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

Home , Macrobaenidae, Sinemys

NorntatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3048, 13 pp., 6 figures, 1 table July 28, 1992

Dracochelys, a New Cryptodiran Turtle from the Early Cretaceous of China

EUGENE S. GAFENEY' AND XIANGKUI YE2

ABSTRACT A turtle skull from the Early Cretaceous Tugulu internal nares, pterygoid forming a cleft-shaped series in the Wuerho district ofXinjiang province, opening for the palatine artery, and flattened skull People's Republic of China, is a new genus of eu- show a close similarity to Hangaiemys, usually cryptodiran turtle. Dracochelys bicuspis is unique placed in the Sinemydidae/Macrobaenidae. How- among cryptodires in possessing a very narrow ever, we place Dracochelys as Eucryptodira incer- maxillary triturating surface with paired cusps on tae sedis, recognizing that the "Macrobaenidae" the labial ridge at the premaxilla-maxilla suture. is presently diagnosed only by characters primitive The very large foramen palatinum posterius, large for Eucryptodira.

INTRODUCTION In 1973, Ye figured and briefly mentioned ilar to Hangaiemys, described by Sukhanov an "amphichelydian skull" from the Early and Narmandakh (1974) from the Early Cre- Cretaceous Tugulu series in the Wuerho dis- taceous (Aptian-Albian) of Mongolia (Mon- trict of Xinjiang province, People's Republic golian People's Republic). Hangaiemys is of China. Further preparation and study of known from well-preserved skulls and shells, this specimen shows that it is a new genus of and, on the basis of the shell morphology, cryptodire. The purpose of this paper is to has been placed in the family Macrobaenidae describe and name this new taxon. (Sukhanov, 1964). Macrobaenidae has been The new genus, Dracochelys, is known from synonymized with the family Sinemydidae a single well-preserved skull that is most sim- by some authors (e.g., Ckhikvadze, 1987),

' Curator, American Museum of Natural History. 2 Institute of Vertebrate Paleontology and Paleoanthropology, Beijing.

Copyright © American Museum of Natural History 1992 ISSN 0003-0082 / Price $1.90 2 AMERICAN MUSEUM NOVITATES NO. 3048 but current work (Brinkman, in progress) sug- Drs. D. Brinkman, L. Nessov, and D. Russell gests that the Sinemydidae can be defined for providing translations of many of the with synapomorphies while the Macrobaen- Russian and Chinese papers. idae or Macrobaenidae plus Sinemydidae The Institute for Paleontology and Paleo- cannot, at least as currently diagnosed (see anthropology, Beijing, allowed us access to discussion). Although the purpose ofthis pa- the specimen described here and greatly aid- per is not to determine the relationships of ed the studies of one of us (ESG) while in Dracochelys, it is certainly prudent to com- Beijing. We particularly thank Dr. Chang pare Dracochelys with possible near relatives. Mee-Mann for her encouragement and sup- We have chosen four taxa (table 1) for these port. comparisons, based on similarity to Dra- We are grateful to Ms. L. Meeker for the cochelys, purported relationships, and avail- high quality of the figures and for fine prep- ability of skull material. aration of the type specimen. Our knowledge ofHangaiemys and Macro- We also thank Drs. H. Hutchison, D. baena is based solely on the papers describing Brinkman, and P. Meylan for reading and and figuring them, Sukhanov and Narman- improving the manuscript. dakh (1974) and Tatarinov (1959), respec- tively. Information on Sinemys sp. is derived primarily from skulls (IVPP V9532-1 1 from ABBREVIATIONS Laolonghuoze, and IVPP V9538 from Otog Anatomical Qi, Inner Mongolia, of undescribed species ofthat genus; Brinkman, personal commun.). bo basioccipital The fourth skull, TMP 87.2.1, represents an bs basisphenoid undescribed genus from the Late Cretaceous epi epipterygoid ex exoccipital of Dinosaur Provincial Park, Alberta, prob- fr frontal ably similar to but not the same as "Clem- ju jugal mys" backmani Russell, 1934 (Hutchison, mx maxilla personal commun.). While our knowledge of na nasal Hangaiemys and Macrobaena is based only op opisthotic on the literature, one of us (ESG) has had the pa parietal opportunity to examine the Sinemys skulls pal palatine and TMP 87.2.1. pf prefrontal Some regions of the skull in Dracochelys pm premaxilla cannot yet be determined in the fossil taxa; Po postorbital pr prootic therefore, comparisons are also made with pt pterygoid chelydrids. Dracochelys is most similar to qj quadratojugal Chelydra among the Recent fauna, and skulls qu quadrate of that genus are widely available. so supraoccipital sq squamosal vo vomer ACKNOWLEDGMENTS Institutional We are particularly grateful to Drs. How- IVPP Institute of Paleontology and Paleoan- ard Hutchison and Don Brinkman, who made thropology, Beijing available many undescribed specimens, which PRC People's Republic of China they collected with some difficulty. Both TMP Royal Tyrrell Museum of Paleontology, workers have invested considerable time and Drumheller effort in studying this material, and we ap- preciate their unrestricted sharing of specimens and conclusions with us. The paper was SYSTEMATICS considerably improved by a meeting in Feb- ORDER TESTUDINES ruary 1992 of the Asian Study Section (ASS) of the World Paleochelonological Society GIGAORDER CASICHELYDIA (WPS) in Berkeley, CA. We are grateful to MEGAORDER CRYPTODIRA 1 992 GAFFNEY AND YE: DRACOCHELYS 3

