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

Home , Anosteira, Palatobaena, Taphrosphys

AMERICAN MUSEUM Norntates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2737, pp. 1-22, figs. 1-1 3 June 29, 1982 Cranial Morphology of the Baenid Turtles

EUGENE S. GAFFNEY'

ABSTRACT The family Baenidae is a group of extinct cryp- living cryptodires, the Eucryptodira. The arterial todiran turtles known from more than a dozen canals of the skull are preserved in enough spec- well-preserved skulls from the Cretaceous, Paleo- imens to conclude that baenids had an arterial cene, and Eocene. The baenid skull is character- pattern characterized by a large and well-devel- ized by significant generalized features in the skull oped canalis stapedio-temporalis and a small but roofand basicranium that are consistent with their persistent canalis caroticus lateralis. hypothesized relationship as the sister group to the

INTRODUCTION The Baenidae is an important group of included diagnoses, synonymies, lists ofspec- Cretaceous to Eocene non-marine turtles that imens, skull and shell figures, and a review have been considered as the sister group to of the literature up to 1972. Descriptions all living cryptodires (Gaffney, 1975). While were lacking and the present work is an effort information on the systematics of this group to fill this lack for skulls. The relationships has been available (Gaffney, 1972), a more of taxa within the Baenidae were also dis- complete study ofthe cranial morphology has cussed in Gaffney (1972), but in 1975 I dealt been lacking. The present paper is intended with the relationships ofbaenids to other tur- to fill this void, particularly in the description tles and the question of the "Amphichely- of basicranial regions. A number of impor- dia," a supposedly ancestral turtle group that tant figures of baenid crania have already was characterized by baenids. Evans and appeared in Gaffney (1972, 1979b) and Ar- Kemp (1975, 1976) described two skulls, chibald and Hutchison (1979). To eliminate Mesochelys and Dorsetochelys, and discussed repetition, I assume that the reader has these relationships of these taxa to baenids and three papers available and is familiar with other turtles. Gaffney (1979a) described the contents of Gaffney (1979b). Glyptops and presented an analysis of relationships involving this genus, Mesochelys, Dorsetochelys, and baenids. The most im- PREVIOUS WORK portant work on taxa within the Baenidae Gaffney (1972) provided the primary basis since 1972 is Archibald and Hutchison's pa- for current systematic work on baenids and per (1979) on Palatobaena which described ' Curator, Department of Vertebrate Paleontology, American Museum of Natural History.

new, well-preserved material of this hereto- MORPHOLOGY fore poorly known form and proposed a new NASAL: Nasals are characteristically re- scheme of its phylogenetic relationships. An tained in baenids and lost in most eucryp- important but as yet unpublished work is todires. In baenids the nasals retain the prim- Archibald (MS), which describes the first itive (for amniotes) posterior contact with the skull-shell associations of Eubaena and Sty- frontals, whereas in those eucryptodires with giochelys and proposes a baenid phylogeny nasals the prefrontals meet medially and pos- based on shell features. Gaffney's cranial terior to the nasals. morphology review (1979b) has whole skull In most baenids (Gaffney, 1979b, fig. 1) drawings of all the baenid genera (but does the nasal contacts the maxilla laterally, the not repeat the important palatal stereopho- frontal posteriorly, and the other nasal me- tographs from Gaffney, 1972), as well as a dially. The anterior edge is free and forms the series of figures showing more detailed, in- dorsal margin ofthe apertura narium externa. ternal features of baenid skulls. Lower jaws Within the baenids, Hayemys is unique in of baenids are dealt with to some extent in having exceptionally large nasals that are lat- Gaffney (1979b) and Archibald and Hutch- erally and posteriorly expanded in compari- ison (1979) but a detailed study is in prepa- son to other baenid nasals. In Baena the na- ration. Also under way is a new phylogenetic sals are either lost or fused to the frontals. analysis and classification of the family. Sutures differentiating nasals are not visible in any ofthe numerous specimens (including ACKNOWLEDGMENTS presumedjuveniles) ofBaena, suggesting loss I am grateful to the curators and collection rather than fusion. The anterior margin of managers of the institutions from which I the apertura narium externa in Baena is have borrowed baenid specimens over the thickened with an anterior facing concavity, years (a complete institutional list is in Gaff- similar to one seen in Palatobaena bairdi. ney, 1972). Dr. David Archibald, Yale Uni- Chisternon is unique among baenids in lack- versity, very kindly allowed me to use infor- ing a nasal-maxilla contact, this being caused mation from his unpublished Ph.D. thesis by an anterior extension of the frontal. and with the aid of Dr. Howard Hutchisop, In Palatobaena the preorbital part of the University of California, Berkeley, made skull is highly modified from the general specimens available to me that I otherwise baenid pattern by extreme shortening antero- would not have been aware of. posteriorly and widening laterally. Nasals are Ms. Lorraine Meeker drew figures 6 and known only in one specimen of Palatobaena 13, Mr. Chester Tarka took the photographs bairdi (CCM 77-11, figured in Gaffney, in figures 3 and 4, Mrs. Jennifer Emry drew 1979b; Archibald and Hutchison, 1979) figures 1 and 2, the American Museum where they are seen to be widely separated Graphic Arts Department did figure 8. 1 very from each other by the extension ofthe fron- much appreciate the assistance of these peo- tals into the apertura narium externa. Al- ple. I took the remaining photographs. though the limits ofthe nasals are not entirely clear in this specimen they do show a dor- ABBREVIATIONS solateral and medial contact with the frontal and ventral contact with the maxilla. The rim AMNH, American Museum of Natural History of the entire apertura narium externa is re- CCM, Carter County Museum cessed in Palatobaena bairdi, whereas in P. MCZ, Museum ofComparative Zoology, Harvard University gaffneyi (Archibald and Hutchison, 1979) PU, Princeton University only the ventral margin is recessed. There are UCMP, University ofCalifornia at Berkeley, Mu- no signs of nasals in this last specimen. seum of Paleontology The ventral morphology of the nasal area USNM, National Museum of Natural History (Gaffney, 1979b, fig. 3) has paired troughs UW, University of Wyoming concave ventrally and separated by a median YPM, Yale Peabody Museum ridge along which the two nasals meet. The 1982 GAFFNEY: BAENID TURTLES 3

FIG. 1. Disarticulated baenid bones from the Late Cretaceous, Hell Creek Formation, Bug Creek Anthills, McCone County, Montana. A. Ventral view of right frontal (MCZ 355 1), a presumed juvenile, 9.6 mm. in length. B. Ventral view of right frontal (MCZ 3554), 22.7 mm. in length. C. Dorsal view of supraoccipital (MCZ 3558), 12.8 mm. in length. D. Ventral view of squamosal (AMNH 8558), 16.6 mm. in length. E. Lateral view of left parietal (MCZ 3526), presumed juvenile, 21.4 mm. in length, anterior is to left, dashed line indicates foramen nervi trigemini. troughs housed the olfactory bulbs and are primitive chelonian condition, whereas eu- an anterior extension of the sulcus olfacto- cryptodires typically have conspicuously rius. large dorsal prefrontal plates that meet in the PREFRONTAL: The prefrontal in turtles usu- midline. Hayemys, however, does have a ally consists oftwo plates, a horizontal dorsal moderately large dorsal plate that probably one and a vertical ventral one. As empha- approximates the primitive condition, but at sized by Evans and Kemp (1976) and dis- present I interpret this as an autapomorphy cussed by Gaffney (1979a) baenids are gen- for Hayemys. erally characterized (see Gaffney, 1979b, fig. The ventral process of the prefrontal is a 1) by a reduction of the dorsal plate in com- broadly curved plate oriented in a roughly parison with what is hypothesized as the vertical plane, extending anterolaterally, and 4 AMERICAN MUSEUM NOVITATES NO. 2737

