3 Special Collections mo HARVARD UNIVERSITY VErlRI LIBRARY OF THE Museum of Comparative Zoology UuLLetln of the Museum of Comparative Zoology iVrt-cb' J Volume 134 1965-1966 HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 U.S.A. ^ CONTENTS Page No. 1. Relationships and Evolution within the Cracidae (Aves, Galli- formes). By Fran9ois Vuilleumier. March, 1965 1 No. 2. Tlie Distribution of the Oceanic Fish Brama brama. By Giles W. Mead and Richard L. Haedrich. June, 1965 29 No. 3. Evolution of the Tapiroid Skeleton from Heptodon to Tapirus. By Leonard B. Radinsky. June, 1965 69 No. 4. The Species Bufo granulosus Spix (Salientia: Bufonidae) and its Geographic Variation. By Jose M. Gallardo. September, 1965 .. 107 No. 5. The Mesopelagic Fishes Collected During Cruise 17 of the R/V Chain, with a Method for Analyzing Faunal Transects. By Richard H. Backus, Giles W. Mead, Richard L. Haedrich, and Alfred W. Ebeling. September, 1965 139 No. 6. New Species of Hemicyclops (Copepoda, Cyclopoida) from Madagascar. By Arthur G. Humes. November, 1965 159 No. 7. New Oceanic Cheilodipterid Fishes from the Indian Ocean. By Giles W. Mead and J. E. De Falla. November, 1965 261 No. 8. Italian Wolf Spiders of the Genus Pardosa ( Araneae-Lycosidae). By Paolo Tongiorgi. February, 1966 275 No. 9. Wolf Spiders of the Pardosa monticola Group (Araneae, Lycosi- dae). By Paolo Tongiorgi. February, 1966 335 No. 10. The Lower Triassic Formations of the Salt Range and Trans-Indus Ranges, West Pakistan. By Bernhard Kummel. August, 1966 361 No. 11. The Taxonomy, Cytology, and Evolution of the Genus Rhagoletis in North America (Diptera, Tephritidae ) . By Guy L. Bush. September, 1966 431 SuLUtln OF THE lUseum o1 Comparative Zoology • I3RARY iviat^ i ^^ 1965 HARVARD UNIVERSITY, Relationships and Evolution within the Cracidae (Aves, Galliformes) by FRANCOIS VUILLEUMIER HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS, U.S.A. VOLUME 134, NUMBER 1 10 AAARCH 1965 PUBLICATIONS ISSUED BY OR IN CONNECTION WITH THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- MEMoms 1864-1938 JoHNSONLv, Department of MoUusks, 1941- OocAsiONAL Papers on Mollusks, 1945- Other Publications. Bigelow, H. B. andW. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50, Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Lyman, C. P. and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 15. Whittington, H. B. and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948 ( complete sets only). Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138 U. S. A. IUJI\/-M^ I MAR 1 a 1965 HARVARD UNIVERSITY RELATIONSHIPS AND EVOLUTION WITHIN THE CRACIDAE (AVES, GALLIFORMES)' FRANCOIS VUILLEUMIER INTRODUCTION The latest systematic treatments of the Cracidae by Peters (1934), and by Hell- nia>r and Conover ( 1942 ) , showed no real improvement over the older one by Sclater and Salvin (1870). All recent authors who have dealt with the Cracidae have done so in the course of regional faunal studies, reviewing only part of a genus or a species without consideration for variation in the total range of the taxon. This is also true for other groups of neotropical birds. The only original studies on Cracidae that brought new and much needed data on ecology and behavior were by \\^agner ( 1953 ) on Mexi- can, and by Schiifer (1953) on Venezuelan species. During an investigation of speciation pat- terns in the Cracidae, I first undertook a character analysis to re-evaluate generic limits within the family. This led to an arrangement slightly at variance with the classifications of Peters and of Hellmayr and Conover. Within each genus, inter- and intraspecific \'ariations were then studied to determine the stage of the speciation proc- ess, evaluating thus the significance of mor- phological differentiation. Finally, a review of the existing forms and their distribution, coupled with the fossil record, ser\'ed as a ^ This paper was published \\ ith the aid of Na- tional Science Foimdation funds from grant GB- 3167 to the Department of Biology, Harvard Uni- \ersitv. basis for a reconstruction of the origin and evolutionary sequence leading to the mod- em species of the genera Penelope and Crax. I must emphasize that I ha\'e not at- tempted a formal revision of the species of Cracidae, for two reasons. First, because whether a form is treated as a species or as a subspecies was considered irrelevant to the aims of this study; second, because much more work on museum skins, supple- mented with collecting in critical areas, must be done before a definitive arrange- ment of the species can be reached, and before the \ ariation nithin each species can be properly assessed. The list of recognized species, particularly in the genus Ortalis, must therefore be taken as tentative, pend- ing the result of studies currently under- taken by others. The concepts adhered to in this paper are those of broad genera and of the biological species, as defined by Mayr, Linsley, and Usinger ( 1953 ) . In the case of allopatric, non- or slightly overlapping forms, the cri- teria for specific recognition are those men- tioned by Mayr (1963: 345): "complete lack of intergradation in a continental area, in the absence of geographic barriers, is now taken as proof of specific distinctness." I am deeply indebted to Dr. Ernst Mayr for his help throughout this study, and for his critical comments on the manuscript. I am most grateful to Prof. Bryan Patterson for his assistance with osteological prob- 2 Miisctini of Comparative Zoolou.!/. Vol. 134. \o. 1 lems and with fossil Cracidac. Dr. Ernest \\'illiams. Prol". Hryan Patterson, Drs. Wal- (i r I. H(nk. W . John Sniitli. Mr. Reginald E. Mori'aii. and Mrs. H. P. Hall read tlie manu- script and suggested niau\ iniprox enients for whieli I am ver\- thankful. I am mucli ohliged to Drs. Philip S. Humphrey ( United States National Museum), .\lden II. Miller ( Museum of N'ertebrate Zoology, Berkeley ), and Pxobert W. Storer (Unixersity of Michi- gan Museum of Zoology) for lending me skeletons of Cracidae. I must thank the offi- cials of thi' Museum of Comparative Zoology for their cooperation during the course ol m\ work, and those of the Academy of Natural Sciences of Philadelphia for kindly placing the collections at my disposal during a \isit to this institution. 1 am grateful to Dr. Oli- verio Pinto for his valuable information on the distribution of some Cracidae in Brazil. Mr. John A. Criswold has most kindly sup- plied photographs of Crax iiiitii and has given man\' useful details about these birds living at the Philadelphia Zoological Gar- den. FinalK- I want to thank Mrs. Alice J. Daiiitls for preparing f^igures 1 through 9. THE CRACIDAE \'ari()us and recent kinds of evidence tend to indicate that the Cracidae are more closeK- related to the other groups of Galli- formes (in particular to the pheasants) than has been suggested in the past (Sibley, H;r-i(); Mainardi, 196.3). The cracids differ from all other gallinaceous birds b\- their arboreal habits, their nests of sticks built in trees and bushes, and their clutches of 2 to 3 (rarely 4) big, rough-shelled white eggs. The majority of species are found within the limits of the American tropics; the great- est number occur in South .\merica, par- ticularK in the northwestern eonieiv Most CJracidae li\e in the forest interior, some in more open situations: forest-edgi's ,111(1 brushy habitats. Some species of curas- sow s ((Wax) and guaus (Penelope) require humid conditions, such as are found in low- land rain forest and montane rain forest (cloud forest). ()nl\ a few species of cha- chalacas (Ortalis) prefer drier habitats. No species is found far from trees, however, which provide them with both shelter and food. The latter is mostly vegetal, and con- sists of fruits, leaves, buds, vine tendrils, and an occasional insect. Several species are quite gregarious, especially after the nesting period, and loose flocks of Penelope or 0/7C///.S can be seem feeding on frnit-laden trees. CLASSIFICATION Huxley (1868) divided the Cracidae on till" basis of a difference in the proportions of the pelvis in some genera, concluding (p. 297): "The Penelopinae (Penelope, Oreo- phasis\ OrtaU(la) have the postacetabular area broad; but in the Cracinae (Crax, Paiixi) it is narrow." Sclater and Salvin ( 1870) adopted this "trenchant difference," coupled with minor differences in bill shape, to distinguish three subfamilies, essentially the same as Hu.xley's except that Orcophasis was placed in a subfamily of its own. Their subfamilies were: 1) Cracinae (Crax, Notho- cm.v, Mitu, Pauxi) 2), Penelopinae (Stegno- laema, Penelope, Penelopina, Pipile, Ahur- ria, Chamaepetes, and Ortalis), and 3) Oreophasinae (Oreophasis) . Peters (1934) brought only minor changes in the generic arrangement of Sclater and Salvin. Stegno- hteina was correctly assigned to Penelope, and the name Ortalis replaced Ortalida, since the latter seems to be the accusative of the former. He did not recognize any subfamilies, nor did Hellmayr and Conover (1942) who, while differing somewhat in the treatment of species, did not alter Peters' generic grouping. Among recent authors who revived the subfamilies are Ridgway and Friedmann (1946), and Tordoff and Maedonald (1957); Verheyen (1956) sub- dixided the faniiK' into four tribes: Cracini, Oreophasini, Penelopini, and Pipilini. I have restudied the osteological differ- ence cited by Huxley and have been able to confirm his findings. I examined the fol- lowing species (number of skeletons in parentheses): Ortalis vetula (4), O. ?ii,ar- Relations of Cracid Genera • ViiiUcumier ruhi (1), O. ruficouda (1), Chanwepefes miicolor (1), Fcndopina nigra (2), Fe- nelope purpurascens (4), Crax urumutuin (1), C. scdvini (1), C. fo?/!c/ifo,srz (2), C. //i/7f/ (3), C. fasciokita (4), C. (///;c;Yi (1), C. alccfor (2), and C. rubra (5). On the basis of measurements of total pelvic length, length of the preacetabular area, length of the postacetabular area, and width of the postacetabular area, it appears that the post- acetabular area is relatively broader and shorter in Ortalis, Chamacpefes, Fenelopina, and Fenelope than in Crax. It seems, how- ever, that Huxley emphasized this differ- ence without enquiring into its possible significance. As a rule, birds with more terrestrial habits tend to evolve a propor- tionately longer pelvis than birds of more arboreal habits ( B. Patterson, pers. comm.). This probably applies to the Cracidae, where the more terrestrial Crax have a relatively longer pelvis than the more arboreal Orfalis, Chaniacpcfcs, and Fcnclope (the rather ter- restrial Fcndopina being an exception). The adaptiveness of this difference in pelvic proportions, and further the very close rela- tionships within the Cracidae do not, in m>' opinion, justify a subdivision of the family into subfamilies or tribes. My main disagreement with the classifi- cations of Peters, and of Hellmayr and Con- o\er, concerns the generic limits in the curassows: as I explain later in more detail, I see no reason to maintain Nothocrax, Mitu. and Fauxi as separate genera; all the curas- sows are united under the generic name Crax. Similarly, I am convinced that Fipile belongs in Penelope, of which it should be- come a synonym. Abiirria and Chamac- pctcs are kept as separate genera for lack of better evidence, but their close relationship to Fenelope is indicated by their position in the list. Fenelopina is also kept separate, but its intermediate position between Fe- nelope and Orfalis must be emphasized. Insofar as possible, the sequence of genera expresses the degree of relationship and the amount of specialization (from least to most specialized ) . Thus it differs consider- ably from Peters' and Hellmayr and Con- over's sequence. In the main, Hellmayr and Conover's spe- cies limits have been accepted, except for the following: Ortalis cohimbiana is here considered a full species, on the authority of Miller (1947, 1952), and not a subspe- cies of guttata. Ortalis poliocephala and O. vetula are considered distinct species, on the basis of field data from Wagner ( 1953 ) ( see also Moore and Medina, 1957 ) . Pend- ing furtlier evidence, O. leucogastra is re- tained as a subspecies of vetula. The close relationships of the Ortalis poliocephala- vctula-ruficauda-ruficrissa complex are in- dicated by treating tliem as members of a superspecies. Ortalis icagleri is considered conspecific with poliocephala ( Moore and Medina, 1957). Fenelope eujubi and jaeu- tinga are treated as subspecies of pipilc. Similarly Fenelope ochrogaster and pileata are merged with jacucaca. Fenelope jac- quagu, F. granti, and F. obscura are con- sidered conspecific with P. purpurascens. Fenelope albipennis is believed to be a color variant of ortoni, and ortoni is treated as a subspecies of F. montagnii. Crax uni- cornis is treated as a subspecies of C pauxi, as already done by \\'etmore and Phelps (1943). The classification outlined below is based on the various kinds of evidence later dis- cussed in detail; it is presented at this time for imiformity of nomenclature throughout the paper. Superspecies, as defined in Mayr (1963: 499, 672) are included in braces. Classification of the Cracidae Genus Ortalis Menem 1786 motmot ( Linnaeus ) supcrciliaris ( Gray) guttata (Spix) \columhiana Hellmayr {poliocephala ( Wagler ) \ vetula (Waffler) \ruficrissa (Sclaterand Salvin) [ruficauda ( Jardine ) garrula ( Humboldt ) enjthroptera (Sclater and Salvin) canicollis (\\'agler) Genus Fenelopina Reiehenbach 1862 nigra ( Fraser ) Muscuni of Comparative Zoology, Vol. 134, Xo. 1 Genus Penelope Merrom 1786 1 ) pipile spccic's-group \pipih' (Jacquin) j ciimaneiisis ( Jaccjuin ) lacucaca species-sroup jacucaca Spix •i ) purpurascens species-group purpurascens \\'agler 4) nionta^nii species-group »ionta- in cracids that are sexually dimorphic and monogamic like C. rubra, or hardly dimor- phic but polygamic like C. alector. Alone among the members of the genus Crax, the species urumutum does not have any black in its plumage. The black is re- placed by mottled brown, and the paler lower belly is uniform buff. Thus, the same pattern is visible, though not so obvious, as in the other species. Other characters such as bill shape, crest feathers, and tracheal loop, all point to its close relationship with other species of Cra.v, especially of the mitu species-group. In the genus Penelope, the pipile and jacueaea species-groups differ somewhat from the other species-groups in color pat- tern. The most conspicuous difference lies in the color of the crest feathers: white in pipile, brown in jacueaea, in both cases paler and in contrast with the color of the rest of the plumage. In the other species- groups the crest feathers are the same color as the neck and most of the body plumage. The broad white patch on the wing of pipile is subject to extensive individual var- iation, and can be much reduced in some birds producing an effect not very different from the wliite spotting of other species of Penelope. Aburria aburri and the two species of Clmmaepetes all have unifonn black or brown coloration, thus differing somewhat from Penelope. They are more generalized than any species of the genus Penelope, \\ ithout erectile crest feathers, and without wattles or areas of naked skin in Chamae- peies. Of all cracids Oreophasis is the only one to have a peculiar plumage pattern, with a conspicuous white bar across the middle of the tail, and white plumage on the breast. DISCUSSION OF CHARACTERS \\'ing ( 1946 ) brought together infonna- tion on the courtship perfomiance of North American species of Tetraoninae, showing the correlations between morphological characters such as erectile feathers and throat sacs, and behavioral ones such as dancing and drumming. It is probable that when the Cracidae are better known a simi- lar analysis will be feasible. At present only some comments are possible. 1) A casque or a frontal protuberance is present in most species of Crax. This struc- ture is always brightly colored, usually yel- low or red, more rarely greenish or bluish. The species that have no casque have a red or yellow bill. Long crest featliers are found in most species of the genus, the exceptions being C. pauxi (casque) and C tomentosa (no casque). The trachea is convoluted in the males only. The voice is deep, described by \arious authors as 10 Miis( iini of Coniparativc 7ah)Io'J,\I . Vol. 134. \o. 1 "hoots," hoominu." etc. (exception: C. (laiihcnioni. see Schiifer, 1953). There is e\icU'nei' that loud caHiiiti; is done only by the males and that it has territorial func- tions, rraeheal sliiKtme and \()iee, as well as color pattern, exhibit a remarkable uni- forniitv thronghont the genus. Clo.se rela- tion.ship of the species is suggested by the almost conipleic huk of genetic isolating mechanisms ( postmating mec'hanisms of Ma\r. 1963: 92) shown in h\ bridization experinu-nts in capti\ity, as explained later. It seems that the species of the genus C.rax are in a stage dl (he speciation process where ecological and ethological isolation in the narrow /.ones of o\'erlap is sufficient to prexent interbreeding ( premating mech- ani.sms (a) and (b) of Nhiyr, 1963: 92). 2) All .species of the genus Fciiclopc ha\e a wattle, in the form of a dewlap or an area of naked skin on the throat. .\11 spe- cies have erectile crest feathers used appar- ently in connection with the wattle, pos- sibl\- during the territorial calling, as rein- forcement of signal characters, 'llie tra- cheal structure \aries, from species lacking convolutions in both sexes to species that have them in both sexes. There is an ap- parent correlation between s]')ecies with a simple trachea and a special drumming flight, the significance of the latter being as \ et obscure. 3) TJic ()nl\- species of Pcuclopiiui is much more sexually dimorphic than any other species in the other genera, and there is no pair Ixmd (Wagner, 1953). Flight of the drumming kind and voice similar to that ol Orlalis vctula (Pitelka, cited by Leopold, 1959 ) ha\e been described and make Pe- nel(>))ina n/gra somewhat intermediate be- tween Penclo])t' and (^rtalis, but maybe still distinct enough to be maintained in a dil- ferent genus. 4) (hidliw hke ('rr/.v, exhibits a great uni- formity in several characters, such as plu- mage pattern, voice, tracheal structure, and tendency to polygamy. There is no ({ues- tion as to the limits of the trenus. 