PARVORDER EUCRYPTODIRA INCERTAE SEDIS Dracochelys bicuspis cf FAMILY "MACROBAENIDAE" TYPE SPECIMEN: Institute ofVertebrate Pa- leontology and Paleoanthropology (IVPP) Dracochelys, new genus V4075 (figured in Ye, 1973, p1. 1, figs. 3, 4, TYPE SPECIES: Dracochelys bicuspis, new as "amphichelydian skull"), collected 1964 genus and species. by IVPP expedition. KNoWN DISTRIBUTION: Wuerho district, LocALrry: Wuerho district, northwestern Xinjiang province, People's Republic ofChi- part of Dzungar (Junggar) Basin, north Xin- na, Early Cretaceous, Tugulu Series. jiang province, People's Republic of China ETYMOLOGY: Draco, dragon; chelys, turtle. (see Dong, 1973). DIAGNOSIS: Member of Eucryptodira, HoRIzON: Upper part (possibly equivalent known only from skull, with canalis caroticus to the Lianmuxin Fm.) of the Tugulo series internus only partially floored by bone and of lacustrine sediments, thought to be Early foramen caroticum basisphenoidale com- Cretaceous. Dracochelys and "Sinemys" pletely formed by basisphenoid and visible wuerhoensis both come from the Upper Tu- in ventral view as in Sinemys, but in contrast gulu series but from different horizons and to most eucryptodires in which canalis caroti- localities (Dong, 1973). cus internus is completely floored by bone DIsCussIoN: Dracochelys is clearly a cryp- and foramen caroticum basisphenoidale is todire: it has the otic trochlea synapomorphy formed at least in part by pterygoid; skull of that group (Gaffney and Meylan, 1988; most similar to those of Chelydra and Han- Gaffhey et al., 1991), and it is here interpreted gaiemys in shape but relatively flatter and as a eucryptodire. The critical synapomorphy wider, with fossa orbitalis and apertura nar- for eucryptodires is that the posterior portion ium externa opening dorsally to a much of the pterygoid encloses the internal carotid greater extent; frontal bone enters orbit in artery. In Dracochelys the bone forming the contrast to that ofchelydrids; degree of tem- floor of the canalis caroticus internus is very poral emargination similar to that in Han- thin, much as in plesiochelyids, and does not gaiemys and Chelydra; degree ofcheek emar- seem to extend for most of the length of the gination as in Chelydra; triturating surfaces groove containing the internal carotid. How- distinct from those of all chelydrids, Sine- ever, this is not definite, as the bone making mys, Hangaiemys, Macrobaena, and Macro- up the floor is broken and missing in many baenidae (sensu Hutchison and Archibald, places, so the floor could be more extensive 1986) in being very narrow with paired cusps than shown in the restoration (fig. 5). The at the premaxilla-maxilla suture; foramen limited amount of bone in the floor of the palatinum posterius relatively large in com- canalis could be interpreted as primitive parison with that of Chelydra but similar to within Eucryptodira, but nonetheless Dra- that ofHangaiemys; vomerine ridge extend- cochelys would be a eucryptodire. ing length of bone as in Hangaiemys but in Within the Eucryptodira the relationships contrast to Chelydra; processus trochlearis ofDracochelys are more uncertain. It is most oticum relatively small in contrast to that of similar to a series ofAsian and North Amer- chelydrids; incisura columellae auris narrow ican eucryptodires that have come to be re- and containing only stapes but not complete- ferred to the family Sinemydidae and/or ly closed by bone as in chelydrids; vertical Macrobaenidae. The Sinemydidae was cre- plate ofprocessus pterygoideus externus larg- ated by Ye in 1963 and Macrobaenidae by er than in any other turtle; pterygoid-basi- Sukhanov in 1964. The perturbations and occipital contact seen in chelydrids and most usage ofthese two names is beyond the scope cryptodires greatly reduced in Dracochelys by of this paper, and we will refer only to the posterolateral process of basisphenoid; deep most recent diagnosis by Ckhikvadze (1987). concavity on quadrate process of pterygoid Ckhikvadze proposed the most inclusive def- as in Hangaiemys but in contrast to that of inition, uniting the two families in Sinemydi- Sinemys, TMP 87.2.1, and chelydrids; differs dae. He included only Asian taxa, but Hutch- from Sinemys wuerhoensis in being more than ison in Hutchison and Archibald (1986) and twice the size and having a much wider skull. in McKenna et al. (1987) has identified spec- 4 AMERICAN MUSEUM NOVITATES NO. 3048

