The putative cormorant Phalacrocorax subvolans Brodkorb 1956, from the early Miocene of Florida, is moved from the Phalacrocoracidae to the Anhingidae and should be known as Anhinga subvolans (Brodkorb 1956). This species is the earliest known anhinga and demonstrates that the family Anhingidae has been present in North America for at least 18 million years. It has been at least 30 million years since the Anhingidae and the Phalacrocoracidae shared a common ancestor. Received 18 December 1985, accepted 17 March 1986.

Department of Zoology, University of Florida, Gainesville, Florida 32611 USA IN a study of fossil birds from the Hawthorn Formation, Brodkorb (1956) described a new species of cormorant, Phalacrocorax subvolans, from the Thomas Farm local fauna, Gilchrist Co., Florida. This species, known only from the holotypical proximal end of a humerus (Brod- korb 1956), "agrees with Phalacrocorax wetmorei Brodkorb (1955) in conformation of caput hu- meri and bicipital crest, but differs as follows: proximal width less; width of shaft less; liMa- mental furrow [sulcus ligamentosus transver- sus] shorter and less deep; deltoid crest [crista pectoralis] longer; internal tuberosity [tuber- culum ventrale] sharper and capital groove [in- cisura capitis] correspondingly deeper; bicipi- tal furrow [impressio m. coracobrachialis cranialis] wider." Brodkorb noted that the greater width of the bicipital furrow in P. sub- volans leaves a larger surface for the attachment of M. coracobrachialis anterior (= M. coraco- brachialis cranialis) and suggested that this in- dicates that P. subvolans was a better soarer than living or fossil cormorants. Brodkorb also not- ed that this condition of the humerus in P. sub- volans approached that of anhingas, birds that customarily soar for long intervals. Reexamination of the holotype indicates that Phalacrocorax subvolans should be moved to the genus Anhinga in the family Anhingidae. , MATERIALS AND METHODS Fossil specimens included in this study are housed in the Vertebrate Paleontology collections of the Florida State Museum (UF). Comparative material of Present address: Division of Birds, National Mu- seum of Natural History, Smithsonian Institution, Washington, D.C. 20560 USA. living species is in the collections of P. Brodkorb; Florida State Museum; National Museum of Natural History, Smithsonian Institution; American Museum of Natural History; University of Michigan; and Roy- al Ontario Museum. Anatomical terminology follows Baumel et al. (1979). Measurements are described in Table 1. SYSTEMATICS Family Anhingidae Ridgway 1887 The proximal ends of humeri of the Anhin- gidae may be distinguished from those of the Phalacrocoracidae using two characters (Miller 1966). In cormorants the crus dorsale fossae overhangs the fossa pneumotricipitalis (see Fig. 1) and fully covers its proximal end, whereas in anhingas the less extensive fossa is well ex- posed. The sulcus ligamentosus transversus on the cranial surface is longer, deeper, and ex- tends transversely to, but is narrowly separated from, the impressio M. coracobrachialis crani- alis in cormorants; the sulcus is shorter and deep only ventrally in anhingas. In addition, anhin- gas have a strong sulcus on the cranial face of the humerus paralleling the distal portion of the crista pectoralis. In cormorants this sulcus is absent, causing the crista pectoralis to merge more smoothly with the shaft. Also, anhingas tend to have a proportionally longer crista pec- toralis than do cormorants. Genus Anhinga Brisson 1760 Anhinga subvolans (Brodkorb 1956) Holotype.