Fossil and Subfossil Birds from the Bahamas Storrs L. Olson and William B. Hilgartner Introduction Rico that yielded many thousands of bird bones, this is still a rather small sample, but it is none¬ Up to the present, three papers have dealt with theless quite a significant one. significant fossil and subfossil avifaunas from the The collection from Crooked Island reported Bahamas. Pleistocene birds were reported by Wetmore (1937b) from Little Exuma, and by by Wetmore (1938) consisted of 21 specimens Brodkorb (1959) from New Providence. Bird re¬ that were referred to 11 species. Although of no mains from an archeological site on Crooked great antiquity, several of these specimens repre¬ Island were discussed by Wetmore (1938). The sent significant range extensions for certain taxa. geological and historical setting of these collec¬ In addition to identifying and analyzing the tions are detailed elsewhere (pp. 3-5). newly collected material from Banana Hole, we Wetmore’s (1937b) material from Exuma con¬ also obtained and re-examined all of the previ¬ sisted of 30 specimens that he assigned to 13 ously described specimens mentioned above, in¬ species, 3 of which were described as new. The cluding all types. Our investigation did not reveal collection that Brodkorb (1959) studied from Ba¬ any new species but instead resulted in the alter¬ nana Hole on New Providence consisted of 65 ation of taxonomic status of 6 of the 9 new species specimens that were referred to 15 species, 6 of described by Wetmore (1937b) and Brodkorb which were described as new. Because of the high (1959), as well as the correction of a number of number of supposedly extinct avian taxa and misidentifications of specimens. their zoogeographical significance, the Banana We have disregarded the collection of bird Hole fossil avifauna is among the most interesting bones from Abaco (presumably Great Abaco) yet reported from the West Indies. The Smithson¬ reported by Conklin (1971) as “post-Columbian” ian expedition of 1978 obtained 219 additional on account of their association with remains of bird fossils from Banana Hole, which, together Rattus. First of all, these cannot be regarded as with 70 previously unstudied specimens in the fossils, and secondly, many of the identifications collections of the Florida State Museum and of do not seem credible. Passerines, mainly small Pierce Brodkorb, yielded a total of 354 speci¬ species, constitute 87% of the sample, yet despite mens—more than 5 times as many as were avail¬ the manifest difficulties in identifying bones of able for Brodkorb’s original study of the avifauna. small Passeriformes, Conklin assigned to species By comparison with other Pleistocene deposits, even such undiagnostic elements as shafts of long such as some of those collected by Olson in Puerto bones. Furthermore, 4 of the 10 nine-primaried oscines reported by Conklin are either rare, or at Storr LsO. lson D, epartmen o tVfertebrat eZoolog yN, ationa Ml useum best uncommon, migrants to the Bahamas (Vireo o fNatura lHistory ,Smithsonian Institution ,Washington ,D.C. 20560 .William B .Hilgartner ,Irvine Natura lScience Center ,St. solitanus , Protonotana citrea , Vermwora pinus, and Timothy ’Sschoo Sl,tevenso nM, arylan 2d1153. Wilsoma citnna ), whereas none of the resident 22 NUMBER 48 23 Parulidae, or such common species as Coereba Wetmore from Exuma were made available by flaveola , Vireo crassirostris, Geothlypis trichas, Den- Raymond A. Paynter, Jr., Museum of Compar¬ droica palmarum, or D. discolor , were said to be ative Zoology (MCZ), Harvard University, and represented. The faunal composition of Conklin’s the archeological specimens from Crooked Island sample of bones seems highly unlikely for the were supplied by Eleanor Stickney, Peabody Mu¬ Bahamas and in our opinion this report cannot seum of Natural History, Yale University (YPM). be considered reliable. Unfortunately, Conklin’s Skeletal material of Buteo mtidus was lent by Mary paper has already been cited by Buden (1981:2) C. McKitrick from the University of Arizona as providing fossil evidence for a long occurrence collections. Line drawings are by Randall W. of the tanager Spindalis zena on Great Abaco, but Blackburn, maps are by Sigrid James Bruch, and the recent age of Conklin’s sample precludes this photographs are by Victor E. Krantz. Gary S. interpretation. That faunal remains from recent Morgan kindly provided information and refer¬ owl pellets may have some distributional signifi¬ ences, and permitted us to use unpublished ma¬ cance, however, is attested to by Buden’s (1974) terial from his master’s thesis. We are grateful to discovery of two partial skulls of Spindalis zena and Gregory K. Pregill, David W. Steadman, and the partial skull of the rail Porzana Carolina in Barn Charles A. Woods for their comments on the Owl ( Tyto alba ) pellets from the Caicos Islands, manuscript and we are especially indebted to where neither species had been recorded previ¬ Mary H. Clench, whose knowledge of the Baha¬ ously. mas made her detailed criticism of the manuscript Acknowledgments.— The impetus for this of particular benefit. study came from David Campbell, former direc¬ tor of the Bahamas National Trust, who indicated Species Accounts that further collecting at the Banana Hole local¬ ity would be fruitful and who suggested numerous At the beginning of each account we have valuable contacts in the Bahamas. Permits to listed the identifications as given by Wetmore collect fossils and modern comparative material (1937b, 1938) and Brodkorb (1959), where appro¬ were kindly provided by the Ministry of Agricul¬ priate. When no such entry appears, the speci¬ ture, Bahamas, through Colin Higgs. In the field, mens constitute the first fossil record of a species Olson was expertly assisted by Frederick V. for the Bahamas. The specimens of Pterodroma cf. Grady, Helen F. James, and Charles A. Meister. hasitata , Buteo sp., Corvus palmarum, and Sturnella Field work and Hilgartner’s museum studies were sp. also constitute first records for the Bahamas supported by grants from the Fluid Research but are based in part on previously misidentified Fund, Smithsonian Institution, through S. Dillon specimens. Minimum number of individuals was Ripley. Grady’s participation was made possible determined by the standard method of counting by funds from the Department of Paleobiology, the most abundant element from a particular through Martin A. Buzas. Ben Bohl and Rose side. We have omitted this figure when it is Blanchard, International Field Studies, were in¬ obvious that only one individual was present. strumental in providing Olson and Meister with facilities at Forfar Field Station while collecting Family Procellariidae birds on Andros. Pierce Brodkorb, University of Florida (PB), generously lent a collection of unstudied fossil Pterodroma cahow (Nichols and Mowbray) birds from Banana Hole. The previously identi¬ Pterodrom acahow —. Wetmore 1, 938:5 1[part p; re-Columbian fied material from this site, including all types, midden C. rooked Island], was kindly supplied by S. David Webb, Florida Puffinus puffinus .—Wetmore ,1938:51 [pre-Columbian mid¬ State Museum (UF). The fossils described by den C, rooked Island], 24 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Puffinus Ihermimen .—Wetmore ,1938:51 [part ;pre-Colum¬ handling, as there is no coracoid of Columba leu¬ bian midden ,Crooked Island]. cocephala among the Crooked Island bones and Columb aleucocephala —. Wetmore 1, 938:5 2[par tp; re-Colum¬ bian midden ,Crooked Island]. the specimen is correctly numbered. The Cahow is known to breed only on Ber¬ Material Examined.— Crooked Island: cora¬ muda, where it once occurred in vast quantities coid (part of YPM 2439), radius and ulna (part but is now nearly extirpated. It is smaller than of YPM 2434), 2 carpometacarpi (YPM 2432), Pterodroma hasitata (Table 5) or P. caribbaea , the tibiotarsus (part of YPM 2433); minimum num¬ only congeners that breed in the West Indies. ber of individuals, 1. Virtually nothing is known of its range at sea and Remarks.— Wetmore (1938) referred a radius, the Crooked Island material is the only indication ulna, and carpometacarpus to this species, but of the species away from the known breeding the last is too large for P. cahow and belongs grounds. 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 cf. hasitata (Kuhl) Pterodroma cahow and may be distinguished from Pterodrom acahow —. Wetmore 1, 938:5 1[part p; re-Columbian those of Puffinus puffinus by the following charac¬ midden C, rooked Island], ters: (1) pisiform process more widely separated from pollical facet, nearly parallel with interme- Material Examined.— Crooked Island: car¬ tacarpal space; (2) caudomedial margin of carpal pometacarpus (part of YPM 2434). trochlea more truncate, not decidedly rounded; Remarks.— This specimen was included (3) fossae around pisiform process faint or lacking. among those that Wetmore (1938) referred to P. Bond (1956:188) attributes osteological re¬ cahow , but it is too large for that species (Table mains of Puffinus puffinus to St. Croix, as well as to 5). The Black-capped Petrel is known to breed in Crooked Island, but we have been unable to find mountainous areas of Hispaniola, Guadeloupe, any other mention of this record in the literature and Dominica. Very recently it was discovered and we suspect that an error may have arisen offshore and flying about sea cliffs in the Sierra through confusion with Puffinus Iherminieri , which Maestra region of southeastern Cuba (Bond, was twice reported from middens on St. Croix 1978; Nicasio Vina, Academia de Ciencias, San¬ (Wetmore, 1918; 1937a). Wing et al. (1968) re¬ tiago de Cuba, pers. comm.). An all dark form, P. corded the distal portion of a humerus from an caribbaea Carte from Jamaica, thought to be ex¬ archeological site on Antigua as Puffinus cf. puffi¬ tinct, is variously regarded as a separate species, nus, but the specimen cannot now be located or as a subspecies or color phase of P. hasitata. (Wing, pers. comm.) and we were not able to These represent the only known resident popula¬ check the identification. Thus, there appears to tions of Pterodroma in the West Indies. be no verifiable record of Puffinus puffinus , prehis¬ Wetmore (1918) assigned a tibiotarsus (USNM toric or otherwise, from the West Indies. 225842) from a midden on St. Croix to this genus Of the two bones Wetmore assigned to Puffinus but did not identify it to species in the absence of Iherminieri , the tibiotarsus actually belongs to P. comparative material. A fragmentary proximal cahow , as shown by the much smaller inner cne¬ end of a humerus (USNM 428289) from a midden mial crest and more terete shaft. We can offer no in Martinique he referred to P. hasitata (Wetmore, explanation for Wetmore’s inclusion of a coracoid 1952). We examined these specimens and con¬ of P. cahow with three other specimens correctly firmed their generic placement. As with the car¬ identified as Columba leucocephala , this element pometacarpus from Crooked Island, these speci¬ being utterly different in these two species. The mens can be referred to the species P. hasitata only error appears not to be one of subsequent mis- on the basis of size and geographic probability, NUMBER 48 25 Table 5.— Measurements (mm) of Plerodroma to show the size of Crooked Island specimens compared with a series of fossils of P. cahow from Bermuda and skeletons of P. hasilala from Hispaniola as they are not otherwise sufficiently diagnostic Material Examined. —Crooked Island; com¬ to permit specific identification in the complex plete carpometacarpus (part of YPM 2433). genu sPterodroma. Remarks.— This specimen, which is from a The Crooked Island specimen is the first indi¬ juvenile, was correctly identified by Wetmore cation of the species in the Bahamas. Elsewhere (1938), although the tibiotarsus he associated in the West Indies, it presently requires high with it was not (see Pterodroma cahow). Audubon’s mountainous terrain or sea cliffs for nesting, but Shearwater is a circumtropical species that this may be an artifact of human intervention, as breeds, or formerly bred, in a number of localities the closely related P. cahow in Bermuda does not. in the West Indies, including the Bahamas, where It is difficult to imagine that the former human it still occurs on suitable cays and in adjacent inhabitants of Crooked Island would have been waters. able to capture these birds at sea, although the presence of P. cahow in the same deposits may be Family Sulidae indicative of some such capacity unless both taxa were picked up as hurricane wrecks. It seems Sula leucogaster (Boddaert) equally improbable that both species bred sym- patrically in the Bahamas, but the possibility Sul aleucogaster —. Wetmore 1, 938:5 2[pre-Columbia nmidden, cannot be ruled out. Crooke Idsland], Material Examined.