A Pt fr fossa orbitalis qj .c -

apertura -L narium externa nasomaxillary sinus endocast B

2cm

foramen 'condylus stapedio-temporale occipitalis Fig. 1. Dracochelys bicuspis, described and di- agrammed under figure 2. Fig. 2. Dracochelys bicuspis, new genus and species, IVPP V4075, Early Cretaceous, Xinjiang imens as belonging to the Sinemydidae and province, PRC. A, Anterior view; B, right lateral Macrobaenidae (using the names as syn- view; C, dorsal view. onyms) in North America, one ofwhich was earlier described as "Clemmys" backmani ent time as mostly, if not solely, primitive at (Russell, 1934). There has been no diagnosis the level of Eucryptodira. The characters are of this family/families including the North all shell features and primarily refer to the American taxa, but Ckhikvadze (1987) di- cruciform plastron without medially extend- agnosed the combined families consisting of ing buttresses. This shell pattern also occurs the Asian taxa. His diagnosis consists ofchar- in more advanced eucryptodires, and has been acters that can best be interpreted at the pres- retained in the chelonioids in such derived 1992 GAF'FNEY AND YE: DRACOCHELYS 5

TABLE 1 Comparison of Dracochelys with Similar Genera Character Dracochelys Hangaiemys TMP 87.2.1 Sinemys Macrobaena Skull relatively flat and wide yes ?no no yes no Nasals present indet indet yes yes indet Fossa orbitalis opens dorsally yes yes no yes no Fossa nasalis opens dorsally yes no ?no no no Prefrontals meet on midline yes ?yes yes no yes Frontal enters orbit yes indet yes yes ?no Extent of temporal emargination moderate moderate moderate greater than much less Dracoche- than Dra- lys cochelys Postorbital widely exposed on tem- yes indet no yes ?no poral margin Maxillary triturating surface relative- yes yes ?no no no ly narrow Paired cusps on premaxilla-maxilla yes no no no no suture Foramen palatinum posterius rela- yes yes no yes no tively large Processus trochlearis oticum and yes no no no no otic chamber relatively small Processus pterygoideus externus very yes no no no no large Canalis caroticus internus only par- yes yes yes yes indet tially floored by bone Foramen carotico-pharyngeale absent absent present absent indet Path of palatine artery exposed ven- yes yes no yes yes trally Foramen caroticum basisphenoidale yes ?yes no yes indet completely formed by basisphenoid and visible in ventral view Pterygoid-basioccipital contact small yes indet no yes indet or absent Basisphenoid with posterolateral yes indet no yes indet processes Deep concavity on quadrate process yes yes no no ?no of pterygoid cheloniids as Osteopygis (phylogeny based on taxa, and is best identified as Eucryptodira, skull features in Gaffhey and Meylan, 1988). incertae sedis. Recently, Don Brinkman (personal commun.) has discovered specimens that may DEsCRIPTION show Sinemydidae to be a monophyletic group (consisting ofSinemys, Manchuroche- PRESERvATION: IVPP V4075 consists of a lys, and newly discovered taxa). The remain- partial skull with well-preserved bone. The ing "macrobaenids," once skulls, vertebrae, palate is almost complete but the skull roof etc., are described, may prove to be mono- is extensively damaged. However, the matrix phyletic or, more likely, a mixture of taxa, underlying the missing bone retains the im- some monophyletic and some belonging to pression of the internal surface of the bone more advanced eucryptodiran groups. Dra- in many places and provides information on cochelys does not have obvious synapomor- bone shape and limits. Although the skull has phies in common with known eucryptodiran been subjected to a certain degree of dorso- 6 AMERICAN MUSEUM NOVITATES NO. 3048 ventral compression, the generally flattened faces dorsally to a much greater extent than appearance is most likely the original shape. in chelydrids where it opens almost directly The crista supraoccipitalis shows no signs of anteriorly. The type skull of Dracochelys is dorsoventral telescoping and the ear region probably flattened slightly but this would not does not appear to be crushed, supporting cause the apertura opening to be so different this interpretation. The right quadrate artic- from that of chelydrids. ulation appears to have been pushed laterally The ventral process ofthe prefrontal reach- but not to a great extent. es the vomer. The prefrontal and the other DERMAL RoOFING ELEMENTS: The bone in bones forming the fossa orbitalis are not vis- the anterior portion of the snout is missing ible in IVPP V4075 because both orbits are in most ofthe area where nasals would occur, entirely filled with matrix. The orbits open and small ones could have been present. There dorsally a bit more in Dracochelys than in is nothing interpretable as nasals. The un- Sinemys and a great deal more dorsally than derlying matrix forms a good impression of in Macrobaena, and much as in Hangaiemys. the ventral surface of the missing bone, and The frontal ofDracochelys is roughly rect- there is no sign of a suture separating nasal angular and similar in shape to that of che- and prefrontal. Sinemys sp. has nasals, Mac- lydrids and in contrast to the triangular fron- robaena is described as having nasals, and tal ofSinemys sp. Only the medial halfofthe chelydrids lack nasals. The other fossil taxa left frontal is preserved but on the right side are indeterminant. only the posterolateral corner is missing. In The prefrontal is preserved in part on both Dracochelys the frontal enters the orbit as in sides. Both prefrontals are missing much of TMP 87.2.1, and Sinemys sp., but in contrast their anterior region but the right one shows to Macrobaena and chelydrids. In Protoche- most of the contacts. Assuming nasals are lydra, a primitive chelydrid, the prefrontal absent, the prefrontal forms the dorsal mar- and postorbital meet to exclude the frontal gin of the apertura narium externa, contacts from the orbit. the maxilla anterolaterally, forms the antero- The parietal ofIVPP V4075 is incomplete dorsal margin of the orbit, and contacts the on both sides. The lateral third is missing on frontal posteriorly. The prefrontal appears to the left side, but only a small part of the lat- contact its mate for the entire length of the eral limits are missing on the right. The pos- bone. The prefrontal of Dracochelys is less terior margin is preserved for a limited sec- extensive than in living chelydrids due to the tion on the right side, but