--UF 4500, proximal half of right humerus. Florida State Museum, Vertebrate Pa- leontology collection. From the Thomas Farm TABLE 1. Measurements of humeri of living and fossil Anhinga species. Data are means + SD and observed ranges. Number of specimens for A. anhinga and A. rufa = 10, all other n = 1. Measurements of the humerus are as follows: W-SHAFT = transverse width of midshaft; D-SHAFT = depth of midshaft; W-PROX = transverse width of proximal end from the external tuberosity (tuberculum dotsale) to the most ventral face of the bicipital crest (crista bicipitalis); D-PROX = depth of proximal end, from the bicipital surface (facies bicipitalis) to the internal tuberosity (tuberculum venttale), measured at right angles to the long axis of the shaft; D-HEAD = depth of head, measured parallel to the axis of the head; L-DELTOID = length of deltoid crest (crista pectoralis), measured from the external tuberosity to the most distal extension of the deltoid crest. novaehol- melano- Measurement anhinga rufa landiae gaster grandis subvolans W-SHAFT 6.66 + 0.39 6.81 + 0.55 7.0 6.4 8.7 7.6 5.7-7.1 6.2-7.9 .... D-SHAFT 5.76 + 0.38 6.06 + 0.54 5.8 6.2 7.7 6.7 5.1-6.2 5.4-7.1 .... W-PROX 18.02 + 0.80 19.85 + 1.11 20.0 18.5 23.1 21.4 17.2-19.8 18.1-21.6 .... D-PROX 8.62 + 0.35 9.85 + 0.54 9.8 8.9 -- 9.7 8.0-9.1 8.8-10.5 .... D-HEAD 6.71 + 0.25 7.22 + 0.49 7.7 6.8 8.0 7.4 6.1-7.1 6.5-7.9 .... L-DELTOID 35.19 + 1.88 37.86 + 2.61 40.7 38.5 42.3 37.5 31.7-37.8 35.0-43.3 .... locality, early Miocene (early Hemingfordian; approximately 18 million years before present), Gilchrist Co., Florida. Collected by R. Bader in the spring of 1955. Webb (1981) reviewed this local fauna. The Thomas Farm locality repre- sents a high-sided sinkhole that was at least partially water filled (A. E. Pratt pers. comm.). The fossil birds of Thomas Farm local fauna were studied by Wetmore (1943, 1958), Brod- korb (1954, 1956, 1963a), Cracraft (1971), Olson and Farrand (1974), and Steadman (1980). Emended diagnosis.--Referable to the family Anhingidae by the characters listed above. The type of Anhinga subvolans differs from the prox- imal ends of the humeri of all species of An- hinga examined (A. grandis; UF 25739, Love Bone Bed locality, Alachua Co., Florida; A. rufa; A. melanogaster; A. anhinga) in having a deeper fossa pneumotricipitalis and impressio M. cor- acobrachialis cranialis, a more prominent crus dorsale fossae, and a better-developed and sharper ridge that extends distally down the shaft from the crus dorsale fossae. The proxi- mal end of the humerus of A. subvolans is sim- ilar in size to that of A. rufa (A. anhinga smaller, A. grandis larger; Table 1). DISCUSSION The two implied generic characters (config- uration of caput humeri and bicipital crest) originally used by Brodkorb (1956) are found in both the Anhingidae and the Phalacrocora- cidae. Two of the original specific characters of A. subvolans (sulcus lig. transversus and length of the crista pectoralis) are actually diagnostic of the family Anhingidae. All other characters in the original description are either size de- pendent or serve only to distinguish Anhinga subvolans from Phalacrocorax wetmorei. Olson (1985) reviewed the fossil history of this family. I can add that Anhinga grandis Mar- tin and Mengel (1975) is now known from three additional localities in the late Miocene of Flor- ida (Becker 1985). Ballman (MS) reported a species of Anhinga from the Pliocene Sahabi Formation of Libya. It is about the size of An- hinga grandis and therefore can be distin- guished from Anhinga subvolans on the basis of size. There is also a large, indeterminate species of anhinga from the earliest Pliocene (early Hemphillian) Bone Valley Mining District (Becker 1985) and from the early Pleisto- cene (Irvingtonian) Coleman III locality (Ritch- ie 1980). Based on only a few ulnae, this species does not appear to be referable either to the living Anhinga anhinga or to Anhinga grandis (contra Ritchie 1980). Fossils representing the living species, Anhinga anhinga, are known from numerous localities in the late Pleistocene (Rancholabrean) of Florida (Brodkorb 1963b). A B C D Fig. 1. Caudal