Brasilemys josai Lapparent de Broin, 2000b

GAFFNEY, EUGENE S, TONG, HAIYAN & MEYLAN, PETER A, 2006, EVOLUTION OF THE SIDE-NECKED TURTLES: THE FAMILIES BOTHREMYDIDAE, EURAXEMYDIDAE, AND ARARIPEMYDIDAE, Bulletin of the American Museum of Natural History 300 (300), pp. 1-698 : 44-56

Brasilemys josai Lapparent de Broin, 2000b

TYPE SPECIMEN: Museu de Geologia de Barcelona, Catalunya, Spain, MGB 37911, partial skull and partial carapace ( Lapparent de Broin, 2000b).

TYPE LOCALITY: Chapada do Araripe, Ceará State, Brazil ( Lapparent de Broin, 2000b) (fig. 13).

HORIZON: Romualdo Member, Santana Formation, Albian ( Lapparent de Broin, 2000b).

DIAGNOSIS: As for genus; see also Lapparent de Broin (2000b).

ETYMOLOGY: For Joan Josa ( Lapparent de Broin, 2000b).

REFERRED MATERIAL: None.

PREVIOUS WORK: Only the type description ( Lapparent de Broin, 2000b).

DISCUSSION: This partial skull and partial carapace provide enough characters for a resolution showing it as the sister taxon to Hamadachelys + Podocnemididae , a result that agrees with the branching diagram (no character matrix is presented) in Lapparent de Broin (2000b) and the computer-assisted cladogram of Fuente (2003).

Hamadachelys Tong and Buffetaut, 1996

TYPE AND ONLY INCLUDED SPECIES: Hamadachelys escuilliei Tong and Buffetaut, 1996 .

DISTRIBUTION: Cenomanian of Morocco.

ETYMOLOGY: From Hamada du Guir ( Tong and Buffetaut, 1996).

REVISED DIAGNOSIS: Podocnemidinura with a partially developed cavum pterygoidei, as in Brasilemys , but in contrast to Podocnemididae ; quadratojugal-parietal contact present, as in Podocnemididae , but in contrast to Brasilemys ; incisura columellae auris fully enclosing stapes and eustachian tube, as in Podocnemididae , but in contrast to Brasilemys ; dentary widely exposed laterally, as in Brasilemys , but in contrast to Podocnemididae ; fossa precolumellaris deep, as in most Podocnemididae , but in contrast to Brasilemys ; foramen jugulare posterius completely enclosed, as in Podocnemididae , but in contrast to Brasilemys ; exoccipital-quadrate contact absent, as in Podocnemididae , but in contrast to Brasilemys .

Hamadachelys escuilliei Tong and

Buffetaut, 1996

TYPE SPECIMEN: MDEt-T-03, a nearly complete skull and jaws.

TYPE LOCALITY: Hamada du Guir, near Taouz, Morocco ( Tong and Buffetaut, 1996).

HORIZON: Kem Kem red beds, Cenomanian ( Sereno et al., 1996; Tong and Buffetaut, 1996).

DIAGNOSIS: As for genus; see also Tong and Buffetaut, 1996.

ETYMOLOGY: For F. Escuillié ( Tong and Buffetaut, 1996).

PREVIOUS WORK: Tong and Buffetaut (1996).

DISCUSSION: This taxon is represented by skull material but no shells. A significant number of characters (but not all) can be coded. The resolution in the MPC shows Hamadachelys as the sister taxon to the Podocnemididae , in agreement with Fuente (2003). Hamadachelys , as well as Brasilemys , could be included in the Podocnemididae . We exclude both from an expanded Podocnemididae to reflect current usage (e.g., Tong and Buffetaut, 1996; Lapparent de Broin, 2000b; Fuente, 2003), which is more restricted (see table 4).

FAMILY PODOCNEMIDIDAE COPE, 1868

TYPE GENUS: Podocnemis Wagler, 1830 .

INCLUDED GENERA: Podocnemis Wagler, 1830 ; Peltocephalus Duméril and Bibron, 1835 ; Erymnochelys Baur, 1888 ; Bairdemys Gaffney and Wood, 2002 ; Dacquemys Williams, 1954c ; Neochelys Bergounioux, 1954 ; Shweboemys Swinton, 1939 ; Bauruemys Kischlat, 1994 ; Stupendemys Wood, 1976 ; Stereogenys Andrews, 1901 .

DIAGNOSIS: Member of the epifamily Podocnemidinura uniquely possessing a fully developed, medially extensive cavum pterygoidei and a dentary covered laterally by the surangular, in contrast to Brasilemys and Hamadachelys ; agreeing with Hamadachelys in basioccipital-opisthotic contact (not known in Brasilemys , may be at Podocnemidinura level), incisura columellae auris enclosing stapes and eustachian tube, usually deep fossa precolumellaris, completely closed foramen jugulare posterius, and exoccipital-quadrate contact absent, all in contrast to Brasilemys .

DISCUSSION: The formation of the family name has varied between ‘‘Podocnemididae’’ ( Broin, 1977, 1988; Lapparent de Broin, 2000a, 2000b, 2001, and many earlier papers; Fuente, 2003) and ‘‘Podocnemidae’’ ( Meylan, 1996). We choose to use ‘‘Podocnemididae’’, mostly because it is the name most frequently used and has the longest usage. It has been the experience of the senior author that classical authorities have different opinions on the correct root for ‘‘-nemis’’ or ‘‘-emis’’, which is usually (but not always) considered comparable to ‘‘-emys’’ (as in Emydidae ). Yes, it’s true. I don’t give a rat’s ass which is used.