5 ) To sum up, it is evident that characters such as the casque, relied upon heavily in the past to establish genera, are only of specific xalue; there is thus no reason to keep Noiliocmx, Paiixi, and Mitu separate from Cirix, where they undoubtedly belong, bi Cra.v, color pattern, voice, and tracheal structure are much better clues to the rela- tionships than wattles and or casque. In Penelope, there is much more variation in color pattern, as shown by the pipile spe- cies-group, and there is also more variation in tracheal structure. It becomes thus more difficult to give diagnostic characters for the genus. The throat wattle seems to vary less throughout Penelope than other char- acters. On the basis of such evidence it is unnecessary to maintain a separate genus for th(^ pij)ile superspecies, which is only a distinct species-group within Penelope. It is i-)ossible that after more study, Aburrki and Chamaepetes will be also included there. HYBRIDS As far as I know no cracid hybrid has ever been reported in nature. All the cases mentioned in the literature concern hybrids obtained in captivity (Gray, 1958). Even in southern Mexico where Ortalis vetula and O. poliocepJwhi are sympatric (Wag- ner, 1953), or in the Upper Sinii Valley of northern Colombia, where Crax ruhni and C alherti have been found in the same area ( Blake, 1955 ) , no intergradation seems to take place. The latter case is quite inter- esting in view of hybrids produced in cap- tixity between C. alherti and rubra (Taibel, 1950). Other crosses include C. alberti X C. fasciolata (Bronzini, 1940, 1946), and C. (ileetor X C. rubra ( Ackermann, 1S9S; Przi- bram, 1910), and concern broadly allopatric forms. Taibel (1961 a,b) crossed Crax alherti and C. mitu and found the Fi and F^. hybrids to be completely fertile, as well as the offspring of the backcrosses. In Pe- nelope, Fi hybrids were obtained between nuirail and the rather different jacucaea pileata ( Hopkin.son, 1926, 1939). Relations of Chacid Genera • Viiillcimiicr 11 Tahle I. Fossil Cracidae Fossil, FonMATION- Time Locality Author OriciUs ■'Upper Snake Creek Pliocene South of Asate, Sioux Co., Wetmore (1923) phengitcs Beds" Nebraska Boi CO) talis Hawthorn Formation, Lower Thomas Farm, Gilchrist Co., Brodkorb (1954) lacssh'i Thomas Farm local fauna Miocene I'lorida Oitalis Rosebud Formation Lower Flint Hill, 9 miles west-south- Miller (1944) ]H>Uivaris Miocene west of Martin, Bennett Co., South Dakota Ortalis Harrison Formation Lower Cameffie Hill, Sioux Co., Wetmore (1933) tantala Miocene Nebraska Palaconossax Brule Formation, Upper 5 miles south of Scenic, South Wetmore (1956) sencctu.s Poleslide member Oligocene Dakota Procrax Chadron Formation, Lower Pennington Co., South Dakota Tordoff and hrcvipcs Peanut Peak member Olijiocene Macdonald ( 1957 ) Calliiuiloiclcs Cireen Ri\ er Formation Middle Fossil, \V\oming Eastman (1900) u'l/oiDin^ciisis Eocene True intergeneric hybrids were produced between Crax rubra and Penelope jociicoca piJeatcL but when adult the birds did not show any sexual activity (Poulsen, 1949, pers. comm. cited by Gray, 1958 ) . Further- more, Ortalis (guttata and Galhis domesticus ■'are said to cross readily" (Gray, 1958). From this evidence of hybridization in captivity but not in nature, it appears that allopatric populations have reached a stage of speciation where morphological differ- entiation and also ecological divergence have taken place in isolation so that, as in tlie Francolins Francoliniis "interbreed- ing is unlikely (but not impossible) if the two populations rejoin" ( Hall, 1963 ) . FOSSIL CRACIDAE' Se\'en fossil cracids have been described in the literature. The oldest, Galliniiloidcs wyomingensis, from the Green River shales of Fossil, ^^Voming, attributed to the Mid- ^ After this paper was completed, I learned of Brodkorb's Catalogue of Fossil Birds, Part 2 ( Bull. Florida State Mas., 8 (3): 195-335, 1964) in wjiicli a number of fossil birds were removed from \arious families and referred to the Cracidae. If these changes are accepted, it means that cracids were present in France from the LTpper Eocene to die Eocene, was placed by Lucas (1900) in a family Gallinuloididae mostly because of the absence of a well-developed post- angular process. Professor Bryan Patterson and I have re-examined the type but cannot be satisfied that the process is actually lack- ing or that it was broken off. As Tordoff and Macdonald ( 1957 ) rightly pointed out, however, the process "does not seem to be a valid basis for establishment of a separate family." I agree with them that GalJinu- loides belongs in the Cracidae. Procrax brevipes is the only other fossil of the family known from more than a bone fragment. In their careful study, Tordoff and Macdonald (1957) compared Procrax with a number of modern genera. On the basis of their description there seems to be no doubt that Procrax is a valid genus, but I would question putting it into a subfam- ily together with GaUimdoidcs. In view of its numerous similarities to modern cracids, the Lower Miocene, in Argentina in the Middle Miocene, and in North America as early as the Lower Eocene. The presence of cracids in the Miocene in South America (if proved) is not in conflict with the \iew expressed in this paper that the modern cracid genera originated in North America. 12 Mtis( inn of Comparative Zoolov i) "^^ XX.. ,^a { .' c. ^~^->^ ) '^ o ; \ °o( /c"^^W»^-^^^ ^^°V .^m^J^ F'°^-^''Tr'- "'' ■' a_ ^°7%,lr-^^V [ >-r^A <7 o^ --,/ ^ c. ^ ^^T~ / o {7^1 .0 r~' o I y^ \-X:^ ^^ s o o ( ( s - \ "/ O/- ^L. ' -^ ^ ' ■'' " ' "^ / ' 1 Y I'-j;' 1, , ^••^// r") 7 ""l-'- V ^\ *^ ^ :^ v— --^V / (■ ', \ ~!-'f\ ^ \ ^ -^ ( \ I '-•'' '' ') o ^^ ^ / ; ^v^ ^.'"^ )i \- ^ - -^ ■'' ''^ \ ■• ' ^ / ~ "^V '---1 _ J \ ■\ ' ( ■ _ 1 / "' V ■* /^ / I >^^-^'' ■'.^- •-'' '- ■ ^r D PURPURASCEIMS } l"''"''^^-' ' "" "^ ^ ! 7^'"> '' '*^ A^"^""^^ O JACQUACU / / ^^'v'""-"', A^'^.'^V7■" -'^ fT • GRANTI 1 ;■ ^ , V' r ^ ^- ^\ ■- ''\^ ' ■^ *- X ^ OBSCURA ) 1 -^ N'^^-V ^ /f, -/ ( / ^ s '. Z*"^ - -V^ i'-./S ^ ^'^^^yfr ^ \ , 1 ' 1,^ AC^ r;' : ■ ^ Fig. I2. Distribution of the Penelope purpurascens species-group. increase of more xeropliytic vegetation nn- favorable to these l)ir(ls. In view of the lack of marked chfferentiation between hricJ^ S'^ ' '-^"■-^■■'■^. ^ i 1 Am \C./'^'r'"l V- !_• V i ^'— ".' \ V ^ -^ ^ s ' -v' 1 .--'•■ ' — ; P-./ //■ ; ,' ^ ^ ^ yr >'/v .'--' s ^- * . / '""^- V*-^'^'- ^/> ■■ '^ '" / ' l^^ » — '' *" "^ ^. ^ -^ - / >- ■ ■' C ^x"^ "'^' ' / -J'';"\ -^ s ^ ^^y ' '' ^"\ -■ / (^7 ^..^ L^ '-•'- ^■' V» ^' 1 ^ - , r- Fig. 13. Distribution of Penelope argyrotis. Brazil {'^granti-orienticok/), and of south- em Bolixia, Argentina, Paraguay, Urugua\", and southern Brazil {^'bridiicsi-ubsciira- bronziiur ) are slightly different in color and size, respectively, from the birds of Colombia, Ecuador, Peru, western Brazil, and northern Bolivia {"pitrpiirascens-jac- quagu"). These peripheral populations may have been isolated for some time as sug- gested by the probable stepped clines. Of course, only examination of more extensive material will pennit a definite conclusion about the variation ^^•ithin this complex, and its interpretation. The o\erall pattern of \'ariation and of distribution of populations of the ptiipiir- ascens species-group suggests that the pro- purpurascens ancestor probably reached South America from the north in the Pleis- tocene and, once there, dispersed rapidly into all suitable habitats. The total range of the group corresponds closely to the extent of dense humid forest (rain forest, montane forest, gallery forest along water- courses) in Central and South America. 4) Monta ' S^ ) • . •• '.A'"--- ^ I ^-- \ ) • :J -- ' s . ' • / ■ /~"-v ''"''-- * y j-'^^ '^■\,--'\.^5l ,' "i- , — ^^^"^ "V.--- >lP^>J-~'^^ / ' \ }.-^f..'--^^ ; ^- / ^ ^ •/ ' -^ v/ ~^ ^ K/ Fig. 14. Distribution of Penelope superciliaris. species, and totally lacking in the others. The species in this group are generalK smaller than the populations of purpur- ascens. P. argyrotis is distinguished from all other fomis of the montagnii group by the cinnamon color of the tip of the rec- trices. It contains six named subspecies, morphologically quite similar. The most distinct is columhiana from the Sierra de Santa Marta in northern Colombia, and not, as one would expect, the isolated barbata of southern Ecuador and northern Peru (Fig. 13). Brown edges of the secondaiy wing cov- erts and innennost secondaries are distinc- tive of P. superciliaris, much more so than the silvery supercilium, which is quite vari- able; some subspecies of P. montagnii, in particular plumosa and sclateri of Peru and Bolivia, have a stripe reminiscent of the one in superciliaris. This species comprises four subspecies showing clinal color varia- tion, occurring in Brazil over a vast area, west to the Rio Purus, south to Rio Grande do Sul, and across into Argentina (xMisiones) and Paraguay ( Fig. 14 ) . The third species contains P. montagnii. with 7 described subspecies from northern 18 Mii\cin)i of Compdrdtiic Zoolo<:ii, Vol. 134, A'o. 1 O MONTAGNI • ORTONI ALBIPENNIs/ '■■y^j Fig. 15. Distribution of Penelope montagnii. For discussion of ortoni and o/b/pennis see text. Colombia and Venezuela to Argentina (Fig. 15). Olrog ( 1960) rightly showed that dab- benei of southern Bolivia and northern Ar- gentina is only a subspecies of inonfagnii and not a distinct species, as formerly bc- licxcd ( Ilelhna\r and Conover, 1942). All Nariatioii in inonUiiinn seems to be clinal but this should be studied with more ma- terial than 1 have seen in order to check whether the variation between brooki, ])}umom, marcapatensis, and sckiteri, and possibK also dabbenei, represents one or more stepped clines. There is the problem ol ihc allocation of three other forms: marail, ortoni, and albi- pcnnis. P. marail is widely distributed in Brazil north of the Amazon, in the Guianas, and a small area in Venezuela (Fig. 16). For a long time this fonn was confused w ith P. purpitrascens iiranti, for example by Peters (1934), because the two are rather similar in the color of the upper parts, glossed with metallic green. In other color characters and proportions, however, marail is distinct from ?^ Vv • •■■^ \-~-^ 7 - ;• -w rv /■ \ S-.---" V '\ • ^ *. ( \ \, \ i ■^ ^ ..• \ -w V ■'^ . •^ ^' -^X.y \ l' \ / \ y ""j \ -//^ Kr^^ i__ \ /T/^ 5^^ • •• ^ ^ y 1 1 - . Fig. 16. Distribution of Penelope marail. Relations of Cracid Genera • VtiiUeumicr 19 Prado" in Lima ( Ridoutt, 1939). In this form, the eight outer primaries are white with the base and the tip dark. This is reminiscent of partial albinism found in a population of P. monta^nii ortoni in Colom- bia. Six specimens in the Academy of Nat- ural Sciences of Philadelphia collected by A. de Buey in the Choco present various amounts of whitish-isabelline in several parts of the body: neck, breast, back, wing coverts, primaries, secondaries, rectrices, and the head (crown feathers). It is note- worthy that two other birds collected at the same locality by de Buey are similar to nor- mal specimens of ortoni. This suggests that some popidations of oi'toni have a tendency for developing whitish areas in the other- wise dark plumage. I strongly suspect that the so-called albipennis is nothing else than a similar variant of ortoni. Sympatry between the species of the nioiit(ii:,nii group occurs only in a few nar- row zones: in the Perija Mountains (Colom- bia-Venezuela boundary), and in south- western Venezuela, between P. tn(>nt(i. '\ ( _ _ _ > T/^ • ALECTOR ( \-~.'' \ 'i -■^■^ o I. ^ '-",-0 t// ▼ BLUMENBACHII \-A / -/ VJ o "V ' /"^ O FASCIOLATA / l.>?:~l c. r. J ' - ''^ /.'■,:.. N<-o,- %' ' ' i /-'^"'^'"'^ 1 .■_ ' >.o 1 ^ ' ~ i)- 0 j~y. - 1 : \^ - - - 1 / ; -' ( ; \- ; o<- ; Fig. 17. Distribution of the Crax rubra species-group. only wliite barrings in the black plumage). Both .sexes of all forms have a crest of long, always erected, and curly feathers. All forms except hlumcnhadiu have an area of bare skin on the face, sometimes restricted to a small /one around the eye (rubra), sometimes much more extensive (fascio- lata). Nh)st forms have a frontal protuber- ance, sometimes (juite big {rubra), usu- ally bright yellow. .Some forms have in addition to the frontal knob a pair of wat- tles at the base of the lower mandibles, that have the same color as the knob. It seems that forms that ha\'e no frontal protuber- ance ha\t' a bigger naked area on the face than the othcis. There is much \ariation in iK)th wattles and knob, however, especially in C. I ^ — — 1 ■^. ( .. .1 . ^ / 1 Fig. 18. Distribution of the Crox mitu superspecies (mifu-fomentoso-so/vin/ and territorial, while the (almost) non- dimorphic C. alector seems to be polyga- mous, and C. daiibentoni, where the female is much masculinized in aspect, seems to occupy an intermediate position, in which polygamy occurs but may not be absolute. For the present I consider it best to treat the various fomis as specifically distinct. Crax ruhia and C. globitJosa look more alike than other forms, so I have included them in a superspecies. Similarly, the three forms alberti, annulata, and dauhcntoni are closer to each other than to any other form, and I include them, too, in a superspecies. The status of anniiluta requires further study, however, and its position as a species must be considered provisional, pending further evidence. It is possible that a detailed study of variation in this species-group \\'ill show all fonns to be members of a single superspecies. The genetic compatibility of \arious spe- cies crossed in captivity- indicates that spe- ciation is in an early stage. The rapid de- velopment of structures such as wattles, frontal protuberance, and areas of naked skin on the face suggests a rapid and recent diversification in isolated, allopatric popu- lations, and should not be taken as an index of the degree of differentiation of popula- tions. Rather, uniformity in tracheal struc- ture and voice, in color pattern, in habitat, and in diet points to the rubra species-group as a very young one. The distribution of the group suggests a rapid dispersal from a pro-rubra ancestor that was able to colonize widely in South America in the Pleistocene. This ancestral population may have been quickly broken do\\'n into se\'eral subpopu- lations isolated from each other by some ecological barriers and resulting in the pat- tern observable now. 22 Miisciint of Comparative ZooJo.^y \~\ ^^ -^r N / - -v ^V^ ^ >-- L'^^im^y^-'Y^'" :5 > *''~V^^_^ 1 A ' ! ^'^ ''V/ ^ ' * ._/■"■ N ' ^-^ ' -^ " •• ;^* / » 1 /-/ '/ 1 '■' ^ \ { ■\ .—..^ X r^'-X,-'""' ^ 3 *-^ / — ^ , /-' /--^\ V\' ^v '■ ''^■■■^.y \ >., '^■;\>->;-^ \ i' \ /- h J h>-'^^-., V r/ _ / -'-•'^ / ^ • v^ Ci^^H •'' ^'^-1 ■•' i-l ^^v ••^"^ /' / (-*.•->.. ' l ' .-— ■ ^' 1 1 i V "l '' .C 1 J\ \ \'^-' 1 ^■■•^1 x' 1 V^^l'^'^A^" — '/ A \ ■-'•-^ ' ■; \ ''^ ) t \ ' / V' ' >'N \ Vw y i.-^ f f \ ''' ^ \1^ ' *v V ::^ (. ^^ / 7 .■' ' "~v // -\ V.—^'^V .^-' •"! \ \_ / ■■■' \_ J, j ; V - '^~' Fig. 19. Distribution of Crox pouxi. 2) .)///// Species-group The fi\ e forms of this group are also veiy closely related to each other and obviously derive from a common stock which in turn is common \vith the ancestral stock of the rubra species-group. The differentiation is so slight that the separation into the two species-groups must have been recent. Here again, allopatr\ is the rule, espe- cially in the mitu superspecies (see Figs. 18 and 19). As far as I can tell, the onl\ area of svmpatiy is in the southern Sierra de la Macarena, Colombia, where Crax salvini and C. tomentosa were taken within a few days of each other (Olivares, 1962). C. unmiiihim (Fig. 20) broadly overlaps the ranges of rnifii in upper Amazonian Brazil and Peru, salvini in Ecuador, and tomentosa in southwestern Colombia and western N'enezuela. These five forms show a great uniformity in plumage pattern in both se.xes, and in tracheal structure. \^iri- ation is found in the absence or presence of elongated crest feathers, in the presence or absence of a casc|n(\ and in the color of the lower belly. The females are always similar to the males in plumage color, but the\ are somewhat smaller. A brown color phase in Crax paiixi was first described by Sclater and Salvin (1870), and studied by Phelps and Phelps ( 1962 ) . Four birds out of 64 in European, American, and Venezu- elan museums showed this rufous colora- tion. Two of them were unsexed, and the other two were females. This plumage does not seem to be juvenile since an unsexed brown bird kept in the Maracaibo Zoo for years never changed color. There is little doubt that three of the species, tomentosa, mitu, and salvini, are closer to each other than to the other two. WHiether they should be considered a single species is debatable in view of the possible s\ nipatn- of two of them in the Sierra de la Macarena. For tliis reason, I have treated them as a superspecies. Crax pauxi differs from the preceding three in its enoiTnous casque and should be regarded as a distinct species exhibiting some variation in the dimensions of the casque and the shape of the crest feathers. The disjunction (Fig. 19) may be less striking than it seems, be- cause it would not be surprising to discover intermediate populations in Ecuador or Peru. This disjunction reminds one of that of Tinamus osgoodi, allied to T. tao, which has been found in the Rio Aguas Claras, X _^ f ^9 '^^ Fig. 20. Distribution of Crax urumuium. Relations of Crac:id Genera • ViiiUnimier 23 Fig. 21. Variation in the size and shape of the casque of Crox mitu. Upper left: young bird (age d'A months). Upper right; adult female. Lower, left and right: adult male. (Photographs courtesy John A. Griswold and Zoological Society of Philadelphia.) Huila, Colombia, and in the Marcapata Val- ley, Cuzco, Peru. In either case it is diffi- cult to account for the disjunctions, but one may sumiise that among such sedentary and territorial birds as Crox paiixi and Ti- namiis osgoodi extinction can occur in con- tinental forest areas as readily as in island areas. DISCUSSION From the analysis of variation and distri- bution, it can be seen that Penelope and Crax show the same basic patterns of speci- ation. In both these genera, the members of a species-group are mostly allopatric representatives of one or more superspecies. These representatives, or species, are, for the main part, in a relatively early stage of speciation. Whether they should be called species or subspecies is not always easy since the criterion of interbreeding in na- ture can not be applied as it could with sympatric species. In the case of Crax in particular, moq^hological divergence might lead to the belief that the various forms have reached levels of speciation beyond those actually attained. The evidence of hybridization in captivity, but not in the wild, is here a far better criterion than mor- phology alone since it indicates that an intennediate stage in the development of the isolating mechanism has been reached (i.e., there is little genetic isolation). 