tyler (1974). Ilia of 0. septentrionalis were espe- cially abundant in Banana Hole and were iden- tified by their having the preacetabular angle more obtuse than that of 0. brunneus of Jamaica or 0. dominicensis of Hispaniola. The cranial frag- ments were identified by their dermal sculptur- ing. The largest fossil individual had a snout-vent length of 98 mm, based on an estimate from an ilium. Trueb and Tyler (1974:11) give an average of 75.2 mm for females of this species (males to 55.3). Several specimens in the USNM alcohol collection exceed 90 mm, however. Osteopilus septentrionalis occurs throughout Cuba, the Cayman Islands, and the Bahamas, and has been introduced in Puerto Rico and Florida. As a fossil, it is known only from the Cayman Islands (Morgan, 1977). Nothing is known of its habits on New Providence, although elsewhere in its range this frog is encountered in a variety of mesophytic situations, breeding in any suitable body of water, from ponds and drainage ditches to other man-made entrap- ments. Family TESTUDINIDAE Geochelone sp. MATERIAL EXAMINED.—None. REMARKS.—Fossils of tortoises from the Baha- mas were referred to as Geochelone sp. by Auffen- berg (1967). All of this material was collected from Banana Hole, except for the distal end of a single pleural bone that was recovered from Mor- gan's Point, Andros Island, at a cite designated as "Cave 1." The fossils from Banana Hole consist of several miscellaneous elements. Fragments of a carapace and plastron came from an individ- uals) estimated to have had a carapace length of at least 600 mm (Auffenberg, 1967). Three cer- vical and three caudal vertebrae were described, but none was complete or retained any specifi- cally diagnostic features. A single femoral shaft, 10 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY the only limb element among the referred mate- rial, was described by Auffenberg (1967:41) as "proportionately stouter than the femoral shaft of either [Geochelone] denticulata or elephantopus, and is slightly less compressed than that oi denticulata." The only cranial bone in the sample was a right squamosal, which Auffenberg thought was dis- tinct from that of either G. denticulata or G ele- phantopus. He concluded that several individuals were represented by the Banana Hole fossils, but did not estimate the minimum number. Family GEKKONIDAE Aristelliger sp. Tarentola americana.—Etheridge, 1966:349 [Pleistocene, New Providence]. MATERIAL EXAMINED.—15 dentaries (USNM 283422, UF 10179-10180); 5 maxillae (USNM 283422, UF 10181); 2 premaxillae (USNM 283422). REMARKS.—A small sample of fossil dentaries and maxillae can be recognized as belonging to a large gecko by their shape and closely spaced, pointed teeth. The fossils are most similar in general proportions to Tarentola or Aristelliger, whereas the teeth in the other large West Indian gecko, Thecodactylus, are less pointed and more recurved. Etheridge (1966) referred some of these specimens to Tarentola americana, but re-examina- tion of them, along with the new material, sug- gests instead that they represent a species of Aristelliger. The most conspicuous feature of the dentaries is the wide groove separating the base of the teeth from the lingual edge of the crista dentalis. In Tarentola americana, this groove is nar- row and the base of the teeth are nearly flush with the lingual edge. In Aristelliger, and in the fossils, the groove is noticeably wider, being greater than the diameter of the tooth bases themselves. In addition, the teeth in the fossils and in Aristelliger are nearly cylindrical, whereas in Tarentola they are more anteroposteriorly com- pressed. Another similarity between the fossil den- taries and those oi Aristelliger is the short, wedge- shaped scar of the coronoid-overlap on the labial surface. In Tarentola the coronoid-overlap is longer and more lanceolate. The fossil premaxillae resemble those oi Aristel- liger in having a shorter nasal process and seven teeth or empty alveoli. The maximum number of premaxillary teeth observed in Aristelliger was nine (A. lar) whereas the minimum in Tarentola americana is eleven. The maxillae of Tarentola americana can be dis- tinguished from those of Aristelliger by the shape of the posterior edge of the nasal process, but this structure is missing from the two fossil maxillae, in which only the tooth-bearing portions remain. The posterodorsal edge is depressed and notched in Aristelliger in order to accommodate the pre- frontal in its participation in the nasal-frontal articulation. The prefrontal does not contact the nasal bone in Tarentola and the nasal process of the maxilla simply curves upward and back along its posterior edge. On the basis of size, the fossils could belong either to Aristelliger or to small individuals of Tarentola, as they came from individuals with snout-vent lengths between 80 and 85 mm. This is the approximate size of Aristelliger praesignis or A. hechti. In the Bahamas, Tarentola americana warreni achieves a snout-vent length of 92 mm, and on Cuba T. a. americana reaches 110 mm (Schwartz, 1968b). I am inclined to refer the fossils to Aristelliger on the basis of the qualitative characters cited above. If correct, some interesting zoogeographic problems are posed. Aristelliger does not now in- habit the Great Bahama Bank, but two species occur south of the Crooked Island Passage: A. barbouri on Great Inagua, and A. hechti in the Caicos Islands (see Schwartz and Crombie, 1975). On the other hand, Tarentola americana presently inhabits Cuba and the islands of Eleuthera, An- dros, the Exuma cays and the Long and Ragged groups, which are on the Great Bahama Bank. The only other published fossil occurrence of T. americana is a dentary from cave deposits from Camaguey Province, Cuba, identified by Koop- man and Ruibal (1955), who provided no descrip- NUMBER 48 11 tion or catalog number for the specimen. The other fossils in their fauna included Rattus and Mus, both of which are post-Columbian introduc- tions to the West Indies, suggesting that at least some of the deposit is essentially recent. However, among pre-Columbian fossils from a cave near Daiquiri, Oriente, Cuba, there are specimens of Tarentola americana, thus confirming that this spe- cies, the single New World representative of an otherwise Old World genus, is an indigenous member of the Antillean herpetofauna (Pregill, pers. obs.). Sphaerodactylus sp. MATERIAL EXAMINED.— 1 frontal (USNM 283423). REMARKS.—A single frontal bone is easily as- signed to a small gecko of the genus Sphaerodactylus because of its tiny size and closed olfactory canal, although it cannot be identified to species. The bone is 3.6 mm in total length, 2.2 mm across the supratemporal processes posteriorly, and came from an individual with an estimated snout-vent length of 32 to 35 mm. There are now four species of Sphaerodactylus on New Providence: S. copei, S. argus, S. nigropunctatus, and S. notatus. Schwartz and Thomas (1975) con- sidered S. copei to have been introduced from Hispaniola, and S. argus is thought to have been introduced from Jamaica (Thomas, 1975). Sphae- rodactylus notatus amaurus is the smaller of the two indigenous species, obtaining a maximum snout- vent length of 26 mm (Schwartz, 1966:172). Sphaerodactylus nigropunctatus reaches a maximum snout-vent length of 36 mm, making it the best guess as to the fossil's identity. It is surprising that these diminutive lizards even turn up in fossil owl pellets, but the New Providence specimen marks the second fossil rec- ord for the West Indies. Sphaerodactylus was found among cave remains from Puerto Rico, and, as in the present case, in very low numbers, only two individuals being represented (Pregill, 1981). Fos- silization in both these instances probably re- sulted from ingestion by an intermediate preda- tor. Family IGUANIDAE Anolis angusticeps Hallowell MATERIAL EXAMINED.—12 dentaries (USNM 283424); 1 maxilla (USNM 283424); 1 premaxilla (USNM 283424). REMARKS.—Identification of this small lizard in the Banana Hole herpetofauna is based pri- marily on a premaxillary bone. This element in Anolis angusticeps is distinguished from that of other Bahaman anoles by the low, flat profile that results from the nasal process being nearly horizontal to the dentigerous portion of the bone. In A. smaragdinus, A. distichus, and especially A. sagrei, the nasal process rises at a less acute angle. In A. smaragdinus, the process is bordered on either side by a ridge; in A. distichus the process is constricted medially. Dentaries of Anolis angusticeps are very similar to those of juveniles of A. smaragdinus because both species have elongate snouts. There are, however, subtle differences in tooth crown morphology; the crowns in A. angusticeps are slightly more pointed and, on the average, teeth are more numerous (26 vs 23). The largest fossil individual, estimated from a dentary, had a snout-vent length of about 40 mm. According to Schwartz and Thomas (1969), the modern subspecies on New Providence, A. a. olig- aspis, achieves its greatest size in males on Eleuth- era at 53 mm. The largest females reach 47 mm on South Bimini. Anolis a. oligaspis is not com- monly encountered on New Providence. It is a tree denizen, cryptic and shy like its conspecifics elsewhere in the Bahamas and Cuba. Anolis distichus Cope MATERIAL EXAMINED.—269 dentaries (USNM 283425, UF 2999, UF 10137, and 8 uncataloged); 116 maxillae (USNM 283426, UF 10135-10136, 12 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY and 3 uncataloged); 33 frontals (USNM 283427, UF 10138). REMARKS.—Numerous fossil jaws of a small anole are assigned to Anolis distichus on the basis of sharply pointed tricuspid tooth crowns. The tooth crowns of other Bahaman anoles are more blunt, especially on the central cusps. The frontal bones of A. distichus are more concave than those of other Bahaman anoles, although this is a sub- jective assessment, to be used with caution. The largest complete fossil dentary has a tooth row 7.5 mm long and came from an individual with an estimated snout-vent length of 50 mm. Schwartz (1968a:265) gives a maximum of 49 mm for males of A. d. distichus now inhabiting New Providence, although the maximum size varies between islands, and on Hispaniola the species may reach a snout-vent length of 58 mm. Anolis sagrei Dumeril and Bibron MATERIAL EXAMINED.—823 dentaries (USNM 283428-283429, UF 10149-10151, and 58 unca- taloged); 359 maxillae (USNM 283430, UF 10152-10154, and 38 uncataloged); 4 premaxillae (USNM 283430, 2 UF uncataloged); 206 frontals (USNM 283431, UF 10161-10162). (See Ether- idge, 1966:352, for additional material). REMARKS.—Remains of Anolis sagrei are the most abundant fossils in the Banana Hole her- petofauna, although identifying the various skel- etal elements is difficult. In general, the dentary of A. sagrei is robust, has a wide dental shelf, the tooth bases are essentially parallel-sided, and the teeth extend well above the parapet of the jaw. The shapes of the crowns are variable between individuals—some have well-developed lateral cusps, whereas others do not. The dentaries can be distinguished from those of A. smaragdinus or A. angusticeps on size and robustness, and from A. distichus by the less pointed teeth, but they are indistinguishable from those of other species, such as A. scriptus on the bank islands south of the Crooked Island Passage or A. homolechis on Cuba. Frontal bones are best identified by their flat surface and narrow interorbital width. A sample of 35 fossil dentaries had lengths ranging from 7.8 to 9.2 mm (estimated snout-vent lengths from 48 to 60 mm). Although the largest of 29 USNM alcohol specimens of Anolis sagrei ordinatus from New Providence has a snout-vent length of 54 mm, Etheridge (1966) reported one fossil dentary from Banana Hole with a tooth row of 11.0 mm, which would have come from an individual with a snout-vent length of about 77 mm. Apparently, individuals in the fossil popu- lation exceeded modern ones in size by 15 to 50%. Anolis cf. smaragdinus Barbour and Shreve Anolis carolinensis.—Etheridge, 1966:350 [Pleistocene, New Providence]. MATERIAL EXAMINED.—112 dentaries (USNM 283432, UF 10140-10142, and 4 uncataloged); 19 maxillae (USNM 283443, UF 10144); 3 fron- tals (USNM 283434). (See Etheridge, 1966:35, for additional material). REMARKS.