— Crooked Island: 1 cor¬ Puflinus lherminieri Lesson acoid and 1 distal end of humerus (YPM 2435). Remarks.— The circumtropical Brown Booby Puffinus lherminieri .—Wetmore ,1938 :51 [part ;pre-Colum¬ breeds in the Bahamas and elsewhere in the bian midden ,Crooked Island], Caribbean. 26 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Family Ardeidae of Accipiter striatus endemic to Cuba, Hispaniola, and Puerto Rico. A. s. velox is now only a rare Nyctanassa violacea (Linnaeus) winter visitor to the West Indies, including the Bahamas. The discovery of fossils of an Accipiter Material Examined.— New Providence: 1 fur- of this size on both Exuma and New Providence cula (UF 25642); 1 distal end of ulna (USNM is a fairly good indication that this hawk was once 283200). either resident, or of more regular occurrence, in Remarks.— The distal end of an ulna of this the Bahamas than at present. species p obably cannot be safely separated from that of Nycticorax nycticorax, but the entire furcula Buteo sp. is easily distinguished by the narrower shaft and longer and more slender symphysis. The Yellow- Calohiera xquadratu sWetmore 1,937b:42 9[Pleistocene “,Great crowned Night Heron is found throughout the Exuma” = Little Exuma], West Indies and is common in the Bahamas. Calohiera xquadratus —. Brodkorb 1, 959:35 1[Pleistocene N, ew Providence]. Family Threskiornithidae Material Examined.— Little Exuma: 1 distal end of tarsometatarsus (MCZ 2256, holotype). Eudocimus albus (Linnaeus) New Providence: 1 distal end of tibiotarsus (UF 3152). Guard alba .—Wetmore, 1938:52 [pre-Columbian midden. Remarks.— Wetmore (1937b:429) founded a Crooke Idsland]. new genus and species of hawk from Little Ex¬ Material Examined.— Crooked Island: 1 dis¬ uma, Calohierax quadratus , on the distal portion of tal end of ulna, 1 distal end of tibiotarsus (YPM a tarsometatarsus, lacking the outer trochlea. This 2436). was diagnosed as being similar to Buteo “but with Remarks.— The White Ibis occurs only occa¬ the projecting outer portion of the second [inner] sionally in the Bahamas at present, probably as trochlea much reduced so that the trochlea ap¬ a wanderer from the mainland, although the pears square and block-like.” Examination of the species breeds in the Greater Antilles. holotype with a dissecting microscope revealed that this supposed character is due entirely to wear, the medial edge of the inner trochlea being Family Accipitridae clearly abraded. This specimen cannot, therefore, be separated generically from Buteo. Accipiter striatus velox (Wilson) The same is true of the distal end of a tibiotar¬ sus from New Providence that Brodkorb (1959: Accipiter striatus velox .—Wetmore ,1937b:428 [Pleistocene, “Great Exuma” = Little Exuma], 351) referred to Calohierax quadratus on account of its having the “raised edges of peroneal groove Material Examined.— Little Exuma: 1 distal forming ridges, not shelves; internal rugosity of end of humerus (MCZ 2255). New Providence: 1 oblique ligament situated slightly lower on shaft.” proximal end of carpometacarpus (USNM These characters are variable, both between and 283232). within, species of Buteo. In addition, the specimen Remarks.— Our new specimen from New is quite worn and has little diagnostic value be¬ Providence corroborates Wetmore’s assignment yond the generic level. We regard Calohierax Wet¬ of the Exuma fossil to the subspecies of the North more 1937 as a junior synonym of Buteo Lacepede American mainland. Both fossils are the size of 1799. females (the larger sex) of A. s. velox and are thus It is difficult to determine the identity of the much larger than the three diminutive subspecies two fragmentary Bahaman fossils within the com- NUMBER 48 27 plex of Buteo and its relatives. On the basis of size seum, we found three fragments of shaft that fit alone, the corresponding bones of Buteo jamaicensis together to constitute most of a right ulna, lacking and Buteogallus anthracinus are larger than the the ends. With these was a fragment of a shaft of fossils, and those of Buteo platypterus , B. magnirostris, a left ulna. We estimate the total length of the and B. ridgwayi are smaller. The middle trochlea more complete specimen to have been about 210 in the fossil tarsometatarsus is shorter and broader mm; the greatest diameter at midshaft is 10 mm. than in Parabuteo unicinctus , and in distal view does These specimens are too large for any of the other not have the external rim extending farther pos¬ species of birds known from Banana Hole and teriorly as in that species. The fossils are within their structure conforms well with that in the the size range of Buteo brachyurus or B. albonotatus , Accipitridae; therefore we have tentatively re¬ but the hindlimb elements in B. brachyurus are ferred them to T. gloveralleni , thus extending its more robust than in the fossils, and in B. albono¬ known range to New Providence. tatus the ends of the bones are wider but the shafts Titanohierax is not related to the eagles of the narrower than in the fossils. The two Bahaman genera Aquila or Hahaeetus. Instead it is a gigantic specimens are very similar in size and details to buteonine hawk that Wetmore (1937b) consid¬ both Buteo lineatus and B. mtidus. Although the ered to be most similar to the much smaller tarsometatarsus is relatively longer in B. lineatus species of Hypomorphnus (= Buteogallus). We found than in B. nitidus, the distinction cannot be made as much or more similarity between Titanohierax in the fossil because of its fragmentary condition. and the South American genus Geranoaetus, how¬ Buteo lineatus would be more likely on geographi¬ ever. Recently, Campbell (1979) erected a new cal grounds, but the nature of the Pleistocene genus, Amphbuteo, for two species of very large, environment in the Bahamas would not have late Pleistocene accipitrids: Amplibuteo hibbardi been incompatible with the occurrence of a xe- Campbell from tarpits in Peru, and Amplibuteo rophilous species, such as B. nitidus. Without fur¬ (Morphnus auct.) woodwardi (L. Miller) from the ther fossil material, the specific identity of the Rancho La Brea tarpits in southern California. Bahaman Buteo will have to remain undeter¬ He regarded Amplibuteo as being most closely mined. The existing specimens nevertheless doc¬ related to Geranoaetus. Although Campbell (1979: ument the presence of a species that has become 83) stated that Amplibuteo “is quite distinct from extinct in the Bahamas. . . Titanohierax ” he did not elaborate further and gave no indication that he had actually examined specimens of Titanohierax in his comparisons. A Titanohierax gloveralleni Wetmore possible relationship between Titanohierax and Titanohierax gloverallen iWetmore ,1937b:430 [Pleistocene, Amplibuteo should be explored further. “Great Exuma” = Little Exuma]. There are additional remains, probably of Ti¬ tanohierax, from elsewhere in the West Indies. A Materal Examined. —Little Exuma: 1 tarso¬ giant “eagle” was described as Aquila borrasi by metatarsus lacking proximal end (MCZ 2257, Arredondo (1970), from two Pleistocene cave de¬ holotype), 1 proximal end of carpometacarpus posits in Havana Province, Cuba. Part of the (MCZ 2258). New Providence: 1 nearly complete material of this species had originally been sent and 1 shaft of ulnae (UF 25640-25641); mini¬ to the late Bryan Patterson at the Museum of mum number of individuals, 1. Comparative Zoology, who intended to collabo¬ Remarks. — Titanohierax gloveralleni is a valid rate with Arredondo in describing it. Unfortu¬ taxon of very large Accipitridae hitherto known nately, Patterson never resumed his correspond¬ only from the two specimens from Little Exuma. ence on the matter (Arredondo, 1972:2) and Ar¬ Among some unidentified material from Banana redondo was constrained to proceed with the Hole in the collections of the Florida State Mu¬ description on the basis of the specimens still 28 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY available to him. He placed the species in the two mandibular symphyses from a cave on Grand genus Aquila , in its classical sense, owing to the Cayman that were tentatively referred to T. glov¬ lack of opportunities to examine extensive com¬ eralleni by Morgan (1977) on account of their parative material. Olson recently examined Pat¬ great size and similarity to Buteogallus. terson’s notes and the specimens upon which they In spite of the taxonomic problems associated were based: two distal portions of tibiotarsi, a with Titanohierax at the generic and specific levels, femur lacking the distal end, and some pedal the fossil record shows that giant hawks once phalanges. The two tibiotarsi are quite different existed in the Bahamas, Hispaniola, Cuba, and in size, and the smaller one is so incomplete that Grand Cayman, where no such birds exist today. no meaningful measurements could be taken of Their extinction may be attributable to the same it. The larger one measures 36.1 mm in width factors that caused the extinction of other large across the condyles. If these two specimens rep¬ raptors and mammalian elements of the “mega¬ resent a single species, then there must have been fauna” in the West Indies. At least in the Baha¬ considerable sexual dimorphism in size. The holo- mas, Titanohierax could not have survived the type of A. borrasi is a tarsometatarsus lacking the virtual extirpation of Geocapromys, which must distal end. It has a proximal width of 22.4 mm, have been its principal prey (see account of Tyto which seems rather small, even for the smaller of pollens , p. 36). the two tibiotarsi just mentioned. It is nevertheless from a very large accipitrid, but one with a Family Pandionidae tarsometatarsus more gracile than that of Aquila and more similar to that of Titanohierax. Patter¬ son’s notes indicate that he regarded the Cuban Pandion haliaetus (Linnaeus) bird as a relative of Hypomorphnus ( = Buteogallus) Pandion haliaetus —. Wetmore 1, 938:52 [pre-Columbian mid¬ and that it was probably a species of Titanohierax. den C, rooked Island]. The species is clearly not referable to Aquila and we believe that it should be listed as Titanohierax Material Examined.— Crooked Island: 1 dis¬ borrasi (Arredondo) until there is sufficient fossil tal end of tibiotarsus (YPM 2437). material to decide questions of species-level tax¬ Remarks.— The osprey is found throughout onomy within Titanohierax. the Bahamas today and is resident on most is¬ That Titanohierax also occurred on Hispaniola lands. is demonstrated by a tarsometatarsus lacking the proximal end (USNM 244573) collected by Family Falconidae Charles Woods under 60 cm of red earth in Cueva de las Abejas, 6 km SE of Pedernales, Dominican Polyborus creightoni (Brodkorb) Republic, 24 April 1978. Olson compared this with the holotype of T. gloveralleni and found it to Caracara creightoni Brodkorb, 1959:353 [Pleistocene, New be somewhat smaller and less robust, with a less Providence], distinctly grooved middle trochlea, and a less Polyboru sfilancus —. Olson 1, 976a:36 3[holotyp eo fC c. reightonirestudied]. pronounced lateral rim of the outer trochlea in distal view. The total length of this specimen Material Examined.— New Providence: 1 ma¬ would have been about 135 mm, roughly com¬ jor metacarpal (UF 3153, holotype of Caracara parable to that of the holotype of T. borrasi. Its creightoni '), 1 distal end of tibiotarsus (USNM specific identity cannot yet be determined and it 283281), 1 quadrate lacking orbital process should be listed only as “ Titanohierax sp.” (USNM 283289). Further evidence that Titanohierax may have Remarks.