Norntatesamerican MUSEUM PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

Home , Adocidae, Adocus, Basilemys, Chelydropsis, Glarichelys, Kinosterninae

NorntatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2941, 60 pp., 22 figs., 3 tables June 26, 1989

The Skeletal Morphology of the Cretaceous Cryptodiran Turtle, Adocus, and the Relationships of the Trionychoidea

PETER A. MEYLAN1 AND EUGENE S. GAFFNEY2

CONTENTS Abstract ...... 2 Introduction ...... 2 Materials and Methods ...... 4 Materials ...... 4 Methods ...... I.....S5 Abbreviations ...... 6 Descriptions ofAdocus with Comparisons to Other Members of the Trionychoidea ...... 6 Skull ...... 6 Skull Roof ...... 6 Palatal Elements ...... 14 Palatoquadrate and Braincase ...... 17 Canalis Carotici Internus and Related Structures ...... 19 Lower Jaw ...... 21 Carapace ...... 24 Carapace Form ...... 24 Nuchal ...... 24 Neurals ...... 24 Costals ...... 27 Peripherals ...... 27 Suprapygals ...... 28 Carapacial Scutes ...... 29 Internal Shell Morphology ...... 29 Plastron ...... 30 Plastral Bones ...... 30 Plastral Scales ...... 32 Inframarginal Scales ...... 35 The Cervical Vertebrae ...... 38 Pectoral Girdle and Forelimb ...... 39 Pelvic Girdle ...... 41

'Research Associate, Department of Vertebrate Paleontology, American Museum of Natural History. 2 Curator, Department of Vertebrate Paleontology, American Museum of Natural History.

Discussion ...... 43 The Trionychoidea ...... 43 The Trionychoidae ...... 48 The Adocidae ...... 49 The Nanhsiungchelyidae, Peltochelys, and the Trionychia ...... 50 The Nanhsiungchelyidae ...... 51 Peltochelys and the Trionychia ...... 52 The Trionychia ...... 52 The Kinosternoidae ...... 52 The Dermatemydidae, Agomphus, and the Kinosternia ...... 54 The Dermatemydidae ...... 54 Agomphus and the Kinostermia ...... 55 The Kinosternia ...... 55 The Kinosternidae ...... 56 Acknowledgments. 56 References ...... 56

ABSTRACT The osteology of the Cretaceous turtle, Adocus, and the sister-group ofthose four taxa, the Adoci- is described, based on a nearly complete specimen dae. The epifamily Kinosternoidae includes the from the Hell Creek Formation of Fallon County, genus Hoplochelys as the sister-group to a mono- Montana. An analysis of this morphology, com- phyletic Kinosternidae; the possible sister-group bined with comparisons with all other fossil and to those two taxa, Agomphus; the sister-group to living trionychoids, provides the basis for a phy- those three taxa, a restricted Dermatemydidae (in- logenetic analysis of the superfamily Trionychoi- cluding only Baptemys and Dermatemys); and the dea, which consists of two large, monophyletic sister-group to those four taxa, Emarginachelys. groups, the epifamilies Trionychoidae and Kino- The inclusion of fossils in this study results in a sternoidae. The epifamily Trionychoidae includes hypothesis for the relationships among the living the sister families Trionychidae and Carettoche- families that is fundamentally different from that lyidae; their sister-group, Peltochelys; the sister- based on a previous study of the living families group ofthose three taxa, the Nanhsiungchelyidae; alone. INTRODUCTION During his short career, George Baur (1859- dae (his Trionychia) was "not an original, but 1898) proposed several hypotheses for rela- a highly specialized group" and that it was tionships among turtles that were novel for related via the Carettochelyidae to stauro- the day but have proved over the years to be typine and kinosternine kinostemids and well supported. In contrast to Cope (1871), dermatemydids (Baur, 1891 c). Thus he pre- Dollo (1886), Boulenger (1889), and Lydek- dicted the existence of the Trionychoidea of ker (1889), he considered Dermochelys to be Gaffney (1975, 1984), a superfamily whose a derived chelonioid (Baur, 1889a) and not recognition is supported by the work of Al- the sister-group to all living turtles (i.e., the brecht (1967, 1976), McDowell (1961), and Atheca). He challenged the ideas of Owen Meylan (1987). (1881) and Boulenger (1887b) and suggested The subject of this paper is the relation- that Meiolania was a cryptodiran turtle (Baur, ships of the members of the Trionychoidea 1889b). And he suggested that Glyptops based primarily on the study of a remarkable should be placed in the family Pleurosterni- new fossil of Adocus. Although membership dae, close to the Baenidae (Baur, 1891a, of living families in this superfamily is suf- 1891 b). All of these ideas have recently been ficiently established, there are two funda- supported by attempts to develop complete mentally different hypotheses for interrela- phylogenetic hypotheses for turtles (Gaffney tionships among them (Gaffney, 1975, 1984; 1972, 1975, 1984; Gaffney and Meylan, Meylan, 1987) and the inclusion of extinct 1988). Baur also recognized that Trionychi- forms has not been completely explored 1 989 MEYLAN AND GAFFNEY: ADOCUS 3

(Hutchison and Bramble, 1981). A nearly complete and well-preserved specimen ofthe genus Adocus provides critical new data and the incentive for a detailed review of the relationships of fossil and living Trionychoi- dea. The specimen described below (CCM 60- 15) was made available through the generos- ity ofMarshall Lambert ofthe Carter County Museum, Ekalaka, Montana. It was located and collected by his son, Brice, during the summer of 1961 from the Late Cretaceous (Lancian), Hell Creek Formation in the NW 1/4 ofSec 32 T6N R60E, Fallon County, Mon- tana (figs. 1, 2). Details of the shell morphology of this 00. specimen support its assignment to the genus Adocus of Cope (1868). The type species of Cope's Adocus is Emys beatus Leidy, 1865, based on fragmentary remains from the Cre- taceous Greensands of New Jersey. Cope (1868) gave this species a new generic name, Adocus, and suggested that it resembled Stau- rotypus and Dermatemys. He later (Cope, 1870) coined the family name Adocidae for this new genus. He included the genera Zy- goramma and Homorophus in this family, both ofwhich appear to be synonyms ofAdo- cus. Wieland (1904) provided the first thorough Fig. 1. Collecting the skeleton of Adocus sp. description of the shell of Adocus based on (Carter County Museum 60-15) in the Hell Creek Marsh's (1890) Adocus punctatus. Wieland Formation, Fallon County, Montana, 1961. Above, also recognized the affinity ofAdocus to Der- Brice Lambert exposing carapace; below, dorsal matemys, Staurotypus, and Claudius. Hay view of carapace in situ showing that the skeleton (1908a) first referred Adocus to the Derma- was found dorsal side up. (Photographs courtesy temydidae, and considered it the least mod- of Marshall Lambert, Carter County Museum) ified member of the family. Gilmore (1919), who described five new species ofAdocus from the Cretaceous-Tertiary boundary of New recognized Cope's Adocidae. Mlynarski Mexico, considered Adocus, and also Hop- (1976) included Adocus in the family Der- lochelys, to be representatives of the Der- matemydidae but recognized the subfamily matemydidae. White (1972) reviewed the Adocinae for Adocus, Basilemys, Peishan- New Jersey Cretaceous Greensand specimens emys, and Tretosternon (which he considered and considered them to represent a single a senior synonym ofPeltochelys, but see Mey- species, Adocus beatus, but did not address lan, 1988). Nessov (1977, 1986) revised the the relationships of Adocus. Hutchison and Adocidae and added to it the Asian genera Bramble (1981) included Adocus in the Der- Ferganemys and Shachemys. Narmandakh matemydidae and considered the scale pat- (1985) interpreted the evidence provided by tern of Adocus and a related form, Fergan- Khosatzky and Nessov (1977) as suggesting emys, to represent the primitive condition for that the Adocidae be recognized as a taxon the family. distinct from the Dermatemydidae. Al- Several Old World paleontologists have though Narmandakh considers the question 4 AMERICAN MUSEUM NOVITATES NO. 2941

Fig. 2. Ventral view of the Adocus sp. skeleton (CCM 60-15) during preparation. Note articulated forelimbs, cervical series, and skull. (Photograph courtesy of Marshall Lambert, Carter County Museum) incompletely resolved, he suggested that the cases we have had to use published descrip- Adocidae belong to the Testudinoidea rather tions. Our treatment ofthe systematics ofthe than the Trionychoidea. Trionychoidea extends to the generic level Gaffney and Meylan (1988) suggested a except for the clearly monophyletic families novel hypothesis ofrelationships for Adocus. Trionychidae and Carettochelyidae which This hypothesis, the subject of this paper, is have been treated elsewhere (Meylan, 1987, that Adocus is part ofan epifamily, Trionych- 1988). The specimens on which we base our oidae, which is the sister-group to the epi- concepts of the basic taxa are listed below. family Kinostemoidae, and that together these The genus Agomphus Cope, 1871, as cur- two epifamilies make up the superfamily rently recognized, may not be monophyletic. Trionychoidea. The species Agomphus alabamensis Gilmore, 1919, has features not present in the remain- MATERIALS AND METHODS ing species and some of these features are derived and shared with other genera (e.g., MATERIALS Hoplochelys and Staurotypus). The genus This study is based primarily on specimens Agomphus is therefore included here exclu- offossil and Recent turtles, although in some sive ofA. alabamensis. Our concept ofAgom- 1989 MEYLAN AND GAFFNEY: ADOCUS 5 phus is based on the nearly complete shells of 7 Staurotypus triporcatus, 2 Staurotypus ofANSP 15359 and NJSM 13753; the partial salvini, 12 Claudius angustatus, and 20 Kino- shells of AMNH 1478, 1479, and 1481; and sternon representing 12 different species. descriptions ofAgomphus tardus Wieland and The genus Adocus Cope, 1868, is included Agomphus masculinus Wieland. All of this on the basis of CCM 60-15, the well-pre- material is from the Cretaceous ofNew Jersey served specimen that is the focal point ofthis and may, in fact, represent a single species. study. Additional skull data, especially for The skull and nonshell postcrania of Agom- characters of the braincase, were taken from phus are unknown. CM 3428. Other specimens examined in- Baptemys Leidy, 1870, has as its type Bap- clude: AMNH 1204 (type of A. substrictus), temys wyomingensis Leidy, 1870, from the AMNH 2528 (type of A. pravus), AMNH Bridger Formation (ANSP 10074) which was 1844 (type ofA. lacer), AMNH 2260, USNM examined. In addition we have studied the 6539 (type ofA. annexus), USNM 8594 (type type specimens ofB.fluvatilus (AMNH 4913) of A. onerosa), USNM 8650, USNM 8649, and B. tricarinatus (AMNH 6109); skulls of USNM 8613 (type ofA. bossi), USNM 8577, DMNH 511, USNM 13437, AMNH 5967, USNM 8596 (type of A. hesperius), USNM and YPM 3754; and the shells and other post- 8593 (type of A. kirtlandius), AMNH 1844 cranial elements of USNM 5000, USNM (type ofA. lineolatus), and AMNH 1204 (type 13437, USNM 13438, AMNH 5934, AMNH ofA. substrictus). 5967, AMNH 6004, UCMP 45477, as well Ifone ignores the undiagnosable taxa, CCM as the excellent descriptions and figures in 60-15 keys out to Adocuspunctatus in the key Hay (1908a). provided by Hay (1908a) for North Ameri- Dermatemys is represented by a single liv- can species. A. punctatus is a junior synonym ing species Dermatemys mawii, for which of A. beatus, the type species of the genus there is ample material. We studied the fol- (White, 1972). The closest named species, lowing specimens: BMNH 1911.1.28.1 (two stratigraphically and geographically, is Ado- specimens), BMNH 1984-1291, MCZ 85551, cus lineolatus Cope (1874). However, the type UF 29168, USNM 51072, USNM 66666, and of this species consists of a single costal and USNM 66669. We also consulted descrip- a partial pleural, and is not diagnostic. tions and figures in Bienz (1895). Basilemys Hays, 1902, is included on the Our knowledge of Emarginachelys Whet- basis ofrestudy ofAMNH 5448, NMC 8890, stone, 1978, is based on reexamination ofthe USNM 8804, and USNM 11084; NMC 376 type of E. cretacea (KU 23488) as well as and FMNH P12008, the type of B. sinuosa Whetstone's (1978) description ofthis species Riggs; and literature accounts ofRiggs (1906), which was originally referred to the Chelydri- Hay (1908a), Langston (1956), and Estes et dae. al. (1969). Hoplochelys Hay, 1908, is included in our Nanhsiungchelys Yeh, 1966, is included on analyses on the basis of restudy of the excel- the basis of the original description, as well lent series of shells in the USNM including as examination ofthe type and only described 6549 (type ofH. bicarinata Hay, 1910), 8524, specimen ofN. wuchingensis. A second spec- 5958 (type of H. caelata Hay, 1908b), 8641, imen ofNanhsiungchelys, said to exist in the 8643, 8525, 8553 (type of H. elongata Gil- Natural History Museum of Shanghai, was more, 1919), 8608, 8609, 8527 (type of H. not available. laqueta Gilmore, 1919), 8605, and 8646. The The Carettochelyidae and Trionychidae are skull and nonshell postcrania of Hoplochelys included using data published in our previous are unknown. studies ofthe taxa (Meylan, 1985, 1987, 1988; The family Kinostemidae is included in Gaffney and Meylan, 1988). analyses of cladistic relationship as two separate genera, Staurotypus and Kinosternon. Kinosternon is considered to include Ster- METHODS notherus, following Seidel et al. (1986). Data Phylogenetic Analysis Using Parsimony for this family are based on the examination (PAUP; Swofford, 1984) was used to analyze 6 AMERICAN MUSEUM NOVITATES NO. 2941 the morphological data. The complete data DESCRIPTIONS OF ADOCUS set (table 1) was examined using various com- WITH COMPARISONS TO binations to find all equally parsimonious OTHER MEMBERS OF THE minimum-length trees. Analysis of the com- TRIONYCHOIDEA plete set of 15 terminal taxa and 48 characters SKULL was undertaken with characters 2, 22, 26, 27, 40, and 46 unordered. Use of the mulpars Discussions of the skull morphology of option and the swap = global or swap = al- Adocus are based on CCM 60-15 (figs. 3-5) ternate option indicates that there are nine and CM 3428. A thorough comparison ofthe equally parsimonious explanations for the skull of Adocus with Baptemys (based on observed character distribution (table 1). DMNH 511, USNM 13437, AMNH 5967, These shortest-length trees are 92 steps and and YPM 3754; fig. 6) and Dermatemys (fig. have a consistency index of .641. These nine 7) is made throughout this section. Compar- trees are identical except in their placement isons to other trionychoids and other cryp- ofthe genera Agomphus and Peltochelys, two todires are made where they are important shell taxa. There are six alternative solutions to understand the phylogenetic relationships for the relationships of Agomphus and four among the Trionychoidea. for Peltochelys. The choice of the solution SKULL ROOF: The prefrontal is preserved cladogram (fig. 22) was made by an a pos- in both Adocus skulls. The dorsal plate ofthis teriori weighting of characters. The solution bone roofs the fossa nasalis and forms the chosen for Peltochelys is discussed in Meylan anterior margin ofthe skull roof(fig. 4). There (1988); that forAgomphus is discussed below. is no indication ofnasal bones. The principal IfAgomphus and Peltochelys are deleted from distinction of the dorsal plate of the prefron- the analysis, PAUP produces a single-solu- tal between Adocus, Baptemys, and Derma- tion cladogram of88 steps, with a consistency temys is the orientation of the prefrontal- of .670 and a topology for the remaining taxa frontal suture. In Adocus the suture is identical to that shown in figure 22. transverse, whereas in Baptemys and Der- matemys it extends posterolaterally from the ABBREVIATIONS midline, reducing the degree of frontal ex- AMNH American Museum of Natural History posure in the orbit. In all kinosternids and ANSP Academy of Natural Sciences, Phila- some specimens of Dermatemys, this pre- delphia frontal extension reaches the postorbital and AUMP Auburn University Museum of Paleon- excludes the frontal from exposure on the tology orbital margin. The suture orientation in Der- BMNH British Museum of Natural History matemys is posterolateral in its medial half CM Carnegie Museum and transverse in its lateral half, usually al- CCM Carter County Museum lowing a small amount offrontal exposure in DMNH Denver Museum of Natural History the FMNH Field Museum of Natural History orbit. IRSNB Natural History Museum of Belgium Adocus and Baptemys also differ in the JI John Iverson (private collection to be shape ofthe external narial opening. The pre- deposited at UF) frontal of Baptemys extends anteriorly in re- KU Museum ofNatural History, University lation to the dorsal process of the maxilla, of Kansas resulting in a distinct lip, seen in many tur- MCZ Museum ofComparative Zoology, Har- tles. In Adocus, such an anterior projection vard of the prefrontal is absent and the narial NJSM New Jersey State Museum opening lies in one plane and appears straight NMC National Museum of Canada in lateral view. Anterolaterally the dorsal plate UCMP University of California Museum of of the maxilla in Adocus and Baptemys su- Paleontology UF Florida State Museum, University of tures to the prefrontal as in most turtles. Florida The ventral process ofthe prefrontal is best UNAM Mexico University seen in USNM 13437 and AMNH 5967 for USNM U.S. National Museum Baptemys and CM 3428 for Adocus. This YPM Yale Peabody Museum process forms the posterior wall of the fossa 1989 MEYLAN AND GAFFNEY: ADOCUS 7

TABLE 1 Data Matrix for Trionychoidea

..N CE,E N

1. 0 9 9 9 0 0 0 0 0 2 2 9 9 1 2. 0 9 9 9 9 1 1 1 0 2 2 9 9 2 2 1 3. 1 9 9 9 0 1 1 0 1 0 9 9 0 0

1 1 4. 0 9 9 9 9 1 1 1 1 9 9 1 1 5. 1 9 9 9 0 0 0 0 1 1 1 9 9 0 0 6. 0 9 9 9 0 0 0 0 0 1 1 9 9 0 0 7. 0 9 9 9 9 1 1 1 0 0 0 9 9 0 0 8. 0 9 0 9 9 1 1 1 0 0 0 9 9 1 0 9. 0 9 1 9 1 1 1 1 0 0 0 9 9 0 0 10. 0 9 1 9 1 1 1 1 0 0 0 9 9 0 0 11. 0 9 1 9 1 1 1 1 0 0 0 9 9 1 1 12. 0 9 1 9 1 1 1 1 0 0 0 9 9 0 0 13. 0 9 1 9 1 1 1 1 0 0 0 9 9 0 0 14. 0 9 0 9 1 1 1 1 1 1 1 9 9 1 1 15. 0 9 1 9 1 1 0 0 0 0 0 9 9 0 0 16. 0 9 9 9 0 0 0 0 0 0 0 9 9 1 1 17. 0 9 0 9 0 1 0 0 0 0 0 9 9 1 1 18. 1 9 9 9 9 1 1 1 9 0 0 9 9 0 0 19. 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 20. 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 21. 1 0 9 0 1 1 1 1 0 0 1 0 0 0 1 22. 0 0 9 1 1 1 2 2 0 0 0 0 0 1 1 23. 9 9 9 9 0 0 0 0 9 9 0 9 9 1 1 24. 0 0 9 9 0 0 0 0 0 1 1 0 0 0 0 25. 0 0 9 9 1 1 1 1 0 0 0 0 0 0 0 26. 0 9 9 9 9 1 2 2 0 9 0 9 9 1 1 27. 0 0 9 0 1 1 2 2 0 0 0 0 0 1 1 28. 0 0 0 0 1 1 1 1 0 0 0 0 0 1 1 29. 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 30. 1 1 9 1 9 9 9 9 0 0 0 1 0 0 0 31. 0 0 1 0 9 9 9 9 1 1 1 1 1 1 32. 0 0 0 0 9 9 9 9 1 1 1 1 1 1 33. 1 1 1 0 9 9 9 9 9 9 9 9 9 9 34. 0 0 9 0 9 9 9 9 0 0 0 0 1 2 2 35. 0 2 9 1 9 9 9 9 1 1 2 1 2 2 36. 0 0 9 9 9 0 0 0 0 1 0 9 9 1 1 37. 0 1 1 9 9 0 0 1 0 0 1 9 9 0 0 38. 9 9 9 9 9 9 9 9 0 1 9 9 1 1 39. 1 1 9 9 9 1 1 1 0 0 0 9 9 0 0 40. 0 1 1 9 9 0 2 2 0 0 0 9 9 0 0 41. 0 0 9 1 1 1 0 0 0 0 0 0 0 0 0 42. 0 9 1 9 1 1 1 1 0 0 1 9 9 1 1 43. 1 9 9 9 9 1 1 1 9 0 0 9 9 0 0 44. 1 9 9 9 1 1 1 1 0 0 0 9 9 0 1 45. 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 46. 0 0 0 0 1 1 2 2 0 0 0 0 0 0 0 47. 2 1 9 0 9 9 9 9 0 0 0 0 0 0 0 48. 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 8 AMERICAN MUSEUM NOVITATES NO. 2941