—A number of upper and lower jaws of large size, with an elongate profile, and short, round teeth with blunt crowns are similar to Anolis smaragdinus of the A. carolinensis group of anoles. The dentaries differ from those of the modern species, however, in having the coronoid scar on the labial surface of the jaw not extending as far forward, the teeth slightly smaller, and on the average fewer in number (24 vs 21). These differences are difficult to interpret because of the strong similarity of A. smaragdinus to its allies A. carolinensis (southeastern U. S.), A. porcatus (Cuba), and A. brunneus (Crooked and Acklins Islands). These species do not seem to differ from one another osteologically any more than most Anolis differ individually within species. The New Prov- idence fossils are referred to A. smaragdinus simply because that species currently inhabits the island. The fossils are mainly from lizards with an estimated snout-vent length of 60 mm, the aver- age size for Anolis smaragdinus on New Providence. Anolis species indeterminate MATERIAL EXAMINED.—39 articular + suran- gulars (USNM 283435); 10 jugals (USNM NUMBER 48 13 283436); 2 postorbitals (USNM 283436); 3 basi- sphenoids (USNM 283437); 4 quadrates (USNM 283437); 57 pelves (USNM 283438); 65 vertebrae (USNM 283439). REMARKS.—Additional fossils of Anolis are un- identifiable to species, but presumably they be- long to those already discussed. They are included here for the sake of completeness. Cyclura sp. MATERIAL EXAMINED.—2 dentaries (USNM 283440); 2 articulars (USNM 283441); 2 frag- ments of maxillae (USNM 283441); 2 quadrates (USNM 283441); 1 parietal (USNM 283441); 1 vertebra (USNM 283441). (See Etheridge, 1966: 353, for additional material). REMARKS.—Etheridge (1966) referred several fossils of a large lizard from Banana Hole to the rock iguana, Cyclura, but the specimens were too fragmentary for specific determination. Unfortu- nately, the new material is scant and adds little to a resolution of taxonomic problems. A nearly complete dentary, lacking the coronoid and sur- angular processes, is 65.8 mm in total length and has a tooth row 53.5 mm long. The tooth crowns are typical of Cyclura in that the central cusp is compressed and flanked fore and aft by one or more smaller cusps; teeth from the middle and posterior portions of the row have two posterior and three anterior flanking cusps. Among species of Cyclura, there is considerable variation in the morphology of the tooth crown, although in the Bahaman forms C. cychlura, C rileyi, and C. cari- nata, and in the Cuban species C nubila, there are usually two lateral cusps on either side of the central one. The fossil dentary came from a lizard with an estimated snout-vent length of 425 to 450 mm. Cyclura cychlura is the largest of the three Bahaman species and the only one to exceed 400 mm in snout-vent length (females to 465 mm; Schwartz and Carey, 1977:39). Cyclura is not known histor- ically from New Providence, but C. cychlura occurs on adjacent Andros Island, and to the south on the Exuma cays. The West Indian rock iguanas have become extinct throughout much of their former range in both prehistoric and modern times. They now persist as relicts in dry areas that are protected from predators and human inter- ference. Leiocephalus cf. carinatus Gray MATERIAL EXAMINED.—17 dentaries (USNM 283442, UF 10175); 13 maxillae (USNM 283442, UF 10173, UF 10176); 1 premaxilla (USNM 283442); 4 frontals (USNM 283442, UF 10177). REMARKS.—The fossils of Leiocephalus from Ba- nana Hole agree with recent L. carinatus of Cuba and the Bahamas, but the specimens are too incomplete to permit certain assignment. One dentary is estimated to have come from an indi- vidual with a snout-vent length of 110 mm. Lei- ocephalus carinatus ranges in size from 73 mm snout- vent length on Cat Island to 130 mm on the Isle of Pines. Family TEIIDAE Ameiva auberi Cocteau MATERIAL EXAMINED.—3 dentaries (USNM 283443); 2 maxillae (USNM 283443); 3 pelves (USNM 283443). (UF material is listed by Eth- eridge, 1966:355) REMARKS.—Most skeletal elements of the ground lizard, Ameiva, are sufficiently distinct from those of other West Indian lizards as to be easily identified. Dentaries are recognizable at a glance by the short, slightly recurved bicuspid teeth with enlarged posterior crowns that may become triconodont posteriorly in some species. The largest fossil from Banana Hole has a tooth row 13.8 mm in length and came from an indi- vidual estimated at 95 mm snout-vent length. Males of Ameiva auberi thoracica now living on New Providence reach a maximum of 116 mm snout- vent length. Externally, Ameiva auberi is one of the most variable lizards in the West Indies. Eight subspecies are recognized in the Bahamas, and on Cuba there are 24 more (Schwartz and McCoy, 14 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 1970). Osteologically, however, Ameiva is conserv- ative, and differences even between species are usually quite subtle. Family TYPHLOPIDAE Typhlops sp. MATERIAL EXAMINED.—3 midbody vertebrae (USNM 283444). REMARKS.—Three very small vertebrae are re- ferred to the blind snake, Typhlops, because of their depressed neural arches, rudimentary neural spines, and well-developed accessory processes. It is difficult to estimate the size of the individual (s) from which the vertebrae came, but a snout-vent length of 250 mm is a reasonable approximation. Typhlops b. biminiensis occurs on New Providence and throughout the Great Bahama Bank from Bimini south to Little Ragged Island (Thomas, 1968). This subspecies is also known from Bahia de Guantanamo, Cuba. Typhlops lumbricalis occurs in the northern Bahamas, including New Provi- dence, and on Cuba and northern Hispaniola. I found no characters that would separate the ver- tebrae of these two species from one another or from the fossils. Fossils of T. biminiensis were re- ported from Cayman Brae (Morgan, 1977), and an unidentified species of Typhlops is known from cave remains on Puerto Rico (Pregill, 1981). Family BoiDAE Epicrates striatus Fischer MATERIAL EXAMINED.—2 dentaries (USNM 283445); 153 vertebrae (USNM 283446, UF 2995, and 133 uncataloged). REMARKS.—Two dentaries and numerous ver- tebrae are assigned to Epicrates because of their large size and general boid morphology. They are referred to E. striatus for no other reason than that the species presently occurs on New Providence. The bones are similar to those of the Puerto Rican boid, E. inornatus, which I have described in detail elsewhere (Pregill, 1981). The largest and most complete of the two fossil dentaries has a tooth row length of 19.5 mm, bears 18 teeth or empty alveoli, and came from a snake with an estimated snout-vent length of 1160 mm. The New Providence subspecies, E. s. striatus, attains a maximum snout-vent length of 1800 mm (She- plan and Schwartz, 1974:133). The vertebrae are from various regions of the vertebral column. The smallest midbody element has a centrum length of 3.0 mm, including the condyle; the largest is 8.8 mm. Five other verte- brae are about the same size as the latter; the remainder are small. Tropidophis can us Cope MATERIAL EXAMINED.— 1 dentary (USNM 283447); 70 vertebrae (USNM 283448, UF 3001, and 27 uncataloged). REMARKS.—Fossil vertebrae of the dwarf boa, Tropidophis, are distinguished from Epicrates by their much smaller size, proportionately longer neural spines, and elliptical, more anteriorly di- rected prezygapophyses. All the fossil vertebrae are nearly equal in size and have an average centrum length of 1.9 mm, being from individuals with a snout-vent length of about 250 mm. The fossil right dentary has a tooth row 5.0 mm in length and positions for 19 teeth. The bone is distinguishable from that of juvenile Epicrates by having the lingual shelf produced medially above a deep Meckelian groove. In small individuals of Epicrates, the lingual surface of the dentary is essentially flat. Tropidophis canus curtus now in- habits New Providence (also South Bimini and the Cay Sal Bank) and males, the larger sex, attain a snout-vent length of 283 mm (Schwartz and Marsh, 1960:82). Family COLUBRIDAE Alsophis vudii Cope MATERIAL EXAMINED.—34 vertebrae (USNM 283449, UF 3001, and 13 uncataloged). REMARKS.—Several fossil vertebrae are referred NUMBER 48 15 to Alsophis vudii because of their elliptical prezy- gopophyses, long, thin, neural spines, and their overall similarity to those of modern specimens from New Providence. There are two size classes of fossils. A single vertebra has a centrum length, including condyle, of 4.9 mm, and is 9.6 mm across the accessory processes. The individual represented by this bone had an estimated snout- vent length of 880 mm. The other size class includes vertebrae from the caudal and trunk region of a smaller snake(s) estimated at about 515 mm snout-vent length. I have no data on maximum size attained by Alsophis vudii through- out its range, but the largest of 16 specimens of A. v. vudii from New Providence in the USNM alcohol collection is 610 mm snout-vent length, a size exceeded by the larger fossil individual by over 40%. Family CROCODYLIDAE cf. Crocodylus MATERIAL EXAMINED.— 1 vertebral fragment (UF 2994). REMARKS.—This specimen is from the posterior trunk region of a small crocodile that had an estimated snout-vent length of 450 to 475 mm. The procoelous element has a total length of 15.5 mm and a condylar diameter of 9.3 mm. Most of the condyle is worn, and nothing remains of the arch laminae except for a small projection on the right side. The ventral surface is smooth and round. The fossil is too fragmentary for positive generic assignment, but Crocodylus is the obvious choice, as it is the only genus of crocodilians in the Antilles. There are two species in the Greater Antilles: Crocodylus rhombifer, a freshwater marsh inhabitant restricted to the Cienaga de Zapata on Cuba and the Cienaga de Lanier on the adjacent Isle of Pines; and Crocodylus acutus, which occurs in brackish and near-shore situations throughout most of Cuba, Jamaica, and parts of Hispaniola. Crocodiles are not known from the Bahamas except for adults that occasionally wan- der into surrounding waters (Neill, 1971:324). For years these waifs have encouraged speculation about local populations, as in the last century when Gardiner (1886) reported an eight foot crocodile shot near shore on Great Inagua Island. He expressed doubt that the animal was anything more than a stray, but did suggest that Andros might support indigenous populations because of the extent of fresh and brackish lagoons on the island. Juveniles of Crocodylus acutus are intolerant of salt water and the species breeds in brackish areas or marshes subject to sufficient run-off of fresh water. The fact that the fossil came from a juve- nile, or at least a very young adult, increases the possibility that crocodiles actually inhabited the Bahamas at one time. At present, the majority of Bahaman ponds and mangrove swamps appar- ently are too saline to support breeding popula- tions, but they may have been less so during times of lowered sea levels. Recently, a crocodilian tooth was found in a cave deposit on Grand Cayman Island, some 300 km south of Cuba (Morgan and Patton, 1979), further suggesting that crocodiles probably were more widespread in the Antilles in the not too distant past. Taphonomy A means of assessing the relative abundance of a fossil species is by determining the minimum number of individuals (MNI) present in the fauna. The estimate of MNI is made from the most abundant skeletal element of a species, with right and left components counted separately. The procedure is routine for lizards and frogs because they are represented amply by paired structures. Most snake fossils, however, are ver- tebrae, and distinguishing numbers of individuals is therefore difficult. For these animals, size is used with caution, in conjunction with any cra- nial bones available. Similarly, the Banana Hole tortoises are represented by shell fragments and vertebrae of "several individuals," which I have taken as representing a minimum of three from Auffenberg's (1967:41) descriptions. 16 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY From the analysis of the minimum number of individuals (Table 1), it appears that the preda- tors responsible for depositing the fossils may have been selective in their choice of amphibian and reptile prey. Out of the 17 species in Banana Hole, Anolis sagrei and Anolis distichus constitute more than 76% of the total number of individuals; A. sagrei alone accounts for nearly 57%. The two frogs, Eleutherodactylus ci. planirostris and Osteopilus septentrionalis, contribute almost 12%, Anolis ci. smaragdinus about 6%, and the remaining 12 spe- cies make up less than 6% of the total. These last 12 species are included in the Banana Hole assem- blage probably as much by chance or secondary ingestion as by predation by barn owls. For ex- ample, it is unlikely that the small, leaf-litter lizard, Sphaerodactylus, fell prey to either of the two barn owls (p. 37). Instead, the individual repre- sented as a fossil probably was consumed first by another predator, perhaps a snake, or a burrow- ing owl, which in turn was eaten by a barn owl. The tortoises probably fell into the sink acciden- tally. Adults of Cyclura and Epicrates probably TABLE 1.