— The species P. creightoni was based been widespread in the West Indies is offered by 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 o fP .plancus ,cranial view; c, same, distal view; b, distal end of left tibiotarsus of P. creightom, cranial view (USNM 283281) (arrow indi¬ cates broken bridge tha tdelim¬ 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. creightom ,media lview (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 creightom 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 * n = 1. served almost no diagnostic characters. Olson area on medial surface, between and ventral to (1976a) examined this specimen and adventured the two heads of the otic process, much more to agree that it was from a caracara but consid¬ deeply excavated. ered the evidence insufficient for the recognition The holotypical carpometacarpal fragment of of a species distinct from P. plancus, which at P. creightom is near the minimum size of that present occurs in Cuba, southern Florida, and element in P. plancus (Olson, 1976a), whereas the parts of the southwestern United States south to tibiotarsus and the quadrate from Banana Hole Tierra del Fuego. However, the two additional equal or exceed the maximum size in P. plancus. specimens from Banana Hole provide definite If these three specimens are from the same species proof, not only of a caracara in the Bahamas, but (and they could well be from the same individ¬ also that this bird was indeed distinct from exist¬ ual), then it would appear that P. creightom was a ing species. large caracara with reduced wings. The distal end of a tibiotarsus (Figure 2c,d) There is a possibility that P. creightom could shows the three openings diagnostic of most Fal- prove synonymous with P. latebrosus Wetmore conidae. It is near the maximum size for Polyborus (1920) from the Pleistocene of Puerto Rico, but plancus (Table 6), from which it differs in having P. latebrosus is as yet based only on two undiagnos¬ a markedly broader intercondylar sulcus (Figure tic fragments of wing bones that Olson (1976a) 2d). also considered to be insufficiently differentiated The most diagnostic specimen is a left quadrate from P. plancus to merit specific recognition. The (Figure 2g,h) that is considerably larger than in new material of P. creightom shows that there any of the specimens of P. plancus examined (Ta¬ definitely was at least one species of Polyborus ble 6). It also presents the following qualitative endemic to the West Indies in the Pleistocene, differences: (1) mandibular articulation in ventral and, therefore, the conspecificity of P. latebrosus view much broader craniocaudally, with the cau¬ with P. plancus must now be regarded as uncer¬ dal flange conspicuously expanded and the lateral tain. projection relatively narrower; (2) pneumatic for¬ Caracaras of the genus Polyborus are largely amen at medial base of orbital process larger; (3) terrestrial scavengers that characteristically in- NUMBER 48 31 habit open brushlands, savannas, and prairies. Providence: 1 complete coracoid, 1 complete and The presence of a caracara in the Banana Hole 1 distal end of humerus, 1 proximal end of ulna, deposits is thus indicative of such habitats in the 2 distal ends of tibiotarsi (USNM 283266-283269, Pleistocene of the Bahamas. PB 9044-9045); minimum number of individuals, .2 Remarks.— Wetmore (1937b:435) assigned the Falco sparverius Linnaeus fossils he examined to Capella [gallinago] delicata with “some mental reservation,” as they were Material Examined.— New Providence: 1 dis¬ larger than comparative material available to tal end of tibiotarsus (USNM 283265). him then. With additional fossils from Banana Remarks.— The American Kestrel is an un¬ Hole and more comparative material, we still common winter visitor from the North American find this to be the case. All of the fossils are larger mainland (F. s. sparverius) to most of the main (Figure 3) than C. gallinago delicata or C. g. para- islands of the Bahamas. Paulson (1966) showed guaiae (Table 7). The coracoids fall well within that the Cuban race, F. s. sparverioides, breeds at the size range of the South American species C. least on Great Inagua and San Salvador. Resident nobilis, although the humeri are somewhat forms of F. sparverius , divided into three subspe¬ smaller, thus being intermediate being C. nobilis cies, are found virtually throughout the remain¬ and C. gallinago. The two fossil tarsometatarsi are der of the Antilles. In the fossil, the shaft is even more equivocal, as they differ from each narrower than in six specimens of F. s. dominicensis other fairly considerably in size. or one specimen of F. s. canbaearum that we ex¬ There is other evidence of large snipe from the amined, and it more closely resembles F. s. spar¬ West Indies. Morgan (1977) reported a humerus verius. We did not examine skeletons of the Cuban and portions of a femur from fossil deposits on race F. s. sparverioides , however, and the single Cayman Brae that exceed C. gallinago in size. We fragmentary fossil is probably insufficient for sub¬ were able to examine these specimens and mea¬ surements of the humerus are included in Table specific identification in any case. 7. The latter specimen (UF 22968, Figure 3a) is larger than the fossil humeri from the Bahamas Family Rallidae and thus falls within the size range of humeri of C .nobilis. Porzana Carolina (Linnaeus) The taxonomic and distributional status of these fossil snipes is quite uncertain. The Com¬ mon Snipe, C. gallinago , breeds in North America Material Examined.— New Providence, Ba¬ mainly in boreal peat bogs and is found in the nana Hole: 1 distal end of humerus (PB 9043). West Indies as a southbound migrant, although Remarks.— In the Bahamas, the Sora Rail is a it may overwinter when and where suitable eco¬ fairly common visitor during the winter months. logical conditions persist. According to Tuck (1972:145), Capella nobilis “is restricted to the Family Scolopacidae montane region in northern and central Andes” in “dry steppe-like places of the Paramo Zone.” Snipes in general tend to breed in temperate Capella sp. zones or montane areas and do not have endemic Capella delicata .—Wetmore ,1937b:435 [Pleistocene ,‘"Great populations restricted to islands, except for the Exuma” = Little Exuma]. very aberrant genus Coenocorypha of the New Zea¬ land region. A supposed fossil species of snipe Material Examined.— Little Exuma: 2 com¬ from Puerto Rico (Wetmore, 1920) was later plete coracoids, 1 complete humerus (MCZ shown to be a woodcock, Scolopax anthonyi (see 2260); minimum number of individuals, 1. New Olson, 1976b). 32 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY What were the large fossil snipes from the cies now extinct? Were they endemic species that Bahamas and Cayman Brae? Were they migrant bred in the West Indies when North American individuals of one or more North American spe¬ biomes were depressed southward in the last gla- l r 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 delicala. (Scale = 3 cm.) Table 7.— Measurements (mm) of bones of snipe ( Capella) to show large size of West Indian fossils Character NUMBER 48 33 ciation? Do they represent large temporal forms literature it was regarded as a separate species. of C. gallinago or are they conspecific with snipes Brodkorb (1959) assigned fossil burhinid ma¬ now found only in South America? If these ques¬ terial from Banana Hole to a distinct species, tions can be answered, it will only be through the Burhinus nanus , which was distinguished from B. accumulation of additional fossil material. Re¬ b. domimcensis by its smaller size. Our material gardless, the presence of a snipe in the Pleistocene largely duplicates the elements that were used in of the Bahamas is indicative of open country, Brodkorb’s original description, and measure¬ probably grassy, with moist organic soil suitable ments of the new specimens are virtually identical for probing. to those given by Brodkorb (1959:355, table 1). Among the new specimens, the distal end of a Family Burhinidae humerus (PB 9047) measures 11.7 mm in width, whereas three specimens of B. b. domimcensis range Burhinus bistriatus nanus Brodkorb from 12.5 to 13.1 mm. The size differences be¬ tween B. b. nanus and B. b. domimcensis are real but Burhinus nanus Brodkorb ,1959:354 [Pleistocene ,New Provi¬ dence]. are not much greater than those between B. b. domincensis and B. b. bistriatus. Therefore, we have Material Examined.— New Providence: 2 scap¬ chosen to reflect these differences at the subspe¬ ular ends of coracoids, 1 proximal end of scapula, cific rather than specific level. Nonetheless, there 1 distal end of humerus, 1 proximal end of radius, definitely was a distinct, small form of Burhinus in 1 distal end of femur, 1 proximal and 1 distal the Bahamas in the Pleistocene, where none exists ends of tibiotarsi, 1 proximal and 2 distal ends of now. tarsometatarsi (USNM 283282-283284; PB The distribution of B. bistriatus domimcensis and 9046-9047, PB 9097, UF 3154-3156 [holotype B. b. nanus is obviously relictual and the presence and paratypes of Burhinus nanus], UF 25655); min¬ of the species in Central America would suggest imum number of individuals, 2. that some form of Burhinus probably occurred in Remarks.— The only member of the Burhini¬ Cuba in the past. This was confirmed by Olson, dae now found in the West Indies is Burhinus who in November 1980 briefly examined fossils bistriatus domimcensis, restricted to Hispaniola. This of a Burhinus from Pleistocene cave deposits in subspecies is markedly smaller than any of the Cuba in the collections of the Instituto de Zool- forms of B. bistriatus from the mainland of Central ogia of the Academia de Ciencias de Cuba and and South America, and in most of the earlier in the personal collection of Oscar Arredondo. Table 7.—Continued Character * n = 1. 34 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY The New World species of Burhinus are inhab¬ 283271-283273, 283376, 283397; PB 9059-9067; itants of open plains and prairies and may be UF 3163-3173, 3175-3182, 25643-25644); mini¬ significant ecological indicators when they appear mum number of individuals, 11. in the fossil record, as stated by Feduccia (1980) Remarks.— The White-crowned Pigeon is in describing a new species of Burhinus from the found throughout the Caribbean area and is still Pleistocene of Kansas. Thus, Burhinus bistnatus fairly common in the Bahamas. It is the most nanus strongly augments other information point¬ abundantly represented bird in the Banana Hole ing to the presence of open prairie-like habitat at deposits both in terms of specimens and in mini¬ the time of deposition of the Banana Hole fossils. mum number of individuals. Family Laridae Columba squamosa Bonnaterre Sterna fuscata Linnaeus Columb saquamos a—. Wetmor e1,937b:43 [5Pleistocen e“,Great Exuma” = Little E.xuma],—Brodkorb, 1959:355 [Pleisto¬ Sterna fuscata .—Wetmore ,1938:52 [Pre-Columbian midden, cene N, ew Providence], Crooke Idsland]. Corvus wetmore iBrodkorb ,1959:363 [part ;Pleistocene ,New Providence]. Material Examined.— Crooked Island: 1 com¬ plete and 1 partial coracoid (YPM 2438); mini¬ Material Examined.— Little Exuma: 1 com¬ mum number of individuals, 2. plete coracoid; 1 complete, 2 proximal, and 2 Remarks.— The pantropical Sooty Tern nests distal ends of ulnae; 1 complete carpometacarpus; today on scattered small cays in the Bahamas. 1 complete femur (MCZ 2259); minimum num¬ ber of individuals, 3. New Providence: 1 cranial Family Columbidae portion of sternum; 3 complete, and 2 scapular and 1 sternal ends of coracoids; 3 proximal ends Columba leucocephala Linnaeus of scapulae; 2 proximal ends of humeri; 2 com¬ plete, and 2 proximal and 1 distal ends of ulnae; Columb aleucocephala —. Wetmore 1, 938:5 2[par tp; re-Colum¬ 2 distal ends of radii; 1 nearly complete and 1 bian midden, Crooked Island],—Brodkorb, 1959:356 [Pleistocene N, ew Providence]. distal end of carpometacarpi; 3 proximal phalan¬ ges of major digit of wing; 1 proximal and 2 distal Material Examined.— Crooked Island: 1 com¬ ends of femora; 1 distal end of tibiotarsus; 1 plete humerus, 1 complete carpometacarpus, 1 complete and 2 proximal ends of tarsometatarsi complete tarsometatarsus (YPM 2439); mini¬ (USNM 283185, 283190, 283193, 283257- mum number of individuals, 1. New Providence: 283259, 283274-283276, 283378; PB 9048-9058; 2 complete, 4 nearly complete, and 3 scapular UF 3157-3162, 3215); minimum number of in¬ and 2 sternal ends of coracoids; 1 complete and dividuals, 3. 1 proximal end of scapula; 1 proximal and 2 Remarks.— We have confirmed the identifica¬ distal ends of humeri; 2 proximal and 1 distal tions of Wetmore and Brodkorb and found the ends of ulnae; 6 complete, 2 nearly complete, and Scaled Pigeon to be among the three most abun¬ 4 proximal and 9 distal ends of carpometacarpi; dant species of birds in Banana Hole. Size alone 1 proximal phalanx of major digit of wing; 4 permits this species to be separated from the proximal and 3 distal ends of femora; 4 distal smaller C. leucocephala in the majority of instances; ends of tibiotarsi; 2 complete, 2 nearly complete, in addition there are discernible qualitative dif¬ and 7 proximal ends of tarsometatarsi (USNM ferences between the two species in some ele¬ 283184, 283186, 283191-293192, 283194-283197, ments. The proximal end of a femur of this species 283234-283237, 283239, 283241, 283243, 283245- was included among the paratypes of Corvus wet¬ 283247, 283253-283255, 283260, 283262-283264, morei Brodkorb (1959). NUMBER 48 35 Columba squamosa is now distributed throughout by Brodkorb, the elements of these two species virtually all of the Antilles, except Jamaica, where are extremely difficult to separate, and indeed we it has strayed on occasion but otherwise appears were unable to identify certain specimens. Four to be replaced by C. fasciata (see Lack, 1976:238). of the skeletal elements could be segregated as C. squamosa , which is essentially a bird of forest follows: treetops, is unknown in the Bahamas today and Tarsometatarsus: Easily distinguished in its en¬ evidently has retreated from there since the Pleis¬ tirety, that of G. chrysia having a much longer, tocene. more slender shaft, and narrower proximal end, whereas in Z. aurita the bone is short and stout. Zenaida aurita (Temminck) Coracoid: Viewed laterally, G. chrysia has a smaller glenoid facet and the scapular facet is not Zenaida aurita.