Type Genus

Podocnemis Wagler, 1830.

Included Taxa

Hamadachelys Tong and Buffetaut, 1996; Brasilemys Lapparent de Broin, 2000b; Portezueloemys de la Fuente, 2003; and the family Podocnemididae.

Diagnosis

A member of the superfamily Podocnemidoidea uniquely possessing a cavum pterygoidei formed by basisphenoid, pterygoid, prootic, and quadrate, underlain by pterygoid and basisphenoid (in contrast to the fossa pterygoidea of some Bothremydidae); processus retroarticularis of articular oriented posteroventrally; basioccipital-opisthotic contact present (also in Pelomedusidae and some Chelidae, not known in Brasilemys).

Discussion

See Gaffney et al. (2006) for revised diagnoses and phylogeny of included taxa. We include here a new restoration and new figures of two specimens of Hamadachelys escuilliei (figs. 1–6) for comparison with Podocnemididae. A description of Hamadachelys can be found in Tong and Buffetaut (1996).

Fig. 1

Hamadachelys escuilliei Tong and Buffetaut, 1996. Partially restored skull based on MDE T03 and AMNH 30644. A, dorsal; B, ventral; C, lateral. [C. Facella, del.]

Fig. 2

Hamadachelys escuilliei Tong and Buffetaut, 1996. Partially restored ventral view based on MDE T03 and AMNH 30644. [A. Phillips, del.]

Fig. 3

Hamadachelys escuilliei Tong and Buffetaut, 1996. MDE T03. A, dorsal; B, ventral; C, right lateral; D, anterior; E, left lateral; F, posterior. [A. Phillips, del.]

Fig. 4

Hamadachelys escuilliei Tong and Buffetaut, 1996. MDE T03. A, dorsal; B, ventral; C, right lateral; D, anterior; E, left lateral; F, posterior. [A. Phillips, del.]

Fig. 5

Hamadachelys escuilliei Tong and Buffetaut, 1996. AMNH 30644. A, dorsal; B, ventral; C, right lateral; D, anterior; E, left lateral; F, posterior. [C. Facella del.]

Fig. 6

Hamadachelys escuilliei Tong and Buffetaut, 1996. AMNH 30644. A, dorsal; B, ventral; C, right lateral; D, anterior; E, left lateral; F, posterior. [C. Facella del.]

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; Stereogenys Andrews, 1901; Caninemys Meylan, Gaffney, and Campos, 2009; Turkanemys Wood, 2003; Neochelys Bergounioux, 1954; Papoulemys Tong, 1998; Peiropemys, n. gen.; Lapparentemys, n. gen.; Pricemys, n. gen.; Mogharemys, n. gen.; Cordichelys, n. gen.; Latentemys, n. gen.; Brontochelys, n. gen.; Albertwoodemys, n. gen.; Lemurchelys, n. gen.

Incertae sedis within family: Stupendemys, Wood, 1976; Cerrejonemys Cadena et al., 2010; Roxochelys Price, 1953; Cambaremys, França and Langer, 2005; Kenyemys Wood, 1983.

Diagnosis

Member of the epifamily Podocnemidinura uniquely possessing a fully developed, medially extensive cavum pterygoidei with a completely developed pterygoid flange, 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, and exoccipital-quadrate contact absent, all in contrast to Brasilemys; palatine forming moderate amount of triturating surface in contrast to little or none as in Euraxemys; pectoral scales contact epiplastra.

Discussion

We are not aware of any contrary hypothesis to podocnemidid monophyly.

Subfamily Bauruemydinae, new

Diagnosis

Member of the Podocnemididae; quadratojugal-parietal contact short; temporal emargination extensive; cheek emargination does not reach above level of orbit; foramen jugulare posterius open laterally; tuberculum basioccipitale narrowly spaced; cervical centra not saddle shaped.

Discussion

This new subfamily is created to reflect the position of Bauruemys as sister taxon to all other Podocnemididae. See table 1 for comparison of the basal Podocnemididae described here.

TABLE 1

Comparison of Skulls of Mesozoic and Early Tertiary Podocnemididae

Included Taxa

Bauruemys.

Bauruemys Kischlat, 1994

Type Species

Podocnemis elegans Suárez, 1969a.

Included Species

Bauruemys elegans.

Distribution

Late Cretaceous of northern South America.

Revised Diagnosis

A podocnemidid known from the skull and postcrania; temporal emargination extensive, second neural consistently four sided, uniquely among the Podocnemididae; skull relatively wide and flat in contrast to Podocnemis, orbits facing dorsally rather than dorsolaterally as in Podocnemis; interorbital groove such as found in Podocnemis absent; postorbital large in contrast to Podocnemis; parietal-quadratojugal contact small; cheek emargination does not reach above level of orbit; medial expansion of triturating surface, median maxillary ridge, absent, wide concavity on the midline, formed by the premaxillae and anterior maxilla present; accessory ridge in triturating surface absent; vomer present; vomer-maxilla contact narrow; fossa precolumellaris deep as in Peiropemys; foramen jugulare posterius open laterally; basioccipital-opisthotic contact narrow in contrast to all other Podocnemididae in which it is wider; tuberculum basioccipitale not widely separated as in all other Podocnemididae; interparietal scale equilateral triangle; cavum pterygoidei with small anterior opening for foramen cavernosum; horizontal occipital shelf present as in Podocnemis; chorda tympani not enclosed in processus retroarticularis; cervical centra not saddle shaped. See table 1.

Small shell (less than 150 mm carapace length) differing from other South American Cretaceous and Early Tertiary Pelomedusoides in having distinct surface sculpture of concentric rings, nuchal bone width equals length; six neurals (rarely seven) extending to costal seven; second neural always four sided; axillary buttress extending onto second costal and reaching to third peripheral anteriorly; suture for axillary buttress broad medially and narrow laterally; second costal thickened to support axillary buttress; bridge peripherals guttered dorsally; D-shaped iliac scar that does not cross from the eighth onto the seventh costal; internal gutter of posterior peripherals and pygal absent; gular scales usually not restricted to epiplastra; intergular scale narrow in contrast to bothremydids; short contact of humeral scales on midline; pectoral scales contact entoplastron and epiplastra, but do not contact mesoplastra.

Bauruemys elegans (Suárez, 1969a)

Podocnemis elegans Suárez, 1969a

Bauruemys elegans (Suárez, 1969a) Kischlat, 1994

Type Specimen

The number of the type specimen of this species is in some ambiguity, but we conclude with the assistance of our colleagues, G. Oliveira and P.S.R. Romano, that the type is MCT 1492-R, an associated shell and skull. The original author did not designate a holotype by number but he figured the shell, carapace, and plastron. However, the skull that was figured (Suárez, 1969a: figs. 6, 7) does not belong to the type specimen; it is MCT 1753-R (figs. 11, 12), and should be considered a paratype based on the text of Suárez (1969a; P.S.R. Romano, personal commun.). Candeiro et al. (2006: 927) confusingly state “holotype of B. elegans is UFRGS 148 and MN-V 4487, a postcranium….” Oliveira and Romano (2007) say (in translation) that although these specimens are representative of the species, they have concluded, following an extensive examination of the collections involved, that neither specimen should be referred to the type material of B. elegans. We agree that the type of Podocnemis elegans Suárez, 1969a, is MCT 1492-R, a skull (figs. 9, 10) and shell with scapulocoracoids, femur, and humerus. This identification is based on the shell figures in Suárez (1969a: figs. 2–4; 1969b: figs. 1–5).

The type skull, MCT 1492-R (figs. 9, 10), is dorsoventrally crushed and much of the bone is broken. The right side is better preserved than the left in general. The prefrontals are present but broken anteriorly. The frontals are both present with clear sutures. Both parietals are broken laterally and posteriorly. Portions of the jugal where it contacts the maxilla are present on both sides, but its posterior contacts are damaged or missing. Only small parts of the quadratojugals are identifiable, and the postorbitals are either missing (left) or broken and overlapped with the jugal and quadratojugal (right). Only a small portion of the anterior margins of the left squamosal are preserved and none of the right. The cheek emargination is partially preserved on the right side but badly damaged, and it is not determinable on the left side.

The premaxillae are represented only by a part of the horizontal plate visible in dorsal view. The left maxilla has the portion below the orbit preserved with the labial ridge intact, but its anterior portion is missing and its posterior margins broken. The right maxilla is more complete, but is covered ventrally by matrix, which helps hold it together along with some loose bone fragments on the ventral surface. The vomer is missing. The left palatine is cracked and missing its anterior margins but is otherwise intact. The right palatine is badly broken and represented only by the posteromedial piece.

The right quadrate is preserved medially, but the cavum tympani and processus articularis are gone. The left quadrate has a partially crushed and broken cavum tympani and the processus articularis present. The medial contacts of both quadrates are preserved. The left pterygoid is nearly complete except for some cracking and slight displacement. The right pterygoid lacks most of its lateral portion and is partially obscured by displaced bone fragments. On both sides, the cavum pterygoidei is preserved, although its margins are broken edges. The anteromedial part of the supraoccipital is preserved, but all of the crista supraoccipitalis is gone. Both exoccipitals are present but slightly broken along the margin of the foramen magnum, and the condylus occipitalis is broken off. The basioccipital is complete except for the condylus occipitalis. Both prootics are present and complete but mostly covered by matrix. The left opisthotic lacks some of its posterior process and the right one lacks all of it. The basisphenoid is complete, with clear sutures.

To the extent that it is preserved, the type skull agrees in detail with the other skulls ascribed to this species that are described below.

Type Locality

“Tartaruguito” locality, near kilometer 736 of the old Sorocabana Railway branch, Pirapozinho, São Paulo State, Brazil, 22°13′08″S, 51°25′59″W (Romano and Azevedo, 2007; also see Bertini et al., 2006, for map; Oliveira and Romano, 2007). Suárez (2002), Henriques et al. (2005), Henriques (2006), Bertini et al. (2006), discuss the taphonomy and depositional history of this locality.

Horizon

Adamantina Fm. fide Suárez (2002), Bertini et al. (2006).

Diagnosis

As for genus.

Referred Material

MCT 1753-R, figured by Suárez (1969a: figs. 6, 7; 1969b: figs. 5, 6), but not associated with type specimen. This skull (described below, figs. 11, 12) is the best preserved of the Bauruemys we have seen; all bones, except the vomer, are known from either one side or the other. The horizontal plates of the premaxillae are cracked and a few small pieces are missing. The crista supraoccipitalis is cracked and broken and has been repaired by reconstructing the available pieces. The result shows all the limits of the crista, but it is slightly deformed by crushing. Both temporal margins are dorsoventrally crushed but not collapsed, only depressed from the level of the skull roof. A small portion of the posteromedial part of the left parietal is missing. The posterior end of the right squamosal is missing. On the ventral surface, the pterygoid flanges are broken away and the floor of the cavum pterygoidei has been removed by preparation to expose the internal features of the cavum pterygoidei.

DGM uncataloged (collected 1969, Campos and Silva), carapace and plastron disarticulated with postcranial material in carapace part, including an eighth cervical vertebra.

MCZ 4123, a more complete skull than DGM uncataloged (above), but it is slightly more distorted and deformed by crushing (figs. 13, 14). MCZ 4123 has the vomer intact and the pterygoid flanges preserved on both sides. It is the only Bauruemys we have seen to have the flanges intact. The fossa nasalis has been deformed by lateromedial crushing, but the elements are all present and the only area lost is part of the horizontal plates of the premaxillae. The right prefrontal is crushed and displaced onto the frontal. As in DGM uncataloged, the posterior parts of the temporal roof have been pushed ventrally below the level of the skull roof, but the bones are missing only a few small pieces and the temporal margins are intact. Some missing areas of the parietals, left cheek, and supraoccipital have been filled in or repaired with resin.

AMNH 7888 consists of the skull, attached lower jaws, and attached fragments, including a centrum and phalange, to the right of the crista supraoccipitalis. The skull is missing the vomer, both prefrontals, both frontals, anterior parts of the parietals, dorsal parts of the premaxillae, anterior parts of the maxillae, and some of the pterygoid flanges. Although the vomer is missing and the lower jaws are in place, the ventral surface of the premaxillae is well preserved and visible. The lower jaw is missing some pieces anteriorly.

AMNH 30643 consists of roughly the right half of a skull, showing the fossa orbitalis, the sulcus palatinopterygoideus, and the fossa temporalis. The right cheek, cavum tympani, crista supraoccipitalis, and temporal roof are missing. The occiput is damaged and it lacks the condylus occipitalis.

AMNH 30642 consists of roughly the left half of the skull without the snout region. The left otic chamber and cavum tympani are well preserved. The dorsal surface of the basicranium is exposed from the rostrum basisphenoidale anteriorly to the condylus occipitalis posteriorly, showing the area of the dorsum sellae and associated structures.

AMNH 30774, two partial lower right jaws and miscellaneous fragments associated with AMNH 30642 and AMNH 30643.

MN 4322-V, nearly complete skull, lacking the premaxillae, anterior parts of the maxillae, vomer, lateral parts of the pterygoids, and some of the occipital surfaces. The more internal areas are still encased in matrix.