EPIFAMILY BOTHREMYDINURA BAUR, 1891

TYPE GENUS: Bothremys Leidy, 1865 .

INCLUDED TAXA: Family Bothremydidae .

DIAGNOSIS: Same as family Bothremydidae .

FAMILY BOTHREMYDIDAE BAUR, 1891

TYPE GENUS: Bothremys Leidy, 1865 .

INCLUDED GENERA: Kurmademys Gaffney, Chatterjee, and Rudra, 2001 ; Sankuchemys Gaffney, Sahni, Schleich, Singh, and Srivastava, 2003 ; Cearachelys Gaffney, Campos, and Hirayama, 2001 ; Galianemys Gaffney, Tong, and Meylan, 2002 ; Foxemys Tong, Gaffney, and Buffetaut, 1998 ; Polysternon Portis, 1882 ; Elochelys Nopcsa, 1931 ; Zolhafah Lapparent de Broin and Werner, 1998 ; Rosasia Carrington da Costa, 1940 ; Araiochelys , n. gen.; Bothremys Leidy, 1865 ; Chedighaii , n. gen.; Taphrosphys Cope, 1869a ; Labrostochelys , n. gen.; Phosphatochelys Gaffney and Tong, 2003 ; Ummulisani , n. gen.; Rhothonemys , n. gen.; Azabbaremys Gaffney, Moody, and Walker, 2001 ; Nigeremys Broin, 1977 ; Arenila Lapparent de Broin and Werner, 1998 .

DISTRIBUTION: Late Cretaceous of India, Late Cretaceous and Paleocene of Europe, Late Cretaceous and Paleogene of North and Central Africa (including Middle East), Late Cretaceous of Madagascar, Late Cretaceous and Paleocene (if Taphrosphys from New Jersey is Paleocene) of North America, Early Cretaceous of South America.

REVISED DIAGNOSIS: Member of the superfamily Podocnemidoidea with wide prefrontals, in contrast to Pelomedusidae and Euraxemydidae , wide premaxillary depression (narrow in some Bothremydini and some Taphrosphyini ); triturating surfaces highly diverse but primitively moderately wide; moderate to large palatine contribution to triturating surface (except in Taphrosphyini ), in contrast to all other pleurodires except Araripemys ; maxilla-quadratojugal contact present (except in Taphrosphyini ); wide exoccipital-quadrate contact present, in contrast to all other pleurodires, which have narrow or no contact; processus paroccipitalis does not project posterior to squamosal, in contrast to all other Pelomedusoides; eustachian tube separated from stapes by bone, in contrast to all other pleurodires; incisura columellae auris closed and stapes contained in bony canal (except Cearachelys , Foxemys , Polysternon ), in contrast to all other pleurodires; fossa precolumellaris absent (except in Kurmademys ), in contrast to all other pleurodires; cavum pterygoidei as seen in Podocnemididae absent; fossa orbitalis posterior enlargement present (except in Cearachelys ), in contrast to all other pleurodires; supraoccipital-quadrate contact present (except in Taphrosphyini and Zolhafah ); prootic partially or completely covered in ventral view; foramen posterius canalis carotici interni not in prootic, in contrast to Araripemydidae , Chelidae , and Pelomedusidae ; basisphenoid-quadrate contact present, as in Podocnemididae , but in contrast to all other pleurodires; high lingual ridge on lower jaw.

PREVIOUS WORK: In 1891, George Baur recognized three living families of pleurodires, the Sternothaeridae (the Pelomedusidae of former usage), the Podocnemididae , and the Chelyiidae ( Chelidae of current usage) ( Baur, 1891: 420). Based on the presence of deep pits in the jaws, which he thought might be alveoli for large tusks, he thought that Bothremys Leidy should be placed in a family of its own, and he coined the term Bothremydidae ( Baur, 1891: 424) . Thus, Baur’s (1891) Pleurodira consisted of four families rather than the two that have been recognized for much of the last century.

Baur (1893) reiterated the uniqueness of the Bothremydidae and used the superfamily name Pelomedusoidea for a group of three of his four pleurodire families: the Pelomedusidae , which he restricted to two genera, Pelomedusa and Sternothaerus (5 Pelusios ); the Podocnemididae , in which he placed Podocnemis , Peltocephalus and Erymnochelys ; and ‘‘the intermediate extinct family Bothremydidae’’. This is very close to the current usage developed for ‘‘Pelomedusoides’’.

Baur’s Bothremydidae was adopted by Hay (1908) in his classic monograph on North American fossil turtles. It was also used by Nopcsa (1923) and Dollo (1924) but then nearly disappeared from use for about 65 years. In his treatment of North American bothremydids, Hay (1908) included Bothremys Leidy , Taphrosphys Cope , and two new genera, Amblypeza and Naiadochelys (both considered invalid here). Schmidt (1940), in his description of Podocnemis barberi (now Chedighaii barberi ), assigned his new taxon to a Pelomedusidae , which he acknowledged was used in an inclusive sense. He specifically mentioned the Bothremydidae , stating that Taphrosphys , which was placed in the Bothremydidae by Hay (1908), may be allied to Podocnemis , but he left the issue of revision of the Pleurodira for future workers.

Carrington da Costa (1940) tentatively referred his new genus Rosasia to the family Pelomedusidae and the ‘‘subfamily’’ Bothremydidae . He was explicit in his conclusion (apparently in consultation with F.-M. Bergounioux) that the Bothremydidae should be treated as a subfamily. He contemplated referring Rosasia to the Bothremydidae again in 1958 ( Carrington da Costa, 1958), but instead referred it to the Pelomedusinae. Williams (1950) used Bothremydinae as a subfamily of the Pelomedusidae , along with the Pelomedusinae. However, in the turtle classification of Romer (1956), which was significantly influenced by Williams, the Bothremydinae is gone; only the Pelomedusidae and Chelyidae make up the Pleurodira .