24 Muscurt) of Comparative 7a)oIooij, Vol. 134, No. 1 Tliis general pattern of allopatric species most likel)- ethologically isolated in the few narrow areas of overlap, suggests that the ancestral stocks of the species-groups in both Penelope and Cra.x reached, or have been in South America only since the Pleis- tocene. Although the speciation pattern of OiidJis lias not been described, a prelimi- nar\- inxcstigation rexealed a situation simi- lar to that of Penelope and Crax. This same pattern of radiation in the three genera that comprise the majority ( 87 9^^ ) of the species of the Cracidae strongly suggests that a nearly simultaneous dispersal preceded dif- ferentiation in allopatric populations ( prob- ably isolated by slight ecological differ- ences) during the Pleistocene. The fact that each genus has such a broad distribu- tion in all favorable habitats of Central and South America, together with a considera- tion of the stage of speciation in which they occur, suggest further that the habitats were empt)- at the time of dispersal and that the ancestral forms probably came from the north outside South America. The four other genera that do not exactly correspond to the patterns of OrtaJis, Pe- nelope, or Crax are: 1) Oreo])]}(i6-i.s, restricted to the high mountains of southern Mexico and Guatemala, 2) Penelopimi in moun- tains from southern Mexico to Nicaragua, 3) Chomaepefe.s; with one species (iini- color) in the mountains of Costa Rica and Panama, and the other (gondotii) in the Andes from Colombia to Peru, and 4 ) Ahiirria in the Andes from Colombia to Peru. These four genera appear to be descendants of a stock that underwent an early radiation in North America. Two of them, Ch(nnae))ete.s- and Ahinria, have con- served some of the supposed primitive char- acters of the ancestral form: veiy arboreal habits, ecological association with humid forests, lack of patterned plumage, and, in the case of Charmiepetes, absence of head or neck ornaments or areas of naked skin. Penelopinu shows what is certainly sec- ondarily accjuired sexual dimorphism, com- plete polygamy, and partialK' terrestrial habits (nest regularly on the ground). In appearance, Oreophasis is the most unique cracid and also exhibits some secondary spe- cialization, such as ground-nesting habits. At some point during the late phases of tliis North American radiation the forms that were going to invade South America and give rise to the modem species of Penel- ope and Crax evolved from a stem closely related to Penelopina, Chainaepetes, and Ahiirrki, while Orfalis had already under- gone radiation when some stock reached South America. SUMMARY 1 ) The majority of species ( 87% ) of Cracidae belong in the three genera Ortalis, Penelope, and Crax. The former genus Pipile has been included in Penelope, and the former genera Nothocrax, Mitu, and Paiixi synonjanized with Crax. Among the genera retained, Chamaepetes and Ahurria were found to be very close to Penelope, Penelopinu to be intennediate between Orfalis and Penelope, while Oreophasis is somewhat aberrant. 2 ) Variation and function of casque, crest feathers, wattles, attenuated primaries, tra- cheal loop, and color pattern are discussed. It is shown that structures such as casque, wattles, and areas of naked skin on the throat are probably of use as species-spe- cific recognition marks, and hence should not be utilized as taxonomic characters for distinguishing genera. In Ortalis and Crax, plumage pattern, tracheal morphology, and voice are very uniform. In Penelope, color pattern and tracheal structure are variable, but all species have a wattle (dewlap or area of naked skin on the throat ) and erec- tile crest feathers. 3) No hybrids have been found in nature, Init hybridization in captivity is not rare. In only one series of experiments was hybridi- zation attempted by the aviculturist beyond the F] generation. In this case, crosses be- tween Crax mitu and Crax alberti produced completely interfertile F2 offspring. From this evidence it seems probable that in Relations of Cracid Genera • VuiUeumier 25 nature isolating mechanisms are ethological (as well as ecological) in the narrow zones of sympatry. 4) Seven fossil cracids show that the family occurred in North America from the Eocene to the Pliocene. This record indi- cates that OrtaIis-\ike cracids were present in North America since the Upper Oligo- cene. The only fossil remains found in South America are of Crax and Penelope in Pleistocene Brazilian cave deposits. It seems probable that the Cracidae originated in the warmer parts of early Tertiary North Amer- ica and radiated in this region long before they reached South America. 5) An analysis of speciation in Penelope and Crax reveals that in each genus the "species" are in an intermediate stage of the speciation process, i.e., the "species" are allopatric members of superspecies. In no species-group is there any broad sympatry between related species. It is postulated that the three main genera of Cracidae, Orudis, Penelope, and Crax, have colonized South America in the Pleistocene, by means of ancestral forms that originally came from North America. REFERENCES CITED AcKERMANN, G. E. 1898. Thierlxistarde. II Theil: Die Wirbelthiere. Weber and Weiden- meyer, Kassel, 79 pp. Alvabkz del Toro, i\I. 1952. Las animales sil- vestres del Chiapas. Ediciones del Gobiemo del Estado, Tuxtia Gutierrez, Chiapas, 247 pp. AxELROD, D. I. 1958. Evokition of the Madro- Tertiary geoflora. Bot. Rev., 24: 433-509. Beebe, Mary B. and C. William Beebe. 1910. Our search for a wilderness. Holt and Co., New York. Blake, E. R. 1955. A collection of Colombian gamebirds. Fieldiana, Zool., 37: 9-23. Bock, W. J. 1958. A generic review of the plovers ( Charadriinae, Aves). Bull. Mus. Comp. Zool., 118: 27-97. Brodkorb, p. 1954. A chachalaca from the Miocene of Florida. Wilson Bull., 66: 180- 183. Bronzini, E. 1940. Descrizioni di un ibrido Crax fasciolata X C. alherti. Riv. ital. Orn., Ser. 2, 10: 11-21. . 1946. Dieci anni di attivita avicola del Giardino Zoologico di Roma. Riv. ital. Orn., Ser. 2, 16: .3-24. Cain, A. J. 1954. Subdivisions of the genus Ptilinopiis (Aves, Columbae). Bull. Brit. Mus. (Nat. Hist.), 2(8): 265-284. Chaney, R. W. 1947. Tertiary centers and mi- gration routes. Ecol. Monogr., 17: 141-148. CoNovER, B. AND W. H. Phelps. 1946. La dis- tribucion geografica de las subespecies de la Pava de Monte Penelope granti. Bol. Soc. Venezol. Cienc. Nat., 10: 321-325. Delacour, J. 1951. The pheasants of the world. Country Life, Ltd., London. DuGAND, A. 1952. Algunas aves del Rio Apa- poris. Lozania, No. 4: 1-12. Eastman, C. R. 1900. New fossil bird and fish remains from the Middle Eocene of Wyoming. Geol. Mag., Ser. 4, 7: 54-58. Forbes, W. A. 1882. On the convoluted trachea of two species of manucode ( Manucodia atra and Phontjgama goiildi); with remarks on similar structures in other birds. Proc. Zool. Soc. London, 1882: .347-353. Friedmann, H. AND F. D. Smith, Jr. 1950. A contribution to the ornithology of northeastern Venezuela. Proc. U. S. Nat. Mus., 100: 411- 538. Garrod, a. H. 1879. On the conformation of tlie thoracic extremity of tlie trachea in the class Aves. Part I. The Gallinae. Proc. Zool. Soc. London, 1879: 354-380. Gray, A. P. 1958. Bird hybrids. A check-list with bibliography. Commonwealth Agric. Bureau, Farham Royal, Bucks, England. Greenway, J. C, Jr. 1958. Extinct and vanish- ing birds of the world. Spec. Publ. No. 13. Amer. Comm. Intern. Wild Life Protection. New York, N. Y. GuLLioN, G. W. 1951. The frontal shield of the American coot. Wilson Bull., 63: 157-166. Hall, B. P. 1963. The francolins, a study in speciation. Bull. Brit. Mus. (Nat. Hist), 10 (2): 105-204. Haverschmidt, F. 1956. Notes on a nest of the Guianan chachalaca. Condor, 58: 29.3-294. Hellmayr, C. E. and B. Conover. 1942. Cata- logue of birds of the Americas. Zool. Ser., Field Mus. Nat. Hist., 13, Part I, Number 1. Publ. No. 514, 636 pp. Herklots, G. a. C. 1961. The birds of Trinidad and Tobago. London. Holman, J. A. 1961. Osteology of living and fossil New World quails (Aves, Gallifonnes ) . Bull. Florida State Mus., 6: 131-233. 26 Mitsciini of Compciralivc Zooloiiij. Vol. 134, \o. 1 Hoi'KiNsoN, K. 1926. lU'Lords ai l)ii(ls Iiicd in capti\it\. II. and Ci. Withcrhy, London. . 1939. Breeding rccord.s: SummaiA \ 111. Avic. Mag., Ser. 5, 4: 162-169. Huxley, T. H. 1868. On the classification .uid distribution of the Alecteroniorphac and Ilet- eromorphae. Proc. Zool. Soc. London, 1868: 294-319. LArnM.\NN', A. 1939. Die Xciyel \ou Para^ua\ . Hand I. \'on .Streeker and .Stlnodcr, Stntt- .Uart. Leopold, A. S. 1959. Wildlife of Me.\ico. Univ. California Press, Berkeley and Los Angeles. Lucas, F. A. 1900. Characters and relations of Gallitiuloidcs, a fossil gallinaeeons bird from the Green River shales of Wyoming. Bull. Mus. Conip. Zool., 36: 79-84. MacGinitie, H. D. 1962. The Kilgore flora: a late Miocene flora from northern Nebraska. Univ. California Pnbl., Geol. Sci., 35: 67-158. Mainahdi. D. 1963. Immunological distances and phylogenetic relationships in birds. Proc. XIII Intern. Ornith. Congr., Ithaca, N. Y., 1962, vol. I: 10.3-114. Mayaud, N. 1950. Integuments et plianeri's, in Traite de Zoologie, P. P. Grasse, ed.. Vol. XV, Oiseaux, .Masson, Paris. \l\vn. E. 1940. Birds collected dining the Whitney South Sea expedition. XIJ. Notes on New Guinea birds. \T. Aiiicr. Mus. Novit., No. 10.56: 1-12. . 196.3. Animal species and cNolution. Har\ard Uni\'ersit\' Press, Canil)ridgc, Nhissa- chusetts. , E. G. Ll.NSLEV, AND R. L. USINGER. 1953. Methods and principles of sy.stematic zoology. McC;raw-Hill Co., New York. Miller, A. H. 1944. An avifauna from the Lower Miocene of South Dakota. Univ. Cali- fornia Publ. Geol., 27: 8.5-99. . 1947. The tropical a\ifauna of tiie up- per Magdalena Vallc\. (Colombia. Auk, 64: 351-381. . 19.52. Supplcmeiitar> data on the tropi- cal avifauna of the arid upper -Magdalena \'al- ley of Colombia. Auk, 69: 450-457. 1963. Seasonal activity and ecology of the a\ifauna of an American ('(juatorial cloud forest. Uni\\ California l^ubl. Zool., 66: 1- 78. Moore, R. T. and D. H. Medina. 1957. The status ol the chachalacas in western Mexico. Condor, 59: 230-234. Olu AHi:s, A. 1962. A\cs de la region sur di' la Sierra de la Macarena, Meta, Colombia. Rev. Acad. Col. Ciene., 11: 305-345. Oliuk:, C. C. 1960. Penelope montagnii en la Argentina ( A\es, Galliformes). Neotropica, 6: 58-59. Peters, J. L. 1934. Check-list of the birds of the world. Vol. II. Harvard University Press, Cambridge, Massachusetts. PiiELi's, W. H. AND W. H. Phelps, Jr. 1962. Two new subspecies of birds from Venezuela, the nifous phase of Pauxi paiixi and other notes. Proc. Biol. Soc. Washington, 75: 199- 204. Przibram, H. 1910. Experimental-Zoologie. 3. Phylogenesc inklusive Hereditiit. Kap. IV (h) Artbastarde der Vogel (Aves). Franz Deu- ticke, Leipzig and Vienna, pp. 79-89. RuKiWAY, R. AND H. Friedmann. 1946. Birds of North and Middle America. Bull. U. S. Nat. Mus., 50(10): 484 pp. RiDOUTT, C. A. 1939. Aves peruanas que van desapareciendo. Bol. Mus. Hist. Nat. "Javier Prado," Lima, 3(10): 91-95. Salvin, O. and F. D. Godman. 1901. Biologia Centrali-Americana. Aves. Vol. 3, 519 pp. Taylor and Francis, London. Sch.\fer, E. 1953. Estudio bio-ecologico com- parativo sobre algunos Cracidae del Norte y Centro de \'enezuela. Bol. Soc. Venezol. Ciene. Nat., 15: 30-63. . 1954. Der Vogel mit dem Stein auf dcni Kopf. Kosmos, 50: 9-13, 118-124. ScL.\TER, P. L. AND O. Salvin. 1870. Syuopsis of the Cracidae. Proc. Zool. Soc. London, 1870: .504-544. SiHLEV, C. C;. 1960. The clectrophoretic patterns of avian egg-white proteins as taxonomic char- acters. Ibis, 102: 215-284. Skutch, a. F. 1963. Habits of the Chestnut- Winged Chachalaca. Wilson Bull., 75: 262- 269. SiHESEiMANN, E. 1954. Ausgestorbcne, Auster- bende Vogelarten, vertreten im Zoologischen Museum Berlin. Mitt. Zool. Mus. Berlin, 30 ( 1 ) : 38-53. SuMox, G. M. 1955. CIreat Curas.sow. Wilson Bidl., 67: 75-77. AND O. S. Pettincill. 1942. Birds of the Gomez Farias region, south-western Ta- maulipas. Auk, 59: 1-34. Tac:/.anowskl L. 1877. Supplement a la liste des oiseaux rcciicillis an nord du Perou occi- dental par MM. Jelski et Stolzmann. Proc. Zool. Soc. I.oikIciii, 1877: 744-754. Relations of Cracid Genera • VitiUcumicr 27 Taibel, a. M. 1950. Gencsi della specie Crax viridirosiris Sclater, alia luce della sperimeii- tazione ibridologica. Boll. Zool., Xapoli, 17 (Suppl.): 543-547. . 1961a. Espeiimenti ibridologici tra .spe- cie di generi distinti: Mitu e Crax. Nota 1': Ibridi di prima generazione. Arch. Zool. (Ital.), Xapoli, 46: 181-226. . 1961b. Esperimenti ibridologici tra spe- cie di generi distinti: Mitu e Crax. Nota 2'': Ibridi di seconda generazione i ibridi di rein- crocio. Arch. Zoo!. (Ital.), Xapoli, 46: 291- 324. ToRDOFF, H. B. AND ]. R. Macdonald. 1957. A new bird (family Cracidae ) from the early Oligocene of South Dakota. Auk, 74: 174- 184. Verheyex, R. 1957. Contribution a Tanatomie et a la systematique des Gallifomies. Bull. Inst. Sci. Nat. Belgique, 32(42): 1-24. Wagner, H. O. 1953. Die Hockohuhner der Sierra .\[adrc de Chiapas, Mexiko. Veroff. Uebersee Mus. Bremen, Ser. A, 2: 10.5-128. Wetmore, A. 1923. Avian fossils from the Mio- cene and Pliocene of Nebraska. Bull. Amer. Mus. Nat. Hist., 48: 483-507. . 1926. Observations on the birds of Ar- gentina, Paraguay, Uruguay and Chile. Bull. U. S. Nat. Mus., 133: 1-448. . 1933. A fossil gallinaceous bird from the Lower Miocene of Nebraska. Condor, 35: 64-65. . 1951. Recent additions to our knowl- edge of prehistoric birds, 1933-1949. Proc. X Intern. Ornithol. Congr., Uppsala, 1950: 51- 74. . 1956. A fossil guan from the Oligocene of South Dakota. Condor, 58: 234. AND W. H. Phelps. 1943. Description of a third form of curassow of the genus Paiixi. J. Washington Acad. Sci., 33: 142-146. Wing, L. 1946. Drumming flights in the Blue Grouse and courtship characters of the Tetra- onidae. Condor, 48: 154-157. (Received 9 October 1964.) SulLatln OF THE Museum of Comparative Zoology A/l -C, -*W£^W^ I MUS. CO MP. JuN 1719 The Distribution of the Oceanic Fish Brama brama by GILES W. MEAD and RICHARD L HAEDRICH Museum of Comparative Zoology, Harvard University; and Woods Hole Oceanographic Institution HARVARD UNIVERSITY VOLUME 134, NUMBER 2 CAMBRIDGE, MASSACHUSETTS, U.S.A. 17 JUNE 1965 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- MEMoms 1864-1938 JoHNSONiA, Department of Mollusks, 1941- OccAsioNAL Papers on Mollusks, 1945- Other Publications. Bigelow, H. B. andW. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50. Bnies, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Lyman, C. P. and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 15. Whittington, H. B. and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948 (complete sets only). Pubhcations of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. MUS. COMP. LIBRARY JUN1719( HARVARC UNIVERSn THE DISTRIBUTION OF THE OCEANIC FISH BRAMA BRAMA GILES W. MEAD AND RICHARD L. HAEDRICH INTRODUCTION This study is chiefly an analysis of the dis- trihution and seasonal migration of Brama hrama and, to a lesser extent, of B. japonica. These are relatively large mesopelagic fishes of temperate and subartic waters, the hab- its of which are largely unknown except for the published and original data summarized here. The genus also includes several tropi- cal species, which will not be considered. Brama as a whole will be treated in a later paper together with Taractes, Taracticlithys, Einnc have chosen to use the curves closest to one another in temperature, identifying each, and hoping that the discrepancy so produced will not be too destructive to our description. Although listed as a commercial species throughout the eastern North Atlantic, the magnitude of the catch of Brama reaches significant proportions only off the north- western coasts of Spain and northern Por- tugal (see above). A study of conditions off northwestern Spain has suggested that fish concentration is related to temperature, although both the physical and the biologi- cal data are inadequate. The temperature-controlled distribution of Brama may be as follows: The adult fish seem to prefer water with temperatures higher than 55°F (12.8°C), seek water of such temperature at depths from the sur- face to about 500 meters, and cannot lixe indefinitely in waters below 50°F (10°C). (The depths of capture reported for the southern catches are in general much deeper than those to the north: the "tropical sub- mergence" of authors and an expected phe- nomenon.) A part of the oceanic Brama population will expand into the North At- lantic as seasonal \\'arming of the upper layers pennits. But during the coldest months (in terms of sea-surface tempera- ture) the range will be restricted, if not sharply limited, by a "thermal curtain" sep- arating water colder than co. 55°F ( 12.8° C ) to the north from the warmer waters to the south. The greatest catch of Brama brama is 200 180 - 160 - 140 - 120 - 100 80 60 - 40 20 1 1 1 1 1 ■ 1 Brama brama /\ 1 1 1 / r TOTAL ■ K LIVE CAPTURE / ''^"^l - : / \ \ STRANDING^ ! A^ \ 1 X t 1 ^ \ ' \ .■4£i^--i r" 1 1 1 1 1 1 ""r - ^ CO 1— Q_ h- > ( ) 7^ □D ce < cc Z3 UJ C) o UJ < UJ LL < O) o z o -o Li- S - notes from \Miitby, May 1894 to May 1896. Naturalist, No. 253: 233-239. Stevenson, J. A. 1926. A blue shark and a Ray's bream at Filey. Naturalist, No. 828 (602 C.S.): 26. Thompson, D'Aucy Wentworth. 1918. The scarcer fishes of the Aberdeen market — Part 3. Scottish Naturalist, No. 75: 59-68. Thompson, Willlj^jni. 1856. The natural history of Ireland. Vol. 4. London, 516 p. TuRTON, William. 1807. British Fauna, c-ontain- ing a compendium of the zoology of the Brit- ish Islands: arranged according to the Linnean s\:steni. \'ol. 1, including the classes Mam- malia, Birds, Amphibia, Fishes, and Worms. Swansea, 230 p. \'er\\i:v. J. 1953. Annual report of the zoologi- cal station of the Netlierlands zoological so- ciety for the year 1952. Arch. Neerl. Zool., 10(3): 343-354. . 1956. Annual report of the zoological station of the Netherlands zoological society for the year 1955. Arch. Neerl. Zool., 12( 1): 89-104. . 1958. Annual report of the zoological station of the Netherlands zoological society for the year 1956. Arch. Neerl. Zool., 12(4): 537-550. . 1960a. Annual report of the zoological station of the Netherlands zoological society for the year 1957. Arch. Neerl. Zool., 13(4): 540-555. . 1960b. Annual report of the zoological station of the Netherlands zoological society for the year 1958. Arch. Neerl. Zool., 13(4): 556-571. Went, Arthur E. J. 1958. Ray's bream, Brama rail Bloch, from Irish waters. Irish Natural. Jour., 12: 246. . 1962. Rare fishes taken in Irish waters in 1961. Irish Natural. Jour., 14: 3.3-35. Willgohs, Johan F. 1954. On Brama rail (Bloch) in Norwegian waters. Univ. Bergen Arbok, 1954(6): 1-9. (Received 29 April 1964.) Brama • Mead and Haedrich 45 c "o _y JZ ^ t 2 rt Mh C 7^ *C rt > C3 r3 .0 cS < 0 u u '5 ^ c ^ 0 0; ■5 OJ ^ .3 vT 3 1^ tt QJ E c z 0 ^ Ui -£ _; 0 a c c 0 u 2 ^ j^ Q i~ ^ z ;- CO r^ 0 ^ ■w jj M c Z 0 0 0 ffi 'c ^ .„ c r> .^ 73 c 0 « u 0 a a C/2 "£ c 0; "« c 2 3j CS' ^ 0 c 1 -J s c 1 c>5 6 0 H ll 0 tj tfl v: ^ c ^ •^ 0 Oj c C ,;i a OJ 0 0 u 0 . _ I < u 05 ^ 5 s ^ 10 tc _ r*! c 0 05 X 3 0 a en w i-T CD 0^ 0 a ■J 1; 0 X -a C3~ C C5 _c rv c H ai C5 ^ -? c 0 05 >, •< 3 6 a] < i-H 1— ' fc rj ^ 0 0) 0 0^ rt 6 a 1; -^ 2 Tt^ C v: rt 5: u c 0 X 0 w 0 m > £ C ^ g —J ^•- ^' .3 z ^ ''^ -i y: 'G '5 Oj 3 ): c C 1- B 0, s 5 rt _o u C u b ^ § J 2 ^ ■ ^-^ C , ■^^ 3 — ^J> 13 < 0 c 5 .;^ 0 z H z c _c 5 ^ ■^ c ^ S r^ < ,-1 C 0 1- c C s a; H 0 X "rt ' \^ 0 i X a ;j IJ rt z < y 0 c c75 0 a v5 ^73 SL| ^ -c C c Cj Jj ^0 0 c3 ^ CS u « Q^ < >> w y >> • tf Li a < 5 i- H ^ -2 < U ScU 00 a> ci c^ CO C ^ 05 TT 30 10 c C CO CO 01 00 CD Ol CO CD CD CD CO in CO CM CD CQ CD t- 10 d CD -H oi in CD t-I cd' CO Ol -^ oi -^ 00 CD ;/; 5 CQ 0^ 05 01 10 O -^ in ^ d Ol CO 00 d 0> 05 CO oi T}5 oi CD in ai d g m o .j5 ■^ c — — ^ — 0; CO -^ -H CI CD • S cc = -^ -= ^-i 4-» 4-t '"' c3 u u 46 Bulletin Museum of Comparative Zoology, Vol. 134, No. 2 ;-H )-; o lo "T3 ^ ^ pH aj P r^ X "C 'S 'S cs ;5 ^ o y: c/2 Z ■< c O _E as 0) O >. ►7 <-4-l ^ -n + . 01 ^ CT> in 05 ^ ?3 oc i—H S ^" >• ^- a 10 ;^ ■,^ ^ ^ ^ c 10 0 — 1 ^ rt 05 ^ ;r Uh 1; ;/:) '3 £ ^ n > o IS ^ -^ ^ I 0=5 ^ « = -^ - ■" -^ g, a 3 2. o 5 :i^ o ~ t< ^ ^di <; c/: cr; |ji + O *-' + + 3 J ? rH ■—1 6" 2 "T tP Tf in ^^ in 05 -^ 0 in ,-( TjH tP . . "^ 10 05 . in in 00 ■• o> 7^ ^ ^ a> CD 05 in 0 > - "^ 05 ■^ 1 m 1^ 9 ' rw 5 1^ -M S 3 _M ii; c ;i; 05 hJ c o c a ai cfia iXa uiosat^ ^^ ^ (X 1) dJ 0; -u C Q C/5 ^ U ^ O J in 05 ^ 30 O CO -H ^ in tin S ■—' 1-1 'S :■- -G « £ J 2 ,^ 1- J-; in — ' in 0 in 0 oi 01 CO in in in in CO c^ in in 05 in 00 in r— ;B, ^ u 55 ■^ r; -O J3 • ^ c bC c w A- T D 03 3 0 c« i: « u. fe ^ U ^ U ^ s. ^ ^ 0 J 0 -^ W uT > M j= i3j 0) 1) !1) t* < E^ s — -1 "2 0 c^ s 0 < J 0; 0 < ' ' ^ c«? 0 dj P3 3 c 3 ^ Z 0 A _c a ■ v-" 0 05 in 0 C/2 J aj CO in £ in -i^ in -H 05 O CD 00 in in Tf m in 1-c aj aj aj OJ ^ z w w w z ^ I ^ ^ ■* 05 -H Brama • Mead and Haedrich 47 ? + c S CO a ± CD ~ CM 2 05 ^ c CC -J s r. G S o 0; O -13 (^ r^ 1 1 1 1 .-r TT in r^ or ^ ^ 0^ in in cc CO •■^ n^ ^ 3 o^ <^ 4J >, >, >^ >. Oj i> OJ 1) ••s- :S ^ M b OJ 11 V OJ ^ > > > > ^ in oc o r- .— 1 a> oc o^ "-H CI r. C ,-^ c q^ Oj r^ Ci. U c K « C/3 C/D ^ -, ^ H -:= 3^ ^ -- "-" rs £ aJ C ^ ^. ^ o W X 1 . ;2 .^ -J. U C S S S X X ^ J^ ^ 5 W „- _' _' 5 ^' 9 "S ^ Z • • c; > _ ^ ^ W W "S i ^ OJ tH <1> K 1* ^ C X -c S "3 x( s X -^ liajiiaji^iiiii;!; HcHHHgHcc «p in .— (incOr^Om (MOCCC'CD inininioCD'^ T} ^•~ 05 Z + OT ^ Ij CO i-H -H .' ^ .. ^ ■ .. 05 -• ^ CO m ■* in 05 in 05 !:^ --i r-H + .. !Lj vr J- OJ r- X t£ r- -£ tj: "z: oi -= .= *i -t; *; 2 K • 3 ^ S '^ >^ z > CO -f. z r - o -3 CD 3 ? C C ^ -D in i^ S 'm< cc a rt a ec 1 I Qj ^ c z 3 5 *j O r- Z .5 Si S T^ >i -^ 1^ t: fa o J2 o rC C i tiC U S U i^ !y5 c>5 H " ^ St 3 fa cj 3 '^ CO o o in Tf in fM in in in ^ CO in in in in ^ *o *o Ot Cl ^H ^ r-H 1— < .—1 ^^ CD I> CO CO in in in in CO CD in m CO in 48 Bulletin yfuscum of Cotnparativc Zoolofiy, Vol. 134, No. 2 jj- c o M ^c J2 rt .ti J o X c rt A -i; jj y ^ X K a CO 5 f CO CO < + 1 o in CD 1 O] in ^ Tf Tf c/: ^ -* • • CT^ c~ l^ S c CO ^ in s ^ S in -^' •• G, in •"^ r^ m .— 1 ^ o? in 00 ^ 05 c> ^ CO ^ p cr> in ^ t— 1 *"• u in '""' cc '^ ^ r-H o; cc r. •^ o^ - in < 3; C o c c c a a of - - s 2 li 'o 1 u £i n ^ > W W H > 'C c u z a; ■< u h --^ c M CV. ^ < ;:i rt o "^ ■^ _^ g Qj ^ ^ rj «-G U ^ % tc 2- 2- S- ^ c — _• »^ U H U U U u ^ ^ ^ c T; "^ V5 ^ c >. aj •^ >. 03 O « 1 ■^ Z u