MN 4323-V, partial skull, lacking the right quadrate, occipital surfaces, and much of the palate. It still has matrix on many of the surfaces.

MN 4324-V, posterior half of a skull, lacking everything anterior to the pterygoid-palatine suture. The center of the cavum cranii is exposed in the anterior surface of the break. The posterior temporal margin is preserved.

MN 4487-V, two cervical vertebrae referred to by Kischlat (1994: 348).

MCZ 4125, anterior half of a shell, with costals 1–4, neurals 1–4, anterior half of plastron with right mesoplastron. Only a few peripherals are preserved.

MCZ 4122, anterior margin of a shell, with the nuchal, peripherals 1–3, and the anterior plastral lobe.

MCZ 4127, right maxilla and palatine with ventral surfaces exposed.

MCZ 4128, partial skull lacking the prefrontals, frontals, postorbitals, and quadratojugals, left quadrate, on the dorsal surface. The posterior rami of the lower jaws are in place on the ventral surface. Most of the elements visible in ventral view are present, but are cracked and not well preserved.

MCZ 4129, partial skull consisting of the snout region exposed in ventral view and the otic chambers and cavum cranii exposed in a break.

MCZ 4124, small block of matrix with some limb elements in situ.

Previous Work

Although Suárez (1969a) appears to be the legitimate publication for the type announcement for Podocnemis elegans, there are two other possible publications listed by Oliveira and Romano (2007) that contain a type announcement. One of these, not listed in our bibliography but listed in Oliveira and Romano (2007: 127), is a short paper that we interpret as an unpublished abstract, not a legitimate taxonomic description. The other, Suárez (1969b), has the same text and figures as Suárez (1969a) with a different page and figure arrangement, but exactly the same content. We do not know the relative dates of publication.

In addition to the generic name proposed in Kischlat (1994), Kischlat (1996a, 1996b), and Kischlat et al. (1994) provide other hypotheses about the relationships and geologic distribution of Bauruemys. Romano and Azevedo (2007) present a morphometric study of this species, possible due to the relatively large numbers of specimens available. They support the hypothesis of a single species in the sample studied.

Discussion

Peirópolis B, the smaller of the two unnamed shell taxa from Peirópolis, is known only from isolated shell elements. It shares derived features with Bauruemys elegans, but is too poorly known to be referred to this species.

Kischlat (1994: 348) described two cervical vertebrae of Bauruemys elegans as lacking saddle-shaped or heterocoelous articulations (in the sense of Williams, 1950; Hoffstetter and Gasc, 1969). We have also examined these vertebrae and a number of similar cervicals in the uncataloged collections of the DNPM. The articulations are clearly similar to the cervical morphology of the outgroups, such as in Euraxemys and pelomedusids, in being a relatively simple, hemispherical, concave-convex articulation, not the more complex surface (“saddle-shaped”) seen in Podocnemis and Peltocephalus, and not the wide, surface seen in Turkanemys and Erymnochelys.

Subfamily Podocnemidinae Cope, 1868

New Diagnosis

Members of the Podocnemididae; quadratojugal-parietal contact long; foramen jugulare posterius closed; tuberculum basioccipitale widely spaced; crista supraoccipitalis with horizontal plate on ventral edge (except in Erymnochelys and Peltocephalus); cervical centra saddle shaped (except in Erymnochelys and Turkanemys, which have a third state).

Included Taxa

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; Stereogenys Andrews, 1901; Caninemys Meylan, Gaffney, and Campos, 2009; Turkanemys Wood, 2003; Neochelys Bergounioux, 1954; Papoulemys Tong, 1998; Peiropemys, n. gen.; Lapparentemys, n. gen.; Pricemys, n. gen.; Mogharemys, n. gen.; Cordichelys, n. gen.; Latentemys, n. gen.; Brontochelys, n. gen.; Albertwoodemys, n. gen.; Lemurchelys, n. gen.

Discussion

Two subfamilies of the Podocnemididae have been named by Broin (1988 [1989]) the “subfamily Podocnemidinae” and the “subfamily Erymnochelinae.” Although used in a number of additional papers (Broin, 1991; Lapparent de Broin, 2000b, 2001, 2003a, 2003b), the most complete expression of the hypothesis is Lapparent de Broin (2000b).

The “subfamily Podocnemidinae” was considered by Lapparent de Broin (2000b) as containing Podocnemis, Peltocephalus, Bauruemys, “aff. Roxochelys vilavilensis” ( =  Lapparentemys), and Stupendemys. This group of Lapparent de Broin (2000b) was not confirmed in our analysis. Besides the successive sister-group relationship of (Bauruemys (Lapparentemys (Podocnemis))), we have found a number of characters linking Peltocephalus with the taxon below, Lapparent de Broin's “Erymnochelyinae.” In our analysis, the “subfamily Podocnemidinae” of Lapparent de Broin (2000b) is simply the South American podocnemidids. The character that seems to be dominant in Lapparent de Broin (2000b, and other papers) for this taxon is the saddle-shaped cervicals, but biogeography may have been an influential phylogenetic character. There is a series of shell characters as well, but the distributions are not clear. Apparently, the fact that Bauruemys lacks the cervical character was not yet known and the author used the shell features and geography to place this species in this group.

The “subfamily Erymnochelinae” was considered by Lapparent de Broin (2000b) as containing Erymnochelys, Neochelys, Stereogenys, Shweboemys, Dacquemys, and “Carteremys” (the last genus a nomen dubium in Gaffney et al., 2006). This group is essentially the present authors' magnatribe Erymnochelydand. The reflection of the cladogram in the classification requires its change in rank from Lapparent de Broin (2000b), but we agree in most of its content and its principal character. The primary character used for this group (Lapparent de Broin, 2000b: 70) is the “Much eroded roof of the enlarged carotid canal, the prootic and quadrate being so much eroded that the floor of the canalis cavernosus is broken and this canal is anteriorly confluent with the part of ‘enlarged canal’ leading to the sulcus cavernosus (not known in Dacquemys, homoplastic but with a less eroded roof in the podocnemidine Peltocephalus).” The interesting fact that the character does actually occur in a member of Lapparent de Broin (2000b)'s other subfamily (i.e., Peltocephalus) is recognized by Lapparent de Broin (2000b), but the geographic distribution was perhaps a more important consideration. There are now a few cervicals showing that extinct members of Lapparent de Broin's “Erymnochelyinae” did have saddle-shaped centra.

When analyzing only the Recent genera, none of the molecular results reproduce the Lapparent de Broin (2000b) resolution of (Erymnochelys (Podocnemis, Peltocephalus)); rather they prefer the (Peltocephalus (Podocnemis, Erymnochelys)) arrangement, while the present paper agrees with neither and finds the (Podocnemis (Peltocephalus, Erymnochelys)) resolution to be most attractive (see Phylogenetic Analysis).

Gaffney and Meylan (1988) used “Podocnemidinae Williams, 1954179180 (as ‘podocnemides’)” as an equivalent to what we now call the family Podocnemididae. In the present paper we redefine the subfamily Podocnemidinae to be the monophyletic group consisting of all Podocnemididae except Bauruemys.

Infrafamily Peiropemydodda, new

Diagnosis

Members of the subfamily Podocnemidinae; cheek emargination reaches above level of orbit; characters also found in Bauruemys: medial expansion of triturating surface absent; accessory ridges absent from the triturating surface.

Distribution

Late Cretaceous of Brazil, Paleocene of Bolivia.

Included Taxa

Peiropemys, n. gen.; Lapparentemys, n. gen.; Pricemys, n. gen.

Discussion

This grouping of three genera that are outside the common ancestor of the living species in our analysis really rests only on the combination of the smaller cheek emargination, as the absence of the medial expansion of the triturating surfaces with accessory ridges is primitive for the family. One step away from the shortest cladogram makes this group a multichotomy with Podocnemis and the magnatribe Erymnochelydand.

Peiropemys, new genus

Type Species

Peiropemys mezzalirai, new genus and new species.

Included Species

Peiropemys mezzalirai.

Distribution

Late Cretaceous, Brazil.

Etymology

Peiropos, in allusion to the locality, Peirópolis; emys, Greek for “freshwater tortoise.”

Diagnosis

A podocnemidid known only from the skull; condylus occipitalis formed only by exoccipitals uniquely among Podocnemididae (except for some Peltocephalus); skull relatively high and narrow in contrast to Bauruemys; orbits facing dorsolaterally; interorbital groove as found in Podocnemis absent; temporal emargination slight, in contrast to Bauruemys but greater than in Peltocephalus; postorbital large in contrast to Podocnemis; parietal-quadratojugal contact long; cheek emargination reaches above level of orbit; medial expansion of triturating surface, median maxillary ridge, absent, wide concavity on the midline, formed by the premaxillae and anterior maxilla present; accessory ridge on triturating surface absent; vomer present; vomer-maxilla contact wide; fossa precolumellaris deep as in Bauruemys; foramen jugulare posterius closed; interparietal scale equilateral triangle in contrast to parallel sided in Pricemys; cavum pterygoidei with small anterior opening for foramen cavernosum; horizontal occipital shelf present as in Podocnemis.

Peiropemys mezzalirai, n. gen. et sp.

Type Specimen

MCT 1497-R (figs. 15–18). This skull is a well-preserved and nearly complete specimen, lacking the right maxilla-jugal area and most of the premaxillae. The only parts unknown from one side or the other is the labial ridge area of the premaxilla, which is missing on both sides, and the distal edges of the pterygoid flange. The cavum cranii has been cleaned out, and the skull is free of most matrix. No crushing or distortion is visible.

Type Locality

Caiera Quarry (of Campos and Kellner, 1999, and Kellner et al., 2005; same locality is “outcrop 1” of Carvalho et al., 2004, and Novas et al., 2008; “Point 1” of Ribeiro and Carvalho, 2007), at the southern end of the hill, Serra do Veadinha, near the old railroad station of Peirópolis, near the city of Uberaba (Municipality of Uberaba or Uberaba County), Minas Gerais State, Brazil. MCT 1497-R, type of Peiropemys mezzalirai, n. gen. et sp., was collected in 1959 by a team led by L.I. Price, DNPM, from the area excavated in that year as shown on the quarry map in Kellner et al. (2005: fig. 3).

Caiera Quarry, or Locality/Outcrop 1, has also yielded MCT 1498-R (the skull here named Pricemys caiera, n. gen. et sp.), MCT 1499-R (a large, nearly complete, shell of Peirópolis A), and many disarticulated turtle elements, here informally referred to as Peirópolis A and Peirópolis B.

According to Kellner et al. (2005), the original fossils from Caiera Quarry were found in 1947, with the main collecting efforts being in 1949, 1950, 1953, 1955, 1957, 1958, 1959, and 1961 (based on the published Caiera Quarry map in the DGM done under L.I. Price's direction, Kellner et al., 2005: fig. 3). Original discoveries apparently date back to 1945 when L.I. Price responded to reports of fossil discoveries made during the construction of a railway at Mangabeira, north of Uberaba (Ribeiro and Carvalho, 2007). During the time this quarry was in operation, a large number of fossils were collected, and over 185 square meters of the fossiliferous layer were exposed (Ribeiro and Carvalho, 2007: figs. 1, 2). Although the work at Caiera Quarry was apparently ended in 1961, there was later collecting around the base of the same hill, Serra do Veadinha, and in the same stratigraphic unit, Serra da Galga Member, Marília Fm., that held the initial quarry. D.A. Campos and L.I. Price in 1967 and 1969 collected turtles in this area, according to labeled specimens. Collections have continued to be made in this area and a paleontological museum has been built near the site (Ribeiro and Carvalho, 2007).

Other faunal elements from the Caiera Quarry include the prominent sauropods, but a frog, a lizard, crocodiles, and theropods, are also described or referenced in Carvalho et al. (2004), Kellner et al. (2005), Salgado and Carvalho (2008), Novas et al. (2008), and Candeiro (2009).

Horizon

Serra da Galga Member, Marília Fm. (see Candeiro, 2009, and references).

Diagnosis

As for genus.

Etymology

For Sérgio Mezzalira, a DNPM geologist and invertebrate paleontologist who was very active in exploration and study of the Bauru Group and São Paulo State geology.

Referred Material

None. However, it is possible that the shell taxon informally referred to here as Peirópolis B belongs to Peiropemys. Peirópolis B is the smallest of the three shell taxa from this locality and the one for which there is the most limited amount of material. The material is all disarticulated (figs. 92, 93, 95, 97), but some of it is associated. At present we have the nuchal, costals 1, 5, 7, and 8, peripherals 1 and 8–11, the suprapygal, and pygal. We cannot assign any plastron material to this taxon at this time, although, based on size and a different morphology than in Peirópolis A, we speculate that some of the smaller plastral elements may belong to this species. Cambaremys langertoni, found near the Caiera Quarry, is apparently a juvenile (França and Langer, 2005; Romano et al., 2009), and its small size may not be characteristic of the species. As far as we know at present, although the type and only known specimen of Cambaremys langertoni was found near the Caiera Quarry, only two taxa of turtles definitely occur in the quarry site. These are Peiropemys and Pricemys, based on skulls, and Peirópolis A and Peirópolis B, based on shells.