Baur’s Bothremydidae appeared next in a review of the genus Bothremys ( Gaffney and Zangerl, 1968) , which was prompted by the discovery of the first skull–shell association of a bothremydid, a specimen of Bothremys (now Chedighaii ) barberi from the Selma Formation of Alabama. In this paper several shell taxa previously assigned to Podocnemis were referred to Bothremys . Although formal recognition of the family Bothremydidae was deferred, the uniqueness of the Bothremys lineage was noted. Gaffney (1975b) gave the group formal recognition as a subfamily of the Pelomedusidae . The subfamily was formally resurrected for the genera Taphrosphys and Bothremys in this paper, which gave a detailed description of the shell of Taphrosphys from the Late Cretaceous (now considered Paleocene) of New Jersey.

This family group name was also briefly considered by Broin (1977) in her treatise on French fossil turtles. In that work she reviewed the higher categories of pleurodires and considered the taxonomy of Baur (1888, 1891) to be the most satisfactory. She suggested that reuniting the genera Bothremys , Nigeremys , and Taphrosphys in the Bothremydidae would be desirable, but she did not follow the lead of Gaffney (1975b) until she published a further review of fossil pleurodires ( Broin, 1988). In this stratigraphic treatment of geographic dispersion of pleurodires, she referred a series of genera to the family Bothremydidae , which she included in the hyperfamily Pelomedusoides along with the Podocnemididae . The genera referred to the Bothremydidae included Taphrosphys , Bothremys , Nigeremys , Rosasia , Apertotemporalis , Apodichelys , and Sokotochelys ( Broin, 1988). We consider the last three to be invalid or incertae sedis.

The redescription of the bothremydid Rosasia by Antunes and Broin (1988) provided the first review of the family Bothremydidae . It included a tentative phylogeny and a detailed description of the skull of Rosasia , a taxon that was previously known only from shell material. The cladogram showed that the Bothremydidae is the sister group of the Podocnemididae . Furthermore, within the Bothremydidae , Bothremys , Nigeremys , and Taphrosphys informal groups were recognized. However, the phylogenetic position of many named taxa, especially those known only from shells, remained unresolved. The genera treated as members of the Bothremydidae in Antunes and Broin (1988) included Apodichelys , Apertotemporalis , Bothremys , Elochelys , Nigeremys , Rosasia , and Taphrosphys .

In Antunes and Broin (1988), the Bothremydidae was characterized by five skull, seven shell, and one vertebral character. Skull characters included (Broin’s character numbers) (C 1) eustachian tube excluded from stapedial canal by quadrate, (C 2) strong development of the jaws and vomer, and secondary closure of temporal and cheek emargination, (C 3) low wide external narial openings, (C 4) ventral coverage of the prootic by the pterygoid, and (C 5) flattening of the anterior part of the muzzle. Shell features characteristic for the family included (C 6) large and depressed form of the subquadrangular carapace, with a smooth semicircular arc without inflations or keels and a trapezoidal ventral profile, (C 7) anterior plastral lobe particularly short and wide at its base, trapezoidal or subtrapezoidal, (C 8) vascular sculpture of the carapace consisting of dichotomous (branching) grooves well developed, and more so in the marine Taphrosphys group, (C 9) frequent suturing of the ilium to the suprapygal (this was considered to be a reversal to the primitive condition), (C 10) reduction in the number of neurals to seven or fewer, which is acknowledged to occur several times in pleurodires, (C 11) reduction of the width of vertebral 1, which does not cover all of the nuchal bone, and with the following vertebrals having an anterior width less than or equal to their median length (parallelism in this character was acknowledged), and (C 12) pectoral–abdominal sulcus crossing the anterior part of the mesoplastra (parallelism in this character was acknowledged). A single character of the vertebrae (C 13) was acknowledged to be variable in the three taxa for which cervical vertebrae were known.

In the same paper, the Bothremys group of the Bothremydidae was characterized by three skull and two shell characters: (D 1) maximum enlargement of the triturating surfaces with a deep fossa present in the palatine and jugal and in the dentary, (D 2) secondary covering of the cheek emargination by posterior extension of the maxilla, ventral enlargement of the quadratojugal, and anterior expansion of the quadrate, (D 3) foreshortening of the bones at the back of the skull, particularly the supraoccipital, the basisphenoid, and the basioccipital, such that the occipital condyle is in line with the articular facets of the quadrate, (D 4) tendency for elongation of the bridge, and (D 5) formation of a more-or-less strongly developed nuchal embayment that affects the nuchal scute.

The Nigeremys group was characterized by four characters: (F 1) deep ventral premaxillary pit but without formation of a recurved beak, (F 2) deep carotid fossa in the area of the quadrate–basisphenoid suture, (F 3) flattening and anterior enlargement of the snout with pronounced posterior elevation of the skull roof with a tectiform or subtectiform profile, and (F 4) marked increase in size.

Nearly all of these characters from Antunes and Broin (1988), or modifications of them, have been used in the dataset presented here (appendices 2 and 3). This work provided the basis for future analyses of the Bothremydidae .