view of the proximal end of humeri of anhingas. (A) Anhinga anhinga, USNM 500870. (B) A. subvolans, UF 4500, holotype. (C) A. grandis, UF 25739. (D) Phalacrocorox auritus, USNM 500819. All photos are 1 x. Protoplotus beauforti Lambrecht from the mid- dle Eocene of Sumatra is probably not referable to the Anhingidae (Rich in litt., cited in Olson 1985). Protoplotus beauforti is smaller and has dif- ferent limb proportions than any anhinga. Re- gardless of its familial affinities, little detail can be seen on the humerus of this species (Lam- brecht 1931). The humerus is not known in the fossil species Anhinga pannonica Lambrecht from the late Miocene of Hungary, A. hadarensis Brodkorb and Mourer-Chauvir from the Plio- Pleistocene of Ethiopia and Tanzania, and A. laticeps Devis from the Pleistocene of Australia. Considering their geographic and geologic provenances, none of these species is likely to be conspecific with Anhinga subvolans. Olson (1985), citing cranial and tarsometatar- sal characters, showed that the New World An- hinga anhinga is distinct from the Old World species of Anhinga and that all members of this genus should not be viewed as a single super- species. A number of additional characters sup- port Olson's view. The Old World species share a similar structure of the proventriculus (glan- dular tissue in two separate patches; proven- tricular glands collected in a diverticulum in A. anhinga; Garrod 1876, 1878; Forbes 1882), py- loric lobe (conical and retractile pyloric plug present; absent in A. anhinga; Gatrod 1876, 1878; Forbes 1882), structure of temporal fossa (fos- sae boundaries distinct; indistinct in A. anhin- ga; Beddard 1892), development of the postor- bital process (smaller; well developed in A. anhinga; Beddard 1892), and structure of the bridge of Dnitz (ossified; not ossified in A. anhinga; Garrod 1876; pers. obs., n = 10+). Ad- ditionally, the sexes are dimorphic in Anhinga novaehollandiae and in A. anhinga but are similar in A. rufa and A. rnelanogaster (Vaurie 1965). Fossil evidence shows the Anhingidae to be present in the early Miocene and the Phalacro- coracidae to exist in the Eo-Oligocene (Phos- phorites du Quercy; Mourer-Chauvir 1982). Therefore, these two groups have not shared a common ancestor for at least 30 million years (see Savage and Russell 1983 for information and references on the age of the Phosphorites du Quercy), and probably much longer. Some authors (Dorst and Mougin 1979, Cra- craft 1985) have reduced the Anhingidae to a subfamily of the Phalacrocoracidae without comment. Anhingas have a feeding behavior and a straight, laterally compressed rostrum with serrated tomia that is unique in the Pel- ecaniformes. Other studies have shown that cormorants and anhingas differ significantly in their habitat preference, locomotion, ecology, and arrangement of the carotid arteries (Garrod 1876, 1978; Beddard 1892; Owre 1967). Given that anhingas and cormorants have had a long fossil history, that the magnitude of morpho- logical difference between cormorants and an- hingas is comparable to that found among oth- er pelecaniform families, and that the ranking of anhingas at the family level is consistent with the taxonomy of the order as a whole, it seems more reasonable to maintain the Anhingidae and the Phalacrocoracidae at their traditional family ranks, as recently suggested by Brod- korb and Mourer-Chauvir (1982) and Olson (1985). ACKNOWLEDGMENTS For the loan of fossil or Recent specimens, or access to collections, I thank J. Barlow, Royal Ontario Mu- seum; P. Brodkorb, Department of Zoology, Univer- sity of Florida; M. Voorhies, University of Nebraska State Museum; G. Barrowdough and F. Vuilleumier, American Museum of Natural History; and J. W. Hardy, B. J. MacFadden, G. S. Morgan, S. D. Webb, and T. Webber, Florida State Museum. I thank Pierce Brodkorb for his encouragement and comments and S. L. Olson and R. L. Zusi for their comments on this paper. Photographs are by Victor E. Krantz. LITERATURE CITED BAUMEL, J. J., A. S. KING, A.g. LUCAS, J. E. BRFZILE, & H. E. EVANS. 