I and Dermatemys, except for the orientation 4b: of the frontal-prefrontal suture described above. The sulcus olfactorius is relatively wide and delimited by thick frontal ridges in all three genera. The frontal of kinosternids is distinctly smaller in comparison to Adocus, Baptemys, and Dermatemys. Its small size is apparently correlated with the large size and posterior extension of the prefrontal in kinosternids. In these three genera, the frontal sends a process anteriorly along the midline below the prefrontals to roof the fissura ethmoidalis. This process is reduced or absent in most kinosternids. The frontal morphology ofthese three genera is similar in ventral view to that figured for Portlandemys (see fig. 5 in Gaffney, 1976), except that the sulcus olfactorius is proportionately wider. The parietal is preserved in both Adocus skulls. The relations and morphology of the dorsal plate of the parietal are very similar in Adocus, Baptemys, and Dermatemys. The parietal has a transverse suture with the frontal anteromedially, and a short suture with the postorbital anterolaterally. The dorsal portion of the parietal that covers the ad- ductor musculature in the temporal fossa in chelonioids and baenids is absent in Adocus, Fig. 3. Ventral view ofAdocus sp. skull (CCM Baptemys, and Dermatemys, as well as in kin- 60-15). Both cornu branchiale II and the lower osternids. Adocus has a rounded ridge that jaw are shown in situ before preparation. (Pho- marks the edge ofthe temporal emargination. tograph courtesy of Marshall Lambert, Carter This ridge is absent in Baptemys and Der- County Museum) matemys. Posteriorly, the parietal overlaps the supraoccipital and posterolaterally contacts the prootic. Adocus differs from Bap- nasalis and the lateral margin of the fissura temys and Dermatemys, but is like the Tri- ethmoidalis. The fissura is completely pre- onychia, in possessing a lateral process ofthe served only in one skull ofBaptemys (AMNH parietal that supports the medial third of the 5967) and is partially broken in both Adocus processus trochlearis oticum (fig. 4B). In Bap- skulls. Dermatemys has a distinct fissura eth- temys and Dermatemys the parietal does not moidalis that is nearly circular in anterior bear a significant portion of the processus view with a narrow ventral extension merg- trochlearis oticum and there is no lateral pro- ing with the sulcus vomeri. In Baptemys the cess. dorsal portion of the fissura is semicircular; The processus inferior parietalis forms the the ventral limits are wide and merge grad- posterior margin ofthe foramen interorbitale ually with the sulcus vomeri. The fissura eth- and the side wall of the braincase. Its lower moidalis in the Adocus specimens, although limits reach the foramen nervi trigemini and incompletely preserved, appears to be similar form complex contacts with the palatine and to that of Baptemys, but wider. prootic. The descending process of the pari- The frontal is preserved in both skulls of etal in Adocus, Baptemys, and Dermatemys Adocus. The contacts and general relations of contacts a well-developed dorsal process of the frontal are the same in Adocus, Baptemys, the palatine anteriorly, and forms a long, un- 1 989 MEYLAN AND GAFFNEY: ADOCUS 9

A pa so ir In

foramen orbito-nasale / \ \qu I-/I mx / incisura columella auris pmx columella auris

C processus trochlearis oticum

juc

foramen stapedio- temporalis 0 1 2 cms

Fig. 4. Partially restored skull ofAdocus sp. in (A) lateral, (B) dorsal, and (C) posterior views. Figure is based primarily on CCM 60-15 with additions from CM 3428. Abbreviations are: bo, basioccipital; bs, basisphenoid; ex, exoccipital; fr, frontal; ju, jugal; mx, maxilla; op, opisthotic; pa, parietal; pal, palatine; pf, prefrontal; pmx, premaxilla; po, postorbital; pr, prootic; pt, pterygoid; qj, quadratojugal; qu, quadrate; so, supraoccipital; sq, squamosal; vo, vomer. even suture with the epipterygoid posteriorly. edge of the foramen. About half of the Der- In Dermatemys the parietal forms the dorsal matemys specimens examined also have a limits of the foramen nervi trigemini and ventral parietal process along the postero- sends a process ventrally along the anterior dorsal margin of the foramen. This dorsal 10 AMERICAN MUSEUM NOVITATES NO. 2941

Fig. 5. Skull ofAdocus (CCM 60-15) in ventral view. Figure is based primarily on CCM 60-15 with additions from CM 3428. process is definitely absent in Baptemys YPM reach the postorbital and excludes the frontal 3754 and Adocus CCM 60-15. The antero- from exposure on the orbital margin. In the ventral margin ofthe foramen nervi trigemini Trionychia this frontoparietal suture is nearly is poorly preserved in all specimens, but the transverse or trends posteromedially. The epipterygoid appears to be the element that orientation ofthis suture in Nanhsiungchelys forms the anteroventral margin, and the pa- is not clear. rietal does not seem to have a ventral process The postorbital is not entirely preserved in in either Adocus or Baptemys. any of the Adocus specimens, but CCM 60- Further variation of the parietal among 15 provides some useful information. The trionychoids involves the orientation of the postorbital lies posterodorsal to the orbit and frontoparietal suture described above. In kin- forms part of the orbital margin and part of ostemids (and some specimens of Derma- the temporal margin in Adocus, Baptemys, temys) the prefrontal extends posteriorly to and Dermatemys. In all three genera its con- 1 989 MEYLAN AND GAFFNEY: ADOCUS I1I

foramen praepalatinum commissual ridge foramen orbito-nasale maxillary 'tooth' foramen palatinum posteri us

processus foramen pterygoideus Pt externus

sq fenestra postotica op Iu foramen, posterior canalis carotici interni

Fig. 5-continued. Key for figure 5. See figure 4 for abbreviations.

tacts are as follows: anteromedially with the ture of the Chelomacryptodira (Gaffney, frontal, posteromedially with the parietal, an- 1984; Gaffney and Meylan, 1988). teroventrally with the jugal, and posteroven- The jugal is nearly intact in CCM 60-15; trally with the quadratojugal. The size and internal features are also visible in CM 3428. relations of the postorbital agree in all three The jugal forms the posteroventral margin of genera except that in Adocus the postorbital the orbit and the extent of exposure in the extends posteriorly above the quadratojugal orbit is about the same in Adocus as in Bap- more than in Baptemys or Dermatemys. It is temys and Dermatemys. The jugal is also ex- possible that the postorbital extends poste- posed ventrally where it forms part of the riorly to contact the squamosal, although the cheek emargination, but its shape varies more likely condition, in which the quadra- among the three genera. Cheek emargination tojugal is exposed on the temporal margin, is most extensive in Adocus. The anterior part is used in figure 4A. The absence of postor- of the emargination forms a pocket in lateral bital-squamosal contact is a diagnostic fea- view that is absent in the other two genera. 12 AMERICAN MUSEUM NOVITATES NO. 2941

Fig. 6. Partially restored skull of Baptemys in ventral view. Figure is based on YPM 3758, with reference to DMNH 51 1.

The zygomatic arch is also thinnest in Ado- nachelys, and Baptemys have cheek emargi- cus, but this is reflected more in the shape of nation that reaches anteriorly to the level of the quadratojugal. The extent ofcheek emar- the orbit. gination is similar in Baptemys but is not as The medial process of the jugal extends on great posteriorly. In Dermatemys the cheek top of the maxilla and reaches the pterygoid emargination is least extensive partly due to just anterior to the processus pterygoideus the ventral extension of the jugal bone. The externus. The medial process forms a dorsal posterior margin of the jugal in Dermatemys ridge marking the posterior margin ofthe fos- has a much greater contact area with the sa orbitalis. This ridge is large and well de- quadratojugal than in Adocus or Baptemys. veloped in Adocus, somewhat less developed The anterior contact with the maxilla is the in Baptemys, and distinctly smaller in Der- same in the three genera. matemys. Although the development of this Cheek emargination is present in chelyd- ridge is quite variable among turtles, kino- rids, testudinoids, and primitive chelonioids. sternids have unusually large ones, even larg- Among trionychoids only Adocus, Emargi- er than in Adocus. 1989 MEYLAN AND GAFFNEY: ADOCUS 13

foramen praepalatinum aperatura narium interna commi ssual ridge foramen orbito-nasale maxillary ' tooth"

pal iu ~~~~~~~~palatinum posterius

pa-- Pt ~~~~~pterygoideus~~~~~~~processus pr ~~~externus

roraoramen

fenestra postotica qu~~~~~4

posterior canalis carotici interni _

Fig. 6-continued. Key for figure 6. See figure 4 for abbreviations.

The quadratojugal is preserved in both mosal is a cone-shaped bone attached to the Adocus skulls. It is a flat bone having a posterodorsal portion of the quadrate and C-shaped contact with the quadrate poste- forms the antrum postoticum. As in most riorly and meeting the postorbital antero- turtles, it has contacts anteriorly with the dorsally and the jugal anteroventrally. Der- quadratojugal and posteromedially with the matemys-Baptemys-Adocus form a series in opisthotic. It lacks any contact with the post- which the quadratojugal decreases in height orbital and does not participate in the skull along the cheek. Adocus has an anterior pro- roof as seen in some chelydrids, chelonioids, jection of the quadratojugal that is absent in and baenids. The squamosals of Derma- both Baptemys and Dermatemys. Caret- temys-Baptemys-Adocus form a sequence in tochelys and members of the Kinostemidae which the posterior crest of the squamosal is are unique in having contact between the progressively larger. This crest is very well maxilla and the quadratojugal. In a skull of developed in Adocus. Although not as elon- either Anosteira or Pseudanosteira (FMNH gate as in the Trionychia or Nanhsiungchelys, PR 966), there is no such contact. the condition could be considered to be de- Significant portions of the squamosal are rived relative to that of Baptemys and Der- preserved in both Adocus skulls. The squa- matemys. Claudius and Staurotypus also have 14 AMERICAN MUSEUM NOVITATES NO. 2941

Fig. 7. Skull ofDermatemys in ventral view. Figure is based primarily on USNM 66666 with additions from USNM 66669. somewhat elongate squamosals. The pres- among the Trionychoidea, scute sulci are ence of such a crest is common in baenids, present only in Nanhsiungchelys. All other but the extent of it is greater in Adocus than fossil and living trionychoids lack scutes on in baenids. the skull roof. Although the skull roofofAdo- The dermal roofing bones of Adocus, like cus and most trionychoids is unsculptured, those of Dermatemys and Baptemys, are un- there are two groups oftrionychoids in which sculptured. The external surfaces ofall ofthese the skull roofis distinctly sculptured, the Ca- elements are nearly smooth. Unlike those in rettochelyidae and Nanhsiungchelys. This is Dermatemys and Baptemys, the skull roof of a derived condition. Adocus was clearly covered by discrete scales PALATAL ELEMENTS: The premaxilla is pre- (fig. 4B). The presence of scales covering the served in both Adocus specimens. Laterally, dermal roofing elements is primitive for tur- the premaxilla meets the maxilla to form the tles. They are present in pleurodires, baenids, triturating surfaces; medially, the premaxil- chelonioids, and in Meiolania. Elsewhere lae meet on the midline; and posteriorly, the 1989 MEYLAN AND GAFFNEY: ADOCUS 15

commissual ridge , pmx

foramen praepalatinum ,aperatura narium interna mx // maxillary tooth' ,foramen palatinum posterius

VI pa externus

(qu

fenestra postotica sq

S( foramen posterior canalis E |, carotici interni

Fig. 7-continued. Key for figur'e 7. See figure 4 for abbreviations. premaxilla meets the vomer. In Dermatemys, tochelyids are unique among the Trionychoi- but not in Adocus or Baptemys, the maxillae dea and among cryptodires in having the meet dorsal to the premaxillae, preventing premaxillae fused (Meylan, 1987). entry ofthe premaxillae into the narial open- The ventral portion ofthe premaxilla forms ing. The premaxilla in anterior view is much the anteriormost portion of the labial ridge higher in Dermatemys than in Baptemys or and triturating surface. All three genera have Adocus. In Dermatemys this dorsal extension a high and well-developed labial ridge with forms an anterior wall for the ventral portion a slight concavity posterior to it on the tri- ofthe fossa nasalis that is absent in Baptemys turating surface (figs. 5-7). The posterior con- and Adocus. Because of its wide occurrence tact of the premaxilla with the vomer is of in other turtles, the Baptemys-Adocus con- the usual cryptodire pattern in Adocus and dition is hypothesized to be primitive. Baptemys, but in Dermatemys this contact is The paired condition ofthe premaxillae of considerably narrowed transversely due to Adocus, Baptemys, and Dermatemys is typ- maxillary expansion. A sagittal view of the ical of most turtles. Trionychoids and caret- premaxilla in Dermatemys is provided in 16 AMERICAN MUSEUM NOVITATES NO. 2941

Gaffney (1979, fig. 73) and shows the median anteromedially, separating the basisphenoid concavity followed by a distinct thickening from the palatines. This condition occurs in ofthe premaxilla (as well as the vomer). This all trionychoids except members of the thickening also occurs in Baptemys (AMNH Trionychia and possibly Nanhsiungchelys. 5967 and YPM 3754) in which it is not quite The palatine is preserved in both Adocus as well developed. This structure is only ru- specimens. The anterolateral part ofthe bone dimentary in Adocus where the premaxilla in forms part of the triturating surface and en- this region is relatively thin. This thickening closes a small foramen palatinum posterius. is termed the commissural ridge and is dis- In Baptemys and Dermatemys, the palatine cussed below. In Adocus there is no foramen is less extensive and does not encroach onto intermaxillaris, an opening in the palate be- the triturating surface. Anteromedially this tween the premaxillae and vomer found in element forms a complete roofto the apertura carettochelyids, trionychids, and many spec- narium internum; it completely separates this imens of staurotypine kinosternids. structure from the fossa orbitalis. The pala- The maxilla is preserved in both Adocus tine does not completely floor the apertura specimens. The contacts and general struc- narium internum in all trionychoids. In ture ofthe maxilla conform to the generalized Nanhsiungchelys, as in the Trionychia, the pattern seen in most cryptodires that lack sec- palatines are truncate anteriorly. Participa- ondary palates. The maxilla may be divided tion by the palatine in the triturating surface into two plates: a dorsal-vertical one and a appears to be autapomorphic for Adocus as ventral-horizontal one. The dorsal-vertical it does not occur to this extent in any other plate is exposed on the lateral surface of the trionychoids (except in the secondary palate skull and articulates with the premaxilla an- of some kinosternids). However, contribu- teriorly and the prefrontal dorsally. This dor- tions ofthe palatine to the triturating surface sal process of the maxilla is much wider in are common in other turtle groups. Dermatemys than in Adocus and Baptemys The two Adocus skulls have large dorsal (AMNH 5867) and is correlated with the rel- processes of the palatines that form the an- atively large choanal opening and large ven- teroventral portion ofthe cavum cranii. This tral area of the fossa nasalis in Dermatemys. process forms part of the anterior margin of One specimen ofBaptemys (DMNH 51 1) has the lateral wall ofthe braincase, and is tallest a broad dorsal process of the maxilla and is just posterior to the margin of the foramen similar to Dermatemys in this respect. interorbitale. The processus inferior parie- The vomer is nearly complete but cracked talis forms the rest of the cavum cranii wall. in both specimens ofAdocus. It has the gen- The dorsal process of the palatine in Adocus eralized cryptodiran morphology as seen in is best seen in CM 3428 but is also preserved Chelydra. It is a long element extending pos- on the right side of CCM 60-15. It is larger teriorly from the premaxillae and maxillae, than in Baptemys or Dermatemys but smaller separating the palatines, to touch the ptery- than in kinosternids or Trionychia. The pro- goid. In Adocus the ventral ridge ofthe vomer cess occurs in all known Trionychoidea and runs for nearly the entire length of the bone appears to be a synapomorphy for this group. whereas in dermatemydids there is an ante- A similar but less extensive dorsal spine also rior ridge only. occurs in some broad-jawed batagurines and In having posterior contact to the ptery- some Pseudemys. But these examples are here goids, the vomer of Adocus is like those of interpreted as homoplasy. most other trionychoids. It is unlike the vo- The triturating surfaces of Adocus are mers of the Trionychia, which are quite re- formed by the maxilla and premaxilla in the duced and fail to contact the pterygoids. Pre- skull and the dentary in the mandible. Adocus liminary observation of Nanhsiungchelys has a high and sharp labial ridge that is suggests that it has a reduced vomer that may marked, at least in CCM 60-15, by toothlike not reach the pterygoids. processes along the anterior portion of the The palate ofAdocus is like those of most maxilla. Baptemys also appears to have some turtles in having the pterygoids in contact undulation to the labial ridge and in Der- 1989 MEYLAN AND GAFFNEY: ADOCUS 17 matemys it is serrated, but neither has the maxilla is in contact with the vomer due to distinct projections ofAdocus. The maxillary this medial expansion ofthe anterior portion surface of the triturating area has a posterior of the triturating surface. The commissural expansion in the region ofthe maxillary-pal- ridge of Dermatemys is high and straight, as atine suture with a long lingual ridge. More opposed to the lower, cusplike commissural anteriorly the lingual ridge rises, but the width ridge of Baptemys. of the surface of the maxillary is reduced. The maxilla ofAdocus meets only the jugal Adocus has a short, but well-developed, on the cheek. This appears to be the primitive toothlike ridge or cusp in the midline of the condition for trionychoids. Contact between posterior expanded area between the labial the maxilla and quadratojugal, as found in and lingual ridges. This ridge, here referred Carettochelys (but not anosteirines, FMNH to as the maxillary tooth, is intermediate in PR 966; Gaffhey, 1979: fig. 173) and kino- height between the high labial ridge and the stemids, is considered derived. nearly absent lingual ridge and extends along PALATOQUADRATE AND BRAINCASE: The half the length of the maxillary surface. quadrate is well preserved in both Adocus In Baptemys the maxillary tooth is cren- skulls, its general shape being quite similar ulated but otherwise identical to that ofAdo- to that of other eucryptodires. The incisura cus. Baptemys differs from Adocus in having columella auris is nearly closed by a dorsal a transverse or commissural ridge just pos- process extending upward just behind the terolateral to the premaxilla-maxilla suture, stapes. In neither specimen, however, does lying at right angles to the labial ridge. The the process meet bone above it, resulting in commissural ridge is separated from the a comma-shaped opening. In Dermatemys maxillary tooth by a trough but is nearly con- and Baptemys, the incisura is more open; in tinuous with the lingual ridge. Baptemys also kinosternids it is like that in Adocus; and in differs from Adocus in having a slight medial Nanhsiungchelys and the Trionychia it is expansion ofthe triturating surface (presum- completely closed. The right quadrate is pre- ably correlated with the commissural ridge). served in a badly crushed skull of Basilemys This is most evident in the increased contact (NMC 8890). It appears to have had a closed between premaxilla and vomer. The anterior or nearly closed incisura columella auris with margin of the apertura narium interna in a strong postcolumellar ridge not unlike that Adocus is distinctly embayed and the pre- ofNanhsiungchelys. The antrum postoticum maxilla has a very limited suture with the of Adocus is developed to the extent seen in vomer. In Baptemys the embayment is re- most eucryptodires and is very similar to that duced by the medial expansion ofpremaxilla of Dermatemys and Baptemys. Among the and maxilla. Trionychoidea, the curved shape of the in- Dermatemys is apparently more derived cisura columella auris is unique. than Baptemys orAdocus in having a division The processus trochlearis oticum ofAdocus ofthe maxillary tooth so that two short ridges is unusually large and well developed, com- appear in this position. The groove separat- parable in extent to that seen in the Tri- ing the two maxillary ridges does not reach onychia. In Adocus the processus is formed the level ofthe grooves delimiting these ridges laterally by the quadrate and more medially from the labial and lingual ridges, respec- by the the usual condi- tively. This substantiates the notion that the prootic, cryptodiran two maxillary ridges of Dermatemys are the tion. However, medial to the prootic the pa- homolog ofthe single maxillary tooth in Bap- rietal sends a process laterally to form the temys and Adocus. Dermatemys also has an medial third of the processus trochlearis oti- expanded anterior portion of the maxillary cum. The processus in Adocus is troughlike, and premaxillary triturating surface, corre- concave dorsally, and protrudes into the ad- lated with a commissural ridge that is nearly ductor fossa to a greater extent than in der- twice the length ofthat in Baptemys. In Bap- matemydids, kinosternids, and other eucryp- temys the maxilla does not extend medially todires, except the Trionychia. Carettochelys to meet the vomer, but in Dermatemys the is similar to Adocus in the shape and extent 18 AMERICAN MUSEUM NOVITATES NO. 2941 of the processus and the distinct lateral pa- in other specimens. Some kinosternids ex- rietal process forming part of the structure. amined also have this process (Staurotypus, Contribution of the parietal to the proces- BMNH 1879.1.7.5; Claudius angustatus, UF sus trochlearis oticum is insignificant or, more 57909, AUMP 231; and Kinosternon scor- typically, absent in most cryptodires. The piodes, UF 583917) but others do not (Kino- large contribution made by the parietal in sternon herrerai, UF 57916; K. integrum, UF Adocus, about one-third ofthe total, is there- 83-51-95). fore of interest. Among other trionychoids, The occipital bones are preserved in both only the trionychids and carettochelyids have Adocus specimens and are similar to those of a similar parietal morphology. The kinoster- Dermatemys and Baptemys. The crista su- nids, Staurotypus and Claudius, are like Der- praoccipitalis lacks the horizontal portion matemys, Baptemys, chelonioids, and che- seen in Trionychia. The basioccipital ofAdo- lydrids in having no parietal contribution at cus differs from those of Dermatemys, Bap- all. In the Kinosterninae and Emarginachelys temys, and other trionychoids in having a there is minimal contribution of the parietal basis tuberculi basalis. In all other trionych- to the processus trochlearis oticum. The pro- oids this structure is absent. cessus ofNanhsiungchelys remains unknown. CANALIS CAROTICUS INTERNUS AND RE- The opisthotic is complete in CCM 60-15 LATED STRUCTURES: The systematic utility of but only visible externally and through the the basicranial arteries in phylogenetic work fenestra postotica. In CM 3428 it is also vis- with Dermatemys and Baptemys was ex- ible through the cavum cranii but is not well pounded by McDowell (1961) who argued preserved. The opisthotic ofAdocus is similar that the loss or reduction of the stapedial ar- to that in Dermatemys. In Dermatemys and tery in Dermatemys, Baptemys, and kino- Baptemys, the processus interfenestralis is sternids suggested common ancestry. Al- visible ventrally and laterally to a greater ex- brecht (1967, 1976) elaborated this idea and tent than in Adocus and Trionychia due to presented more detailed work on a broader the less extensive ossification ofthe pterygoid array of taxa, including trionychids. Gaffney in this area. The fenestra postotica in Adocus (1975) interpreted this information in the fol- is open laterally rather than closed as in near- lowing way: the primitive chelonian arterial ly all other trionychoids, except members of pattern is that seen in pleurodires and tes- the Trionychidae. Baptemys is another pos- tudinoids. In these taxa the stapedial artery sible exception; in the only specimen pre- is large and well developed relative to the serving this area, YPM 3754, the fenestra is more medial branches of the carotid, the ce- nearly closed on the right side but this may rebral carotid, and the palatine artery. A de- be the result of damage. rived condition exists within the Trionych- Despite the fact that the foramen nervi tri- oidea: the stapedial artery is reduced or ab- gemini is preserved on both sides of both sent and either the cerebral carotid artery (= Adocus specimens, the contacts of the ele- the pseudopalatine of Albrecht) or the pala- ments in the ethmoid region in Adocus are tine artery is distinctly enlarged. The Kino- not clear. The region is similar in Adocus, sternidae and Dermatemydidae have this last Baptemys, and Dermatemys, with the prootic pattern (fig. 8B, C) and it is a character supporting the monophyletic nature ofthis group. forming the posterodorsal margin, the pari- In Adocus the foramen stapedio-temporale etal forming the anterodorsal margin, and the is large and well developed as in most turtles pterygoid forming the ventral margin of the and is formed by the prootic and quadrate. foramen nervi trigemini. In Adocus the epi- Kinosternids usually have a minute but dis- pterygoid seems to form the anteroventral tinct foramen stapedio-temporale reflecting margin, whereas in Dermatemys (BMNH the presence ofa minute stapedial artery (Al- 1911.1.28.1, 2 skulls; UF 29168) the parietal brecht, 1967: 92). But Baptemys and Der- sends a posteroventral process between the matemys, as noted by McDowell (1961: 36), foramen and the epipterygoid. This process lack a foramen stapedio-temporale (as well appears to be present on the left side ofBap- as the canalis stapedio-temporalis) and a sta- temys, YPM 3754, but is not determinable pedial artery. Although Baptemys is extinct, 1 989 MEYLAN AND GAFFNEY: ADOCUS 19