—Relative abundance of fossil amphibians and reptiles from Banana Hole, New Providence, Bahamas (MNI = minimum number of individuals, * = extinct on New Providence) Species MNI % of total Eleutherodactylus cf. planirostris 40 6.0 Osteopilus septentrionalis 39 5.9 Geochelone sp.* 3 0.4 Aristelliger sp.* 7 1.0 Sphaerodactylus sp. 1 0.1 Anolis angusticeps 8 1.2 Anolis distichus 134 19.9 Anolis sagrei 382 56.9 Anolis cf. smaragdinus 38 5.8 Cyclura sp.* 2 0.3 Leiocephalus cf. carinatus 8 1.2 Ameiva auberi 3 0.4 Typhlops sp. 1 0.1 Epicrates striatus 2 0.3 Tropidophis canus 1 0.1 Alsophis vudii 2 o.:'. cf. Crocodylus* 1 0.1 Total 672 100.0were not regular prey items because of their large size, but their preference for limestone crevices and outcrops may have made them more likely to be preserved. West Indian fossil owl pellets can potentially yield a remarkable sampling of the local herpe- tofauna during the time of deposition, as was demonstrated in Blackbone Cave on Puerto Rico (Pregill, 1981) and as seems to be the case here. There are no obvious absences from Banana Hole, as the fauna contains practically all species pre- sent on New Providence today in addition to extinct taxa. Species richness in a particular de- posit is a function of time as well as predator habits, because the longer a cave or sink hole remains in use by owls, the more likely it is that a given species will turn up in the pellet samples. A comparison of fossil and modern pellets in the West Indies indicates that the percentage of amphibians and reptiles consumed by owls was higher in the past. With the introduction of Rat- tus, the number of nonmammalian species in the diet has, in general, declined. For example, in an examination of modern pellets of Tyto alba from Crooked Island, Acklins Island, Mayaguana, and two of the Caicos (Providenciales and Middle Caicos), the only amphibian or reptile present was Anolis scriptus, represented by three individu- als from Mayaguana (Buden, 1974). These sam- ples had a total of 357 individuals of 19 species of rodents, bats, and birds, and of these, Rattus accounted for 76%. No amphibians or reptiles were found in the recent pellets of Tyto alba examined by Johnston (1974) from Grand Cay- man, and only a few specimens of Osteopilus sep- tentrionalis were recovered from similar pellets on Hispaniola (Wetmore and Swales, 1931). There is, however, one noteworthy exception to this trend towards a decline in amphibians and rep- tiles as prey. Anolis and Osteopilus made up over 75% of individuals of vertebrates in a large, mod- ern pellet deposit analyzed from the Cayman Islands (Morgan, 1977). Thus, the West Indian barn owls are, and apparently always have been, opportunistic feeders. That only two species dominate the Banana NUMBER 48 17 Hole herpetofauna is typical of the composition of other West Indian fossil assemblages studied so far (Pregill, 1981). Usually one or two species are far more abundant, despite the diversity of species that may otherwise be present. This is a function of prey abundance, size, conspicuousness, and general vulnerability. Although there are no pub- lished ecological data on Anolis for New Provi- dence, the same four species (A. sagrei, A. distichus, A. angusticeps and A. smaragdinus) have been stud- ied on South Bimini (Schoener, 1968). Anolis sagrei is the most common and terrestrial, occupying nearly all vegetational communities on the island. On this basis its dominance in the Banana Hole herpetofauna could be expected. Anolis distichus is a tree trunk dweller and exhibits the most ecolog- ical overlap with A. sagrei. Anolis angusticeps is a canopy dweller and the most ecologically re- stricted of the four; it is rather surprising that it occurred as a fossil at all. Apparently, the avian predators were feeding at night or near dusk. Ameiva and Leiocephalus, poorly represented by fossils, are diurnal species that sleep underground or in concealed places. On the other hand the anoles, also diurnal, sleep mostly on exposed perches. The Banana Hole material was carefully screened and sorted, thus permitting some discus- sion of skeletal completeness and the amount of recoverable bone. This is a useful exercise because of the differing diagnostic value of the various elements and because it illustrates additional as- pects of cave taphonomy. All skeletal components except limb bones were counted and identified. Limb bones were not included because of their usually fragmentary nature, and the difficulty of identifying them. The compound radius and ulna of frogs was the limb element most commonly found in one piece. Table 2 lists the quantity and relative abundance of all other skeletal elements. Dentaries of lizards are nearly three times as abundant as any other part of their skeleton and are also the most taxonomically useful structures. By contrast, the various bones of frogs occur at a somewhat more even frequency relative to one another in the sample, although over half of the TABLE 2.—Relative abundance of various skeletal elements of amphibians and reptiles from fossil owl pellet deposits in Banana Hole, New Providence, Bahamas Bone n % of total LIZARDS Dentaries 1213 57.5 Articular -1- surangular 39 1.8 Maxillae 471 22.3 Premaxillae 6 0.3 Frontals 220 10.4 Jugals 6 0.3 Other cranial 13 0.6 Girdles 66 3.1 Vertebrae 78 3.7 Total 2111 100.0 FROGS Maxillae 28 4.9 Mandibles 89 15.6 Sphenethmoids 15 2.6 Basicrania 3 0.5 Squamosals 28 4.9 Other cranial 20 3.5 Ilia 167 29.3 Vertebrae 146 25.7 Total 569 100.0 SNAKES Dentaries 2 0.5 Vertebrae 385 99.5 Total 387 100.0 specimens are ilia and vertebrae. Frog ilia are simple structures by contrast to lizard jaws and consequently are more difficult to interpret. Man- dibles and sphenethmoids are the anuran skull bones most often preserved intact, while roofing bones are usually fragmented. Except for two dentaries, all of the snakes are represented only by vertebrae. Relative skeletal completeness in owl pellets is a function of a predator's eating habits. North American barn owls (Tyto alba) fed on a diet of mice regurgitated over 80% of the skulls intact, and mandibles and femora were the most abun- dant skeletal elements returned (Dodson and Wexler, 1979). On the other hand, screech owls (Otus asio) and great-horned owls (Bubo virginianus) returned very few complete skulls. The small 18 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY vertebrates ingested by West Indian owls proba- bly were consumed relatively intact so that most of the skeleton was regurgitated. With time, the pellets break down and the elements become disassociated. Estimating the minimum number of individu- als of represented taxa, and the relative abun- dance of osteological components should be useful in subsequent faunal and chronological compar- isons between different cave deposits. Zoogeography The modern herpetofauna of New Providence is typical of that found on other islands of the Great Bahama Bank, from Bimini south to the Ragged Island chain. A few species that are erratically distributed or absent from the smaller islands and cays on the bank are present on larger ones such as Andros. Of the 19 extant species on the Great Bank, Cyclura cychlura, Anolis smaragdinus, and Chrysemys fe lis are endemic there, but eleven, possibly twelve species also occur on Cuba (Table 3). The uniformity of this herpetofauna and its association with Cuba is undoubtedly a reflection of the Great Bank having formed a single, large island that was much closer to Cuba during lowered sea levels of the late Pleistocene. Hence, most of the Great Bank species are only subspe- cifically different from Cuban forms. Compared to the Great Bank, the herpetofauna of the Little Bahama Bank is remarkably impov- erished, being composed essentially of a small assemblage of seven Great Bank species plus one endemic, Epicrates exsul. The island of Grand Ba- hama, on the Little Bahama Bank, has an area of 1136 km2 and seven indigenous species of am- phibians and reptiles: Osteopilus septentrionalis, Eleutherodactylus planirostris, Anolis sagrei, Leioce- phalus carinatus, Sphaerodactylus notatus, Typhlops lumbricalis, and Alsophis vudii. The Bimini Islands, located on the Great Bank immediately to the south across the Northwest Providence Channel, have an area of about 40 km2 and harbor the same seven species plus eight others: Ameiva auberi, Anolis angusticeps, Anolis distichus, Anolis smaragdinus, Sphaerodactylus nigropunctatus, Epicrates striatus, Tro- pidophis canus, and Typhlops biminiensis. South of TABLE 3.—Extralimital West Indian distributions of the extant indigenous amphibians and reptiles of the Great Bahama Bank (LB = Little Bahama Bank; R/S = Rum Cay (R) and San Salvador (S); SB = Southern Bahama Islands; Hisp = Hispaniola; x = present; dash = absent; ? = questionable record from Camaguey Province, Schwartz and Thomas, 1975:192) Species* Cuba LB R/S SB Hisp Other Eleutherodactylus planirostris X X s X Cayman Is. Osteopilus septentrionalis X X X X Cayman Is. Sphaerodactylus notatus X X - X S. nigropunctatus X - R - Tarentola americana X - - - Anolis angusticeps X - - - A. distichus - X - X A. sagrei X X X - Cyclura nleyi - X X Leiocephalus carinatus X X Cayman Is. Ameiva auberi X Typhlops lumbricalis X X - T. biminiensis X - X Cayman Is. Epicrates striatus - - X Tropidophis canus ■> - - X Alsophis vudii X - X * The species Anolis smaragdinus, Cyclura cychlura, and Chrysemys fe lis are not listed because they are endemic to the Great Bank. NUMBER 48 19 the Great Bank, the islands of the southern Ba- hamas, made up of the Crooked-Acklins Bank, Samana Cay, Mayaguana, Turks-Caicos banks and the Inaguas, have a combined area somewhat greater than the Little Bahama Bank and boast 26 species, 17 of which are endemic to one or more of the islands of the group (Table 4). In other words, the herpetofauna increases dramat- ically in number of species and degree of ende- mism as one progresses from north to south in the three major regions of the Bahaman archipelago. The number of species in each group is summa- rized as follows: Little Bahama Bank 8, Great Bahama Bank 19, Southern Bahamas 26. Even though the Little Bahama Bank was effectively cut off from direct colonization by Cuban pro- pagules, by the intervening Great Bank and two deep channels, land area and relative isolation alone will not account for this drastic cline from the northern to the southern islands. There is, however, an interesting inverse cor- relation between species diversity in each of these three regions and the amount of rainfall. The southern Bahamas receive just over half the an- nual precipitation of the Little Bahama Bank and have broadleaf scrub as the predominant vege- tation. The Little Bank is almost completely dom- inated by pine forest (Henry, 1974; Young and Cant, 1977). Pinus caribaea probably began invad- ing the Little Bahama Bank some time after pine forest returned to central Florida, about 13,000 years ago (Watts and Stuiver, 1980). Previously, TABLE 4.