the edge ofthe skull larger dorsal process ofthe maxilla, but oth- roof emargination is seen clearly on the ma- erwise the prefrontals are very similar. Dra- trix impression of the ventral surface of the cochelys differs significantly from Sinemys sp. parietal (fig. 1). The parietal in Dracochelys in the degree ofprefrontal exposure dorsally. is very similar to that bone in Chelydra and In Sinemys sp. the prefrontal is only a rela- Sinemys sp. The temporal emargination of tively small lappet bordering the fossa orbi- Dracochelys is restored from the matrix im- talis and separated on the midline by the print. The precise limits of the parietal and frontal. This is similar to the condition in postorbital are not determinable but we have baenids, but the lappet is smaller in baenids. confidence in the approximate position ofthe In Dracochelys the prefrontals meet on the skull roof margin. As restored, Dracochelys midline for their entire length and are not has a relatively well-developed temporal separated by the frontals. The prefrontals in emargination with the parietal and postor- Macrobaena are figured as meeting on the bital well exposed. Sinemys sp. is more emar- midline for their length. TMP 87.2.1 is in- ginate, Macrobaena has virtually no emar- determinant. gination, and TMP 87.2.1 is less emarginate The anterior expansions of the sulcus ol- than Dracochelys. In TMP 87.2.1 there is a factorius that form the roof of the fossa na- parietal-squamosal contact, very different salis are preserved as molds in the matrix from the exposed postorbital as restored in filling the skull of IVPP V4075. These ex- Dracochelys. pansions are similar to those in chelydrids. The jugal is incomplete on both sides of The apertura narium externa in Dracochelys IVPP V4075, the posterior portions being lost 1 992 GAFFNEY AND YE: DRACOCHELYS 7 in both. The jugal widely enters the orbital terior view the labial ridge has paired ven- margin between maxilla and postorbital, to trally projecting "teeth" formed by the pre- a slightly greater extent than in Chelydra but maxillae plus maxilla. Medially, the similar to that in Sinemys and TMP 87.2.1. premaxillary labial ridge rises higher than the Based on the position of a fragment of the level of the maxillary labial ridge to form a quadratojugal and postorbital suture, it is midline notch. Two labial cusps are uncom- likely that the jugal in Dracochelys extended mon in turtles but do occur in some forms, posteriorly to some extent as in Chelydra and such as Dermochelys and Pseudemys nelsoni Protochelydra rather than being restricted as (Ernst and Barbour, 1972). However, this in Sinemys sp. and TMP 87.2.1. However, morphology does not occur in Hangaiemys, this cannot be determined definitely. The ex- Sinemys sp., TMP 87.2.1, Macrobaena, che- tent of cheek emargination is also unclear, lydrids, plesiochelyids, or any non-eucryp- but the matrix on the left side shows an im- todire. On the ventral surface both premax- pression that strongly suggests that it was rel- illae form the anterior half of a circular atively extensive as in Protochelydra and concavity or commissural depression on the Chelydra. The medial process of the jugal is midline ofthe triturating surfaces, similar to only visible as a small area adjacent to the that seen in other chelydrids and suggesting maxilla-pterygoid contact in ventral view. that Dracochelys may have had a midline The quadratojugal is entirely absent on the "hook" or process on the lower jaw. The fo- left side and preserved only as a fragment on ramen praepalatinum is formed almost en- the right. The fragment contains the postor- tirely by the premaxilla with a small contri- bital-quadratojugal suture and is in a position bution from the vomer. consistent with a relatively small, C-shaped Both maxillae of IVPP V4075 are pre- quadratojugal, the generalized cryptodiran served almost completely; only a small por- condition found in primitive chelydrids. tion ofthe labial ridge is missing. The vertical Squamosals are lacking in IVPP V4075, part ofthe maxilla is an anterodorsal process except for a possible fragment in what ap- forming the anterior rim of the orbit and pears to be the quadrate-squamosal suture on reaching dorsally to meet the prefrontal. This the right side. process is somewhat more extensive in Dra- The postorbital in IVPP V4075 is absent cochelys than in Sinemys sp. and chelydrids, except for an impression ofits ventral surface being visible in dorsal view in Dracochelys on the left side, and present on the right side but not in chelydrids. The general flattening for about half of its length anteriorly. The of the skull in Dracochelys accentuates this postorbital is a large element forming the pos- condition but the maxilla still extends dor- terodorsal margin ofthe orbit between frontal sally more than in chelydrids. andjugal. Most ofthe parietal-postorbital su- The posterolateral corner of the fossa na- ture is preserved but a small section of its salis has a shallow but distinct sinus that can posterior end is missing. Nonetheless, the be seen in the matrix on the left side. The medial edge ofthe postorbital in Dracochelys dorsal part of the maxilla seems to form this seems to have had a shape very similar to sinus. that of Chelydra. The lateral and posterior The horizontal part of the maxilla bears limits of the postorbital, however, are much the triturating surfaces. In Dracochelys the more ambiguous and it is only the small frag- labial and lingual ridges are parallel and very ment of quadratojugal-postorbital contact close to each other, being separated by a shal- that allows some idea of its extent. At least low but distinct trough. The narrowness of some of the temporal emargination was the triturating surface in Dracochelys is ex- formed by the postorbital but how far pos- treme, and among the taxa being compared teriorly it extended cannot be determined. As here, only Hangaiemys approaches it. The restored, the postorbital is most similar in triturating surface ofDracochelys is relatively size and extent to that in Chelydra. narrower than the narrow-jawed batagurids PALATAL ELEMENTS: Both premaxillae are and any other turtle. The labial ridge forms preserved in IVPP V4075 and are seen in a large cusp or "tooth" at the maxilla-pre- anterior and ventral views (figs. 1, 3). In an- maxilla suture and extends posteriorly as a 8 AMERICAN MUSEUM NOVITATES NO. 3048