A byproduct of Meylan’s (1996) phylogenetic analysis of Araripemys was the first computer-assisted cladistic analysis of the Bothremydidae . Three bothremydids with available skull information were included in that analysis: Bothremys , Rosasia , and Taphrosphys . The Bothremydidae was monophyletic in this analysis and was the sister group of the Podocnemididae , as suggested by Antunes and Broin (1988) and Broin (1988). Beginning with this work, paleontologists have generally followed the taxonomy of Baur (1891) as suggested by Broin (1988) in recognizing more family-level taxa in the pleurodires. Neontologists have generally followed this development by splitting the living Pelomedusidae (sensu lato) into a more restricted Pelomedusidae and recognizing the family Podocnemididae ( Pough et al., 1998, 2001; Zug et al., 2001).

In her description of Brasilemys, Lapparent de Broin (2000b) provided an explicit hypothesis for the phylogenetic position of the Bothremydidae . This family, including an undescribed form from Erfoud, Hamada de Guir, Morocco (now known as Galianemys Gaffney, Tong, and Meylan, 2002 ), was argued to be the sister group of the Podocnemididae plus the genera Hamadachelys and Brasilemys , a group that she termed the Podocnemidoidae (here renamed the Podocnemidinura).

A diverse fauna of bothremydid turtles was described by Lapparent de Broin and Werner (1998) from the Maastrichtian of Egypt. They described what they thought to be five taxa representing three separate lineages of bothremydid turtles. Two were described from skulls and three from shell fragments. Their new genus Zolhafah was based on a single skull and allied with members of the ‘‘ Bothremys group’’, Bothremys and Rosasia . A second new genus, Arenila , was based on a poorly preserved skull and allied with Nigeremys in the ‘‘ Nigeremys group’’. A carapacial disc was also provisionally referred to this new taxon. Two other poorly known genera, Sokotochelys and Apertotemporalis (here considered nomina nuda), were referred to this group. The ‘‘ Taphrosphys group’’ was represented by two carapace fragments.

Lapparent de Broin and Werner (1998) characterized the skull of members of the Bothremys group as having pits in the triturating surface and as having a tendency to a flat posterior palatal surface with reduction of the ‘‘podocnemidoid fossa’’. Lapparent de Broin and Werner (1998) characterized the shell of this group as ‘‘a rather wide rectangular-rounded shell with a notched nuchal, an anterior trapezoidrounded [plastral] lobe, short and posteriorly wide, a posterior [plastral] lobe narrower at its base than the anterior lobe and with straight posteriorly converging borders.’’

They characterized the skull of members of the Nigeremys group as having: (1) a much enlarged depression in the area of the pterygoid-basisphenoid and pterygoid-quadrate suture (this depression, the fossa pterygoidea, was considered by them to be homologous to the ‘‘podocnemidoid fossa’’ or cavum pterygoidei, a conclusion we dispute), (2) an enlarged snout, (3) basioccipital participating in occipital condyle, and (4) an enlarged and posteromedially projecting trochlear process (here interpreted as an artifact of preservation in Arenila ). The shell of members of the Nigeremys group differed from other podocnemidoids in having the axillary buttress cross the second rather than the third peripheral to reach the first costal bone.

Lapparent de Broin and Werner (1998) relied entirely on shell characters to distinguish the Taphrosphys group for which they said skulls were too poorly known. Referral of new fossil shell material appeared to be based largely on the ‘‘typical very marked decoration of small prominent polygons and granulation and rounded crests.’’ Elochelys and Taphrosphys were reported to share a shell that is ‘‘elongated rounded ovoid anteromedial shell, intergular separating the gulars and meeting the pectorals, rounded posterior lobe in most Taphrosphys specimens and in Elochelys , well trapezoid anterior lobe.’’ They considered Gafsachelys Bergounioux, 1956 to be a member of this group on the basis of shell decoration.

These authors suggested that the shell shapes common to the Taphrosphys and Bothremys groups indicated that they formed a monophyletic sister group to the Nigeremys group ( Lapparent de Broin and Werner, 1998: 164). This was a refinement of the trichotomy indicated for these groups in Antunes and Broin (1988).

Since 1998, the number of genera referred to the Bothremydidae has more than doubled. In addition to Zolhafah and Arenila discussed above ( Lapparent de Broin and Werner, 1998), Tong et al. (1998) described the French bothremydid Foxemys and presented a cladogram with a resolution of ( Foxemys ( Taphrosphys ( Bothremys , Rosasia ))), a result (provided by the senior author) that is inconsistent with our current results, due to the smaller data set. More recently, Gaffney and colleagues have described six additional genera. Phosphatochelys ( Gaffney and Tong, 2003) was described on the basis of a single skull of a short-faced member of the Nigeremys group. Another new member of the Nigeremys group is the remarkable skull taxon Azabbaremys from the Paleocene of Mali ( Gaffney, Moody, and Walker, 2001). The large and strongly roofed skull of the type is quite similar in many respects to Nigeremys itself. The oldest bothremydid described to date is the Albian genus Cearachelys ( Gaffney, Campos, and Hirayama, 2001) , which was described from two nearly complete skeletons from the Chapada do Araripe in Ceara, Brazil. The three well-preserved skulls and shells of this taxon provide excellent evidence for the primitive skeletal morphology of the family. This taxon was referred by its describers to the Bothremys group of Lapparent de Broin and Werner (1998). A morphologically similar and apparently related genus, Galianemys ( Gaffney, Tong, and Meylan, 2002) , has been described from the Cenomanian Kem Kem Redbeds of Morocco. The two Indian bothremydids Sankuchemys (Gaffney et al., 2003) and Kurmademys ( Gaffney, Chatterjee, and Rudra, 2001) considerably extend the geographic range of the Bothremydidae .