— Brodkorb ,1959:356 [part ;Pleistocene ,New situated as far sternally on the shaft as in Z. aurita. Providence], Humerus: Virtually indistinguishable except Material Examined.— New Providence: 1 that the ectepicondylar prominence in G. chrysia complete and 4 proximal ends of coracoids; 1 is usually closer to the external condyle than in Z. complete, 2 nearly complete, and 1 distal end of aurita. humeri; 2 shafts of carpometacarpi (USNM Carpometacarpus: In its entirety, the carpo¬ 283183,283187-283188, 283244, 283248, 283250, metacarpus of G. chrysia is shorter and stouter 283253, 283371; UF 3187, 3192, PB 9071); min¬ than that of Z. aurita ; also, the shaft of the major imum number of individuals, 4. metacarpal is straighter and slopes up proximally Remarks.— The Zenaida Dove is restricted to towards the pollical facet, whereas in Z. aurita the the West Indies, except for a subspecies inhabiting shaft of the major metacarpal is very slightly the Yucatan Peninsula and certain adjacent is¬ bowed and does not slope up as markedly to the lands. It occurs throughout the Bahamas. pollical facet. We were forced to regard the following speci¬ Geotrygon chrysia Salvadori mens as indeterminate, although they certainly pertain either to G. chrysia or to Z. aurita. New Oreopeli achrysia —. Wetmore 1, 937b:43 6[Pleistocene “, Great Providence: 1 scapula, 3 ulnae, 1 radius, 2 car¬ Exuma” = Little Exuma], pometacarpi, 3 femora, 2 tibiotarsi, and 1 tarso¬ Zenaida aurita .—Brodkorb ,1959:356 [part ;Pleistocene ,New metatarsus (USNM 283238, 283240, 283256, Providence]. 283372, 283373-283375; UF 3184, 3188, 3189, Material Examined.— Little Exuma: 1 com¬ 3191; PB 9070, 9072); minimum number of in¬ plete humerus and 1 complete tarsometatarsus dividuals, 3. (MCZ 2261). New Providence: 1 nearly complete Geotrygon chrysia , the Key West Quail-Dove, is and 2 proximal ends of coracoids, 3 distal ends of found in Cuba, Hispaniola, Puerto Rico, and the humeri, 1 complete and 2 proximal ends of car¬ Bahamas, where it has been recorded from Grand pometacarpi, 2 complete and 2 proximal ends of Bahama, Andros, New Providence, Eleuthera, tarsometatarsi (USNM 283189, 283233, 283242, and San Salvador, but not the Exumas. It is a 283249, 283251, 283270; UF 3183, 3185, 3186, forest bird that is found in several types of wood¬ 3190; 25645; PB 9068-9069); minimum number land but evidently is “most frequently encoun¬ of individuals, 3. tered in dry lowland forest and scrub, often on Remarks.— Two humeri, a coracoid, and a limestone” (Sorrie, 1979:728). The species is rare carpometacarpus that Brodkorb (1959) referred in parts of its range, including the Bahamas, and to Zenaida aurita appear to belong instead to G. evidently is becoming more restricted in distri¬ chrysia. Apart from the tarsometatarsus, which bution. It may once have bred in the Florida was not included among the material examined Keys, where it is now only accidental. (U. S. 36 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY records have been summarized in detail by Sorrie, Family Tytonidae 1979). Tyto pollens Wetmore Family Psittacidae Tyto pollens Wetmore, 1937b:436 [Pleistocene, “Great Ex- uma” = Little Exuma].—Brodkorb, 1959:357 [Pleisto¬ cene N, ew Providence]. Amazona leucocephala (Linnaeus) Material Examined.— Little Exuma: 1 nearly Amazon aleucocephala —. Wetmore 1, 938:5 2[pre-Columbian complete coracoid, 1 proximal end of ulna, 1 midden ,Crooked Island].—Brodkorb ,1959:356 [Pleisto¬ major metacarpal lacking proximal end, 1 com¬ cene N, ew Providence]. plete femur (holotype), 1 proximal end of tibio¬ tarsus (MCZ 2262, 2263); minimum number of Material Examined.— Crooked Island: 1 pre¬ individuals, 1. New Providence: 1 distal end of maxilla (YPM 2440). New Providence: 1 nearly carpometacarpus; 2 distal ends of tibiotarsi; 1 complete and 1 distal end of ulnae, 1 distal end complete, 1 proximal and 1 distal ends of tarso¬ of radius, 1 proximal and 1 distal ends of carpo- metatarsi, and 1 inner and 1 outer trochleae; 2 metacarpi, 1 proximal end of femur, 1 complete pedal and 1 ungual phalanges (USNM 283287- and 1 distal end of tarsometatarsi (USNM 283288; PB 9077; UF 3195-3199, 25646-25647, 283277-283279, 283290-283291; PB 9078, UF 25656); minimum number of individuals, 2. 3193, 3194); minimum number of individuals, 2. Remarks.— This giant, extinct barn owl ap¬ Remarks.— The Cuban Parrot occurs in Cuba pears to be the Bahaman counterpart of T. ostologa and the Cayman Islands, in addition to the Ba¬ Wetmore (1922) from the Pleistocene of Hispan¬ hamas, where the only extant populations are on iola, and T. noeli Arrendondo (1972) from the Abaco and Great Inagua. Worthington (Todd Pleistocene of Cuba. The exact taxonomic status and Worthington, 1911) obtained specimens on of these three owls is rendered difficult to deter¬ Acklins, where they no longer occur, and Bond mine by the relative scarcity of remains of T. (1956) also mentions Long and Fortune Islands pollens and because it has not yet been possible to in the historical range of the species. The fossil make direct comparisons with specimens of T. record indicates an even more extensive distri¬ noeli. Nevertheless, Wetmore’s diagnosis of T. pol¬ bution in the Bahamas; presumably much of the lens still holds: it is somewhat larger and the shafts reduction in range is due to human disturbance. of the limb elements are noticeably more robust than in T. ostologa (Table 8); also, the trochanteric Family Cuculidae 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. Saurothera merlini d’Orbigny ostologa or T. pollens (Table 8). Sympatric with T. noeli in Cuba was an even larger species of barn Material Examined.— New Providence: 1 dis¬ owl, T. riveroi (see Arredondo, 1976). tal end of humerus, 1 distal end of tibiotarsus, 1 The tremendous number of bones belonging to complete and 1 proximal end of tarsometatarsi rodents of the genus Geocapromys in the Banana (USNM 283198, 283199; PB 9073, 9096); mini¬ Hole deposits is almost surely the result of pre¬ mum number of individuals, 1. dation by Tyto pollens. Geocapromys has disappeared Remarks.— In the Bahamas, the Great Lizard from all of the islands of the Bahamas except East Cuckoo is known to be resident only on New Plana Cay, a tiny arid islet near Acklins Island in Providence, Eleuthera, and Andros; it is found the southern Bahamas, where the only living elsewhere in Cuba and on the Island of Pines. 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 population is well adapted to dry, scrubby envi¬ Remarks.— The Barn Owl today is a rather ronment (Clough, 1972). It is possible that late uncommon resident known from most of the Pleistocene reduction of arid environments in the larger Bahaman islands, including New Provi¬ Bahamas, combined with the later effects of pre¬ dence. dation and habitat alteration by man, may have been responsible for the near total extirpation of Family Strigidae 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¬ Athene cunicularia (Molina) cipal prey. Speotyto cunicularia. —Wetmore, 1937b:440 [Pleistocene, “Great Exuma” = Little Exuma]. Tyto alba (Scopoli) Glaucidium dickinsom Brodkorb ,1959:358 [Pleistocene ,New Providence]. Tyto alba. —Brodkorb, 1959:358 [Pleistocene, New Provi¬ Otu sprovidentiae Brodkorb 1, 959:360 [Pleistocene N, ew Prov¬ dence). idence], Material Examined.— New Providence: 1 pre¬ Material Examined.— Little Exuma: 1 proxi¬ maxilla, 1 distal end of humerus, 1 distal end of mal and 1 distal ends of tarsometatarsi (MCZ tarsometatarsus (USNM 283286; UF 3200, 3201); 2264). New Providence: 4 complete and 4 par¬ minimum number of individuals, 1. tial coracoids, 1 proximal and 2 distal ends of 38 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY humeri, 3 proximal ends of ulnae, 1 complete and are quite certain that none of the fossil strigid 2 proximal ends of carpometacarpi, 1 complete bones collected thus far from the Bahamas are and 4 proximal ends of femora, 1 proximal and referable to Glaucidium. 10 distal ends of tibiotarsi, 3 proximal and 5 In addition to the holotype, Brodkorb (1959) distal ends of tarsometatarsi (USNM 283201- referred four other bones to Otus providentiae (prox¬ 283228, 283367-283370; PB 9074-9076; UF 3202 imal end of tibiotarsus, incomplete coracoid, [holotype of Glaucidium dickinsoni] and 3203-3207 proximal end of ulna, and proximal end of car- [holotype and paratypes of Otus providentiae]); pometacarpus). Because we have much more ad¬ minimum number of individuals, 6. ditional strigid material from Banana Hole and Remarks. —The Burrowing Owl, Athene cuni- because the problem of separating Athene cunicu¬ cularia, is the only strigid owl that occurs in the laria from Otus also arises with fossils from Puerto Bahamas today. It has been found on most of the Rico and Cuba, we have undertaken to point out northern islands, including the Exumas and New the differences between these taxa in some detail. Providence, but is known in the southern Baha¬ Although the differences between A. cunicularia mas only on Samana Cay and Great Inagua and Otus are not as striking as those separating (Buden, 1979). Paulson (1966) gives details con¬ these two genera from Glaucidium , they are nev¬ cerning the apparently changing status of the ertheless quite pronounced. species on various islands of the Bahamas. Coracoid (Figure 4): Otus with procoracoid pro¬ When all 32 of our newly collected specimens cess narrow and more nearly parallel-sided, not of small owls from Banana Hole proved to belong broad and triangular as in A. cunicularia ; shape of to this species, we became highly suspicious of the head very different, in ventral view more bulbous identity of the two species of Strigidae named in A. cunicularia and lacking process that extends from this site by Brodkorb (1959). Examination towards and sometimes fuses with procoracoid in of the type material of Otus providentiae (based on Otus; in dorsal view, glenoid facet extends much 5 specimens) and Glaucidium dickinsoni (based on farther laterally than in Otus. 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 Figure 4.—Scapular end of left coracoid in dorsal view: a, pygmy owls of the genus Glaucidium is very distinct Athene cunicularia ;b ,Otus nudipes .(Lines indicate some o fthe from that of either Otus or Athene cunicularia. We difference sdiscussed in text.) NUMBER 48 39 Humerus (Figure 5): Bicipital surface more Ulna: Internal cotyla in proximal view not extensive in A. cunicularia; deltoid crest more tri¬ distinctly pointed internally in A. cunicularia; ra¬ angular, less rounded and not extending as far dial depression deeper and better defined than in down the shaft as in Otus; bicipital furrow rela¬ Otus. tively larger and extending farther internally in Carpometacarpus: Proximal end in A. cunicu¬ Otus; entepicondylar area in A. cunicularia not as laria not rotated relative to major and minor pointed and not extending as far medially as in metacarpals, and nearly parallel with them, re¬ Otus; ectepicondylar process situated farther prox- sulting in pisiform process being on a line with imally than in Otus; brachial depression not as middle of shaft of major metacarpal, not elevated long and narrow. above it as in Otus. Femur: Trochanteric ridge in A. cunicularia much better developed than in Otus , extending farther distally and somewhat swollen at distal end. Tibiotarsus (Figure 6): Tendinal groove in A. cunicularia deeper, distinctly undercutting medial side of intercondylar bar, unlike Otus; condyles in cranial view nearly parallel, whereas in Otus the external condyle is tilted medially at the distal end; caudal rims of proximal articular surface more undercut. Tarsometatarsus (Figure 7): Cranial surface of proximal end in A. cunicularia not deeply exca¬ vated; external margin of shaft sloping steeply caudolaterally, not extending equally far cra- A B A B Figure 5. —Proximal (top row) and distal (bottom row) portions o flef thumerus in palmar view :a ,Athene cunicularia-, Figure 6.— Distal end of left tibiotarsus in cranial view: a, b, Otus nudipes. (Lines indicate some of the differences dis¬ Athene cunicularia- ,b ,Otus nudipes .