—Distribution of amphibians and reptiles on the main islands of the southern Bahamas (x = present, - = absent, I = introduced) Crooked- Maya- Turks- Extra- Species Acklins guana Caicos Inagua limital Eleutherodactylus planirostris - - X I X Osteopilus septentrionalis X - - X X Aristelliger barboun - - - X - A. hechti - - X - Sphaerodactylus caicosensis - - X - S. corticola X - - X S. inaguae - - - X - S. manguanae - X X - - S. notatus - - - I X S. underwoodi - - X - Anolis brunneus X - - - - A. scriptus x1 X X X - Cyclura carinata - x2 X - - C. rileyi X - - - X Leiocephalus greenwayi x3 - - L. inaguae - - - X - L. psammodromus - - X - - L. punctatus X - - - Ameiva maynardi - X Typhlops biminiensis - X X T. richardi - X - X Epicrates chrysogaster X - X X Tropidophus canus - - X X T. greenwayi - X - - Alsophis vudii X - X X Chrysemys malonei X - 1 West Plana Cay. 2 Booby Cay. 3 East Plana Cay. 20 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY during the late Pleistocene, a cooler (about 4° C below present), more arid climate is believed to have typified much of the West Indies (p. 2 herein and Pregill and Olson, 1981). At that time, the Bahamas probably consisted in part of vast, arid savannas and grassland, an interpretation rein- forced by fossils from Banana Hole of birds rep- resenting such xerophilous or open-country spe- cies as caracaras (Polyborus), thick-knees (Burhinus), and meadowlarks (Sturnella) (p. 54). Stands of broadleaf forest must have been present too, as indicated by the tree-dwelling lizard Anolis angus- ticeps, and such birds, now extinct on New Prov- idence, as the woodpeckers Colaptes and Mela- nerpes, the crows Corvus palmarum and C. nasicus, and the pigeon Columba squamosa (p. 54). If we assume that such habitats were distributed throughout all of the Bahaman islands, then the herpetofauna now inhabiting the dry southern islands, or xeric habitats in Cuba, would probably have had a wider distribution in the Bahamas in the late Pleistocene. If the rainfall gradient now observed in the archipelago began to establish itself with the retreat of the Wisconsinan ice sheet, then pine forest would have started replacing the xeric habitats on the northern islands first. As these are presently the wettest of the Bahaman islands, we could expect that more of the xeric- adapted members of the herpetofauna would have become extinct, thus resulting in the depau- perate herpetofauna observed on the Little Bank today. As mesophytic conditions moved south- ward, extinctions would have taken place on the Great Bank as well, whereas south of the Crooked Island Passage, where arid scrub still predomi- nates, the herpetofauna suffered less attrition and has remained more or less intact as a relict fauna. Unfortunately, there are as yet no fossils from the Little Bahama Bank with which to test this hypothesis. On the Great Bank, the fossils from New Providence offer some support, though rather niggling. All the species now occurring naturally on New Providence were recovered as fossils in Banana Hole (we may assume that one of the two species oi Sphaerodactylus and one of the two Typhlops were missed because of sampling bias). In addition, there were four extinct species in this deposit, for an extinction rate of slightly over 20%. Of these four taxa, two (Geochelone and Crocodylus) no longer occur anywhere in the Ba- hamas, but one (Cyclura) is still found elsewhere on the Great Bank. The remaining taxon, Aristel- liger, has its only Bahaman relatives south of the Crooked Island Passage, thus establishing that at least one presumed member of the southern Ba- haman herpetofauna actually did have a more northerly distribution in the Pleistocene. Without additional fossils, we cannot know if any of the other endemic amphibians and reptiles of the southern Bahamas were once part of a Great Bank assemblage. As theorized by Schwartz (1968c), several of these species were probably derived from Hispaniola and Puerto Rico and may never have occurred farther north. Two other vertebrates show positive evidence of a contraction of range southward in the Ba- hamas. The Cuban crow, Corvus nasicus, and the bat Brachyphylla cavernarum, occur in the Bahamas today only in the Caicos Islands. Yet both of these species were found as fossils on New Provi- dence (pp. 6, 46). This corroborates the evidence from the distribution of modern and fossil rep- tiles. Because we would expect reptiles to be good indicators of changes in climate, even at the microclimatic level, it seems somewhat puzzling that we observe only a 20% extinction in the herpetofauna of Banana Hole, whereas the rate of extinction in the avifauna is much higher at 50% (p. 53). Sampling bias could account for some of this; for example, the herpetofauna ap- pears to be nearly "complete" at 17 species, whereas the avifauna, with 32 species, is certainly not representative of the total number of species of birds present at the time of deposition. How- ever, certain of the birds, particularly some of the prime ecological indicators, are represented by very few individuals. If some of the species of reptiles were very rare on New Providence during the late Pleistocene, then it seems possible that they would have been less likely to be deposited or recovered as fossils than a comparably rare NUMBER 48 21 species of bird of the size of the caracara, thick- knee or meadowlark. Actually, if amphibians and reptiles are more sensitive to changes in climate than birds, then the Banana Hole fossils could be regarded as having been deposited after the her- petofauna had already undergone most of the extinction it was to suffer, but before climates had changed sufficiently to have altered the avi- fauna as drastically. The preceding discussion is admittedly based on precious little paleontological information, but if nothing else the ideas advanced here should emphasize the potential importance of the fossil record of vertebrates in the Bahamas and the need to expand our knowledge of this record in order to document further the effects that chang- ing climates and sea level had on populations of amphibians and reptiles. Fossil and Subfossil Birds from the Bahamas Storrs L. Olson and William B. Hilgartner Introduction Up to the present, three papers have dealt with significant fossil and subfossil avifaunas from the Bahamas. Pleistocene birds were reported by Wetmore (1937b) from Little Exuma, and by Brodkorb (1959) from New Providence. Bird re- mains from an archeological site on Crooked Island were discussed by Wetmore (1938). The geological and historical setting of these collec- tions are detailed elsewhere (pp. 3-5). Wetmore's (1937b) material from Exuma con- sisted of 30 specimens that he assigned to 13 species, 3 of which were described as new. The collection that Brodkorb (1959) studied from Ba- nana Hole on New Providence consisted of 65 specimens that were referred to 15 species, 6 of which were described as new. Because of the high number of supposedly extinct avian taxa and their zoogeographical significance, the Banana Hole fossil avifauna is among the most interesting yet reported from the West Indies. The Smithson- ian expedition of 1978 obtained 219 additional bird fossils from Banana Hole, which, together with 70 previously unstudied specimens in the collections of the Florida State Museum and of Pierce Brodkorb, yielded a total of 354 speci- mens—more than 5 times as many as were avail- able for Brodkorb's original study of the avifauna. By comparison with other Pleistocene deposits, such as some of those collected by Olson in Puerto Storrs L. Olson, Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. William B. Hilgartner, Irvine Natural Science Center, St. Timothy's School, Stevenson, Maryland 21153. Rico that yielded many thousands of bird bones, this is still a rather small sample, but it is none- theless quite a significant one. The collection from Crooked Island reported by Wetmore (1938) consisted of 21 specimens that were referred to 11 species. Although of no great antiquity, several of these specimens repre- sent significant range extensions for certain taxa. In addition to identifying and analyzing the newly collected material from Banana Hole, we also obtained and re-examined all of the previ- ously described specimens mentioned above, in- cluding all types. Our investigation did not reveal any new species but instead resulted in the alter- ation of taxonomic status of 6 of the 9 new species described by Wetmore (1937b) and Brodkorb (1959), as well as the correction of a number of misidentifications of specimens. We have disregarded the collection of bird bones from Abaco (presumably Great Abaco) reported by Conklin (1971) as "post-Columbian" on account of their association with remains of Rattus. First of all, these cannot be regarded as fossils, and secondly, many of the identifications do not seem credible. Passerines, mainly small species, constitute 87% of the sample, yet despite the manifest difficulties in identifying bones of small Passeriformes, Conklin assigned to species even such undiagnostic elements as shafts of long bones. Furthermore, 4 of the 10 nine-primaried oscines reported by Conklin are either rare, or at best uncommon, migrants to the Bahamas (Vireo solitanus, Protonotana citrea, Vermivora pinus, and Wilsonia citrina), whereas none of the resident 22 NUMBER 48 23 Parulidae, or such common species as Coereba flaveola, Vireo crassirostris, Geothlypis trichas, Den- droica palmarum, or D. discolor, were said to be represented. The faunal composition of Conklin's sample of bones seems highly unlikely for the Bahamas and in our opinion this report cannot be considered reliable. Unfortunately, Conklin's paper has already been cited by Buden (1981:2) as providing fossil evidence for a long occurrence of the tanager Spindalis zena on Great Abaco, but the recent age of Conklin's sample precludes this interpretation. That faunal remains from recent owl pellets may have some distributional signifi- cance, however, is attested to by Buden's (1974) discovery of two partial skulls oi Spindalis zena and the partial skull of the rail Porzana Carolina in Barn Owl (Tyto alba) pellets from the Caicos Islands, where neither species had been recorded previ- ously. ACKNOWLEDGMENTS.—The impetus for this study came from David Campbell, former direc- tor of the Bahamas National Trust, who indicated that further collecting at the Banana Hole local- ity would be fruitful and who suggested numerous valuable contacts in the Bahamas. Permits to collect fossils and modern comparative material were kindly provided by the Ministry of Agricul- ture, Bahamas, through Colin Higgs. In the field, Olson was expertly assisted by Frederick V. Grady, Helen F. James, and Charles A. Meister. Field work and Hilgartner's museum studies were supported by grants from the Fluid Research Fund, Smithsonian Institution, through S. Dillon Ripley. Grady's participation was made possible by funds from the Department of Paleobiology, through Martin A. Buzas. Ben Bohl and Rose Blanchard, International Field Studies, were in- strumental in providing Olson and Meister with facilities at Forfar Field Station while collecting birds on Andros. Pierce Brodkorb, University of Florida (PB), generously lent a collection of unstudied fossil birds from Banana Hole. The previously identi- fied material from this site, including all types, was kindly supplied by S. David Webb, Florida State Museum (UF). The fossils described by Wetmore from Exuma were made available by Raymond A. Paynter, Jr., Museum of Compar- ative Zoology (MCZ), Harvard University, and the archeological specimens from Crooked Island were supplied by Eleanor Stickney, Peabody Mu- seum of Natural History, Yale University (YPM). Skeletal material of Buteo nitidus was lent by Mary C. McKitrick from the University of Arizona collections. Line drawings are by Randall W. Blackburn, maps are by Sigrid James Bruch, and photographs are by Victor E. Krantz. Gary S. Morgan kindly provided information and refer- ences, and permitted us to use unpublished ma- terial from his master's thesis. We are grateful to Gregory K. Pregill, David W. Steadman, and Charles A. Woods for their comments on the manuscript and we are especially indebted to Mary H. Clench, whose knowledge of the Baha- mas made her detailed criticism of the manuscript of particular benefit. Species Accounts At the beginning of each account we have listed the identifications as given by Wetmore (1937b, 1938) and Brodkorb (1959), where appro- priate. When no such entry appears, the speci- mens constitute the first fossil record of a species for the Bahamas. The specimens of Pterodroma cf. hasitata, Buteo sp., Corvus palmarum, and Sturnella sp. also constitute first records for the Bahamas but are based in part on previously misidentified specimens. Minimum number of individuals was determined by the standard method of counting the most abundant element from a particular side. We have omitted this figure when it is obvious that only one individual was present. Family PROCELLARIIDAE Pterodroma cahow (Nichols and Mowbray) Pterodroma cahow.—Wetmore, 1938:51 [part; pre-Columbian midden, Crooked Island]. Puffinus puffinus.—Wetmore, 1938:51 [pre-Columbian mid- den, Crooked Island]. 24 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Puffinus Iherminieri.—Wetmore, 1938:51 [part; pre-Colum- bian midden, Crooked Island]. Columba leucocephala.—Wetmore, 1938:52 [part; pre-Colum- bian midden, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: cora- coid (part of YPM 2439), radius and ulna (part of YPM 2434), 2 carpometacarpi (YPM 2432), tibiotarsus (part of YPM 2433); minimum num- ber of individuals, 1. REMARKS.