DISCUSSION: As a result of the discovery of many new bothremydid skulls that are used as the basis for a phylogenetic analysis of the Bothremydidae presented here, four monophyletic groups are recognized as tribes within the Bothremydidae (figs. 1, 2; table 5). The idea of the ‘‘ Bothremys group’’ has been documented and confirmed in this analysis, but the ‘‘ Nigeremys group’’ and ‘‘ Taphrosphys group’’ of Lapparent de Broin and Werner (1998) are combined in the tribe Taphrosphyini .

The possible extension of the Bothremydidae into the Miocene is based on a single questionable record. Roger et al. (1994) reported on a damaged braincase from Miocene rocks from Oman in the Arabian Peninsula. The braincase, an uncataloged specimen with no institutional collection indicated, was not described in detail and was figured only in ventral view, but it was reported to have a ‘‘fosses ptérygoido-carotidiennes obliques’’ ( Roger et al., 1994: 11) and an incisura columellae auris separated

TABLE 5

Tribes of Bothremydidae from the eustachian tube. Unfortunately, we have been unable to see this specimen and cannot confirm these characters. The photograph ( Roger et al., 1994: pl. 1, fig. 1) shows a worn braincase in ventral view that certainly has large depressions in the position of what we would call the fossa pterygoidea. This is a character found in only a few bothremydids as a deep fossa, namely Foxemys , Polysternon , Nigeremys , and Arenila . However, the specimen is clearly eroded on its surface and it is possible that the fossae are actually cavum pterygoidei, with the ventral covering eroded away along the anterior and medial edges, making this a podocnemidid and not a bothremydid character. There is some evidence for this conclusion in the photograph, which shows the two depressions with different shapes, requiring at least some erosion on the right fossa. The second character, the closed incisura columellae auris, is restrict- ed only to bothremydids and does not occur outside the group. However, this character is not visible in ventral view. Furthermore, the specimen is eroded extensively on both sides and is missing most if not all of each cavum tympani, so this character may not be determinable at all in this specimen. It is unfortunate that this specimen is not available for further study, as it represents the best possibility for bothremydids extending past the Eocene. As it stands, however, the published claim, while certainly possible, is inadequate for a range extension past the Eocene into the Miocene.

Extension of the range of bothremydids into the Late Cretaceous of Madagascar is based on the report of Gaffney and Forster (2003). This consists of a partial lower jaw (fig. 247) that has the symphyseal edge and lateral pits typical of the tribe Bothremydini .

A short, wide anterior lobe of the plastron has been used as a diagnostic character for the Bothremydidae ( Broin, 1977, 1988; Lapparent de Broin, 2000a, 2001). However, Bairdemys venezuelensis has a short, wide anterior plastral lobe and a plastral morphology (fig. 275) that is very similar to that in such bothremydids as Chedighaii (fig. 264). However, Bairdemys is clearly a podocnemidid based on the skull morphology ( Gaffney and Wood, 2002). Therefore, this character is not diagnostic for the family Bothremydidae and the records based on this feature are in doubt.

SUBFAMILY KURMADEMYDINAE , NEW

TYPE GENUS: Kurmademys Gaffney, Chatterjee, and Rudra, 2001 .

INCLUDED GENERA: Kurmademys Gaffney, Chatterjee, and Rudra, 2001 ; Sankuchemys Gaffney, Sahni, Schleich, Singh, and Srivastava, 2003 .

DIAGNOSIS: As for tribe Kurmademydini .

TRIBE KURMADEMYDINI , NEW

TYPE GENUS: Kurmademys Gaffney, Chatterjee, and Rudra, 2001 .

INCLUDED GENERA: Kurmademys Gaffney, Chatterjee, and Rudra, 2001 ; Sankuchemys Gaffney, Sahni, Schleich, Singh, and Srivastava, 2003 .

DISTRIBUTION: Late Cretaceous Maastrichtian of Peninsular India.

DIAGNOSIS: Bothremydid pleurodires with the following unique characters: extreme degree of posterior temporal emargination characterized by a short postorbital and no parietal-quadratojugal contact; large fossa precolumellaris (unknown for Sankuchemys ); condylus mandibularis well anterior to main body of basioccipital; small part of prootic exposed on ventral surface at junction of basisphenoid, pterygoid, and quadrate, containing foramen nervi facialis; other differen-

TABLE 6

Genera of Kurmademydini tiating characters are: preorbital part of skull narrow, in contrast to Bothremydini ; jugal-quadrate contact absent and supraoccipital-quadrate contact (unknown for Sankuchemys ) present, in contrast to Taphrosphyini ; foramen stapedio-temporale not very close to foramen nervi trigemini, in contrast to Bothremydini and Taphrosphyini ; fenestra postotica closed, in contrast to Cearachelyini ; basisphenoid pentagonal, in contrast to Cearachelyini ; jugal not retracted from orbit, in contrast to Cearachelyini ; triturating surfaces narrower than in Bothremydini ; parietal contacts pterygoid (unknown for Sankuchemys , also occurs in some Bothremydini and Taphrosphyini ).

DISCUSSION: The sister-group relationship of Kurmademys and Sankuchemys becomes unresolved at two steps in the MPC (fig. 288). The most serious problem is the flattened Sankuchemys skull, preventing nearly all quadrate characters from being determined. The two taxa are united only on the basis of the temporal emargination and the exposure of the quadrate-pterygoid-basisphenoid contact, both of which appear a number of times within pleurodires. Nonetheless, when the tribe Kurmademydini becomes unresolved, these two genera are still sister taxa to the subfamily Bothremydinae (consisting of the tribes Cearachelyini , Bothremydini , and Taphrosphyini ), which holds together for three steps. Table 6 compares the two genera Sankuchemys and Kurmademys .