straight ridge that ends at the posterior edge ium intema, contacts the prefrontal, and ofthe maxilla. The labial ridge ofDracochelys forms the medial margin of the foramen or- is comparable in relative height and sharp- bito-nasale. Anterolaterally the palatine has ness to the labial ridge in Chelydra. The lin- a process that contacts the maxilla between gual ridge in Dracochelys is lower than the the apertura narium interna and the foramen labial ridge. The lingual ridge begins near the palatinum posterius. The palatine forms the maxilla-premaxilla suture at the margin of medial halfofthe margin ofthe large foramen the midline concavity formed just posterior palatinum posterius. The entire width of the to the labial ridge, but it trends away from palatine posteriorly is a transverse suture with the labial ridge along the palatine-maxilla su- the pterygoid. Medially the palatine has a long ture. suture with the vomer. The palatine of Dra- Anteromedially the maxilla has a long su- cochelys agrees in most of its features with ture with the premaxilla, and medially a short that of Chelydra but it is relatively wider, suture with the vomer just behind the pre- more like that of Hangaiemys. maxillary contact, as in most turtles. The short PALATOQUADRATE ELEMENTS: The left process ofthe maxilla that curves posteriorly quadrate is represented only by the medial- to reach the vomer forms the anterior margin most portion ofthat bone, but the right quad- ofthe apertura narium interna. In Dracoche- rate has much of the cavum tympani and lys the apertura is unusually large, compar- processus articularis. The quadrate of Dra- atively much larger than in chelydrids but cochelys is similar to that bone in Chelydra. similar to the size of the apertura in Han- The medial part contacts the pterygoid ven- gaiemys. The position and shape of the pal- trally and the prootic dorsally. As in Chely- atine-maxilla contact is similar to those in dra, the quadrate forms most ofthe processus Chelydra and most turtles. The foramen pal- trochlearis oticum, but in contrast to the large atinum posterius is formed laterally and pos- processus characteristic of chelydrids, this terolaterally by the maxilla. The foramen in structure is unusually low and hardly delim- Dracochelys is unusually large, much larger ited from the surrounding bone in Dracoche- than in Sinemys sp., Macrobaena, TMP lys. The processus is indeterminant in the 87.2.1, and chelydrids, but similar in size to published figures ofHangaiemys but appears that in Hangaiemys. Posteriorly the maxilla large in TMP 87.2.1 and in Sinemys sp. On has a roughly transverse suture with the pter- the dorsal surface the quadrate contacts the ygoid. prootic anteriorly and the opisthotic poste- The vomer is complete in IVPP V4075 but riorly. As in Chelydra the foramen stapedio- it is visible only in ventral view. The vomer temporale is formed in the prootic-quadrate is generally similar to that in Chelydra in hav- suture close to the opisthotic contact. Pos- ing the normal chelonian anterior expansion teromedially the opisthotic would be expect- with premaxillary and maxillary contacts, and ed to extend a process posterolaterally along a posterior extension separating the palatines the quadrate but this process is broken offon and reaching the pterygoids posteriorly. In both sides, exposing the quadrate sutural sur- Dracochelys as well as in Hangaiemys and face. TMP 87.2.1, the ventral midline ridge ofthe The pterygoid has a process that extends vomer is relatively deep and continues for along the ventral surface ofthe quadrate pos- the length ofthe bone, rather than being shal- terolaterally in Dracochelys as in all selma- low and restricted to the anterior portion as cryptodirans. The position of the pterygoid- in chelydrids. The former condition also oc- quadrate suture in Dracochelys is the same curs in plesiochelyids and may be primitive as in TMP 87.2.1, Sinemys sp., and Chelydra. for eucryptodires. Anterodorsally the lateral The processus articularis of the quadrate is contacts of the prefrontals with the vomer positioned slightly farther posteriorly in Dra- can be seen in the roofofthe apertura narium cochelys than in Chelydra, resembling Mac- interna. roclemys to a certain extent. Only the right The palatines on both sides ofIVPP V4075 condylus mandibularis is preserved and it is are preserved almost entirely. Anteriorly the the same as in Chelydra. palatine forms the roof of the apertura nar- The cavum tympani ofIVPP V4075 is pre- 1 992 GAFFNEY AND YE: DRACOCHELYS 9 served only in part on the right side. The dially. The