  • • r- 0/ c u Z c/: 0; ^ '"-' -:^ ^ ^ ^ rt c ~s. X ;- m m i-H o % (2; in ^ in Tf " ^ G S s Hi? u ** il ^ O Tf CO , — t C l-J in in -t in Tf D i; *o o+ CH- <"><" , . ■ u d =" ii o CO CO ^ o> 05 lO CD O CO u in m in in o o 1—1 in in in )m^ HH HH i^^ l-H I^H ^H >_H HH H-l Q HH HH >-H •-H »-H HH 1— 1 t— t hH h— I CO O) in O) c fM O) Oh < ^ U '% g >^ C Z o OJ 1—1 C3 O Z > o Z M ^ D 3 >. >. "X; m ^ il o Z >> c 1 c l-H o C C 5 C3 C < o U c c o c < .c > u J "c C CO CI o o Ji % ji ^" "S 'I a Z c o -* :> U C G •j^ r^ — f in yn u '"' ^^ ^ '"' i-H '"' ^ ■^ '"' •"^ '-' o CD t-H ^ C c Gi r~ii ^ on in l/J in in in irj in o ^ in -^ c H s H K S j:^ s 5 H 5 5 Brama • Mead and Haedrich 49 c .a S o CS c c/2 N > ^ — lO CO ^ ci ■• 00 '-* c c u 0 3 ii ^ = ^ « I' '3 ^ i" "C ? '^ 3 *^ C H Z Q ^ CO in 00 00 > > > > CO CO Ol + »C c« c Oi o c i-M 00 ^" of CD CD 05 O! Ol 05 ^' g jj- (— 1 J2 r-" c ^ O 3 o C^ P P ^ O ^ rS j ;-! j c 3 c a ■^ rt c3 C-l Oj n CD o; 03 c3 C5 c3 C/3 05 « 13 + in 00 6i in ^ CT> CD --^ ^ . ^ J, ' 3 05 '^l 3 = 3 ;$; "^^ lo ° 5 "^ ^" - 3 c^ .^' r^ in rt ig 00 00 ^ i; > t^ --ell « 2 u c5 c 3 C3 Ol 'k CD O >3 r^ o Oj J^ p 1 , 1 ■ ■ 0 /^ _•'— ; ;3 -3 ^ •p Oj tc M t£ tt tJG M 1-3 s 3 3 3 3 3 /^ Si; coo 5 ii ^ g c « O o o o o 'S tc a; > > > ^ m c/^ < < < j^ a; • c/3 c/:) en . t£ , , C/3 C/^ en W > .. ^ en i; in c o . ° " O CD ^'V ^ in O CI in in rt ^_ oj VD t ^ 3 '? . Q^ 3 ?:; >- i: 5i "^ J- ^ 1; C -J 3 3m o+ o ■— - — < ^ ^ ^ ^ ^ oi 1—1 ^ ^ 05 m ■^ Tf in CD l> It- o) oi CM o^ oi in CO CD 00 00 00 00 )— 1 >— 1 p— 1 1— ( ►— 1 > > >>>>>> > > > in 00 CD CO CO CD 00 00 ^ O] CM ^ Ol -H -<*< ;r; CO Tf • • in in r;^ 2 2 S S i^ in X '^ o-j •-* s 3 in .. - - - 05 CM - c« c« 1— ( CM 3 3 .0> ^ rt C i-H -£ -£ 3 ' ^o ^o 'J3 o h* t^ l<^ ^ :/^ wT ^ {- Ij OJ C 3 c3 c3 Ji a) CQ P3 IX cn rt 3 3 3 3 3 ti O O O O c5 H J U U J U 03 fe T3 fe 3 '~Z J- ^' rt C e; Tj t« rt o o o "5^ _rt cc k^ o "oj ^ fe d 1) o 0^ -3 3 ^ ^ 3 3 3 en ffl i O "o ^ 3j ^ S ^ 'I ij en lb V en 1^ en £ oi > >>>>>> "« 50 Btillctin Museum of Comparative Zoolofiy. Vol. 134, No. 2 u u :^ oa cQ p= 'X, lO •■' ^ a^ OJ CC + 05 .#v ^ lO CO J2 ?y^ 00 p r^ n a> 1— < lO »n in n> 05 05 V! .— I + a-' ^' ?: .._," ^ .lI :S tU :^ ". ^^ »^ > 22 o X X •-H •• ^H en F;; ^ Tj< iS 05 : 00 05 X^ ^ i-{ ' — ' s .- t- _' *i 00 CO ^j -^ _W -^ r— ' Qj .— 1 T) jj- «N C c c c ul^ CS rt o 0; i; « i; tH r3 n Ui tf « hJ « = ;: C C p "lis , ^ 1) , . —. V a ;i a r-] _« 05 cb 3^ a.* CO :1h OS !X ^ u u "zt is ii z. ^ — c ^ ^ -t; -^ c j^ C c ^^ »-^ 1^-1 •— 1 ^.^ 4_. 4.J •t-' 4_t 4_» 4_> •p. " ^ " ^ tt y. iJ, y tJ tJ U u u u U CJ u U u u u u 1, (M O ^ M t£ tt > St " 3 3 5 § C = U U J H J U PS 1 2 J h4 J H H J 0 •£ x .'^ ^ P3 Jj 3= ^ 73 « ^ ;zr 0; A3 ;= i: a 3 3 ^ ^ 3 1^ 3 • N.N. N.N. N.W S.E. f N.\ CO > ^ f . c/D ib - ~ ~ . a:; . i^ 00 ^ oi o] -^ oi -^ lo Ui -,■ c/^ ^ o ' s .,■ X S ffl ;^ m/^ 73 '^ 1; 3 ^ >> lb 2 m f>i J ^ ^ ffl :=( J « ^ 3 O o c J a: in in i6 i6 in o\ lO in CD ID in '-O -o o+ ,-H Ol O) oi Ol I^ X 0> Dl Ol .-fC t^ c) CI -^ lo in ■' in ^i^oooincoo-^inTfcn^cor-i cot>t^05in co Ol fM Ol ^ ^ ^ CO I— I i-H fM CO (N > 00 t-- — ■£ tc Ph >■ - — 3 >v ^ tt c« ^ M ^ t: 5 o< 3 3 « J5 1^ E o ^ 35 Q CD U C J a uckl ooe, ewly c o X Z 2 J Z —I CD oi 00 05 00 CO Brama • Mead and Haedrich 51 a OS ;j U J 1 sJ C 1 1 1 1 1 1 '^ IJ ^ Ji ^ ^ ^ 1 ^ ij >; x X >; >. >, >. >. >. s 23 M fC 3S 22 rr* as S3 b >• >> >• >• >. — ' >. >, _^ ^ _^ _^ _^ _^ _^ ^ _^ it ^ _c a, w Z 01 35 IC t^ CO O Ol ~ — m Ol Cl l^ CC IC l> CC mmccmvCTfio-^iOTriO'^io CO ic "^ ic -^ o+ CH o o "-O •^ •-0 ^ -H in CO o ;o CD CO CO CO CO ^ CO S K S 2 t:^ s S s a s 1:5 > > > > > > > i4 B PQ W ^ - c - - ^ m m -^ -/■ u tl ti tl t/ ii 1) 1; Oj - - s - ^ is ^ ■5 t: ri; ^ rs '^ 11 ;-< ^ rt ^ JS J2 ^ oc53ooocc5o5o UUUUUUUUUUUUU 5 bO &L^ --5 fe J- C J. C ^ C Ti rT >. r-^ I— -; >* ;= c ;= i; c 0; *: oj -/ -C C > -^ Jj " K < ^ ^ ^ ^Tj '^4h "^^ _ _•■ c . _r CJ J^ ^ 1^ cr; > CC :/: > T r o :o !- :c; ►^ - — J£ O J2 s c u o y^ CO s> M as i-i o CO I CO CO X 30 ao in in in CO Ol Ol -H I— I Tt< CO ro CO C^ CO I> 1-- l^ 01 (M (M C-^l Ol inmininminminminiomm (MX-^xo-HOTfiin-^io^oc 52 Bulletin Muscinu of Comparative Zoolofiy. Vol. 134, No. 2 CO CO in in 2 o, 2 -^ CO + + + 'S ^ a; >-- 5 + .^ in + ic s; — T}< "5 ^ -H ■- a> 2 2 "' " "" 5 ^ '-^ "" 3 ;.^ > ^ 1.^ CS y; c ^ n 05 bi cc CQ 3^335 5 c S £ ■" en P3 1- "^ « , x ^ —1 0? 3C E3 '^ it •• X 0^ -— > _ ^ CS ■■^ p-H IC 00 Ol 1-H h— , ^ in IC 10 u 2 c>i C^ ^', 05 o> C^ rM CO I— < i-H I-H ^ ■^ • 3 a^ Ot' •*S ^ ^ ^ ^ ■*-' 4-» z C c S 3 T)< 0 £ ^- c ^ ^ ^ X 3 in • ^ r^ J; ^ ^^ TZ^ 0 c:^ 0 3 C3 s "^ S ^ 1 2 '^ ^ ■^ 0 ^ ^ 3 3 c ^ 0 3 3 c -' 5 c S u ■.o CD 1-1 D 0; of CO S 1^ C/2 ^ cc; 1) a c8 05 CC 05 i ^« •-; ^ 01 in rr: 05 »— < ~ ^ , g *i 3 3 0 c ^ c ^M 3 ri ^ ;L7^ ■-- cr* ?i J c« -4 (1) 3 3 a 3 c: « 0 05 ^ 0^ aj ^ ^ s I' ;=■ "3 (JiJU « ffl a w CO 'y;" cC -j^ •^ J- oi CO cc CO in Tf 05 c^ CO in CO in "> ^ CO 0 *o CH- 0 -o 0+ r-, -H ^ — '— ^ — ' ^ OI CC '"' ""^ ,— 1 ^ Ol ^ 1— 1 l-H Tf CO CO -^ \n in \n < j;;1 j< X >< >< j;< X >< >< X '^ >1 X X X in in CO CO CI ^ CO 00 fM i-H 0 CO Tf- CO O] -H CO oi 01 ^ 01 CO — I 01 '^ ^ 03 bo ^ 0) ^ y J; i; 0 r— 3 t U4 ^ en ^ r- 3 i' Oj 03 3 -3 2 0 >, y^ x* X K 0 P K «H en '^ e^ en E m m '^■' "^ « 0 CC t£ en CC Z X Z X ii z 13 3 J2 en 0 z' 0^ 0 > 03 3 0 (X 0 1 Z i z 0 b X ?; X l-H u 3 0; W. % S. of N.W. X N. 60' W.S.W. z z z _2 z z v: ^ ^ ^ o lO s '^ Tf u c c rv .> £ £ ^ ^ s r- H ^ o a5 CD ^ o C2 2 05 £ cc Tf C^ '-A •4->' '^ Ol .. TfH V C 3 01 .. CM CO a 2 S Tf in „ ^ 'oj X r>N 05 oi Ol -^ 0) ^x in oS "— ' cc CD J ,-C ^ o^ ^ „ 05 cn ^ o s c a: J "5 o U IJ c c c "c tc 1/ a a :^ j: aj (Ti ul K o Mh T3 d w >-l O u- 1 "S -^ 3 U./ m >; -5 X :^ c aj '■A "^ p^ i^ ^ o V) o 2" 3 ^ "o i:; Q u :^ X X X X z X in m Z Z oc o> in TfH Z CO in 3 2 (M z Z z Z z u s a a t^ Ui 14 £ § 5 S U H in U! £ ^ - in £ en CM ro 05 35 ^ cS tc u o o> 1 9 y X «>1 •- ii tJ: -w ^ *-' in — 1 xH ^ -5 -5 ^ ci Q 3^ CO C5 lO 3^ 3> 0) C ^3 C > n c aj S^ S S 52 " £ £ Bf 5f ^ ^ U U 0 U H H H U ^ ?^ C/3 PC ^ jj "c Z -i 3j K^ [Jh ><^ -^ 2 in rt ol in 'x i > Z > ^ 1j z CO 1 [JH " Z 1 "o Ol m z o £ ^ ^ £ > ^ •ti 1; a in Z Si 1 i I ^ o ^ *^ 3 30 J2 s > ■— 1 r: [il -H ^H in < U .£ Z ~ K lis C C '^ Ol in oc in X in in CO in m -H Ol CD m Ol in in in Ol Ol I- Ol Z Z Z Z Z Z Z Z Z Z Z Z Z CO ol CO ^ in -H a> t-H t-H 00 54 BiiUctiii Miiscuiu of CoDiixirativc J.oolofUj. Vol. 134, No. 2 fe in ^ cq ^ M ■<* c in 00 I— i^ sc o^ '^ ^^ _c ^ ^ 7t lf^ Iff vT C-l in 00 in m c; c^ lO ;>; "i; 3^ ^ Ol _' r-'~ 5 5 S ^ iq ^ oc Wilsoi '-^ ' ^ -i o '— ' ' ir i) c" r-3 -^ Zj v. r^ rv ~ 1; 1.' Jl ii. ■^ — . /V> X x y; cc X + + ,-1 CO in in o^ C^ C5 ill Q 5 o o c o o c o x a! OS 5 in H Q ^ 3 ^ ^ iC ^ tu - u _^ t Q 2 S C rt r; ^ — ;^ 0) IJ Q ~ Q M J^ r- r- ^, ; M O •£ •£ ^ Z Si Z -o :/) £ _!; V: JS" „' c ^ _j 3^ ~ H ~ 0/ -J < 3 Z X fe fc Z ._; rT W W f^ ^- ^ ^ r z a a = ° a z z x; b o -s ^ CC CD CC O 1 1 1 1 in t- 00 in in 00 in in a> c> o> CT in in in in -* 00 X Ol 1-- 00 m in in in in in 0+ 'O CH o o+ l^ l^ I- I- C? X 01 f>l 01 01 CM 01 CO t- t-- X CO t^ i^ i^ i^ ^ Cl Cl 01 CI c] Cl c^ CO CO in in m fM Cl 01 Cl in m in in in in in in > 0; 1; 0) !- CS C3 rt 5 X QC Ph O -^ '^- ^ Cl "^ ^ '"^ c c c 00 -- Cw '-^ J J U C J H Q r^ -rr '-a ^ '> ni ^ CO ^ ^ ?: X z z > "'. &H = Z CD 2 ^-^^ ^ *s Z •— I Cl m XXXXXX XXXZXXXXXXXXXXXX^^ X XXX XXX CO (X ■* c p p 3 ;i CO 05 S2 ID a r—< a a a -C CO in CO in 00 o5 ^ - ,c in cr> 0^ in 00 lO o Brama • A/e«c/ fl/ifi HaedricJi 55 G C C C C " o o o o c • • « G C G =" OJ 15 i s « ^ ^1 J hJ J ;^ C G C C C OJ c. . C C . ^^ K 0; Qj -'^ u 1 1 '^ o :^ A ^ 2 -s -s ^ ^ u c^ i: « ^ 2 i3 Oj ;-< S ■ ^ •5 O fa C/^) fa X c^ 'o o c M-l c lO in fa fa Z o OS Z tii aj a d fa '2. > in in O 'IJ fa fa •S .s u fa aj c^ fa o fa c fa CO fa 5 s 1 a; G C fa _ M ■ ^ ^ -Si c75 Si 'I '^ 0; fa o a o G G ^ 2 "^ ^' fa fa X ^ ^ ^~0 CO fa C c cn 3i fa o m X o v5 m Z M-l o c/^ <-lH O X qa >> O m 6 G in L^ fa' ^ E X fa C/2 - fa fa in o fa c fa 0; OJ aj if- fa' in fa' CO z z CO Z Z z C/} 1 ^ in G ^ ty3 ^ ^ CO ^ C » ^ .^ cb 00 CD in C TP ^ < ^ c ^ o i-H in Q in cc ^ .—1 CM So 00 fa cc in oi CD l> CO '— ' tT ^ CD I- ^ CD CH O i-H (M ^ ^ r-H r^ fM i> in t^ 00 i^ CD b- in in in in in in oi X x; x X x: X X lf> -rf — ( CO -^ CD I-H ir)inCDl>l>0000O^C^'— '^^'^1^"^CDCDC>C50> CD Ol OI (M CI CM fM (M OI Ol CO CO CO CO CO CO CO ^ ^ ■* x; xxxxxxxxxxxxxxxxx^ix in ■-Ha>^o^o>cocococoinc>coin^cDO 56 Bulletin Mu.scinn of Comparative Zoology. Vol. 134, No. 2 c X! « Z Q ^ 14 CO C5 3 1/ O o _>> -t; "r i ^ j; HH-J0-(C^Hi>oHHJi-Jh-l uj X -2 Ch a 2 ° z i- ii 1 ■^ 3 a l^ « z ^'^ w Z^ in z z i So ih 2 <: ^ ^ aM r- w' '5^ -^ '"' "^ ■ ?5 -^ -^ J o — ^ O dj G5 — ' •^ CO :/o TT -S 00 C/2 -^ ?£in ^ 1, ^• !^ C J' -i r;T z ^ 3 « .^ ^. > ^ ,^ ^ ^ ^ ^ o c [— io Z c-i in f>i oi CO d 00 •-0 Ot 0+ O CH- OI —< -H ,-H ^ r-l ,—1 Ol ■>^r}<;c.:S':ccoooccinioin Ol -* X) 05 C5 r-H CO 35 I- ^f Ol 01 -H ^H Ol . —I -H -s ?i c; ;:: C l> t-- ^3 C ^ w w ,-z CC in in in ^^- Ol Ol ^ ^ Z Z 00 00 O] Ol U 5^ m i3 •ti u: ^ ti ^ ^ ~ U 1= u u 00 oi ^ 00 ^ CD lO CO CO .. ^' CO -^ tr- ^ =o .. g §8 .. no CO ^ in 05 --I „ ■^00 " oj ^ .03 ^- ?i '£ 9 ~ -^ aj ~ 3 is ^^ 02 U (1. J tL 0; +j •" *- C tX "' 2 3^3 c^ U H U Oj ? o -^1 K IT-I o ^ ^^. >-, f >. , X ' fJ3 ci ^ ^ ■^ ^ § 2 oa "1 pa "ij "g :/5 i-H Oj ^ ;-, C ^ '"' ^ "_^ C o S .« c/i X U yi fc4 g >-l -^ -^ rt CA) O « C OS OJ *i ot ot ^ '=o 1— (■— (i— ii— (i— ii— i-^i— I in CDin-OXO5CD00-^ inTfiTf<-. ^ 0 v: ^~" '—' s 00 tf ^ of CI CD 'u 6 ;5 1 ^ 0^ 'J-. CD 05 00 1 o Cl <>{ 1> r;; 0) ■£ c 1 — 1 2" 05 CD CO 10 2 '^ g 10 30 '.-1 CO 0; 3 3 S 00 fM 1-1 00 CI in CO 00 CD 00 1—1 06 ^ Ol rv of 3d ^ ^ Q^ c « rt r^ ::j ^ CD Oj ^ C3 o s ^ 2^ U cr; U u u J H-! H J ^ ^ r— III I 0 aj 0 rt 1) t-^ c VD H-l CO K 2m y^ m -f- CM 5 "c -S (b CD 00 in Ol X >1 'A i> oi -H ^ 5 »^ »J^ CD -b IC CO X X 01 in CO C> 20 '^ X 00 00 X >^ X ^ ^ •S t ^ •S ^ jj « HH a^ 0; t; ' S ^ -5 4-; 01 t^ t- oc CO X c f^i ^ ■IS 01 0 tt ^^ C IJ J ^ J > -1^ 0 = Z u ^ ^ rt ^ 0 r^' CS rT ^ r-H. 0 0 "ti 'ti ►^ ^ 2- ^ z ^ ^ OS W W b x w X, c r^ g 5 "c ^ ^ c "c ^ 'if. Q c/^ X X 2 <-l-H 0 X 0 w Oj iiT X S^ ^ X ^ ^ > 0 -c w z z z ~ in « CD 06 CO CD m T}* *© 0+ 1— li— ICO"— I^HC0^01.-H^!— ( CD co"^ininincDCDC^i^T-Hoo oi oi oi OI oi oi oi o] Tf 01 ^ XXXXXXXXXXX rr CD oi oi oi Tf in -^ 00 01 ^ 01 01 O] .-H a o U :: 10 cc 1/ 0 ^ ^ '^ G^ o^ tt ^ 2 2 c ^ l.* 0 '^ ^i: ^ N 0 >^. in ,_£ U U -^ ■^ u = Q 0 '~\ hr ii ffl ^ -fl ^ bJ. — I> X X X 58 BiilUiin Mtiscum of Comparative Zoology, Vol 134, No. 2 Q M U u 5 c t D o ^ y ,- o S tt 3 — 1 ,5i C c ^ c y .'-: c u: N ^ s -^ ^ M 4; O /^ a N i: Z a U! :^ -J in w in ,1= ^rl 1 P ■^ 4; CJ tT -o u ^"^ P 4_> -4 c c/:) a w = - J ^^^^^5:5 « = rt rt « cs rt HHUOJo^cJ5hHHHHHHHHc/^ Oj c 0; ■*-» Oi -*H rt !.; rt V-i Qj 2j CT) C/2 cyn c o W O b ^ ti c« O j: rt I- _ o s^ - ,:i: 3 -r: cC _:i Ji ^ -£ .b ZH E 5 05 ^ t^ 'S ^ _ -Ti C/D !" c 5 oT 5 a; C/2 c 3 ► C ro ^ EC tx 3- o ^ r^ f7! -J '^ ,5 3-1 fcr ^ o ^ S' in Z -c < o "^ [ij 2 ^ g ^ W § Z Z oc g o b X ^1 30 in U:: r-i CO O* — I — I en 01 r-l fe t: Z '"^ *: ,j= -f 2 ■« -3 06 CD cn -^ 1 [IH tX CO xt< o in ;c T m in Tt< Tf in rr ■<1< Tj< ^ Tf Z Z Z z Z Z Z Z Z Z Z ^ cn ^ 1-H ^ 01 ^ 00 CO CO cn 01 O] -^ ^ 01 (M ^ ■* ooooooc2a>c>a5C750^ "^ T^ "^ "^ j5 O i> o c in in CO CO in o o c o Z Z Z Z -H ^ ^ ^ 01 oi oi C CO 05 ^ 05 CO Tt< ^ CO -^ rk ^ ^r, ^. ^ i^ 00 t- ^ in o (75 Tf T— 1 05 1— 1 >< ^ U C :$ Oj > C in oi ^- cr > t— I o n ccS i-H n ^ c« in •'' -= O CO Si -^ /C ^ 02 .2^ ii « = =1 ^IIJ i . '^ 5>>J bQ OS'S -^^ S'i^;^£ ^£m ?Sc i-=^J = .^S '^•5i;- = '--i£'p'c'-^fcr =S ,. 5^-5 ?«jij 2b^ ^i^sm'B^**^ 5 C* 5 O CO '~0 c o c; w to c tc S H i ^ ii 2 a ^ o rv a; c/0 1 ? 1 'l^ > •^ _i2 ^ ^•■^ ^ r-i r-"~- ;^' ■i. y^ ^-- < J^ v: r^ Z "P ~ t^ -"^ r" Jt: ^ = C: 30 ^ i i; "^ > O ^ i ,^ ;- Tti t' ;^ -i2 O tf- -^ "S- ^^'Sr~^^'$^'-*r\ ioxcci33com-Hc: en 00 t^ Ol Ol oi fM Ol CM Ol i-H Ol >< X X X X X X X X in CO tP m 00 05 in XXXXXXXXXXXXXX XX XXXX XXX xxxxxxxxxxx ooocO';DGi ^^oico'^cc inco Tpinoo 05 in oioirtb--HOinincci:— oo 60 BtiUctin Mtisciim of Comparative Zoology, Vol. 134, No. 2 A + ? CO CM 05 i-H 00 05 Oi o 1—1 ^'■ Tt< rs ^ O^ 's O Oj C '-J ►^H c ^ "z; ^ -:i« ^ "S -S ^ 0 '-' C c^ c ^^~* Uh ^ V 0; « '.Jh -t; p o (1; r! m -t: 3 — , ^ cd < <; a S "^ -5; Uh J£ c a w X -c K 0 i^ 'a; 3 X X w o 2; c W z c/^ C/l CT) C^ z c= ib Ol ^ 2 ! O « X-^ in i^ -^ »rj i-^ CD in in in in in 10 •-0 "-o o "-o > 52 o ^ en ^ S i; M _, c/^ ? t. iH 2 ^ rl O S " QO C/3 ttH PQ J -H in CO m CO PCi ^ US (U i: 2 * (yD c^ O 1; "->^ r- S in in r-l ^ ^ pC 'X' a; c c 03 ^ 0 ^ •« Oj 1 n 0 < 0 c5 X C w W 0 TtI c 'p z' CO « ^ Q /!^ m ( ; w O) W 0 in I- CAl OT c v; fc rH -S-n-fJ ■^ >> t- 'i 3 c:! o T^ < ^ fc '^ rii 5 rT3 _ rt C a> ti o ^ m rr ^ i33 5 ^ o .i -s td CS X ^ X ^ C - 't; 'C O C 0; J = c :- u O 'y W W Z w 3 oi - in H "i^^—j-icidr'- iinc^inM* ^ :/D < in O 06 O CD .-H CD in CD in CD •^ "^ 0+ <^ Ch- U t- CO Cl CI CI CI CI CI CI C] CI CI CI CI a> ^^ in in in in in in in in m m in in CI 1-H y^ y, >l y. y. >^. y. P, y, y. >< >< P. >< X C^ ^ CO ' 01 Cl CI Cl Cl t Tf i^ i^ t^ t- i> i> t^ 00 a> Tt< "rfi m CD ,_l ^ ^ ^ ^ 01 01 Cl Cl Cl Dl 01 C-l 01 01 Cl CO CO CO CO yyyyyyyyyyyyyyyyypPP. cinrooocococo CI CI CI CO CO in 1-- ^ in c Cl CD o CD in Cl cl Cl CO Cl cl Cl Brama • Mead and Haedrich 61 00 CD 00 2 S ^^ CO CO in in OS OT ,-J > r- c ^ c CM rt ^ ^_» 0; Oj '^ r-] PS o; c75 ■30 05 OS S ^ O Oh .it; c^ 05 at> CO 03 ^ § 2 '^^ ^ c-1 ?^ s i5 i5 S ^ 05 03 J35 ^ S § P c ^^^ CD 00 in of CD 05 in 00 in 03 . J CCCC-rtS- oi CD 03 00 in 05 ^ 5 CD 03 in 00 OJ in OS hJ '^ a i^ '^ Hi CD ^ 05 '^ 00 03 05 9 c ^ 5 ^' p 5 '-' ^ S ^ ^ S CO 3 i-i j^i_-«x;p'-'- ^ i^ y, t- >^ tij •— ^ ■-' He/: (y)HH;/Dl:^HH c/Dc/^.-tlHHcAiHHH 3 ^ U H c/2 H ^ % H H ^ U C a; "^ tj 1) r 1i ~ rs v: r o c;) ^ ^ ,- ^ ^ •— Uh j; r— , < ^:] z y o 05 6 0; ■0; a 'tic v: tx < X a a o CO z _2 a '^ a a > a z s 3 ^ -^ » Ij 00 '— *- w in oa o f • — Cj ;'j i ^ ^ < ~ti c?5 1; .« a a 1 — ' ^ _3 c < CO < Zj ii 3 Z CO i2 a -Tj O ^ 35 u 1) Z O a 0; C5 OJ 3 O a ^ en in a 00 in ■jj < 05 ^ a a 00 1 a z z > 05 C8 < en Z c/^ a z z CO 1 Z a iSa a c in CO 05 ^ m 1) ^ V? ■^ CD ?1 Z 5c z 05 05 5i a, ^ ^3 ■S o O ^ -c t— m o ■*~j CO ■^ *s m in .— 1 in C: in CD ci in in m m in 00 in ro t^ C^ in ^ in 00 in rr in cocccocot^i^ooxinx inm mm mmmco^H^t^ cDt> cococococococomm-H mm mm mmmmmmm if^ lo xxxxxxxxixx — — 1 ^ f>1 m CD X ;^ >^ x; X X X X >< X X CO CD X X X X X X X 62 BtiUctin Museum of Comparative Zoology. Vol 134, No. 2 cc Tf 25 35 G> ;:' X X: '~' IJ 1/ 1; 'x '^ D U S U in CM 1 00 S S 35 CD r ^_r ^ 1— 1 5 1-. tr. >. ^_ p •;=; 3: Oj ^ c a> > C/D (yj ^ ca 3 s c -^ • o ►^ ■£^^Jii-i'£ -^-'^rS-S rS "^ ^ o ti. S E U tt M tt tt c c c c o m c/: x r- »- 01 3 3 o C >^ t/3 co m tr, 'Br C5 o h T O o < -^ ^ ^ y. j^. x j^ x .Tj bC TS •r" f— - w ;-< ^ 3 3 M bC ^ 3: ■iS.S r= c "^ ^ c a '^ U( 1— v: &E ' ^ c5 2 § 1 a; rt o aj ■-^ c-l >> " a; i 1^ Q K C-1 o ^- r-* "Tin c ^ o o 3 C>D .3 4-< ' ^ Z K ^ C3 ^ Oj rt 3 ib rt -^ -C C ^ Ol Tj "5 ^ ^ .^ i- 3 -t: '^ r ^ ^ "^ r r •— K-* ^^ cc* CO C c O CC CD CD I I I ' ■r^ _( ^H ^H rH O lO lO IC lO ic lo CO t^ C5 cc 00 t- I- I- I- I- i^ r-H O ■— ' ^^ ^M <^1 f^' fM ^M '^1 g; Tt" CO in I- I- oi CD CI 1- I- Ol I- CI 1- cl 1^ C) Cl CI Cl x >< >< Pi )< >< >< >< y. Pi 2 t- CI 00 CI TP 05 CD CD 05 CI a> o t^ Tt t^ m m c! Cl i^ C] CI r-H lo ic lo lo in y. yr/ 05 ■— I S o 8x O C C ii 05 TP TJ< C „ ,— I T— I .y >• - - .ti a; C C tH > OJ OJ 0; 5 c c t^ > nil. o <^ 00 ^ O) ^- S 0; CO a .. :- ^ Tf H CD 05 " j_p y;, C5 ^ • I-] C Hi , r* o CO in 00 c ^ 05 5 W ;5! ^ c^' ^ ^ '-' hS ^ O'' K o^ oi . CC C5 I — , ,_| ,— I 05 . . .- O oj oj a .. F £ S 20 Ji. GOG 1^ c/2 U x; X >< o y; ^ _ 1" ;-! -C ^ ■? « c ii c 0; Oj S ri ,s 1) ^ ^ gj ^ 4_i r^ c tl a r^' 2 > ^ O CC ^ ^, < Q r^ cJ: U _c O fl u Q ;m oT OJ C — o c c c s =1J .V U ^ o *^ Q "=^ ^^ aj u s c c m l-U( c75 c C ffl c ?i J 1; C 1) c ^ o "« o I) ?i o C/3 o u ^ C/^ ;^ H U lO p: in ■>* 00 1^ >^ X X X X X X ^ CM oi lo c; CO 05 u U G G ° -? 4-r Ot - - ^ - <<) - io Tf Ol lO X CI Ol Of) ;^ ;.^ or a^ or CO CO r- lO fM 01 o lO Tf _ Ol X ^ c 00 CO 1— 1 iJ en in 05 1—1 s It- t^"" ^ '^ ^^ CD 05 Ol in 05 CO m o> ■5 -'in S:? -i? ^ U! ^^ .^ . c; 5 tt; ij ti P3 ch K CC IX Wl c 0 h- Tt< r- i-H i-H oi c^ r>\ in in 05 05 ^ ^ Tf ^^ .V -x -^ 1) in 05 i^ Z '"' , , tr ^ C 2 c rt ^ tL (U ■J— "-^ tr; Uh ^ .S t£ O _^ -TO c a r- c > >. = ■"^ " II 25 oS§S2S£S-=22;i: GGG;:5c/2c/2c/2c/2c/2E-'E-'I5^ Ul ^ rC 0 "i; ■j_i 0 cn >> m 0 Uh •£ W ^ c/5 0 "c c fc CD Z ,— c •p 3 tr n OJ c -< -t; c 'a; 0 1 a fe ^ 0 c 0 ,0 0 ■^ 0 0 G ^ -^ S or' < H-l m ^ 0 H fc i5 w G C 0 W Z >> ^ ^ 0 OJ rt ri c Z !•) ^ ^ -C 0/ ^ m c/^ ^ G c CO "^ Ol r— ( in 1—1 ^ *^ 4-1 CO C a^ « a Q -5 E ^ K c c o I -t; Z = a -H in c; m —1 in in in in in Ol in 01 01 Ol in in in 01 in in m X X X >< >< X X ><><>< X X >< CO CD -H 01 01 01 ^ CO CD 05 CO 64 Biilhtin Museum of Comparative Zoolof^y, Vol. 134, No. 2 ^^ Q in or or^ c in sO^ *-• iO o ,—1 G^ ^ CI <-H Tf us -^ ^ Q': 1— ♦ "r- o s ■^ ^ f— ( Cw as -n J _c ■"W a< 0( c 5 :r. ^ ; ! 1; ^ « ■w .s = ^ ^ -b' JZ U IS. ^^ o u ^- w u ^2 s, c W u -^ >- X CO OS OS OS ^'■ 5 E £ C OS )^ ? S - 9 o 9 o c = c c c 'i' ::; rs rt rt CS *J Zj li 0.) OJ U "~ s: tf rt « « c^ X y. ii Ji "S -i ^ ^ ^ U ^ -C Jh = fc ■? J= bE K I- ,aj T? "t^ r- u rT" r- ^ ^ __ ^ Oj £ ^ '-^ ^ >■ P 3 _C '^ I' O jy w -^ , ,^ 1^ "^^ ^' ^ K _o o 1/ a < < X w w. fe ^ 0; J- l^ CO c -:= O _^ « OS ^ o c r, z rt xu rH w Cj X H-l o Tf in C; 01 CD oi oi t-^ in in lo -^ 00 '— ' »,^ ^ OS in in ■ ^ 01 o o .[^ OS CO o ^H ^ ,__, ■rrS C o tl. in 1^ "S" TT ^■~ - a •;= o .-H ^ c 5 c c .. T O rt CO Z o ^ OS Z Q a s wi i-i o o H c/2 c/3 ^ -s -^ Q o ^ -^ 02 M 1; ii Z ■= -o '^ aj oj c c S 5 Q -^ -2 bQ I- I- n) 3 .