Previous Work

Possibly the material referred to in Broin (1991: 515).

Discussion

Peiropemys is the sister group to Pricemys + Lapparentemys in our analysis. These three genera are the sister taxon to all other podocnemidids, except Bauruemys, which is outside (Peiropemys (Pricemys, Lapparentemys)) and is the sister taxon to all podocnemidids. The morphology of these four genera and the Recent Podocnemis has “remained in all particulars in a pristine and unmodified condition (at least as regards their structure) through a grievous long period of time on Earth…” (Tedwelle in Gaffney, 1979: 69).

Lapparentemys, new genus

Synonymy

?Roxochelys vilavilensis Broin, 1971.

Type Species

?Roxochelys vilavilensis Broin, 1971.

Included Species

Lapparentemys vilavilensis (Broin, 1971).

Distribution

Paleocene of Bolivia.

Etymology

Restudy of the type and two additional specimens of Roxochelys wanderleyi indicate that the species vilavilensis Broin, 1971, is not referable to the genus Roxochelys. Thus, we erect a new generic name, Lapparentemys, in honor of the many important contributions of France de Lapparent de Broin to our understanding not only of pleurodires, but of all turtles.

Revised Diagnosis

A podocnemidid known from the skull and postcrania; pectoral scales contact mesoplastra uniquely among Podocnemididae; skull relatively high and narrow in contrast to Bauruemys; orbits facing dorsolaterally; interorbital groove as found in Podocnemis absent; temporal emargination slight in contrast to Bauruemys but deeper than in Peltocephalus; postorbital large in contrast to Podocnemis; parietal-quadratojugal contact long; cheek emargination reaches above level of orbit; medial expansion of triturating surface, median maxillary ridge, absent; wide concavity on the midline, formed by the premaxillae and anterior maxilla, present; accessory ridge in triturating surface absent; vomer present; vomer-maxilla contact wide; fossa precolumellaris shallow as in Podocnemis expansa; foramen jugulare posterius closed; interparietal scale equilateral triangle; cavum pterygoidei with small anterior opening for foramen cavernosum; horizontal occipital shelf present as in Podocnemis; chorda tympani not enclosed in processus retroarticularis; nuchal bone width greater than length; seven neurals extending to costal seven.

The shell of this medium-size Pelomedusoides pleurodire (to 400 mm straight carapace length) differs from other South American Cretaceous and early Tertiary Pelomedusoides in having a relatively thick, elongate shell with deep scute sulci but no distinct surface sculpture; vertebral scales are narrow and parallel sided; a relatively short, wide nuchal bone is narrow anteriorly; first neural four sided; axillary buttress not extending onto second costal, reaching to third peripheral anteriorly; suture for axillary buttress broad and curving with parallel sides; second costal not thickened to support axillary buttress; bridge peripherals weakly guttered dorsally; iliac scar with anterior margin concave and crossing from the eighth onto the seventh costal both medially and laterally; internal gutter of posterior peripherals and pygal absent; gular scales restricted to epiplastra; intergular narrow; short humeral contact on midline; pectoral scales contact epiplastron and entoplastron; pectoral scale contact to mesoplastron variable.

Lapparantemys vilavilensis (Broin, 1971), new combination

?Roxochelys vilavilensis Broin, 1971

Type Specimen

MNHNP VIL 1 (Broin, 1971: pl. 31, figs. 1, 2), a complete carapace and plastron.

Type Locality

2 kms south of Vilavila, later designated Villa Viscarra, Cochabamba Province, Bolivia.

Horizon

Originally (Broin, 1971: 445) described from the late Cretaceous El Molino Formation, the horizon was later (Muizon et al., 1983; Broin, 1991) identified as the early Paleocene Santa Lucía Fm.

Diagnosis

As for genus.

Etymology

Named for the type locality near Vilavila (Broin, 1971: 445).

Referred Material

Specimens from the type locality listed in the hypodigm (Broin, 1991) include: MNHNP VIL 2, partial shell; MNHNP VIL 3, plastron; MNHNP VIL 4, partial plastron; MNHNP VIL 5, juvenile shell; MHNC 6902, partial shell; MHNC 6903, plastron.

AMNH 14444, a well-preserved and nearly complete skull (figs. 21, 22), lacking the distal end of the crista supraoccipitalis and the edges of the pterygoid flange. The skull is slightly distorted by dorsolateral compression from the right side, ventromedial compression from the left side. The fossa temporalis on both sides is free of matrix, but the cavum cranii is not. The lower jaw is present and is complete except for the right articular region. A small portion of the carapace is associated with the skull and it agrees with the other Lapparentemys shells. Locality and Horizon: “El Molino, Vila vila” (label).

WUS 2160, a complete shell (fig. 90) with skull (figs. 23–25) lacking the right skull roof and cheek. Before preparation the skull was articulated to the shell by the cervical vertebrae. The left premaxilla lacks its medialmost portions and the right one is mostly missing. The right maxilla lacks its labial ridge and the area below the orbital margin. The pterygoid flanges are broken, the right one is missing, and the left one is nearly all present. Only the anteroventral portion of the crista supraoccipitalis remains, the rest is missing. The lower jaws are present and nearly complete; only part of the left prearticular is missing. Locality and horizon: Santa Lucía Fm., Cochabamba Department, Bolivia, purchased George Helsp Fossils, 1994 (label).

RM 20.5155, skull, shell (fig. 89), cervicals, and limb bones in partial articulation. Specimen poorly preserved due to small cracks and displacements throughout. Locality and horizon: “Santa-Lucia, Tiupampa (label),” collector, Pierre-Yves Gagnier.

Previous Work

Both Broin (1971) and (1991) provide important descriptions and figures of skulls and shells of this taxon. Numerous papers have referred to this taxon but provided no additional information.

Discussion

We refer two additional specimens to Roxochelys wanderleyi (following Romano et al., 2009; see discussion below) greatly expanding our knowledge of the shell. As a result of this redescription based on new shell material of R. wanderleyi, the type species of Roxochelys, it is apparent that “?Roxochelys vilavilensis” Broin, 1971, cannot be referred to Roxochelys. Therefore, we have created a new genus, Lapparentemys.

Lapparentemys and Pricemys are linked by the possession of a shallow fossa precolumellaris.

Our examination of MNHNP material identified by Broin (1971, 1991) as “?Roxochelys vilavilensis,” leads us to believe that there may be more than one taxon in the original hypodigm and subsequently assigned specimens. In order to avoid possible problems of overlapping species, we have restricted our concept of vilavilensis to the holotype and other specimens listed above. We consider MNHNP VIL-6 and MHNC 6904 to represent two possible, additional unnamed taxa that are excluded from our concept of Lapparentemys vilavilensis.

MNHN VIL-6 is a small round carapace with an associated plastron (Broin, 1991: pl. I, figs. 1–4). It has a very different overall shell shape than the type of vilavilensis or other specimens with skulls referable to vilavilensis. In our opinion the small size is not due to immaturity as all sutures are closed and the shell is well ossified. In MNHNP VIL-6 the pleuro-marginal sulci are on the costoperipheral sutures, which is not the case in the type of vilavilensis or other specimens with skulls that are referable to vilavilensis. The epiplastron differs in having a long straight anterior margin and a more transverse posterior margin. Where the posterior margin of the epiplastron articulates with the hyoplastron it is straight and without a major posterior projection that can be seen in MNHNP VIL-3, which was collected with the vilavilensis type specimen. Furthermore, on the ventral surface, the pectoral scales reach the epiplastral suture, but do not cross onto the epiplastron as in vilavilensis. The vertebral scales in MNHNP VIL-6 are shorter and somewhat more hexagonal than in Lapparentemys vilavilensis.

MNHC 6904 is a specimen from Tiupampa (Broin, 1991: pl. II, figs. 8–11). It has a longer, narrower shell more similar to vilavilensis in overall form, but differs from vilavilensis in having a robust axillary buttress located on the posterior part of the first costal and producing some thickening of the second costal. The epiplastron has the same basic shape as vilavilensis but the pectoral scale appears to not reach the epiplastron as it does in material of vilavilensis. Unlike vilavilensis, the vertebrals of this specimen are more hexagonal. Resolution of the phylogenetic position of MNHN VIL-6 and MNHC 6904 is beyond the scope of this study. They are discussed to stress that they are not used in our concept of Lapparentemys vilavilensis.

Pricemys, new genus

Type Species

Pricemys caiera, new genus and species.

Included Species

Pricemys caiera, new genus and species.

Distribution

Late Cretaceous, Brazil.

Etymology

For Llewellyn I. Price, influential Brazilian paleontologist and leader of the team that collected the type specimen and the other Caiera Quarry, Peirópolis, turtles.

Diagnosis

A podocnemidid known only from the skull; skull relatively high and narrow in contrast to Bauruemys; interorbital groove as found in Podocnemis absent; temporal emargination slight in contrast to Bauruemys but deeper than in Peltocephalus; postorbital large in contrast to Podocnemis; parietal-quadratojugal contact long; cheek emargination reaches above level of orbit; medial expansion of triturating surface, median maxillary ridge, absent; wide concavity on the midline, formed by the premaxillae and anterior maxilla, present; accessory ridge in triturating surface absent; vomer present; fossa precolumellaris shallow as in Podocnemis expansa; foramen jugulare posterius closed; interparietal scale parallel sided in contrast to equilateral triangle in Peiropemys; cavum pterygoidei with small anterior opening for foramen cavernosum; horizontal occipital shelf present as in Podocnemis.

Pricemys caiera, n. gen. et sp.

Type Specimen

MCT 1498-R, a braincase (figs. 28, 29), left quadrate-squamosal (figs. 32, 33), and right maxilla-jugal (figs. 30, 31). The quadrate piece articulates with the braincase but the maxilla piece has no contacts with the rest of the skull. The braincase consists of the right quadrate, the medial part of the right squamosal, both prootics, both opisthotics (the left one is incomplete posteriorly), the supraoccipital, and most of both parietals. The lateral edges of both parietals, however, are incomplete. The right one is separated from the quadrate by matrix, but the left one shows the posterior part of the quadratojugal contact when articulated with the left quadratojugal-squamosal piece. MCT 1498-R preserves both exoccipitals, the basioccipital, and the basisphenoid. The pterygoids are present but lacking their anterolateral areas, the processus trochlearis pterygoidei and pterygoid flange in particular, but also the palatine contact. This specimen has been prepared with only the matrix supporting the remaining portions of skull roof. The internal structures of the cavum cranii, cavum acustico-jugulare, cavum acusticum, and cavum pterygoidei, are all visible and well preserved. The left stapes is preserved in situ.

The left quadrate-squamosal piece (figs. 32, 33) of MCT 1498-R preserves nearly all of the squamosal. The quadrate is complete in the area of the cavum tympani, but its medial contacts with the braincase piece have been lost. The skull appears to have been slightly damaged by dorsoventral compression, although not deformed just broken along the parietal and quadrate contact areas. A fragment of stapes remains in the incisura columellae auris.

The right maxilla-jugal piece of MCT 1498-R (figs. 30, 31) preserves the maxilla from the premaxilla contact to the jugal. The labial ridge is broken except for a short section under the orbit, which gives the depth of the ridge. The dorsal process of the maxilla appears to end at the prefrontal contact. The lateral section of the palatine is also preserved where it forms part of the triturating surface and enters the floor of the fossa orbitalis. The jugal is preserved anteriorly, but its dorsal and posterior margins are broken edges. The anteroventral margin is natural and shows the depth of the cheek emargination.

Type Locality

Caiera Quarry, near village of Peirópolis, near city of Uberaba, Minas Gerais State, Brazil. Collected in 1958 from base of Serra do Veadinha that yielded Peiropemys and MCT 1499-R, a large, nearly complete, shell in DNPM. See locality discussion under Peiropemys mezzalirai, n. gen. et sp., for more information.

Horizon

Serra da Galga Member, Marília Fm. (see Candeiro, 2009, and references).

Diagnosis

As for genus.

Etymology

For locality of Caiera, near village of Peirópolis, near city of Uberaba, Minas Gerais, Brazil.

Referred Material

It is likely that an uncataloged DGM dentary (fig. 34) belongs to this species, as it fits well into, and is the same size as, the MCT 1498-R maxilla-jugal piece.

It is possible that the shell taxon referred to here as Peirópolis A (see Shell Morphology below) belongs to Pricemys. Peirópolis A is represented by a large number of disarticulated shell bones (figs. 91–93, 95–97), mostly complete single bones but not associated or articulated with others, collected mostly during the DNPM expeditions of 1949, 1950, 1953, 1955, 1957, 1958, 1959, and 1961. Included with these are three articulated specimens: (a) DGM Peirópolis 321, a complete plastron, at least some carapace, and some cervical parts, all disarticulated. The plastron of 321 is in a drawer labeled: “Desmonte 1967 Peirópolis Mun. Uberaba”; (b) MCT 1499-R (fig. 91B), a large shell with most of the carapace, the plastron, pelvis, and first thoracic present. From Peirópolis, Caiera Quarry, locality #1, the classic Peirópolis locality at lime plant, Bauru Fm., it is prepared with the plastron side up, the carapace is unprepared, and some of the anterior lobe is separate; (c) another Peirópolis shell, collected by Langerton, has a thickened dorsal lip of the epiplastra, which appears to be characteristic of Peirópolis A.