faci faci rb rb faci fcl

cci (cl

--"

fpci

B Fig. 8. The internal carotid canals and foramina of selected trionychoid turtles. (A) Adocus sp. (based on CM 3428), (B) Baptemys sp. (based on DMNH 511), (C) Dermatemys mawii (based on USNM 66666). Abbreviations are as follows: cci, canalis caroticus internus; ccl, canalis caroticus lateralis; faci, foramen anterius canalis carotici interni; fcb, foramen caroticum basisphenoidale (= foramen carotico- pharyngeale); fcl, foramen caroticum laterale; fpci, foramen posterius canalis carotici interni; rb, trabecula of basisphenoid. we are confident that the stapedial artery was The canalis caroticus intemus extends an- absent, but the condition ofthis artery cannot teriorly within the pterygoid bone. In all the be so easily hypothesized for Adocus. A large genera being considered here, as well as in foramen stapedio-temporale also occurs in most turtles, the prootic forms a portion of trionychids which, nonetheless, have a re- the roofofthe canalis. In Adocus the foramen duced stapedial artery (Albrecht, 1967) so that caroticum basisphenoidale lies in the floor of the presence ofa large foramen stapedio-tem- the canalis caroticus internus in the suture porale is consistent with both a large and a between basisphenoid and pterygoid, near the small stapedial artery. anterior end ofthe basisphenoid (fig. 5). This The foramen posterius canalis carotici in- foramen has been enlarged somewhat during terni in Adocus, Baptemys, and Dermatemys preparation, but some of the original edge is is quite large, much larger than in plesioche- visible on the left side ofCCM 60-15 and the lyids, for example. This may indicate that left side of CM 3428. The foramen diameter even in Adocus, the canalis caroticus intemus was about 1/3 or 1/4 the diameter ofthe canalis bears the main arterial supply (as opposed to caroticus internus. The foramen occurs at the the stapedial artery). The foramen posterius point of entry into the canalis ofthe foramen canalis carotici interni in Adocus, as in Bap- pro ramo and can be seen on the lateral mar- temys and members of the Trionychia, is gin of the canalis. A groove confluent with formed entirely by the pterygoid (figs. 5, 6) the foramen pro ramo nervi vidiani extends and is slightly anterior to the posterior mar- ventrally through the foramen onto the ven- gin of the pterygoids. In contrast, Derma- tral surface of the skull, suggesting that this temys and kinostemids have the foramen foramen in the floor of the canalis caroticus posterius canalis carotici intemi at the pos- intemus contained a branch of the palatine terior limits of the pterygoid (fig. 7). The po- or vidian nerve. It is possible that this fora- sition in Baptemys is not a particularly sig- men is the foramen carotico-pharyngeale de- nificant character contradiction, because the scribed by Albrecht (1967) in Chrysemys and relative position of the foramen with regard Sternotherus. The foramen carotico-pharyn- to other structures appears to be the same in geale, however, is minute and formed only Adocus, Baptemys, and Dermatemys. The lat- by the pterygoid. We have found what ap- ter genus is characterized by less ossification pears to be this foramen in Baptemys, Der- in the fenestra postotica, resulting in the ap- matemys, Claudius, Anosteira (or Pseuda- parent difference in foramen position. nosteira), and Staurotypus, as well as in many 20 AMERICAN MUSEUM NOVITATES NO. 2941

other turtles. Albrecht described branches of presence of an enlarged canalis caroticus lat- the vidian nerve and the arteria carotico- eralis or an enlarged foramen anterior canalis pharyngealis as traversing the foramen (and carotici interni can both be considered de- canalis) carotico-pharyngeale. rived. With some misgivings, we tentatively Dermatemys and kinostemids have a cana- homologize the foramen caroticum basi- lis caroticus lateralis (and foramen caroticum sphenoidale in Adocus with the foramen ca- laterale) that is at least twice the diameter of rotico-pharyngeale described by Albrecht the canalis caroticus internus (fig. 8B, C). In (1967). This is primarily because they both Baptemys the canalis caroticus lateralis is vis- extend ventrally from the canalis caroticus ible only in part of the disarticulated basi- internus to the ventral surface of the skull in sphenoid in USNM 13437 and the foramen the vicinity of the foramen pro ramo nervi caroticum laterale is visible, although bro- vidiani. The differences are that the structure ken, only on the left side of DNHM 511. in Adocus is larger (possibly a primitive con- From the specimens, the size of the canalis dition) and on the pterygoid-basisphenoid caroticus lateralis and foramen caroticum lat- suture rather than within the pterygoid. The erale is at least as large as the foramen an- latter feature, however, is due to the fact that terius canalis carotici interni and quite likely the canalis caroticus internus at the position is distinctly larger, as in Dermatemys and in Adocus is formed in the basisphenoid-pter- kinostemids. In members of the Trionychia ygoid suture. Plesiochelyids (Gaffney, 1976) for which data are available, the foramen an- also seem to have a foramen in this position; terior canalis carotici interni is significantly most of the specimens do have one, but at larger than the foramen caroticum laterale. present it is difficult to determine whether or In most turtles, the medial branch of the not this is the result of breakage in the thin canalis caroticus internus goes through the floor of the canalis caroticus intemus. The basisphenoid and enters the cavum cranii at foramen posterius canalis carotici interni of the foramen anterius canalis carotici intemi. baenids would appear to be at the position These foramina in Adocus are quite close to- ofthe foramen caroticum basisphenoidale in gether and posterior to the anterior edge of Adocus because the foramen pro ramo nervi the dorsum sellae. In Baptemys and kino- vidiani in baenids exits from the skull at the sternids, the foramina are also closer togeth- foramen posterius canalis carotici intemi. er, but are just under the anterior edge of the Therefore, there is some basis for arguing that dorsum sellae. In Dermatemys the foramina the presence of this foramen in Adocus is are farther apart and slightly anterior to the primitive for eucryptodires. anterior edge of the dorsum sellae. The dor- Just anterior to the foramen caroticum ba- sum sellae is low in all ofthese taxa and there sisphenoidale in Adocus, the canalis caroticus is no sign ofthe condition seen in chelonioids internus divides to form the canalis caroticus in which a distinct area of the bone is de- lateralis. The more medial canal is still termed veloped between the dorsum sellae and the the canalis caroticus intemus (fig. 8). This foramina anterius canalis carotici intemi. point of division is apparently the same in The rostrum basisphenoidale and sella tur- Baptemys, Dermatemys, and kinostemids. cica are similar in the taxa being dealt with The canalis caroticus lateralis in Adocus is here and comparable to that seen in Testu- about the same diameter as the more medial dinoidea. Adocus has a short rostrum with branch of the canalis caroticus internus that each trabeculum having a sharp ridge (fig. enters the basisphenoid (fig. 8A). This ap- 8A), whereas in Baptemys and kinosternids pears to be the primitive condition for turtles. the trabeculae are long (fig. 8B), fully forming It is present in pleurodires, chelydrids, prim- the sella turcica but with no ridges. The sella itive chelonioids, and testudinoids. In other turcica of Dermatemys, however, has short trionychoids, the canalis caroticus lateralis trabeculae (fig. 8C). In general, the structure may be distinctly smaller (in carettochelyids of the sella turcica, dorsum sellae, and fora- and trionychids) or distinctly larger (in Der- men anterius canalis carotici intemi agrees matemys, Kinosternidae) than the medial with the presumed primitive condition ofeu- branch of the canalis caroticus intemus. The cryptodires. 1989 MEYLAN AND GAFFNEY: ADOCUS 21

The pterygoid of Adocus has a well-devel- turcica, and no sign of a small foramen caro- oped processus pterygoideus extemus (fig. 5). ticum laterale on either side of the rostrum It extends laterally well beyond the palate. basisphenoidale anteriorly as indicated by The same condition occurs in Baptemys, Der- Whetstone (1978: fig. 12). The foramen caro- matemys, and Emarginachelys. In all other ticum laterale cannot be seen, probably as a trionychoids this process is weakly developed result ofpostmortem damage, but the general or absent. The presence of this structure in relations of the braincase sidewall and pro- most testudinoids, primitive chelonioids, cessus clinoideus on the better preserved right chelydrids, plesiochelyids, and many other side indicate enough room for a large fora- higher taxa suggests that its occurrence among men caroticum laterale opening ventrolat- trionychoids is plesiomorphic. The absence erally (and not visible dorsally) as in Stau- ofthis structure in certain trionychoid genera rotypus, or more posteriorly, as in Baptemys. is considered derived. Unfortunately, the sole specimen of Emar- At this point it may be useful to comment ginachelys does not allow definite determi- on the braincase of Emarginachelys because nation ofthe relative sizes ofthe canalis caro- there are some errors in the literature. Whet- ticus lateralis and canalis carotici intemi, stone (1978) described the only known skull however, a large canalis caroticus lateralis is as a chelydrid. Based on cranial and postcra- probable. nial synapomorphies, however, we consider it to be a trionychoid (see discussion of monophyly of the Kinosternoidae). Whet- LOWER JAW stone (1978: 552) figured a restored dorsal view ofthe braincase but further preparation The lower jaws of Adocus, Baptemys, and and examination of this area makes some of Dermatemys are characterized by a distinc- his restoration unlikely. The entire region is tive "pocket" (here termed the dentary pock- not well preserved and much is broken, as et) in the posterior portion of the dentary stated by Whetstone. The right side of the (figs. 9, 10). It is bounded by an anterolat- braincase is better preserved than the left. erally directed ridge anteriorly and by the The right braincase wall, sulcus cavernosus, processus coronoideus posteriorly. The max- processus clinoideus, and foramen nervi tri- illary tooth fits into this pocket and together gemini are not seriously damaged. On the left they form a distinctive triturating surface side, however, the braincase wall has been morphology. In Baptemys and Dermatemys broken by dorsoventral compression, pro- the ridge forming the anterior margin of the ducing a flange running below the foramen dentary pocket forms a cusp where it meets nervi trigemini. This flange is pushed onto the labial ridge ofthe dentary. In Adocus this the processus clinoideus of the left side. The point is marked only by a small cusp. The actual position of the foramen cavernosum dentary pocket ofDermatemys has a low ridge is not preserved on either side, contrary to running down its midline that fits into the Whetstone's statement (1978: 551) that its trough dividing the maxillary ridges. The position is anterior and very unusual. His dentary triturating surface of Dermatemys is interpretation is based on the crushing and expanded anteriorly, matching the maxillary distortion of the left braincase wall and the expansion, and the distinctive symphyseal postmortem contact of the processus clino- pocket in Dermatemys is entirely absent in ideus and the braincase wall. The dorsal sur- Adocus and Baptemys. Dermatemys has cren- faces of the basioccipital and basisphenoid ulations and cuspules on both the mandibular are well preserved on the midline; and there and maxillary triturating surfaces that are is no evidence of a basis tuberculi basalis, as lacking in Adocus and appear only on the indicated in figure 12 of Whetstone (1978). maxillary tooth of Baptemys. The dorsum sellae and sella turcica are pre- Adocus and Baptemys have a relatively high served as described by Whetstone and there and well-developed coronoid process (figs. do appear to be trabecular fragments some 9A, 1OA), whereas Dermatemys has a very distance anterior to the sella. There is no in- low one (fig. 1OB), a condition probably au- dication of the lateral margins of the sella tapomorphic for Dermatemys. However, the 22 AMERICAN MUSEUM NOVITATES NO. 2941

processus coronoideus car processus coronoideus

pra art

den

ang foramen nerv auiuoeprls7 Nn foramen nervi auriculotemporalis

foramen intermandibularis foramen intermandibularis foramen intermandibularis oral is caudalis caudalis Fig. 9. Lower jaws of (A) Adocus sp. (CCM 60-15) and (B) Amyda cartilaginea (RH 129) in dorsal (top), lateral (middle), and medial (bottom) views. The area articularis mandibularis is shaded; the section through the symphysis (medial view) is hatched. Abbreviations are as follows: ang, angular; art, articular; cor, coronoid; den, dentary; pra, prearticular; sur, surangular. coronoid ofAdocus reaches its greatest height the fossa meckelii which is particularly small in the middle of the lower jaw. This is unlike in Baptemys, in contrast to the more normal the condition in Baptemys and Dermatemys proportions in Adocus and Dermatemys. The in which the processus coronoideus is more whole postcoronoid portion of the jaw in posteriorly located. The condition in Adocus Baptemys is shortened relative to that in Ado- is quite similar to that in trionychids (fig. 9B) cus and Dermatemys, the fossa meckelii is and carettochelyids (fig. 174 in Gaffney, small, and the area articularis mandibularis 1979). A tall processus coronoideus, located tilts posteriorly so that it faces more poste- at the middle ofthe lowerjaw, does not occur riorly than vertically. in dermatemydids, kinosternids, chelydrids, A retroarticular process that is about half chelonioids, or testudinoids. the length ofthe area articularis mandibularis Posterior to the processus coronoideus is extends posteriorly from the lowerjaw in both 1 989 MEYLAN AND GAFFNEY: ADOCUS 23

A B den dentary

cor

-pra

-art

processus coronoideus cor pra \ -I 2 0 art

den sur

foramen nervi auriculotemporalis foramen nervi auriculotemporalisan

car den art 0 pra

foramen intermandibularis oralis foramen intermandibularis caudalis foramen intermandibularis caudalis Fig. 10. Lower jaws of (A) Baptemys sp. (YPM 3754) and (B) Dermatemys mawii (USNM 66669) in dorsal (top), lateral (middle), and medial (bottom) views. The area articularis mandibularis is shaded; the section through the symphysis (medial view) is hatched. See figure 9 for abbreviations.

Adocus specimens. It is composed entirely of This is typical for members ofthe Trionychia the articular (fig. 9A). In two specimens of in which the surangular comprises more than Baptemys (DMNH 511 and YPM 3754), the halfofthis area (Meylan, 1987). In most oth- articular extends ventrally from the area ar- er turtles, except for some chelonioids, the ticularis mandibularis but does not form a surangular makes up little or none ofthe area retroarticular process. The same is true of articularis mandibularis. lowerjaws ofDermatemys. Among other tur- In both Adocus specimens the foramen ner- tles, only members ofthe Trionychia (fig. 9B) vi auriculotemporalis is divided up into sev- and Eubaena (Gaffney, 1982) have retroar- eral small foramina. In Dermatemys and ticular processes as well developed as those Baptemys, the foramen is a single large lateral of Adocus. opening. However, this morphology is highly The area articularis mandibularis of Ado- variable within higher taxa [both morphol- cus has a large contribution by the surangular. ogies are present in some species of the Tri- 24 AMERICAN MUSEUM NOVITATES NO. 2941 onychidae (Meylan, 1987)] and may not be features that are important in determining of systematic value. the relationships ofAdocus, its general shape The presence of the maxillary tooth/den- is unmodified. It is low and wide and without tary pocket in Adocus, Baptemys, Derma- keels. Among trionychoids, most carettoche- temys, and also Emarginachelys is significant lyids, Zangerlia, and Baptemys have a mid- because we hypothesize this structure as syn- dorsal keel that is best developed posteriorly. apomorphic for all Trionychoidea, even In anosteirine carettochelyids, this midline though it is absent in Trionychia and Kino- crest is developed into at least one dorsal sternidae. Based on other characters (see spine. Baptemys tricarinata, all Hoplochelys, Gaffney and Meylan, 1988, and fig. 22), we Staurotypus, Xenochelys, Agomphus ala- propose Adocus as the sister taxon to the bamensis, and some kinosternines have three Trionychia, and Baptemys plus Dermatemys well-developed keels. Other members of the (the Dermatemydidae) as the sister-group to genera Baptemys (e.g., B. wyomingensis, the Kinosternidae. The maxillary tooth/den- AMNH 5967) and Kinosternon retain some tary pocket would then have to be derived evidence of all three keels but these may be independently in both groups or be a com- lost with age. It is possible that tricarination mon feature. Because of the very close sim- has appeared once in the Kinosternoidae as ilarity between the triturating surfaces in proposed by Hutchison and Bramble (1981) Adocus and Baptemys, we prefer the inter- and has been lost repeatedly or that it has pretation of this structure as a common fea- appeared on several occasions independent- ture of all trionychoids that is lost indepen- ly. dently in Trionychia and Kinosternidae. The sculpture pattern of the CCM 60-15 is typical for the genus. All Adocus-as well as Basilemys, Zangerlia, Nanhsiungchelys, CARAPACE Peltochelys, carettochelyids, and trionych- The shell of the Carter County Museum ids-have some type of wave-form or punc- Adocus specimen (CCM 60-15) is nearly tate sculpturing. In Basilemys and Zangerlia, complete (fig. 11). The carapace lacks only it is so well developed that it is best described the seventh and eighth costal bones (the term as "pock-mark" sculpturing (Mlynarski, pleural is reserved for scales), the suprapy- 1976). In the Trionychinae, wave-form, rath- gal(s), and the peripherals (the term marginal er than punctate, sculpture predominates. is reserved for scales) posterior to number 7 NUCHAL: The nuchal bone is wider than on the right and posterior to number 9 on the long and is approximately pentagonal. It ap- left. The neural series is complete through pears to have had a nuchal scale but the an- neural 6. The plastron is missing only the terior margin ofthis element is badly broken most posterolateral portions of the right and the existence of this structure is uncer- xiphiplastron. tain. Costiform processes do not appear to CARAPACE FORM: The overall shape of the be present and ventral processes ofthe nuchal carapace appears to be relatively undistorted. are absent. It is fractured across most of its surface but The presence of a cervical scale on the nu- is not badly distorted although somewhat chal bone ofturtles is the primitive condition. flattened anteriorly. It forms a low, smooth It is present in all trionychoids that retain arc in cross section and lacks midline or lat- their scales except for the genus Shachemys eral keels. If the carapace were complete, it Kuznetsov (1976) which was considered by would undoubtedly be considerably longer Nessov (1977) to be closely related to Adocus. than wide. The entire surface is finely sculp- Paired ventral processes of the nuchal occur tured with a very regular arrangement of mi- only within the Trionychoidea. They are well nute tubercles lying close enough together that known in the Carettochelyidae and also occur the pattern could also be described as rows in the extinct genus Peltochelys. of small depressions. NEURALS: Neural bones 1 through 7 form The carapace of CCM 60-15 is represen- a continuous row in CCM 60-15 (fig. 1 A). tative ofa primitive eucryptodiran in its basic Neural 1 is hexagonal with the short sides structure. Although it shows several derived facing posteriorly and contacting the second 1989 MEYLAN AND GAFFNEY: ADOCUS 25