> 5 5 o ^ c/^ -5 Q I z Q o oo tlH Ol in-Hcc I— CI X in X Ol CO X in in in in m in X in y y >< >< Pt y Pi y >< >< y >< 00 y Cl in cc CO 00 in in CD in 1—1 Brama • Mead and Haedrich 65 aj 3; ^^ aj O CC G5 3J CD ^ 2 2 '^ a 2 "^- UJ Ui 3J ^0; 3 I' '2 aiS Q l2 O X CD Q .. of oc CD 10 OS ec .i:3 « « 2 + s u TT CO 1 0; rn a en „ fM 00 CD 05 05 -H j_r »v c c n 2 a cj h-l 0 ^ H rt 0 in 00 0 0 ic 05 i-H 00 ^ 01 C : : ■ z z Ci M U u u u qj Zi ij ij ih Qj U H c/2 i75 c^ CD cJ: ■^ "^ ' ci -^ w Q^ cu 0^ a^ 0 in LD C/^ 1^ f^ C 3 a V 3 >. C •5 K CQ CO TS ^" tt b >. c CQ 3 C CD 0 ^ < < £ j^ = _2 1^ if ^ v: . .J: >^ 1*-. "r^ 3 1' 2 1^ a CQ •■r. z 0 U CQ C/2 X y? -5 0 oc S ffi a 0 iJ5 '•J oi < 3j '-/^ *" ri u it •- V -r: . OJ W a K cP=: z 13 J a a OS 3 X ^ a '-^ -^' 10 • u z J CD 73 .-^ ■5 ib = -^ 3; V c Oj ■ — < •Oj 73 ' w Oj < ^ >. ""^ C rt Oj a <; S --ri Ui ■ X c a z -^ X 5 z 5 5 cH ^ "C E K 't5 ~ = ^' ^ li 00 7 r,^ ii X o c lb a a a ^ CD » l^ . ~ ^ 10 ^ ^ ^ ^ 10 10 tT ffl == .S .£ S 2 3 a^ Xy'S'i;'!;^ St- c/5 Z'^ c/i -."f. -S « M -i W >< £ t != ^ != 1) oj 3^ i; 1/ o i^ H H r- c c CD in I- CC' C: 00 10 Tt< in 00 in TT in in in 01 01 >< Z Z Z z z 35 in ^ CD t^ inCDCDCDOt^t-t^CDXOCCiOiOC 01 in in in in c-i 01 01 c^ CO fc ■* ■^ o 00 0> 05 C O ZZZZZZZZZZZZZZ ZZZZZZZZZZZZ^'tZ -H ^ t- X ^ ^ ^ (M C^OOCCD^-i05C001>05C35a>01 ' fM -H r? 01 .— I (M 01 O] PI 66 Bulletin Museum of C.euipardihc 7a)oIo<^[i, Vol. 134, No. 2 K ■■■« c^" 9 1— I 00 ij ^ 2 2 X 17 S ^ '^ 10 I s i S ^ i" ^ z J 2 2 — tt — r ^ •• ^ C c >- >' 2 •" ~I IJ 3j "* -5 = ^ CD .J2 1) in >- rt fr, rt -J OJ OS : IX C > O c: CO 2S G> -^ O] £1 ^-< c> a ^ u w ^ ^ -0 aj « 0; a; OJ u u. u« 0 0 0 0 0 A a a X ITj tn t« cS "^ 0 '" p ^ ^ '^ 0 1^ 0; OJ ^ V Oj V u 0; -C '^ 'o c C c CQ u o Qj ^ iD I' ^ "*"* "si 0 uTi uri 0 'JO — in eJo U-C?^ >^ K=5 0 c/T P ^ -5 ^ & X '0 •^ ^ X ^ j_, 1^ c 'i ^ tf. "o jj v; ^ x E "o c p ,_j^ .0 K 0 ij ^ c .t: r- i: '-' -^ -r ±, ^ t, hJ N N X H Q -c o W c/5 :> X CO c^ c^ O O CO ^ CO CD CD fM I I I I I ■■ ^^^^^OCOCD < 2 5 »o 06 3 m r^ ^H -H 1— I CO t^ CO Ol — I ^ ^ C t^ 00 01 Ol fM x O^ r-H C z u W 00 1 II I £ ' > > - '-t: .5 i ^ .5 ° -7 -^ c ^ t- 2 ^ ^ = c ffl H ^ m ■> >o 0 0 c O] ro tn in ^ 66 00 00 ^^ ^^ 1— 1 J 0; 1; Q Z Brama • Mead and HaedricJi 67 ::; o -i — o z ^ >. y. ^ ^ ^ ^, 0 'r-> > C i-C IC c:2 *" o X x: > •—p f-H - — ■ ^ J w r- ^ c r— •-J t£ "S ^ — S c O O 0^ >> 0; 0/ Qj -J p J J 3, if^ "-^ IC C^ ^ ^H ;j :^ ii. ^ -ii -9 1; VD V] U K ^ T}< rH ^1 ^ in S 00 oc tt oi 1 ■S oc oc IC J, < ^ 11 SulUtln OF THE seum Comparative Zoology S- Uf\' Cj^^: OOl Evolution of the Tapiroid Skeleton from Heptodon to Tapirus by LEONARD B. RADINSKY HARVARD UNIVERSITY VOLUME 134, NUMBER 3 CAMBRIDGE, MASSACHUSETTS, U.S.A. 29 JUNE 1965 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- MEMoms 1864-1938 JoHNSONiA, Department of Mollusks, 1941- OocAsiONAL Papers on Mollusks, 1945- Other Publications. Bigelow, H. B. andW. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Lyman, C. P. and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 15. Whittington, H. B. and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948 (complete sets only). Pubhcations of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. MUS. COMP. ZC LIBRARY " ^ 1965 HARVARD UNIVERSITY, EVOLUTION OF THE TAPIROID SKELETON FROM HEPTODON TO TAPIRUS LEONARD B. RADINSKY^ INTRODUCTION The purpose of this study is two-fold: first, to describe and ilhistrate the best-pre- served known skeleton of an early Eocene perissodactyl and, second, to investigate the morphological changes which occurred dur- ing about 50 million years of evolution from the late early Eocene tapiroid Ilcpio- (lon to the recent genus Tapiius. Although equoid and tapiroid remains are fairly common in early Eocene deposits, good skeletons of these or other early Eo- cene perissodactyls are discouragingly rare. Most of the skeletal elements which have been found are incomplete, crushed, or coated with a hard deposit which makes preparation difficult. For these reasons no well-documented description of the osteol- ogy of an early Eocene perissodactyl has heretofore been published. The most de- tailed previous study, Kitts' revision of Hyracotherium (1956), is inadequately il- lustrated. The present paper therefore should fill a critical gap in knowledge of perissodactyl evolution. This study is based on an almost com- plete, well preserved, and excellently pre- pared skeleton of Heptoclon posticus, an early Eocene helaletid tapiroid. All early Eocene perissodact\'ls appear to have been ' Department of Biology, Brooklyn College, Brooklyn, N. Y., and Department of Vertebrate Paleontology, the American Museum of Natural History, New York. extremely similar in skeletal morphology, with the main differences between the major groups expressed in the dentition. The osteological differences which do exist apparently reflect relatively small differ- ences in size and do not appear to have phylogenetic significance. Therefore, the skeleton of Hepfodon described below rep- resents, probably fairly closely, the ancestral morphology not only of all later tapiroids, but of all other perissodactyls as well. To facilitate future studies by workers to whom the original material is not readily available, the bones described here are illustrated with figures drawn from several views. Heptoclon probably was directly ancestral to the modern tapirs, as well as to several extinct tapiroid lineages (see Radinsky, 196'3a). In the following description com- parisons are made between the skeleton of Heptoclon and that of a modern species of Tcipirus, T. pinchacjue, to detennine the ex- tent of the osteological changes which oc- curred in this most conservative of surviving perissodactyl lineages. To aid interpreta- tion of the functional significance of the observed changes, reference was made to Bresson, 1961, and to Campbell, 1936 and 1945, the most recent accounts of the myol- ogy of Recent species of Tapinis. In addi- tion, I dissected the facial and fore hmb musculatiu-e of a specimen of Tapiius indi- cus. My own observations and the pub- lished accounts confirm Gregory's ( 1929 ) restoration of the relatively unspecialized Bull. Mus. Comp. ZooL, 134(3): 69-106, June, 1965 69 70 Mii.scum of Conipanitivc Zoo/o^'i/, Vol 134, No. 3 imisculature of a titaiiothere. Gregory's work is still the best one to consult for illus- trations of muscle origins and insertions on the perissodactyl skeleton. The skeleton of Ileptodon posticus here described, MCZ 17670,- was collected in Lostcabinian ( late Wasatchian ) beds in the Wind River basin, \\'yoming, by Mr. Henry Seton, \\ho published a preliminary note on the specimen in 1931. Mr. Seton has gener- ously allowed me to complete the descrip- tion.' The scientific xalue of MCZ 17670 has been greatly enhanced by Seton's skill and painstaking work in preparing it. Post- cranial remains of Heptodon are known in only one other specimen, AMNH 294, an incomplete skeleton of H. calciciihis. This was brief!)' described by Osborn and Wort- man (1892), and differs from H. posticus primarily in its smaller size and more slen- der build. The systematics of Heptodon have been discussed in a recent revision of North American tapiroids (Radinsky, 1963a). Of the four living species of Tapirus, T. ))incJuique was chosen for comparison with Heptodon mainly because it is the least spe- cialized of the surviving tapirs, and also because it is the rarest and least well known of the four species. Illustrations of its post- cranial skeleton are published for the first time in this paper. Specimens of Tapirus pineluique examined for this study include: MCZ:M 6037, YPM 204, and AMi\H:M 149331 and 149424. For knowledge of intermediate stages be- tween Heptodon and Tapirus, comparisons were made with Protapirus, a primitive Oligocene tapirid. Although known species of Frotapirus occur too late in time to be directly ancestral to later tapirids ( see Schaub, 1928, p. 13), the genus is probably -Abbreviations of institutions are: AMNH, the American Museum of Natural History, Department of Vertebrate Paleontology; AMNH:M, same. Department of Mammalogy; MCZ, Museum of Comparative Zoology, Division of Vertebrate Pale- ontology; MCZ:M, same, Division of Mammalogy; PU, Princeton University; YPM, Yale Peabody Nhiseuin. representative of a morphological stage through which the main line of tapirid evo- lution passed. Of the skeleton of Frotapirus, only the skull, atlas, fore limb (PU 10899), and manus (AMNH 662) are known. These were originally described by Hatcher (1896), and Wortman and Earle ( 1893 ) , respec- tively, and were most recently redescribed by Scott (1941). To avoid repetition in the comparisons of bones of Heptodon and Tapirus, differences in size will not be mentioned. These may be calculated from the illustrations. It will suffice to note here that Heptodon posticus is about 40 per cent smaller than Tapirus pinchaque. I am extremely grateful to Mr. Henry Seton, whose encouragement and generous support made this paper possible. I wish to thank Professors B. Patterson and A. S. Romer for making available to me the re- sources of the Division of Vertebrate Pale- ontology of the Museum of Comparative Zoology, Miss B. Lawrence and Mr. C. Mack for facilitating my studies of recent tapirs in the Division of Mammalogy of the Museum of Comparative Zoology, and Dr. R. Van Gelder for permission to study tapir skeletons in the collections of the Depart- ment of Mammology of the American Mu- seum of Natural History. The drawings of Heptodon limb bones are the excellent work of Mr. N. Strekalovski. The plates were drawn by Llewellyn I. Price. This study was supported by National Science Founda- tion Grant No. GB 2386. OSTEOLOGY Skull and Mandible The skull of Heptodon posticus included in MCZ 17670 is one of the best-preserved early Eocene perissodactyl skulls ever dis- covered (see Pis. 1-3). A slight anterior displacement of the right side of the skull, and a crushing of the right side of the ros- trum, which resulted in a slight dorsal dis- placement of the right nasal bone, are the only evident distortions in proportions. Evolution of Tapiroid Skeleton • Rodinsky 71 Fig. 1. Skull and mandible. Above, Hepfodon posticus, MCZ 17670, X 'A (after Hatcher, 1896, pi. 5), X 'A ■ Below, Tapirus pinchaque, YPM 204 The skull ( see Fig. 1 and Pis. 1-3 ) is rel- atively long and narrow, measuring 195 mm from the anterior tip of the premaxilla to the dorsal edge of the foramen magnum, and about 80 mm wide across the zygomatic arches. The occiput is 56 mm high (from basioccipital to top of nuchal crest ) and has a maximum width of 46 mm (across the post-tympanic processes). The preorbital portion of the skull is slightly longer than the postorbital portion. The nasal incision extends back to a point over the postcanine diastema, closer to P' than C\ The pre- maxillae contact the nasals dorsallv and ex- clude the maxillae from the nasal incision. The infraorbital foramen is located above the front of P-* and the anterior edge of the orbit is above the anterior border of M-. The lacrimal has a relatively small facial portion, and bears a triangular tubercle on the rim of the orbit. The supraorbital proc- esses are relatively massive and are followed by a pronounced postorbital constriction. The braincase is slightly expanded and the sagittal and lambdoidal crests are promi- nent. The zygomatic arches are relatively slender. The external auditory meatus is widely open ventrally; it is bounded ante- 72 Mu.sciDii oj (U))iii)(ir(ilii(' Zo()/o,i,'!/, Vol. 134, No. 3 riorl\ b\ a R'latixcl) [)r()inim'iil postgk'iioid process and posteriorly by a short, antcro- xentralh' projecting post-tympanic process. The paroccipital process is relatively long and thin, slightK flattened anteroposteriorh', and extends postero\entrall\' and slighth laterally. The palate is almost flat from front to back and only slightly arched transversely. It has a pair of small elongate incisive fo- ramina between the third incisors, and a long deep anterior median groove which may have opened into the nasal passage. The internal nares open at the posterior border of M-. The interpterygoid fossa is relatively deep and narrow, with no trace of a vomer on the presphenoid almost as far forward as the anterior border of the choanae. The glenoid fossa is relatively flat. It is bounded posteriorly by a strong post- ?scf Fig. 2. Hepfodon poiticus. MCZ 17670. Above, lateral view of orbital region with zygomatic arch removed. Below, anterolateral view of same region. X 'A- '^o'" abbrevia- tions see p. 102. glenoid process which is oriented postero- laterally at an angle of about 30 degrees from the long axis of the skull. The basi- cranial axis is inclined slightly anterodor- sally relative to the plane of the palate. At the anterior edge of the orbit (see Fig. 2) the lacrimal bone is too broken to allow accurate determination of the number or configuration of the openings into the nasolacrimal canal. The maxillary foramen is visible in the extreme anteroventral cor- ner of the orbit. The sphenopalatine fora- men is located above the anterior edge of M'', and is relatively large; a small posterior palatine foramen is located posteroventral to the sphenopalatine foramen. Sutures be- tween the bones of the orbit cannot be definitely determined, partly due to the advanced age of MCZ 17670 (the teeth are heavily worn), and partly because the bones are extensively fractured. The posterior part of the medial wall of the orbit (see Fig. 2) is pierced by four foramina, arranged in a posteroventrally descending row. Beginning with the most anterior one, which is located a short dis- tance behind the posterior palatine fora- men, these are: a small ethmoid foramen, a very large optic foramen, a smaller, ver- tically-bilobed, slit-like, anterior lacerate foramen ( = orbital fissure or sphenoidal foramen), and the anterior opening of the alisphenoid canal. Examination of the in- terior of the braincase revealed a small fora- men, apparently the foramen rotundum, opening into the alisphenoid canal. The op- tic foramen is unusually large and Simpson (1952, p. 200) suggested that the equally large, apparently homologous foramen in Hijracotherium was the confluent opening of the optic and anterior lacerate foramina. However, inside the braincase of Heptodon it can be seen that the left and right internal openings of the large foramen in {question are confined entirely within the bounds of the optic chiasiua and there are no grooves suggesting that nerves other than the optic left through the foramen. The presence of Evolution of Tapiroid Skeleton • Radinsky 73 a foramen opening into the alisphenoid canal confirms the interpretation that the sht-hke opening behind the optic foramen is the anterior lacerate foramen. Above and slightly posterior to the ante- rior lacerate foramen on each side of the skull is an irregnlarly-shaped opening, at least part of which is artificial, but which may include a natural opening into the braincase ( presumably the sinus canal fora- men ) . There is a pit immediately above the anterior opening of the alisphenoid ca- nal. A shallow vertical groove which ascends Fig. 3. Heptodon posticus. MCZ 17670. Left, restoration of basicranial region of skull. Right, ventrolateral view of restored left petrosal, witfi tympanic and tegmen tympani removed. X 'A- from the laterodorsal border of the anterior opening of the alisphenoid canal may mark the course of the supraorbital artery, or pos- sibly the deep temporal artery. The foramen ovale (see Fig. 3) is sepa- rated from the middle lacerate foramen by a strip of alisphenoid about 7.5 mm wide, the posterior border of which is notched by two grooves. The more lateral of the two, located just medial to the anterior edge of the postglenoid process, is relatively broad and probably marks the passage of the in- ternal maxillary artery. The second, more medial groove, is narrow and slit-like, and probably contained the chorda tympani. The postglenoid foramen is relatively large. In ventral view the basisphenoid overlaps the ventral border of the petrosal. There is a medium-sized posterior lacerate foramen, a relatively large hypoglossal (or condyloid) foramen, and a medium-sized mastoid fora- men. Both petrosals of MCZ 17670 lack small portions, but fortunately the missing parts of each petrosal are preserved on the oppo- site side so that it is possible to obtain a composite picture of the entire bone (see Fig. 3 ) . The ventral border of the petrosal is relatively long and convex, and the ante- rior border is relatively short and concave. Seen in ventrolateral view, the surface of the petrosal is flat anteroventrally, and swells posterodorsally towards a relatively low promontorium. The surface of the promontorium is smooth, and shows no traces of grooves for blood vessels. The tegmen tympani extends relatively far ven- trally, forming the lateral wall of a deep groove for the facial nerxe. This is similar to the condition in modern tapirs in \\'hich the facial nerve runs posteriorly along the lateral face of the petrosal in a ventrally- open groove, rather than in an enclosed ca- nal. Two small foramina are present on the lateral face of the tegmen tympani. The lower and slightly more medial one proba- bly is the facial hiatus, and thus served for the exit of the great superficial petrosal nerve (which joins the great deep petrosal nerve anteriorly to form the vidian nerve). The higher and more laterally situated fora- men may have transmitted the small super- ficial petrosal nen^e. The tegmen tympani is broken off above the level of the facial canal on the right 74 Mtiscinu of Coniiunafivc Zoolo<:,y. Vol. 134, No. 3 petrosal, (■.