Previous Work

Probably referred to in Broin (1991: 515).

Discussion

Our resolution of this species is (Peiropemys (Pricemys, Lapparentemys)). It differs from its contemporary in the Caiera Quarry, Peiropemys, in a series of characters listed in table 2.

TABLE 2

Comparison of Peirópolis Skulls (measurements in mm)

TABLE 3

Species of the Tribe Stereogenyini

Infrafamily Podocnemidodda Cope, 1868, new rank

Diagnosis

Members of the subfamily Podocnemidinae; medial expansion of triturating surface with median maxillary ridge; accessory ridges present.

Included Taxa

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; Stereogenys Andrews, 1901; Caninemys Meylan, Gaffney, and Campos, 2009; Turkanemys Wood, 2003; Neochelys Bergounioux, 1954; Papoulemys Tong, 1998; Mogharemys, n. gen.; Cordichelys, n. gen.; Latentemys, n. gen.; Brontochelys, n. gen.; Albertwoodemys, n. gen.; Lemurchelys, n. gen.

Magnatribe Podocnemidand Cope, 1868, new rank

Diagnosis

Same as for Podocnemis.

Included Taxa

Podocnemis Wagler, 1830.

Discussion

The category magnatribe is created here to reflect the cladogram. We use the Latinized ending -and.

Podocnemis Wagler, 1830

Type Species

Emys expansa Schweigger (1812).

Included Species

Podocnemis expansa, P. vogli, P. unifilis, P. erythrocephala, P. lewyana, P. sextuberculata, P. bassleri.

Distribution

Northern South America.

Diagnosis

A podocnemidid with postorbital small allowing jugal-parietal contact; jugal unusually large; interorbital groove present, uniquely among Podocnemididae; skull relatively high and narrow in contrast to Bauruemys; orbits facing dorsolaterally; temporal emargination slight in contrast to Bauruemys but deeper than in Peltocephalus; parietal-quadratojugal contact long; cheek emargination reaches level of orbit; medial expansion of triturating surface, median maxillary ridge, present; accessory ridge or ridges present in triturating surface; vomer absent (except in P. vogli); fossa precolumellaris varying from shallow to moderate but not as deep as in Bauruemys; foramen jugulare posterius closed; interparietal scale equilateral triangle or elongate triangle; cavum pterygoidei with small to moderate anterior opening for foramen cavernosum; horizontal occipital shelf present as in Peiropemys; chorda tympani not enclosed in processus retroarticularis.

Postcrania with cervical centra saddle shaped; nuchal bone width greater than length; seven neurals extending to costal eight; axillary musk duct absent from buttress (except in P. erythrocephala), axillary buttress reaching second peripheral (except in P. vogli), pectoral scales do not contact mesoplastra, but do contact entoplastron and epiplastra.

Previous Work

Baur (1890) gave a comparative diagnosis for Podocnemis, Peltocephalus, and Erymnochelys (repeated under Erymnochelys discussion), when he authored Erymnochelys. Boulenger (1889) gives a diagnosis for Podocnemis sensu lato, synonymizing Peltocephalus and Erymnochelys, for the Recent species of the family as currently constituted. The same is true for Siebenrock (1902), Williams (1954a), and Wermuth and Mertens (1961, 1977). The Pritchard and Trebbau (1984) diagnosis was published after the separation of the three genera was widely accepted. All of these emphasize characters visible in the preserved whole animal with limited if any osteologic characters. Although there is no systematic information, Groombridge (1982) has references for biology of the various living species. Fritz and Havas (2007) and Bickham et al. (2007) are the most recent to produce species lists and synonymies of variable usefulness.

Discussion

We follow Iverson (1992) who, in turn, followed Williams (1954a) and Pritchard and Trebbau (1984), in the recognition of six living species of Podocnemis sensu stricto. See all three references for an introduction to the literature on this genus. As our present work is not primarily concerned with these species, we do not review this literature. The specimens listed below as Referred Material are only a few of the ones we have actually seen and do not reflect the material available in collections or referred to in the literature.

Although there have been few suggestions of relationships among the living Podocnemis species, Mittermeier and Wilson (1974) argued that erythrocephala, unifilis, lewyana, vogli, form a monophyletic group within Podocnemis. They called this the “vomerine group” (ibid.: 157), based on the possession of at least two parallel ridges on the triturating surface (three in vogli, which they interpreted as primitive) and elongate heads. Frair et al. (1978) produced a resolution of: (P. expansa, P. vogli, P. lewyana, P. unifilis, P. erythrocephala (P. sextuberculata)), based on a serological analysis. Cadena et al. (2010) did not resolve these species.

Our resolution of these species is: (P. vogli (P. lewyana (P. unifilis (P. erythrocephala (P. sextuberculata, P. expansa))))). There is virtually nothing in common with this resolution and the molecular study of Vargas-Ramírez et al. (2008). Most of the characters we have used that actually resolve these species are frequently subject to homoplasy in other turtle taxa, and are reversed within our own study, so we do not consider this part of the analysis to be strongly supported as the resolution collapses into a multichotomy at one step more than the shortest resolution (fig. 98). Osteologically, these Recent species are very similar to each other and have relatively conservative morphologies even within a group that has changed little since the late Cretaceous.

In our resolution of the character list presented here, the following characters are unique synapomorphies for Podocnemis: interorbital groove, jugal-parietal contact, postorbital smaller than orbit (except in P. unifilis), and foramen nervi abducentis large (also in Pricemys, and not determinable in many fossil skulls).

For the living species of Podocnemis, our analysis produces the following characters (see also Character List under Conclusions) supporting the resolution in fig. 98:

(P. expansa, P. sextuberculata) – character 6, parietal and pterygoid contact in septum orbitotemporale present and narrow; character 30, fossa precolumellaris present but shallow.

((P. expansa, P. sextuberculata) P. erythrocephala) – character 36, anterior opening of cavum pterygoidei moderate in size.

(((P. expansa, P. sextuberculata) P. erythrocephala) P. unifilis) – character 70, axillary musk duct with three openings.

((((P. expansa, P. sextuberculata) P. erythrocephala) P. unifilis) P. lewyana) – character 8, interparietal scale elongate triangle; character 22, vomer absent; character 68, axillary buttress reaches peripheral two.

(((((P. expansa, P. sextuberculata) P. erythrocephala) P. unifilis) P. lewyana) P. vogli) – characters for genus Podocnemis listed above.

Podocnemis expansa (Schweigger, 1812)

Synonymy

See Boulenger (1889), Siebenrock (1902, 1904), Wermuth and Mertens (1961, 1977), Pritchard and Trebbau (1984), and Iverson (1992).

Type Specimen and Locality

See Iverson (1992).

Diagnosis

See Boulenger (1889), Siebenrock (1902), Williams (1954a), Pritchard and Trebbau (1984).

Referred Material

YPM Herpetology 15401, AMNH Herpetology 58098, AMNH Herpetology 62947, NMV 1852, USNM 222470, YPM Herpetology 204, FMNH Herpetology 98958 (49), USNM Herpetology 65112, USNM Herpetology 65113, USNM Herpetology 28976, USNM Herpetology 28975.

Previous Work

Siebenrock (1902) and Williams (1954a) provide useful information. Groombridge (1982) and Valenzuela (2001) have general literature reviews. Williams (1950; redrawn in Hoffstetter and Gasc, 1969) figures the cervical vertebrae. Williams (1954a) figures the palate. Good skull figures are in Grey (1855, repeated in Gaffney, 1979) and Hay (1908). Gaffney (1990) figures most of the postcranial osteology except the shell. Pritchard and Trebbau (1984) figure skull and shell with sutures.

Discussion

Although relatively common in collections and referred to in many systematic works, there is no rigorous, comparative morphologic diagnosis for this species (or the genus Podocnemis for that matter). Among the references indicated, Williams (1954a) and Pritchard and Trebbau (1984) do provide descriptive information particularly useful in preserved specimens but few comparative osteologic characters.

Podocnemis vogli Müller, 1935

Synonymy

See Wermuth and Mertens (1961, 1977), Pritchard and Trebbau (1984), Iverson (1992).

Type Specimen and Locality

See Iverson (1992).

Diagnosis

See Williams (1954a) and Pritchard and Trebbau (1984).

Referred Material

UF 39100, MCZ 53626, USNM Herpetology 266206.

Previous Work

Müller (1935) has figures of skulls in dorsal and ventral views, shells figured without sutures. Pardo (1969) has figures of cranial as well as shell scales. Williams (1954a) figures the palate. Pritchard and Trebbau (1984) figure the skull and shell with sutures, and review the literature. Groombridge (1982) has a general review of the literature. We have included new figures of the skull of Podocnemis vogli for comparison (figs. 35, 36).

Podocnemis lewyana Duméril, 1852

Synonymy

See Siebenrock (1902), Boulenger (1889), Wermuth and Mertens (1961, 1977), Pritchard and Trebbau (1984: table 3), Iverson (1992).

Type Specimen and Locality

See Iverson (1992).

Diagnosis

Boulenger (1889), Siebenrock (1902) and Williams (1954a) have some comparative information, but there is no thorough comparative diagnosis for this species.

Referred Material

FMNH 73790, MCZ 53281, DGM 283RR, USNM Herpetology 108580.

Previous Work

Siebenrock (1902) and Williams (1954a) provide useful information. Williams (1954a) figures the palate. Groombridge (1982) has a general literature review.

Podocnemis sextuberculata Cornalia, 1849

Synonymy

See Boulenger (1889), Siebenrock (1902), Wermuth and Mertens (1961, 1977), Pritchard and Trebbau (1984: table 3), Iverson (1992).

Type Specimen and Locality

See Iverson (1992).

Diagnosis

Siebenrock (1902) and Williams (1954a) have some comparative information, but there is no thorough comparative diagnosis for this species.

Referred Material

USNM 65109, USNM Herpetology 65110, USNM Herpetology 65111, USNM Herpetology 51979, AMNH Herpetology 58108, FMNH Herpetology 16075.

Previous Work

Boulenger (1889), Siebenrock (1902) and Williams (1954a) provide useful information. Williams (1954a) figures the palate. Groombridge (1982) has a general literature review. We have included new figures of the skull of Podocnemis sextuberculata for comparison (figs. 37, 38).

Podocnemis unifilis Troschel, 1848

Synonymy

See Boulenger (1889), Siebenrock (1902, 1904), Wermuth and Mertens (1961, 1977), Pritchard and Trebbau (1984), Iverson (1992), Mittermeier and Wilson (1974).

Type Specimen and Locality

See Iverson (1992).

Diagnosis

See Siebenrock (1902), Williams (1954a) and Pritchard and Trebbau (1984).

Referred Material

AMNH Herpetology 58200, NMV 1824, AMNH Herpetology 58104, AMNH Herpetology 58198, FMNH Herpetology 98865, FMNH Herpetology 98926, FMNH Herpetology 45657, FMNH Herpetology 211758, USNM Herpetology 313861.

Previous Work

Boulenger (1889), Siebenrock (1902, 1904), Williams (1954a) provide useful information. Siebenrock (1904) has descriptions of the whole animal with figures; Pritchard and Trebbau (1984) figure skull and shell with sutures. Williams (1954a) figures the palate. Groombridge (1982) has a general literature review.

Discussion

Although relatively common in collections and referred to in many systematic works, there is no rigorous, comparative morphologic diagnosis for this species. We have included new figures of the skull of Podocnemis unifilis for comparison (figs. 39, 40). Among the references indicated, Siebenrock (1902), Williams (1954a), Pritchard and Trebbau (1984), do provide descriptive information particularly useful in preserved specimens but few comparative osteologic characters.

Podocnemis erythrocephala (Spix, 1824)

Synonymy

Authors before 1974 used P. cayennensis (Schweigger, 1812) for this species. See Wermuth and Mertens (1961, 1977), Mittermeier and Wilson (1974), Pritchard and Trebbau (1984), Iverson (1992), Bickham et al. (2007; however, usage of cayennensis was not “brief,” ibid.: 187) for further discussion and references.

Type Specimen and Locality

See Mittermeier and Wilson (1974), Iverson (1992).

Diagnosis

See Siebenrock (1902), Mittermeier and Wilson (1974), Pritchard and Trebbau (1984).

Referred Material

USNM 135474, AMNH 111069, DGM 45RR.

Previous Work

Earlier authors (e.g., Boulenger, 1889; Siebenrock, 1902; Williams, 1954a) did not recognize this species, which is involved in the “cayennensis” type question discussed in Mittermeier and Wilson (1974), see also Pritchard and Trebbau (1984) and Iverson (1992). Mittermeier and Wilson (1974) figure and describe skull and shell. Groombridge (1982) has general references.

Podocnemis bassleri Williams, 1956

Type Specimen

AMNH 1662, a nearly complete skull figured by Williams (1956: figs. 1, 3–5).