The phylogenetic analysis of the tribe Kurmademydini (essentially just Kurmademys , as Sankuchemys is poorly preserved) shows it as the sister group to the remaining Bothremydidae . In geographic and stratigraphic terms, this is surprising, as the Kurmademydini is Asian and Maastrichtian, while the oldest bothremydids ( Cearachelys ) are South American and Albian and the next oldest ( Galianemys ) are African and Cenomanian. Furthermore, the tribe Cearachelyini , which contains these bothremydids, also has skull morphology that seems more plesiomorphic for bothremydids than for the kurmademydines. However, the emarginated skull and the deep fossa precolumellaris of Kurmademys are primitive for Pelomedusoides, and the Cearachelyini have such subfamily Bothremydinae synapomorphies as the anteriorly opening foramen stapedio-temporale (fig. 309), large postorbital, maxilla-vomer contact, as well as slight temporal emargination and small or absent fossa precolumellaris. So at the present time, the Kurmedemydini are relatively well-supported as sister taxa to all remaining bothremydids. This indicates that the biogeographic history of the bothremydids is more complex than some recent workers would suggest (e.g. Lapparent and Werner, 1998). The senior author hopes to provide a biogeographic hypothesis for bothremydids in the near future that reflects the cladogram presented here.

Kurmademys Gaffney, Chatterjee, and

Rudra, 2001

TYPE AND ONLY INCLUDED SPECIES: Kurmademys kallamedensis Gaffney, Chatterjee, and Rudra, 2001 .

DISTRIBUTION: Maastrichtian of southern India.

ETYMOLOGY: Kurma, ‘‘turtle’’ in Sanskrit, in allusion to the second-stage incarnation of Lord Vishnu as a turtle in Hindu mythology.

REVISED DIAGNOSIS: Bothremydid of the tribe Kurmademydini with extensive temporal emargination seen also in Sankuchemys but absent in all other Bothremydidae ; differing from Sankuchemys in having a smooth, expanded triturating surface rather than a narrower one with an accessory ridge; foramen posterius canalis carotici interni in basisphenoid unique among Bothremydidae ; foramen stapedio-temporale visible in dorsal view in contrast to all other bothremydids (unclear in Sankuchemys ). Carapace low and oval; shell surface texture weak granulated polygons; six neurals, sixth to eighth costals meeting on the midline; plastron with anterior lobe longer than in Bothremydini and Taphrosphyini ; bridge longer than anterior and posterior plastral lobes; posterior plastral lobe short; humeropectoral sulcus posterior to epihyoplastral suture, crossing entoplastron; pectoroabdominal sulcus far anterior to mesoplastron, as in Araiochelys ; pubic and ischiac scars as in Bothremydini .

DISCUSSION: Kurmademys is well known from a series of skulls and postcrania, including a shell (figs. 255–257) that provides most of the characters for the tribe Kurmademydini in the dataset. The other postcrania have not yet been studied. Comparison with Sankuchemys is given in table 6.

Kurmademys kallamedensis Gaffney,

Chatterjee, and Rudra, 2001

TYPE SPECIMEN: ISI R152 (figs. 56, 57, 63, 64, 282B), a nearly complete skull lacking the dorsal part of the prefrontals, the posterior part of the crista supraoccipitalis, and part of the left quadratojugal.

TYPE LOCALITY: Near the village of Kallamedu, Tamil Nadu, southern India. Map of locality is in Sastry et al. (1972) (fig. 12).

HORIZON: Kallamedu Formation of the Ariyalur Group. Formation named and described by Sastry et al. (1972), who correlated it with the uppermost Maastrichtian; the Cretaceous–Tertiary boundary is its upper limit ( Sastry et al., 1972). Dinosaurs have been described from other exposures of the Kallamedu ( Matley, 1929; Yadagiri and Ayyasami, 1987). The Kurmademys locality is a small pocket of fine-grained sandstone and clay, about 6 in. thick. It also contained crocodiles, gar scales, and freshwater gastropods and bivalves and is interpreted as a freshwater pond deposit ( Gaffney, Chatterjee, and Rudra, 2001).

DEPOSITIONAL ENVIRONMENT: Interpreted as a freshwater pond deposit ( Gaffney, Chatterjee, and Rudra, 2001).

DIAGNOSIS: As for genus.

ETYMOLOGY: For the Kallamedu Formation.

REFERRED MATERIAL: ISI R155A, partial skull (fig. 60C); ISI R155B, partial skull (fig. 58C); ISI R155C, partial skull; ISI R158, partial skull (fig. 58A); ISI R159, partial skull (fig. 60A); ISI R155D, lower jaws; ISI R155E, lower jaws (fig. 233); ISI R155F, right ramus; ISI R152, 20 shell elements; ISI R153, 8 shell elements; ISI R157, 74 shell elements; ISI R278, partial shell (figs. 255, 256). See specimen list in Shell Description section for individual shell elements.

Sankuchemys Gaffney, Sahni, Schleich,

Singh, and Srivastava, 2003

TYPE AND ONLY INCLUDED SPECIES: Sankuchemys sethnai Gaffney, Sahni, Schleich, Singh, and Srivastava, 2003 .

DISTRIBUTION: Maastrichtian of Bombay, India.

ETYMOLOGY: Sankuch, ‘‘compressed’’ in Sanskrit, in allusion to the truck that ran over the type specimen.