Fig. 1 1. Shell of Adocus sp. (CCM 60-15) in (A) dorsal and (B) ventral views. Heavy lines indicate scute sulci; light lines indicate bone sutures. costal bones. Neural 2 is small and four sided; merous chelonioids (Allopleuron, Argilloche- it contacts only neurals 1 and 3 and the sec- lys, Caretta, Chelonia, Desmatochelys, Er- ond costals. Neurals 3 through 5 are hexag- quelinnesia, Eurysternum, Glarichelys, onal with short sides anterior (contacting the Lophochelys, Osteopygis, Porthochelys, previous pair of costal bones). Neural 6 is Thinochelys, and Toxochelys), plesiochelyids similar but has a diagonal suture with the (Plesiochelys, Eurysternum, and Thalasse- seventh neural. The shape of the seventh mys), some trionychoids (Dermatemys, Bap- neural and the presence ofadditional neurals temys, Agomphus, and Anosteira), and many in this specimen remain unknown. testudinoids (Chinemys, Cuora, Emys, Ech- The neural formula found in the CCM Ado- matemys, Geomyda, Kachuga, Mauremys, cus (6 4 6 6 6 6 ?) is uncommon among eu- Ocadia, Pseudemys, Sakya, Stylemys, and cryptodires. It occurs in every Adocus spec- Terrapene). This neural formula has also been imen for which a neural series is known. This shown for Proganochelys based on the Jaekel neural formula also occurs in the trionychoid (1918) reconstruction, but recent work by genera Basilemys (Langston, 1956) and Zan- Gaffney (in prep.) can only confirm that neur- gerlia (Mlynarski, 1972), a series of London als 4 to 7 had contact to anterior costal bones Clay cheloniids (Owen and Bell, 1849), and (short sides anterior). a Jurassic pleurodire, Platychelys (Briim, The common occurrence of a 4 6 6 6 6 6 6 1965). In the trionychoids Basilemys and neural formula in all possible outgroups to Zangerlia the neural series is not divided pos- the Trionychoidea suggests that such a neural teriorly by costal bones meeting on the mid- formula is primitive for trionychoids and in line, as it is in Adocus. fact may be primitive for all turtles. Altera- The most common neural formula for casi- tion ofthis condition has occurred numerous chelydians is 4 6 6 6 6 6 6. A survey of neural times. At least five independent occurrences patterns in Hay (1908a) and Mlynarski (1976) can be hypothesized in the Trionychoidea. reveals this pattern in Pleurodires (Pelusios, However, these modifications have usually Phrynops, Podocnemis, Stereogenys, Taph- resulted in differing and unique new neural rosphys), pleurosternids (Glyptops), baenids patterns, and, as in the case of a 6 4 66 66 (Baena, Boremys, Chisternon, Neurankylus, formula in Adocus, Basilemys, and Zanger- Plesiobaena), chelydrids (Chelydropsis), nu- lia, are evidence ofshared common ancestry. 26 AMERICAN MUSEUM NOVITATES NO. 2941

Fig. 12. Diagrammatic medial cross sections of the shells of selected chelomacryptodiran turtles. (A) Adocus sp. (a composite reconstruction based on AMNH 1204 for external shell form and CCM 60-15, USNM 8613, and AMNH 22600 for internal contacts), (B) Dermatemys mawii, and (C) Pseudemys floridana. The ribheads have not been included in Adocus. 1 989 MEYLAN AND GAFFNEY: ADOCUS 27

Among trionychoids the primitive neural configuration is widespread. It occurs in Bap- temys, Dermatemys, Emarginachelys, Agom- phus, Hoplochelys, Kizylkumemys, Anos- teira, some Claudius, some Staurotypus, and some Carettochelys. Deviation from this primitive pattern among trionychoids includes the development of numerous posterior-facing six-sided neurals in trionychids and kinosternids. It also includes the development of at least one eight-sided neural, probably number 2, in Nanhsiungchelys, and sometimes number 2, but almost always number 7, in Staurotypus. Peltochelys has a unique 4 6 4-4 6 6 formula. The advanced anosteirine, Pseudanosteira, has a 6 4-4 8 4 6 6 formula and among specimens ofCarettoche- lys insculpta there are a variety of neural arrangements including several in which numerous anterior costal bones meet on the midline. Primitively, the neural series in turtles is complete between the nuchal and suprapygals. Adocus is like several other trionychoids in having posterior costal bones that meet on Fig. 13. Partial shell of Agomphus (NJSM the midline, interrupting the neural series. In 13753). Hatched area indicates missing bones; some species of Adocus there are no neurals dashed lines mark the limits of broken elements. in evidence posterior to number 6 (A. bossi, Scute sulci are not visible on this specimen but A. hesperius) or posterior to number 7 (KU scutes are present in Agomphus. unnumbered). In others there is what is usually identified as an isolated neural posteriorly (A. kirtlandius, A. punctatus). Costal distal parts ofcostal 6 are present on the left. bones meet on the midline posteriorly in all All preserved costal bones are subequal in carettochelyids, all trionychids, Derma- length except the first which is nearly twice temys, and all kinosternids except most Stau- as long laterally as the others. Since the plas- rotypus triporcatus (UF uncat., JI 1053, 83- tral buttresses do not reach the costal bones JI-20 1). Based on new material from the New in Adocus (fig. 12A), there is no lateral mor- Jersey State Musem (NJSM 13753; fig. 13), phology worthy of mention. As in all turtles Agomphus from the Cretaceous of New Jer- the first costal receives thoracic ribs 1 and 2. sey appears to have posterior costal bones PERIPHERALS: The peripheral bones are well completely divided by neurals (contra Hay, sutured to the costal bones. Posteriorly, the 1908a; Mlynarski, 1976). Basilemys nobilis series is incomplete, and those peripherals Hay has been reconstructed with this con- that are preserved are lacking their distal dition, but the critical area is represented by margins. In the majority ofturtles, peripheral dashed lines, suggesting that it is hypotheti- bones are present and are sutured to the cos- cal. Basilemys is therefore considered to pos- tals. This is the case in Adocus and most other sess the primitive condition as indicated by trionychoids. Peripherals are absent in most the best preserved specimen, B. sinuosa Riggs members of the Trionychidae. Within this (FMNH P12008). family, only the genus Lissemys retains them. COSTALS: Costal bones 1 through 4 are pre- In this taxon they are completely free of the served on both sides ofCCM 60-15. Number costals (see Meylan, 1987, for a discussion of 5 is complete on the left but only the proximal the homology of the peripheral bones of part is present on the right. Proximal and Lissemys to those ofother turtles). Peripheral 28 AMERICAN MUSEUM NOVITATES NO. 2941

mation (AMNH 5983). The weakness or absence of costoperipheral sutures may be re- sponsible for the frequent absence of peripheral bones from fossil carapaces of ca- rettochelyid genera that almost certainly had them (Dollo, 1884; Hummel, 1929). Outside the Trionychoidea, absence of costoperipheral sutures in adults may occur in chelydrids and is common in chelonioids. Costoperipheral sutures are present in Pro- ganochelys, pleurodires, pleurosternids, baenids, and Meiolania and thus seem likely to be primitive for the Eucryptodira. The exact number of peripherals in the CCM Adocus specimen is not known. Other complete specimens representing this genus have the primitive number for casichelydians, 11 on each side. Portions of eight are preserved on the left side ofCCM 60-15 and, as in other specimens of Adocus, number 8 is in contact with costal bone 5. Nothing suggests that peripherals 9, 10, and 11 were absent. However, certain other trionychoids do not retain the primitive number of 1 1 peripherals. All genera of the Kinosternidae (Claudius, Kinosternon, Staurotypus, and Xenochelys), the Carettochelyidae (Al- laeochelys, Anosteira, Carettochelys, Kizyl- umemys, and Pseudanosteira), and the genus Peltochelys have 10 peripheral bones on each side. All trionychids except for the genus Lissemys lack peripherals altogether. Liss- emys has a variable number but never more than nine on each side. The remaining trionychoids under consideration here have the primitive number, 11 peripheral bones per side (the condition in Nanhsiungchelys is unknown). Other described specimens ofAdocus (types of A. beatus and A. lacer) have the anterior peripheral bones very finely recurved (Hay, 1908a; White, 1972). The condition of this character is unknown in the Carter County Museum specimen due to the incomplete Fig. 14. Internal view ofthe carapace of Stau- nature of the anterior peripherals. rotypus triporcatus (83-JI-201) showing the free SUPRAPYGALS: The number of suprapygals 9th and 10th thoracic vertebrae. in the CCM Adocus is unknown. All figured specimens of Adocus, including A. bossi, A. bones appear to be free of, or very weakly hesperius, A. kirtlandius, and A. punctatus (= connected to, the costals in all carettoche- beatus), as well as all material ofAdocus ex- lyids. This is certain for the living genus Ca- amined during this study, have at least one rettochelys. However, the first peripheral may large suprapygal, and it is always in contact be sutured in Anosteira from the Bridger For- with the pygal. In some specimens, this ele- 1 989 MEYLAN AND GAFFNEY: ADOCUS 29 ment is isolated anteriorly by the eighth costal. On the right side the sulcus formed by tals meeting on the midline. But in others marginal scutes 6 and 7, at theirjunction with (including types of A. punctatus Marsh and pleural scutes 2 and 3, lies entirely on the A. kirtlandius Gilmore), a smaller midline costal bones. On the left side the common element meets this large suprapygal. In the sulcus of marginals 5 and 6 and pleurals 1 past this small midline element has been con- and 2 is sinuous, crossing the costoperipheral sidered a neural. But based on its position suture a total of five times. Marginal scutes relative to the site ofarticulation ofthe pelvic 7 through 9 have their medial sutures entirely girdle to the shell, where the thoracic verte- restricted to the costals. It is likely that the brae posterior to number 10 are free of the most posterior marginal-pleural sulci re- shell, this element would not have had a neur- mained entirely on the costal bones, as in al arch fused to it. Lacking this definitive other Adocus specimens (see pl. 1 in Wieland, character of a neural, this element must be 1904; fig. 296 in Hay, 1908a; and figs. 7, 9, considered a suprapygal. Therefore we con- and 11 in Gilmore, 1919). sider those Adocus with a small midline ele- The number of carapacial scutes on the ment in contact with a large suprapygal to CCM Adocus specimen is considered to be have two suprapygals. A similar argument the primitive number as far as it can be dis- can be made for the same small element in cerned. Primitively, the five vertebral scutes Dermatemys. of turtles were quite wide (Zangerl, 1969), Elsewhere among the Trionychoidea, a sin- and approximately equal in width. The nar- gle suprapygal is present in carettochelyids rowed vertebrals 2 through 5 in Adocus are and kinosternids. Two suprapygals are pres- derived relative to the primitive condition, ent in Basilemys (may have two or three) but the level ofgenerality ofthis derived con- (Estes et al., 1969), Zangerlia (Mlynarski, dition is difficult to determine. 1972), Peltochelys (Meylan, 1988), Agom- The unique relationship of the marginal phus (fig. 13), Baptemys and Hoplochelys scutes to the costoperipheral suture line in (Hay, 1908a), and rarely in Dermatemys. The the CCM Adocus is consistent with the con- absence ofsuprapygal elements is one ofmany dition described in the type of Marsh's Ado- unique features of the Trionychidae. cuspunctatus (Hay, 1 908a) as well as A. bossi, CARAPACIAL SCUTES: The carapace of the A. kirtlandius, and A. hesperius (Gilmore, CCM Adocus is incomplete but there is no 1919). Although in some other turtles (Pleu- evidence that anything other than 5 verte- rosternon) the pleuromarginal sulcus is lo- brals, 4 pairs ofpleurals, and 12 pairs ofmar- cated on the costal bones, in no turtles known ginal scutes were present. The anterior part to us does this sulcus move so abruptly and of the nuchal bone is badly damaged and the completely from being entirely on the pe- presence of a cervical scute could not be de- ripheral bones to entirely on the costals, as termined. The first vertebral is very wide; its in the case of Adocus. This configuration of lateral edges do not touch the nuchal bone. the marginal scutes appears to be an auta- The remaining vertebrals are quite narrow, pomorphy for the genus. about 1.5 times longer than wide. The first INTERNAL SHELL MORPHOLOGY: The inte- pleural scute is also the broadest, reaching rior surface of the carapace of CCM 60-15 is well onto the peripheral bones. The second well preserved and has most of the thoracic pleural makes some contact with the periph- vertebral centra in place. Ribheads of most erals on the left side only, the remaining pleu- costals are preserved and are narrow and ral scutes do not contact the peripheral bones. moderately developed. Plastral buttresses ap- The relationship ofthe peripheral bones to parently do not rise onto the first costal bone. the marginal scutes changes along the length No suture is visible for such a structure on of the shell (see fig. 1 A). Marginal scutes 1 the internal surface of this costal. Only parts through 4 lie entirely on the peripheral ele- of the fifth and sixth costal bones are pre- ments. The fifth marginals lie entirely on the served, and therefore, we cannot be sure if fourth peripheral anteriorly, but posteriorly there were sutures on the visceral surfaces of they extend toward the midline crossing the these elements to receive an inguinal buttress. costoperipheral suture onto the second cos- But based on other specimens ofAdocus, es- 30 AMERICAN MUSEUM NOVITATES NO. 2941 pecially AMNH 22600, USNM 8613, and centrum across a broad surface just posterior USNM 8593, we have reconstructed the shell to the cotyle. The first rib becomes quite nar- with both axillary and inguinal buttresses row laterally. It appears to have articulated failing to reach the costal bones (fig. 12A). with the rib ofthe second vertebra on the first Cope's name, Adocus, meaning "without costal bone. rafters," refers to the extreme weakness of the ribheads in the type species A. beatus. The ribheads in CCM 60-15 are very well pre- PLASTRON served and are moderately developed. There PLASTRAL BONES: The plastron ofCCM 60- is a possibility that the condition of the rib- 15 is broad, well ossified, and akinetic (fig. heads in the type ofA. beatus is due to post- 11 B). Its finely sculptured appearance is due depositional wear. Moderately to weakly de- to numerous closely spaced tubercles. The veloped ribheads which are not firmly sutured bridges are quite long (almost exactly one- to the adjacent vertebral centra seem to be halftotal plastron length), and the anal notch the common condition for eucryptodires. Al- is absent. The plastron consists ofthe typical though the ribheads ofAdocus may be slightly nine elements with the standard contacts reduced relative to those of most trionych- found in nearly all eucryptodires (mesoplas- oids, a more remarkable condition is found tra are absent). The epiplastra share a long in the Trionychidae and Carettochelyidae common suture on the midline. They contact (Meylan, 1987). In members of these two the entoplastron posteromedially and the families, the ribheads are greatly enlarged and hyoplastra posteriorly. The entoplastron is strongly sutured to the vertebral centra. large and lies between the paired epiplastra Description of the condition of the 10th and paired hyoplastra. The hyoplastra extend body vertebra in Adocus cannot be based on posteriorly to meet the hypoplastra. Laterally CCM 60-15. This portion of the carapace is they contact peripheral bones 2, 3, 4, and 5. missing. In his redescription ofAdocus punc- The hyo-hypoplastral suture lines up laterally tatus, Wieland (1904) marked a location on with the suture between the fifth and sixth his Plate 1 where the "rib of tenth dorsal peripherals. The hypoplastra extend poste- vertebra unites suturely with eighth pleural riorly to meet the paired xiphiplastra. Lat- [= costal], or else point of iliac support." His erally the hypoplastra contact peripherals 6, commitment to a complete transverse pro- 7, and 8. The hyoplastra and hypoplastra ap- cess on the 10th body vertebra seems equiv- pear to have been well sutured to the cara- ocal. However, Baur (1891 a) stated that pace. Axillary buttresses extend anteriorly, transverse processes of the 10th body ver- nearly to the level of the anterior end of the tebra do reach the eighth costal. The mem- entoplastron. Inguinal buttresses extend pos- bers of the family Kinosternidae are unique teriorly beyond the level of the hypo-xiphi- among turtles in having both the 9th and 10th plastral suture. Neither axillary nor inguinal thoracic ribs failing to reach the carapace buttresses are markedly produced dorsally. It (Meylan, 1987: fig. 14). is unlikely that they reached dorsally to con- The first body vertebra ofthe CCM Adocus tact the interior surface of the costal bones is procoelous with a wide cotyle (about three (see descriptions of costals above and Hay, times wider than tall). The articular areas of 1908a: 238). the prezygapophyses are restricted to their The xiphiplastra have ridges that lie across dorsal surface. They are wide and approach the center of their dorsal surface anterome- each other medially where they are divided dially. These would have contacted the pec- by a large U-shaped emargination of the tinial processes of the pubic bones. The in- neural arch. Posterior to this emargination is ternal surface ofthe anterior plastral lobe also a broad dorsal neural process which appears shows some relief. The epiplastra are quite to be provided with sutures for firm anchor- thickened laterally, as is the central portion age to the overlying carapacial element (prob- of the entoplastron. A median ridge appears ably the first neural). The first rib is slightly on the midline of the entoplastron with separated from the first vertebra on the right depressions lateral to it on either side. These side, but in life it would have sutured to the depressions are probably the insertion site for 1989 MEYLAN AND GAFFNEY: ADOCUS 31 the acromial ligament which fixes the acro- present in other trionychoid taxa with ante- mial process of the scapula to the plastron. rior-lobe kinesis, carettochelyids, and stau- The plastra ofnearly all other trionychoids rotypines, it forms an elongate triangle with consist of the same nine elements found in the apex anterior. Adocus. The general shapes and contacts of Hyo- and hypoplastra are quite constant in the bones of the Adocus plastron are primi- general shape and contacts among trionych- tive for turtles and nearly uniform among oids. They always form the majority of the trionychoids. Only in members ofthe family plastron although in many cases (trionych- Trionychidae are the epiplastra reduced to an ines, anosteirines, Emarginachelys, Hop- I- or J-shaped deep dermal element which lochelys, and kinosternids) these elements may or may not be covered by a superficial are somewhat reduced. They are sutured to dermal callosity. Neither the callosity nor the one another in all trionychoids and in some deeper element in trionychids is sutured to trionychids they are fused. The hyo- and hy- any other elements. The epiplastra of the poplastra are usually firmly sutured to the CCM Adocus, other examples of the genus, carapace, as is the case in the CCM and all and other trionychoids in general are not large. other Adocus, but in certain taxa these sutures Those ofEmarginachelys, Claudius, and kin- are absent and only a ligamentous connection ostemines are much longer than those ofoth- exists. This is the case for all trionychids and er trionychoids. In Emarginachelys and carettochelyids, as well as Emarginachelys Claudius, these elements extend laterally to and Claudius. the entoplastron, nearly to the bridge. In kin- The plastral buttresses fail to reach the cos- ostemines the entoplastron is absent and the tal bones in all trionychoids (fig. 1 2A) except epiplastra make up the entire anterior plastral Baptemys (AMNH 1078, 1103, 5934) and lobe. Dermatemys. In these two genera the very The epiplastra of Adocus and most tri- long axillary buttresses barely reach the first onychoids are not thickened or elaborated on costal (fig. 12B). The inguinal buttresses do the dorsal surface. Only Basilemys and Zan- not reach the costals in these genera (contra gerlia have thickened epiplastra analogous to Hay, 1 908a). Plastral buttresses extend from the condition in testudinids. the hyo- and hypoplastra to the costal bones The entoplastron in the majority of tri- ofpleurodires, baenids, and testudinoids (fig. onychoids is similar in shape to that ofAdo- 12C). cus. However, in the Trionychidae it is so The xiphiplastra of other Adocus speci- different that Williams and McDowell (1952) mens and most trionychoids are like those of questioned the homology of this unpaired CCM 60-15, being broadly rounded poste- element in the anterior lobe ofthe trionychid riorly and lacking an anal notch. Only in Der- plastron. Bramble and Carr (Ms) have shown matemys, Xenochelys, and some Kinosternon that it is in fact the entoplastron. In trionych- is an anal notch present. Several genera in- ids this element consists of a deep V-shaped cluding all of the Anosteirinae, some Bap- or boomerang-shaped element with the free temys, Hoplochelys, Emarginachelys, and the ends oriented posteriorly. In about halfofthe staurotypines have elongate xiphiplastra that living trionychid species, a circular superfi- are narrow and come to a point posteriorly. cial dermal callosity is centered on the an- The ridges noted on the dorsal surface of the teriormost part ofthe entoplastron. This cal- xiphiplastra in CCM 60-15 have also been losity and the underlying element never suture noted in other Adocus including A. agilis, A. to other plastral elements in living forms. lacer, A. syntheticus, and A. bossi, as well as However, in one fossil form the entoplastral Basilemys sinuosa and B. variolosa. They are callosity is tightly sutured to those ofthe hyo- also present in nontrionychoids such as Pleu- plastra (Hutchison, personal commun.). rosternon and Plesiochelys, and are probably The entoplastron of other trionychoids is primitive for Cryptodira. a simpler single ossification that is usually In all trionychoids except for the Trionych- round or oval. In Emarginachelys it is antero- idae, all plastral elements are sutured in a posteriorly elongate. As discussed above, this normal pattern. Sutures are kinetic in certain element is absent in kinostemines. Where ofthese taxa. In trionychids the anterior lobe 32 AMERICAN MUSEUM NOVITATES NO. 2941 is greatly reduced, and the plastral elements ment. Posteriorly and laterally, they contact do not normally suture on the midline (they only the pectorals; they do not contact any do so regularly in Plastomenus and rarely in inframarginals. Apalone and members ofthe Cyclanorbinae). The pectoral scales (set 4) are narrow. They The epiplastra never suture to the entoplas- are slightly wider medially than laterally and tron in trionychids. In only one highly de- are restricted entirely to the hyoplastra. The rived trionychid does the entoplastron suture right pectoral extends broadly across the to the hyoplastra (Hutchison, in prep.). Con- midline onto the left hyoplastron. The left tact between the hyo- and xiphiplastra is min- pectoral also overlaps the midline onto the imal in members ofthis family. The absence right hyoplastron but to a lesser extent. Lat- of sutures between so many of the elements erally each pectoral contacts the two ante- of the plastron in trionychids allows longi- riormost of four inframarginals and extends tudinal kinesis along the midline and trans- anteriorly into the axillary notch. The pec- verse kinesis along the anterior edge of the torals reach the entoplastron but do not over- hyoplastra (Bramble and Carr, MS). lap it. Anterior-lobe kinesis also occurs along the The abdominals (set 5) are the largest of anterior edge ofthe hyoplastra in living kino- the plastral scales. They cover the posterior sternids (except Claudius). As is the case in third of the hyoplastra and the anterior two- trionychids, the hyoplastra are more or less thirds of the hypoplastra, thus crossing the immobile relative to the carapace, and the hyo-hypoplastral suture. Laterally they con- entoplastron (ifpresent) and epiplastra move tact the second, third, and fourth inframar- anterior to them. Anterior-plastral-lobe ki- ginals. They reach the inguinal notches pos- nesis in the testudinoids (except Pyxis) and teriorly and therefore prevent contact of the Pelusios occurs along the posterior edge of femorals (set 6) and the inframarginals. the hyoplastra. In Claudius, plastral kinesis The femoral scales cover the posterior third occurs at the bridge (Bramble et al., 1984). ofthe hypoplastra and the anterior halfofthe In most members of the genus Kinoster- xiphiplastra. They extend across the hypo- non, the posterior plastral lobe is also kinetic. xiphiplastral suture. Anteriorly, they contact In these forms mobility occurs along a trans- only the abdominals and posteriorly, they verse suture between the hypo- and xiphi- contact only the anals (set 7). plastra. The anal scales, set 7, are restricted to the Unlike the plastron of Adocus, the plastra xiphiplastra. The femoro-anal sulcus makes of some trionychoid genera do not join the about a 600 angle to the midline. The cast carapace by a strong bony suture. This is the from which these descriptions are being made condition in members of the Trionychia and is not complete enough to allow a description in Emarginachelys (fide Whetstone, 1978) and of anal overlap onto the dorsal surface of the Claudius. xiphiplastra; similarly, descriptions of the PLASTRAL SCALES: CCM 60-15 has seven femoral, humeral, extragular, and gular scales pairs of plastral scales (fig. 11 B). Set number on the dorsal surface ofthe plastron are lack- 1, the gulars, are restricted largely to the epi- ing. The midline sulcus ofthe plastron, though plastra but do extend across the anterior not highly sinuous, does not follow the mid- quarter of the entoplastron. Lateral to the line sutures very closely. The gular scales are gulars, set 2, the extragulars, are restricted offset to the left. Both pectorals extend onto entirely to the epiplastra. They are separated the opposite hyoplastron, the right pectoral from each other by the gulars and from the remarkably so. The interabdominal seam lies entoplastron by contact of set 3, the humer- on the left hyoplastron but then crosses over als, and set 1, the gulars, on the epiplastra. to the right hypoplastron. Both interfemoral The humeral scales extend from the pos- and interanal sulci remain close to the mid- terior half of the epiplastra across the ento- line suture but undulate slightly over their plastron and onto the anterior third of the entire length. Among trionychoids, Agom- hyoplastra. Their medial contact is restricted phus, Basilemys, and Peltochelys are like entirely to the entoplastron and they exclude Adocus in having a sinuous midline sulcus. the set 4 scales, the pectorals, from that ele- Outside of the Trionychoidea this unusual 1 989 MEYLAN AND GAFFNEY: ADOCUS 33 feature is also present in some baenids and they are largely confined to the epiplastra some Compsemys. It is, therefore, considered and barely overlap onto the entoplastron in to be derived within the Trionychoidea. Basilemys, Nanhsiungchelys, Peltochelys, and All other specimens of Adocus for which Xenochelys. One-half or more of the ento- the plastron is known have seven pairs of plastron is covered by set 1 scales in Bap- plastral scales, as do all known Basilemys and temys, Dermatemys, Emarginachelys, Hop- Peltochelys. Mongolemys has six pairs of lochelys, Staurotypus, and Claudius; a smaller plastral scales due to the absence of the ex- portion (about one-third) is covered in tragulars (set 2). This genus was originally Agomphus. described as a dermatemydid (Khosatzky and Set 1 scales are normally paired as is seen Mlynarski, 1971) and has been used as a in CCM 60-15 and most other Adocus spec- model for the ancestral condition ofthe plas- imens. However, fusion of set 1 scales occurs tron in the dermatemydid-kinosternid clade among trionychoids in some Adocus (A. kirt- (Hutchison and Bramble, 1981). However, landius), Basilemys (the type ofB. nobilis Hay examination ofa well-prepared skull ofMon- and NMC 376), and in Nanhsiungchelys golemys suggests to us that this genus is not (Zangerlia unknown). a trionychoid but a testudinoid. Scale set 2 As mentioned above, set 2 scales are pres- is also missing in Nanhsiungchelys but the ent only in Peltochelys, Zangerlia, Basilemys, posterior portion ofthe plastron is unknown; and Adocus among the Trionychoidea. In all therefore a complete account of the plastral of these taxa, except some Basilemys, this scales of this genus cannot be given. pair of scales is isolated on the epiplastron The members of the Kinosternidae and by contact between the gular (set 1) and hu- Dermatemys have six pairs of plastral scales, meral (set 3) scales as described in the CCM but Hutchison and Bramble (1981) argued specimen. Some Basilemys also show this that these scales are not homologous to pairs condition (B. variolosa, B. nobilis; see Lang- 1, 3, 4, 5, 6, and 7 found in Mongolemys. ston, 1956), but others have the set 2 scale They suggested that a new pair, intergulars elongate posteriorly, meeting on the midline (set 8), has been added to a condition seen and thus isolating set 1 from set 3 scales. in Baptemys (sister-group to Dermatemys Set 3 scales are persistent in trionychoids. based on shared absence of foramen stape- They are one of the pairs that remains even dio-temporale) in which scale set 4, the pec- in the most reduced scale pattern of torals, has apparently been lost. Following Hoplochelys and the Staurotypinae. In all the model ofHutchison and Bramble (1981), cases, as in CCM 60-15, they are restricted scale sets 2 and 4 are apparently also absent largely to the hyoplastra although they always from Emarginachelys, Agomphus, Hoploche- cover part of the entoplastron. In Baptemys lys, and the Kinosternidae. In Hoplochelys, they reach the posterior edge of the epiplas- the abdominals, scale set 5, do not meet on tra. When set 4 scales are absent as in Bap- the midline, so only sets 1, 2, 3, 6, and 7 do temys, Dermatemys, Agomphus, Hoploche- meet. In the Kinosterninae a pair of scales lys, and kinostemids, set 3 scales contact the analogous, but not homologous, to the inter- inframarginals laterally. gulars of Dermatemys has been added and When present, set 4 scales, the pectorals, scale set 3 has been divided by a hinge-line lie between the humerals (set 3) and abdom- to give a total of six pairs of plastral scales. inals (set 5). They always isolate the scales of The staurotypines, Staurotypus and Clau- the anterior plastral lobe from the inframar- dius, are the trionychoids with the fewest ginals as in CCM 60-15. As in that specimen, scales on the plastron. They have only four these scales lie entirely or nearly entirely on pairs, pairs 1, 3, and 5 being completely ab- the hyoplastra. In nontrionychoids with pec- sent. torals these scales are typically well posterior In addition to reduction in number, some to the entoplastron. The close approach of trionychoids lose plastral scales altogether. the pectoral scales to the entoplastron due to This is the case in the Trionychidae and Ca- their medial expansion (as in CCM 60-15) rettochelyidae. Scale set 1, the gulars, is pres- can be considered a derived feature. In some ent in all scaled trionychoids. As in Adocus Adocus (A. lacer, A. syntheticus, A. annexus) 34 AMERICAN MUSEUM NOVITATES NO. 2941