is a mere vertical slit in Glyptops, whereas in the Baenidae it is wider but not so wide as in the Testudinidae. In my earlier paper on baenids I figured Trinitichelys as lacking a dorsal exposure of the prefrontal (Gaffney, 1972, fig. 2). Since then I have been convinced by other students ofthis specimen that a small prefrontal lappet is present and have modified the new recon- struction accordingly (Gaffney, 1979b, fig. 164). Hay (1904, 1905, 1908) reported the presence of a lacrimal in Eubaena, Baena, and Chisternon. Hay's "lacrimal" is here identified as the prefrontal in view ofthe following features held in common between the pre- A - B frontal of baenids and living cryptodires: (1) articulates with the vomer ventromedially (2) borders the foramen orbito-nasale ventrolaterally (3) meets the palatine between the two above structures (4) has a long suture with the maxilla along the anterior border of the fossa orbitalis (5) forms the anterior wall of the fossa orbitalis FIG. 2. A. Dorsal view of left maxilla, Pala- The only significant distinction between tobaena bairdi (AMNH 8550). B. Dorsal view of them seems to be the small or absent dorsal left maxilla, Eubaena cephalica (MCZ 3519). C. plate in baenids. Hay's identification of this Medial view of right maxilla (MCZ 3519, re- element as the lacrimal required further iden- versed), Eubaena cephalica. All are Late Creta- tifications to be consistent with that deter- ceous, Hell Creek Formation, Bug Creek Anthills, mination. A reevaluation suggests the follow- McCone County, Montana. ing changes in Hay's identification:

nasal ...... nasal forming the posterolateral wall of the fossa prefrontal ...... frontal nasalis and the anterior wall of the fossa or- lacrimal ...... prefrontal bitalis. The lateral edge of the bone is a long frontal plus fused parietal ...... parietal curved contact with the maxilla which continues to the floor ofthe fossa orbitalis where Hay's identifications were repeated in Ro- the prefrontal contacts the palatine, just lat- mer's figure of Chisternon (1956, fig. 50). eral to the foramen orbito-nasale. The pala- FRONTAL: The frontal ofbaenids (Gaffney, tine contact continues medial to the foramen 1979b, figs. 152-164) is essentially a flat bone orbito-nasale for a short distance until it con- situated between the midline and the orbit. tacts the vomer. The latter forms the ventro- Its contacts are usually as follows: anteriorly medial limit of the prefrontal. Dorsal to the with the nasal, anterolaterally with the max- vomer the prefrontal has a curved edge which illa, anteroventrolaterally with the prefrontal, defines the fissura ethmoidalis. The fissura posterolaterally with the postorbital, poste- 1982 GAFFNEY: BAENID TURTLES 5

riorly with the parietal, and medially with the the attachment of the M. pterygoideus mus- other frontal. The dorsal surface is broadly cle. This pit is usually behind the orbit and convex. The ventral surface (fig. 1; Gaffney, is demarcated by a ridge anteriorly and a 1979b, fig. 3; Archibald and Hutchison, ridge posteriorly (see Gaffney, 1979b, fig. 3; 1979, fig. 6) usually has a parasagittal ridge Archibald and Hutchison, 1979, fig. 6). The defining a medial sulcus olfactorius. A low muscle mass itself parallels and is attached transverse ridge generally separates the fossa to the lateral surface ofthe processus inferior nasalis from more posterior regions of the parietalis. cavum cranii. The vertical wall is situated parasagittally The frontal broadly enters the orbital mar- with-the cavum cranii medial to it and the gin in most baenids, the exceptions being fossa temporalis lateral to it. The anterior Palatobaena, Hayemys, and Eubaena. In part, the processus inferior parietalis (Gaff- Hayemys the large dorsal plate of the pre- ney, 1979b, fig. 13) is usually longer and ar- frontal forms the principal border for the or- ticulates with the crista pterygoides of the bit, and the prefrontal reaches posteriorly to pterygoid. The foramen nervi trigemini per- contact the postorbital. The type specimen forates the posteroventral border of the pro- of Eubaena cephalica (YPM 1785) has no cessus. The parietal articulates with the frontal exposure in the orbit but this is due prootic laterally behind the foramen nervi to closer approximation ofthe prefrontal and trigemini. The posterior margins of both the anterodorsal margin of the postorbital plates of the parietal contact the supraocci- rather than a large prefrontal. Another spec- pital ventromedially. imen (AMNH 4948) of Eubaena has a lim- Within the Baenidae, there is some vari- ited frontal exposure in the orbit. In Pala- ation in size of the skull roof portion of the tobaena the frontal has reduced orbital parietal. Hayemys has a parietal that is rel- exposure due to approximation of maxilla atively smaller than in other baenids and this and postorbital. is interpreted as an autapomorphy due to the The posterolateral margin of the frontal absence of this feature in comparable out- tends to be extended posteriorly to form groups. Hayemys has the posterior temporal "wings" that are particularly apparent in Eu- emargination, however, developed to the ex- baena and Chisternon. Hayemys is unique tent seen in Eubaena, Stygiochelys, Plesio- among baenids in having a frontal that is baena, and Palatobaena. Baena and Chister- proportionally larger than the parietal. The non have relatively large parietals that are well-developed frontal-maxilla contact seen more extensive posterolaterally than in other in most baenids and absent in Glyptops, Me- baenid genera. This feature is hypothesized sochelys, and Trinitichelys is hypothesized as a synapomorphy for Baena and Chister- here as a derived character for the group con- non. taining Plesiobaena, Palatobaena, Eubaena, The posterior temporal emargination shows Stygiochelys, Baena, and Chisternon. an interesting taxonomic distribution within PARIETAL: The parietal (fig. 1) consists es- the baenids (Gaffney, 1979b, figs. 152-164). sentially of two plates of bone, a horizontal The primitive condition for baenids and one and a vertical one, meeting at right an- cryptodires appears to be that seen in Dor- gles; both are well developed in baenids. The setochelys and Mesochelys in which the emar- horizontal plate is dorsal and roofs the fossa gination is relatively shallow (but present in temporalis and cavum cranii in the posterior contrast to Proganochelys) and bordered only region of the skull. It usually contacts the by the parietal and squamosal. The latter frontal anteriorly, postorbital laterally, and bones have a well-developed contact. Trini- sometimes the squamosal posterolaterally. tichelys is advanced in comparison by having The posterior limits of the horizontal plate a more extensive emargination and a minute are usually involved in emargination ofsome to absent parietal-squamosal contact. The sort. On the ventral surface ofthe horizontal Late Cretaceous baenids have the most ex- plate lateral to the cavum cranii is a pit for tensive emargination characterized by no pa- 6 AMERICAN MUSEUM NOVITATES NO. 2737