\[)()siiiu; the opt'iiiiiu; ot tlic facial canal and the fenestra oxalis. The latter is located posteriori)' and slightK' \entrall\' from the former. A ver\' small foramen, possihK a nutrient foramen, is pri'sent on both petrosals a short distance anti'ro\ entral to the fenestra ovalis. The fenestra rotunda is located a short distance posteroventral to the fenestra ovalis. A deep V-shaped groove extends posteroventrally from the fenestra rotunda and separates the main body of the petrosal from a small, globose posterior portion. This groove occupies ap- proximately the same position as the auric- ular branch of the vagus nerve ( which joins the facial nerve near the stylomastoid fora- men ) , but appears too large to have housed that ner\'e alone. The stylomastoid foramen is represented by a broadly open groove anteromedial to the post-tympanic process. No tympanohyal is presei-ved; the broken surface anterior to the stylomastoid foramen on left and right petrosals suggests that one may originally have been present. There are no separate fossae for the stapedius or tensor tympani muscles. These presumably occupied the same groove as the facial nerve, the tensor tympani at the anterior end of the groove, lateral to the facial nerve, and the stapedius at the posterior end, medial to the nerve. On the medial side of the petrosal a shal- low depression adjacent to the concavity of the anterior edge may have accommodated the semilunar ( Gasserian ) ganglion of the trigeminal nerve. The internal auditory meatus is relatively large and elongate in an almost vertical direction. Cochlear and ves- tibular aqueducts are visible near the poste- rior edge of the petrosal, the former poste- roventral and the latter posterodorsal to the internal auditory meatus. The entire tympanic bone was preserved on the left side but was largely destroyed during the preparation of a latex mold of the skull. Fortunately, photographs and drawings made by Henry Seton prior to this mishap provide a record of the config- uration of this bone. The tympanic in this early Eocene tapiroid is a semicircular strip of bone, dorsoventrally flattened and slightly expanded in a horizontal plane. In other words, it is only slightly modified from the simple, narrow, open tympanic ring which is considered primitive for mammals. Kitts (1956, p. 17) stated that the early Eocene equoid Hyracotherimn had a tym- panic bulla which was "... apparently oval in shape and moderately inflated." Re-ex- amination of the specimens involved showed that Kitts mistook the petrosal for a tym- panic bulla. Except for the skull of Hepto- don posticus described here, I know of no other specimen of an early Eocene peris- sodactyl in which the tympanic is preserved. The tympanic probably was a relatively narrow semicircular strip of bone in all early Eocene perissodactyls. Its failure to be preserved in otherwise complete skulls is probably because it was loosely attached to the skull, as in modern tapirs. Portions of at least some of the auditory ossicles are visible dorsal and medial to the tympanic ring, but are too fragile to be removed from the last remnants of the hard sandstone matrix. Their description must await discovery of additional specimens. The symphysis is slightly constricted and extends back as far as Pi. Its dorsal surface is deeply concave transversely. Three small mental foramina are visible, beneath Ci, the middle of the diastema, and P-. The body of the mandible is relatively long and slender, with a very slightly convex ventral border. The angle is moderately convex and extends relatively far behind the con- dyle and slightly below the ventral border of the body. The condyle is located rela- tively high above the tooth row and is inclined medioventrally at an angle of about 20 degrees from horizontal. Its articular surface slopes anteriorly and a small facet continues ventrally onto the medial third of the posterior face of the condyle. The coronoid process is relatively small and ver- tical. On the lateral surface of the ascending Evolution of Tapiroid Skeleton • Radinsky 75 ramus a pronounced fossa for the zygoniat- icomandibularis extends ventrally to about the level of the tooth row. The border of the insertion area of the masseter is marked by a fairh' prominent ridge for most of its length. The medial surface of the angle is slightly concave and bears prominent scars from the insertion of the internal pterygoid. COMPARISON WITH TAPIRUS The skull of Tapiriis pincJutque (see Fig. 1) differs from that of Heptodon in several features, most of which are related to one or more of three basic developments: evolution of a proboscis, relative enlarge- ment and change in proportions of the brain, and increased specialization of the mastica- tory apparatus. The most obvious differences between the skulls of Heptodon and Tapinis are those associated with proboscis develop- ment. Primary modifications for the pro- boscis are the enlargement (or retraction) of the nasal incision, which provides room for the main mass of the proboscis, and shortening of the nasals, to allow flexibilitv to that organ. The nasal incision in Tapinis extends back over the orbits, and the nasals do not reach beyond the first premolar. An additional factor which adds to the vertical dimension of the nasal incision in Tapinis is the higher position of the nasals in that genus, which results from the presence of a frontal sinus. Frontal sinus development is correlated with changes in brain propor- tions, and will be discussed below. The nasal diverticulum (a blind cartilaginous sac which opens into the main nasal pas- sage) has been displaced from the nasal incision in Tapinis and is lodged in a long, broad groove which begins on the ascend- ing process of the maxilla, extends up along the posterior border of the incision on the anterior edge of the frontals, and terminates in a curl on the posterior edge of the nasals. The absence of a groove or fossa for the nasal diverticulum in Heptodon indicates that the diverticulum in that primitive ta- piroid must have been relatively small and lodged in the nasal incision, as in modern horses and rhinos. Secondary cranial modifications associ- ated with proboscis development in Tapinis include a posterior displacement of the nasal cavity, strengthening of the premax- illae, and possibly the anterior shift of the orbit. In Heptodon the nasal cavity is lo- cated in front of the orbit, with the cribri- form plate situated near the anterior border of the frontals, at the anterior edge of the orbit. In Tapinis, the tremendous expan- sion of the nasal incision has removed most of the lateral walls of the rostrum, and the greater part of the chamber which houses the ethmoturbinals is located between the orbits, with the cribriform plate located well behind the orbit, near the posterior border of the frontals. This has resulted in a lateral displacement of the walls of the orbit which obliterates the postorbital con- striction and greatly deaccentuates the su- praorbital processes. Another result of the posterior displacement of the nasal chamber relative to the orbits is that the internal nares, which open posteriorly in Heptodon, open downward as well as backward in Tapinis. The expansion of the nasal incision in Tapinis has left the premaxillae extending far out, relatively unsupported. Presuma- bly to strengthen this area against the ver- tical stresses which result from use of the incisors, the premaxillae are thickened and arch downward and, with the anterior ends of the maxillae, are closer together, forming a relatively deep, narro\\', and arched pro- jection on which the incisors and canines are borne. Associated with this development, the two lateral incisive foramina seen in Heptodon are merged into a single, large, median opening in Tapinis, extending back almost to the first premolar. The large size of this opening is probably related to the fact that it occupies that part of the palate least invobed in transmitting vertical stresses from the incisors and canines, and 76 Museum of Comjuirativc Zoolosiy. Vol. 134, No. 3 thus lU't'ds no boin support. Anotlier factor possihh- relatt'd to the stifngtheuing of the premaxillae in Tapinis is the large size of tlie third incisor, which has been trans- formed into a small tusk which occludes against the front of the large lower canine. The upper third incisor of Tapinis is lo- cated more anteriorly than its functional counterpart, the upper canine, is in Hcpto- dan, and would recjuire stronger premaxil- lae for support. The orbit is located more anteriorly in Tapini.'i than in lleptodon. This may be ad- vantageous for operation of the proboscis since the main muscles involved (the leva- tor nasolabialis and superior and inferior maxillolabialis) take origin on the anterior rim of the orbit and it woidd be mechani- cally more advantageous to have them orig- inate nearer to their insertions. Another result of the anterior displacement of the orbit in Tapinis is that it brings the masseter forward over a greater portion of the tooth row, which increases the mechanical ad- vantage of that muscle in mastication (see discussion below ) . The skull is larger and heavier relative to the body in the modern tapir than in Hcp- todon. Probably reflecting this difference, the paroccipital process is relatively more robust, and is vertically oriented and fused to the post-t\'mpanic process, and the back of the skull is relatively wider, suggesting relatively stronger neck muscles in Tapini.s than in lleptodon (see p. 81). However, stronger cervical musculature might also be correlated with increased stresses resulting from the use of the proboscis. Another fac- tor which may be pertinent here is orienta- tion of tlie liead. In lle])lodou the occlusal plane diverges auteroveutrally from the basicranial axis, while in Tapinis the occlu- sal plane is parallel or slightly anterodor- sally inclined to the basicranium. This suggests that the head is held slightly more horizontally in l\i})inis than it was in Hep- todon, possibly in response to its relatively heavier weight, or possibly because of the proboscis. A second major factor which has been responsible for differences between the skulls of lleptodon and Tapinis is brain evolution. Evolution of the tapiroid brain will be described in a future paper. For the present discussion it will be sufficient to note that in tapiroids, as in equoids, a major feature of brain evolution (described in a classic work by Edinger, 1948) has been the expansion of the cerebral hemispheres and relativ e decrease in size of the olfactory bulbs. In lleptodon, the olfactory bulbs are relatively long, underlying most of the length of the frontals, and lie at the same level dorsally as the cerebrum. In Tapinis, the olfactory bulbs are relatively short, ly- ing under only about the posterior third of the frontals, and the dorsal surface of the olfactory bulb chambers is at a lower level than that of the expanded cerebrum. The anterior part of the frontals is underlain by the posteroventrally-sloping nasal chamber. Since the ventral surface of the frontals im- mediately overlies the olfactory bulbs, and the dorsal surface must remain high to keep the nasal incision open, the external and in- ternal tabulae of the frontal bone in Tapinis have become separated by a space, the frontal sinus. Thus, formation of the frontal sinus in Tapinis may be thought of as a re- sponse to the necessity of maintaining the nasals high above the orbits (to provide room for the proboscis beneath) when growth of the olfactory bulbs failed to keep pace with cerebral expansion and growth of the rest of the skull. ( For a more thorough discussion of frontal sinus formation, see Edinger, 1950. ) As a direct result of cerebral expansion, the braincase of Tapinis is relatively wider than that of Heptodon. Sagittal and lamb- doidal crests are relatively lower and do not project back as far in Tapinis as in Hepto- don, perhaps because the expanded brain- case provides more room for attachment of the temporal muscles in the modern form. Evolution of Tapiroid Skeleton • Radiihsky 77 Or, on the otlier liand, the lower crests may reflect a relatively smaller amount of tem- poral musculature in Tapiiiis. The promi- nent sagittal and lambdoidal crests on the skull of Heptodon create a dorsal profile which diverges posterodorsally from the basicranial axis and occlusal plane. In Tap- irus pinchaque the weaker crests and ex- panded frontal sinus result in a dorsal skull profile which parallels the basicranial axis. The foramen ovale and postglenoid fora- men have shifted posteriorly in Tapirus and are confluent with the foramen lacerum medium. Edinger and Kitts (1954) have pointed out that similar shifts in the posi- tion of the foramen ovale have occurred in equoids and rhinocerotoids. These changes may be related to the cerebral expansion that occurred in all three groups, but the exact reasons are still obscure. The third major area in which changes have occurred during evolution of the skull from Heptodon to Tapirus involves the mas- ticatory apparatus. The molar cusp pattern of Tapirus is similar to that of Heptodon, differing only in the following features: metacone more labially located, providing a slightly longer metaloph; ectoloph shorter ( posterior to metacone apex ) and relatively less prominent; paralophid and metalophid virtually nonexistent; M,-, hypoconulid ab- sent. In Heptodon, molar shear occurs between the lingual side of the ectoloph and labial side of the paralophid and meta- lophid, as well as between upper and lower cross-lophs. As a result of the differences noted above, molar shear in Tapirus is al- most entirely confined to the cross-lophs, \\ith the anterior sides of protoloph and metaloph abo\'e shearing against the poste- rior sides of protolophid and hypolophid below. The functional advantage or adap- ti\e significance of eliminating ectoloph shear and emphasizing cross-loph shear is not immediately apparent. This trend oc- curred in other tapiroid families besides the Tapiridae (Helaletidae and Deperetellidae) while in still other tapiroid families ( Lophi- odontidae and Lophialetidae ) ectoloph shear was retained, or even emphasized. Probably the most important change from the dentition of Heptodon to that of Tapirus has been the molarization of the premolars. In Heptodon none of the premolars are molariform; in Tapirus, P^ and P^ are sub- molariform and the remaining premolars are molariform. As a result, the cheek tooth row is relatively longer and there is a rela- tively larger surface area available for mas- tication in Tapirus than in Heptodon. Per- haps in response to this increase in occlusal area at the front of the tooth row, the ante- rior border of the origin of the masseter ( marked by a scar on the maxilla and malar below the orbit) has shifted forward rela- tive to the tooth row, and is located above the anterior border of M^ in Tapirus, com- pared with about the middle of M- in Heptodon. This increases the mechanical advantage of the masseter by lengthening its lever arm ( distance from the masseter to jaw articulation) relative to the lever ami of the resistance (distance from teeth to articulation ) . The orbit also is located rel- atively more anteriorly in Tapirus than in Heptodon, and the anterior edge of the masseter scar is in the same position relative to orbit in both forms. Since the anterior part of the masseter takes origin from the bones forming the lateroxentral border of the orbit, it is possible that the selective ad- vantages resulting from a forward shift of the masseter were a significant factor in bringing about the change in position of the orbit. However, it should be kept in mind that in horses and other mammals the ante- rior origin of the masseter has moved for- ward independent of the orbit, by shifting onto the maxilla anterior to the orbit, and also that there are other functional advan- tages (related to proboscis development) involved in an anterior shift of the orbit. Another change in dentition between Heptodon and Tapirus has been the atrophy of the upper canine and enlargement and ^caninization of the upper third incisor. 78 Museum of Compdrdtivc Zoolom/, Vol. 134. No. 3 Thus, in 'rii])irus, tlu' iipi)t'r third incisor has rephicecl the caiiiiu> functionally, and the upper tusk occludes in front ol the lower, the reverse of the usual nianinialian condi- tion. Tliis may be related to pr()l)oscis de- \elopment if, with the proboscis extending in front of the premaxilla, the upper canine is located too far back to be effective. An- other consideration is that the atrophy of the upper canine increases the amount of space available in the diastema for manipu- lation of food. Even without this additional space, the postcanine diastema is relatively longer in T(ij)inis than in lleptodon. Finally, there are several features, in which the skull of Tapirus differs from that of Hcptodon. which do not seem to be cor- related with the three major developments discussed above. The optic foramen is smaller relative to the size of the skull and the other orbital foramina in Tapirus than in lleptodon. The tympanic, a simple half- ring in Hcptodon, is expanded anteroven- trally and laterally and forms a short floor to the external auditory meatus in Tapirus, although it is never large enough to form an inflated auditory bulla. The petrosal of Tajyirtis differs from that of lleptodon in having a shallower subarcuate fossa, a teg- men tympani composed of cancellous rather than solid bone, and in ha\'ing a tympano- hyal fused to it. The mandible of Tapirus is fairly similar to that of lleptodon, differing from the lat- ter in having a more procumbent s\'mphysis (in correlation with the downcurved pre- maxillae), a posteriorly cm-yed coronoid process (which may increase the mechani- cal efficiency of the temporalis), a slighth more rounded angle, and a relativelv shorter and wider condyle. SKULL EVOLUTION The few known skulls of fossil tapiroids reveal some of the intermediate stages in the evolutionary developments discussed above and provide information on the rates at which the changes took place. In a late early Eocene specimen of llep- todon caJeiculus (AMNH 294), the nasal incision extends back to a point over V\ and the premaxillae no longer contact the nasals. In the middle Eocene genus Helalctes, which may have included species near the ancestry of the Tapiridae, the nasal incision is tremendously enlarged, both dorsoven- trally and posteriorly, and extends back as far as P-'. The anterior wall of the orbit is over M\ and there is a groove for the nasal diverticulum on the ascending process of the maxilla. However, the nasals of HeJa- letes are unshortened and extend as far for- ward as the anterior border of the premaxil- lae. This suggests that Helaletes did not have a prehensile proboscis, since the long nasals would have restricted its mobility. The next stage in cranial evolution in the line leading to modern tapirs is represented by the latest Eocene or early Oligocene species Colodon? hancocki (known from a few specimens in collections of the Univer- sity of Oregon Museum of Natural History). In Colodon? haneoeki the nasal incision is extended slightly more posteriorly than in llekilefes, but is not quite as deep postero- ventrally. However, the nasals are shorter in C? haneoeki, extending only to to a point over the postcanine diastema. This suggests that by the beginning of the Oligocene an- cestral tapirids had a proboscis which, judging from the degree of retraction of the nasal incision and nasal shortening, may have been almost as long as that of modern tapirs. A Xorth American late Oligocene skull of Protapirus (PU 10899), the most primitive known tapirid, was described and figured by Hatcher ( 1896). Scott ( 1941, p. 754 and pi. 79) provided a few additional observa- tions and new illustrations, with the pre- maxillae, which are missing in the Princeton skull, restored from a second skull (South Dakota School of Mines 2829). Modifica- tions for the proboscis in the late Oligo- cene Protapirus skull are in about the same stage of development as in Colodon? Evolution of Tapiroid Skeleton • Radinsky 79 Fig. 4. Protapirus. Restoration of anterior half of skull, based mainly on PU 10899. Compare with Fig. X 'A. hancocki. The nasal incision extends back to a point over P^, and the anterior edge of the orbit is over the middle of M^ The nasals terminate slightly anterior to P^ A broad groove for the nasal diverticulum ex- tends up the ascending process of the max- illa, arches posteriorly into the dorsal surface of the prominent supraorbital process of the frontal, and continues back to the posterior border of the supraorbital process. This is a marked difference from the condition in Tapirus, where the supraorbital processes are suppressed and the groove for the nasal diverticulum curls medially and anteriorly to terminate on the posterior border of the nasals. Both Hatcher's and Scott's illustrations of the Princeton Protapirus skull ( drawn from the left side) show a deep groove on the anterior portion of the maxilla, parallel to the groove for the nasal diverticulum and separated from it by a high ridge. Examina- tion of the right side of the skull shows that this anterior groove is an artifact, due to a break and inward crushing of the anterior part of the left maxilla. A new restoration