—Wetmore (1938) referred a radius, ulna, and carpometacarpus to this species, but the last is too large for P. cahow and belongs instead to a species of Pterodroma the size of P. hasitata (see below). The two carpometacarpi he identified as Puffinus puffinus are also referable to Pterodroma cahow and may be distinguished from those of Puffinus puffinus by the following charac- ters: (1) pisiform process more widely separated from pollical facet, nearly parallel with interme- tacarpal space; (2) caudomedial margin of carpal trochlea more truncate, not decidedly rounded; (3) fossae around pisiform process faint or lacking. Bond (1956:188) attributes osteological re- mains oi Puffinus puffinus to St. Croix, as well as to Crooked Island, but we have been unable to find any other mention of this record in the literature and we suspect that an error may have arisen through confusion with Puffinus Iherminieri, which was twice reported from middens on St. Croix (Wetmore, 1918; 1937a). Wing et al. (1968) re- corded the distal portion of a humerus from an archeological site on Antigua as Puffinus cf. puffi- nus, but the specimen cannot now be located (Wing, pers. comm.) and we were not able to check the identification. Thus, there appears to be no verifiable record of Puffinus puffinus, prehis- toric or otherwise, from the West Indies. Of the two bones Wetmore assigned to Puffinus Iherminieri, the tibiotarsus actually belongs to P. cahow, as shown by the much smaller inner cne- mial crest and more terete shaft. We can offer no explanation for Wetmore's inclusion of a coracoid of P. cahow with three other specimens correctly identified as Columba leucocephala, this element being utterly different in these two species. The error appears not to be one of subsequent mis- handling, as there is no coracoid of Columba leu- cocephala among the Crooked Island bones and the specimen is correctly numbered. The Cahow is known to breed only on Ber- muda, where it once occurred in vast quantities but is now nearly extirpated. It is smaller than Pterodroma hasitata (Table 5) or P. caribbaea, the only congeners that breed in the West Indies. Virtually nothing is known of its range at sea and the Crooked Island material is the only indication of the species away from the known breeding grounds. Pterodroma cf. hasitata (Kuhl) Pterodroma cahow.—Wetmore, 1938:51 [part; pre-Columbian midden, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: car- pometacarpus (part of YPM 2434). REMARKS.—This specimen was included among those that Wetmore (1938) referred to P. cahow, but it is too large for that species (Table 5). The Black-capped Petrel is known to breed in mountainous areas of Hispaniola, Guadeloupe, and Dominica. Very recently it was discovered offshore and flying about sea cliffs in the Sierra Maestra region of southeastern Cuba (Bond, 1978; Nicasio Vina, Academia de Ciencias, San- tiago de Cuba, pers. comm.). An all dark form, P. caribbaea Carte from Jamaica, thought to be ex- tinct, is variously regarded as a separate species, or as a subspecies or color phase of P. hasitata. These represent the only known resident popula- tions oi Pterodroma in the West Indies. Wetmore (1918) assigned a tibiotarsus (USNM 225842) from a midden on St. Croix to this genus but did not identify it to species in the absence of comparative material. A fragmentary proximal end of a humerus (USNM 428289) from a midden in Martinique he referred to P. hasitata (Wetmore, 1952). We examined these specimens and con- firmed their generic placement. As with the car- pometacarpus from Crooked Island, these speci- mens can be referred to the species P. hasitata only on the basis of size and geographic probability, NUMBER 48 25 TABLE 5.—Measurements (mm) of Pterodroma to show the size of Crooked Island specimens compared with a series of fossils of P. cahow from Bermuda and skeletons of P. hasitata from Hispaniola Co Tiparat ve material Crooked Island specimens Character P. cahow P. hasitata P. cahow P. hasitata Range Mean n Range Mean n Range n Range n Coracoid Length 22.5-23.7 23.2 15 26.4 1 22.3 1 - Ulna Length 87.4-91.7 88.9 10 102.2-104.4 103.3 2 86.5 1 - Proximal 8.1- 9.0 8.6 10 9.3-10.0 9.6 2 8.8 1 - depth Distal 6.5- 7.4 6.9 10 8.1 l 7.0 1 - width Carpometacarpus Length 40.0-44.0 41.9 15 50.1-50.4 50.3 2 41.9-43.2 2 48.8 1 Proximal 9.0-10.4 9.7 15 10.7-11.8 11.3 2 10.0-10.3 2 11.4 1 depth Tibiotarsus Length 52.8-55.8 54.3 12 61.7 1 52.8 1 - Distal 4.6- 5.1 4.8 12 6.2 1 4.9 1 - width as they are not otherwise sufficiently diagnostic to permit specific identification in the complex genus Pterodroma. The Crooked Island specimen is the first indi- cation of the species in the Bahamas. Elsewhere in the West Indies, it presently requires high mountainous terrain or sea cliffs for nesting, but this may be an artifact of human intervention, as the closely related P. cahow in Bermuda does not. It is difficult to imagine that the former human inhabitants of Crooked Island would have been able to capture these birds at sea, although the presence of P. cahow in the same deposits may be indicative of some such capacity unless both taxa were picked up as hurricane wrecks. It seems equally improbable that both species bred sym- patrically in the Bahamas, but the possibility cannot be ruled out. Puffinus Iherminieri Lesson Puffinus Iherminieri.—Wetmore, 1938: 51 [part; pre-Colum- bian midden, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: com- plete carpometacarpus (part of YPM 2433). REMARKS.—This specimen, which is from a juvenile, was correctly identified by Wetmore (1938), although the tibiotarsus he associated with it was not (see Pterodroma cahow). Audubon's Shearwater is a circumtropical species that breeds, or formerly bred, in a number of localities in the West Indies, including the Bahamas, where it still occurs on suitable cays and in adjacent waters. Family SULIDAE Sula leucogaster (Boddaert) Sula leucogaster.—Wetmore, 1938:52 [pre-Columbian midden, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: 1 cor- acoid and 1 distal end of humerus (YPM 2435). REMARKS.—The circumtropical Brown Booby breeds in the Bahamas and elsewhere in the Caribbean. 26 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Family ARDEIDAE Nyctanassa violacea (Linnaeus) MATERIAL EXAMINED.—New Providence: 1 fur- cula (UF 25642); 1 distal end of ulna (USNM 283200). REMARKS.—The distal end of an ulna of this species p :obably cannot be safely separated from that of Nycticorax nycticorax, but the entire furcula is easily distinguished by the narrower shaft and longer and more slender symphysis. The Yellow- crowned Night Heron is found throughout the West Indies and is common in the Bahamas. Family THRESKIORNITHIDAE Eudocimus albus (Linnaeus) Guara alba.—Wetmore, 1938:52 [pre-Columbian midden, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: 1 dis- tal end of ulna, 1 distal end of tibiotarsus (YPM 2436). REMARKS.—The White Ibis occurs only occa- sionally in the Bahamas at present, probably as a wanderer from the mainland, although the species breeds in the Greater Antilles. Family ACCIPITRIDAE Accipiter striatus velox (Wilson) Accipiler striatus velox.—Wetmore, 1937b:428 [Pleistocene, "Great Exuma" = Little Exuma]. MATERIAL EXAMINED.—Little Exuma: 1 distal end of humerus (MCZ 2255). New Providence: 1 proximal end of carpometacarpus (USNM 283232). REMARKS.—Our new specimen from New Providence corroborates Wetmore's assignment of the Exuma fossil to the subspecies of the North American mainland. Both fossils are the size of females (the larger sex) of A. s. velox and are thus much larger than the three diminutive subspecies of Accipiter striatus endemic to Cuba, Hispaniola, and Puerto Rico. A. s. velox is now only a rare winter visitor to the West Indies, including the Bahamas. The discovery of fossils of an Accipiter of this size on both Exuma and New Providence is a fairly good indication that this hawk was once either resident, or of more regular occurrence, in the Bahamas than at present. Buteo sp. Calohierax quadratus Wetmore, 1937b:429 [Pleistocene, "Great Exuma" = Little Exuma]. Calohierax quadratus.—Brodkorb, 1959:351 [Pleistocene, New Providence]. MATERIAL EXAMINED.—Little Exuma: 1 distal end of tarsometatarsus (MCZ 2256, holotype). New Providence: 1 distal end of tibiotarsus (UF 3152). REMARKS.—Wetmore (1937b:429) founded a new genus and species of hawk from Little Ex- uma, Calohierax quadratus, on the distal portion of a tarsometatarsus, lacking the outer trochlea. This was diagnosed as being similar to Buteo "but with the projecting outer portion of the second [inner] trochlea much reduced so that the trochlea ap- pears square and block-like." Examination of the holotype with a dissecting microscope revealed that this supposed character is due entirely to wear, the medial edge of the inner trochlea being clearly abraded. This specimen cannot, therefore, be separated generically from Buteo. The same is true of the distal end of a tibiotar- sus from New Providence that Brodkorb (1959: 351) referred to Calohierax quadratus on account of its having the "raised edges of peroneal groove forming ridges, not shelves; internal rugosity of oblique ligament situated slightly lower on shaft." These characters are variable, both between and within, species oi Buteo. In addition, the specimen is quite worn and has little diagnostic value be- yond the generic level. We regard Calohierax Wet- more 1937 as a junior synonym oi Buteo Lacepede 1799. It is difficult to determine the identity of the two fragmentary Bahaman fossils within the com- NUMBER 48 27 plex oi Buteo and its relatives. On the basis of size alone, the corresponding bones oi Buteo jamaicensis and Buteogallus anthracinus are larger than the fossils, and those of Buteo platypterus, B. magnirostris, and B. ridgwayi are smaller. The middle trochlea in the fossil tarsometatarsus is shorter and broader than in Parabuteo unicinctus, and in distal view does not have the external rim extending farther pos- teriorly as in that species. The fossils are within the size range of Buteo brachyurus or B. albonotatus, but the hindlimb elements in B. brachyurus are more robust than in the fossils, and in B. albono- tatus the ends of the bones are wider but the shafts narrower than in the fossils. The two Bahaman specimens are very similar in size and details to both Buteo lineatus and B. nitidus. Although the tarsometatarsus is relatively longer in B. lineatus than in B. nitidus, the distinction cannot be made in the fossil because of its fragmentary condition. Buteo lineatus would be more likely on geographi- cal grounds, but the nature of the Pleistocene environment in the Bahamas would not have been incompatible with the occurrence of a xe- rophilous species, such as B. nitidus. Without fur- ther fossil material, the specific identity of the Bahaman Buteo will have to remain undeter- mined. The existing specimens nevertheless doc- ument the presence of a species that has become extinct in the Bahamas. Titanohierax gloveralleni Wetmore Titanohierax gloveralleni Wetmore, 1937b:430 [Pleistocene, "Great Exuma" = Little Exuma]. MATERAL EXAMINED.—Little Exuma: 1 tarso- metatarsus lacking proximal end (MCZ 2257, holotype), 1 proximal end of carpometacarpus (MCZ 2258). New Providence: 1 nearly complete and 1 shaft of ulnae (UF 25640-25641); mini- mum number of individuals, 1. REMARKS.