REVISED DIAGNOSIS: Bothremydid of the tribe Kurmademydini with extensive temporal emargination also seen in Kurmademys , but absent in all other Bothremydidae ; differing from Kurmademys in having a narrow triturating surface with an accessory ridge parallel to the labial ridge (unique among Bothremydidae except for some Foxemys ); foramen posterius canalis carotici interni formed by basisphenoid and pterygoid; foramen stapedio-temporale not visible in dorsal view, as in all other bothremydids except Kurmademys .

DISCUSSION: Although many characters are visible in the type skull of Sankuchemys , one important area completely wrecked in SDS/VPL 1125 is the cavum tympani. This region has many characters important in bothremydid systematics that are not determinable for this taxon. The occiput is also reduced to two dimensions, completely obscuring foramina and features in that area. Nonetheless, SDS/VPL 1125 does have enough characters to show its distinctness from all other taxa and to test its relationships. Table 6 compares it with Kurmademys .

‘‘ Carteremys ’’ leithii is also from the Intertrappean beds in Mumbai; although its age is unclear, it could be Cretaceous or Paleogene. This taxon, originally consisting of shells and skull material apparently unavailable for 150 years and presumably lost, is considered a nomen dubium (see below). The similarity of age and locality between Sankuchemys and ‘‘ Carteremys ’’ suggests that the two may be the same taxon. However, the ‘‘ Carteremys ’’ figures seem to show a broad rather than a narrow skull, suggesting that they are not the same taxon. In any case, there is no material available for this taxon.

Sankuchemys sethnai Gaffney, Sahni,

Schleich, Singh, and Srivastava, 2003

TYPE SPECIMEN: SDS /VPL 1125, a complete skull (figs. 66, 67) completely smashed flat.

TYPE LOCALITY: Amboli Quarry, Jogeshwari, Mumbai, India (map and faunal discussion in Singh et al., 1998) (fig. 12).

HORIZON: Green tuff bed of Amboli, Intertrappean beds, late Maastrichtian (discussion in Singh et al., 1998).

DIAGNOSIS: As for genus.

DEPOSITIONAL ENVIRONMENT: Presumed to be fresh water ( Singh et al., 1998).

ETYMOLOGY: ‘‘In honor of the discoverer of the holotype skull, Prof. S. F. Sethna, known for his pioneering work on the geology of the Mumbai region’’ ( Gaffney, Sahni, Schleich, Singh, and Srivastava, 2003:3).

DISCUSSION: See above.

SUBFAMILY BOTHREMYDINAE BAUR, 1891 ,

NEW RANK

TYPE GENUS: Bothremys Leidy, 1865 .

INCLUDED GENERA: Cearachelys Gaffney, Campos, and Hirayama, 2001 ; Galianemys Gaffney, Tong, and Meylan, 2002 ; Foxemys Tong, Gaffney, and Buffetaut, 1998 ; Polysternon Portis, 1882 ; Elochelys Nopcsa, 1931 ; Zolhafah Lapparent de Broin and Werner, 1998 ; Rosasia Carrington da Costa, 1940 ; Araiochelys , n. gen.; Bothremys Leidy, 1865 ; Chedighaii , n. gen.; Taphrosphys Cope, 1869a ; Labrostochelys , n. gen.; Phosphatochelys Gaffney and Tong, 2003 ; Ummulisani , n. gen.; Rhothonemys , n. gen.; Azabbaremys Gaffney, Moody, and Walker, 2001 ; Nigeremys Broin, 1977 ; Arenila Lapparent de Broin and Werner, 1998 .

DIAGNOSIS: Bothremydidae with this unique character: foramen stapedio-temporale opens anteriorly not dorsally on otic chamber (possibly present in Sankuchemys ); other distinguishing characters are: a relatively large postorbital (except in some Taphrosphyini ), in contrast to Kurmademydinae ; maxilla-vomer contact present (except in Chedighaii and Azabbaremys ), in contrast to Kurmademydinae and Podocnemididae ; very small to absent fossa precolumellaris, unique among pleurodires except for some Podocnemididae ; condylus mandibularis posterior to or on level of basioccipital-basisphenoid suture, in contrast to all other Eupleurodira (except for Polysternon and some Taphrosphyini ); condylus occipitalis formed only by exoccipitals (except in Galianemys ), in contrast to all other pleurodires except Pelomedusidae ; ventral outline of basisphenoid usually triangular; where known, iliac scar on costals 7, 8, and suprapygal.

DISCUSSION: The subfamily Bothremydinae consists of the tribes Cearachelyini , Bothremydini , and Taphrosphyini (figs. 1, 2; table 5). It is relatively well supported in figure 288, with a decay value of 3. The alternate possible relationship of Kurmademydini + Taphrosphyini + Bothremydini is discussed under Kurmademydini .

INFRAFAMILY CEARACHELYODDA , NEW

TYPE GENUS: Cearachelys Gaffney, Campos, and Hirayama, 2001 .

INCLUDED GENERA: Cearachelys Gaffney, Campos, and Hirayama, 2001 ; Galianemys Gaffney, Tong, and Meylan, 2002 .

DIAGNOSIS: Same as for tribe Cearachelyini .

TRIBE CEARACHELYINI , NEW

TYPE GENUS: Cearachelys Gaffney, Campos, and Hirayama, 2001 .

INCLUDED GENERA: Cearachelys Gaffney, Campos, and Hirayama, 2001 ; Galianemys Gaffney, Tong, and Meylan, 2002 .

DISTRIBUTION: Early Cretaceous of Brazil and Late Cretaceous of Morocco.