of the anterior half of the skull of Protapirus is shown in Figure 4. The nasal incision is narrower in Protap- irus than in Tapirus because the anterior part of the maxilla is higher (less excavated) and the nasals are lower in the Oligocene genus. Scott (loc. cit.) considered the dif- ferences between the two forms great enough to suggest that in Protapirus the proboscis was only in an incipient stage. However, in Scott's restoration the nasal incision is drawn too narrow. In my opinion the nasal incision is large enough and the nasals are short enough in Protapirus to suggest that it had a fairly versatile pro- boscis. There is still a marked postorbital con- striction and the supraorbital processes are prominent in Protapirus, indicating that the nasal chamber did not extend back between the orbits in that genus. The premaxillae arch downwards, as in Tapirus, although the upper third incisor is not enlarged. The up- per canine, however, is about as small as in the modern tapir. This indicates that the third incisor was enlarged after the atrophy of the canine, probably to functionally re- place the upper canine in shear against the lower canine. The premolars are still non- molariform to submolariform in Protapirus. The foramen ovale and postglenoid foramen are confluent with the foramen lacerum me- dium in Protapirus. Although Protapirus possesses all the fea- tures one would expect to find in the ances- tor of Tapirus, the known species of Protap- irus appear too late in time to be directly ancestral to modern tapirs. Schuab (1928) described under the name Tapirus helvctius the anterior half of an almost modern ta- pirid skull from Europe, which he considered on the basis of lithological correlation to be of middle or late Oligocene age. Cranial 80 Museum of Covijuiidtivc Zooloiiy, Vol. 134. \o. 3 iiiodificatioiis lor the pioboscis arc more ad- vanced ill Tiipiriis hclvctiiis than in Protap- irns in the Following featnres: the nasal incision extends more posteriorly, reaching to a point ov(M- the orbit; the anterior part of the maxilla is low t'r and the nasals appear to be located slightly higher, leaving a deeper nasal incision; Hie groove for the nasal di\ frticuhiin curls onto the posterior border of the nasals, as in modern tapirs; the supraorbital processes are less pro- nounced. The anterior part of the skull of Tapirus hclvctius is basically like that of Tapinis pincJuique, differing only in the following features: the nasal incision is not (|uite as deep posteroventrally, partly because a long posterior process of the pre- maxilla extends back over the maxilla to a point above the posterior edge of P"*; the nasals appear to be located slightly lower in the Oligocene species. Thus on the basis of anterior cranial morphology, Tapirus hclvc- tius is intermediate between Pwtapirus and modern tapirs, but definitely closer to the latter. The same is true for its dentition — the premolars of T. hclvctius are almost, but not quite, molariform. Another tapirid close in age and similar in morphology to Tapirus hclvctius was de- scribed by Schlaikjer ( 1937 ) under the name Miotapirus liarrisoncnsis. This form, known only from the anterior half of a skull and a few limb bones (MCZ 2949) from early Miocene deposits in Wyoming, dis- plays proboscis modification about as in Tapirus hclvctius, except that the posterior process of the premaxilla is shorter, extend- ing back to a point above P-, and the nasals appear to be slightly higher. The skulls of Tapirus hclvctius and Miotapirus harrison- ensis indicate that b\' about the end of the Oligocene, or over 25 million years ago, cranial modifications for the tapirid probos- cis were in essentially the same stage as in the recent species Tapirus pinch(Uiuc. One of the major steps in the evolution of the proboscis was the tremendous enlarge- ment of the nasal incision that occurred during the relatively short period of time between early Eocene Heptodon and mid- dle Eocene Ilclalctcs. Since the nasals were unshortened in Ilclalctcs it seems unlikely that it had a proboscis; this suggests that the reasons for the initial enlargement of the nasal incision were probably not related to proboscis development. Primitively in perissodactyls the nasal diverticulum is lodged in the nasal incision. Enlarged nasal diverticula apparently have created depres- sions in the maxillary walls (preorbital fossae ) and caused enlarged nasal incisions in many extinct perissodactyls (see Gregory, 1920). It therefore seems possible that ex- pansion of the nasal diverticulum may have been responsible for the enlargement of the nasal incision in Ilclalctcs. With the nasal incision enlarged, tapiroids would then have been preadapted for proboscis develop- ment. Hyoid arch: Fragments of the hyoid arch are presei-ved in MCZ 17670, but most of the elements are too incomplete to yield much information. Neither ceratohyals nor thyrohyals were fused to the basihyal in Heptodon, and there is a long, low, irregular lingual process on the basihyal. In Tapirus, the thyrohyals are fused to the basihyal, and there is no lingual process. AXIAL SKELETON Vertebrae (Fig. 5) : The vertebrae known for lleptodon include all the cervicals (ex- cept the fifth ) , the first two thoracic, and the last lumbar. In the atlas, the vertebrar- terial canal enters at the posterior edge of the transverse process and emerges a short distance anteriorly on the ventral side. The atlantal (or alar) groove is open, notching the anterior edge of the transverse process. The neural spine of the last lumbar vertebra is only very slightly cranially inclined, w hich suggests reduced mobility of the ver- tebral column (Slijper, 1946, p. 103). Vertebrae of Ta))irus differ from those of lleptodon in the following featiu-es: cervical vertebrae relativelv shorter and wider, with Evolution of Tapiroid Skeleton • Radinsky 81 Fig. 5. Vertebrae of Heptodon posticus (MCZ 17670) and Tapirus pinchaque (AMNH:M 149424). A, E, dorsal atlases of H. posticus and 7, pinchaque, respectively. B, C, D, lateral views of axis and tfiird and fourth vertebrae of H. posticus. F, G, lateral and anterior vie/zs of sixth cervical vertebra of H. posticus. H, I, J, views of seventh cervical and first and second thoracic vertebrae of H. posticus. K, L, lateral views of last lum tebra of H. posticus and T. pinchaque, respectively. All X Vj. views of cervical lateral bar ver- anterior ends of centra more convex and posterior ends more deeply concave; atlan- tal groove of atlas bridged over by anterior growth of transverse process; odontoid proc- ess of axis relatively shorter and broader; neural spine of fourth cervical vertebra lower; postzygapophyses of first thoracic vertebra facing more laterally (and less ventrally); neural spine of last lumbar ver- tebra slightly caudally inclined. The shorter, wider and more opisthocoel- ous cervical vertebrae of Tapirus indicate a more powerful neck in the modern tapir than in Heptodon, probably in response to the needs of supporting a relatively larger and heavier head. The atlantal groove of 82 Miiscinii oj Coinpaidlivc '/jx^loiiy. Vol. 134, No. 3 [\\v atlas is bridged on cr in all iiiodcni pcris- sodactyls and may simply relloct expansion of trans\'erse processes to provide greater area for muscle attaclinient. The more lat- erallv-facing post/Agapopln ses of the first thoracic \ertebra restrict lateral movement but strengthen articulation between it and th(^ following vertebra. The trend from cra- nial to caudal inclination of lumbar neural spines is correlated with decreasing impor- tance of the longissimus muscles as spinal flexors and a backwards shift in their inser- tion from lumbar to sacral vertebrae. This results in decreased mobility of the verte- bral column and is often correlated with increasing bodv weight ( Slijper, 1946, pp. 103-104)'. APPENDICULAR SKELETON Sca))iil(i (Fig. 6): Only the glenoid end of the scapula is known for Hcptodon but a fairly accurate restoration of the entire bone may be made by extrapolating from com- plete scapulae known for the early Eocene ecjuoid Ihjracotherium (figured in Kitts, 1956, pi. 2, fig. 1) and the middle Eocene tapiroid Ilchilctcs (several specimens in collections of the United States National Museum). The posterior border is straight, the N'crtebral border straight to gently con- vex, and the anterior border more strongly convex. The neck is moderately constricted. The spine extends ventrally almost to the glenoid and is high at its ventral border. It has a small tuber spinae and a relatively pronu'uent acromion. Infraspinatus and su- praspinatus fossae are about equal in area. The tuber scapulae is low and bears a small coracoid process. In Tapirus the spine is reduced ventrally, terminating more dorsally than in Hcpto- don, and bears no acromion. The ventral reduction of the spine accentuates the prom- inence of the tuber spinae. There is no projecting coracoid process and the tuber scapulae is considerably higher and more prominent, and forms the ventral border Fig. 6. Right scapula. Left, Heptodon posticus, hypotheiical restoration, X 'A- Right, Tapirus pinchaque, MCZ:M 6037, X %. Evolution of Tapihoid Skeleton • Radinsky 83 Fig. 7. Right humerus. Above, Heptodon posticus, MCZ 17670, in, from left to rig'if, anterior, posterior, proximal, lat- eral and medial views. X 'A- Below, Tapirus pinchaque, MCZ:M 6037, same views. X 'A- 84 Miiscuiti of Coiiiixiidtirr y.oolo'^ij. \'()l. 134. Xo. 3 of a cU'i'p c'oracoscapiilar ( or stipiaspiiioiis ) notch. Recliiction of the acromion is an advanced feature in perissodactyls and appears to be correlated with loss of the claxicle. Ilepto- don ma\' haxe still had a clavicle but if so, it i)robabl\' was relatix ely small and did not articulate with the scapula. The absence of an acromion in Tapinis is reflected in mod- ifications of the muscles which originally attached to that part of the scapula. These include the levator scapulae ventralis, which typically originates on the transverse proc- esses of the atlas and inserts on the acro- mion, and the acromiodeltoid, which orig- inates on the acromion and inserts on the deltoid tuberosity of the humerus. Accord- ing to Campbell' ( 1936, p. 206), in Tapirus terrestris the levator scapulae ventralis and acromiodeltoid have fused to form the transversohiuneralis, which originates on the wing of the atlas and inserts on the fas- cia of the lateral head of the triceps. A deep coracoscapular notch distin- guishes scapulae of Tapini.s from those of all other perissodactyls. The notch is formed by the dorsal expansion of the prominent tuber scapulae, possibh' to extend the area available for attachment of the biceps. This expansion restricts the space open in front of the narrow neck of the scapula, thus forming the notch. The suprascapular ar- tery and nerve pass through the coraco- scapular notch. In life the notch is closed by a tendinous band which extends from the tip of the tuber scapulae to the antero- ventral edge of the anterior border of the scapula. Humerus (Fig. 7): The humerus of llcpt- odon is relativc^ly long and slender. In MCZ 17670 it is 34 mm wide across the dis- tal epicondyles and, judging from the pro- portions of the parts pr(\served, probably was about 150 mm long. The lateral (greater) tuberosity is raised slightly above the \v\v\ of the head and extends anteriorly and then curves medially. The medial (lesser) tuber- ositv is short and low, and not distinctlv separated from the head. The bicipital groove is undivided and relatively deep and narrow. Much of the proximal half of the shaft is missing in MCZ 17670 and the prox- imal (juarter is gone in AMNH 294 (Heptu- don calciculus), but between the two specimens enough is preserved to indicate that the deltoid crest and tubercle were not prominent. The teres tubercle is not evi- dent in AMNH 294; either the teres major left no attachment scar or it inserted rela- tively far proximally ( on the missing proxi- mal quarter of the shaft). The supinator crest is relatively low but sharp-edged, and flares out posterolaterally along the distal third of the shaft. The coronoid (supratrochlear) fossa is relatively broad and shallow, and the olecranon (an- coneal ) fossa slightly narrower and deeper. The thin wall of bone separating the two is perforated but this may be artificial. The trochlea is asymmetrical, narrowing later- ally. At the proximal end of the trochlea there is a thin strip of lateral condyle which rapidly narrows and terminates distally. Lateral and medial epicondyles are about equally prominent and both are relatively low. The humerus of Tapirus differs from that of lleptodon in the following features: lat- eral tuberosity higher, with a more promi- nent, medially-directed anterior hook which is separated from the main ridge by a broad groove; medial tuberosity higher, raised above the level of the head and almost as high as the lateral tuberosity, separated from the head by a low groove, and with its medial face vertical; bicipital notch rela- tively wider; deltoid tubercle prominent, located almost one-third of the way down the shaft, with a narrow ridge continuing distalK- from it almost to the coronoid fossa; teres tubercle prominent, located one-third to one-half of the way down the shaft; supi- nator ridge not as extended proximally and blunter at its proximal end; distal end of shaft deeper anteroposteriorly; lateral con- dyle wider; coronoid fossa shallower, olec- Evolution of Tapiroid Skeleton • Radinsky 85 :^i ''le is evenly convex and has a promi- nent medial keel on the posterior half. Two pits on the posterior face, immediately above the condyle, accommodated the prox- imal ends of the sesamoids. On each side of the condyle a deep pit sinniounted by a low tuberosit\' marks the attachment of deep 90 Museum of Comparative Zoolo' oriented; posterior lunar- magnum facet relatively smaller. Cunei- form with posterior part of proximal surface lower, resulting in a more distally located pisiform facet. Trapezium relatively higher. Trapezoid relatively shorter ( anteroposte- riorly) and higher, with less of a posterior process. Magnum with a relatively lower and wider anterior face and a shorter and less pointed posterior process; magnum- unciform articulation relatively higher, but lacking a separate posterior facet; magnum- second metacarpal facet higher proximodis- tally. Metacarpals relatixely shorter and wider. The main difference between the front foot of Tapini.s- and that of Ilcptodon is in the radio-carpal joint. The truncation of the posterior end of the scaphoid and lowering of the posterior half of the lunar produces a posteriorly-shortened articulation between the radius and the scaphoid and lunar. This difference, plus the lower pisiform facet on the cuneiform, suggests greater freedom for flexion at this joint in Tapirus. The lat- eral extension of the proximal surface of the scaphoid behind the proximal facet of the lunar strengthens the scapho-lunar articula- tion and also provides additional support for the radius (possibly to compensate for the shortened radio-lunar articulation). Most of the other differences listed above are either related to the modification just discussed or, like the more horizontal lunar- magnum facets, result from the greater weight of Tapirus. The manus of Frotapirus was described and figured most recently by Scott ( 1941, pp. 756-758, pi. 80, fig. 2), based on a late Oligocene specimen (AMXH 662). It is generally quite similar to the manus of Tap- irus, but slightly more primitive in the following features: scaphoid and trapezoid relatively longer anteroposteriorly; posterior process of lunar and cuneifonn relatively Fig. 14. Right innominate. Above, Heptodon posticus, MCZ MCZ:M 6037. X 'A- 17670, lateral X 'A- Below, fapirus pinchaque, Evolution of Tapiroid Skeleton • Radimky 93 higher; distal hinar-cimeiform articulation extends further posteriorly; trapezium ex- tends posterolateral!)' instead of postero- distally. The manus of Piotapinis further differs from that of Tapinis in having a rel- atively larger trapezium, a more posteriorly and more proximally located scaphoid-lunar articulation, and relatixely shorter pha- langes. On the scaphoid the posterior lunar facet is separated from the trapezoid facet by a large fossa. The carpus of Protopinis is not relati\ely longer and narrower than that of Tapims, contrary to the statements of Scott {op. cit., p. 758). A second manus of Protapinis- ( Univ. Calif. Mus. Paleo. 934), representing P. ro- biistiis, a larger species than the one Scott described, differs in no appreciable way from AMXH 662 except in larger size. The lunar, cuneiform, and pisiform are all that is known of the carpus of the early Miocene genus Miotapirus. The posterior processes of the lunar and cuneiform are as low in Miotapirus as in Tapinis, but the distal lunar-cuneiform articulation is still longer and the posterior scaphoid facet of the lunar more posteriorly and proximally located than in the modern tapir. Innominate (Fig. 14): The blade of the ilium is expanded dorsally into a relatively Fig. 15. Heptodon posticus. MCZ 17670. Right femur in, from left to right, anterior, posterior, lateral and medial views. X 'A- 94 Miisctn)} of Conifuntilivc ZooIoi;y. Vol. 134. \o. 3 Fig. 16. Tapirus pincha:,ue. MCZ:M 6037. Right femur in, from left to right, anterior, posterior, lateral and medial views. X V4- Compare with Fig. 15. narrow tuber sacrale, and extends relatively far anterolaterally to terminate in a narrow tuber eoxae. Between these two tuberosi- ties, the iliae crest is straight. The shaft of the ilium is relatively long and narrow. On the lateral side, just in front of the acetabu- lum, a prominent groove and ridge mark the origin of the lateral tendon of the rectus femoris. The ischiatic spine is located on the dorsal border a short distance posterior to the acetabulum. The posterior end of the ischium is broken off, but judging from its condition in Ilyiacliyiis and Ilelaletcs, it probably was considerably shorter than the ilium, and the tuber ischii was probably relatively weak. Pelvic girdles in which the ilium is relatively long and pubo-ischiac portion relatively short are characteristic of cursorial ungulates, and indicate an increase in importance of the gluteal muscles (which originate on the blade of the ilium) over those of the pul)()-ischiac group in extension of the femur (Smith and Savage, 1956, pp. 612-614). The innominate of Tapinis differs from that of Ih'ptodon in having a much wider, more Ncrticaily expanded, tuber coxae, and relatively larger sacral and ischial tuberos- ities. The tuber coxae serves for attachment of the external and internal obli(|uus ab- dominis muscles, the tuber sacrale for sacral ligaments medialK' and parts of the gluteus medius and longissimns dorsi on its lateral surface, while the tuber ischii serves for the origin of biceps femoris, semimembranosus, and semitendinosus. The greater promi- nence of these tuberosities in the modern tapir probably correlates with its larger size and relatively heavier body. Femur (Fig. 15): The femur of Hcpto- (I(>)i posticus measures about 205 mm long from head to distal condyles ) and 37 mm wide across the distal end. The top of the greater trochanter has been broken off in Evolution of Tapiroid Skeleton • Radinsky 95 Fig. 17. Right tibia and fibula. Above, Heptodon posticus, MCZ 17670, in, from left to right, anterior, posterior, lateral and medial views. X Vi- Below, TopiVus p;nchoque, MCZ:M 6037, in same views. X 'A- 96 Museum of Comparative Zoolop.y. Vol. 134, No. 3 MCZ 17670, but judging from its condition in Hcptodon calcicuhis and Hyruclujiis, it must have extended above the level of the head. Tlie licad is hemispherical and ex- tends anteromedially on a relatively long and thin neck. (There may have been a small amount of post-mortem anterior dis- placement of the head in MCZ 17670. ) The trochanteric ridge is fairh' prominent, and extends distalK' almost to tlie level of the lesser trochanter. The lesser trochanter is located slightly less than one-third of the \\a\' down the shaft, and the large third trochanter is situated slightly lower on the opposite side. A supracondyloid fossa, marking the origin of the plantaris, is evi- dent on the posterior side of the shaft above the condyles. At the distal end of the femur, the trochlea is lelatively narrow and almost symmetrical. Medial and lateral epicon- dyles are low. Pits for the popliteus and extensor digitalis longus are present on the lateral epicondyle. Lateral and medial con- dyles are separated by a deep intercon- dyloid fossa. The femur of Tapinis (see Fig. 16) dif- fers from that of Hcptodon mainly in being slightly more robust, with wider trochlea and condyles and more massive epicondyles ( especially the lateral one ) . The lesser and third trochanters in TapUus are located about one-third of the way down the shaft. These differences are probably accounted for mainly by the larger size of the modern tapir. Tibia (Fig. 17): The tibia of Ucpiodon po.