FIG. 3. Occipital view of Trinitichelys hiatti (MCZ 4070, 55 mm. in length), Early Cretaceous, Trinity Sand, Hardee Montague County, Texas. See Gaffney (1979b, fig. 157) for a labeled diagram ofthe occiput in Eubaena. rietal-squamosal contact and a broad expo- 152-164) consists of a flat, vertical cheek sure ofthe postorbital. Baena and Chisternon, plate and a medially projecting ramus form- however, do have a parietal-squamosal con- ing the posteroventral wall of the fossa or- tact which is hypothesized here as a synap- bitalis (fig. 6). The cheek portion of the jugal omorphy for these two taxa. The parietal in is roughly rectangular in outline and is Baena and Chisternon is larger and more ex- bounded dorsally by the postorbital, poste- tensive laterally than in the Cretaceous riorly by the quadratojugal, ventrally by the forms. In Baena the emargination is most cheek emargination, and anteriorly by the limited and a well-developed parietal-squa- maxilla. Thejugal is exposed along the orbital mosal contact is present. Chisternon has a opening between the postorbital and maxilla limited parietal-squamosal contact but a in all baenids except Eubaena. In Eubaena more extensive emargination. Baena, how- the jugal is excluded from the orbital rim by ever, has a distinctive degree of variation in a wide meeting ofthe maxilla and postorbital. its temporal emargination. This has been de- The medial ramus of the jugal is usually ex- scribed in Gaffney (1972, figs. 26, 28, 29) and posed along the floor of the orbit and in the more thoroughly discussed in Archibald anterior wall ofthe fossa temporalis interior. (MS). Ventrally and anteromedially the jugal con- The emargination in Palatobaena is gen- tacts the maxilla, dorsally it contacts the post- erally comparable with that in the other Late orbital, and medially there is a short ptery- Cretaceous and Paleocene baenids, but it is goid contact. This pterygoid contact is at the distinctly less developed than in such forms anterior limits of the processus pterygoideus as Eubaena. The parietal in Palatobaena externus. gaffneyi has the dorsal plate developed lat- QUADRATOJUGAL: The quadratoj ugal erally much as in Plesiobaena putorius but in (Gaffney, 1979b, figs. 152-164) is a flat, tri- Palatobaena bairdi (CCM 77-1 1) the dorsal angular bone forming part of the lateral wall plate is more extensive and its posterior por- of the fossa temporalis. It is sutured ante- tion is aligned parasagittally, producing a riorly to the jugal, dorsally to the postorbital, distinctive parallel sided emargination. and posterodorsally to the squamosal. The JUGAL: The jugal (Gaffney, 1979b, figs. posterior margin of the quadratojugal con- 1982 GAFFNEY: BAENID TURTLES 7