(Lines indicate some o fthe cussed in text.) difference sdiscussed in text.) 40 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY nially as the internal margin, as in Otus; attach¬ ment of M. tibialis cranialis usually divided in A. cumcularia; calcaneal ridge of hypotarsus usually shorter and always located farther proximally than in Otus; shape and orientation of each of the three trochleae decidedly different in the two genera— A. cumcularia with outer trochlea in distal view smaller, caudal flange not extending as far caudally and medially as in Otus; external rim of middle trochlea in distal view extending farther caudally than internal rim, whereas they are of the same extent in Otus. On the basis of the characters outlined above, we identify the types of Otus providentiae and all other strigid fossils from Banana Hole as burrow¬ ing owls. Thus, at the species level, Glaucidium dickinsoni Brodkorb, 1959, and Otus providentiae Brodkorb, 1959, become junior subjective syn¬ onyms of Athene cumcularia (Molina, 1782). The subspecies of burrowing owl currently in¬ habiting the Bahamas is now considered to be the same as that found in southern and central Flor¬ ida, Athene cuniculana flondana (Ridgway, 1874). Although two names have been applied to the Bahaman populations (bahamensis Cory 1891, type-locality Inagua; and cavicola Bangs 1900, type-locality New Providence), these have long been regarded as synonymous with flondana (e.g., Ridgway, 1914; Buden, 1979). On the basis of limited modern skeletal material, it appears that the Pleistocene remains of A. cuniculana from the Bahamas are consistently smaller than either A. c. flondana or A. c. troglodytes of Hispaniola (Table 9). It may well be that the Bahamas once had an endemic subspecies that became extinct and that flondana colonized the islands subsequently. Should the fossil burrowing owl from the Baha¬ mas prove distinct, either of Brodkorb’s names could be applied to it, as they were proposed simultaneously. As first revisors, we choose the name providentiae , in the combination Athene cum- Figure 7. —Proximal end in cranial view (top row), proximal end in caudal view (middle row), and distal end in distal view (bottom row) of left tarsometatarsus :a ,Athene cunicu¬ lana- ,b ,Otus nudipes .(Lines indicate some of the differences discussed in the text.) A B NUMBER 48 41 Table 9 .—Measurements (mm) of burrowing owls ,Athene cumculana ,to show smaller size of the Bahaman fossils as compared to extant taxa in Florida and Hispaniola Character * n = 3. culariaprovidentiae (Brodkorb, 1959), for this taxon. Bahamas and Hispaniola, with extinct subspecies For the present, however, the subspecific status of known historically from Antigua and Nevis (A. c. the Bahaman fossils remains uncertain. amaura ) and Marie Galante ( A. c. guadeloupensis). The burrowing owl is usually placed in the Previously undetected breeding populations are monotypic genus Speotyto. It has been separated now known from three widely separated localities from the Old World species of the genus Athene in Cuba (western Pinar del Rio, northeast Me- primarily by the greatly lengthened tarsometa- tanzas, and in the eastern part of the island near tarsus, correlated with the highly terrestrial and Guantanamo) with specimens of uncertain breed¬ fossorial habits of the bird. We have not detected ing status having been taken on Cayos Coco and other osteological differences of generic value and Guillermo off northwestern Camagiiey (Gonzalez we prefer the more recent treatment synonymiz- and Garrido, 1979). The birds breeding in Pinar ing Speotyo with Athene, as this emphasizes the del Rfo are close to, or inseparable from, A. c. affinities of the species rather than its differences floridana (see Garrido and Garcia, 1975), but those (cf. Mayr and Short, 1970). from Guantanamo are distinct and probably rep¬ Various subspecies of Athene cumcularia occur in resent an unnamed subspecies (specimens exam¬ Florida and western North America, through por¬ ined by Olson in Havana in February 1978 and tions of Central and South America as far south November 1980). This population is probably a as Tierra del Fuego. Until recently, this species relictual subspecies that was more widely distrib¬ was known in the West Indies only from the uted on the island in the past, as fossils of A. 42 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY cunicularia are known from cave deposits near taxon the fossil may belong. This is the first fossil Daiquiri (Olson and Hilgartner, pers. observ.). record of Chordeiles from the West Indies. Fossils of burrowing owls are also known from Jamaica (Olson and Steadman, 1977), Barbuda (specimens examined at Florida State Museum), Family Trochilidae Mona Island (specimens examined at American Museum of Natural History) and Cayman Brae Chlorostilbon ricordii (Gervais) (Morgan, 1977). A very small form, possibly a distinct species, is common in certain of the older Material Examined.— New Providence: 1 Pleistocene cave deposits in Puerto Rico (Pregill- complete humerus; 1 complete ulna (USNM and Olson, 1981). 283393, 283394). Burrowing owls are found in open, often tree¬ Remarks. —Both of the above elements are less, grassland and desert areas in which there is readily distinguished from those of Philodice eve- soft, sandy substrate suitable for burrowing. The lynae , the only other hummingbird now residing species is thus a prime ecological indicator. The in the Bahamas, by their larger size. The shape abundance of fossils of this species in Banana and position of the internal tuberosity and ecte- Hole is probably indicative of prairie-like or dry picondylar process of the humerus are also very scrub habitat during the time of deposition. The different in these hummingbirds (Figure 8). The continued (or possibly resumed) presence of A. nominate subspecies of the Cuban Emerald Hum¬ cunicularia in the Bahamas is perhaps a reflection mingbird occurs in Cuba and the Isle of Pines. of the xeric nature of the present environment. The Bahaman subspecies, C. r. bracei , is generally The more extensive distribution of Athene cunicu¬ restricted to Grand Bahama, Abaco, and Andros, laria in the West Indies in the late Pleistocene and and is now only accidental on New Providence. its subsequent withdrawal into a few relict pop¬ Because of the proximity of New Providence to ulations are strong evidence that much of the Andros, it is perhaps not surprising that C. ricordii West Indies was characterized in the Pleistocene was once present on the former. by a predominance of open, arid habitat that has largely disappeared since the last glacial period. Family Caprimulgidae Chordeiles sp. Material Examined.— New Providence: 1 dis¬ tal end of humerus (USNM 283280). Remarks. —The resident forms of nighthawks in the West Indies are now often considered as forming a separate species, Chordeiles gundlachu , a b c distinct from C. minor of the American mainland (e.g., Eisenmann, 1962). Both occur in the Ba¬ hamas, the latter as a migrant. No comparative Figure 8.— Right humeri of modern and fossil humming¬ material of C. gundlachu was available to us, so we birds in ancona lview: a ,modern Philodice evelynae\ B ,fossil were unable to determine if there are osteological Chlorostilbon ricordi ifrom New Providence (USNM 283393); differences between the two species, or to which c ,modern Chlorostilbon ricordi ibracei .(Scale in mm.) NUMBER 48 43 Trochilidae genus and species indeterminate from undetermined remains from a cave deposit in Brazil (Winge, 1888). Material Examined. —New Providence: 1 ma¬ jor metacarpal, 1 distal end of carpometacarpus Family Picidae (USNM 283395, 283396); minimum number of individuals, 1. Remarks.— Although fragmentary, these two Melanerpes superciliaris (Temminck) fossils are unquestionably from a species of hum¬ Centurus superciliaris.— Wetmore ,1937b:440 [Pleistocene, mingbird larger than any found in the Bahamas “Great Exuma” = Little Exuma], today (Figure 9). They are less robust than in Bathoceleus hyphalus Brodkorb ,1959:362 [Pleistocene ,New Anthracothorax dominicus of Hispaniola and Puerto Providence]. Rico, the geograhically nearest large humming¬ Melanerpe ssuperciliaris —. Brodkorb 1, 959:363 [Pleistocene, bird to the Bahamas. They are somewhat more Ne wProvidence]. similar to Sericotes holosericeus of the Lesser Antilles, Material Examined. —Little Exuma: 1 com¬ but their fragmentary condition does not permit plete tarsometatarsus (MCZ 2266). New Provi¬ identification other than to family. The speci¬ dence: 3 incomplete coracoids; 1 complete, and 1 mens nevertheless represent a species that has proximal and 1 distal ends of humeri; 1 distal become extinct in the Bahamas. end of ulna; 1 proximal end of femur; 1 proximal Anthracothorax dominicus has been recorded from end of tibiotarsus (USNM 283229-283231, the Pleistocene of the Dominican Republic (Bern¬ 283379-283380; PB 9080; UF 3209 [holotype of stein, 1965); this and the hummingbirds from Bathoceleus hyphalus ], 3208, 25648); minimum Banana Hole now constitute the only published number of individuals, 2. fossil record for the family Trochilidae, apart Remarks.— Wetmore (1937b) identified a tar¬ sometatarsus from Little Exuma as Melanerpes superciliaris. Brodkorb (1959) referred the distal portion of a humerus from New Providence to this species and at the same time described a new genus and species of woodpecker, Bathoceleus hy¬ phalus, based on a single, rather abraded coracoid. This specimen is hardly sufficient for the estab¬ lishment of a new genus and comparisons of it are still hampered by the lack of skeletal material of any of the three Bahaman subspecies of M. superciliaris. For comparisons, Brodkorb evidently had only a single unsexed specimen of M. s. superciliaris from Cuba (USNM 347972), which we also examined, augmented by eleven skeletons of M. s. caymanensis recently collected by Olson on Grand Cayman. Our assessment of the generic characters of Figure 9.— Left carpometacarpi of modern and fossil hum¬ Bathoceleus, as given by Brodkorb (1959:362), is as mingbirds in internal view; a, modern Philodice evelynae: b, follows: modern Chloroslilbon ncordu bracer c, f,oss ilTrochilidae g, enus and species indeterminate, from New Providence (USNM 1. “Coracoid with internal margin of upper 283396) (note larger size than in the two extant species of end of shaft nearly straight (as in Xiphidiopicus\ the Bahamas). (Scale in mm). curved medially toward brachial tuberosity in 44 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Colaptes, Nesoceleus, and Melanerpes ).” This was not respect the fossil clearly differs from Colaptes and evident to us; the seeming straightness appears to agrees with Melanerpes. be an illusion caused by the nearly complete loss We conclude that the holotype of Bathoceleus of the brachial tuberosity through abrasion. hyphalus presents no evidence for the existence of 2. “Pneumatic foramina not extending from an extinct genus of woodpeckers in the Bahamas. region of brachial tuberosity onto upper end of We regard Bathoceleus Brodkorb, 1959, as a junior posterior face of shaft (resembling Nesoceleus and synonym of Melanerpes Swainson, 1831. Brodkorb Melanerpes-, area pneumatic in Colaptes and Xiphi- maintained that Bathoceleus hyphalus was larger diopicus) .” This character does not separate Bath- than Melanerpes superciliaris, but in fact the holo¬ oceleus from Melanerpes. The area in question is type is about the same length as, and is decidedly nearly worn away in the holotype and the amount less robust than, the coracoid in our single un- of intraspecific variation in pneumatization is sexed specimen of M. s. superciliaris, although it is rather variable in these genera in any case. much larger than in M. s. caymanensis. All of the 3. “A pneumatic foramen between scapular other Bahaman fossil specimens we have referred facet and procoracoid (foramen present in Co¬ to M. superciliaris are considerably larger than M. laptes, Nesoceleus, and Xiphidiopicus-, absent in Me¬ s. caymanensis, except for the tarsometatarsus from lanerpes) .” We believe that the pit in this area in Little Exuma, which is about the size of males of the holotype of Bathoceleus hyphalus is an artifact M. s. caymanensis, but more robust. The three of wear. There is usually a fairly well-developed existing Bahaman subspecies, based on measure¬ foramen in this position in Colaptes, but in one ments of skins, are smaller than the nominate specimen of Nesoceleus [= Colaptes ] fernandinae (the race and larger than M. s. caymanensis. same specimen examined by Brodkorb) this for¬ The West Indian Red-bellied Woodpecker oc¬ amen is minute on the left coracoid and absent curs on Cuba, Grand Cayman, and on three on the right. Its presence or absence is obviously islands in the Bahamas (Figure 10): M. s. baha- not of generic significance. mensis on Grand Bahama, M. s. blakei on Abaco, 4. “Lower part of shaft slightly compressed in and M. s. nyeanus on San Salvador. With the fossil medial view (as in Melanerpes and Xiphidiopicus ; evidence from New Providence and Little Ex¬ expanded in Colaptes and Nesoceleus) .” Again, this uma, it becomes quite evident that these disjunct character does not separate Bathoceleus from Me¬ populations are relicts and that the species was lanerpes. It was not evident to us in any case. more generally distributed in the Bahamas in the 5. “Lower part of shaft narrowly V-shaped in lateral view (concave on posterior face in allied genera).” The shaft in lateral view is decidedly narrower as it descends to the sterno-coracoidal process in Colaptes, but this is not true of Mela¬ nerpes. In this respect the fossil actually agrees with Melanerpes, contra Brodkorb. 