Type Locality

Rio Aguaytia, eastern Peru, about 8°10′S, 75°15′W fide Williams (1956).

Horizon

Tertiary, probably Late Miocene, Contamana Group fide Williams (1956). See Gaffney et al. (1998) and Meylan et al. (2009) for further comments and references for Contamana Group chronology.

Diagnosis

“Close to Podocnemis expansa, differing only in the apparently larger size, in the relationship of the internal palatal processes of the maxillae (which project abruptly dorsally and do not lie extended anteroposteriorly as in the Recent species), and in the slightly shorter relative distance from snout tip to posterior borders of the orbits” (Williams, 1956: 2).

Previous Work

Only Williams (1956) has dealt with this species. Wood and Diaz de Gamero (1971) state that this species is either ancestral to or closely related to Podocnemis expansa.

Discussion

The nearly indistinguishable skulls of P. bassleri and P. expansa suggest that the age of the fossil was Williams' main criterion for differentiating the taxa. It could be synonymized with P. expansa. In any case, this skull is the oldest definite record for the Podocnemis clade, here termed the magnatribe Podocnemidand, as “P.” negrii (Carvalho et al., 2002) does not preserve any of the diagnostic characters of the genus Podocnemis and is placed as Podocnemididae incertae sedis (see below), and the specimen of Cerrajonemys is too poorly preserved to be sure of its diagnostic characters, also Podocnemididae incertae sedis (see below).

Magnatribe Erymnochelydand Broin, 1988, new rank

Diagnosis

Members of infrafamily Podocnemidodda; cheek emargination slight to absent in contrast to Podocnemis (reversed in Cordichelys and Bairdemys); fossa precolumellaris shallow to absent, not deep (except in Erymnochelys).

Included Taxa

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; Stereogenys Andrews, 1901; Caninemys Meylan, Gaffney, and Campos, 2009; Turkanemys Wood, 2003; Neochelys Bergounioux, 1954; Papoulemys, Tong, 1998; Mogharemys, n. gen.; Cordichelys, n. gen.; Latentemys, n. gen.; Brontochelys, n. gen.; Albertwoodemys, n. gen.; Lemurchelys, n. gen.

Discussion

Rather than completely reflect the cladogram within this group (see fig. 98) in the classification with new categories and taxa, we only list the paraphyletic taxa outside the tribe Stereogenyini.

Caninemys Meylan, Gaffney, and Campos, 2009

Type Species

Caninemys tridentata Meylan, Gaffney, and Campos, 2009.

Included Species

Caninemys tridentata.

Distribution

Late Tertiary, Miocene, of Acre, Brazil.

Etymology

Named for the bulldog appearance of the skull and the large maxillary processes in the position of mammalian canines.

Diagnosis

A podocnemidid pleurodire with a well-developed processus trochlearis pterygoidei, quadrate-basioccipital contact, and a large cavum pterygoidei; unique among podocnemidids (and all other turtles) in having greatly inflated maxillae, each with a ventral, toothlike process that, together with a single process formed on the midline of the premaxillae, form a tridentate condition in the upper triturating surfaces, unique among pleurodires. The entire animal was probably smaller than Stupendemys geographicus Wood.

Caninemys tridentata Meylan, Gaffney, and Campos, 2009

Type Specimen

MCT 1496-R, a nearly complete skull collected by L.I. Price in 1962.

Type Locality

Locality 28 of L.I. Price, Volta de Pedra Pintada, upper Rio Jurua, Acre, Brazil.

Horizon

Late Miocene (Meylan et al., 2009).

Diagnosis

Same as for the genus.

Etymology

The species epithet is based on the tridentate appearance of the skull that is most clearly seen in anterior view.

Previous Work

Meylan et al., 2009.

Discussion

Although this taxon cannot be differentiated from the shell-based Stupendemys geographicus Wood, 1976, because there is no overlap in presently known morphology, it is likely that Caninemys is significantly smaller than the Venezuelan Stupendemys. Using skull-shell ratios of recent specimens of Podocnemis expansa and other recent podocnemidid species, it is hypothesized that the shell of Caninemys would be less than 4–5 ft in length rather than the 7+ ft length of Stupendemys. For comparison, the largest skulls of recent Podocnemis expansa have a condylobasal length of about 12 cm (Williams, 1956) and the Mio-Pliocene Podocnemis bassleri (Williams, 1956; very similar to P. expansa in morphology) is 15.7 cm in length, compared to about 16.5–17.0 cm for Caninemys. It is possible that smaller species of Stupendemys were present in the Acre region and that Caninemys is the skull of one of these, but that is only speculation (Meylan et al., 2009).

Dacquemys Williams, 1954b

Type Species

Dacquemys paleomorpha Williams, 1954b.

Included Species

Dacquemys paleomorpha.

Distribution

Late Eocene to early oligocene of Egypt.

Diagnosis

A genus of podocnemidid pleurodire known only from the skull; characterized by the unique possession of a fully roofed temporal region with a posteriorly extensive parietal and wide supraoccipital completely covering the otic chamber in dorsal view; broad parietal-squamosal contact not seen in other podocnemidids, except the unnamed UCMP 42008; prefrontal extends to anterior margin of apertura narium externa as in Peltocephalus but in contrast to Podocnemis, interorbital distance very wide with orbits facing laterally in contrast to living Podocnemididae; maxillae meet broadly behind premaxillae; premaxillae recessed and not visible laterally in contrast to all other Pelomedusoides; vomer absent; anterior part of triturating surface more extensive than in living Podocnemididae but secondary palate as seen in Shweboemys group absent; two accessory maxillary ridges meet anteriorly to form enclosed trough unique to this genus; antrum postoticum small as in Podocnemis; precolumellar fossa shallow as in Podocnemis expansa; pterygoid-jugal contact absent.

Dacquemys paleomorpha Williams, 1954b

Type Specimen

SMNS 12645, a nearly complete skull without lower jaws, lacking parts of the left side and the central basicranial region (Dacqué, 1912: pl. 2, figs. 6–8; Gaffney, 1979: fig. 128; Williams, 1954b: pl. 1; Gaffney et al., 2002: figs. 2 and 3).

Type Locality and Horizon

Dacqué (1912) gave the locality at first in reference to a shell: “Unteroligocäne Fluviomarinestuffe nördlich der Birket Qerun; nordlich non Tamieh; ostlich vom Schweinforthplateau. Schädel: aus derselben stufe bei Dimêh. Alles Fayum” (ibid., p. 310). Williams (1954b, repeated in Gaffney, 1979) has the locality as “Diieh” which seems to be a misspelling of “Dimêh.” The reference to the skull “from the same sediment near Dimêh” presumably means the Qasr el-Sagha or Birket Qarun Formations of late Eocene age.

Referred Material

DPC 5986 (Gaffney et al., 2002: figs. 4, 5), partial skull lacking anterior part of palate and some of left side. Field No. 86-292, collected by Alex van Nievelt. B-4 (lower Jebel Qatrani Fm., 1/4 mile southeast of A), AMNH quarry B area, early Oligocene (Seiffert, 2006).

Possibly BMNH R3346, a lower jaw figured by Williams (1954b; previously identified by Andrews, 1906, as Podocnemis fajumensis) that fits the triturating surface of the maxilla in Dacquemys, as argued by Williams.

Previous Work

Williams (1954b). Gaffney et al. (2002) provides a redescription based on two specimens and a reconstruction of the skull.

Discussion

There is no support for the earlier suggestions that the shell for Dacquemys paleomorpha is the shell-based species, Podocnemis fajumensis Andrews, 1903 (Williams, 1954b; Lapparent Broin, 2000a). We argue below that fajumensis is the shell of a Neochelys species. We have concluded that based on skull morphology, Dacquemys is the sister taxon to UCMP 42008, which has a very unusual shell similar to that of Albertwoodemys. The possibility exists that Dacquemys is the skull for Albertwoodemys, as both are known from the early Oligocene of the Fayum.

Albertwoodemys, new genus

Albertemys nomina nuda Auffenberg, 1981.

Type Species

Albertwoodemys testudinum, new species.

Distribution

Oligocene of Egypt.

Etymology

For Albert E. Wood, father of one of the authors, Roger C. Wood, and a founding member of the Society of Vertebrate Paleontology, as well as a past president, honorary member, and recipient of the Romer-Simpson medal of that society.

Diagnosis

A podocnemidid pleurodire known only from the shell with these unique features: single pair of scales on plastron between gular scales and abdominal scales (could be interpreted as absence of pectoral or humeral scale or fusion of the two); very small anal scales not meeting in midline accompanied by deep, C-shaped anal notch; scale sulci lying on raised ridges rather than incised on bone surface; large raised, swellings on lateral edges of posterior plastral lobe; entire shell more highly domed than any other podocnemidid; other distinguishing features: arcuate ischial contacts, with convex sides facing midline; ridge on visceral surface of posterior peripherals; large intergular scale separating small gular scales that do not reach entoplastron.

Discussion

The specimen here named Albertwoodemys testudinum was described and named in an unpublished Ph.D. dissertation by R. Wood in 1971. Wood coined the name “Albertemys” for this specimen, but the dissertation and the name were never published by Wood. Auffenberg (1981), in his description of Olduvai fossil turtles, had access to Wood's dissertation and compared his new taxon, Latisternon, to a number of pleurodires, including “Albertemys Wood, 1970” (1981: 512, although the dissertation date is actually 1971). It is apparent in the bibliography of Auffenberg (1981: 522) “Unpubl. PhD Diss….” that the author was aware of the fact that the dissertation was not an actual publication. A species name is not mentioned but there are characters for “Albertemys” indicated in the comparison with Latisternon. As it lacks a type species and type specimen, the name “Albertemys” is a nomen nudum.

Albertwoodemys testudinum, n. gen. et sp.

Type Specimen

AMNH 5088 (figs. 83, 84), plastron with articulated left lateral and posterior peripherals.

Type Locality

One mile west of AMNH quarries in the Lower Fossil Wood zone (Simons, 1968: 21), Fayum Depression, Egypt.

Horizon

Jebel Qatrani Fm., early Oligocene (Seiffert, 2006).

Diagnosis

As for genus.

Etymology

In reference to the high-domed, tortoiselike nature of the shell.

Referred Material

None, but a shell, KNM-WK17139, from the Miocene of Kalodirr, Kenya, is very similar and may be the same species or genus.

Previous Work

Described in Wood (1971).

Discussion

It may seem the epitome of self-indulgent hypocrisy to name another new shell taxon, particularly after thoroughly criticizing those who have done so (Gaffney et al., 2006). However, we have a good reason (doesn't everyone?) for breaking our own rules. Albertwoodemys testudinum is both diagnosable and subject to phylogenetic analysis, just not right now. The key to the relationships of Albertwoodemys lies in the skull of UCMP 42008 (currently under study by Bramble, Hutchison, and Gaffney), which has a shell with a number of synapomorphies in common with Albertwoodemys. The skull, however, is not yet named and described, although we have enough information to place it in the cladogram as the sister taxon to Dacquemys (fig. 98). Albertwoodemys is probably the sister taxon to UCMP 42008 or it could be the shell of Dacquemys itself (although Dacquemys is Eocene, it is very late Eocene and Albertwoodemys is very early Oligocene), in any case, it is clearly closely related to UCMP 42008 and Dacquemys. So we name the plastron with the presumption that description of the skull of UCMP 42008 will clarify its relationships.

Description

There are several unique features on the plastron (figs. 83, 84). The humeral and pectoral scales are either fused or one or the other of these pairs of scales is totally missing. The anal scales are so small and posteriorly placed that they make no contact at all along the midline. On the inner surface of the posterior lobe of the plastron parallel to and just inside the lateral borders are bulbous ridges of unknown significance. All of these features are unique in pleurodires.

Excellent preservation characterizes the only known specimen of Albertwoodemys. The anterior lobe of the plastron is relatively short and rounded; the posterior lobe is somewhere between 1.5 and 2 × as long as the anterior one. The bridge is unusually long, being nearly twice as long as the posterior lobe. The lateral borders of the posterior lobe converge only slightly toward the midline as they extend posteriorly. The tips of the xiphiplastra are sharply pointed and separated from each other by a deeply incised U-shaped notch. Based on UCMP 42008 and specimens associated with it, a caudal buckler or osteoderm apparently fits here. The ventral surface of the plastron is a convex bulge curving gently from anterior to posterior. On either side of the midline in the area of the hypo-xiphiplastral sutures are localized oval depressions reversing this trend. A slight transverse concavity is present between the bridges at right angles from the midline. The plastron measures 43.1 cm in length.

In anteroposterior cross section (fig. 84B, C) the epiplastra are wedge shaped, so that the greatest thickness occurs posteriorly at the junction between the epiplastra and the entoplastron. This thickening, however, is not uniform, and the resultant structure is not truly comparable to the epiplastral thickening and excavation present in nearly all Testudinidae. Instead, between the laterally situated ridges there is a concavity on the midline suture anteriorly (figs. 83, 84). The entoplastron affords a striking example of how the shape of a bone may vary between the exterior and interior of a chelonian shell. Ventrally it is roughly diamond shaped, while dorsally it is a trapezoid whose leading edges are so shallow that its shape might almost be described as that of an equilateral triangle (fig. 83B).