processus pterygoideus externus dorsal section, including all ofthe squamosal, is very large in Dracochelys with a huge ver- and the anteroventral parts are missing. One tical plate. The size of the processus and ex- section of the rim of the cavum tympani is tent ofthe plate are greater than in any other preserved anterodorsally and this provides turtle. It is possible that the large size of the an idea of the size. The cavum tympani in processus pterygoideus externus and the very Dracochelys is relatively larger than in Chel- narrow triturating surface are correlated be- ydra. A small outward-facing concavity (in- cause the processus bears the mundplatte and dicated by an asterisk in fig. 2) lies at its an- guides the lower jaw. A narrow triturating terior margin; this concavity is absent in surface may require a more restricted area of Chelydra and Sinemys sp. Sinemys does have lateral movement when the jaws are closed. a deep concavity in the ventral part of the Medially both pterygoids in Dracochelys cavum tympani, but the cavum in Sinemys are in contact anteriorly, but for a much is smooth in the region of the concavity in shorter length than in Chelydra and most oth- Dracochelys. The cavum tympani is not vis- er eucryptodires. This is because the basi- ible in Hangaiemys. Although the upper part sphenoid in Dracochelys is exposed ventrally ofthe incisura columellae auris is broken pos- to a much greater extent. Posteriorly the pter- teriorly, it can be seen from the lower area ygoid-basisphenoid suture runs into an area that the incisura was not closed as in chely- of broken bone and then emerges postero- drids. The incisura does not seem to have laterally showing that the basisphenoid has a been widely open in Dracochelys and this posterolateral extension not seen in chely- condition seems close to Sinemys sp. and drids, Sinemys sp., and TMP 87.2.1 (inde- TMP 87.2.1. The area of the antrum posto- terminant in Macrobaena and Hangaiemys). ticum formed by the quadrate is visible on The anterior part of the pterygoid-basi- the right side of IVPP V4075, and it is very sphenoid suture shows a vertical separation similar in Chelydra. ofthe basisphenoid to form a cleftlike open- The ethmoid region of IVPP V4075 is not ing on each side. This opening would appear well prepared but the ventral part of the re- to be for the palatine artery which in most gion is visible. An epipterygoid can be seen, eucryptodires runs between the pterygoid and best on the left side, and with ventral limits basisphenoid to emerge lateral to the rostrum very similar to those of Chelydra. The an- basisphenoidale. In Dracochelys the palatine terodorsal-posteroventral ridge found below artery would appear to have branched at or the foramen nervi trigemini in most turtles near the foramen basisphenoidale. Although is very well developed in Dracochelys, and is it is hard to be sure from the published pho- much more pronounced than in Chelydra. tographs, it appears that Hangaiemys has a The foramen nervi trigemini ofIVPP V4075 posteromedial flange on the pterygoid, essen- is visible completely on the left side but only tially an extension of the cleft seen in Dra- ventrally on the right. It is relatively small cochelys that nearly covers the path of the compared with that in Chelydra, about one- palatine artery. Sinemys sp. could be inter- third the width. It is possible that crushing preted as being intermediate in this area with has reduced the width slightly. Although the more of the palatine artery path exposed. sutures around the foramen nervi trigemini TMP 87.2.1 has no exposed path of the pal- are not clear dorsally, it appears that the epi- atine but has a relatively small foramen in pterygoid forms the anteroventral margin and the pterygoid as well as a eucryptodiran pter- the prootic the posterior margin. The entry ygoid entrance of the internal carotid artery. ofthe parietal and the extent ofthe pterygoid The condition ofthis pterygoid flange in Dra- in the margin are not determinable. cochelys is similar to that postulated for the Both pterygoids are preserved in IVPP meiolaniid Crossochelys by Gaffiiey (1983) V4075 and they have the general form of and may be primitive for cryptodires. selmacryptodiran pterygoids. Anteriorly the A careful examination ofthe basisphenoid- pterygoid has a roughly transverse margin, pterygoid contact area in IVPP V4075 shows meeting the maxilla anterolaterally, the pal- that there is a eucryptodiran foramen pos- atine anteriorly, and the vomer anterome- terior canalis carotici intemi and its associ- 10 AMERICAN MUSEUM NOVITATES NO. 3048