— Titanohierax gloveralleni is a valid taxon of very large Accipitridae hitherto known only from the two specimens from Little Exuma. Among some unidentified material from Banana Hole in the collections of the Florida State Mu- seum, we found three fragments of shaft that fit together to constitute most of a right ulna, lacking the ends. With these was a fragment of a shaft of a left ulna. We estimate the total length of the more complete specimen to have been about 210 mm; the greatest diameter at midshaft is 10 mm. These specimens are too large for any of the other species of birds known from Banana Hole and their structure conforms well with that in the Accipitridae; therefore we have tentatively re- ferred them to T gloveralleni, thus extending its known range to New Providence. Titanohierax is not related to the eagles of the genera Aquila or Haliaeetus. Instead it is a gigantic buteonine hawk that Wetmore (1937b) consid- ered to be most similar to the much smaller species of Hypomorphnus (=Buteogallus). We found as much or more similarity between Titanohierax and the South American genus Geranoaetus, how- ever. Recently, Campbell (1979) erected a new genus, Amplibuteo, for two species of very large, late Pleistocene accipitrids: Amplibuteo hibbardi Campbell from tarpits in Peru, and Amplibuteo (Morphnus auct.) woodwardi (L. Miller) from the Rancho La Brea tarpits in southern California. He regarded Amplibuteo as being most closely related to Geranoaetus. Although Campbell (1979: 83) stated that Amplibuteo "is quite distinct from . . Titanohierax" he did not elaborate further and gave no indication that he had actually examined specimens of Titanohierax in his comparisons. A possible relationship between Titanohierax and Amplibuteo should be explored further. There are additional remains, probably of Ti- tanohierax, from elsewhere in the West Indies. A giant "eagle" was described as Aquila borrasi by Arredondo (1970), from two Pleistocene cave de- posits in Havana Province, Cuba. Part of the material of this species had originally been sent to the late Bryan Patterson at the Museum of Comparative Zoology, who intended to collabo- rate with Arredondo in describing it. Unfortu- nately, Patterson never resumed his correspond- ence on the matter (Arredondo, 1972:2) and Ar- redondo was constrained to proceed with the description on the basis of the specimens still 28 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY available to him. He placed the species in the genus Aquila, in its classical sense, owing to the lack of opportunities to examine extensive com- parative material. Olson recently examined Pat- terson's notes and the specimens upon which they were based: two distal portions of tibiotarsi, a femur lacking the distal end, and some pedal phalanges. The two tibiotarsi are quite different in size, and the smaller one is so incomplete that no meaningful measurements could be taken of it. The larger one measures 36.1 mm in width across the condyles. If these two specimens rep- resent a single species, then there must have been considerable sexual dimorphism in size. The holo- type of A. borrasi is a tarsometatarsus lacking the distal end. It has a proximal width of 22.4 mm, which seems rather small, even for the smaller of the two tibiotarsi just mentioned. It is nevertheless from a very large accipitrid, but one with a tarsometatarsus more gracile than that of Aquila and more similar to that of Titanohierax. Patter- son's notes indicate that he regarded the Cuban bird as a relative of Hypomorphnus (= Buteogallus) and that it was probably a species of Titanohierax. The species is clearly not referable to Aquila and we believe that it should be listed as Titanohierax borrasi (Arredondo) until there is sufficient fossil material to decide questions of species-level tax- onomy within Titanohierax. That Titanohierax also occurred on Hispaniola is demonstrated by a tarsometatarsus lacking the proximal end (USNM 244573) collected by Charles Woods under 60 cm of red earth in Cueva de las Abejas, 6 km SE of Pedernales, Dominican Republic, 24 April 1978. Olson compared this with the holotype of T gloveralleni and found it to be somewhat smaller and less robust, with a less distinctly grooved middle trochlea, and a less pronounced lateral rim of the outer trochlea in distal view. The total length of this specimen would have been about 135 mm, roughly com- parable to that of the holotype of T borrasi. Its specific identity cannot yet be determined and it should be listed only as "Titanohierax sp." Further evidence that Titanohierax may have been widespread in the West Indies is offered by two mandibular symphyses from a cave on Grand Cayman that were tentatively referred to T glov- eralleni by Morgan (1977) on account of their great size and similarity to Buteogallus. In spite of the taxonomic problems associated with Titanohierax at the generic and specific levels, the fossil record shows that giant hawks once existed in the Bahamas, Hispaniola, Cuba, and Grand Cayman, where no such birds exist today. Their extinction may be attributable to the same factors that caused the extinction of other large raptors and mammalian elements of the "mega- fauna" in the West Indies. At least in the Baha- mas, Titanohierax could not have survived the virtual extirpation of Geocapromys, which must have been its principal prey (see account of Tyto pollens, p. 36). Family PANDIONIDAE Pandion haliaetus (Linnaeus) Pandion haliaetus.—Wetmore, 1938:52 [pre-Columbian mid- den, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: 1 dis- tal end of tibiotarsus (YPM 2437). REMARKS.—The osprey is found throughout the Bahamas today and is resident on most is- lands. Family FALCONIDAE Polyborus creightoni (Brodkorb) Caracara creightoni Brodkorb, 1959:353 [Pleistocene, New Providence]. Polyborus plancus.—Olson, 1976a:363 [holotype of C. creightoni restudied]. MATERIAL EXAMINED.—New Providence: 1 ma- jor metacarpal (UF 3153, holotype of Caracara creightoni), 1 distal end of tibiotarsus (USNM 283281), 1 quadrate lacking orbital process (USNM 283289). REMARKS.—The species P. creightoni was based on a fragment of a carpometacarpus that pre- NUMBER 48 29 FIGURE 2.—Comparison of modern and fossil caracaras (Polyborus): A, distal end of left tibiotarsus of P. plancus, cranial view; c, same, distal view; B, distal end of left tibiotarsus of P. creightoni, cranial view (USNM 283281) (arrow indi- cates broken bridge that delim- its the second and third tendi- nal openings); D, same, distal view (note wider intercondylar sulcus); E, left quadrate of P. plancus, medial view; G, same, ventral view of mandibular ar- ticulation; F, left quadrate of P. creightoni, medial view (USNM 283289); H, same, ventral view of mandibular articulation. (Scale = 1 cm.) 30 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY TABLE 6.—Measurements (mm) of the two new specimens of Polyborus creightoni compared with the modern forms of Polyborus plancus and the extinct species P. lutosus of Guadalupe Island, Mexico (quadrate measurements are as follows: A = length from squamosal articulation to medial knob of mandibular articulation; B = transverse (lateromedial) diameter of mandibular articulation; C = depth (craniocaudal diameter) of mandibular articulation) Character P. pl incus P. lutosus Mean* P. creightoni Range Mean n Mean* Quadrate A 15.3-18.5 16.8 18 17.2 20.9 B 9.7-12.8 11.3 18 11.5 14.0 C 5.3- 6.4 5.7 18 5.7 7.7 Tibiotarsus Shaft width at proximal 6.1- 8.7 7.1 13 7.5 8.5 end of ligamental scar Depth of inner trochlea 9.5-12.4 10.8 13 10.7 12.5 Width through trochleae 12.0-15.0 13.2 13 13.7 14.2 + (abraded) n = 1. served almost no diagnostic characters. Olson (1976a) examined this specimen and adventured to agree that it was from a caracara but consid- ered the evidence insufficient for the recognition of a species distinct from P. plancus, which at present occurs in Cuba, southern Florida, and parts of the southwestern United States south to Tierra del Fuego. However, the two additional specimens from Banana Hole provide definite proof, not only of a caracara in the Bahamas, but also that this bird was indeed distinct from exist- ing species. The distal end of a tibiotarsus (Figure 2c,d) shows the three openings diagnostic of most Fal- conidae. It is near the maximum size for Polyborus plancus (Table 6), from which it differs in having a markedly broader intercondylar sulcus (Figure 2d). The most diagnostic specimen is a left quadrate (Figure 2g,h) that is considerably larger than in any of the specimens of P. plancus examined (Ta- ble 6). It also presents the following qualitative differences: (1) mandibular articulation in ventral view much broader craniocaudally, with the cau- dal flange conspicuously expanded and the lateral projection relatively narrower; (2) pneumatic for- amen at medial base of orbital process larger; (3) area on medial surface, between and ventral to the two heads of the otic process, much more deeply excavated. The holotypical carpometacarpal fragment of P. creightoni is near the minimum size of that element in P. plancus (Olson, 1976a), whereas the tibiotarsus and the quadrate from Banana Hole equal or exceed the maximum size in P. plancus. If these three specimens are from the same species (and they could well be from the same individ- ual), then it would appear that P. creightoni was a large caracara with reduced wings. There is a possibility that P. creightoni could prove synonymous with P. latebrosus Wetmore (1920) from the Pleistocene of Puerto Rico, but P. latebrosus is as yet based only on two undiagnos- tic fragments of wing bones that Olson (1976a) also considered to be insufficiently differentiated from P. plancus to merit specific recognition. The new material of P. creightoni shows that there definitely was at least one species of Polyborus endemic to the West Indies in the Pleistocene, and, therefore, the conspecificity of P. latebrosus with P. plancus must now be regarded as uncer- tain. Caracaras of the genus Polyborus are largely terrestrial scavengers that characteristically in- NUMBER 48 31 habit open brushlands, savannas, and prairies. The presence of a caracara in the Banana Hole deposits is thus indicative of such habitats in the Pleistocene of the Bahamas. Falco sparverius Linnaeus MATERIAL EXAMINED.—New Providence: 1 dis- tal end of tibiotarsus (USNM 283265). REMARKS.—The American Kestrel is an un- common winter visitor from the North American mainland (F. s. sparverius) to most of the main islands of the Bahamas. Paulson (1966) showed that the Cuban race, F. s. sparverioides, breeds at least on Great Inagua and San Salvador. Resident forms of F. sparverius, divided into three subspe- cies, are found virtually throughout the remain- der of the Antilles. In the fossil, the shaft is narrower than in six specimens of F. s. dominicensis or one specimen of F. s. caribaearum that we ex- amined, and it more closely resembles F. s. spar- verius. We did not examine skeletons of the Cuban race F. s. sparverioides, however, and the single fragmentary fossil is probably insufficient for sub- specific identification in any case. Family RALLIDAE Porzana Carolina (Linnaeus) MATERIAL EXAMINED.—New Providence, Ba- nana Hole: 1 distal end of humerus (PB 9043). REMARKS.—In the Bahamas, the Sora Rail is a fairly common visitor during the winter months. Family SCOLOPACIDAE Capella sp. Capella delicata.—Wetmore, 1937b:435 [Pleistocene, "Great Exuma" = Little Exuma]. MATERIAL EXAMINED.—Little Exuma: 2 com- plete coracoids, 1 complete humerus (MCZ 2260); minimum number of individuals, 1. New Providence: 1 complete coracoid, 1 complete and 1 distal end of humerus, 1 proximal end of ulna, 2 distal ends of tibiotarsi (USNM 283266-283269, PB 9044-9045); minimum number of individuals, 2. REMARKS.—Wetmore (1937b:435) assigned the fossils he examined to Capella [gallinago] delicata with "some mental reservation," as they were larger than comparative material available to him then. With additional fossils from Banana Hole and more comparative material, we still find this to be the case. All of the fossils are larger (Figure 3) than C. gallinago delicata or C g. para- guaiae (Table 7). The coracoids fall well within the size range of the South American species C. nobilis, although the humeri are somewhat smaller, thus being intermediate being C. nobilis and C. gallinago. The two fossil tarsometatarsi are even more equivocal, as they differ from each other fairly considerably in size. There is other evidence of large snipe from the West Indies. Morgan (1977) reported a humerus and portions of a femur from fossil deposits on Cayman Brae that exceed C. gallinago in size. We were able to examine these specimens and mea- surements of the humerus are included in Table 7. The latter specimen (UF 22968, Figure 3a) is larger than the fossil humeri from the Bahamas and thus falls within the size range of humeri of C. nobilis. The taxonomic and distributional status of these fossil snipes is quite uncertain. The Com- mon Snipe, C. gallinago, breeds in North America mainly in boreal peat bogs and is found in the West Indies as a southbound migrant, although it may overwinter when and where suitable eco- logical conditions persist. According to Tuck (1972:145), Capella nobilis "is restricted to the montane region in northern and central Andes" in "dry steppe-like places of the Paramo Zone." Snipes in general tend to breed in temperate zones or montane areas and do not have endemic populations restricted to islands, except for the very aberrant genus Coenocorypha of the New Zea- land region. A supposed fossil species of snipe from Puerto Rico (Wetmore, 1920) was later shown to be a woodcock, Scolopax anthonyi (see Olson, 1976b). 32 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY What were the large fossil snipes from the Bahamas and Cayman Brae? Were they migrant individuals of one or more North American spe- cies now extinct? Were they endemic species that bred in the West Indies when North American biomes were depressed southward in the last gla- FIGURE 3.