FIG. 4. Occipital view of Stygiochelys estesi (AMNH 2601, 63 mm. in length), Late Cretaceous, Hell Creek Formation, Glendive, Montana. See Gaffney (1979b, fig. 157) for a labeled diagram ofthe occiput in Eubaena. tacts the quadrate along the anterior edge of bital medially in all baenids except Chister- the cavum tympani forming a broadly curved non. In Chisternon the anterior margin is suture. There is a posteroventral process cov- formed completely by the postorbital with no ering some of the lateral surface of the pro- certain quadratojugal contact. Although most cessus articularis ofthe quadrate. The ventral of the medial edge in Chisternon borders the margin ofthe bone is part ofa limited ventral temporal emargination, the parietal makes emargination of the cheek, which reaches its a short contact anteromedially, whereas greatest extent in Plesiobaena. Baena has a long medial suture with the pa- There is not much variation of the qua- rietal. The antrum postoticum is well devel- dratojugal among the baenid genera, but oped in all baenids. Posteromedially the op- Trinitichelys has a somewhat longer qua- isthotic covers the margin of the squamosal. dratojugal and Baena and Chisternon lack The squamosal sends a process ventrally that the dorsal projection seen in other baenid covers the external surface of the quadrate genera. This latter feature is interpreted as a posterior to the incisura columellae auris. synapomorphy uniting Baena and Chister- POSTORBITAL: As in other Casichelydia, the non. postorbital is a flat, rectangular bone forming SQUAMOSAL: The squamosal of baenids is part of the dorsolateral wall ofthe fossa tem- quite similar to that of other casichelydians, poralis. It forms the posterior border of the it is a roughly cone-shaped bone forming a eye, contacts the frontal anteromedially, and cap above and somewhat behind the quad- borders the parietal medially. The postorbital rate (Gaffney, 1979b, figs. 152-164). The in- contacts the maxilla anteroventrally in Eu- terior of the squamosal (fig. 1) forms part of baena but is prevented from doing so in other a cavity, the antrum postoticum, which is baenids by the jugal, which reaches the or- continuous with the cavum tympani. The bital margin. The posterolateral contact is a squamosal is exposed along the lateral mar- long suture with the quadratojugal and is al- gin of the posterior temporal emargination. ways longer than the contact with the jugal. The posteroventral external surface usually Posteriorly the postorbital reaches the squa- has a trough for the attachment of the M. mosal, and in Chisternon the postorbital may depressor mandibulae. also contact the quadrate at the dorsal edge Anteriorly, the squamosal is bounded by of the cavum tympani. the quadratojugal laterally and the postor- In Trinitichelys, Baena, and Chisternon the 8 AMERICAN MUSEUM NOVITATES NO. 2737 postorbital does not reach the temporal sal anteriorly and the frontal posteriorly on emargination, but in the remaining baenids the outer surface of the skull. The process in the posteromedial edge ofthe bone forms the Trinitichelys seems to contact only the nasal, anterolateral border of the temporal emar- but in all other baenids this dorsal process gination. contacts the frontal. In Chisternon an ante- PREMAXILLA: The premaxilla of baenids rior extension ofthe frontal prevents a nasal- (Gaffney, 1972, figs. 47-58; 1979b, figs. maxilla contact, whereas the apparent nasal- 152-164) does not have any systematic dif- frontal fusion (or loss of the nasal) is ferences from other Casichelydia. The bone characteristic of Baena. is roughly triangular in ventral view and Posterolateral to the dorsal process the forms the ventral margin ofthe apertura nar- lower rim ofthe orbit forms the dorsal margin ium externa. The premaxilla is bordered by ofthe vertical plate ofthe maxilla. In Baena, the other premaxilla medially and by the Chisternon, Palatobaena, Stygiochelys, Ple- maxilla posterolaterally. The premaxilla siobaena, Trinitichelys, and probably Hay- forms the anteriormost portion of the labial emys (but this area is obscured by damage), ridge along its ventral edge and is continuous the vertical plate of the maxilla ends poste- with the labial ridge of the maxilla. The riorly in a suture with the jugal which begins anterodorsal edge ofthe premaxilla forms the in the posteroventral corner of the orbit. In ventral margin of the apertura narium ex- Eubaena the maxilla sends up a dorsally di- terna. The dorsal surface of the premaxilla rected process behind the eye which articu- is the main element in the floor of the fossa lates with the jugal posteriorly and also with nasalis. Palatobaena (Archibald and Hutch- the postorbital dorsally. The postorbital con- ison, 1979) is characterized by a sulcus or tact is lacking in the other taxa. recession of the premaxillae and maxillae On the internal side of the maxilla (figs. 2, along the margin of the apertura narium ex- 6), in the posterior part of the wall of the terna. Posteriorly the vomer contacts the pre- fossa nasalis, is the foramen alveolare supe- maxilla near the midline. The lingual ridge rius, which leads into the canalis alveolaris usually becomes reduced before reaching the superior. In Eubaena the foramen is large premaxilla but some remnant of it usually and lies at the head of a groove leading pos- continues nearly to the midline. The lingual teriorly into the foramen orbito-nasale. In ridges ofthe right and left sides are separated Chisternon, Stygiochelys, and Palatobaena in the midline by a trough which is formed the foramen alveolare superius lies near the mostly by the premaxillae and the vomer. edge ofthe foramen orbito-nasale rather than This trough in Eubaena is narrow in the re- farther away from the edge of the foramen gion of the vomer and widens anteriorly in orbito-nasale as in Eubaena. Plesiobaena has the region of the premaxillae. In the other the foramen alveolare superius in about the baenids the trough is wide posteriorly and same position as it is in Chisternon but no becomes indistinct anteriorly. The foramen groove seems to be present. None of the nu- praepalatinum usually lies in, or near, the merous specimens of Baena have this area suture between the premaxilla and the vomer. sufficiently well preserved to determine the MAXILLA: As in other turtles, the maxilla presence or absence ofthe foramen alveolare of baenids (figs. 2, 6; Gaffney, 1972, figs. superius. Trinitichelys and Hayemys also 47-58; 1979b, figs. 152-164) consists of two yield no information about this area. regions, one vertical and one horizontal, The horizontal plate of the maxilla makes joined at right angles. The vertical portion up the main part of the triturating surface, makes up much of the facial region and is forms some ofthe floor ofthe fossa orbitalis, limited anteroventrally by the premaxilla. A and contacts buttresses of the palate. Trini- vertical process of the maxilla slopes pos- tichelys and Plesiobaena have quite similar terodorsally from the premaxilla-maxilla su- triturating surfaces. In these forms the labial ture to form the lateral edge of the apertura ridge is straight and a deep channel is present narium externa, and usually contacts the na- just medial to the ridge. The lingual ridge 1982 GAFFNEY: BAENID TURTLES 9 borders this channel anteriorly but quickly tura narium interna, with the anterior and turns medially to form a posteromedial ex- posterior portions expanded to articulate pansion of the triturating surface. The area with the other palatal bones. The anterior where this expansion begins to take place is expansion lies at a lower level than the pos- raised higher than the rest of the surface. terior one and forms a roof for the ventrally Trinitichelys seems to differ from Plesio- concave trough developed between the trit- baena in having a narrower triturating sur- urating surfaces. The area exposed dorsally face posteriorly and in not having the labial in the fossa nasalis (fig. 6) is convex upward ridge as high but the specimen is not well and is wider in area than the ventral surface preserved and further precision is not pos- of this expansion. The anteriormost contact sible. of the vomer is with the premaxillae. There Eubaena has a labial ridge which is not is a lateral contact of the anterior expansion straight as in most of the other forms, but of the vomer with the maxilla that ends at has a bend anteriorly which results in the the anterior edge of the apertura narium in- anterior portion of the ridge being oriented terna. in a more anteroposterior direction than the Between the apertura narium internae the rest of the ridge. The lingual ridge is so low vomer is rodlike and narrow, except in Pal- as to be nothing more than the medial edge atobaena where it is broad. Above and be- ofthe triturating surface. This surface begins hind the apertura the posterior portion is ex- quite close to the labial ridge just behind the panded into a flat plate of bone. This plate premaxilla-maxilla suture and curves medi- is above the level of the anterior expansion ally and then laterally to form a gently con- and carries a groove in its midline, the sulcus cave broad plate. Stygiochelys has the same vomeri, which forms the ventral limits ofthe form ofthe triturating surface as in Eubaena fissura ethmoidalis. At its posteriormost lim- but the former differs in having the two max- its the posterior plate contacts the pterygoid illae set farther apart and at greater angles to in a short, transverse suture. For most of its one another thereby increasing the width of length the posterior plate has a lateral contact the snout region. Stygiochelys lacks the in- with the palatine. Anterodorsolaterally, on cipient secondary palate developed in Eu- either side ofthe sulcus vomeri, the posterior baena. plate contacts the prefrontal. Palatobaena (Archibald and Hutchison, Eubaena differs from the other baenids in 1979) has the broadest and most distinctive having a narrower anterior vomerine expan- triturating surfaces in the Baenidae. The la- sion. Baena and Chisternon differ from Ple- bial ridge is lower than in the other baenids siobaena and /Eubaena in having a vomer and the lingual ridge is virtually absent. The that extends posteriorly beyond the palatines, bone is thicker than in other forms but the whereas the palatines extend farther poste- overall morphologic relationships are the riorly in the last two genera. In Stygiochelys same as in Stygiochelys. and possibly Palatobaena gaffneyi (Archi- The maxilla in Hayemys has a very narrow bald and Hutchison, 1979, figs. 5, 7) there is dorsal process that is constricted between the a minute vomer-basisphenoid contact. In expanded nasal and prefrontal. The limits of Hayemys the anterior half of the vomer is the vertical plate are difficult to determine badly damaged but the posterior half has an but the maxilla probably did not reach the expansion that is wider than in other baenid postorbital in life and its present contact with genera. the postorbital is probably due to crushing. PALATINE: The palatine of baenids (Gaff- The triturating surface is obscured by the la- ney, 1972, figs. 47-58; 1979b, figs. 152-164) bial edge being folded over it during preser- does not differ significantly from other cryp- vation (Gaffney, 1972, fig. 7). todires. In baenids it is a platelike element VOMER: The vomer of baenids (Gaffney, forming the portion of the palate anterior to 1972, figs. 47-58; 1979b, figs. 152-164) is a the pterygoid between the vomer and the single, median element separating each aper- maxilla and having a straight anteroposterior 10 AMERICAN MUSEUM NOVITATES NO. 2737 suture with the vomer medially. Anterome- The canalis stapedio-temporale trends dor- dially it contacts the descending process of sally where it reaches the fossa temporalis. the prefrontal, and lateral to this contact the The lateral surface of the quadrate is the palatine forms the posterior half of the fo- cavum tympani, a nearly circular cavity ramen orbito-nasale. The palatine has a long opening laterally. The incisura columellae lateral suture with the maxilla. Ventrally this auris is a relatively narrow slit enclosing both suture borders the triturating surface and ex- the stapes and the eustachian tube in baenids. tends posteriorly to the transverse suture with Above the incisura, the antrum postoticum the pterygoid. In so doing, the foramen pal- extends posterodorsally into the squamosal. atinum posterius is formed at least in part by The suture with the squamosal is roughly the palatine. The palatine forms most of the oval in dorsal view. Ventrolaterally the quad- dorsolateral wall of the internal nares ven- rate sends down a short processus articularis trally, and part of the floor of the fossa or- which bears the condylus mandibularis, the bitalis dorsally. lower jaw articulation. There is some difference in palatine mor- The quadrate of baenids is typically cryp- phology among baenids but it is not exten- todiran in its basicranial relations, that is, it sive. Most of the variation concerns width has a broad sutural contact with the posterior ofthe area between lingual ridges, that is, the part of the pterygoid rather than having a choanal channel. The form of the choanal medial process that reaches the basisphenoid. channels, paired grooves just posterior to the The incisura columellae auris is elongate, en- aperturae narium internae formed by the closing the stapes and eustachian tube, a fea- vomer and palatines, ranges from narrow ture common to nearly all pleurodires (but (e.g., Eubaena) to broad (e.g., Chisternon). not Bothremys and Taphrosphys, see Gaffney In Chisternon and Stygiochelys the foramen and Zangerl, 1968; Gaffney, 1979b). Al- palatinum posterius is formed entirely by the though many eucryptodires (chelonioids, for palatines, and in the remaining baenid genera example) do not restrict the incisura, when the pterygoid forms some of the margin of they do, the restriction usually separates the foramen. stapes and eustachian tube. Meiolaniids are The anteroventral limits of the processus the only eucryptodires that have an incisura inferior parietalis come in contact with the columellae auris enclosing stapes and eusta- posterior edge of the palatine in Eubaena, chian tube in the baenid manner. At present, Palatobaena, and Chisternon. The condition the baenid condition can be interpreted either is indeterminate in the other forms. as a derived character for baenids or as a QUADRATE: The quadrate is a large C- primitive feature ofCasichelydia retained in- shaped element lying at the posterolateral dependently in baenids and meiolaniids. I corner of the skull (Gaffney, 1979b, figs. prefer the former interpretation. 152-164). Its dorsal surface floors the lateral PTERYGOID: A disarticulated right ptery- part of the fossa temporalis. Anterolaterally goid (fig. 5; Gaffney, 1979b, fig. 24; also in the quadrate has a broad vertical contact with Gaffney, 1975, fig. 21) from the Hell Creek the prootic (Gaffney, 1979b, fig. 13). At the Formation of Montana has been particularly anterior end of this contact they form the useful and forms the basis for much of the processus trochlearis oticum, the suture run- following description. This bone (and MCZ ning down the middle of the process. Ven- 3566, a small pterygoid fragment that pre- trally there is a vertical suture with the pter- serves the crista pterygoidea particularly ygoid. The quadrate forms much of the well) is impossible to identify to genus, but lateral floor and roof of the cavum acustico- Palatobaena can be ruled out on the basis of jugulare (Gaffney, 1979b, fig. 94). In the an- the pterygoideus muscle scar (see below) and terior wall of the cavum, the canalis caver- the closest similarity is with Eubaena and nosus extends anteromedially and the aditus Plesiobaena which occur in the Hell Creek canalis stapedio-temporalis connects to the Formation. well-developed canalis stapedio-temporale. The crista pterygoidea rises sharply on the 1982 GAFFNEY: BAENID TURTLES 11