.: N

Fig. 3. Dracochelys bicuspis, new genus and species, IVPP V4075, Early Cretaceous, Xinjiang province, PRC. Stereophotograph of ventral view. '.~~~l

Fig. 4. Dracochelys bicuspis, new genus and species, IVPP V4075, Early Cretaceous, Xinjiang province, PRC. Stereophotograph of occipital view. ated canal. The ventral floor of the canal is is very small or absent. A process ofthe pter- very thin and broken but its presence is es- ygoid curves medially around the basisphe- tablished. The canal is relatively short, how- noid to nearly reach the exoccipital. ever, and does not seem to extend all the way The posterolateral or quadrate process of to the foramen caroticum basisphenoidale, the pterygoid in Dracochelys has a well-de- similar to the condition in Sinemys sp. Al- veloped concavity (indicated by an asterisk though the canalis caroticus internus seems in fig. 5) opening ventrally and partially over- to be formed mostly by the pterygoid, the hung medially. This concavity lies in an area sutures are not clear and it is possible that usually occupied by the attachment of part the basisphenoid forms some part of it. Be- of the pterygoideus muscle in recent cryp- cause of the broken condition of the canalis todires and this may have been the case in caroticus internus, the exact position of the Dracochelys. A similar concavity is present foramen posterior canalis carotici interni in in Hangaiemys but not in chelydrids, Sine- Dracochelys could be more posterior than in- mys sp., or TMP 87.2.1. dicated. BRAnNCASE ELEMENTS: The supraoccipital Due to the posterolateral processes of the is visible only on the midline just above the basisphenoid in Dracochelys, the pterygoid- foramen magnum in IVPP V4075. Most of basioccipital contact seen in most cryptodires the crista supraoccipitalis is broken off, but 1 992 GAFFNEY AND YE: DRACOCHELYS I1I

E|labinatgera praepalatinum

I labial ridge mxforamen orbito-nasale lingual ridge foramen a.\ / ug ~~~palatinum posterius foramen caroticum pal altnmpotru laterale processus foramencaroticum pt pterygoideus externa foramen caroticum " basisphenoidale , bs fossa cartilaginis epipterygoidei

foramen posterius X foramen nervi hypoglossi canalis caroti interni condylus occipitalis Fig. 5. Dracochelys bicuspis, new genus and species, IVPP V4075, Early Cretaceous, Xinjiang province, PRC. Ventral view, partially restored. Asterisk is concavity referred to in text.