6. “Scar of coracobrachialis longer and shal¬ lower than in allied genera.” Variation within our series of M. s. caymanensis encompasses a sim¬ ilar configuration of this scar. 7. “Anterior sternal facet curved (as in Xiphi¬ diopicus-, straight in Colaptes, Nesoceleus, and Mela¬ nerpes ).” This facet is in fact curved in Melanerpes, Figure 10. — Modern and fossil distribution of Melanerpes in which it is also longer and extends farther superciliaris in the Bahamas .(Dagger (f )indicates Pleistocene medially, onto the internal distal angle. In this occurrences.) NUMBER 48 45 Pleistocene, when it occurred on the Great Ba¬ We had only one unsexed skeleton of C. fernandinae hama Bank, where it is now absent. The bird is for comparison, and its bones were somewhat possibly extinct on Grand Bahama (Bond, 1980) larger and definitely stouter than the Bahaman but persists in low numbers on Abaco. On both fossils. On the other hand, the tibiotarsus from islands it apparently prefers coastal broadleaf Little Exuma was larger and stouter than in a scrub. On San Salvador the species is found in single unsexed example of C. a. chrysocaulosus. We somewhat mesic broadleaf forest bordered by are not prepared to say that some or all of the scrub and often by mangroves; it nests mainly in Bahaman flicker bones could not be from a rep¬ the palm Sabal palmetto. Miller (1978:285) esti¬ resentative of C. fernandinae. Regardless, the genus mates the population to be between 100 and 160 no longer occurs in the Bahamas. pairs, but gives no details of how this figure was Barbour (1943:87) considered C. a. chrysocaulo¬ derived. Our experience in San Salvador leads us sus in Cuba to be a rare bird that “loves the wide to believe this to be an overestimation. The de¬ savannas and open pastures with scattered groves crease and fragmentation of populations of M. of guasimas and other shabby trees which struggle superciliaris in the Bahamas could be explained by along on sterile lands.” Garrido and Garcia (1975: the deterioration of broadleaf habitats and their 77) mention only forests (“ bosques”) as its habitat. subsequent replacement by other vegetation. On Grand Cayman, C. a. gundlachii occurs in low, dense broadleaf forest and in mangroves. Accord¬ ing to Barbour (1943:88), C. fernandinae was “lo¬ Colaptes sp. cally abundant only in the dry pastures of retired communities in southern Santa Clara and Ca- Colapte schrysocaulosus —. Wetmore 1,937b:43 9[Pleistocene, “Great Exuma” = Little Exuma], magiiey;” Bond (1956:100) lists “open country” in Las Villas as well; and Garrido and Garcia Material Examined.— Little Exuma: 1 distal (1975:77) say that it is now a rather rare inhab¬ end of humerus, 1 distal end of tibiotarsus (MCZ itant of savannas, swamp, and forest edge. Thus, 2265). New Providence: 1 proximal end of ti¬ the extinction of Colaptes in the Bahamas could biotarsus (PB 9079). well be correlated with the apparent loss of sa¬ Remarks.— We have confirmed Wetmore’s vanna habitats since the Pleistocene. (1937b) identification of two bones from Little Exuma as Colaptes, but refrain from any further Family Tyrannidae refinement of their taxonomic status. The proxi¬ mal end of a tibiotarsus from Banana Hole was Myiarchus sagrae (Gundlach) clearly separable from Melanerpes and we have referred it to Colaptes as well. Material Examined.— New Providence: 2 Three taxa of flickers currently reside in the scapular ends of coracoids (USNM 283381, West Indies, the systematics of which has been 283382); minimum number of individuals, 1. dealt with by Short (1965): Colaptes (Nesoceleus Remarks.— The presence of pneumatic foram¬ auct.) fernandinae of Cuba; C. auratus chrysocaulosus , ina in the scapular end of the coracoid separates also of Cuba; and C. auratus gundlachu , a small tyrannids from other passerines occurring in the subspecies endemic to Grand Cayman. The last Bahamas. The fossils agree in size with M. sagrae two were long regarded as constituting a species lucaysiensis, which is a fairly common woodland (C. chrysocaulosus ) separate from C. auratus of the resident in the northern islands but is known only North American mainland. from Crooked, Acklins, and Inagua in the south¬ Wetmore’s (1937b) assignment of the Exuma ern Bahamas (Buden, 1979). We have followed bones to C. chrysocaulosus, as opposed to C. auratus, Lanyon’s (1967) revision of West Indian Myiarchus was doubtless based on geographical probability. in separating M. sagrae of Cuba, Grand Cayman, 46 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY and the Bahamas, from the taxa formerly in¬ phalus from three pre-Columbian archeological cluded under M. stolidus. sites in St. Croix, Virgin Islands, and one in Puerto Rico. We have examined these specimens Family Corvidae and confirmed the identifications. C. leucogna¬ phalus is not known in historic times from St. Croix and it became extinct in Puerto Rico dur¬ Corvus nasicus Temminck ing this century. Corvus nasicus .—Wetmore ,1937b:440 [Pleistocene ,“Great Wetmore (1920) named a new species of crow, Exuma” = Little Exuma.]. Corvus pumilis, based on a complete ulna from a Corvus sp.—Wetmore, 1938:52 [pre-Columbian midden, Pleistocene cave deposit in Puerto Rico. Later, Crooke Idsland]. Wetmore (1937a) referred a tibiotarsus from a Corvus welmore iBrodkorb ,1959:363 [part ;Pleistocene ,New Providence], midden deposit on St. Croix to the same species. C. pumilis was diagnosed on the basis of size, being Material Examined. —Little Exuma: 1 com¬ larger than C. palmarum and smaller than C. leu¬ plete (juvenile) and 1 nearly complete humerus, cognaphalus. 1 nearly complete ulna, 1 distal end of tibiotarsus Corvid fossils from Exuma were assigned by (MCZ 2267-2268); minimum number of individ¬ Wetmore (1937b) to C. nasicus, whereas a tibio¬ uals, 2. Crooked Island: 1 distal end of tibiotar¬ tarsus from Crooked Island he (Wetmore, 1938: sus (YPM 2441). New Providence: 2 complete 52) referred to as “Corvus sp.,” stating that it was and 1 scapular end of coracoids; 1 shaft, 1 prox¬ from a bird the size of females of C. leucognaphalus imal and 1 distal ends of humeri; 1 distal end of and probably belonged to the same species as the ulna; 1 complete, 1 shaft, 4 proximal and 3 distal Exuma material he had identified as C. nasicus. ends of carpometacarpi; 1 proximal end of femur; Brodkorb (1959) proposed a new species, Corvus 1 proximal and 4 distal ends of tibiotarsi; 1 wetmorei , based on nine fossils from Banana Hole; proximal and 3 distal ends of tarsometatarsi; 1 this was diagnosed as being smaller than C. leu¬ subterminal pedal phalanx; 1 unguis (USNM cognaphalus and by alleged differences in the en- 283292-283293, 283295-283300, 283302-283303, tepicondyle of the holotypical humerus. Brod¬ 283390; PB 9081-9083, 9085-9088, 9091; UF korb’s assessment of this species was clouded by 3210-3213, 3216-3218 [holotype and paratypes his considering C. nasicus to be a subspecies of C. of C. wetmorei ]; UF 25649); minimum number of leucognaphalus , which it definitely is not. individuals, 4, possibly 5. At the time Wetmore described Corvus pumilis, Remarks.— In our analysis of the corvid ma¬ he had no comparative skeletal material of C. terial from the Bahamas we have re-examined all nasicus and he did not even mention this species of the fossil and archeological specimens from the in his discussion. Brodkorb had only a few bones West Indies previously assigned to Corvus. Four of C. nasicus, taken from skins, for comparison species of crows inhabit the West Indies at present with C. wetmorei. Until very recently there simply (Figure 11): Corvus jamaicensis in Jamaica; C. leu- was no useful skeletal material of C. nasicus avail¬ cognaphalus of Flispaniola and Puerto Rico; C. able for examination. Fortunately, Olson ob¬ palmarum in Hispaniola, with scattered, relict pop¬ tained six skeletons in Cuba in October 1979; ulations in Cuba; and C. nasicus in Cuba and the these were used in the present study along with Caicos Islands in the southern Bahamas. two partial skeletons already on hand. Measure¬ We have not seen skeletal specimens of C. ments of these specimens show that C. nasicus is jamaicensis , but it is a small species, the size of C. intermediate in size between C. leucognaphalus and palmarum or smaller. Corvus leucognaphalus is the C. palmarum, except in the length of the tarsome- largest of the three remaining species. Wetmore tatarsus, which is nearly equal in all three species (1918, 1925, 1937a, 1938) recorded C. leucogna¬ (Table 10). Of the species on the adjacent main- NUMBER 48 47 Figure 11.—Modern and fossil distribution of Corvus in the West Indies. (Dagger (f) indicates Pleistocen eo srubfoss iol ccurrences.) land, the Common Crow, Corvus brachyrhynchos , is There is as yet no evidence in the fossil record larger than in any of the Bahaman fossils, and to substantiate the existence of any extinct species the Fish Crow, Corvus ossifragus, differs from the of Corvus in the West Indies. We regard Corvus fossil specimens in numerous qualitative features. wetmorei Brodkorb, 1959, as a junior synonym of The lengths of the two specimens of C. pumilis Corvus nasicus Temminck, 1826. C. pumilis Wet- fall within the range of size variation in C. nasicus more, 1920, could perhaps be added under C. (Table 10). The shaft of the tibiotarsus assigned nasicus , although it might also be a larger repre¬ to C. pumilis is somewhat more slender than in the sentative of C. palmarum. available specimens of C. nasicus, but otherwise it As the Cuban Crow, C. nasicus, still occurs in presents no apparent features of specific value. the Bahamas, although only in the Caicos Islands, Likewise, the shaft of the holotypical ulna is it is perfectly reasonable that most Bahaman somewhat more slender than in C. nasicus and in fossils of Corvus would prove to be of this species. proximal view the external cotyla is less hooked. In Cuba, C. nasicus occurs sympatrically, but ap¬ These differences are difficult to evaluate with parently not syntopically, with C. palmarum. The such poor fossil representation. two species may have been in more direct associ¬ Most of the type material of Corvus wetmorei, ation on New Providence, as the remains of both including the holotype, is within the size range of were found in Banana Hole. Of greater interest C. nasicus. The exceptions are the two paratypical is the possible occurrence of C. nasicus in Puerto femora, one of which (UF 3215) we believe to Rico and the Virgin Islands, where it presumably belong to Columba squamosa , and the other to Corvus was sympatric with C. leucognaphalus. If so, this palmarum (see following account). We find no would dispel any further consideration of the idea consistent differences between the entepicondyle that C. nasicus and C. leucognaphalus are conspe- of the holotype of C. wetmorei and that in speci¬ cific, as had been suggested in some of the earlier mens of C. nasicus. literature. It should also provide a good example 48 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Table 10. — Skeletal measurements (mm) of modern and fossil Corvus from the West Indies of the perils of “lumping” seemingly allopatric Material Examined. —New Providence: 1 insular species solely on the basis of present zoo- proximal and 1 distal ends of carpometacarpi, 1 geographical patterns. complete femur, 1 distal end of tibiotarsus, 2 In Cuba today, Corvus nasicus is an inhabitant distal ends of tarsometatarsi (USNM 283294, of forests and scrub, and in the Caicos it may be 283301, 283304; UF 3214 [paratype of C. wet- found in dry scrub and dry forest (M. H. Clench, more *]; PB 9084, 9089-9090); minimum number pers. comm.). It is evident that the species is in of individuals, 1. the process of retreating from a formerly greater Remarks.— We have identified six specimens range that once included most of the Bahamas, of Corvus from Banana Hole as C. palmarum by and possibly Puerto Rico and the Virgin Islands their distinctly smaller size (Table 10). This is the as well. One can only speculate as to whether it first fossil record for the species. One of the spec¬ may have occurred on Hispaniola, where two imens (femur UF 3214) was included among the other species of crow co-exist at present. type material of Corvus wetmorei ; in Brodkorb’s (1959:365) table of measurements it stood out as being much smaller than C. leucognaphalus , al¬ Corvus palmarum Wiirttemberg though Brodkorb did not include data for C. Corvus wetmore iBrodkorb ,1959:363 [part ;Pleistocene ,New palmarum. Providence]. The Palm Crow is most numerous in Hispan- NUMBER 48 49 Table 10.—continued Character 1 Pleistocene Bahamas ,include sC .wetmorei. 2Pleistocen eBahamas. 3 n = 1. iola, where it evidently occurs in a variety of Mimus gundlachii .