DIAGNOSIS: Member of the subfamily Bothremydinae with the following unique characters: jugal nearly or completely retracted from orbital margin; fenestra postotica open and formed as a short slit; foramen jugulare posterius open (also in

Foxemys and Polysternon ); other differentiating characters are: temporal emargination greater than in Bothremydini and Taphrosphyini but less than in Kurmademydini ; fossa precolumellaris absent, in contrast to Kurmademydini ; preorbital skull narrow, in contrast to Bothremydini ; jugal-quadrate contact absent, in contrast to Taphrosphyini ; shelf below cavum tympani absent, in contrast to Bothremydodda ; supraoccipital-quadrate contact present; exoccipitals do not completely form neck of condylus occipitalis; foramen stapedio-temporale not visible dorsally; foramen stapedio-temporale and foramen nervi trigemini not very close, in contrast to Bothremydodda ; condylus mandibularis anterior to condylus occipitalis but not as anterior as in Kurmademydini ; palatine forms greater part of triturating surfaces than it does in Taphrosphyini ; triturating surfaces wider than in Taphrosphyini but not as wide as in Bothremydini . Six-sided first neural with short posterolateral sides; four-sided second neural not contacting the first costal; plastron with longer anterior lobe than in other bothremydids; pectoral scale posterior to entoplastron.

DISCUSSION: The tribe Cearachelyini unites Cearachelys (Albian, Brazil) and Galianemys (Cenomanian, Morocco) by a number of unreversed characters. The three species included are similar in morphology but distinguishable by a number of characters (table 7), although Galianemys probably has a much larger shell, if the suggested association with AMNH 30550 and 30551 proves to be correct. These two genera, found on either side of the early Atlantic Ocean, seem to be a vicariant pair, divided by the opening of the Atlantic, possibly comparable to the situation for Hamadachelys and Brasilemys , and for Euraxemys and Dirqadim .

See table 5 for comparison of Cearachelyini with other tribes, and table 7 for comparison of the members of the Cearachelyini .

Cearachelys Gaffney, Campos, and

Hirayama, 2001 TYPE AND ONLY INCLUDED SPECIES: Cearachelys placidoi Gaffney, Campos, and Hirayama, 2001 .

TABLE 7 Genera of Cearachelyini

DISTRIBUTION: Albian of Brazil.

ETYMOLOGY: In allusion to the type locality that is in Ceará State, Brazil, and to chelys, turtle in Greek.

REVISED DIAGNOSIS: Bothremydid pleurodire of the tribe Cearachelyini with the following unique character: jugal nearly or completely retracted from orbital margin but not widely separated from orbit by broad postorbital-maxilla contact, as in Galianemys ; other differentiating characters are: incisura columellae auris open not closed, in contrast to Galianemys ; fossa pterygoidea present, as in Galianemys emringeri , not absent, as in G. whitei ; cheek with slight emargination not straight, as in Galianemys ; interorbital distance narrower than in Galianemys ; labial ridge thin in contrast to thick in Galianemys ; sulcus olfactorius ridge shallower than in Galianemys ; antrum postoticum larger than in Galianemys ; tuberculum basioccipitale small and blunt not larger and shelflike, as in Galianemys ; jugal-palatine contact more extensive than in Galianemys .

Carapace moderately domed as in Pelomedusa , oval in outline, with eight neurals completely separating all eight costals, in contrast to at least one pair of costals meeting in midline as in most other bothremydids; second neural does not contact first costals. Plastron with anterior lobe rounded and broader than in other Santana Pelomedusoides; pectoral scales do not extend anteriorly onto entoplastron, but do extend posteriorly onto mesoplastra; mesoplastron small and laterally placed, as in Podocnemis .

DISCUSSION: Do the three skulls of Cearachelys (table 8) represent more than one species? It could be argued that the swollen triturating surfaces, jugal depression, pinched snout, deep fossa pterygoidea, and more medial foramen posterius canalis carotici interni sufficiently differentiate BSP 1976 I 160 from THUg 1798 to make them separate taxa. However, we interpret these differences as individual variation, possibly growth related, and recognize the three skulls as belonging to one species, Cearachelys placidoi .

Among recent taxa, intraspecific variation, including maxillary swelling and snout pinching, has been reported in emydids, trionychids, and chelids ( Cann, 1998; Carr, 1952; Dalrymple, 1977; Lindeman, 2000). Maxillary swellings typically occur in larger indi-

TABLE 8 Comparison of Cearachelys placidoi Specimens viduals and have been associated (sometimes ambiguously) with age, sex, and dietary differences. Similarly, muscle attachment sites can be exaggerated in any vertebrate, and some individuals show greater ridging, thicker or thinner bone, and deeper concavities. The three Cearachelys skulls show a size progression (table 8). If BSP 1976 I 160 is placed at 100 %, then MPSC is 88 % of BSP 1976 I 160 and THUg 1798 is 73 % of BSP 1976 I 160. This suggests an age correlation with BSP 1976 I 160 as oldest, but it may also be due to gender, as in some living emydids ( Carr, 1952; Dalrymple, 1977; Lindeman, 2000). Nonetheless, the three Cearachelys skulls form a series showing degrees of variation that we interpret as representing a single species.

The swollen maxilla, triturating surface depression, and pinched snout occur in other taxa, but the difference in position of the foramen posterius canalis carotici interni does not occur outside the Bothremydidae . To a certain extent, this may be due to the unusual juxtaposition of a thin-walled canalis caroticus internus within a muscle attachment depression. As described in the Cranial Morphology section ( Cearachelys, Pterygoid ), an older or larger individual often has better defined muscle attachments, and it is likely that the difference in foramen position is related to the increase in depth of the fossa pterygoidea. At present, it seems best to include all the Cearachelys skulls in the same taxon and to interpret their differences as intraspecific variation.