Superimposed on the dorsal surface of the entoplastron is a Y-shaped ridge whose open end faces anteriorly. The distal ends of the bifurcated arms lead into the transverse ridges just described at the posterior edges of the epiplastra. Along the lateral border of the inner surface of the posterior lobe of the plastron are paired bony protuberances. These structures extend over an area covering the posterior portion of the hypoplastron and an adjacent region on the anterior end of the xiphiplastron (fig. 84D).

Prominent pelvic scars are also present on the inner surface of the plastron. The posterolateral ends of the ischial scars are continuous with the U-shaped anal notch and then curve forward until the proximal ends are nearly parallel to each other and the midline. This entire paired structure is slightly raised on a pedestal above the inner plastral surface. The pubic scars are spaced in the form of obtuse-angled triangles whose apices are directed posterolaterally. A very small portion of the right pubis has been preserved, which shows that the basal thickness quickly diminishes to produce a bladelike leading edge. Noting else is known of the pelvis.

The bone sutures on the plastron of Albertwoodemys are consistent with a typical podocnemidid. Roughly trapezoidal epiplastra form the anteriormost portion of the plastron. Immediately behind these paired elements is the more or less diamond-shaped entoplastron, whose posterior extension lies behind the level of the axillary notches of the bridge. This same condition can be seen in Neochelys. Sutures between the hyoplastra and hypoplastra bisect the middle of the bridge and terminate laterally in a junction with the mesoplastral sutures. The mesoplastra themselves are roughly hexagonal elements. Sutures separating the hypoplastra from the xiphiplastra diverge laterally and posteriorly from the midline at an angle of approximately 75°.

Scale sulci on the plastron of Albertwoodemys are very clearly incised. Consequently, the absence of a humero-pectoral sulcus cannot be attributed to inadequate preservation. Much as in most living and extinct Testudinidae, the scale furrows are not simply grooves on the bony surface of the plastron, but are instead enclosed within slightly raised parallel ridges. At the front of the plastron, the pentagonal intergular scale is extremely large and broad. Small triangular gulars are wedged in on either side of the anterior end of the intergular. The pectoral-abdominal sulcus traverses the belly of the plastron anterior to the mesoplastra. The position of the abdominal-femoral sulcus is in no way remarkable. But at the posterior end of the plastron, the anal scales are so reduced that they are prevented from meeting at the midline by the xiphiplastral notch.

At both the anterior and posterior ends of the plastron, the scales extended far onto the dorsal surface of the plastron (fig. 83B). Anteriorly, most of the visceral surface of the epiplastra was covered by scales. The trapezoidal prolongation of the intergular was flanked on either side by rhombic gulars. In contrast to the situation for these scales on the external surface of the plastron, the combined areal extent of the two gulars exceeded that of the intergular. Triangular wedges of the pectoral scales stretched from the lateral borders of the gular scales back to the axillary notch. On the posterior lobe of the plastron, the femoral scales extend dorsally to cover the outer surface of the lateral bony ridges. Apparently the anal scales completely encompassed the tips of the xiphiplastra, encircling them in glovelike fashion.

Part or all of six peripheral bones have been preserved on the left side. Assuming that Albertwoodemys had the usual chelonian complement of 11 peripherals, it is probable that numbers five through 10 are represented. A lateral carina is lacking, and the vertical orientation of the peripherals shows that the carapace was high vaulted and rounded (as in UCMP 42008 and the Kalodirr shell) rather than being low and flat as in typical aquatic forms. Unlike any tortoise, however, the distal ends of the pleural scales extend well down onto the peripheral bones in agreement with the pattern encountered in pleurodires, emydids, and baenids. The sulcus separating what is probably the sixth marginal scale from the abdominal scale crosses the distal portion of the mesoplastron in a manner similar to that seen in the type of “Stereogenys” libyca (Andrews, 1906: 303, fig. 97). An unusual feature on the inner surface of the posterior peripherals is a thick ridge (fig. 84E), which continues forward and around the circumference of the inguinal notch as a bladelike crest that ultimately connects to the anterior extension of the thick ridge on the inner surface of the posterior lobe of the plastron. Such a structure is often found in testudinids but never in pleurodires.

The depositional environment of the Jebel Qatrani Fm. has been described as “fluviomarine,” but Simons (1968: 15, 16) and Bown and Kraus (1988) indicate a more terrestrial depositional environment, with channel sands and conglomerates together with clayey flood plains deposits reinforcing this interpretation. The tortoiselike structure of the shell of Albertwoodemys, suggests that it (and UCMP 42008) was a terrestrial form.

UCMP 42008, unnamed new genus and species

Diagnosis

Podocnemidid known from skull and shell; differing from all other pleurodires on the basis of these characters: triangular-shaped and fully roofed skull, cheek and temporal emargination completely absent, due to unique possession of unusually large squamosal with ventral flange on occipital surface and extensive parietal contact, laterally and anteriorly expanded supraoccipital with extensive squamosal and parietal contacts (longer than in Dacquemys, the only form to approach it), large quadratojugal with extensive squamosal and jugal contacts; anteriorly extensive prefrontal with very long maxilla contact and ventral flange forming snout region; apertura narium externa small and partially subdivided on midline; other distinguishing characters: orbits facing anterolaterally; interorbital groove absent; medial expansion of triturating surface present, median maxillary ridge present, accessory ridge present; complete cavum pterygoidei present; processus trochlearis pterygoidei present and at right angles to midline; carapace high domed with deep nuchal embayment; caudal buckler or cloacal cover formed by fused osteoderms.

Specimen

UCMP 42008, skull lacking major portions of palate and braincase, partial shell with carapace and plastron elements.

Locality

UCMP locality V4898, Moruorot Hill (often alternately spelled Muruarot Hill), 20 miles NE of Lodwar, northcentral Kenya.

Horizon

Early Miocene Lothidok Formation (Boschetto et al., 1992).

Referred Material

A caudal buckler, UCMP 41918, and a peripheral, UCMP 42141, both from UCMP locality V48100, are identified with this taxon. It should be mentioned that National Museum of Kenya WK-17139, a shell from the Miocene of Kalodirr, Kenya, is very similar to the shell of UCMP 42008 and may be the same species or genus.

Discussion

UCMP 42008 is currently being studied by J.H. Hutchison, D. Bramble, and E.S. Gaffney. Enough work has been done to include it in the phylogenetic analysis. The unusual shell morphology of this specimen links it with Albertwoodemys, based on the high-domed carapace, the deeply incised anterior margin of carapace, thickened plastral edges, and the fusion (or loss) of the humeral and pectoral scales. See discussions under Dacquemys and Albertwoodemys.

Turkanemys Wood, 2003

Type Species

Turkanemys pattersoni Wood, 2003.

Included Species

Turkanemys pattersoni Wood, 2003.

Distribution

Late Miocene Nawata Formation, early Pliocene Kanapoi Formation, and Mio-Pliocene sediments west of Ekora, northwestern Kenya (Wood, 2003).

Diagnosis

A podocnemidid known from the skull and postcrania; skull relatively high and narrow in contrast to Bauruemys; orbits facing laterally in contrast to Podocnemis; interorbital groove such as found in Podocnemis absent; temporal emargination less than in Podocnemis; parietal-quadratojugal contact long; cheek emargination present but not reaching to dorsal edge of orbit; postorbital large in contrast to Podocnemis; medial expansion of triturating surface, median maxillary ridge, present; two accessory ridges present on triturating surface in contrast to Erymnochelys and Peltocephalus, ridges deep in contrast to all other Erymnochydand; vomer absent; fossa precolumellaris present and shallow as in Peltocephalus but in contrast to Erymnochelys; foramen jugulare posterius closed; jugal-quadrate contact absent in contrast to Erymnochelys and Peltocephalus; horizontal occipital present in contrast to Erymnochelys and Peltocephalus; chorda tympani enclosed in processus retroarticularis.

Postcrania with cervical centra saddle shaped but not identical to those in Peltocephalus; nuchal bone width greater than length; six neurals extending to costal six; first neural four sided; pectoral scales do not contact mesoplastra, but do contact entoplastron and epiplastra.

Discussion

Turkanemys provides another glimpse of the podocnemidid record in the African Miocene providing further evidence that there is much to be learned about the relationships of Erymnochelys and Peltocephalus from the fossil record. The shell of Turkanemys is very similar to that of Erymnochelys, particularly in the classically influential area of the anterior plastral lobe, in that it has medially meeting gular scales rather than separated ones as in Peltocephalus. In our resolution, the MPC shows Turkanemys as the sister taxon to (Erymnochelys, Peltocephalus) + remaining magnatribe Erymnochelydand. A close alternative would be Turkanemys as the sister taxon to (Erymnochelys, Peltocephalus). Unfortunately some potentially decisive characters in Turkanemys are missing due to lack of preparation, in particular the condition of the anterior opening of the cavum ptergoidei. So future work may resolve these relationships. Nonetheless, Turkanemys is clearly a member of the magnatribe Erymnochelydand, with close similarities to both Erymnochelys and Peltocephalus.

Cervical vertebrae 4–7 of both Erymnochelys and Turkanemys are very similar to each other in being wider than high and differ from other Podocnemidinae (as defined here) in lacking the well-developed heterocoely or saddle-shaped centra. This condition could be interpreted as an intermediate between the Podocnemis fully heterocoelous centra that wrap around posterolaterally and the condition seen in the basal podocnemidid, Bauruemys. Nonetheless, our analysis still places Turkanemys outside Erymnochelys + Peltocephalus requiring the wide articular condition to be acquired twice, despite the proximity of these taxa in the cladogram. Again, this is only a few steps away from a group containing Erymnochelys, Peltocephalus, and Turkanemys. We feel that although our MPC resolves these three taxa, in view of the missing data for Turkanemys it would be more realistic to conclude that the three are an unresolved trichotomy at present.

Another as yet incompletely described piece of the Erymnochelys-Peltocephalus puzzle may be a specimen from Kenya. Witmer (1990) published an abstract announcing a new fossil from the lower Miocene of Rusinga Island, Kenya. This well-preserved specimen, KNM-RU 18401, consists of an articulated skull, shell, cervicals, and appendicular elements, and has been examined by us in the KNM. This as yet unnamed specimen has what appears to be a quadrate-jugal contact that partially closes the check emargination, similar to that seen in Erymnochelys and Peltocephalus. It has a shallow fossa precolumellaris and a posterior cervical that is almost identical with one from Turkanemys. The Rusinga skull has a surface coating of matrix obscuring sutures, which needs to be removed before the very tentative sutures reported here can be accepted. Because of this and the fact that the lower jaws are still attached to the skull by matrix, we feel that there are insufficient characters to use this specimen in our character matrix. Although Witmer mentions the presence of an interorbital groove characteristic of Podocnemis, our own examination suggests that the widely spaced orbits of the Rusinga skull have a shallow midline depression not homolgous to the distinctly formed, interorbital groove seen between the very closely spaced orbits of Podocnemis. Nonetheless, we agree completely with Witmer's conclusions (1990: 49A): “the Rusinga turtle is a member of a clade that was relatively diverse on mainland Africa, but that today exhibits a relict distribution.”

Turkanemys pattersoni Wood, 2003

Synonymy

None.

Type Specimen

KNM-LT 569, a shell, skull, and associated axial and partial appendicular skeleton (see Wood, 2003: 117, 118) for more detailed description of type.

Type Locality

Lothagam, Kenya (Wood, 2003).

Horizon

Lower member of the late Miocene Nawata Formation (Wood, 2003; see also other articles in same volume).

Referred Material

Specimens listed in Wood (2003). See also discussion below under Erymnochelys.

Previous work

See Wood (2003) for figures and description of skull and shell.

Erymnochelys Baur, 1888

Type Species

Dumerilia madagascariensis Grandidier (1867).

Included Species

Erymnochelys madagascariensis.

Distribution

Recent of Madagascar, possibly Miocene-Pliocene of eastern Africa. There have been a number of records of shell material identified as Erymnochelys sp. or as something similar to Erymnochelys (see Lapparent de Broin, 2000a, for summary and literature). Most of these are relatively fragmentary and, as far as can be determined from the publications, lack generic level diagnostic characters. However, Hirayama (1992) described shells and a posterior cervical from the Mio-Pliocene Sinda beds of the Democratic Republic of Congo (formerly Zaire) that are better preserved. The shells have a small intergular with medially meeting gular scales, diagnostic of Erymnochelys among the Recent fauna. The cervical is nearly identical to one of the cervicals in Erymnochelys, in that it is wider than high and lacks the saddle-shaped centra. However, considering the more recent additions to the fossil record, particularly Turkanemys and possibly related taxa, it is less likely that the Sinda material can be reliably identified as Erymnochelys. Among the published records, it is clear that the magnatribe Erymnochelydand has a long record in Africa that is only now being documented. The relatively conservative shell morphology in this group must be kept in mind when identifying shell material.