^ticus preserved in MCZ 17670 is missing al)out one-tjuarter of the shaft and it is not possible to determine its length. In //. cal- cicidm (AMNH 294) the tibia is about as long as the femur. The proximal end of the tibia is about as deep anteroposteriorly as it is wide. The spine, or intercondyloid em- inence, is broken off. The cnemial crest is fairly prominent and bears a large depres- sion proximally for the middle patellar liga- ment. At the distal end of the shaft the medial malleolus is prominent and bears a large smooth scar from attachment of the medial ligament. Posterior to the medial malleolus is a groove for the tendon of the flexor digitalis longus. On the lateral side there is a facet for articulation with the distal end of the fibula. The articular grooves for the astragalus are relatively deep, with the medial one narrower than the lateral one. The tibia of Tapirus differs from that of Hepiodon mainly in being relatively slightly more robust, with relatively wider proximal and distal articular surfaces and a more laterally expanded tuberosity (at the head of the cnemial crest). These differences may be accounted for by the larger size of the modern tapir. The laterally expanded tuberosity provides a greater area for at- tachment of the lateral patellar ligament and the fascia lata. In Tapinis the medial intercondyloid tubercle is higher than the lateral one, the reverse of the condition in Hcptodon. The same change in relative heights of these tubercles occurred in equoid evolution (Kitts, 1956, p. 26). Increase in height of the medial tubercle would help prevent lateral dislocation of the femur relative to the tibia and might be correlated Fig. 18. Hepiodon posticus. MCZ 17670. Right tarsus and proximal portions of metatarsals. X 'A- Evolution of Tapiroid Skeleton • Rodinsky 97 Fig. 19. Right astragalus. Above, Heptodon posticus, MCZ 17670, in, from left to right, posterior, anterior, laterol and medial views. Natural size. Below, Tapirus pinchaque, YPM 204, in same views. X 'A- with the relative increase in size of the lateral epicondyles (indicating more power- ful extensor musculature) on femora of large perissodactyls. Fibula (Fig. 17): Proximal and distal ends of the fibula are present in MCZ 17670 and indicate a relatively thin but complete bone which articulated with, but was not fused to, the tibia. The proximal articular surface is elliptical in outline and horizon- tally oriented. The distal end is expanded and has a vertical groove in its posterolat- eral side for the tendon of the peroneus brevis (lateral digital extensor). Distally, the medial surface forms the lateral side of the lateral articular groove for the astrag- alus. The fibula of Tapirus is similar to that of Heptodon, differing only in having an oblique facet for the proximal articulation with the tibia. The lateral condyle of the tibia entirely covers the head of the fibula in Tapirus, while in Heptodon it is partly exposed. Tarsus (Fig. 18): Seen in anterior view the tarsus of Heptodon is relatively high and narrow. The astragalus rests on the cuboid posteriorly but anteriorly the two are separated by a small calcaneum-navic- ular articulation. The ectocuneifonn ex- tends further distally than the cuboid or mesocuneiform so that it articulates with second and fourth metatarsals as well as the third. In Heptodon posticus the trochlea of the astragalus ( see Fig. 19 ) is about as high as it is wide, and has a relatively broad, shal- low median groove. In H. calciculus the trochlea is relatively higher and narrower. The neck is relatively long and diverges slightly from the line of the trochlear crests. On the posterior (plantar) side of the astrag- alus the two main faces of the proximal calcaneal articulation are approximately perpendicular to each other. The susten- tacular facet is relatively long and slightly proximodistally convex. The distal calca- neal facet is a long, low strip which in MCZ 17670 ( H. posticus) does not, but in AMNH 294 (H. calciculus) does, contact the sus- tentacular facet. On the distal end of the astragalus the navicular facet is slightly wider than long, and is convex anteroposte- riorly and slightly concave mediolaterally. 98 Mii.sc'ttm of Comparafivc Zoolom/, Vol. 134, No. 3 Fig. 20. Right calcaneum and astragalus. Above, Heptodon posticus, MCZ 17670, m, from left to right, lateral, anterior and medial views of calcaneum and distal view of astragalus and calcaneum. Natural size. Below, Tapirus pinchaque, YPM 204, in same views. X 'A- The cuboid facet is a relatively narrow, laterodistally-facing strip and does not reach the anterior edge of the astragalus. On the lateral side of the trochlea a broad pit marks the insertion of the short lateral ligament. On the medial side a shallow proximal pit indicates the insertion of the short medial ligament, while a tuberosity at the distal end of the neck marks the at- tachment area of the dorsal ligament. The calcaneum (see Fig. 20) is relatively long and thin, with the lateral astragalar facet located slightly closer to the distal than proximal end. The tuber calcis is mediolaterally compressed and slightly ex- panded at its free end. The two major planes of the lateral astragalar facet meet at a right angle. A prominent pit just above this facet accommodated the distal end of the fibula in extreme flexion of the tibio- tarsal joint. The sustentaculum is slightly higher than wide and bears an irregularly oval, slightly concave facet for the astrag- alus. A long, low facet on the distal end of the medial side of the body articulates with the navicular anteriorly and with the astrag- Evolution of Tapiroid Skeleton • Radinsky 99 Fig. 21. Hepfodon posticus. MCZ 17670. Right distal tarsals. Top row: (from left to right) cuboid, medial view; na- vicular, ecto- and mesocuneiforms, lateral view; cuboid and na/iculor, proximal view. Middle row: distal tarsals in, from left to right, posterior, lateral, anterior and medial views. Bottom row: entocuneiform in anterolateral view and cuboid, ecto- and mesocuneiforms in distal view. All natural size. Letters indicate articular contacts. aliis for most of its length. An irregularly oval, saddle-shaped cuboid facet occupies the entire distal end of the calcaneum. A pit on the lateral side of the calcaneum at the level of the lateral astragalar facet pre- sumably marks the insertion of part of the short lateral ligament. The cuboid ( see Fig. 21 ) is relatively high and narrow, with a moderately large posterior tuberosity. The saddle-shaped calcaneal facet shares the proximal surface with a small elongate astragalar facet which does not reach the front of the cuboid. The proximal half of the medial side of the cuboid bears a small anterior facet and a large slightly concave posterior facet for the navicular. Just below the latter is a medio- distally inclined facet for the ectocuneiform. There is also a small anterior facet for the ectocuneiform. The facet for the fourth metatarsal covers the distal surface. The navicular is about as wide as it is deep ( anteroposteriorly ) with a saddle- shaped proximal surface for articulation with the astragalus. A small oblique facet at the anterolateral corner articulated with the calcaneum. The lateral side bears a large, gently convex posterior facet and a small anterior facet for the cuboid. On the posterior side there is a posterodorsally- facing facet for the entocuneiform. The distal surface bears a large triangular facet for the ectocuneiform and a small, roughly quadrangular facet for the mesocuneiform. The ectocuneiform is triangular in proxi- mal view and slighth' higher than the navic- ular. In the articulated tarsus it extends further distallv than the cuboid or meso- 100 Mu.sciiiu of Conijxiidlivc Zoo/oa;/, Vol. 134, No. 3 Fig. 22. Tapirus pinchaque. YPM 204. Right distal tarsals. Compare with Fig. 21. X 'A cuiu'ifonii and lias small distal facets on medial and lateral sides for the second and fourth metatarsals, respectively. It has a vertical anterior facet and a proximolater- ally-facing posterior facet on the lateral side for articulation with the cuboid, and a large proximal facet for the mesocuneiform on the medial side. Distally, the ectocuneiform articulates with the third metatarsal by a relatively flat facet. The mesocuneiform is small and roughly triangular in horizontal section. On its lat- eral side it has a proximal facet for the ectocuneiform and on its medial side a large proximal facet for the entocuneiform. The entocuneiform and vestigial first metatarsal were briefly described in a pre- vious paper (Radinsky, 1963b). The ento- cuneiform is roughly circular in posterior view, and relatively flat and anteroposteri- orly compressed. On its anteromedial edge it bears adjacent facets for the navicular and mesocuneiform, and a short distance distally from these, a long facet for the sec- ond metatarsal. On the anterolateral edge there is an elongate prominence which artic- ulated with the vestigial first metatarsal. Metatarsals: The vestigial first meta- tarsal is roughly oval and flat in postero- medial view, with a raised area on the lateral end of the anterior side. It bridged the gap between the lateral edge of the entocuneiform and the back of the head of the third metatarsal, articulating with the latter by a large facet. The contact with the entocuneiform left no facet. Only the proximal ends of the second, third, and fourth metatarsals are preserved. The configuration of their proximal facets complements that of the distal facets of the distal tarsals ( see Fig. 21 ) and need not be described here. The pes of Tapirus (see Figs. 19-22) differs from that of Hcptodon in the follow- ing features: astragalus relatively lower and \\'ider, lacking a flange on the distal end of the lateral side of the trochlea; astragalo-navicular and astragalo-cuboid articulations relatively shorter (anteropos- teriorly) and wider, and the latter more horizontal; proximal astragalo-calcaneal ar- ticulation shallower, with a lower proximal face; distal astragalo-calcaneal facet rela- tively higher and bent slightly; calcaneum lacking a pit above the proximal calcaneal- astragalar facet for the fibula, and with a Evolution of Tapiroid Skeleton • Radinsky 101 Fig. 23. Heptodon posficui. MCZ 17670. Skeleton in lateral view to show general proportions. Restored portions indi- caied by hatching. Slightly less than '/,, size. much shallower pit for the short lateral ligament; calcaneo-cuboid articulation rel- atively narrower; no calcaneo-navicular con- tact; cuboid relatively lower and longer ( anteroposteriorly ) ; anterior cuboid-navic- ular contact relatively larger; articulation between cuboid and fourth metatarsal more saddle-shaped and less flat; navicular rela- tively lower; ectocuneiform relatively lower and wider, without the small proximal posterior tuberosity seen in Heptodoii; ectocuneiform-mesocuneiform articulation relatively smaller; mesocuneiform relatively lower and narrower, lacking a small proxi- mal posterior tuberosity; entocuneiform more elongate. Metatarsals presumably relatively shorter and wider. The differences listed above result in a relatively lower and wider tarsus in T a pirns, presumably in correlation with its heavier weight, compared to Heptodon. The ab- sence of a pit on the calcaneum to receive the fibula in extreme flexion, and the lack of a flange on the astragalus for the fibula in extreme extension, suggests that a lesser degree of extension and flexion occurs at the tibiotarsal joint in Tapinis than it did in Heptodon. CONCLUSIONS The most striking osteological changes which occurred during fifty million years of evolution from Heptodon to Tapinis are modifications of the skull correlated with development of a proboscis. Primary changes were a tremendous enlargement of the nasal incision and a shortening of the nasals. These, in turn, caused several sec- ondary changes. There were also less drastic cranial modifications which resulted from dental and brain evolution. By the end of the Oligocene, over 25 million years ago, the evolutionary changes which resulted in the modern tapirid skull were essentially com- pleted, and there has been little cranial evo- lution in the Tapiridae since then. The postcranial skeleton of Tapinis has remained basically similar to that of Hepto- don, and most of the differences observed are correlated with the larger size of Tap- inis. These differences include relatively broader and more robust limb bones and vertebrae, a less flexible vertebral column, and a much more expanded iliac blade in Tapinis. Apparently, at some point in evolution 102 Museum of Compaidtivc Zoolo'i.y, Vol. 134. No. 3 from Heptodon to T(i))inis, there was a trend toward increasing cursorial specialization. This is indicated b\ such features in Tapiru.s as the reduction of the acromion of the scapula (correlated with loss of the clav- icle), a widened lateral condyle on the humerus, the fusion of radius and ulna, and a relatively wider and shorter radio-carpal articulation than in Heptodon. However, modern tapirs are relatively heavy-bodied and short-legged, which suggests a more recent emphasis on large size, rather than speed, for defense against predators. It is significant that the cursorial modifi- cations mentioned above are confined to the fore limb; the same is true in other tapiroid lineages descended from Heptodon. This fact suggests that the hind limb of Hepto- don was more specialized than the fore limb and had in fact approached its biomechan- ical limit of specialization for running (ex- cept for lengthening of distal limb segments in some tapiroid lineages). Thus, further modifications for running would be more likely to appear in the less specialized front limb. ABBREVIATIONS A astragalus aac anterior opening of alisphenoid canal all anterior lacerate foramen C cuneiform Ca calcaneum Cu cuboid Ec ectocimciform ef ethmoid foramen En entocuneiform feo fenestra ovalis fo foramen o\'ale fr tenestra rotunda ga groove for artery gav groove for auricular branch ol xagus get groove for chorda tympani gfn groove for facial nerve hf liypoglossal foramen iof infraorbital foramen b hinar M luagniun Me mesocuneiform nif maxillary foramen mlf middle lacerate foramen N navicidar of optic foramen ofc opening of facial canal P pisiform pac posterior opening of alisphenoid canal pgf postglenoid foramen plf posterior lacerate foramen pp paroccipital process ppf posterior palatine foramen ptp post-tympanic process R radius S scaphoid scf sinus canal foramen sf stylomastoid foramen spf sphenopalatine foramen T trapezoid Tm trapezium tt ventral edge of tegmen tympani ty tympanic U imciform Ul ulna I-V" metapodials REFERENCES CITED Bhessou, C. 1961. La myologie du tapir {Tap- inis indictis L. ). Mammalia, vol. 25, no. 3, pp. 358-400. Campbell, B. 1936. The comparative myology of the forelimb of the hippopotamus, pig and tapir. Amer. Jour. Anat., vol. 59, no. 2, pp. 201-247. . 1945. The hindfoot musculature of some basic ungulates. Jour. Mammalogy, vol. 26, pp. 421-424. Edingek, T. 1948. Evolution of the horse brain. Mem. Cleol. Soc. Amer. vol. 25, pp. 1-177. . 1950. Frontal sinus evolution (particu- larly in the E(iuidae). Bull. Mus. Comp. Zool., vol. 103, no. 8, pp. 412-496. Edingeh, T. and D. B. Kitts. 1954. The fora- men ovale. Evolution, vol. 8, no. 4, pp. 389- 404. Gregohv, VV. K. 1920. Studies in comparative myology and osteology, No. V. On the anat- omy of the preorbital fossae of Equidae and other ungulates. Bull. Amer. Mus. Nat. Hist., vol. 42, art. 3, pp. 26.5-284. . 1929. The nniscular anatomy and the restoration of (lie titanotheres. Chapter 8, in: The titanotlieres of ancient Wyoming, Dakota, and Nebraska, by H. F. Osborn. U. S. Geol. Surv., Monograph 55, pp. 1-953. Hatcher, J. B. 1896. Recent and fossil tapirs. Amer. Jour. Sci., 4th ser., vol. 1, no. 3^ pp. 161-180. Evolution of Tapiroid Skeleton • Radinsky 103 Krns, D. B. 1956. American Ilyracothcrium ( Pcrissodactyla, Equidae). Bull. Anier. Mus. Nat. Hist., vol. 110, art. 1, pp. 1-60. OSBORN, H. F. AND J. C. WoRTMAN. 1892. FoS- sil mammals of the Wahsatch and Wind River beds. Collection of 1891. Bull. Amer. Mus. Nat. Hist., vol. 4, art. 11, pp. 81-147. Radinsky, L. 1963a. Origin and early evolution of North American Tapiroidea. Bull. Yale Peabody Mus., no. 17, pp. 1-106. . 1963b. The peri.ssodaetyl hallux. Amer. Mus. Novit., no. 2145, pp. 1-8. — . 1965. Early Tertiary Tapiroidea of Asia. Bull. Amer. Mus. Nat. Hist., noI. 129, art. 2, pp. 181-264. ScHAUB, S. 1928. Der Tapirschiidel von Haslen. Ein Beitrag zur Revision der Oligocanen Ta- piriden Europas. Abh. Schvveiz. Palaeont. Gesellsch., vol. 47, pp. 1-28. ScHLAiKjER, E. M. 1937. A new tapir from the Lower Miocene of Wyoming. Bull. Mus. Comp. Zool., vol. 80, pp. 231-251. Scott, W. B. 1941. Perissodactyla. In: The mammalian fauna of the White River Oligo- cene, by W. B. Scott and G. L. Jepsen. Trans. Amer. Philos. Soc, new ser., vol. 28, pt. 5, pp. 747-980. Seton, H. 1931. A new Heptodon from the Wind River of Wyoming. Proc. New En- gland Zool. Club, vol. 12, pp. 45-48. Simpson, G. G. 1952. Notes on British hyra- cotheres. Jour. Linn. Soc. London (Zool.), vol. 42, no. 284, pp. 195-206. Slijper, E. J. 1946. Comparative biologic-ana- tomical investigations on the vertebral column and spinal musculature of mammals. Verb. K. Nederl. Akad. Wetensch., Afd. Natuurk., vol. 42, no. 5, pp. 1-128. Smith, J. M. and R. J. G. Savage. 1956. Some locomotory adaptations in mammals. Jour. Linn. Soc. London (Zool.), vol. 42, no. 288, pp. 603-622. Wortman, J. L. and C. Earle. 1893. Ancestors of the tapir from the Lower Miocene of Da- kota. Bull. Amer. Mus. Nat. Hist., vol. 5, art. 11, pp. 1.59-180. (Received 26 October 1964.) 104 Museum of Cotuptiidtivr Zoolofnj, Vol. 134, No. 3 Plate 1. Heptodon posticus. MCZ 17670. Lateral view of s'