canalis caroticus lateralis

foramen posterius

FIG. 5. Medial view ofa baenid right pterygoid (MCZ 3563), Late Cretaceous, Hell Creek Formation, Bug Creek Anthills, McCone County, Montana. Compare with Gaffney (1 979b, fig. 24) for a dorsal view of this specimen. Crista pterygoidea restored from MCZ 3566. anterior part ofthe dorsal surface ofthe pter- rietal and epipterygoid but there is no com- ygoid then drops to form the lower border of pelling reason to accept this. The processus the foramen nervi trigemini. Behind the fo- inferior parietalis extends ventrally and ramen the crista again rises, forming the pos- lodges in a groove on top of the crista pter- terior margin of the foramen nervi trigemini goidea. At the posteroventral end ofthis pro- then gradually tapers down to the level ofthe cess is a space, the fossa cartilaginis epipter- rest ofthe pterygoid at its posterior end. The ygoidea, which held the epipterygoid cartilage. contacts ofthe crista pterygoidea are ofsome Directly behind the foramen nervi trigemini interest systematically but due to the unfor- is the laterally expanded portion ofthe crista tunately common habit of sutural fusion in pterygoidea which bears the prootic (best baenids, information on this area is scarce. seen in Baena arenosa, AMNH 5952, and The available material, however, does fit a AMNH 5977). Posterior to this contact the single pattern. In nearly all other cryptodires lateral margin ofthe pterygoid is a broad sur- (except Dermochelys) the epipterygoid ossi- face sutured to the quadrate. fies as a distinct element and is sutured to the On the medial side ofthe crista pterygoidea crista pterygoidea, but in baenids there is no lies the sulcus cavernosus. At the anterior separate epipterygoid ossification (Gaffney, point of prootic contact the pterygoid forms 1979b, fig. 13). I have (ibid., p. 106) inter- the foramen cavernosum and posterior to preted this condition as due to fusion of pa- this structure the sulcus is called the canalis 12 AMERICAN MUSEUM NOVITATES NO. 2737

FIG. 6. Dorsal view ofhorizontally sectioned skull ofEubaena cephalica (composite, based primarily on AMNH 4948 but with additions from YPM 1785). From Gaffney (1979b). cavernosus because it is roofed over by the occipital ventrally and the exoccipital dor- prootic. The canalis cavernosus ends poste- sally. riorly on the floor ofthe cavum acustico-jug- In medial view MCZ 3563 and AMNH ulare. Just medial to this point lies a circular 8555 furnish important information about concavity which represents the bony floor of the arterial circulation (figs. 8, 9). The fora- the cavum labyrinthicum, the surface of men posterior canalis carotici interni has its which was probably covered with cartilage lateral wall formed by the pterygoid. The ca- as in living turtles. The distinction between nalis caroticus internus curves medially into cavum labyrinthicum and cavum acustico- the basisphenoid very soon after entering the jugulare is not obvious in the disarticulated skull. The canalis caroticus lateralis, how- pterygoid (MCZ 3563) as the elements ofthe ever, forms a groove continuous with the ca- bony inner ear (prootic and opisthotic) are nalis caroticus internus and this groove ex- not fused to the pterygoid as in most other tends anteriorly on the medial edge of the specimens. The posteromedial area of the pterygoid until it exits at the foramen carot- pterygoid is a sutural contact with the basi- icum laterale. This structure is best seen in 1982 GAFFNEY: BAENID TURTLES 13

foramen alveolare superius

foramen orbito-nasale foramen caroticum laterale ,foramen palatinum posterius

rostrum basisphenoidale processus clinoideus

foramen anterius ~-canalis carotici interni dorsum sellae foramen nervi abducentis

foramen stapedio-temporale foramen nervi hypoglossi

FIG. 6. Continued. the following specimens: Eubaena cephalica Although information on the vidian canals (AMNH 4948, fig. 7), Plesiobaena putorius in baenids is limited, some conclusions can (PU 14984), and Baena arenosa (AMNH be drawn from the disarticulated pterygoid 5977, YPM 3941, USNM 211143). In these (MCZ 3563), a disarticulated skull of Chi- specimens the foramen caroticum laterale is sternon (YPM 3930), and a fragmentary skull formed between the pterygoid and palatine of Baena (YPM 3941). In the last specimen bones (as would be expected from an ex- (YPM 3941) the canalis nervi vidiani has amination of MCZ 3563) and is just medial been prepared out revealing that it extends to the anteriormost part of the processus in- posteriorly from the foramen palatinum pos- ferior parietalis-crista pterygoidea suture. At terius through the palatine and into the pter- the posterior end of the foramen posterior ygoid to the vicinity ofthe foramen pro ramo canalis carotici intemi lies the ventral open- nervi vidiani. There may be a communica- ing of the foramen pro ramo nervi vidiani. tion between the foramen and the canalis This foramen extends dorsally to enter the nervi vidiani but the material is not well sulcus cavemosus. enough preserved to confirm this and I have 14 AMERICAN MUSEUM NOVITATES NO. 2737