Op U-so foramen stapedio-temporale El Nj4Ut

fenestra _ postotica I* __.. , foramen condylus magnum mandibularis foramen jugulare foramen posterius nervi hypoglossi Fig. 6. Dracochelys bicuspis, new genus and species, IVPP V4075, Early Cretaceous, Xinjiang province, PRC. Occipital view. at the position of the foramen magnum it is base of the condylus occipitalis, much as in slightly lower than in Chelydra. The ventral Chelydra. As in chelydrids, the ventral flange expansion of the supraoccipital containing of the exoccipital reaches the pterygoid. In part ofthe cavum labyrinthicum contacts the Dracochelys, due to the posterolateral pro- opisthotic laterally and the exoccipitals on cesses of the basisphenoid, the exoccipital either side of the foramen magnum in Dra- comes very close and seems to just contact cochelys, much as in Chelydra. the basisphenoid on the right side. The exoccipitals are preserved on both sides The basioccipital is complete in IVPP of IVPP V4075, the right one being more V4075 and forms the condylus occipitalis complete. Both are broken laterally and it posteroventrally as in chelydrids. Anteriorly cannot be determined whether or not there there is a long V-shaped contact with the ba- was a foramen jugulare posterius entirely sisphenoid that differs from the usual trans- formed by bone as in Chelydra or whether it verse contact seen in Sinemys sp., TMP was open as in the Tyrrell skull. Two foram- 87.2.1, and most turtles. ina nervi hypoglossi are present close to the The prootic is present and apparently com- 12 AMERICAN MUSEUM NOVITATES NO. 3048 plete on both sides of IVPP V4075, but ma- character analyses that could resolve this trix makes it difficult to determine its limits problem, although this work is underway. and no internal structures are visible. Dor- The close similarity of Dracochelys and sally it seems to enter the medial margin of Hangaiemys, i.e., narrow triturating surfaces the foramen stapedio-temporale and lies me- and large foramina palatinum posterius, sug- dial to the quadrate. Posteriorly it contacts gests that they are closely related. The best the opisthotic and medially the supraoccip- estimate of the relationships of Dracochelys ital. Portions ofthe opisthotic are present on at present is that it is a eucryptodire below both sides of IVPP V4075 but neither are the level of Polycryptodira (Gaffney and complete. The medial parts of the bone are Meylan, 1988) and probably close to Han- present and contact the supraoccipital, proot- gaiemys. ic, and exoccipital, but the lateral area, the processus paroccipitalis, is largely missing. The basisphenoid of IVPP V4075 is present and complete except for the damaged sec- REFERENCES tion around the canalis caroticus internus and Ckhikvadze, V. M. basisphenoid-pterygoid contact. The basi- 1987 (1988). Sur la classification et les carac- sphenoid is arrow-shaped, pointed anterior- teres de certaines tortues fossiles d'Asie, ly, and broadening posteriorly to end pos- rares et peu etudiees. Stud. Palaeoche- terolaterally in a blunt process. The loniol. 2(3): 55-86. basisphenoid lies between the pterygoids for Dong, Zhi-ming most ofits length and has a posterior contact 1973. (Cretaceous stratigraphy of Wu'erhe with the basioccipital for most of its width. (Urho) region). In Reports of the Pale- Anteriorly, the basisphenoid lies in a plane ontological Expedition to the Xinjing; slightly dorsal to the pterygoids, allowing the (2), the pterosaurian fauna from Wu'erhe (Urho) (China, Academia Sinica, Insti- formation of the openings discussed above. tute ofVertebrate Paleontology and Pa- In the middle of the basisphenoid are paired leoanthropology), Acad. Sin., Inst. Ver- foramina for the medial branch ofthe carotid tebr. Palaeontol. Palaeoanthropol., arteries. These foramina are entirely formed Mem., 11: 1-7. [in Chinese] by the basisphenoid and are not close to the Ernst, C. H., and R. W. Barbour pterygoid suture. Posterolateral grooves ex- 1972. Turtles ofthe United States. Lexington: tend from the foramina and are continuous University Press of Kentucky, 347 pp. with the canalis caroticus internus of each Gaffney, E. S. pterygoid. 1983. Cranial morphology of the extinct RELATIONSHIPS: Dracochelys has the otic horned turtle, Meiolania platyceps, from trochlear of and the Pleistocene of Lord Howe Island. synapomorphy cryptodires Bull. Am. Mus. Nat. Hist. 175(4): 361- the posterior pterygoid process ofselmacryp- 480. todires (Gaffhey and Meylan, 1988). The eu- Gaffhey, E. S., and P. A. Meylan cryptodires are characterized by an internal 1988. A phylogeny ofturtles. In M. J. Benton carotid contained in the pterygoid, but in (ed.), The phylogeny and classification some eucryptodires the bone flooring the ca- of the tetrapods, pp. 157-219. Oxford: nal may be very thin (plesiochelyids) or only Clarendon Press. partial. Among the taxa with partial floors to Gaffney, E. S., P. A. Meylan, and A. R. Wyss the canalis caroticus internus are Sinemys, 1991. A computer assisted analysis of the re- Hangaiemys, and Dracochelys. Sinemys has lationships of the higher categories of more extensive flooring than Dracochelys, turtles. Cladistics 7: 313-335. which has the least degree of carotid enclo- Hutchison, J. H., and J. D. Archibald sure in any eucryptodire. Whether or not this 1986. Diversity of turtles across the Creta- ceous/Tertiary boundary in northeast- represents a character phylogeny, or reflects ern Montana. Palaeogeogr., Palaeocli- variation within taxa with partial carotid matol., Palaeoecol. 55: 1-22. floors, is hard to test at present. Sinemys, McKenna, M. C., J. H. Hutchison, and J. H. Hart- Hangaiemys, and other relevant taxa are not man yet described in sufficient detail to develop 1987. Paleocene vertebrates and nonmarine 1992 GAFFNEY AND YE: DRACOCHELYS 13

Mollusca from the Goler Formation, biostratigraphy of Mongolia. The Joint California. In Brett F. Cox (ed.), Basin Soviet-Mongolian Paleontological Ex- analysis and paleontology of the Paleo- pedition. Trans. 1: 192-200. [in Rus- cene and Eocene Goler Formation, El sian] Paso Mountains, California. Los An- Tatarinov, L. P. geles: Society of Economic Paleontolo- 1959. (A new turtle of the family Baenidae gists and Mineralogists, Pacific Section. from the Lower Eocene of Mongolia). Russell, L. S. Paleontol. Zh. 1: 100-1 13. [in Russian] 1934. Fossil turtles from Saskatchewan and Ye, Xiangkui Alberta. Trans. R. Soc. Canada 28: 101- 1963. (Fossil Turtles of China). Palaeontol. 110. Sinica n. ser. C, 18: 1-112. [in Chinese, Sukhanov, V. B. text also in English] 1964. (Testudinata). In I. A. Orlov (ed.), Am- 1973. (Chelonia fossils from Wuerho). In Re- phibians, reptiles, and birds, Osnovy ports of Paleontological Expedition to Paleontol. 12: 354-438. [in Russian] Sinkiang (II). Pterosaurian fauna from Sukhanov, V. B., and P. Narmandakh Wuerho, Sinkiang. Mem. Inst. Vertebr. 1974. (New Early Cretaceous turtle from con- Paleontol. Paleoanthropol., Acad. Sin- tinental deposits of the northern Gobi). ica 1 1: 8-1 1. [in Chinese] Mesozoic and Cenozoic Faunas and

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.

THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. |