TABLE 2 TABLE 2-(Continued) Characters and Character States Used in the Anal- ysis of Relationships Among the Members of the 22. Number of peripheral 0) 11 per side Trionychoidea bones* 1) 10 per side 2) 9 or fewer per (The primitive condition for the Chelomacryptodira, side based on outgroup considerations, is given as state 0. 23. Transverse processes of 0) yes The states of those multistate characters marked with thoracic vertebra 9 su- no an asterisk are not considered to form a transformation 1) series and were run unordered in the phylogenetic anal- tured to overlying costal yses.) 24. Axillary buttress of hyo- 0) no plastron reaches overlying 1) yes 1. Foramen stapedio-tempo- 0) large costal bones rale 1) small 25. Rib ends strongly articu- 0) no 2) absent lated to vertebral centra 1) yes 2. Size of foramen caroticum 0) FCL = FACCI 26. Ventral process of eighth 0) single laterale (FCL) relative to 1) FACCI > FCL cervical vertebra* 1) double size of foramen anterius 2) FCL > FACCI 2) absent canalis carotici interni 27. Number of suprapygals 0) two (FACCI)* 1) one 3. Foramen posterius canalis 0) no 2) none carotici intemi completely 1) yes 28. Plastral kinesis present 0) no surrounded by pterygoid along anterior edge of hyo- 1) yes 4. Basis tuberculi basalis 0) present plastra 1) absent 29. Plastron strongly sutured 0) yes 5. Maxillary "tooth" 0) absent to carapace at bridge 1) no 1) present 30. Midline plastral scute sul- 0) no 6. Commissural ridge 0) absent cus sinuous 1) yes 1) present 31. Extragular scales (set 2) 0) present 7. Premaxillae fused 0) no 1) absent 1) yes 32. Pectoral scales (set 4) 0) present 8. Foramen intermaxillaris 0) absent 1) absent 1) present 33. Pectoral scales reach ento- 0) no 9. Vomer reduced 0) no plastron 1) yes 1) yes 34. Scale set 5 (abdominals) 0) meet medially 10. Palatines truncated ante- 0) no 1) displaced laterally riorly 1) yes 2) absent 11. External process of ptery- 0) present 35. Number of inframarginal 0) 5/5 to 4/4 within goid 1) absent scales one genus 12. Basisphenoid-palatine 0) absent 1) 4/4 to 3/3 within contact 1) present one genus 13. Incisura columellae auris 0) no 2) 3/3 to 2/2 within closed 1) yes one genus 14. Scutes sulci of skull roof- 0) present 36. Notch present in ilium 0) no ing bones 1) absent just posterior to acetabu- 1) yes 15. Skull roofing bones sculp- 0) no lum tured 1) yes 37. Thelial process 0) absent 16. Frontal bones enter orbit 0) yes 1) present 1) no 38. Biconvex cervical 0) is number 4 17. Maxilla contacts quadra- 0) no 1) is number 3 tojugal 1) yes 2) is number 2 18. Retroarticular process 0) absent 39. All intercervical articula- 0) no 1) present tions opisthocoelous 1) yes 19. Three keels on carapace 0) no 40. Phalangeal formula* 0) 2-3-3-3-3 1) yes 1) 2-2-2-2-2 20. Neural formula 6 > 4 < 0) no 2) shows hyperphy- 6 < 6 < 6 < 6 1) yes langy 21. Costal bones meet on the 0) no 41. Paired ventral processes 0) absent midline 1) yes of the nuchal 1) present 1989 MEYLAN AND GAFFNEY: ADOCUS 35

TABLE 2-(Continued) the posterior portion of the inframarginal series. 42. Cheek emargination 0) yes Scale sets 6 and 7, the femorals and anals, reaches level of orbit 1) no in all scaled 43. Coronoid tall and located 0) no are present trionychoids. The near middle of mandible 1) yes femorals normally cover the hyo-xiphiplas- 44. Parietal contribution to 0) little or none tral suture and the anals are usually restricted processus trochlearis oti- 1) large to the xiphiplastra. In Hoplochelys and stau- cum rotypines, set 6 approaches or reaches the 45. Costoperipheral sutures 0) present hyo-hypoplastral suture. In one species of 1) absent Hoplochelys (H. elongata Gilmore), the anal 46. Shell scales* 0) present scales overlie the hypo-xiphiplastral suture. 1) absent from plas- In kinosternines with a mobile posterior plastron tral lobe the anterior edge of scale set 6 lines 2) absent from plas- suture. tron and carapace up with the hypo-xiphiplastral 47. Marginal scales reach cos- 0) no INFRAMARGINAL SCALES: There are four in- tal bones 1) posteriorly only framarginals on each side ofCCM 60-15 (fig. 2) posteriorly and 1 B). They are approximately equal in size, laterally the first being only slightly narrower than the 48. Antrum postoticum re- 0) no remaining three. The first inframarginal, the duced 1) yes axillary, lies entirely on the hyoplastron. It contacts at least the fourth and possibly the third right marginal scales laterally, the sec- and in Basilemys and Nanhsiungchelys, the ond inframarginal posteriorly, and the pec- pectorals extend across as much as halfofthe toral medially. The fourth inframarginal, the entoplastron. In those Adocus in which the inguinal, lies completely on the hypoplas- pectorals do not extend onto the entoplastron. It contacts the abdominal medially, the tron, they are always quite close to its pos- third inframarginal anteriorly, and the sev- terior edge (A. agilis, A. kirtlandius, A. bossi, enth and possibly the eighth marginals lat- CCM 60-15). There is also variation in the erally. The second and third inframarginals contacts between the pectorals and infra- lie between the fifth and sixth marginals marginals (these will be discussed below un- laterally and the pectorals and abdominals der variation of inframarginals). Set 4 scales medially. The second lies entirely on the hy- are absent in the scaled plastra of Emargi- oplastron; the third crosses the hyo-hypo- nachelys, Dermatemys, Baptemys, Agom- plastral suture. phus, Hoplochelys, Staurotypus, Claudius, and The number of inframarginal scales varies Kinosternon (Hutchison and Bramble, 198 1). among the taxa ofthe Trionychoidea but also They are also absent from the unscaled plas- varies within genera assigned to this super- tra of Anosteira, Carettochelys, and all tri- family (table 2). Based on the presence of onychids. complete inframarginal series (four scales per Set 5 scales (the abdominals) always cover side) in pleurosternids, plesiochelyids, baen- the hyo-hypoplastral suture when they are ids, chelonioids, and some trionychoids, a present. Following the work ofHutchison and low number ofinframarginals can be consid- Bramble (198 1), it is proposed that set 5 scales ered derived within the Trionychoidea. The are completely absent from the Kinostemi- problem is recognizing states in the character, dae. In Hoplochelys, set 5 scales are displaced given the known variability. Adocus has four laterally onto the bridges and do not meet on scales per side with only one exception. Der- the midline. This pair of scales, if present, matemys has either four or five per side. Bap- separates the scales of the posterior plastral temys has three or four, Hoplochelys has three lobe from the inframarginals except in Basi- per side, Agomphus has two or three, and all lemys and Zangerlia (Langston, 1956; Mly- kinosternids only two per side. Taxa are con- narski, 1972) and Dermatemys (Mlynarski sidered to have the primitive condition ifmost 1976; BMNH 1911.1.28. 1) in which the fem- individuals had at least one side with four orals extend forward sufficiently to contact inframarginals. Those with three or fewer per 36 AMERICAN MUSEUM NOVITATES NO. 2941

atlas 2 3 4

C', co 0

,-1 1- -- ,

7-, I a) N---- (zi7§;X

r.. I -;.. 9 K>_ :L-- c -z=-======-=i (1)

II -)

CZ,

)

/g I r~

0 1 2 cms Fig. 15. The cervical vertebrae of Adocus sp. (CCM 60-15). Vertebrae are arranged from anterior (atlas) on the left to posterior (cervical 8) on the right.

side are considered derived; those with two Ideally this character would be treated as par- per side are further derived. Those taxa that tially ordered in numerical analysis. lack any shell scales were scored separately. In most trionychoids the inframarginals 1989 MEYLAN AND GAFFNEY: ADOCUS 37

5 6 7 8

.1 -,

1, - 1-

( )

t ==2_ '.

I ,'

J I,,, 4

ii K' \\

\(-

1:I" II

Fig. 15-continued. completely separate the plastral and carapa- scales appears to occur when the number of cial scales. This condition appears to be prim- inframarginals is reduced. Within the tri- itive for turtles. Division ofthe inframarginal onychoid genera treated here, it occurs in all series by contact of plastral and carapacial Basilemys and in some Kinosternon. Else- 38 AMERICAN MUSEUM NOVITATES NO. 2941

TABLE 3 The Number of Inframarginals in Members of the Trionychoidea (All kinosternids have two inframarginals and they are not included in this table) Number of inframarginals (left/right) Taxon 5/5 5/4 4/4 4/3 3/3 2/2 Adocus CCM 60-15 USNM 8577 AMNH 6712 USNM 6539 USNM 8593 Agomphus ANSP 15359 YPM 671 AMNH 1478 YPM 774 Baptemys AMNH 6004 USNM 13437 UCMP 45477 USNM 13438 AMNH 6109 MCZ Seton, AMNH 5967 1951 AMNH 5934 YPM 3754 Basilemys FMNM 11084 Dermatemys UNAM 2612 UNAM 2613 UNAM 2479 UNAM 2480 UNAM 2483 UNAM 51072 Emarginachelys KU 23488 Hoplochelys USNM 8646 USNM 6549 USNM 8525 USNM 8553 Peltochelys IRSNB 1536 where it occurs in all Pleurodira except Pro- and only weakly projecting on centra 2 terochersus and in all Testudinoidea. through 4, but they are very broad on centra 6 and 7. The keel is missing on centrum 5 but appears to have been moderately devel- THE CERVICAL VERTEBRAE oped. The keel on the eighth cervical differs The cervical series of CCM 60-15 is com- from the others in having small accessory plete (fig. 15). The first thoracic vertebra is keels lying on either side. preserved in place within the carapace. The The neural arches are typical of eucryp- anterior cervicals are lightly built, long, and todiran turtles having widely separated pre- narrow. The more posterior ones are broad and postzygapophyses. The postzygapophy- and robust. Centra 2 through 5 are at least ses become more robust posteriorly and are twice as long as they are wide; 6 and 7 are quite thick on vertebrae 6, 7, and 8. The two somewhat wider; 8 is as wide as long. These most angular contacts between successive centra are remarkable in that all are opistho- pairs ofzygapophyses are those between ver- coelous. Centra 1 through 5 have single con- tebrae 5 and 6, and 7 and 8. cave posterior articular surfaces; in 6 and 7 The paired atlantal neural arches meet on these surfaces are doubled, and number 8 is the midline but are neither sutured nor fused. single. Centra 1 through 6 have single, con- They articulate posterolaterally with the pre- vex, anterior articular surfaces; numbers 7 zygapophyses ofthe second cervical (the axis). and 8 have clearly developed double anterior Beyond this articulation extends a narrow, articular surfaces. Centra 2 through 5 are rect- free process that reaches a point about half- angular in ventral view with well-developed way across the axis. Anteromedially the at- and anteriorly placed parapophyses; centra 6 lantal neural arches contact the first vertebral and 7 are hourglass-shaped, and 8 is square centrum; anteroventrally they contact the at- with large parapophyses. Ventral keels are lantal intercentrum. Together with the atlan- present on centra 2 through 8. They are thin tal intercentrum the two neural arches form 1989 MEYLAN AND GAFFNEY: ADOCUS 39 the atlantal ring, which articulates with the lan, 1987), this process is accompanied by occipital condyle. The atlantal intercentrum accessory ridges on either side, but only in is braced posteriorly by a thick ventrally di- certain trionychoids is this structure paired. rected portion of the first vertebral centrum. As shown in Meylan (1987: fig. 19), the ven- Opisthocoely of the cervical series ofAdo- tral process of the eighth cervical is paired in cus is a remarkable feature. Complete cer- kinosternids and carettochelyids. vical opisthocoely has been reported elsewhere among turtles only in the Trionychidae and Carettochelyidae. New evidence shows PECTORAL GIRDLE AND FORELIMB that Basilemys, like Adocus and the Tri- The coracoid of Adocus (CCM 60-15) is onychia, has opisthocoelous posterior ver- very long and narrow (fig. 17). It is substan- tebrae. Two different specimens ofBasilemys tially longer than the acromion process ofthe (AMNH 5448 and NMC 8890) each include scapula and only slightly shorter than the body a single cervical vertebra. In both cases it of the scapula. appears to be number 6. This element has a The coracoids of most aquatic turtles are broad, opisthocoelous centrum in both ex- long and narrow, as they are in all of the amples, which can be seen to be double in trionychoids examined during this study. NMC 8890 (fig. 16A, B). The anterior artic- Those of trionychids are curved, those of ular surface is a large, convex structure in Basilemys are distally expanded and very long both specimens. The prezygapophyseal facets (NMC 8890, AMNH 5448), and those ofkin- are inclined at a high angle to the centrum as osternids are only very weakly sutured to the they are in the sixth centrum ofAdocus. scapula. No feature of the coracoid was dis- In all other trionychoids for which the cer- tributed in such a manner that it might be vical series is known, one to three anterior useful in suggesting relationships among the vertebrae are opisthocoelous. These are fol- taxa of interest here. lowed by a single biconvex cervical; the re- The right humerus of CCM 60-15 is pre- mainder of the cervicals are procoelous. The served in nearly perfect condition (fig. 18). fourth cervical is biconvex in the single avail- The caput humerus is large and is located at able cervical column ofEmarginachelys (KU the end of a short neck that extends from the 23488). In Baptemys (USNM 13437) and shaft ofthe humerus at an angle ofabout 900. nearly all kinostemids (Williams, 1950), the It has a weakly developed groove in its an- third cervical is biconvex, with all more posterior edge. The main axis of the articular terior cervicals being procoelous (fig. 16). In surface of the caput humerus lies at a slight Dermatemys the second cervical is biconvex. angle to the main axis ofthe shaft. The medial Among cryptodirans with formed cervical process of the humerus is about three times centra (the Centrocryptodira of Gaffney and larger than the lateral process; between them Meylan, 1988), typically the fourth cervical is a well-developed intertubercular fossa. The is biconvex. Of those eucryptodires exam- medial process extends posteriorly from the ined by Williams (1950), 115 of 117 chel- shaft at an angle of nearly 400, and extends ydrids (98.3%), 647 of 694 emydids (93.2%), as far medially as does the caput humeri. The 70 of 72 cheloniids (97.7%), and 136 of 358 lateral process diverges anteriorly from the oftestudinids (40.0%) have a biconvex fourth shaft at an angle of about 100. It does not cervical. The occurrence of this feature in extend medially to the level of the medial Meiolania and its broad distribution process or caput humeri. The two distal tro- throughout the Centrocryptodira lead Gaff- chanters, the capitellum and trochlea, are ap- ney (1985a) to propose that it is the gener- proximately equal in size and do not com- alized condition for this group. The distri- pletely cover the distal end of the element. bution ofthis feature in a complete phylogeny The ectepicondylar foramen is closed. of turtles (Gaffney and Meylan, 1988) is in The humerus ofAdocus is essentially iden- agreement with this hypothesis. tical to that of Baptemys (USNM 13437), In most turtles, as in Adocus, the ventral Staurotypus, and Dermatemys. The humerus process of the eighth cervical is not a paired of Basilemys (CM 8890) is similar but the structure. In certain other taxa (e.g., Trach- lesser trochanter curves posteriorly, further emys scripta, AMNH 11961; fig. 19 in Mey- enclosing the intertubercular fossa. It also has 40 AMERICAN MUSEUM NOVITATES NO. 2941

anterior posterior

\ I I I I

I I I 3 I I Fig. 16. Posterior cervical vertebrae of selected trionychoid turtles. A, Anterior and posterior views of the sixth cervical of Basilemys sp. (NMC 8890 with reference to AMNH 5548). B, Anterior and posterior views of the seventh cervical of Baptemys sp. (USNM 13437). large distal trochanters that cover the entire radius and ulna are complete and appear to distal portions of the element and the main lie close to their correct natural position with axis ofthe caput humeri lies at a higher angle respect to the manus. They are about equal to the shaft. The lateral process in trionychids in length and both are distally expanded. The is larger than in other trionychoids. In ca- ulna is broader throughout its length than the rettochelyids it is smaller and displaced dis- radius. tally on the humeral shaft. The similarity of The contacts between the radius and ulna the humeri among trionychoids (except for and the carpus are best preserved in the right autapomorphies in Basilemys, carettoche- foot. It appears that the radius contacted the lyids, and trionychids) restricts their utility intermedium and a medial centrale and is in the cladistic analysis. No characters were only slightly separated from the first carpal. developed using humeral morphology. The ulna is separated from the radius by the Both the right and left forelimbs of CCM intermedium. It contacts only the interme- 60-15 are well preserved (figs. 2, 19). Both dium and the ulnare. 1989 MEYLAN AND GAFFNEY: ADOCUS 41