FIG. 7. Dorsal view of basicranium, Eubaena cephalica (AMNH 4948), Late Cretaceous, Hell Creek Formation, Bug Creek Anthills, McCone County, Montana. See figure 6 for identification of structures. left out a connection in the restoration (fig. 8). There is a connection between the canalis nervi vidiani and the canalis caroticus lateralis in all three specimens as well as in AMNH 8555 but this is a short, lateral canal anterior to the foramen pro ramo nervi vid- FIG. 8. Diagrammatic restoration of basicra- iani. The vidian canal and its relations then nial canals in a baenid. See figures 5 and 6 for are quite similar to those described by Al- further identification ofstructures, see text for dis- brecht (1967) for Chrysemys (see also Gaff- cussion. ney, 1979b, fig. 28). This pattern could be primitive for Casichelydia or Cryptodira but at present our information on the taxonomic unidentified Paleocene baenid, but very low distribution of vidian canal types is too lim- in Chisternon. Baena and Trinitichelys seem ited to be particularly useful. to be intermediate. The crista supraoccipi- SUPRAOCCIPITAL: The supraoccipital is ex- talis is short in all baenids when compared posed on the skull roof ofbaenids at the pos- with most casichelydians but quite compa- teriormost extension of the midline in those rable to chelids and Proganochelys, and is specimens with determinable sutures. A free presumably the primitive condition. The supraoccipital (fig. 1) from an unidentified crista is shortest in Chisternon, where it does Hell Creek baenid shows the flat plate on top not extend beyond the condylus occipitalis. of the crista supraoccipitalis that is part of In other forms it does extend beyond the con- the skull roof. dylus but not beyond the squamosal. The crista supraoccipitalis (Gaffney, 1979b, Anterodorsally the supraoccipital is over- figs. 94, 155, 157) forms a major attachment lapped by the paired parietals, anterolaterally area for fibers ofthe M. adductor mandibulae it meets the prootic, posterolaterally it meets externus, portio profundus of Schumacher the opisthotic, and posteriorly it has a short (1973). The crista is high in Eubaena, Ple- suture with the exoccipital. The surface ofthe siobaena, and especially in PU 16838, an supraoccipital which is exposed in the cavum 1982 GAFFNEY: BAENID TURTLES 15 carnii is trough-shaped and is continuous with the dorsal border of the foramen mag- canal to canalis nervi vidiani num. The hiatus acusticus (Gaffney, 1979b, fig. 71) is ossified in baenids but the area is often broken due to the thinness of the bone or fused with the adjacent element so that the ventral suture of the supraoccipital is not known. The foramen aquaducti vestibuli can be seen just above and behind the prootic contact. The supraoccipital contribution to the cavum labyrinthicum is best seen in Baena (AMNH 5977) and an unidentified Paleocene baenid (PU 16838). The supraoccipital forms the dorsal portion of the cavum labyrinthicum and it is surrounded by a circular suture, the anterior halfofit being a contact with the prootic and the posterior halfbeing a contact with the opisthotic. The supraoccipital forms the posterodorsal portion ofthe canalis semicircularis anterior and the anterodorsal portion of the canalis semicircularis posterior. The recessus labyrinthicus supraoccipitalis takes up most of the area formed by the supraoccipital in the cavum labyrinthicum. The dorsal plate of the supraoccipitalis varies from narrow to wide within Palato- baena (Archibald and Hutchison, 1979) with most baenids having a relatively narrow plate. This variation does not seem to have systematic significance. EXOCCIPITAL: The exoccipital of baenids (Gaffney, 1979b, figs. 71, 94, 155, 157) does not differ significantly from that bone in other casichelydians. It is a paired element lying lateral to the foramen magnum, and can be described in two parts, a dorsolateral process and a ventral process. The dorsolateral portion contacts the supraoccipital dorsally but most of its lateral face is sutured to the opisthotic. The medial surface forms most of the lateral wall of the foramen magnum FIG. 9. Ventral view ofbaenid basicranial frag- and part of the cavum cranii. Inside the ca- ment (AMNH 8555) partially eroded and showing vum cranii the exoccipital forms the poste- canalis caroticus internus and associated struc- rior border of the foramen jugulare anterius tures. Late Cretaceous, Hell Creek Formation, Bug as well as the medial wall ofthe canal between Creek Anthills, McCone County, Montana. 27.5 the foramen jugulare anterius and the fora- mm. in length. menjugulare posterius. The foramenjugulare posterius is incomplete laterally and the canal The exoccipital is sutured to the basisphenoid connecting the two foramina is also incom- by a dorsolateral contact that runs the length plete as it approaches the posterior foramen. of both bones. The exoccipitals do not meet 16 AMERICAN MUSEUM NOVITATES NO. 2737

z, . 7,7. A-1 P4" 0t'*

?7M r

*rV,..LN'N.. V.,

FIG. 11. Dorsal view of baenid basicranial fragment (AMNH 8554), same data as figure 9. 28.8 mm. in length.

FIG. 10. Dorsal view of baenid basicranial in Plesiobaena tends to be smaller fragment (MCZ 3532), same data as figure 9. 21.8 tribution mm. in length. than in other forms and Chisternon has a distinctly larger foramen magnum in comparison with other baenids. at any point thus exposing the basioccipital BASIOCCIPITAL: The basioccipital (Gaffney, dorsally in the foramen magnum and on the 1979b, figs. 71, 94, 155, 157) is a small, me- condylus occipitalis. The two paired foram- dian element at the posteroventral end ofthe ina nervi hypoglossi penetrate the exoccipital basicranium. It forms most of the condylus between the foramen magnum and the path occipitalis. A disarticulated basioccipital ofthe vena cerebralis posterior, the foramen (MCZ 3543) from a juvenile Hell Creek jugulare anterius. baenid shows the form of the element. The The condylus occipitalis is made up ofboth anterior contact is a transverse suture with exoccipitals and the basioccipital. Ventro- the basisphenoid. Two winglike processes medially the exoccipital-basioccipital contact extend posteroventrally and are joined to the extends posteriorly to the well-developed tu- exoccipitals dorsally. Anterolaterally these berculum basioccipitale which includes a processes meet the pterygoids which extend dorsolateral contribution from the exoccipi- posteriorly where both elements form the tal. Laterally, the lower portion of the exoc- paired tubercula basioccipitales. The basioccipital contacts the pterygoid above the basi- cipital is excluded from the fenestra post- occipital contact. otica. The condylus occipitalis has small dor- The variation ofthe exoccipital within the solateral contributions from the exoccipitals Baenidae is not easily observable since most but otherwise is formed from the basioccip- specimens have the sutures fused in this re- ital in all specimens that are determinable in gion, however, there seem to be only two this area. The condylus occipitalis has a pit differences of interest. The tubercular con- at its posterior end for the ligamentum apicis 1982 GAFFNEY: BAENID TURTLES 17