—Humeri in anconal view of modern and fossil snipe (Capella) to show differences in size: A, Capella sp., fossil from Cayman Brae (UF 22968); B, C, C. nobilis; D, Capella sp., fossil from Little Exuma (MCZ 2260); E, Capella sp., fossil from New Providence (USNM 283266); F, G, C. gallinago delicata. (Scale = 3 cm.) TABLE 7.—Measurements (mm) of bones of snipe (Capella) to show large size of West Indian fossils Character C. gallinago delicata C. g. paraguaiae Range Mean n Range Mean n Coracoid Length 17.9-19.2 18.4 14 19.1-20.4 19.8 4 Humerus Length 35.2-40.3 37.7 14 37.5-40.0 38.4 4 Proximal width 8.2- 9.7 9.0 14 9.3- 9.5 9.4 4 Distal width 6.0- 6.8 6.4 14 6.4- 6.7 6.6 4 Ulna Proximal width 4.7- 5.6 5.0 14 5.0- 5.4 5.2 4 Tarsometatarsus Distal width 4.3- 5.2 4.9 14 5.4- 5.9 5.6 4NUMBER 48 33 ciation? Do they represent large temporal forms of C. gallinago or are they conspecific with snipes now found only in South America? If these ques- tions can be answered, it will only be through the accumulation of additional fossil material. Re- gardless, the presence of a snipe in the Pleistocene of the Bahamas is indicative of open country, probably grassy, with moist organic soil suitable for probing. Family BURHINIDAE Burhinus bistriatus nanus Brodkorb Burhinus nanus Brodkorb, 1959:354 [Pleistocene, New Provi- dence] . MATERIAL EXAMINED.—New Providence: 2 scap- ular ends of coracoids, 1 proximal end of scapula, 1 distal end of humerus, 1 proximal end of radius, 1 distal end of femur, 1 proximal and 1 distal ends of tibiotarsi, 1 proximal and 2 distal ends of tarsometatarsi (USNM 283282-283284; PB 9046-9047, PB 9097, UF 3154-3156 [holotype and paratypes oi Burhinus nanus], UF 25655); min- imum number of individuals, 2. REMARKS.—The only member of the Burhini- dae now found in the West Indies is Burhinus bistriatus dominie ensis, restricted to Hispaniola. This subspecies is markedly smaller than any of the forms of B. bistriatus from the mainland of Central and South America, and in most of the earlier literature it was regarded as a separate species. Brodkorb (1959) assigned fossil burhinid ma- terial from Banana Hole to a distinct species, Burhinus nanus, which was distinguished from B. b. dominicensis by its smaller size. Our material largely duplicates the elements that were used in Brodkorb's original description, and measure- ments of the new specimens are virtually identical to those given by Brodkorb (1959:355, table 1). Among the new specimens, the distal end of a humerus (PB 9047) measures 11.7 mm in width, whereas three specimens of B. b. dominicensis range from 12.5 to 13.1 mm. The size differences be- tween B. b. nanus and B. b. dominicensis are real but are not much greater than those between B. b. domincensis and B. b. bistriatus. Therefore, we have chosen to reflect these differences at the subspe- cific rather than specific level. Nonetheless, there definitely was a distinct, small form oi Burhinus in the Bahamas in the Pleistocene, where none exists now. The distribution of B. bistriatus dominicensis and B. b. nanus is obviously relictual and the presence of the species in Central America would suggest that some form of Burhinus probably occurred in Cuba in the past. This was confirmed by Olson, who in November 1980 briefly examined fossils of a Burhinus from Pleistocene cave deposits in Cuba in the collections of the Instituto de Zool- ogia of the Academia de Ciencias de Cuba and in the personal collection of Oscar Arredondo. TABLE 7.—Continued Character C nobilis Cape (Bah Range 'la sp. am as) Mean n Capella sp. (Cayman Brae) Mean* Range Mean n Coracoid Length 20.4-22.9 22.1 4 20.5-21.7 21.1 3 Humerus Length 44.6-46.0 45.1 4 41.2-42.2 41.7 3 44.0 Proximal width 10.4-10.7 10.5 4 10.2 1 11.5 Distal width 7.3- 7.7 7.5 4 7.1- 7.5 7.2 3 7.6 Ulna Proximal width 5.9- 6.4 6.1 4 6.0 1 Tarsometatarsus Distal width 5.7- 6.0 5.8 4 5.1- 5.7 5.4 2 n = 1. 34 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY The New World species of Burhinus are inhab- itants of open plains and prairies and may be significant ecological indicators when they appear in the fossil record, as stated by Feduccia (1980) in describing a new species of Burhinus from the Pleistocene of Kansas. Thus, Burhinus bistriatus nanus strongly augments other information point- ing to the presence of open prairie-like habitat at the time of deposition of the Banana Hole fossils. 283271-283273, 283376, 283397; PB 9059-9067; UF 3163-3173, 3175-3182, 25643-25644); mini- mum number of individuals, 11. REMARKS.—The White-crowned Pigeon is found throughout the Caribbean area and is still fairly common in the Bahamas. It is the most abundantly represented bird in the Banana Hole deposits both in terms of specimens and in mini- mum number of individuals. Family LARIDAE Sterna fuscata Linnaeus Sterna fuscata.—Wetmore, 1938:52 [Pre-Columbian midden, Crooked Island]. MATERIAL EXAMINED.—Crooked Island: 1 com- plete and 1 partial coracoid (YPM 2438); mini- mum number of individuals, 2. REMARKS.—The pantropical Sooty Tern nests today on scattered small cays in the Bahamas. Family COLUMBIDAE Columba leucocephala Linnaeus Columba leucocephala.—Wetmore, 1938:52 [part; pre-Colum- bian midden, Crooked Island].—Brodkorb, 1959:356 [Pleistocene, New Providence]. MATERIAL EXAMINED.—Crooked Island: 1 com- plete humerus, 1 complete carpometacarpus, 1 complete tarsometatarsus (YPM 2439); mini- mum number of individuals, 1. New Providence: 2 complete, 4 nearly complete, and 3 scapular and 2 sternal ends of coracoids; 1 complete and 1 proximal end of scapula; 1 proximal and 2 distal ends of humeri; 2 proximal and 1 distal ends of ulnae; 6 complete, 2 nearly complete, and 4 proximal and 9 distal ends of carpometacarpi; 1 proximal phalanx of major digit of wing; 4 proximal and 3 distal ends of femora; 4 distal ends of tibiotarsi; 2 complete, 2 nearly complete, and 7 proximal ends of tarsometatarsi (USNM 283184,283186, 283191-293192, 283194-283197, 283234-283237, 283239, 283241, 283243, 283245- 283247, 283253-283255, 283260, 283262-283264, Columba squamosa Bonnaterre Columba squamosa.—Wetmore, 1937b:435 [Pleistocene, "Great Exuma" = Little Exuma].—Brodkorb, 1959:355 [Pleisto- cene, New Providence]. Corvus wetmorei Brodkorb, 1959:363 [part; Pleistocene, New Providence]. MATERIAL EXAMINED.—Little Exuma: 1 com- plete coracoid; 1 complete, 2 proximal, and 2 distal ends of ulnae; 1 complete carpometacarpus; 1 complete femur (MCZ 2259); minimum num- ber of individuals, 3. New Providence: 1 cranial portion of sternum; 3 complete, and 2 scapular and 1 sternal ends of coracoids; 3 proximal ends of scapulae; 2 proximal ends of humeri; 2 com- plete, and 2 proximal and 1 distal ends of ulnae; 2 distal ends of radii; 1 nearly complete and 1 distal end of carpometacarpi; 3 proximal phalan- ges of major digit of wing; 1 proximal and 2 distal ends of femora; 1 distal end of tibiotarsus; 1 complete and 2 proximal ends of tarsometatarsi (USNM 283185, 283190, 283193, 283257- 283259, 283274-283276, 283378; PB 9048-9058; UF 3157-3162, 3215); minimum number of in- dividuals, 3. REMARKS.—We have confirmed the identifica- tions of Wetmore and Brodkorb and found the Scaled Pigeon to be among the three most abun- dant species of birds in Banana Hole. Size alone permits this species to be separated from the smaller C. leucocephala in the majority of instances; in addition there are discernible qualitative dif- ferences between the two species in some ele- ments. The proximal end of a femur of this species was included among the paratypes of Corvus wet- morei Brodkorb (1959). NUMBER 48 35 Columba squamosa is now distributed throughout virtually all of the Antilles, except Jamaica, where it has strayed on occasion but otherwise appears to be replaced by C fasciata (see Lack, 1976:238). C. squamosa, which is essentially a bird of forest treetops, is unknown in the Bahamas today and evidently has retreated from there since the Pleis- tocene. Zenaida aurita (Temminck) Zenaida aurita.—Brodkorb, 1959:356 [part; Pleistocene, New Providence]. MATERIAL EXAMINED.—New Providence: 1 complete and 4 proximal ends of coracoids; 1 complete, 2 nearly complete, and 1 distal end of humeri; 2 shafts of carpometacarpi (USNM 283183,283187-283188,283244, 283248, 283250, 283253, 283371; UF 3187, 3192, PB 9071); min- imum number of individuals, 4. REMARKS.—The Zenaida Dove is restricted to the West Indies, except for a subspecies inhabiting the Yucatan Peninsula and certain adjacent is- lands. It occurs throughout the Bahamas. Geotrygon chrysia Salvadori Oreopelia chrysia.—Wetmore, 1937b:436 [Pleistocene, "Great Exuma" = Little Exuma]. Zenaida aurita.—Brodkorb, 1959:356 [part; Pleistocene, New Providence]. MATERIAL EXAMINED.—Little Exuma: 1 com- plete humerus and 1 complete tarsometatarsus (MCZ 2261). New Providence: 1 nearly complete and 2 proximal ends of coracoids, 3 distal ends of humeri, 1 complete and 2 proximal ends of car- pometacarpi, 2 complete and 2 proximal ends of tarsometatarsi (USNM 283189, 283233, 283242, 283249, 283251, 283270; UF 3183, 3185, 3186, 3190; 25645; PB 9068-9069); minimum number of individuals, 3. REMARKS.—Two humeri, a coracoid, and a carpometacarpus that Brodkorb (1959) referred to Zenaida aurita appear to belong instead to G. chrysia. Apart from the tarsometatarsus, which was not included among the material examined by Brodkorb, the elements of these two species are extremely difficult to separate, and indeed we were unable to identify certain specimens. Four of the skeletal elements could be segregated as follows: Tarsometatarsus: Easily distinguished in its en- tirety, that of G. chrysia having a much longer, more slender shaft, and narrower proximal end, whereas in Z. aurita the bone is short and stout. Coracoid: Viewed laterally, G chrysia has a smaller glenoid facet and the scapular facet is not situated as far sternally on the shaft as in Z. aurita. Humerus: Virtually indistinguishable except that the ectepicondylar prominence in G. chrysia is usually closer to the external condyle than in Z. aurita. Carpometacarpus: In its entirety, the carpo- metacarpus of G chrysia is shorter and stouter than that of Z. aurita; also, the shaft of the major metacarpal is straighter and slopes up proximally towards the pollical facet, whereas in Z. aurita the shaft of the major metacarpal is very slightly bowed and does not slope up as markedly to the pollical facet. We were forced to regard the following speci- mens as indeterminate, although they certainly pertain either to G chrysia or to Z. aurita. New Providence: 1 scapula, 3 ulnae, 1 radius, 2 car- pometacarpi, 3 femora, 2 tibiotarsi, and 1 tarso- metatarsus (USNM 283238, 283240, 283256, 283372, 283373-283375; UF 3184, 3188, 3189, 3191; PB 9070, 9072); minimum number of in- dividuals, 3. Geotrygon chrysia, the Key West Quail-Dove, is found in Cuba, Hispaniola, Puerto Rico, and the Bahamas, where it has been recorded from Grand Bahama, Andros, New Providence, Eleuthera, and San Salvador, but not the Exumas. It is a forest bird that is found in several types of wood- land but evidently is "most frequently encoun- tered in dry lowland forest and scrub, often on limestone" (Sorrie, 1979:728). The species is rare in parts of its range, including the Bahamas, and evidently is becoming more restricted in distri- bution. It may once have bred in the Florida Keys, where it is now only accidental. (U. S. 36 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY records have been summarized in detail by Sorrie, 1979). Family PSITTACIDAE Amazona leucocephala (Linnaeus) Amazona leucocephala.—Wetmore, 1938:52 [pre-Columbian midden, Crooked Island].—Brodkorb, 1959:356 [Pleisto- cene, New Providence]. MATERIAL EXAMINED.—Crooked Island: 1 pre- maxilla (YPM 2440). New Providence: 1 nearly complete and 1 distal end of ulnae, 1 distal end of radius, 1 proximal and 1 distal ends of carpo- metacarpi, 1 proximal end of femur, 1 complete and 1 distal end of tarsometatarsi (USNM 283277-283279, 283290-283291; PB 9078, UF 3193, 3194); minimum number of individuals, 2. REMARKS.—The Cuban Parrot occurs in Cuba and the Cayman Islands, in addition to the Ba- hamas, where the only extant populations are on Abaco and Great Inagua. Worthington (Todd and Worthington, 1911) obtained specimens on Acklins, where they no longer occur, and Bond (1956) also mentions Long and Fortune Islands in the historical range of the species. The fossil record indicates an even more extensive distri- bution in the Bahamas; presumably much of the reduction in range is due to human disturbance. Family CUCULIDAE Saurothera merlini d'Orbigny MATERIAL EXAMINED.