The carpus is composed of four rows of elements: the intermedium; the medial and A lateral centrale and ulnare; five subequal-size carpals; and five metacarpals. The medial scapula centrale appears to have a partially fused radiale medially. There is no evidence ofa pisi- to form. The phalanges are 2-3-3-3-3, with claws present on all five digits. The manus of Adocus is much like those of living casichelydians. It has a primitive glenoid fossa number ofcarpals and phalanges, and retains evidence of a radiale. The absence of a pisi- acromion form is ascribed to loss during preservation. process Elsewhere among the Trionychoidea, de- coracoid rived manus and pes morphology is found in carettochelyids and trionychids. Both have a reduced number of clawed digits; the latter also exhibit hyperphylangy (Meylan, 1977). In Basilemys there is a reduced number of phalanges in the pes (2-2-2-2-1 based on AMNH 5448). A reduction in the number of phalanges also appears to exist in Nanhsi- B ungehelys, but this needs verification. scapula PELVIC GIRDLE Part of the pelvic girdle of CCM 60-15 is preserved in a block of matrix with the head of the right femur and a series of caudal ver- s tebrae. However, enough of it is exposed or acromion process free from the block that the entire structure can be described and reconstructed (fig. 20). The pubes make up one-third of the acetabulum. They are broad and long, extending anteriorly well beyond the pectineal process- coracoid es. There is broad interpubic contact and the pectineal processes are elongate. These elements project into the thyroid fenestra at the midline but do not contact the ischia. Instead they lie dorsal and anterior to the ischia, which results in an undivided thyroid fenestra. The ischia make up the posteroventral third of the acetabulum. They have extremely well- developed metischial processes as well as anterior projections that extend halfway across the thyroid fenestra. The ilia form the dorsal third of the acetabulum from which they extend posterodorsally and somewhat medially. They are not curved but have weakly defined thelial processes that are best seen in Fig. 17. The right pectoral girdle of Adocus lateral view. (CCM 60-15) in (A) lateral, (B) anterior, and (C) The pelvis ofAdocus is primitive in general ventral views. appearance. The open thyroid fenestra, large metischial processes, dorsomedially directed 42 AMERICAN MUSEUM NOVITATES NO. 2941

intertubercular fossa caput humerus lateral process / medial process

ANTERIOR DORSAL POSTERIOR VENTRAL Fig. 18. The right humerus of Adocus sp. (CCM 60-15).

III Ill

Fig. 19. The right (A) and left (B) manus of Adocus (CCM 60-15) in dorsal view. 1 989 MEYLAN AND GAFFNEY: ADOCUS 43

metischial process

sischium

Fig. 20. The pelvic girdle of Adocus sp. (CCM 60-15) in (A) dorsal, (B) ventral, and (C) lateral views. and dorsally expanded ilia, and unexpanded losses in the Trionychidae and Dermatemys. pectineal processes appear to be primitive for The ilial notch, known to occur in kinoster- eucryptodirans (fig. 21; Baur, 1891a). The nids, is absent from the ilium ofAdocus. This presence of the thelial process is of interest. structure is present in Baptemys (USNM This structure is well developed only in kin- 13437). ostemids (fig. 21 B), carettochelyids (fig. 21 D), and Baptemys (fig. 342 in Hay, 1908a; UCMP 45477, AMNH 5934). It is weakly developed DISCUSSION: in Adocus, Emarginachelys (Whetstone, MONOPHYLY OF THE TRIONYCHOIDEA 1978), and Lissemys (Meylan, 1987). Its dis- (Fig. 22, Node 1) tribution is best explained as a synapomor- The current study indicates that the Tri- phy for the Trionychoidea, with independent onychoidea of Gaffney (1975, 1984) is a 44 AMERICAN MUSEUM NOVITATES NO. 2941 A D epipubis- pectineal

- thyroid fenestra

thelial process

metischial process Fig. 21. Dorsal views of the pelvis of six eucryptodiran turtles. A, Pseudemys nelsoni (AMNH 129736); B, StaurotYpius triporcatus (UF 13482); C, Apaloneferox (AMNH 129737); D, Carettochelys insculplta (AMNH 84212); E, Basileinys sp. (NMC 8890); F, Macroclemys temmincki (AMNH 58251). Abbreviations: il, ilium; isch, ischium; pub, pubis. Stippled areas represent calcified cartilage. 1 989 MEYLAN AND GAFFNEY: ADOCUS 45 monophyletic group that can be considered distribution among the living members ofthe to include the extinct genera Adocus, Basi- superfamily, occurring only in the Kinoster- lemys, Nanhsiungchelys, Peltochelys, Emar- nidae, Carettochelys, and Lissemys and has ginachelys, Baptemys, Hoplochelys, and Xen- been used as evidence for monophyly of the ochelys. This superfamily is diagnosed on the Kinosternidae plus Trionychia (Meylan, basis of a combination of characters used by 1987). However, the observation that this Gaffney (1975, 1984), Meylan (1987), and structure is present in Emarginachelys, Bap- Gaffney and Meylan (1988). The inclusion of temys, and Adocus suggests that its presence data from the fossil record requires a reex- is primitive for the superfamily and lost in amination of the diagnostic features of the the common ancestor of Basilemys and Trionychoidea as summarized for living taxa Nanhsiungchelys, in Dermatemys, and within by Meylan (1987: table 19). the Trionychidae. Thelial processlike struc- The features which led Gaffney (1975, tures occur elsewhere only in testudinoids 1984) to suggest the existence of this natural with kinetic plastra (Cuora, Terrapene, Pyxi- group, the enlarged internal carotid canal and dea, Kinixys). In these taxa they appear to apparent reduction of stapedial circulation, mark the anterior limit of a broad blade for remain ofgreatest importance (Gaffney, 1979; insertion of an expanded m. testoiliacus Albrecht, 1967, 1976; McDowell, 1961). Sur- (Bramble, 1974). They are not homologous veys of new material indicate that significant to the thelial process in trionychoids. participation by the palatine in the braincase The presence of a maxillary tooth is rare is also diagnostic (Gaffney, 1979; Meylan, among living trionychoids, occurring only in 1987). This feature occurs elsewhere sporad- Dermatemys. But it is known to occur in the ically in species of the Deirochelyinae (sensu extinct genera Adocus, Baptemys, and Emar- Gaffney and Meylan, 1988), in the genera ginachelys. Cladistic analysis of about 50 Graptemys and Pseudemys (Meylan, personal morphological characters suggests that these observ.), and in certain batagurines (e.g., Ba- four taxa are basal in the trionychoid clade tagur; Gaffney, 1979). and that the presence of a maxillary tooth is Transverse processes do not reach the cos- primitive for, and diagnostic of, the Tri- tal bones in any trionychoids for which the onychoidea. The maxillary tooth is lost twice 10th vertebra is well known. The only pos- within the Trionychoidea. sible exception is Adocus. Baur (1891 a) sum- The presence of a caudifibularis muscle is marized his observations on the contacts of unique to Dermatemys, kinosternids, and the ribs of the 10th vertebra among the Tes- trionychids (Zug, 1971). Although unknown tudines and indicated that those of the Ado- for Carettochelys and the fossil taxa consid- cidae are present and sutured to the eighth ered in the current study, it is used as evi- costal bone. His observations were based on dence for the monophyly of the Trionychoi- Marsh's Adocus punctatus which he recon- dea here and elsewhere (Gaffney and Meylan, structed (Baur, 1891b). However, Wieland 1988). (1904) thoroughly described the same spec- The addition of new data, especially on imen, and in an accompanying figure indi- fossil forms, suggests that several characters cated the point at which "[the] rib of tenth used by Meylan (1987) to support monophy- dorsal vertebra unites suturally with the 8th ly ofthe Trionychoidea no longer apply. The costal, or else point of iliac support." Wie- distribution of plastral buttresses that reach land's observations, combined with our ex- overlying costal bones among the Cryptodira perience with material preserved in New has been reconsidered (Gaffney and Meylan, Jersey Greensands, suggest that Baur's de- 1988). The absence of enlarged plastral but- scription may have been hypothetical, based tresses in all trionychoids other than Der- on observed scars and breaks. We have not matemys and Baptemys can best be attrib- found decisive material and must consider uted to retention of the primitive condition this feature unknown for Adocus. for the Eucryptodira. The appearance of ax- The possession of a thelial process of the illary buttresses that reach anterior costals in ilium appears to be a derived feature of the Baptemys and Dermatemys is a synapomor- Trionychoidea. This character has a limited phy for a restricted Dermatemydidae. The 46 AMERICAN MUSEUM NOVITATES NO. 2941

TRIONY* 12 A# KINO'Ul I L I ON CHIA STERNI DAE 5 = _ KINOSTERNIA ~DERMAT* 1 CHELYIDAE ~~EMYDIDAE

TRIONYCHOIDAE 2 7 KINOSTERNOIDAE 1 TRIONYCHOIDEA Fig. 22. A cladogram for the Trionychoidea. Characters supporting monophyly at indicated nodes are as follows: Node 1: Trionychoidea 1) decrease of stapedial circulation and increase of carotid circulation 2) significant contribution to braincase made by palatine 3) 10th body vertebra free from costals 5) loss of basis tuberculi basalis (present in Adocus) 6) anal notch absent (may be primitive condition for polycryptodira) 7) maxillary "tooth" present (lost at nodes 6 or 7 and 10) 8) thelial process present 9) presence of a caudifibularis muscle (Zug, 1971) Node 2: Trionychoidae 1) articulations between cervical vertebrae opisthocoelous 2) foramen posterior canalis carotici interni very ventral, completely enclosed by pterygoids (occurs also in Baptemys) 3) large portion of processus trochlearis oticum made up by parietal 4) coronoid process tall and located in middle ofjaw ramus 5) well-developed retroarticular process 6) sinuous midline plastral sulcus (occurs also in Agomphus) Node 3: Nanhsiungchelys, Peltochelys, and the Trionychia (skull characters not known for Peltochelys) 1) vomer reduced 2) palatines truncated anteriorly allowing large area of communication between the apertura narium internum and the fossa nasalis 3) no processus extemus of pterygoid 4) incisura collumella auris closed 5) cheek emargination absent 6) top of skull sculptured (lost in Trionychidae, present in Basilemys) 7) basisphenoid contacts palatines and divides pterygoids Node 4: Nanhsiungchelyidae 1) reduced number of phalanges 2) thelial process absent Node 5: Peltochelys, Carettochelyidae, and Trionychidae 1 989 MEYLAN AND GAFFNEY: ADOCUS 47

1) 10 or fewer peripheral bones 2) ventral processes of the nuchal Node 6: Trionychia (Trionychidae and Carettochelyidae) 1) peripherals not sutured to costals 2) ribheads strongly sutured to vertebral centra 3) premaxillae fused 4) radius and ulna in contact adjacent to manus 5) number of clawed digits three or fewer 6) coracoid longest of three pectoral processes 7) carapace not sutured to plastron Node 7: Kinosternoidae 1) scale set 2 (extragulars) lost from plastron (Hutchison and Bramble, 1981) 2) scale set 4 (pectorals) lost from plastron (Hutchison and Bramble, 1981) 3) no scales on skull roof (occurs also at node 1 1) Node 8: Dermatemydidae, Agomphus, Hoplochelys, and Kinosternidae 1) foramen caroticum laterale larger than foramen anterior canalis carotici intemi 2) biconvex cervical is number 3 or 2 3) foramen stapedio-temporalis reduced or closed Node 9: Dermatemydidae (restricted to the genera Baptemys and Dermatemys) 1) commissural ridge on triturating surface 2) foramen stapedio-temporalis closed 3) axillary buttresses reach costal bones Node 10: Agomphus, Hoplochelys, and Kinostemidae 1) no more than three inframarginal scales present on each bridge (occurs also in Basilemys) Node 1 1: Hoplochelys and Kinosternidae 35) reduction of scale set 5, abdominals (Hutchison and Bramble, 1981) 19) tricarinate carapace Node 12: Kinostemidae 1) 10 peripheral bones on each side of carapace 2) no ribs connecting ninth and tenth body vertebra to eighth costal 3) anterior lobe kineses that includes entoplastron (when it is present) 4) frontal not reaching orbit 5) maxilla contacting quadratojugal 6) ilio-ischial notch in acetabulum (occurs also in Baptemys) 7) only two inframarginals (occurs also in Basilemys) 8) stridulating organs present in males (lost in Kinosternon scorpioides group) 9) scale set 5 (abdominals) absent 10) incised anterior musk ducts

contact ofall four buttresses to overlying cos- family. Similarly, a basis tuberculi basalis is tal bones is a synapomorphy for the Testu- present in Adocus, suggesting that the absence dinoidea (Hirayama, 1985; Gaffney and of this structure is not a synapomorphy of Meylan, 1988). the superfamily. Also, the quadrate appears Cheek emargination is well developed in to make up a large portion of the processus Adocus (fig. 4A), Emarginachelys, and Bap- trochlearis oticum in Emarginachelys, sug- temys. This distribution eliminates reduced gesting that small quadrate contribution to cheek emargination as a synapomorphy for the processus trochlearis oticum is not a syn- the Trionychoidea (Meylan, 1987). Such re- apomorphy of the Trionychoidea. duction appears to occur at least three times Nessov (1977) acknowledged monophyly within the superfamily and is thus no longer of the family Adocidae but placed it within considered a synapomorphy for the super- the Testudinoidea. He supported this posi- 48 AMERICAN MUSEUM NOVITATES NO. 2941 tion by suggesting that the members of this thus terminate use of, the family group name taxon do not have a reduced foramen sta- Dermatemydidae. These criteria limit the in- pedio-temporale typical of the Dermatemy- cluded genera to Dermatemys and Baptemys. didae (which he implies is the sister-group of The priority of the Kinostemidae (Agassiz, the Testudinoidea). However, as the condi- 1857) over Dermatemydidae (Gray, 1870) tion of the foramen stapedio-temporale of reduces the role of this previously broadly Adocus is primitive for turtles, it does not applied name. support Nessov's hypothesis. Other derived features of Adocus discussed above suggest MONOPHYLY OF THE TRIONYCHOIDAE that it belongs within the Trionychoidea (sen- (Fig. 22, Node 2) su Gaffney 1975, 1984), which is the sister- group to the Testudinoidea. Six morphological features are important This study confirms monophyly of the in establishing the monophyly of the Tri- Trionychoidea and presents the first oppor- onychoidae. Five of these are established for tunity to fully explore the relationships of Adocus for the first time in this paper. living and extinct members ofthe group down In all members of the epifamily Trionych- to genus (fig. 22). oidae in which cervical central articulations The relationships among members of the are known, all articulations between cervicals superfamily Trionychoidea are best reflected are opisthocoelous. Based on CCM 60-15, by the recognition oftwo large monophyletic Adocus, like carettochelyids and trionychids, groups, each containing several families of can be shown to be completely opis- turtles (fig. 22). The rank ofepifamily is used thocoelous (fig. 15). Two cervical vertebrae here to acknowledge this level ofrelationship of Basilemys, one from each of two different as suggested by Bour and Dubois (1985, specimens (AMNH 5448 and NMC 8890) 1986). The names Trionychoidae (Fitzinger, suggest that it, too, has opisthocoelous ver- 1826; as Trionychoidea) and Kinosternoidae tebrae in the posterior part of the cervical (Agassiz, 1857; as Kinostemidae) have al- series (fig. 16A, B). The only turtles with opis- ready been employed (Gaffney and Meylan, thocoelous posterior cervicals are those in 1988). which the entire column is opisthocoelous The Trionychoidae are considered to in- (Williams, 1950). Opisthocoely of the cer- clude the monophyletic families Adocidae vicals in this epifamily is unique for the Tes- Cope, 1870 (not including Basilemys; contra tudines. Although the cervical series of Gaffney and Meylan, 1988), Nanhsiungche- Nanhsiungchelys is known, the degree of lyidae (Yeh, 1966, including Basilemys), Ca- preparation of the material does not allow rettochelyidae Boulenger, 1887, and Tri- reconstruction of the cervical central articu- onychidae Fitzinger, 1826, as well as the genus lations. The cervicals of Peltochelys are un- Peltochelys. The Kinostemoidae are consid- known. ered to include the Dermatemydidae Gray, A feature ofthe internal carotid circulation 1870, the Kinosternidae Agassiz, 1857 (in- used by Meylan (1987) as evidence for mono- cluding Xenochelys), and the genera Emar- phyly of the Trionychia (Carettochelyidae ginachelys, Agomphus, and Hoplochelys. The plus Trionychidae) is also present in Adocus. relationships within and among these taxa Like members of the Trionychia, Adocus has are discussed below. the foramen posterius canalis carotici intemi An important taxonomic result ofthis work completely surrounded by the pterygoid (fig. is the proposed restriction ofthe family Der- 5). This condition also occurs in Baptemys matemydidae Gray, 1870. Our results sug- (fig. 6), but the available data suggest that it gest that the taxa assembled under this name is independently derived. The exact position by most authors (e.g., Hay, 1908a; Mlynar- of the foramen posterius canalis carotici in- ski, 1976) form a paraphyletic group. We have terni in Nanhsiungchelys should be deter- reduced the included genera in order to: (1) minable from further study of the type skull. form a monophyletic group; and (2) exclude As pointed out by Gaffney (1979: 115), the members of the Kinosternidae (Agassiz, processus trochlearis oticum oftrionychids is 1857), which would have priority over, and distinct in being made up in large part by the 1 989 MEYLAN AND GAFFNEY: ADOCUS 49 descending process of the parietal. The cur- dorsal views of the carapace; see Wieland, rent study reveals that the same is true of 1904; Hay, 1908a; Gilmore, 1919). The re- carettochelyids (Gaffney, 1979: figs. 173-175) sults ofthe present study suggest that the dis- and Adocus (figs. 4, 5). This region ofthe skull tribution of the character used previously to appears to be the same in Nanhsiungchelys place Basilemys in the Adocidae, the unusual but it is unknown in Basilemys. 6 4 6 6 6 6 neural pattern (Gaffney and Mey- Two characters of the lower jaw used pre- lan, 1988), is most parsimoniously explained viously to support monophyly of the Caret- as two independent occurrences. However, tochelyidae plus Trionychidae (Meylan, 1987) improved knowledge of the cranial mor- are also present in Adocus and thus support phology of Basilemys and Nanhsiungchelys monophyly ofthe Trionychoidae. In this tax- is required to refute or corroborate the cur- on the coronoid process is quite tall and is rent hypothesis. located in the middle of the mandible rather Nessov (1977) revived the name Adocidae than in the posterior half (figs. 9, 10). The and considered this family to include Adocus, known members of the taxon also possess a Zygoramma, and Alamosemys as members well-defined retroarticular process of the of the subfamily Adocinae, and Shachemys mandible. and Ferganemys as members of the subfam- Adocus, Basilemys, and Peltochelys are un- ily Shachemydinae. As the current analysis like other scaled trionychoids, other than suggests, he, too, considered Basilemys to be- Agomphus, in having a sinuous midline plas- long to the Nanhsiungchelyidae ofYeh (Nes- tral sulcus, presumably a derived condition sov, 1986). within the Eucryptodira. Only the anterior Zygoramma Cope, 1870, was differentiat- half of the plastron of Nanhsiungchelys is ed from Adocus by its describer by differences known and the midline plastral sulcus is not in buttress construction. However, it is iden- discernible from available figures and pho- tical to Adocus in having buttresses restricted tographs. to the peripheral bones. It does not deserve There have been previous hypotheses of generic recognition and is placed in synon- relationship of Adocus, Basilemys, and ymy with Adocus. Nanhsiungchelys (Chkhikvadze, 1975), and a Alamosemys Hay, 1908a, was differentiat- distinction drawn between the Adocidae and ed from Adocus in having the marginal scales Dermatemydidae (sensu lato) (Nessov, 1977). wholly excluded from the costal bones. Reex- However, the arrangement employed in Gaff- amination of the type (AMNH 1204) reveals ney and Meylan (1988), and modified slightly that the carapacial sutures cannot be seen in in the present paper, is novel. Available data this badly eroded shell. It has been impos- suggest that the Adocidae (Adocus and related sible to substantiate the diagnostic feature of Asiatic genera) is the sister-group to the the genus, and it is considered a synonym of Nanhsiungchelyidae, Peltochelys, and the Adocus following Gilmore (1919). Trionychia; and that the Nanhsiungchelyidae Evidence can be taken from Khosatzky and is the sister-group of Peltochelys and the Nessov (1977) and Nessov (1977, 1986) that Trionychia. Elsewhere, one of us (Meylan, Ferganemys is an adocid, differentiated from 1988) has argued that Peltochelys can be con- Adocus by its less expanded marginal scales sidered the sister-group of a monophyletic (fewer reach the costal bones than in Adocus) Trionychia. and the extra pair of scales in the axillary region of the plastron. This genus, described from the Albian ofKirgizia, USSR, is MONOPHYLY OF THE ADOCIDAE known from shell and skull material (see fig. 8 in The concept of the Adocidae in this paper Nessov, 1986). Ferganemys retains primitive differs from that used in Gaffney and Meylan eucryptodiran shell features including a com- (1988) in the exclusion of the genus Basi- plete set of plastral scales, complete infra- lemys. The family can be diagnosed by the marginal series, no mesoplastra, plastral but- presence of marginal scales that extend well tresses restricted to peripheral bones, and a up onto the costal bones beginning in the raised area for pubic contact on the dorsal vicinity ofperipheral 4 (fig. 1 lA, and all other surface of xiphiplastra. It also exhibits two so AMERICAN MUSEUM NOVITATES NO. 2941