prootic forms the posterior margin of the foramen nervi trigemini. Just lateral to this foramen the prootic roofs the canalis cavernosus. About midway along the canal the foramen pro ramo nervi vidiani emerges from the prootic to enter the canalis cavernosus. The portion of the prootic contributing to the wall ofthe cavum cranii has a prominent fossa acustico-facialis. Two foramina nervi acustici and one foramen nervi facialis penetrate this fossa. Most ofthe prootic is formed around the cavum labyrinthicum, specifi- cally, the recessus labyrinthicus prooticus. Anterodorsally the prootic forms the canalis semicircularis horizontalis. There is very lit- tle variation in the morphology of this element within baenids. OPISTHOTIC: The opisthotic of baenids (Gaffney, 1979b, figs. 71, 94, 155, 157) does not differ significantly from that bone in other cryptodires. It is a triangular element forming the posterior part of the bony inner ear, much of the roof of the cavum acustico- jugulare, and the posterior part of the fossa temporalis. Anteroventrally it meets the prootic, but anterodorsally this contact may FIG. 12. Dorsal view of basicranial fragment be interrupted by a contact of the supraoc- ofPalatobaena bairdi (UCMP 114539), Late Cre- taceous, Hell Creek Formation, V-75 180, Garfield cipital and quadrate. Laterally there is a County, Montana. 35.8 mm. in length. broad contact with the quadrate and medially the opisthotic is sutured to the supraoccipital. Ventromedially the exoccipital dorsal pro- dentis. The posterior face is usually set offby cess reaches the opisthotic. a distinct neck. The central surface of the A prominent processus interfenestralis ex- basioccipital is broadly concave and has two tends ventrally from the opisthotic to divide small foramina in its center. the cavum acustico-jugulare into two regions. PROOTIC: As in most turtles, the prootic of Posterior to the processus interfenestralis is baenids (Gaffney, 1979b, figs. 13, 71) is a the processus scallae tympani, whereas lateral small cuboid element forming the anterior to the processus is the cavum acustico-jug- portion of the bony inner ear, part of the ulare proper. The foramen extemum nervi lateral wall of the cavum cranii, and the an- glossopharyngei pierces the process dorsally. teromedial portion of the fossa temporalis. The anterior edge of the process forms the Anterodorsally the prootic meets the parietal, posterior margin of the fenestra ovalis. An- posterodorsally it meets the supraoccipital. teromedial to the processus interfenestralis The lateral contact with the quadrate is in- is the recessus labyrinthicum opisthoticus. terrupted by the canalis stapedio-temporalis. The canalis semicircularis posterior extends There is a small posterior suture with the into the supraoccipital from this chamber. opisthotic in the roof of the cavum labyrin- BASISPHENOID: As in other Casichelydia, thicum. Ventrally the prootic meets the pter- the basisphenoid ofbaenids (figs. 6-12; Gaff- ygoid and medially it comes in contact with ney, 1979b, figs. 71, 152-164) is triangular the basisphenoid. The anterior edge of the in ventral view with the pterygoids meeting 18 AMERICAN MUSEUM NOVITATES NO. 2737 anteriorly ventral to the rostrum basisphe- The entire canalis caroticus internus can noidale (except in Stygiochelys). The general be seen in AMNH 8555 (fig. 9), a Hell Creek morphology of the baenid basisphenoid is baenid braincase that has been naturally common to other casichelydian groups and eroded, revealing this structure. The canalis there seem to be only two features of system- is seen to curve medially anterior to the fo- atic significance: first, the position of the foramen posterius canalis carotici interni then ramen posterior canalis carotici interni, and it straightens somewhat before ending at the second, the degree of ossification of the ros- foramen anterius canalis carotici interni. Evi- trum basisphenoidale. Except for these fea- dence for an arterial anastomosis is present tures, discussed below, the baenid basisphe- below the dorsum sellae in the smooth, con- noid agrees with other casichelydians and vex connection between the two foramina differs from Proganochelys in the following anterius canalis carotici interni. All other characters: baenid basisphenoids that I have seen simply have the foramina end directly into the con- 1. Broad sutural attachment of pterygoid cave sella turcica without a smoothly curved to lateral and anterior margins of ba- wall connecting them. An anastomosis could sisphenoid. still have been present in these. 2. Absence ofbasipterygoid processes and The ossification of the trabeculae into a basicranial articulation. median rostrum basisphenoidale is incom- 3. Flattened, platelike ventral surface of plete in baenids. The condition is as yet un- basisphenoid. known in Proganochelys but pleurodires 4. Foramen posterior canalis carotici in- have the rostrum well ossified into a distinc- terni not formed or formed only in part tive median bar anterior to the sella turcica. by basisphenoid. Eucryptodires, however, cannot be charac- The formation of the foramen posterior terized so easily, well-developed rostra being canalis carotici interni midway along the common in chelonioids and testudinoids, length of the basisphenoid-pterygoid suture unossified rostra being typical of triony- is typical in baenids and glyptopsids. This choids. Enough exceptions are present to give character has been interpreted both as de- this character low resolving power. rived (Gaffney, 1972, 1975) and as primitive The degree of overhang of the dorsum sel- (Evans and Kemp, 1975, 1976; Gaffney, lae over the sella turcica in baenids is typi- 1979a) for the Cryptodira. The implications cally that seen in Eubaena cephalica, AMNH of these contradictory hypotheses are dis- 4948 (figs. 6-7). In this specimen the foram- cussed more fully in Gaffney (1979a). Within ina anterius canalis carotici interni are just the Baenidae the size and positions of the barely hidden by the dorsum sellae. In some canalis caroticus intemus and the foramen baenids (e.g., Palatobaena bairdi, UCMP posterius canalis carotici interni do not offer 114539, fig. 12) the foramina are placed more systematically useful variations. The diam- anteriorly in the sella turcica while in others eter of the foramen and canalis are roughly (e.g., unidentified Hell Creek baenid, MCZ one-halfthe diameter ofthe canalis stapedio- 3532, fig. 10; and Plesiobaena putorius, PU temporalis which is the common pattern in 14984) the foramina are more posterior. As chelonioids but in testudinoids the canalis far as I can determine at present, only Chi- caroticus internus is usually less than halfthe sternon has a in this re- diameter of the canalis stapedio-temporalis unique morphology and in trionychoids the canalis caroticus in- gion. In the only two specimens showing this ternus is usually larger than the canalis sta- area, YPM 3939 and AMNH 5904, the pos- pedio-temporalis. The baenid pattern is terior wall of the dorsum sellae has paired probably primitive for cryptodires but I did anteriorly facing concavities just below the not see that any phylogenetic significance can processi clinoideus with a median concavity be obtained from this. between them. 1982 GAFFNEY: BAENID TURTLES 19

ENDOCRANIAL CAST In Plesiobaena the "rider" is W-shaped rather than knoblike as in Bothremys. One of the Plesiobaena antiqua skulls The brain flexure referred to by Edinger (UCMP 49759; Gaffney, 1979b, fig. 162) has (1929) and Zangerl (1960) as occurring be- a matrix-free endocranium and a cast of the tween the olfactory lobes and the cerebrum endocranium has been made by dripping liq- is not as marked in Plesiobaena and Both- uid latex into the various skull openings and remys as it is in sea turtles. However, an en- rotating the skull while the latex dried. The docast made from an unidentified baenid figures of the resulting endocast (fig. 13) are braincase (PU 16838, discussed in Gaffney, partially reconstructed by using the skull as 1972, p. 269) shows a flexture very similar well as the endocast. The reader is referred in degree to that seen in cheloniids. This to Edinger (1934), Zangerl (1960), Gaffney braincase and cheloniids both are character- and Zangerl (1968), and Gaffney (1977) for ized by a deep crista supraoccipitalis and it other descriptions of turtle endocasts. is likely that this feature of the endocast is The roofofthe fossa nasalis consists oftwo more related to jaw muscle architecture than ventrally open concavities housing the olfac- brain architecture. The hindbrain region on tory region in life. Nasal bones occur in Ple- the endocast is marked by a slight median siobaena but the nasal-frontal suture is not rise, absent in most turtle endocasts I have distinct. The hypothesized position of this seen. This raised area would appear to be suture indicates that the major portion ofthe posterior to the optic lobes and lie in the fossa nasalis roofis formed by the nasals and cerebellum region. Whether or not the en- that the frontal has an anterior extension docast morphology reflects brain morphol- along the midline. The dorsal portion of the ogy in this instance is an open question. fissura ethmoidalis is marked by a constric- The side walls of the braincase show the tion (seen in dorsal view) in the endocast cranial nerve exits and vessel positions. Be- between the area of the fossa nasalis and the ginning anteriorly the foramen nervi tri- sulcus ethmoidalis. This region of the fissura gemini and the foramen cavemosum (the was broad and not constricted as in some anterior opening of the canalis cavemosus) emydids. The sulcus ethmoidalis is a ven- extend laterally near the ventral part of the trally open trough extending from the fossa endocast. The distinct swelling formed at the nasalis to the region of the cerebral hemi- junction of these two extensions presumably sphere. The sulcus is broad and without high housed the trigeminal ganglion as in recent boundary ridges in Plesiobaena in contrast turtles. Posterior to the trigeminal area is the to some turtles that have a narrow, well-de- fossa acustico-facialis, a low convexity (con- fined sulcus ethmoidalis. The endocast shows cavity in the bone) containing three foram- the dorsal and posterior margins of the fo- ina. The anteriormost foramen is the fora- ramen interorbitale, a large open space in the men nervi facialis, whereas the posterior two ethmoid region. are presumably the foramina nervi acustici. The dorsal region of the endocast is dom- The hiatus acusticus is the wall that lies inated by relatively well-developed convex- between the cavum labyrinthicum and the ities that presumably housed the cerebral cavum acustico-jugulare. In most turtles this hemisphere in life. The cerebral convexities wall is cartilaginous but in baenids it is char- are relatively well developed in comparison acteristically ossified (Gaffney, 1979b, fig. to the Bothremys barberi endocast figured by 71). The Plesiobaena endocast shows this Gaffney and Zangerl (1968, fig. 21 E) but they typical baenid condition and the region ofthe are smaller than in Bothremys cooki (ibid., hiatus is marked by a shallow concavity. fig. 21C). Between the swellings is the car- Although the occurrence is too poorly known tilaginous "rider" mentioned by Zangerl to generalize, it would appear that most tur- (1960, p. 29) that in life contained the car- tles have a branch ofthe acoustic (VIII) nerve tilaginous anterior end of the supraoccipital. that traverses the hiatus (Gaffney, 1979b). 20 AMERICAN MUSEUM NOVITATES NO. 2737