—Brodkorb ,1959:365 [Pleistocene ,New habitats, including forest and open areas of pine Providence]. and palm, as well as in arid and scrubby regions Material Examined.— Crooked Island: 1 com¬ (Wetmore and Swales, 1931). It is a very rare bird in Cuba, found in arid wastelands with pines plete tibiotarsus (YPM 2442). New Providence: (Barbour, 1943), with disjunct populations in 1 complete and 4 scapular ends of coracoids; 1 Pinar del Rio and Camagiiey (Garrido and Gar¬ complete, 8 proximal and 9 distal ends of humeri; cia, 1975). As with C. nasicus, the fossil record 1 complete, 2 nearly complete, 4 proximal and 2 indicates that C. palmarum is declining, which distal ends of ulnae; 3 complete and 1 distal end would explain the relictual nature of the present of carpometacarpi; 1 complete and 2 proximal distribution of the species in Cuba. ends of femora; 7 proximal and 1 distal ends of tibiotarsi; 2 complete, 2 proximal and 1 distal ends of tarsometatarsi (USNM 283305-283308, Family Mimidae 283314-283316, 283337-283338, 283345-283346, 283349-283352, 283354-283355, 283362-283363, Mimus gundlachii Cabanis 283383-283389, PB 9092-9095; UF 3219, 3220, Mimus gundlachii .—Wetmore ,1938:52 [Pre-Columbian mid¬ 25650-25653); minimum number of individuals, den ,Crooked Island], 9. 50 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Remarks.— This is a common bird in most of coracoids, 1 proximal end of femur (USNM the Bahamas, but it is more abundant on the 283398-283400); minimum number of individu¬ drier southern islands. It is conspicuous and not als, 2. at all shy. Evidently, birds of prey found the Remarks.— The mandibular ramus can be Bahaman Mockingbird easy to capture, as it is identified as belonging to a species of Dendroica the best represented passerine in the Banana Hole (as opposed to other genera of Parulidae) about deposits. The species is practically endemic to the the size of D. virens. Neither this specimen nor the Bahamas except for populations on a few of the other material is sufficient for specific identifica¬ keys off northern Cuba and a subspecies endemic tion, nor can it even be determined with certainty to the driest portions of Jamaica. In the Bahamas that more than one species is present, although it may be found in a variety of habitats, including the two coracoids are sufficiently different in size very arid ones. The modern distribution of the as to suggest this. At least 16 species of Dendroica , species might suggest that it was once more widely both resident and migrant, have been recorded distributed and was originally adapted to xeric from the Bahamas. environments. This is confirmed by fossils of M. gundlachu from Pleistocene cave deposits in Puerto Rico (Pregill and Olson, 1981), which contain other species indicative of more open, xeric hab¬ Family Thraupidae itat than found in Puerto Rico today. That the species persists in the Bahamas may perhaps be attributable to these islands having been more consistently arid through the Pleistocene and Coereba flaveola (Linnaeus) Holocene. Material Examined.— New Providence: 1 Family TURDIDAE complete, 1 proximal and 1 distal ends and 1 Turdus plumbeus Linnaeus partial humeri; 1 nearly complete and 1 distal end of tibiotarsi; 1 proximal end of tarsometatar- Mimocichla plumbed .—Wetmore ,193 7b :441 [Pleistocene, sus (USNM 283310, 283342, 283343, 283347- “Great Exuma’' = Little Exuma]. 283348, 283391, 283402); minimum number of Material Examined.— Little Exuma: 1 com¬ individuals, 2. plete humerus (MCZ 2270). New Providence: Remarks.— The Bananaquit is an abundant 2 distal ends of tarsometatarsi (USNM 283311, resident throughout the Bahamas, where it occurs 283312); minimum number of individuals, 2. in various habitats. The species is also found Remarks.— The Red-legged Thrush (71 plum¬ throughout the West Indies (except Cuba) and beus plumbeus) is a relatively common bird inhabit¬ on the mainland from southern Mexico to north¬ ing wooded country and undergrowth in the ern South America. The subspecies inhabiting northern islands of the Bahamas. Other subspe¬ the Bahamas (C. f bahamensis) is more similar to cies occur in the Greater Antilles (except Ja¬ C. f. sharpei of the Cayman Islands, C. f caboti of maica), Cayman Brae, Dominica, and formerly Cozumel Island, C. f oblita of San Andres, and C. on Swan Island, where the species is now extinct. f tricolor of Providencia, than to the many sub¬ species geographically closer in the West Indies. Family Parulidae These could perhaps be relicts of a separate spe¬ cies, Coereba bahamensis , that is distinct from C. Dendroica spp. flaveola. The absence of any form of Coereba on Material Examined.— New Providence: 1 mainland Cuba is one of the most inexplicable partial mandibular ramus, 2 scapular ends of facts in the entire discipline of zoogeography. NUMBER 48 51 Family Thraupidae or Emberizidae Spindalis zena (Linnaeus) or Loxigilla violacea (Linnaeus) Material Examined.— New Providence: 1 complete, 1 proximal and 1 distal ends of humeri; 1 distal end of ulna (USNM 283339-283341, 283392); minimum number of individuals, 2. Remarks.— The humeri of the tanager Spindalis zena zena and the finch Loxigilla violacea violacea are identical in size and we were unable to detect any qualitative features that would separate them. Both of these birds are found on New Providence today and neither would be unexpected in the Banana Hole deposits. Family Icteridae Sturnella sp. Margarops fuscatus. —Wetmore, 1937b:441 [Pleistocene, “Great Exuma” = Little Exuma], Material Examined. —Little Exuma: 2 com¬ plete humeri (MCZ 2269); minimum number of Figure 12. — Premaxillae of Icteridae in lateral view: a, fossil individuals, 2. New Providence: 1 premaxilla, 1 Sturnella sp. from New Providence (USNM 283364); b, Stur¬ proximal end of scapula, 1 proximal end of hu¬ nell amagna c\I,cteru dsomimcens ipsortoncens di ;sA,gelaiu pshoen- merus (USNM 283336, 283353, 283364); mini¬ icius bryanti .(Scale = 1 cm.) mum number of individuals, 1. Remarks.— A distinctive premaxilla from Ba¬ nana Hole (Figure 12) definitely establishes mea¬ dowlarks as former inhabitants of the Bahamas. The straight, very flat culmen, combined with size, distinguishes this specimen from any other icterids likely to occur in the Bahamas. The two humeri from Little Exuma, identified by Wet- more (1937b) as Margarops fuscatus , definitely do not belong with the Mimidae, as the medial bar separating the tricipital fossae is reduced and does not extend to the shaft, as in that family. Fur¬ thermore, the fossils are too small for Margarops fuscatus. The characters of these specimens place them with the New World nine-primaried oscines, within which they are too large for anything other than icterids, or for any Icteridae now found in 52 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY would doubtless have occupied a similar niche in 13 species were also found in the New Providence the Bahamas. Sturnella magna is partially migra¬ deposits. This faunal similarity might be attrib¬ tory, but does not migrate to any localities where utable to approximate contemporaneity of the there are not resident populations also. Regard¬ two deposits and also to the fact that both Little less of whether the Bahaman birds were migrants Exuma and New Providence are on the Great or residents, their habitat requirements would Bahama Bank and would have been part of a have been the same. single continuous land mass when sea levels were lower during the last glaciation. Of the 13 species Discussion found as fossils on Little Exuma, only one, Athene cuniculana , is known from the Exumas today. If, Our re-examination of the supposedly extinct as we have suggested, the burrowing owl in the fossil species of birds from the Bahamas consid¬ Pleistocene of the Bahamas was different from erably alters previous ideas about the nature of the extant form, then all of the Exuma fossils the Pleistocene avifauna. We have proposed the would represent taxa that are now extinct on the following synonymies or changes in nomencla¬ island. Of the twelve certainly extinct species, ture: Calohierax quadratus = Buteo sp., Burhinus nanus only four presently exist elsewhere in the Baha¬ = Burhinus bistnatus nanus , Glaucidium dickinsom = mas: Geotrygon chrysia , Melanerpes supercilians (ab- Athene cuniculana, Otus providentiae = Athene cunicu- laria , Bathoceleus hyphalus = Melanerpes supercilians, and Corvus wetmorei = Corvus nasicus. Thus, of the nine new species described by Wetmore (1937b) and Brodkorb (1959), only three represent taxa that are certainly extinct at the species level: the giant hawk Titanohierax gloveralleni , the giant barn owl Tyto pollens , and the poorly known caracara Polyborus creightoni. Other possible extinctions of Extinct taxa at either the specific or subspecific levels involve an enigmatic large snipe Capella sp., a small subspecies of thick-knee Burhinus bistnatus nanus , and the Pleistocene burrowing owl of the Bahamas, which, if a valid taxon, would be known as Athene cunciculana providentiae. Of equal or greater importance, however, are the numerous extant species that are represented as fossils but which either no longer occur in the Bahamas or are now absent from the islands where they were found as fossils. The archeological remains from Crooked Is¬ land are of interest in connection with the distri¬ bution of two seabirds, Pterodroma cahow and P. hasitata , and also for the presence of Eudocimus albus , Amazona leucocephala , and Corvus nasicus , which no longer are found on the island. The fossils from Little Exuma are so few in number that generalizations about them cannot be made with confidence. It is significant that all NUMBER 48 53 sent from the Great Bahama Bank), Turdus plum- Table 12.—Analysis of the sample of fossils from Banana beus, and Corvus nasicus (restricted to Caicos). Eight Hole, New Providence, that was studied by Brodkorb (1959), species (62%) in the Exuma fossil fauna no longer and our re-evaluation of the same sample (Corvus nasicus is here regarded as extinct in the Bahamas) occur in the Bahamas (Table 11). Even given the probable bias towards larger species induced by the collecting methods, the extinction rate for this Brodkor (b1959) fauna seems high. The taxa occurring in the Banana Hole depos¬ its are listed according to their current status on New Providence in Table 11. The results of this tabulation are rather striking. Of the 32 species in the fossil fauna, 16 (50%), and possibly 17 or 18 counting Athene cuncicularia and Falco sparvenus , are extinct on New Providence and 13 (40%) no longer occur in the Bahamas (excepting Corvus nasicus in Caicos). Only 3 species (9%) are totally extinct and all of these are large raptors. Birds of prey account for 8 of the species in the fauna (25%) and of these, 5 (possibly 6, counting Athene cuniculana ssp.) are extinct in the Bahamas. In terms of actual numbers of specimens (354), the Banana Hole sample is rather small; it is certainly not comprehensive enough to enable speculation on which species not represented were actually absent from the island at the time of deposition. Nevertheless, sampling bias cannot the substitution of a single extant species for an account for peculiarities in the ratio of extinct to extinct one would result in percentages of extinct extant taxa, as the species represented range in taxa nearly identical to those obtained with the size from giant hawks and owls to hummingbirds larger sample containing 32 species. Thus, even and warblers. Although additional specimens of if we were to obtain additional specimens from smaller species might raise the proportion of ex¬ Banana Hole, the proportion of extinct taxa tant species, they could as easily augment the would be unlikely to change substantially, and number of extinct taxa. we must still account for the extinction of roughly Our reassessment of the species composition of half of the avifauna of New Providence since the Brodkorb’s much smaller original sample from late Pleistocene. Banana Hole (65 specimens) is contrasted with One possible cause, of course, would be the his original identifications in Table 12. Coinci¬ great reduction and fragmentation in the land dentally, there are 15 species in each instance. mass of the Bahamas with the postglacial rise in Although by our analysis there is a lower per¬ sea level following the Wisconsinan glacial stage. centage of extinct species than by Brodkorb’s, the Brodkorb (1959:367) cited this as “the major percentages (60% extinct on New Providence; factor in restricting the size of the avifauna.” For 47% extinct in the Bahamas) are actually quite whatever reasons, isolation and reduction in land similar (50% and 40%, respectively) to those ob¬ area is usually correlated with a decrease in spe¬ tained when the sample was increased to 354 cies diversity on islands, whether attributable to specimens and the