derived shell features that are also found in we consider to be a testudinoid. The possi- Basilemys and Adocus, the 6 4 6 6 6 6 neural bility that Shachemys could be either a tri- pattern, and medially expanded pectoral onychoid or testudinoid suggests that it should scales. Ferganemys shares with Adocus the be referred to Eucryptodira, incertae sedis, presence of a large posterior suprapygal bone until it is better known. and marginal scales that extend dorsally be- Hay (1 908a) characterized Adocus as "the yond the peripheral bones. The latter feature least modified ofthe Dermatemydidae." Since occurs only posteriorly in Ferganemys and his concept of the Dermatemydidae (inclu- not posteriorly and laterally, as in Adocus. sive of Adocus, Basilemys, Dermatemys, The skull ofFerganemys described by Nes- Baptemys, Hoplochelys, and Anosteira) is sov (1977) supplies additional evidence that equivalent to our use of Trionychoidea, we this genus should be referred to the Adocidae. are in general agreement with this statement, The description of the foramen posterius ca- especially applied to shell morphology. How- nalis carotici interni as similar in size to the ever, derived features ofthe skull, lower jaw, foramen stapedio-temporale suggests that and neck indicate that Adocus is more trans- Ferganemys is a trionychoid. Nessov de- formed than previously recognized and is a scribed the processus trochlearis oticum as member of the Trionychoidae. large but with the prootic narrow (see also Khosatzky and Nessov, 1977). This description implies that the parietal contribution MONOPHYLY OF THE must be large, as indicated by figures 1 and NANHSIUNGCHELYIDAE, PELTOCHELYS, 2 in Nessov (1977). This feature is employed AND THE TRIONYCHIA here as a character ofthe Trionychoidae. Fer- (Fig. 22, Node 3) ganemys is also like Adocus in retaining prim- Monophyly ofthe sister-group to the Ado- itive features such as the basis tuberculi ba- cidae is supported by characters ofskull mor- silis, a large foramen stapedio-temporale, and phology. However, the skulls of Basilemys small foramen caroticum laterale. and Peltochelys remain unknown. Basilemys Shachemys is more problematical than is included in this clade because it shares de- Ferganemys. Shachemys Kuznetsov, 1976, rived features of the shell, pelvis, and feet was originally based on a series of isolated with Nanhsiungchelys. Peltochelys is includ- elements ofa carapace and plastron from the ed because it shares derived shell features of Upper Cretaceous ofKazakhstan. This genus the Trionychoidea and Trionychia. The nu- was further described and figured by Nessov merical method employed here (PAUP 2.4; (1986). It is like Adocus in having fine punc- Swofford, 1984) assumes that these and all tate sculpturing, but this sculpturing appar- missing values should take the most parsi- ently differs in detail from that ofAdocus and monious condition based on known charac- Ferganemys. Like Adocus and Basilemys, it ter states. has medially expanded pectoral scutes, and The important skull features at this level like Adocus and Ferganemys, it has an en- are those shared by Nanhsiungchelys, mem- larged posterior suprapygal. Shachemys and bers of the Carettochelyidae, and members Ferganemys also have the posterior marginal of the Trionychidae. In these taxa the vomer scutes extending dorsally beyond the periph- is reduced. It does not reach the palatine- eral elements. However, Shachemys exhibits pterygoid suture as in Adocus and nearly all a series of features that are not known to other cryptodires. The anterior palate ofthese occur elsewhere in the Adocidae. It lacks a taxa is further derived in having the palatines cervical scale and inframarginal scales, as well truncated anteriorly, allowing a large area of as neural bones. Also the epiplastra are ap- communication between the fossa orbitalis parently hinged transversely across the an- and the apertura narium intema, and in hav- terior lobe of the plastron. If Shachemys is ing the processus pterygoideus externus near- considered an adocid, these features must all ly absent. The former feature also occurs in be treated as autapomorphies. An absence of some chelonioids that lack a secondary palate inframarginals and the presence of mobile (Dermochelys); the latter occurs also in some epiplastra also occur in Mongolemys, which kinosternids and some Dermatemys, and 1989 MEYLAN AND GAFFNEY: ADOCUS 51 outside of the Trionychoidea in some che- articulated. They indicate a phalangeal for- loniids and testudinoids (Meylan, 1987). The mula of 2-2-2-2-1 for this genus. The foot of quadrate in Nanhsiungchelys and members Nanhsiungchelys was described as having of the Trionychia completely encloses the short, broad phalanges (Yeh, 1966) and being stapes. This closure of the incisura columel- testudinid-like. It appears to have shortened lae auris is unique within the Trionychoidea. digits, but further preparation of the type is It occurs outside the superfamily in meiolan- necessary to verify this. iids, chelydroids, and testudinids (Gaffney, Nanhsiungchelys and Basilemys are also 1979; Meylan, 1987). This closure is one of similar in having "pock-mark" sculpturing the few features visible on the crushed skull (Mlynarski, 1972), and in the form of their of Basilemys (NMC 8890). In this specimen anterior plastral lobe. In both genera this lobe the incisura columellae auris is closed; un- is a short triangle rounded anteriorly with a fortunately, it cannot be determined with cer- tendency to fuse the intergular sulcus (set 1 tainty that this is not due to crushing. scales) (type ofB. nobilis, USNM 11084, and In its general form, the skull of Nanhsi- Nanhsiungchelys) and to have the extragulars ungchelys is like that of carettochelyids and (set 2) greatly reduced (type of B. variolosa, trionychids in having significantly reduced type of B. nobilis, USNM 11084) or absent cheek emargination. Reduction of cheek (Yeh, 1966; Langston, 1956). emargination occurs also within the Kino- Zangerlia was described from the Upper stemoidae in Dermatemys and in the Kino- Cretaceous Upper Nemegt Beds ofMongolia sternidae. Nanhsiungchelys is also like Ca- and is reported to be similar to Basilemys in rettochelys in having a sculptured skull roof. having "pock-mark" sculpturing, a 6 4 6 6 6 In these taxa the texture of the skull roof 6 neural formula, reduction in the phalangeal sculpture approaches that of the shell. Sculp- formula, and contact between plastral and ture of this type is absent in other trionych- marginal scutes (as in Basilemys; see Lang- oids including members ofthe Trionychidae. ston, 1956). Unlike Basilemys, it retains a full Although Gaffney (1979) considered set offour inframarginals (primitive) and has Nanhsiungchelys to be Cryptodira incertae a midline keel (derived). Sukhanov and Nar- sedis, several authors have anticipated the mandakh (1974) suggested synonymizing this results of the current study by suggesting a genus with Basilemys. However, the derived relationship between this genus and other shell feature which best defines Basilemys, members of the Trionychoidea (Nessov and the absence of an inframarginal series, does Julinen, 1977; Hirayama, personal com- not occur in Zangerlia, and thus it is not mun., 1986). desirable to add Zangerlia to this genus. Zan- gerlia can be recognized by its middorsal keel, MONOPHYLY OF THE but the data are insufficient to allow deter- NANHSIUNGCHELYIDAE mination of its relationships to Nanhsi- (Fig. 22, Node 4) ungchelys and Basilemys. Mlynarski (1976) included the genera Basi- Our data suggest that Nanhsiungchelys and lemys, Zangerlia, Tretosternon (including the Basilemys form a monophyletic group that is synonyms Peltochelys and Helochelydra), and the sister-group to Peltochelys and the Tri- Peishanemys in his Adocinae. The data avail- onychia. This clade is best termed the able to us suggest that Basilemys and Zan- Nanhsiungchelyidae and probably includes gerlia share a common ancestor with Nanhsi- the genus Zangerlia Mlynarski (1972). ungehelys, not shared by Adocus or any other Two features are suggested to support trionychoids; they are better placed in the monophyly ofthis family: a reduced number Nanhsiungchelyidae. Meylan (1988) has ofphalanges and no thelial process. The pres- shown that Peltochelys is not a synonym of ence of a thelial process is considered to be Tretosternon but rather a trionychoid and is primitive for the Trionychoidea. The feet of best treated as the sister-group of the Tri- Basilemys are best known from AMNH 5448 onychia. Tretosternon (including its synonym from the Belly River Formation. Both hind Helochelydra) is best placed in the Pleuro- feet are preserved and are nearly completely sternidae (Meylan, 1988). Peishanemys has 52 AMERICAN MUSEUM NOVITATES NO. 2941 been placed in the Testudinoidea by Chkhik- menez-Fuentes, 1971; Broin, 1976) and vadze (1975) and Nessov (1981) and we know Nanhsiungchelyidae (Nessov and Julinen, of no characters that support its inclusion in 1977). the Trionychoidea. MONOPHYLY OF THE TRIONYCHIA MONOPHYLY OF PELTOCHELYS (Fig. 22, Node 6) AND THE TRIONYCHIA Monophyly ofthe Trionychia is one ofthe (Fig. 22, Node 5) best corroborated hypotheses of turtle phy- Our numerical analysis of data for the logeny. Meylan (1987) listed 17 morpholog- Trionychoidea suggests a number of plausi- ical features that support the recognition of ble hypotheses for the placement ofPeltoche- the Trionychia among living taxa. Inclusion lys among the Trionychoidea. Although these of fossils in the hypothesis of relationship hypotheses are equally parsimonious, one of among the members of the Trionychoidea us (Meylan, 1988) has argued that the genus suggests that about one-half of these char- be considered the sister-group to the Tri- acters have a greater level of generality than onychia. In the arrangements that place Pel- ifonly living taxa are examined. Nonetheless, tochelys closer to the base of the cladogram, we know of no other competing hypothesis it is excluded from placement as sister taxon of sister-group relationship for the families to the Trionychia because it lacks certain rel- Carettochelyidae and Trionychidae. Non- atively insignificant derived characters. In- morphological data including karyology stead it retains such trivial primitive char- (Bickham and Carr, 1983; Bickham et al., acters as costal bones not meeting on the 1983) and serology (Frair, 1985) also support midline, pectoral scutes not reaching the en- monophyly of the Trionychia. The relation- toplastron, and possession of two suprapy- ships among the genera ofthe Trionychia are gals. An equally parsimonious hypothesis, in discussed by Meylan (1985, 1987, 1988) and which homoplasy in the derived features of are not elaborated on in the present study. Peltochelys is minimized, places Peltochelys as the sister taxon to the Trionychia. Char- acters supporting this relationship include the MONOPHYLY OF THE KINOSTERNOIDAE presence ofa sinuous midline plastral sulcus, (Fig. 22, Node 7) the reduction of the number of peripheral Hutchison and Bramble (1981) examined elements to 10 per side (or fewer in the Tri- the relationships among the group of trionychidae), and the presence of ventral pro- onychoids that we have termed the Kino- cesses of the nuchal (Meylan, 1988). Reduc- sternoidae (Gaffney and Meylan, 1988). To tion in the number of peripherals occurs those taxa included in this clade (labeled as elsewhere only in the Kinostemidae but the Dermatemydinae plus Kinosternidae in their presence of a meandering midline plastral figure 4), we would add the genus Emargi- sulcus suggests the Trionychoidae rather than nachelys. Although this genus was originally the Kinosternoidae. referred to the Chelydridae, it exhibits six of The relationship of Peltochelys to the the eight characters that diagnose the Kino- Trionychia was proposed by Hummel (1929: stemoidae as listed by Hutchison and Bram- 428) who figures it as the common ancestor ble (1981) including: absence of scale set 2, of the Trionychia and Chelydroidea. How- absence of scale set 4, reduced width of pos- ever, Lydekker's (1889) synonymy of Pel- terior lobe of the plastron, reduced number tochelys with Tretosternon has resulted in of inframarginals (secondarily increased in some confusion, with both genera being con- Dermatemys), inguinal scale overlapping hyo- sidered members ofthe Dermatemydidae (in hypoplastral suture, and absence of pedicel- the broad sense) (Mlynarski, 1976). But Pel- late sculpturing. tochelys is easily distinguished from Tretos- The numerical analysis ofdata for the cur- ternon which has large, medially meeting me- rent study has not included considerations of soplastra (Meylan, 1988) and the former has the width of the posterior plastral lobe, po- been placed among the Carettochelyidae (Ji- sition of the inguinal scale, or presence/ab- 1 989 MEYLAN AND GAFFNEY: ADOCUS 53

sence of pedicellate sculpturing. Variation in questionable value because they may repre- the shape ofthe posterior plastral lobe among sent primitive conditions. the genera considered in this study is difficult Following Gaffney and Meylan (1988), cru- to divide into discrete morphologies. The po- ciform plastra are primitive for chelydrids sition of the inguinal scale over the hyo-hy- and chelonioids and thus may be considered poplastral suture occurs in all trionychoid taxa primitive for the Polycryptodira. Large cos- with three or fewer inframarginals, appar- tiform processes appear sporadically among ently making these features redundant. Char- cryptodires, in chelydrids, in modified form acters of sculpture are difficult to homologize in dermochelyids and trionychids, and in and have been ignored. Furthermore, the ab- Dermatemys. Their apparent absence in many sence of sculpturing in the Kinosternoidae fossil taxa may only reflect the fact that they may best be explained as retention of the are completely hidden in articulated shells as primitive condition for turtles. is the case in Dermatemys. Large costiform The characters most important for sup- processes are present in hatchlings and ju- porting monophyly of the Kinosternoidea, veniles of many cryptodires and their differ- including Emarginachelys, are those most ential retention in adults appears to be highly thoroughly discussed by Hutchison and variable. A nonsutural contact at the bridge Bramble (1981): the loss of certain scales of occurs both in chelydrids and chelonioids and the plastron. Unlike all of the other scaled could be considered primitive for all poly- members of the Trionychoidae, all chelo- cryptodirans. Similarly, a largejugal with sur- nioids, and all nonpolycryptodiran crypto- rounding contacts like those ofEmarginache- dires, the kinosternoids do not have a com- lys is widespread among eucryptodires and plete set of seven pairs of plastral scales. must be considered primitive. The pectineal Emarginachelys was figured with four pairs processes of Emarginachelys and chelydrids of plastral scales but, based on further prep- (fig. 21 F) are like those of Adocus (fig. 20A, aration of the type and a referred posterior B), Basilemys (fig. 21 E), and plesiochelyids plastral lobe in the UCMP collection, it can (Brim, 1965) in being anteriorly directed. The now be shown to have five pairs (J. H. Hutch- distribution of this feature, like those listed ison, personal commun.). Based on the po- above, suggests that it may be the generalized sitions ofthe figured scales and the postulates condition and thus weakens Whetstone's ar- ofHutchison and Bramble (198 1) concerning gument for the monophyly of the Chelydri- scale loss, the five pairs are considered to be dae inclusive of Emarginachelys. scale sets 1, 3, 5, 6, and 7, or gulars, humerals, Three characters argued by Gaffney (1975) abdominals, femorals, and anals. All other to support monophyly of the Chelydridae do kinosternoids also lack sets 2 and 4, and these not occur in Emarginachelys. All members characters support monophyly ofthe epifam- of the family except Chelydropsis have some ily. degree ofhooking ofthe premaxillae, all have Another derived feature found in all mem- the frontals excluded from the orbits, and all bers of the Kinosternoidae is the absence of have the incisura columellae auris enclosed scale sulci on the skull roof. These sulci are around the stapes. It might be argued, how- present primitively in the Trionychoidea. ever, that these are all features ofthose mem- Emarginachelys was originally described bers of the Chelydridae that are the sister- as a chelydrid (Whetstone, 1978). However, group to Emarginachelys (as in fig. 21 in a reexamination ofthis taxon suggests that it Whetstone, 1978). Therefore, it is the pres- should be placed among the Trionychoidea. ence ofderived features ofthe Chelomacryp- Whetstone (1978) listed five characters which todira, and specifically the Trionychoidea, in he considered to support monophyly of the Emarginachelys that suggest that this genus Chelydridae, including Emarginachelys: (1) should not be placed among the Chelydridae. cruciform plastron, (2) long costiform pro- The monophyly ofthe Chelomacryptodira cesses, (3) ligamentous attachment of the (Trionychoidea plus Testudinoidea) is sup- plastron to carapace, (4) an elongate jugal, ported by a single character, the absence of and (5) pectineal processes of pelvis nondi- postorbital-squamosal contact. This feature vergent. All five of these characters are of is quite apparent in Emarginachelys (figs. 7 54 AMERICAN MUSEUM NOVITATES NO. 2941 and 8 in Whetstone, 1978). Furthermore, temi are not equal in size, the former being Emarginachelys has a maxillary tooth, a fea- significantly larger than the latter. This con- ture primitively diagnostic of the Tri- dition occurs elsewhere in the Euchelonioi- onychoidea. It has a 10th thoracic vertebra dea (Gaffney and Meylan, 1988). Derma- with transverse processes that are free of the temydids and kinosternids also show overlying costal bones (among chelydrids this reduction (kinosternids) or closure (derma- occurs only in some Chelydra). Emargi- temydids) of the foramen stapedio-tempo- nachelys also has a thelial process and lacks rale. This type of reduction is not known to a basis tuberculi basalis, both of which are occur among other turtles. considered here to be features of the Trio- Two other characters might support mono- nychoidea. Participation by the palatine in phyly of the Dermatemydidae and Kino- the braincase cannot be determined from the sternia. With the exception of Dermatemys, type skull. all members of this group (for which a pelvis The position ofEmarginachelys among the is known) have an ilial notch (see fig. 21 B) Kinosternoidea has been considered as part and all except Baptemys lack cheek emargi- ofour numerical cladistic analysis. The most nation. Both of these features could, with parsimonious solution suggests that Emar- equal parsimony, have occurred once within ginachelys be considered the sister taxon to the Dermatemydidae and once in the Kino- the remainder ofthe Kinostemoidae. This is sternia, which eliminates their support for the supported by characters of the cervical ver- arrangement preferred here. tebrae and cranial circulation. Monophyly of The skull, neck, and pelvis of Hoplochelys the Dermatemydidae and Kinosternia, ex- and Agomphus remain unknown. The inclu- clusive of Emarginachelys, is supported by sion of these two genera in this study relies the presence of a biconvex cervical number on features that they share with the Kino- 2 or 3, a reduced or absent foramen stapedio- sternidae. These features are discussed be- temporale, and a foramen caroticum laterale low. significantly larger than the foramen anterius canalis carotici intemi. We can be certain that Emarginachelys has the primitive condition MONOPHYLY OF THE DERMATEMYDIDAE for at least two of these features: the fourth (Fig. 22, Node 9) cervical is biconvex and the foramen stape- is Three characters provide strong corrobo- dio-temporale large. ration for a monophyletic Dermatemydidae. If this family is restricted to the genera Der- matemys and Baptemys, it is diagnosed by MONOPHYLY OF THE DERMATEMYDIDAE, the presence of axillary buttresses that reach AGOMPHUS, AND KINOSTERNIA the overlying costal bones (fig. 1 2B), the pres- (Fig. 22, Node 8) ence of commissural ridges on the maxillary Monophyly of the restricted Dermatemy- triturating surfaces (figs. 6, 7), and the ab- didae (Gaffney and Meylan, 1988), Agom- sence of a foramen stapedio-temporale. phus, and the Kinosternia (Hoplochelys and Plastral buttresses that extend to the costals the Kinosternidae) is supported unequivo- are not found in chelydrids, chelonioids, or cally by three characters. Unfortunately, the elsewhere among the Trionychoidea. When condition of these characters is not known they do occur elsewhere among the Poly- for Agomphus or Hoplochelys. In derma- cryptodira (in the Testudinoidea), both the temydids and kinosternids, the second or the axillary and inguinal pairs reach the overlying third cervical vertebra is biconvex (Williams, costal bones (except in some kinetic forms). 1950; current study). Primitively, the fourth The presence of a commissural ridge and cervical is biconvex in cryptodires. In der- complete closure of the foramen stapedio- matemydids and kinostemids, the cranial ar- temporale may be unique to the Derma- teries differ significantly from those of most temydidae. However, certain species of the cryptodires. The foramen caroticum laterale genus Kinosternon have the foramen staped- and the foramen anterius canalis carotici in- io-temporale reduced to the point that it is 1 989 MEYLAN AND GAFFNEY: ADOCUS 55 barely visible (e.g., Kinosternon herrerai, UF Given the ambiguity in the characters 57916). mentioned above, we are forced to rely on the number of inframarginals, an admittedly MONOPHYLY OF AGOMPHUS weak character, for our decision to propose AND THE KINOSTERNIA that Agomphus be considered the sister taxon (Fig. 22, Node 10) to the Kinosternia. Inspection of table 2 will reveal that Agomphus is like Hoplochelys and Elsewhere we have termed the sister-group kinosternids in never having more than three to the Dermatemydidae as the Kinosternia inframarginals per side; in fact, some speci- (Gaffney and Meylan, 1988). This taxon, of mens have only two. Some Baptemys also unspecified rank, was proposed to recognize have only three inframarginals per side, but the special relationship ofHoplochelys to the most have four on at least one side. Der- Kinosternidae. Examination of the genus matemys specimens have four or five per side. Agomphus suggests that it, too, may be part The variability of this character makes it of the clade that is the sister-group to the less than ideal. However, it is used here to Dermatemydidae. However, in order to provisionally determine the phylogenetic po- maintain nomenclatural stability, we will sition of Agomphus. Node 10 in figure 22 is continue to restrict the name Kinosternia to supported by a single character: never more those taxa it originally included rather than than three inframarginals present. Only the modify its definition to include Agomphus. availability of additional data will allow us The limited data for Agomphus suggest that to retest the phylogenetic position of Agom- it is the sister-group to the Kinosternia. phus among the Trionychoidea. The absence Agomphus is a shell taxon and can be placed of an enlarged axillary buttress suggests that among the Kinosternoidae only because it Agomphus does not belong within the der- has a reduced number of plastral scales (five matemydids as defined here. However, it pairs; Hutchison and Bramble, 1981). It was might be the sister-group to the Derma- considered the sister-group to the Derma- temydidae, the sister-group to the Derma- temydidae plus Kinosternia (Hoplochelys and temydidae plus Hoplochelys and the Kino- the Kinosternidae) (Hutchison and Bramble, sternidae (as suggested by Hutchison and 1981: fig. 4). The monophyly ofthe latter set Bramble, 1981), or perhaps it lies even closer of taxa was supported by the presence of tri- to the genus Emarginachelys. carination, costiform processes that span peripheral 1, and a reduced bridge, all ofwhich MONOPHYLY OF THE KINOSTERNIA are lacking in Agomphus. However, each of 1 these characters is subject to an alternative (Fig. 22, Node 1) interpretation. Tricarination is absent in most The Kinosternia was originally diagnosed species ofBaptemys and in Dermatemys and by the reduction or loss of scale set 5, the is actually more parsimoniously explained by abdominals (Hutchison and Bramble, 198 1), independent occurrences within Baptemys and by the presence of tricarination (consid- and the Kinosternia rather than an appear- ered to occur independently within Bap- ance early in the Kinosternoidae with inde- temys) (Gaffney and Meylan, 1988). Hutch- pendent loss within Baptemys and in Der- ison and Bramble (198 1) also list megacephaly matemys. Short costiform processes (in as evidence for this clade but the skull of Baptemys) and a long bridge (in Derma- Hoplochelys is unknown. temys) are both found within the Derma- The monophyly of this taxon is best sup- temydidae. Thus, these characters could be ported by the reduction and loss of scale set used as evidence of the monophyletic 5, a character whose use depends on the as- Dermatemydidae plus Kinosternia only by signment of plastral scale homologies pro- assuming that a reversal has occurred in each vided by Hutchison and Bramble (198 1). We character within the Dermatemydidae. They can find no characters that conflict directly could otherwise be viewed as occurring in- with the evolutionary scenario for plastral dependently within dermatemydids and in all scales proposed by these authors and several Kinosternia. characters are consistent with it (reduction in 56 AMERICAN MUSEUM NOVITATES NO. 2941 inframarginals, reduction in peripherals). (1) considering the microsomes fused to the However, the discovery of a skull of same group B macrosome to be nonhomolo- Hoplochelys would provide the best test for gous (see King, 1985); (2) considering the mi- the currently favored arrangement. crosomes homologous but the fusion to the same macrosome homoplasious; or (3) con- MONOPHYLY OF THE KINOSTERNIDAE sidering the numerous morphological char- (Fig. 22, Node 12) acters cited here to all be homoplasious. For the present, the first alternative for resolving Monophyly of the Kinosternidae is sup- this character conflict seems most appropri- ported by a series of shell, cranial, and non- ate. shell postcranial characters. However, since the skulls ofAgomphus and Hoplochelys are as yet unknown, it is conceivable that the ACKNOWLEDGMENTS distribution ofcertain characters used here is This study is dedicated to the memory of actually greater than described. Shell char- Brice Lambert. We thank Marshall Lambert acters diagnosing the Kinosternidae include for the long-term loan of the Carter County the presence of only 10 pairs of peripheral Museum Adocus specimen and the photo- bones (considered to occur independently in graphs of its collection and preparation. We Peltochelys and the carettochelyids), abdom- are grateful for his patience during the decade inal scales absent, incised anterior musk duct, of work on this turtle. This study is a result one suprapygal (occurs independently in the of Carter and Thorne fellowships from the Trionychia), plastral kinesis (occurs indepen- American Museum of Natural History to dently in the Trionychia), and never more P.A.M. Further support ofthe work was pro- than two inframarginals present (occurs also vided by N.S.F. grants DEB 800 2885 and in Basilemys). Kinosternids also lack com- BSR 831 4816 to E.S.G. and a short-term plete connections ofthe ribs ofthe ninth and visitors grant from the Smithsonian Institu- tenth body vertebrae to the eighth costal (fig. tion to P.A.M. 14). The presence oftwo articular surfaces on We thank H. Hutchison and G. Zug for the eighth costal in one of the best preserved reading and commenting on our work. Lor- specimens of Hoplochelys (USNM 8527) in- raine Meeker and Frank Ippolito have done dicates that at least the ninth vertebra was the artwork with their usual skill and accu- connected to the eighth costal in this genus. racy. We are grateful for their efforts. The second articular surface may have served as the site of contact of the tenth body ver- REFERENCES tebra, the ilium, or both. Characters that can at present be used to diagnose the Kinoster- Agassiz, L. nidae but are not known for Hoplochelys or 1857. North American Testudinata. In Con- frontal from or- tributions to the natural history of the Agomphus include: excluded United States 1(2): 233-452. Boston: bit, maxilla in contact with the quadratojugal Little, Brown. (occurs also in Carettochelys), absence of a Albrecht, P. W. maxillary tooth, reduction of the external 1967. The cranial arteries and cranial arterial process of the pterygoid, and the presence of foramina of the turtle genera Chryse- a pair of ventral processes of the eighth cer- mys, Sternotherus, and Trionyx: a com- vical (the last three characters occur inde- parative study with analysis of possible pendently in the Trionychia). evolutionary implications. Tulane Stud. The morphological data provide abundant Zool. 14: 81-99. evidence that Staurotypinae is not closely re- 1976. The cranial arteries of turtles and their evolutionary significance. J. Morphol. lated to Testudinoidea, as suggested by Bick- 149: 159-182. ham and Carr (1983). The single character Baur, G. supporting their arrangement, a group B mac- 1889a. Nachtriigliche Bemerkungen iiber die rosome to which a supposedly homologous systematische Stellung von Dermoche- microsome has fused, is called into question. lys Blainv. Biol. Centralbl., Bd. IX, 20- The conflict in the data could be resolved by: 21: 618-619. 1989 MEYLAN AND GAFFNEY: ADOCUS 57