I p

11 f ..

-1.1

-0fer -,

Wi o_ ,,,

A" < :

*s sL-q..; .."-,. .s _ FIG. 13. Endocranial cast ofPlesiobaena antiqua (UCMP 49759), Late Cretaceous, Lance Formation, Lusk (V-5620), Wyoming. Midline skull length 63 mm.

Such a branch seems to be present in.the skull is a well-defined foramen aquaducti vestibuli. of Baena arenosa (YPM 3933) but not in the This foramen is often on the margin or within Plesiobaena endocast. It is possible, however, the cartilage of the hiatus acusticus in other that the foramen was obstructed on both turtles and may be used as a landmark for sides and did not affect the endocast. this region. Two sutures are visible extending Posterodorsal to the fossa acustico-facialis from the foramen, an anterior one separating 1982 GAFFNEY: BAENID TURTLES 21

cartilaginous rider

basis tuberculi basalis foramen nervi abducentis sulcus cavernosus FIG. 13. Continued. the supraoccipital and the prootic and a ven- and the exoccipital posteriorly but sutures tral one separating the prootic from the op- separating these elements are lacking. Two isthotic. Within the opisthotic region lies the foramina nervi hypoglossi are present pos- foramen medialis nervi glossopharyngei and terior to the foramen jugulare anterius, as in the foramen jugulare anterius. That opening nearly all turtles. The endocast of Plesio- is usually formed by the opisthotic anteriorly baena was extended to the margins of the 22 AMERICAN MUSEUM NOVITATES NO. 2737 foramen magnum and around the dorsal and Gaffney, E. S. posterior surface of the condylus occipitalis. 1972. The systematics ofthe North American The ventral surface of the endocast is im- family Baenidae (Reptilia, Cryptodira). perfect anteriorly, so the complexities of pal- Bull. Amer. Mus. Nat. Hist., vol. 147, atal were art. 5, pp. 241-320. morphology ignored. The sella tur- 1975. A phylogeny and classification of the cica seems to be obscured along its left edge higher categories of turtles. Ibid., vol. by matrix but the right foramen anterius can- 155, art. 5, pp. 387-436. alis carotici intemi can be identified. The 1977. An endocranial cast of the side-necked dorsum sellae and the dorsal surface of the turtle, Bothremys, with a new recon- basisphenoid are readily identifiable. The struction of the palate. Amer. Mus. No- sulcus cavemosus and foramen cavernosum vitates, no. 2639, pp. 1-12. are apparent in ventral and lateral views. The 1979a. The Jurassic turtles of North America. right sulcus and foramen are obscured by Bull. Amer. Mus. Nat. Hist., vol. 162, matrix and the figures are restored from the art. 3, pp. 91-1 36. 1979b. Comparative cranial morphology of re- left side. cent and fossil turtles. Ibid., vol. 164, LITERATURE CITED art. 2, pp. 65-375. Gaffney, E. S., and R. Zangerl Albrecht, P. W. 1968. A revision of the chelonian genus Both- 1967. The cranial arteries and cranial arterial remys (Pleurodira: Pelomedusidae). foramina ofthe turtle genera Chrysemys, Fieldiana: Geol., vol. 16, pp. 193-239. Sternotherus, and Trionyx: a compara- Hay, 0. P. tive study with analysis ofpossible evo- 1904. On some fossil turtles belonging to the lutionary implications. Tulane Studies Marsh collection in Yale University Zool., vol. 14, pp. 81-99. Museum. Amer. Jour. Sci., vol. 18, pp. Archibald, J. D. 261-276. MS. Fossil Mammalia and Testudines of the 1905. On the group of fossil turtles known as Hell Creek Formation, and the geology the Amphichelydia, with remarks on the of the Tullock and Hell Creek Forma- origin and relationships of the subor- tions, Garfield County, Montana. Univ. ders, superfamilies, and families of Tes- Calif. Berkeley, Ph.D. dissertation, 1977. tudines. Bull. Amer. Mus. Nat. Hist., Archibald, J. D., and J. H. Hutchison vol. 21, art. 9, pp. 137-175. 1979. Revision ofthe genus Palatobaena (Tes- 1906. On two interesting genera ofEocene tur- tudines, Baenidae), with the description tles, Chisternon Leidy and Anosteira of a new species. Postilla, no. 177, pp. Leidy. Ibid., vol. 22, art. 9, pp. 155-160. 1-19. 1908. The fossil turtles ofNorth America. Car- Edinger, T. negie Inst. Washington Publ., no. 75, pp. 1929. Die fossilen Gehime. Erg. Anat. Ent- 1-568. wicklungsgesch., vol. 28, pp. 1-249. Romer, A. S. 1934. Anton Fritsch's "Grosshirn von Poly- 1956. The osteology ofthe reptiles. Univ. Chi- ptychodon" ist der Steinkern eines cago Press, xxi + 772 pp. Schildkrotenschadels. Psychiatr. Neu- Schumacher, G.-H. rol. Bladen, nos. 3-4, pp. 396-404. 1973. The head muscles and hyolaryngeal Evans, J., and T. S. Kemp skeleton of turtles and crocodiles. In 1975. The cranial morphology of a new lower Gans, C., and T. S. Parsons (eds.), Bi- Cretaceous turtle from southern En- ology ofthe Reptilia, vol.4, pp. 10 1-199. gland. Palaeontology, vol. 18, pp. 25-40. Zangerl, R. 1976. A new turtle skull from the Purbeckian 1960. The vertebrate fauna of the Selma For- of England and a note on the early di- mation of Alabama. V. An advanced chotomies of cryptodire turtles. Ibid., cheloniid sea turtle. Fieldiana: Geol. vol. 19, pp. 317-324. Mem., vol. 3, no. 5, pp. 281-312.