—New Providence: 1 dis- tal end of humerus, 1 distal end of tibiotarsus, 1 complete and 1 proximal end of tarsometatarsi (USNM 283198, 283199; PB 9073, 9096); mini- mum number of individuals, 1. REMARKS.—In the Bahamas, the Great Lizard Cuckoo is known to be resident only on New Providence, Eleuthera, and Andros; it is found elsewhere in Cuba and on the Island of Pines. Family TYTONIDAE Tyto pollens Wetmore Tyto pollens Wetmore, 1937b:436 [Pleistocene, "Great Ex- uma" = Little Exuma].—Brodkorb, 1959:357 [Pleisto- cene, New Providence]. MATERIAL EXAMINED.—Little Exuma: 1 nearly complete coracoid, 1 proximal end of ulna, 1 major metacarpal lacking proximal end, 1 com- plete femur (holotype), 1 proximal end of tibio- tarsus (MCZ 2262, 2263); minimum number of individuals, 1. New Providence: 1 distal end of carpometacarpus; 2 distal ends of tibiotarsi; 1 complete, 1 proximal and 1 distal ends of tarso- metatarsi, and 1 inner and 1 outer trochleae; 2 pedal and 1 ungual phalanges (USNM 283287- 283288; PB 9077; UF 3195-3199, 25646-25647, 25656); minimum number of individuals, 2. REMARKS.—This giant, extinct barn owl ap- pears to be the Bahaman counterpart of T ostologa Wetmore (1922) from the Pleistocene of Hispan- iola, and T noeli Arrendondo (1972) from the Pleistocene of Cuba. The exact taxonomic status of these three owls is rendered difficult to deter- mine by the relative scarcity of remains of T pollens and because it has not yet been possible to make direct comparisons with specimens of T noeli. Nevertheless, Wetmore's diagnosis of T pol- lens still holds: it is somewhat larger and the shafts of the limb elements are noticeably more robust than in T ostologa (Table 8); also, the trochanteric ridge of the femur is longer and better developed in T pollens. Arredondo's (1972) measurements indicate that T noeli was smaller than either T. ostologa or T pollens (Table 8). Sympatric with T noeli in Cuba was an even larger species of barn owl, T riveroi (see Arredondo, 1976). The tremendous number of bones belonging to rodents of the genus Geocapromys in the Banana Hole deposits is almost surely the result of pre- dation by Tyto pollens. Geocapromys has disappeared from all of the islands of the Bahamas except East Plana Cay, a tiny arid islet near Acklins Island in the southern Bahamas, where the only living population of Geocapromys ingrahami survives. This NUMBER 48 37 TABLE 8.—Measurements (mm) of three allopatric taxa of giant barn owls (Tyto) from the West Indies (those for T. noeli are from Arredondo (1976); those for T. ostologa are based on an uncataloged USNM series) Character T. noeli T. pollens T. ostologa Range n Range n Range Mean n Femur Length 73.0-74.3 2 81.2 1 80.7 - 1 Proximal width 13.1-14.4 3 16.3 1 11.9-17.7 16.3 12 Least width 6.0- 6.5 3 8.5 1 7.0- 7.4 7.2 3 of shaft Distal width 14.0-14.8 2 18.0 1 12.1-18.3 17.0 10 Tibiotarsus Proximal width 16.0 1 15.2 1 17.5 - 1 Distal width 15.0 1 16.7 1 16.0-17.6 16.9 12 Least width 6.4 1 8.0 1 7.5 7.5 2 of shaft Tarsometatarsus Length 91.7 1 99.2 1 91.5-95.5 93.8 3 Proximal width 14.6-16.0 2 17.6- 18.0 2 17.1-17.6 17.5 8 Least width 6.7- 7.3 2 8.8 1 5.8- 7.8 7.1 4 of shaft Width through 17.4 1 21.1 1 18.5-20.5 19.6 7 trochleae Depth of outer - - 14.9- 15.0 2 12.7-13.6 13.2 6 trochlea Depth of inner - - 12.1- 13.2 2 11.0-12.0 11.6 5 trochlea population is well adapted to dry, scrubby envi- ronment (Clough, 1972). It is possible that late Pleistocene reduction of arid environments in the Bahamas, combined with the later effects of pre- dation and habitat alteration by man, may have been responsible for the near total extirpation of G. ingrahami. Whatever the reasons for the demise of these rodents, it is certain that Tyto pollens could not have survived the disappearance of its prin- cipal prey. Tyto alba (Scopoli) Tyto alba.—Brodkorb, 1959:358 [Pleistocene, New Provi- dence]. MATERIAL EXAMINED.—New Providence: 1 pre- maxilla, 1 distal end of humerus, 1 distal end of tarsometatarsus (USNM 283286; UF 3200, 3201); minimum number of individuals, 1. REMARKS.—The Barn Owl today is a rather uncommon resident known from most of the larger Bahaman islands, including New Provi- dence. Family STRIGIDAE Athene cunicularia (Molina) Speotyto cunicularia.—Wetmore, 193 7b: 440 [Pleistocene, "Great Exuma" = Little Exuma]. Glaucidium dickinsoni Brodkorb, 1959:358 [Pleistocene, New Providence]. Otusprovidentiae Brodkorb, 1959:360 [Pleistocene, New Prov- idence]. MATERIAL EXAMINED.—Little Exuma: 1 proxi- mal and 1 distal ends of tarsometatarsi (MCZ 2264). New Providence: 4 complete and 4 par- tial coracoids, 1 proximal and 2 distal ends of 38 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY humeri, 3 proximal ends of ulnae, 1 complete and 2 proximal ends of carpometacarpi, 1 complete and 4 proximal ends of femora, 1 proximal and 10 distal ends of tibiotarsi, 3 proximal and 5 distal ends of tarsometatarsi (USNM 283201- 283228, 283367-283370; PB 9074-9076; UF 3202 [holotype of Glaucidium dickinsoni] and 3203-3207 [holotype and paratypes of Otus providentiae]); minimum number of individuals, 6. REMARKS.—The Burrowing Owl, Athene cuni- cularia, is the only strigid owl that occurs in the Bahamas today. It has been found on most of the northern islands, including the Exumas and New Providence, but is known in the southern Baha- mas only on Samana Cay and Great Inagua (Buden, 1979). Paulson (1966) gives details con- cerning the apparently changing status of the species on various islands of the Bahamas. When all 32 of our newly collected specimens of small owls from Banana Hole proved to belong to this species, we became highly suspicious of the identity of the two species of Strigidae named from this site by Brodkorb (1959). Examination of the type material of Otus providentiae (based on 5 specimens) and Glaucidium dickinsoni (based on the holotype only) confirmed that both taxa are referable to Athene cunicularia. Curiously, Brodkorb did not mention the Burrowing Owl in his com- parisons or discussions. The holotypes of G. dickinsoni and 0. providentiae are the distal ends of tibiotarsi. In the diagnoses of both of these species, Brodkorb (1959) mentions a distinct pit or fossa at the distal end of the tendinal groove that undercuts the intercondylar bar. This feature likewise occurs in Athene cunicu- laria. All of the other characters in the diagnosis of G. dickinsoni are also found in A. cunicularia (intercondylar bar oblique rather than transverse; peroneal groove reduced; internal prominence for oblique ligament on anterior face of shaft, instead of protruding medially; shaft only slightly ex- panded distally). It should also be noted that the holotype of G. dickinsoni is extremely abraded and of minimal diagnostic value. The skeleton in pygmy owls of the genus Glaucidium is very distinct from that of either Otus or Athene cunicularia. We are quite certain that none of the fossil strigid bones collected thus far from the Bahamas are referable to Glaucidium. In addition to the holotype, Brodkorb (1959) referred four other bones to Otus providentiae (prox- imal end of tibiotarsus, incomplete coracoid, proximal end of ulna, and proximal end of car- pometacarpus). Because we have much more ad- ditional strigid material from Banana Hole and because the problem of separating Athene cunicu- laria from Otus also arises with fossils from Puerto Rico and Cuba, we have undertaken to point out the differences between these taxa in some detail. Although the differences between A. cunicularia and Otus are not as striking as those separating these two genera from Glaucidium, they are nev- ertheless quite pronounced. ot stapled or bound), and accompanied by original art. Minimum acceptable length is 30 pages. Front matter (preceding the text) should include: title page with only title and author and no other information, abstract page with author/title/series/etc., following the establish- ed format, table of contents with indents reflecting the heads and structure of the paper. First page of text should carry the title and author at the top of the page and an unnum- bered footnote at the bottom consisting of author's name and professional mailing address. Center heads of whatever level should be typed with initial caps of major words, with extra space above and below the head, but with no other preparation (such as all caps or underline). Run-in paragraph heads should use period/dashes or colons as necessary. Tabulations within text (lists of data, often in parallel columns) can be typed on the text page where they occur, but they should not contain rules or formal, numbered table heads. Formal tables (numbered, with table heads, boxheads, stubs, rules) should be sub- mitted as camera copy, but the author must contact the series section of the Press for edito- rial attention and preparation assistance before final typing of this matter. Taxonomic keys in natural history papers should use the alined couplet form in the zoology and paleobiology series and the multi-level indent form in the botany series. If cross-referencing is required between key and text, do not include page references within the key, but number the keyed-out taxa with their corresponding heads in the text. Synonymy in the zoology and paleobiology series must use the short form (taxon, author, year:page), with a full reference at the end of the paper under "Literature Cited." For the botany series, the long form (taxon, author, abbreviated journal or book title, volume, page, year, with no reference in the "Literature Cited") is optional. Footnotes, when few in number, whether annotative or bibliographic, should be typed at the bottom of the text page on which the reference occurs. Extensive notes must appear at the end of the text in a notes section. If bibliographic footnotes are required, use the short form (author/brief title/page) with the full reference in the bibliography. Text-reference system (author/year/page within the text, with the full reference in a "Literature Cited" at the end of the text) must be used in place of bibliographic footnotes in all scientific series and is strongly recommended in the history and technology series: "(Jones, 1910:122)" or ".. . . Jones (1910:122)." Bibliography, depending upon use, is termed "References," "Selected References," or "Literature Cited." Spell out book, journal, and article titles, using initial caps in all major words. For capitalization of titles in foreign languages, follow the national practice of each language. Underline (for italics) book and journal titles. Use the colon-parentheses system for volume/number/page citations: "10(2):5-9." For alinement and arrangement of elements, follow the format of the series for which the manuscript is intended. Legends for illustrations must not be attached to the art nor included within the text but must be submitted at the end of the manuscript—with as many legends typed, double- spaced, to a page as convenient. Illustrations must not be included within the manuscript but must be submitted sepa- rately as original art (not copies). All illustrations (photographs, line drawings, maps, etc.) can be intermixed throughout the printed text. They should be termed Figures and should be numbered consecutively. If several "figures" are treated as components of a single larger figure, they should be designated by lowercase italic letters (underlined in copy) on the illus- tration, in the legend, and in text references: "Figure 9b_." If illustrations are intended to be printed separately on coated stock following the text, they should be termed Plates and any components should be lettered as in figures: "Plate 9b," Keys to any symbols within an illustration should appear on the art and not in the legend. A few points of style: (1) Do not use periods after such abbreviations as "mm, ft, yds, USNM, NNE, AM, BC." (2) Use hyphens in spelled-out fractions: "two-thirds." (3) Spell out numbers "one" through "nine" in expository text, but use numerals in all other cases if possible. (4) Use the metric system of measurement, where possible, instead of the English system. (5) Use the decimal system, where possible, in place of fractions. (6) Use day/month/year sequence for dates: "9 April 1976." (7) For months in tabular list- ings or data sections, use three-letter abbreviations with no periods: "Jan, Mar, Jun," etc. Arrange and paginate sequentially EVERY sheet of manuscript—including ALL front matter and ALL legends, etc., at the back of the text—in the following order: (1) title page, (2) abstract, (3) table of contents, (4) foreword and/or preface, (5) text, (6) appendixes, (7) notes, (8) glossary, (9) bibliography, (10) index, (11) legends.