1979. Nomina Anatomica Avi- urn. London, Academic Press. BECKER, J.J. 1985. The fossil birds of the late Mio- cene and early Pliocene of Florida. Unpublished Ph.D. dissertation, Gainesville, Univ. Florida. BEDDARD, F.E. 1892. Notes on the anatomy and os- teology of the Indian Darter (Plotus melanogas- ter). Proc. Zool. Soc. London 1892: 291-296. BIODKOIn, P. 1954. A chachalaca from the Miocene of Florida. Wilson Bull. 66: 180-183. 1955. The avifauna of the Bone Valley For- mation. Florida Geol. Surv., Rept. Investigations 14: 1-57. 1956. Two new birds from the Miocene of Florida. Condor 58: 367-370. 1963a. Miocene birds from the Hawthorne Formation. Quart. J. Florida Aca. d. Sci. 26: 159- 167. --. 1963b. Catalogue of fossil birds. Part 1 (Ar- chaeopterygiformes through Ardeiformes). Bull. Florida State Mus., Biol. Sci. 7: 180-293. --, & C. MOURER-CHAUVIRfi. 1982. Fossil anhin- gas (Aves: Anhingidae) from Early Man sites of Hadar and Oreo, Ethiopia, and Olduvai Gorge, Tanzania. Geobios 15: 505-515. CRACRAFT, J. 1971. The humerus of the early Mio- cene cracid, Boreortalis laesslei Brodkorb. Wilson Bull. 83: 200-201. 1985. Monophyly and phylogenetic rela- tionships of the Pelecaniformes: a numerical cla- distic analysis. Auk 102: 834-853. DORST, J., & j. L. MOUGIN. 1979. Order Pelecani- formes. Pp. 155-193 in Check-list of birds of the world, vol. 1, 2nd ed. (E. Mayr and G. W. Cot- trell, Eds.). Cambridge, Massachusetts, Mus. Comparative Zool. FOInES, W.A. 1882. On some points in the anatomy of the Indian Darter (Plotus melanogaster), and on the mechanism of the neck in the darters (Plo- tus), in connexion with their habits. Proc. Zool. Soc. London 1882: 208-212. GARROD, A.H. 1876. Notes on the anatomy of Plotus anhinga. Proc. Zool. Soc. London 1876: 335-345. 1878. Note on points in the anatomy of Le- vaillant's Darter (Plotus levaillanti). Proc. Zool. Soc. London 1878: 679-681. LAMBRECHT, K. 1931. Protoplotus beauforti n.g. n. sp., ein Schlangenhalsvogel aus dem Terti'fir von W. Sumatra. Wetenschappelijke Mededeelingen Dienst van den Mijnbouw. in Nederlandsch-In- di 17: 15-24, pls. II and III. MARTIN, L., & R. g. MENGEL. 1975. A new species of anhinga (Anhingidae) from the Upper Plio- cene of Nebraska. Auk 92: 137-140. MILLEl% A. H. 1966. An evaluation of the fossil an- hingas of Australia. Condor 68: 315-320. MOURER-CHAUVIRfi, C. 1982. Les oiseaux fossiles des Phosphorires du Quercy (fiocne sup6rier a oli- gocne sup6rieur): implications pa16obiograph- iques. Geobios m6m. sp6c. 6: 413-426. OLSON, S. L. 1985. The fossil record of birds. Pp. 79-252 in Avian biology, vol. 8 (D. S. Farner, J. S. King, and K. C. Parkes, Eds.). New York, Ac- ademic Press. --, t J. FARRAND, JR. 1974. Rhegminornis restud- ied: a tiny Miocene turkey. Wilson Bull. 86: 114- 120. OWIE, O. T. 1967. Adaptations for locomotion and feeding in the Anhinga and the Double-crested Cormorant. Ornithol. Monogr. No. 6. RITCHIE, T. L. 1980. Two mid-Pleistocene avifaunas from Coleman, Florida. Bull. Florida State Mus., Biol. Sci. 26: 1-36. SAVAGE, D. E., & D. E. RUSSELL. 1983. Mammalian paieofaunas of the world. Reading, Massachu- setts, Addison-Wesley Publ. Co. STEADMAN, D.W. 1980. A review of the osteology and paleontology of turkeys (Aves: Meleagridi- nae). Contrib. Sci., Nat. Hist. Mus. Los Angeles Co. 330: 131-207. VAIJRIE, C. 1965. The birds of the Palearctic fauna. Non-Passeriformes. London, H. F. and G. With- erby Ltd. WEBB, S.D. 1981. The Thomas Farm fossil site. Plas- ter Jacket 37: 6-25. WETMORE, A. 1943. Fossil birds from the Tertiary deposits of Florida. New England Zool. Club 32: 59-68. � 1958. Miscellaneous notes of fossil birds. Smithsonian Misc. Coil. 135: 1-11.