There is, however, undescribed skull material that can possibly be identified as Erymnochelys. Partial skulls and shell material in the MAC identified by label as coming from Miocene beds in the Sinda/Mohari region, western rift valley, Democratic Republic of Congo, consist of MAC RG1368A, MAC RG2626, MAC RG13686B, and MAC uncataloged. All are anterior portions of the skull, and all unfortunately lack any part of the cheek or basicranium. Although the preserved areas agree closely with the Recent Erymnochelys, the diagnostic characters of the cheek and basicranium are missing, making it possible that these specimens represent a taxon different from Erymnochelys.

Diagnosis

A podocnemidid known from the skull and postcrania; skull relatively high and narrow in contrast to Bauruemys; orbits facing laterally in contrast to Podocnemis; interorbital groove such as found in Podocnemis absent; temporal emargination less than in Podocnemis; parietal-quadratojugal contact long; cheek emargination absent; postorbital large in contrast to Podocnemis; medial expansion of triturating surface, median maxillary ridge, present; accessory ridge or ridges present in triturating surface; vomer absent; fossa precolumellaris deep and well defined as in Podocnemis unifilis; foramen jugulare posterius closed; jugal-quadrate contact present in contrast to all other podocnemidids except Peltocephalus and UCMP 42008; cavum pterygoidei with large anterior opening and foramen cavernosum in roof; horizontal occipital shelf absent; chorda tympani enclosed in processus retroarticularis.

Postcrania with cervical centra not saddle shaped but not the same as in Bauruemys; nuchal bone width greater than length; six neurals extending to costal six; first neural four sided; axillary buttress reaches peripheral two in contrast to Peltocephalus; axillary musk duct in buttress not in bridge; pectoral scales do not contact mesoplastra, but do contact entoplastron and epiplastra; gular scales meet on midline.

Discussion

Use of Erymnochelys and Peltocephalus goes back to Dumeril and Bibron (1835) and Baur (1888), but it was the synonymization with Podocnemis by the influential Boulenger (1889) catalog that curtailed their use, despite Baur's (1890) argument against synonymy, until Williams (1954c) began their more recent resurrection. In that paper Williams, agreeing with Baur (1890), Siebenrock (1902), and Müller (1935), proposed that Peltocephalus was allied with the African forms and not with the other South American ones, differing from the hypotheses of Dacqué (1912) and Zangerl (1948). Later, more widespread use of the three genera by Frair et al. (1978), Gaffney (1979), and others normalized the revival of the three genera.

Baur (1890: 483) published the following comparative diagnoses of the osteology of the three Recent podocnemidid genera in opposition to Boulenger's sinking of them into Podocnemis:

“Podocnemis, Wagler

“Peltocephalus, Dum. and Bib.

“Erymnochelys, Baur

Despite examining specimens, we have been unable to substantiate the atlas character in Podocnemis. The other characters became very influential systematically and appear in nearly all phylogenetic treatments of the Podocnemididae over the past 120 years, including the present one.

Erymnochelys madagascariensis (Grandidier, 1867)

Synonymy

See Boulenger (1889), Baur (1890), Tronc and Vuillemin (1974), Wermuth and Mertens (1977).

Type Specimen

MNHN Herpetology 9544 (Iverson, 1992).

Type Locality

West coast of Madagascar (Iverson, 1992).

Referred Material

Specimens that we used are as follows: AMNH Herpetology 63579, AMNH Herpetology 63574, MCZ 5198, MNHN Herpetology 92.494 DD67, NMV 1813, NMV 139, NMV 138, NMV 1839, NMV 1840, NMV 1811, NMV 1810, NMV 1475, NMV 1476, NMV 843/20, NMV 2518/0, NMV 847/1920, YPM 15398.

Possible subfossil specimens from Madagascar: DPC 11870, skull and jaws, “Anjohibe near Q’ exit, 92-M-327” (label); DPC 3797, skull and jaws, “Anjohibe, 83-310” (label).

Previous Work

See Discussion above. Literature on this species can be found in Iverson (1992). Also see Wermuth and Mertens (1961, 1977) for synonymies and further references; Kuchling (1993) has ecologic and other information and references. Williams (1954a) figures the palate. The paper by Tronc and Vuillemin (1974) is difficult to find, but has a useful review of the literature on Erymnochelys madagascariensis and description of its osteology with the most figures of any Recent podocnemidid.

Peltocephalus Dumeril and Bibron, 1835

Type Species

Emys tracaxa Spix (1824)  =  Emys dumeriliana Schweigger (1812), by monotypy fide Iverson (1992).

Included Species

Peltocephalus dumerilianus.

Distribution

Northern South America.

Diagnosis

A podocnemidid known from the skull and postcrania; skull relatively high and narrow in contrast to Bauruemys; elongate, pinched snout in contrast to all podocnemidids except tribe Stereogenyini; orbits facing laterally in contrast to Podocnemis; interorbital groove such as found in Podocnemis absent; temporal emargination less than in Podocnemis; parietal-quadratojugal contact long; cheek emargination absent; postorbital large in contrast to Podocnemis; medial expansion of triturating surface, median maxillary ridge, present; accessory ridge or ridges present in triturating surface; vomer absent; fossa precolumellaris shallow as in Podocnemis expansa and in contrast to deep as in Erymnochelys; foramen jugulare posterius closed; jugal-quadrate contact present in contrast to all other podocnemidids except Erymnochelys and UCMP 42008; cavum pterygoidei with large anterior opening and foramen cavernosum in roof; horizontal occipital shelf absent; chorda tympani enclosed in processus retroarticularis.

Postcrania with cervical centra saddle shaped; nuchal bone width greater than length; six neurals extending to costal six; first neural four sided; axillary buttress reaches peripheral three in contrast to Erymnochelys; axillary musk duct in buttress not in bridge; pectoral scales do not contact mesoplastra, but do contact entoplastron and epiplastra.

Discussion

Previous literature on this species can be found in Iverson (1992). See Erymnochelys discussion for original Baur (1890) diagnoses of Recent genera.

Peltocephalus dumerilianus Schweigger, 1812

Synonymy

See Boulenger (1889, as P. tracaxa), Iverson (1992).

Type Specimen

Apparently lost (Iverson, 1992).

Type Locality

“America meridionali” (Iverson, 1992).

Referred Material

Specimens that we have seen are: NFWFL 337, NFWFL 336, AMNH Herpetology 131886, MCZ 93077, DGM 361-RR, USNM Herpetology 44760 ( =  257687), USNM Herpetology 44761 ( =  257688), FMNH Herpetology 140281.

Previous Work

See Siebenrock (1902, as “Podocnemis dumeriliana”), Wermuth and Mertens (1961, 1977) for synonymies and further references, also Pritchard and Trebbau (1984). Williams (1954a) figures the palate. We have included new figures of the cavum pterygoidei of Peltocephalus dumerilianus to show the details of this structure.

Neochelys Bergounioux, 1954

Synonymy

See Broin (1977), Lapparent de Broin (2003a).

Type Species

Emys capelinii de Zigno, 1889.

Included Species

Neochelys arenarum, N. fajumensis, N. franzeni, N. capellinii, N. eocenica, N. zamorensis. See Broin (1977), Schleich (1993), and Lapparent de Broin (2003a) for more information on species.

Distribution

Eocene, Europe, northern Africa. A possible podocnemidid record of a Neochelys-like epiplastron was reported by Hutchison et al. (2004). Although the epiplastron is very similar to some Neochelys we do not feel this is sufficient for a range extension to Southeast Asia. The epiplastron may not be a pleurodire.

Diagnosis

A podocnemidid known from the skull and postcrania; skull relatively high and narrow in contrast to Bauruemys; orbits facing dorsolaterally; interorbital groove such as found in Podocnemis absent (except as a variation); temporal emargination less than in Podocnemis; parietal-quadratojugal contact long; cheek emargination slight; postorbital large in contrast to Podocnemis; medial expansion of triturating surface, median maxillary ridge, present; accessory ridge or ridges absent in triturating surface; vomer absent; fossa precolumellaris shallow as in Podocnemis expansa in contrast to deep as in Erymnochelys; foramen jugulare posterius closed; interparietal scale equilateral triangle; cavum pterygoidei with large anterior opening and foramen cavernosum in roof; horizontal occipital shelf absent; chorda tympani enclosed in processus retroarticularis.

Postcrania with nuchal bone width greater than length; neural series variable with six or seven neurals extending to meet costals six and seven; first neural is four sided; axillary buttress reaches peripheral three; pectoral scales do and do not contact mesoplastra, but do contact entoplastron and epiplastra.

Discussion

A number of species have been named for this genus, Broin (1977), Schleich (1993), and Lapparent de Broin (2003a) attempt to diagnose and clarify these. Only Neochelys arenarum Broin has had its skull described (except for the basisphenoid of Neochelys cf. arenarum, Lapparent de Broin, 2003a), although Lapparent de Broin (2003a: table 1) compares cranial features among Neochelys arenarum and some undescribed specimens.

Although the species are a problem, the genus is well diagnosed by a unique combination of characters. It possesses a wide parietal-pterygoid contact, as in the tribe Stereogenyini that is absent or narrow in Podocnemis, and absent in Erymnochelys and Peltocephalus. Neochelys has a relatively small interparietal scale that is an equilateral triangle, as in Papoulemys and some Podocnemis, but absent in Erymnochelys, Peltocephalus, and the tribe Stereogenyini. Neochelys and Papoulemys also lack accessory ridges in a group, the infrafamily Podocnemidodda, characterized by them.

Lapparent de Broin (2000a: 68; 2001: 171) questionably suggested the presence of Neochelys in the Fayum Eocene as “Genus indet. (Neochelys group?: ‘Stereogenys’) podocnemoides Reinach, 1903124: Qasr El Sagha beds, Fayum, Egypt Late Eocene” (Lapparent de Broin, 2000a: 68). Although there has not been further discussion of this suggestion, the known material of podocnemoides includes a plastron with a large intergular scale that divides the gulars, intergulars, and reaches the pectoral scales (Reinach, 1903124: pl. 10, fig. 1), in contrast to the material identified in this paper as fajumensis (fig. 47), which has a small intergular not dividing these scales. Whether or not podocnemoides Reinach, 1903124, belongs to Neochelys or another genus is indeterminable.

Neochelys arenarum Broin, 1977

Type Specimen

MNHNP RI 45, a shell (Broin, 1977: text figs. 10–12; pl. 2, figs. 10, 11; see also for discussion and description).

Type Locality

Rians, Var, France (Broin, 1977).

Horizon

Sparnacian (Broin, 1977).

Diagnosis

Broin (1977), Schleich (1993), and Lapparent de Broin (2003a) give diagnoses for the numerous European shell-based species and further references. Broin (1977: figs. 11, 12) shows the anterior plastral lobe of arenarum Broin, 1977, but with some individual variation, specimens all have the intergular scale separating the gular scales, in contrast to the condition in fajumensis Andrews, 1903, discussed below. Only Neochelys arenarum is known from a skull that has been described and it is very similar to that of Neochelys fajumensis.

Referred Material

Same data as type: MNHNP RI 6, nearly complete skull and jaws, slightly deformed, described and figured in Broin (1977: 70–83); MNHNP RI 7, probably not deformed but with few sutures visible; see list of other specimens in Broin (1977: 64); material from Rians, Var, France, in MDE: MDE R1 and MDE R2, skull and lower jaw.

Previous Work

The skull of this species is well described and figured in Broin (1977). A disarticulated basisphenoid associated with other Neochelys fragments is described and figured in Lapparent de Broin (2003a). Schleich (1993) described a new species, Neochelys franzeni, with associated skull material, which is not described. Other known but undescribed Neochelys skulls (Jiménez Fuentes, 1988: 9, fig. 3; Lapparent de Broin, 2001: 170–171; 2003a: table 1) have been mentioned in the literature, but are apparently unavailable for study. The shell of Neochelys arenarum is described and figured in Broin (1977).

Discussion

Although a number of species have been named, Broin (1977), Lapparent de Broin (2001), and some diagnosed comparatively in Broin (1977), Schleich (1993), and Lapparent de Broin (2003a), the shell characters distinquishing the various species seem to be within the individual variation of many recent species of Podocnemis, in our assessment.

Neochelys fajumensis (Andrews, 1903), new combination

Podocnemis fajumensis Andrews, 1903.

Podocnemis blanckenhorni Reinach, 1903124.

Erymnochelys fajumensis (Andrews) Lapparent de Broin, 2000a.

Type Specimen

CGM 10202, anterior portion of plastron, collected by Beadnell, 1902, figured in Andrews (1903: pl. 8, fig. 2C; and Andrews, 1906: fig. 93) possibly now lost (Cyril Walker, R.C.W., unable to locate in Cairo Museum, personal commun.).

Type Locality

“North of Birket-el-Qurun” Fayum Depression, Egypt (Andrews, 1906: 292).