number of species was more “area effects” alone or in combination with re¬ than doubled. In the smaller sample of 15 species, duction in habitat diversity or other factors. Ex- 54 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY amples of this are the reduced numbers of species arid, scrubby areas, often in association with on the Isle of Pines versus western Cuba (Todd, pines, but also with palms. It is not clear to what 1916) and similarly on Tasmania versus south¬ extent the extinction of the rodent Geocapromys eastern Australia (Ridpath and Moreau, 1966), from all of the Bahamas except East Plana Cay both of these islands having been joined to the is due to habitat change, but the only remaining mainland 12,000 years ago or less. This “island population is now well adapted to a particularly effect“ no doubt played a part in the extinction arid, scrubby habitat. The fates of the giant of some species on New Providence, but probably predators Titanohierax gloverallem and Tyto pollens cannot account for the disappearance of 40% of were almost surely linked to the well-being of the Banana Hole fauna from all or most of the Geocapromys ; if habitat changes were largely re¬ Bahaman archipelago. sponsible for the disappearance of the rodent It is quite obvious from the fossil record that from much of its former range, then they were habitats must have changed rather drastically in equally responsible for the extinction of its pre¬ the Bahamas since the Pleistocene. Some of the dators. species represented are certain indicators of open If the late Pleistocene environment of the Ba¬ grassland or prairie environments that no longer hamas was more arid and increased precipitation exist in the Bahamas. Included here are the car- was actually a major cause of extinction, then the acara Polyborus creightoni , the thick-knee Burhinus marked rainfall gradient observed from north to bistnatus nanus , and the meadowlark Sturnella sp. south in the Bahamas today (p. 1) should have In southern Florida and elsewhere, burrowing a marked effect on the fauna. We can then pos¬ owls (. Athene cumculana) are found in association tulate that there would have been a greater num¬ with caracaras and are characteristic of prairies. ber of extinctions on the wetter islands of the If the fossil burrowing owl is in fact taxonomically Little Bahama Bank, whereas more species would distinct from the one inhabiting the Bahamas have persisted on the drier southern islands. Un¬ now, its extinction could be attributed to the fortunately, we do not as yet have any fossil same degradation of prairie-type habitat that faunas from the Little Bahama Bank. Neverthe¬ caused the disappearance of caracaras, thick- less, the pattern of distribution of the Recent knees, and meadowlarks. The large snipe ( Capella reptilian fauna of the Bahamas fits the above sp.) found in the fossil deposits likewise fits well hypothesis remarkably well (p. 19), as, to a lesser with the picture of more open habitats in the extent, do the bats. The only Recent bird that Pleistocene, although it is a less certain indicator, clearly fits this pattern is Corvus nasicus , which is inasmuch as the modern species Capella gallinago restricted in the Bahamas to the Caicos, whereas occurs in the Bahamas under present environ¬ it is known as a fossil from islands of the Great mental conditions. On the other hand, some of Bahama Bank and from subfossils on the the species in the fossil fauna that are absent in Crooked-Acklins Bank. Mimus gundlachii also fits the Bahamas today indicate the former presence this pattern, as it is most abundant in the south¬ of broadleaf forest of some sort. These include ern islands, but is apparently absent from Abaco Corvus nasicus and Columba squamosa. The presence and has only recently been reported from Grand of a flicker ( Colaptes ) would be in accordance Bahama, where it is decidedly uncommon (Em- either with open habitats, broadleaf forest, or len, 1977:122). both. The relictual distribution of Melanerpes su- The avifauna of the Little Bahama Bank is not percilians , and its extinction on some islands, ap¬ proportionately as depauperate as the herpeto- pears to be linked to the disappearance or reduc¬ fauna. Birds, being more vagile, compensate more tion of broadleaf woodland. rapidly for extinctions by subsequent coloniza¬ At least one of the species now absent from the tion. Examples are Dendrocopos villosus piger, Sitta Bahamas, Corvus palmarum , seems to thrive best in pusilla insularis, and Dendroica dominica flavescens , NUMBER 48 55 which are endemic to the pine forests of the Little Our findings have a direct bearing on several Bahama Bank (the nuthatch is restricted to much-discussed modern ecological and biogeo- Grand Bahama alone). All of these are derived graphical hypotheses, such as the equilibrium from the North American mainland but are theory of island biogeography, which predicts markedly distinct from their parental stock, par¬ that immigration balances extinction (Mac- ticularly the nuthatch and the warbler. This dif¬ Arthur and Wilson, 1967). An outgrowth of this ferentiation has presumably taken place since the is the concept of “taxon cycles,” which holds that end of the Pleistocene, when pine forest presum¬ a given species progresses through a series of ably invaded the Little Bahama Bank. The wood¬ stages in which it is first widespread but undiffer¬ pecker has spread to Andros and New Providence, entiated, later develops subspecies endemic to where it has differentiated further ( Dendrocopos certain islands, followed by the extinction of some villosus maynardi). Dendroica pinus achrustera , a pine- populations to leave discontinuous relict popula¬ inhabiting warbler derived from the mainland, is tions. Ultimately, the species is reduced to a form also found on the islands of the Little Bahama endemic to a single island and then the taxon Bank, as well as on New Providence and Andros. eventually becomes extinct (Ricklefs and Cox, Another pine-inhabiting warbler, Dendroica pityo- 1972, 1978). This “taxon cycle” was supposed to phila , occurs only on the Little Bahama Bank but be powered by “counteradaptation” that renders is evidently derived from Cuba, the only other island birds more susceptible to interspecific com¬ place where the species occurs. The two popula¬ petition from new incoming colonizers that ulti¬ tions of D. pityophila have not differentiated sub- mately are responsible for the doom of established specifically, which, with the vegetational history species (Ricklefs and Cox, 1972:217). In contrast outlined above, suggests that the Bahaman pop¬ with this counterintuitive theory is that of Lack ulation is not relictual and that the species has (1976), which states that the habitats on an island colonized relatively recently. This in turn implies determine the number of species, that populations that the species passed over the islands of the of birds on islands are stable, and that once a Great Bahama Bank without colonizing them, population is established, its extinction and re¬ presumably because of a lack of suitable habitat, placement by a new colonist is unlikely. at least at the time of dispersal. Although we do not subscribe to all of Lack’s It is our contention that drastic habitat changes views on island biogeography, the fossil record of caused the extinction of a significant proportion birds from the Bahamas would seem to come of the avifauna of the Bahamas in the late Pleis¬ much closer to supporting the outline of his hy¬ tocene. These species were not replaced by similar pothesis, as presented above, than that of Ricklefs taxa simply because there was no suitable habitat and Cox. Although half of the species occurring for them. There is no place left in the Bahamas as fossils on New Providence are extinct there in which thick-knees or meadowlarks, for exam¬ now, there is not the slightest evidence that these ple, could maintain viable populations. were displaced, or even replaced, by new colo¬ Late Pleistocene extinctions, as a result of the nists. Certainly there is nothing on New Provi¬ loss of arid and savanna-type habitats, are not dence today that occupies the niches of thick- restricted to the Bahamas. Similar extinctions of knees, crows, meadowlarks, flickers, the large reptiles, as well as birds, have also been docu¬ hummingbird, or any of the five extinct raptors. mented for Puerto Rico (Pregill, 1981; Pregill Columba squamosa was certainly not displaced by and Olson, 1981), and it is probable that the C. leucocephala , as both occurred together com¬ relictual pattern of distribution of many species monly as fossils and they co-exist everywhere else of West Indian vertebrates can be explained by in the present range of C. squamosa. It is perhaps changes in habitat caused by a postglacial pluvial conceivable that Philodice evelynae has displaced period (Pregill and Olson, 1981). Chlorostilbon ricordii on New Providence, but this 56 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY seems highly unlikely because both species occur As far as faunal equilibrium is concerned, be¬ together on Andros, Abaco, and Grand Bahama, cause the fossil sample from New Providence is where Chlorostilbon predominates and where Phil- not representative of the entire avifauna at the odice is much less abundant than on islands where time of deposition, it is impossible by comparison it is the only resident hummingbird. It might also with the Recent fauna to say whether extinctions be argued that the Hairy Woodpecker ( Dendroco - have balanced immigrations. On New Providence pos villosus) is displacing Melanerpes superciliaris. we have observed 50% rate of extinction since the The former, however, occurs almost entirely in late Pleistocene, so it would seem reasonable that pine forest, whereas the latter inhabits broadleaf immigrations have not kept pace with extinctions. coppice. Furthermore, D. villosus is present on two Although it appears that the northern islands of of the three islands on which M. superciliaris has the Bahamas are being newly colonized by certain been able to survive in the Bahamas. species of birds, the same is not true for reptiles, The hypotheses of Lack (1976) and Ricklefs the difference probably being due to differential and Cox (1972, 1978) are both predicated on the capabilities for dispersal. Different portions of a unstated assumption that habitats are stable, fauna might reach “equilibrium” at vastly differ¬ which, as we have seen, was certainly not true for ent rates and thus the fauna as a whole might the late Pleistocene of the Bahamas. Changes in never be at “equilibrium.” Furthermore, since climate and habitat much more readily explain the relictual and disjunct distributions of most different habitats in mainland areas support dif¬ West Indian birds than the presupposition that a ferent numbers of species, there is no a prion species is somehow “programmed” to effect its reason why a certain number of species in a given own extinction. There is no reason to assume that habitat on an island would be replaced by the a species will not persist for an indefinite period same number of species if the habitat were to of time provided that its environment does not change and cause the extinction of a large com¬ change, whereas no evidence has been presented ponent of the original fauna. that shows extinction due to competition with We believe that when the faunal history of newly arrived colonists to be a common event. more islands is known, that Pleistocene fluctua¬ Clearly it is habitats and environments that have tions in climates and their subsequent effect on been cycling in the West Indies and it is highly habitats will be seen to have had a greater influ¬ unlikely that taxa themselves are subject to any ence on the composition of island faunas than endogenous or exogenous “cycle” that cannot be any of the more easily quantifiable variables upon linked directly to fluctuations in their environ¬ which MacArthur and Wilson’s equations were ment (Pregill and Olson, 1981). based. Olson, Storrs L. and Hilgartner, W B. 1982. "Fossil and Subfossil Birds from the Bahamas." Fossil vertebrates from the Bahamas 48, 22–56. View This Item Online: https://www.biodiversitylibrary.org/item/266616 Permalink: https://www.biodiversitylibrary.org/partpdf/352110 Holding Institution Smithsonian Libraries Sponsored by Smithsonian Institution Copyright & Reuse Copyright Status: In copyright. Digitized with the permission of the rights holder. Rights Holder: Smithsonian Institution License: http://creativecommons.org/licenses/by-nc-sa/4.0/ Rights: http://biodiversitylibrary.org/permissions This document was created from content at the Biodiversity Heritage Library, the world's largest open access digital library for biodiversity literature and archives. Visit BHL at https://www.biodiversitylibrary.org. This file was generated 13 April 2023 at 20:02 UTC