1889b. The systematic position of Meiolania, perieur et du Tertiaire de France. Mem. Owen. Ann. Mag. Nat. Hist. 3: 54-62. Mus. Natl. Hist. Nat., ser. C, 38: 1-366. 1891 a. The pelvis ofthe Testudinata, with notes Chkhikvadze, V. M. on the evolution ofthe pelvis in general. 1975. [The volume and status ofthe turtles of J. Morphol. 4: 345-359. the suborder Amphichelydia Lydekker 1891b. Notes on some little known American 1889.] Bull. Acad. Sci. Georgian SSR fossil tortoises. Proc. Acad. Nat. Sci. 78(3): 745-748 [in Russian with English Philadelphia 43: 411-430. summary]. 1891c. On the relations ofCarettochelys, Ram- Cope, E. D. sey. Am. Nat. 25: 631-639. 1868. On some Cretaceous Reptilia. Proc. Bickham, J. W., J. J. Bull, and J. M. Legler Acad. Nat. Sci. Philadelphia 1868: 233- 1983. Karyotypes and evolutionary relation- 242. ships of trionychoid turtles. Cytologia 1870. On the Adocidae. Proc. Am. Philos. Soc. 48: 177-183. 11: 547-553. Bickham, J. W., and J. L. Carr 1871. Supplement to the "Synopsis of the ex- 1983. Taxonomy and phylogeny ofthe higher tinct Batrachia and Reptilia of North categories of cryptodiran turtles based America." Proc. Am. Philos. Soc. 12: on a cladistic analysis of chromosomal 41-55. data. Copeia 4: 918-932. 1874. [Description ofAdocus lineolatus.] Bull. Bienz, A. U.S. Geol. and Geogr. Surv. Terrs. 1(2): 1895. Dermatemys mavii Gray. Eine osteo- 30. logische Studie. Rev. Suisse de Zool. 3: Dollo, Louis 61-136. 1884. Premiere note sur les cheloniens de Ber- Boulenger, G. A. nissart. Bull. Mus. R. Hist. Nat. Bel- 1 gique 3: 63-84. 887a. On a new family of pleurodiran turtles. 1886. Premiere note sur les cheloniens du Ann. Mag. Nat. Hist. 19: 170-172. Bruxellien (Eocene moyen) de la Bel- 1 887b. On the systematic position of the genus gique. Bull. Mus. R. Hist. Nat. Belgique Miolania, Owen (Ceratochelys, Hux- 4: 75-96. ley). Proc. Zool. Soc. London, pp. 554- 555. Estes, R., P. Berberian, and C. A. Meszoely 1889. Catalogue of the 1969. Lower vertebrates from the late Creta- chelonians, rhyncho- ceous Hell Creek Formation, McCone cephalians, and crocodiles in the British County, Montana. Breviora 337: 1-33. Museum (Nat. Hist.). London, printed Fitzinger, L. J. by Order of the Trustees, new ed., 31 1 1826. Neue Classification der Reptilien. Wien: PP. J. G. Heubner Verlag, 66 pp. Bour, R., and A. Dubois Frair, Wayne 1985. Nomenclature ordinale et familiale des 1985. The enigmatic plateless river turtle, tortues (Reptilia). In F. DeBroin and E. Carettochelys, in serological survey. J. Jimenez-Fuentes, Studia Geologica Sal- Herpetol. 19(4): 515-523. amanticensia 1: 77-86. Gaffney, E. S. 1986. Nomenclature ordinale et familiale des 1972. The systematics of the North American tortues (Reptilia). Note complemen- family Baenidae (Reptilia, Cryptodira). taire. Bull. Soc. Linn. Lyon 55: 87-90. Bull. Am. Mus. Nat. Hist. 147: 241- Brim, H. 320. 1965. Die Schildkroten aus dem oberen Jura 1975. A phylogeny and classification of the (Malm) der Gegend von Solothum. higher categories of turtles. Bull. Am. Schweiz. Palaeontol. Abh. 83: 1-190. Mus. Nat. Hist. 155: 387-436. Bramble, D. M. 1976. Cranial morphology of the European 1974. Emydid shell kinesis: biomechanics and Jurassic turtles Portlandemys and Ple- evolution. Copeia 3: 707-727. siochelys. Bull. Am. Mus. Nat. Hist. 157: Bramble, D. M., J. H. Hutchison, and J. M. Legler 489-543. 1984. Kinosternid shell kinesis: structure, 1979. Comparative cranial morphology of re- function and evolution. Copeia 2: 456- cent and fossil turtles. Bull. Am. Mus. 475. Nat. Hist. 164: 65-375. Broin, F. de 1982. The lower jaws of baenid turtles. Am. 1977. Contribution a l'etude des Cheliniens. Mus. Novitates 2749: 10 pp. Cheloniens continentaux du Cretace su- 1984. Historical analysis of theories of che- 58 AMERICAN MUSEUM NOVITATES NO. 2941

lonian relationships. Syst. Zool. 33: 283- von Halberstadt. Serie II. Testudinata. 301. Teil 1. Stegochelys dux n. g., n. sp. Pa- 1985a. The cervical and caudal vertebrae ofthe laeontol. Zeit. 2: 88-214. cryptodiran turtle, Meiolania platyceps, Jimenez-Fuentes, E. from the Pleistocene of Lord Howe Is- 1971. Primer pseudotrionyux espanol: Al- land, Australia. Am. Mus. Novitates laeochelys casasecai, nov. sp. del Lu- 2805: 29 pp. teciense de Corrales (Zamora). Estudios 1985b. The shell morphology of the Triassic Geol., Inst. "Lucas Mallada" 27: 153- turtle Proganochelys. Neues Jahrb. Geol. 166. Palaeontol. Abh. 170: 1-26. King, M. In prep. The comparative osteology of the 1985. Chromosome markers and their use in Triassic turtle, Proganochelys. Bull. Am. phylogeny and systematics. In G. Grigg, Mus. Nat. Hist. R. Shine, and H. Ehmann, Biology of Gaffney, E. S., and P. A. Meylan Australasian frogs and reptiles, pp. 165- 1988. A phylogeny of turtles. In M. J. Benton 175. Sydney: R. Zool. Soc. New South (ed.), The phylogeny and classification Wales. of Tetrapods, vol. 1, Amphibians, rep- Khosatzky, L. I., and M. Mlynarski tiles, birds, pp. 157-219. Oxford: Clar- 1971. Chelonians from the Upper Cretaceous endon Press. ofthe Gobi Desert, Mongolia. Palaeon- Gilmore, C. W. tologica Polonica 25: 131-144. 1919. Reptilian faunas of the Torrejon, Puer- Khosatzky, L. I., and L. A. Nessov co, and underlying Upper Cretaceous 1977. Turtles of the genus Adocus from the formations of San Juan County, New late Cretaceous in the USSR. Acad. Sci. Mexico. U.S. Geol. Surv., Prof. Pap. USSR, Proc. Zool. Inst. 74: 116-119. 119: 68 pp. Kuznetsov, V. V. Hay, 0. P. 1976. A freshwater turtle from the Senonian 1902. Bibliography and catalogue ofthe fossil of northeastern Cisaralia. Paleontol. J. Vertebrata ofNorth America. Bull. U.S. 10(4): 498-500. Geol. Surv. 179: 886 pp. Langston, W., Jr. 1908a. The fossil turtles of North America. 1956. The shell ofBasilemys varialosa (Cope). Carnegie Inst. Washington Publ. 75: 1- Bull. Ann. Rep. Natl. Mus. Canada 142: 568. 155-163. 1908b. Descriptions of five species of North Leidy, J. American turtles, four ofwhich are new. 1865. A Cretaceous reptile ofthe United States. Proc. U.S. Natl. Mus. 35: 161-169. Smithsonian Contrib. Knowledge 192: 1910. Descriptions ofeight new species offos- 1-135. sil turtles from west of the one hun- 1870. Description ofBaptemys wyomingensis. dredth meridian. Proc. U.S. Natl. Mus. Proc. Acad. Nat. Sci. Philadelphia 38: 307-326. 1870: 5. Hirayama, R. Loveridge, A., and E. E. Williams 1985. Cladistic analysis of batagurine turtles 1957. Revision of the African tortoises and (Batagurinae: Emydidae: Testudinoi- turtles ofthe suborder Cryptodira. Bull. dea); a preliminary result. In F. DeBroin Mus.Comp.Zool. 115: 163-557. and E. Jimenez-Fuentes, Studia Geo- Lydekker, R. logica Salamanticensia 1: 141-157. Edi- 1889. Catalogue of the Fossil Reptilia and ciones Universidad de Salamanca. Amphibia in the British Musum (Nat- Hummel, K. ural History) Part III. The Order Che- 1929. Die fossilen Weichschildkroten (Tri- lonia. London: British Museum (Nat. onychia). Eine morphologische-syste- Hist.) by order of the Trustees, 239 pp. matische und stammesgeschichtliche Marsh, 0. C. Studie. Geol. Palaeontol. Abh. 16: 359- 1890. Notice ofsome extinct Testudinata. Am. 487. J. Sci. 40(3): 177-179. Hutchison, J. H., and D. M. Bramble McDowell, S. B. 1981. Homology of the plastral scales of the 1961. On the major arterial canals in the ear Kinosternidae and related turtles. Her- of testudinoid turtles and the classifi- petologica 37: 73-85. cation of the Testudinoidea. Bull. Mus. Jaekel, 0. Comp. Zool. 125: 23-39. 1918. Die Wirbeltierfunde aus dem Keuper 1964. Partition of the genus Clemmys and re- 1989 MEYLAN AND GAFFNEY: ADOCUS 59

lated problems in the taxonomy of London Clay, and on the Bracklesham aquatic Testudinidae. Proc. Zool. Soc. and other Tertiary beds. Part I. London 143: 239-279. Chelonia. Palaeontogr. Soc. Monogr. 2: Meylan, P. A. 1-78. 1985. Evolutionary relationships of recent Riggs, E. S. trionychid turtles: evidence from shell 1906. The carapace and plastron ofBasilemys morphology. In F. DeBroin and E. Ji- sinousus, a new fossil tortoise from the menez-Fuentes, Studia Geologica Sal- Laramie beds of Montana. Field Co- amanticensia 1: 169-188. Ediciones lumbian Mus., Publ. 110, Geol. Ser. 2(7): Universidad de Salamanca. 249-256. 1987. The phylogenetic relationships of soft- Seidel, M. E., J. B. Iverson, and M. D. Adkins shelled turtles (family Trionychidae). 1986. Biochemical comparisons and phylo- Bull. Am. Mus. Nat. Hist. 186: 1-101. genetic relationships in the family Kin- 1988. Peltochelys Dollo and the relationships osternidae (Testudines). Copeia 2: 285- among the genera of the Carettochelyi- 294. dae (Testudines: Reptilia). Herpetolog- Sukhanov, V. B., and P. Narmandakh ica 44: 440-450. 1974. [New Early Cretaceous turtle from con- Mlynarski, M. tinental deposits of the northern Gobi.] 1972. Zangerlia testudinimorpha n. gen., n. sp. Mesozoic and Cenozoic Faunas and a primitive land tortoise from the Up- biostratigraphy of Mongolia. Joint So- per Cretaceous ofMongolia. Palaeontol. viet-Mongolian Paleontological Expe- Polonica 27: 85-92. dition. Trans. 1: 192-220 [English ab- 1976. Testudines. In 0. Kuhn, Encyclopedia stract, p. 367]. of Paleoherpetology 7: 1-130. Stuttgart: Swofford, D. L. Gustav Fischer Verlag. 1984. PAUP. Phylogenetic analysis using par- Narmandakh, P. simony. Version 2.3. Unpubl. docu- 1985. A new chelonian species of the genus mentation. Champaign, Ill. Adocus from the upper Cretaceous of Whetstone, K. N. Mongolia. Paleontol. Zh. 2: 85-93 1978. A new genus ofcryptodiran turtles (Tes- [translation]. tudinoidea, Chelydridae) from the up- Nessov, L. A. per Cretaceous Hell Creek Formation of 1977. [Skull morphology of the Early Creta- Montana. Univ. Kansas Sci. Bull. 51: ceous turtle belonging to the family 539-563. Adocidae.] In N. B. Ananjeva, K. L. White, R. S. Borkin, and I. S. Darevsky, Gerpetol. 1972. A recently collected specimen ofAdocus Sborhik, Trudy Zool. Inst., Akad. Nauk. (Testudines; Dermatemydidae) from USSR 74: 75-80. New Jersey. Notulae Naturae 447: 1- 10. 1986. Some late Mesozoic and Paleocene tur- Wieland, R. tles of Soviet Middle Asia. Studia Geo- 1904. Structure of the Upper Cretaceous tur- logica Salmanticensia 2(1): 7-22. tles of New Jersey: Adocus, Osteopygis Nessov, L. A., and V. A. Julinen and Propleura. Am. J. Sci. 17: 112-132. 1977. On phylogenetic relations and history Williams, E. E. of spreading of some families of conti- 1950. Variation and selection in the cervical nental turtles. [English translation from central articulations of living turtles. the Russian by Nessov in letter to Dr. Bull. Am. Mus. Nat. Hist. 94: 505-562. Gaffney]. In Life on ancient continents, Williams, E. E., and S. B. McDowell its appearance and development, pp. 54- 1952. The plastron ofsoft-shelled turtles (Tes- 55. Theses of communications of 23 tudinata, Trionychidae): a new inter- sessions of All-Union Paleontological pretation. J. Morphol. 90(2): 263-280. Society, Leningrad (21-25 March 1977). Yeh, H.-K. Owen, R. 1966. A new Cretaceous turtle of Nanhsiung, 1881. Description of some remains of the gi- northern Kwangtung. [In Chinese and gantic land lizard (Megalania prisca, English]. Vertebrata Palasiatica 10: 191- Owen) from Australia. Part II. Philos. 200. Trans. R. Soc. London 171: 1037-1050. Zangerl, R. Owen, R., and T. Bell 1969. The turtle shell. In C. Gans, A. d'A. 1849. Monograph on the fossil Reptilia of the Bellairs, T. Parsons, Biology ofthe Rep- 60 AMERICAN MUSEUM NOVITATES NO. 2941

tilia 1: 311-339. New York: Academic 1971. Buoyancy, locomotion, morphology of Press. the pelvic girdle and hindlimb, and sys- Zug, G. R. tematics of cryptodiran turtles. Misc. 1966. The penial morphology and the rela- Publ. Mus. Zool. 142: 1-98. tionships of cryptodiran turtles. Occas. Pap. Mus. Zool. 647: 1-24.

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, N.Y. 10024.

I THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. I