Type Horizon

“In the Upper Eocene beds remains of another small Pleurodiran are not uncommon” (Andrews, 1903: 121), who did not specifically designate a type locality or horizon. Later authors have reinterpreted the age assignment and concluded that fajumensis occurs only in the Oligocene part of the Jebel Qatrani Fm. (Simons, 1968; Seiffert, 2006, q.v. for further references). The age of the Jebel Qatrani Fm. is a complex question with current evidence supporting only the base of the Jebel Qatrani as Eocene (Seiffert, 2006). There is no direct evidence that any of the N. fajumensis material is Eocene, but as turtle distributions are frequently undocumented, this remains a possibility.

Diagnosis

Six neurals, the first contacting nuchal; costals six, seven, eight meeting on midline; anterior end of nuchal very narrow, nuchal width-length about equal; intergular scale small, triangular or parallel sided, not completely separating gular scales, in contrast to all other Neochelys, or humeral scales on midline, barely extending onto entoplastron; differs from “Podocnemis” podocnemoides (von Reinach) in the gular and humeral scales meeting on the midline rather than being separated.

The shell is very similar to that of Erymnochelys.

Etymology

Presumably for the Fayum, Egypt.

Referred Material

AMNH 5086 (fig. 86), anterior plastral lobe; AMNH 5093, anterior plastral lobe (fig. 86); AMNH 5087, carapace. Apparently, most of the AMNH specimens came from three quarries: A, B, and C, in the lower Fossil Wood zone of the Jebel Qatrani Fm. (Simons, 1968: 20; see also Moustafa, 1974, for map).

Similarly, the BMNH (including what was sent to Cairo and is now CGM) and YPM material are also probably from the early Oligocene part of the Jebel Qatrani Fm., according to Simons (1968: 19, 20). CGM 8790, shell; CGM 8509, anterior plastral lobe; BMNH R3103, anterior plastral lobe; BMNH R3435, anterior plastral lobe; YPM 6202, shell; YPM 6203, shell. The identifications are based almost entirely on the anterior plastral lobe morphology of the small intergular scale not separating the gulars or humerals.

We list all specimens found by the Duke parties because they provide evidence that the Neochelys skulls and jaws occur in close association with fajumensis shells in the Jebel Qatrani Fm. There is no documented case of an articulated set of cervicals connecting a skull with a shell, but the proximity of the skulls and shells is a matter of a few centimeters or less (D. Deblieux, E. Simons, personal obs.). Shell fragments of fajumensis are the commonest fossil in the Quarry L 41, Quarry M, and Quarry I localities, and only articulated shell material of some completeness was collected. Even the skeptical E.S.G. agrees that this skull-shell association should be accepted. Stratigraphic data for these quarries can be found in Bown and Kraus (1988: fig. 19, q.v. this paper for further references; see also Moustafa, 1974; Gingerich, 1992). The most recent age assessment for the Fayum stratigraphy (Seiffert, 2006) indicates that Quarry L-41 is ca. 35.4 Ma (latest Eocene), Quarries A, B, and C are ca. 34.8, and Quarries I, M, and P are ca. 30.2–29.5 Ma.

DPC 3146 (figs. 44, 45), skull, complete but lacking right temporal roof and cheek, right quadrate and squamosal, Quarry M, 82-1229; DPC 12143, laterally compressed skull, most elements present but considerably displaced by disarticulation, L-41, 92-461; DPC 15310, partial skull, most elements present but considerably displaced by disarticulation, loss of palate and left quadrate and otic area, L-41, 95-350; DPC no number, L-41, 99-245 (fig. 46), nearly complete skull, dorsoventrally flattened and slightly disarticulated but not extensively deformed; DPC 1697, partial skull, consisting of both premaxillae, maxillae, frontals, partial parietals, postorbitals, jugals, palatines, and basisphenoid, Quarry I, 77-39; DPC 3882, skull lacking snout and skull roof, Quarry I, 83-252. These skulls listed above are from the Jebel Qatrani Fm., and are the basis (primarily DPC 3146) for the restorations in figures 42, 43.

DPC 6209, carapace, Quarry O, 86-861; DPC 2491, shell, Quarry R, 81-727; DPC 7747, shell, Quarry M, 85-986; DPC 9483, shell, Quarry I, 89-1028; DPC 4605, partial carapace, Quarry P, 83-1107; DPC 2149, disarticulated shell elements, Quarry M, 80-854; DPC 10686, shell, L-41, 90-936; DPC 10434, partial shell, L-41, 90-531.

All of the following are lower jaws and consist of the fused dentaries indistinguishable from the lower jaw of Neochelys arenarum Broin (1977: fig. 18, pl.2). DPC 1699, Quarry M, 77-240; DPC 2003, Quarry I, 80-813; DPC 2223, Quarry M, 80, 311; DPC 2230, Quarry M, 80-345; DPC 2602, Quarry M, 81-586; DPC 2747, Quarry M, 81-137; DPC 2754, Quarry M, 81-176; DPC 2779, Quarry P, 81-52; DPC 2858, Quarry M, 81-563; DPC 3249, Quarry I, 82-623; DPC 3261, Quarry M, 82-542; DPC 3261, Quarry M, 82-542; DPC 3308, Quarry M, 82-542; DPC 3308, Quarry I, 82-161; DPC 3442, Quarry M, 82-951; DPC 3862, Quarry M, 83-464; DPC 3973, Quarry M, 83-431; DPC 4291, Quarry M, 83-1380; DPC 4400, Quarry M, 83-302; DPC 4413, Quarry M, 83-123; DPC 4495, Quarry M, 83-787; DPC 5068, Quarry M, 84-825; DPC 5193, Quarry I, 84-867; DPC 5520, Quarry M, 85-990; DPC 5749, Quarry M, 85-193; DPC 5786, Quarry I, 85-141; DPC 5892, Quarry M, 85-309; DPC 6205, two lower jaws, Quarry M, 86-933; DPC 6413, Quarry M, 86-336; DPC 6522, Quarry M, 86-95; DPC 6558, Quarry M, 86-332; DPC 7331, Quarry M, 87-300; DPC 7545, Quarry M, 87-758; DPC 7546, Quarry I, 87-780; DPC 7673, Quarry I, 87-859; DPC 9352A, Quarry M, 89-384; DPC 9795, Quarry M, 89-330; DPC 10226, Quarry I, 90-107; DPC 10382, Quarry I, 90-347; DPC 10607, Quarry M, 90-1208; DPC 11192, two lower jaws, Quarry M, 91-362; DPC 11403, Quarry M, 91-780; DPC 12116, Quarry M, 92-407.

Discussion

The identification of Neochelys in the Fayum extends the range of this genus outside Europe to North Africa. The skull material from the Egyptian early Oligocene units is hardly distinguishable from the European Eocene Neochelys. However, the identification of this material with the Oligocene species fajumensis, extends the temporal range of Neochelys, formerly known from the early to late Eocene of Europe (55–34 Ma), adding approximately another 5 Ma to the already long duration of this genus. The recognition of Neochelys in the Paleogene of North Africa also adds another piece to the biogeographic puzzle of Afro-European dispersal in the Paleogene (Lapparent de Broin, 2003a).

The recognition of fajumensis as Neochelys is based not only on the hypothesis that the shells and skulls in the Duke quarries are associated, which is testable but not certain, but that the characters used to diagnose fajumensis actually are restricted to this single species. It is possible that these characters diagnose a wider group of taxa, as the shell characters in the Pelomedusoides (particularly the Podocnemididae), tend to be conservative in comparison to the skulls (discussion above and in Gaffney et al., 2006). Furthermore, fajumensis (fig. 47) and the Duke material differ from all European Neochelys in having small intergular scales allowing the gular scales to contact on the midline. The anterior lobe in the Pelomedusoides is the most variable part of the shell and has figured prominently in pleurodire shell taxonomy. Unfortunately, it is also individually variable (E.S.G., P.A.M., R.C.W., personal obs.), and too much emphasis has been placed on this area for characters in our judgements. Even Andrews (1906) showed that there was variation in the anterior plastral lobe of his podocnemidids.

We conclude that there are no systematic differences between the shells in the Duke quarries associated with skulls of Neochelys, and the shells identified in the literature as fajumensis. One could still argue that this taxon is not Neochelys because all Neochelys known to date have the intergular scales separating the gular scales. But then one would have to argue either that the skulls are not Neochelys or that they are not associated with the shells.

The synonymy of Podocnemis blanckenhorni 123Reinach, 1903, with Podocnemis fajumensis Andrews, 1903, appears to be correct, but the synonymy with Dacquemys suggested by Williams (1954c) is clearly not. The proposed synonymy of Pelomedusa progaleata 123Reinach, 1903, with fajumensis (Lapparent de Broin, 2000a) is hard to support as the type material is inadequate for useful comparison. Pelomedusa progaleata 123Reinach, 1903, is better considered as a nomen dubium.

Description of the Skull

The description of the skull of Neochelys arenarum by Broin (1977) is extremely close to the Egyptian skulls and only differences are noted below. Our description is based mostly on DPC 3146 (figs. 42–45) with some information from DPC 99-245 and other cranial fragments listed above.

The skull figured by Broin, MNHNP RI 6 (1977: figs. 13–18, pl. 2), is one of a series available for this species. All of the skulls in the series are slightly deformed, and the figured skull is slightly wider and lower than the others, which are narrower and almost exactly the same shape as DPC 3146, particularly the apparently undeformed MNHNP RI 7. So the wide shape of the figured Neochelys arenarum seems to be due to preservation and is unique to that specimen.

The prefrontal-frontal suture in the Egyptian specimens is W-shaped in dorsal view (fig. 42) and in MNHNP RI 6 it is transverse. DPC 3146 has an interorbital groove, similar to that seen in Podocnemis, but with a much wider interorbital distance. The other Egyptian skulls, however, do not show this groove, although we have coded Neochelys as present and absent for this character. DPC 3146 has a more emarginate posterior temporal roof with a greater quadratojugal exposure along the edge than in Neochelys arenarum, but this is within the range of variation frequently seen in many turtles. What may be the only significant distinction between the Egyptian skulls and the Neochelys arenarum series is in the cheek. The quadratojugal in DPC 3146 has a greater exposure along the cheek emargination than in MNHNP RI 6 associated with a smaller anterior process of the quadrate. It is possible, however, that even this is due to breakage in DPC 3146. In any case, we have not seen consistent characters that would differentiate the French Neochelys skulls from the Egyptian Neochelys skulls. They would probably be placed in the same species, if only the skulls were known.

Description of the Shell

The reconstruction of the shell of Neochelys fajumensis (fig. 47) is based on a number of specimens: YPM 6202, AMNH 5087, CGM 8790, CGM 10202, AMNH 5086, AMNH 5093, CGM 8509, BMNH R3103, BMNH R3435, and YPM 6203.

The carapace of Neochelys fajumensis is domed and oval in the uncrushed examples in both groups; there are eight costals and 11 peripherals. Peripherals 8–11 and the pygal (YPM 6203 and CGM 8690) bear a ridge on their inner surface, which can also be seen in DPC 2491, DPC 7747, DPC 2225 and DPC 9483. The nuchal is wider than long, with an unusually narrow anterior half, in both groups of specimens. The first neural is four sided, as in most pleurodires. There are either five (YPM 6202, AMNH 5087, CGM 8790), six (DPC 2491), or seven (DPC 6209, DPC 7747) neural bones. This is accompanied by either costals six to eight meeting on the midline or just costals seven and eight meeting on the midline.

The plastron of Neochelys fajumensis has an anterior lobe that is shorter than the posterior lobe and is squared off along its anterior margin. Small swellings are present marking each of the scale areas: intergular, gular, and humeral. The outer border of the posterior lobe curves medially in a gentle arc to the tips of the xiphiplastra. The entoplastron varies slightly from all four sutures being straight to having the posterior pair curved; their margin convex posteriorly. Mesoplastra are preserved in CGM 8790, DPC 2491, and DPC 7747. There is no sulcus on the mesoplastron. The anal notch varies from wide and semicircular to narrower and more V-shaped. On the inner surface of the posterior lobe, along the lateral margin just inside the body wall attachment ridge is a low ridge of thickened bone paralleling the margin. Narrow intergular scales not completely separating the gular scales is one of the diagnostic features used in the literature for “Podocnemis” fajumensis.

Papoulemys Tong, 1998

Neochelys laurenti (Tong) Lapparent de Broin, 2001

Type Species

Papoulemys laurenti Tong, 1998.

Included Species

Papoulemys laurenti.

Distribution

Early Eocene of France.

Diagnosis

A podocnemidid known from the skull; skull relatively high and narrow in contrast to Bauruemys; orbits facing dorsolaterally; interorbital groove such as found in Podocnemis absent; temporal emargination less than in Podocnemis; parietal-quadratojugal contact long; cheek emargination slight; postorbital large in contrast to Podocnemis; medial expansion of triturating surface, median maxillary ridge, present; accessory ridge or ridges absent in triturating surface as in Neochelys; fossa precolumellaris present but shallow as in Podocnemis expansa in contrast to deep as in Erymnochelys; foramen jugulare posterius closed; interparietal scale small equilateral triangle as in Neochelys; horizontal occipital shelf absent; differs from Neochelys in condylus occipitalis consisting mostly of exoccipitals, quadratojugal exposure on cheek wider.

Discussion

Lapparent de Broin (2001: 170) considers this genus a species of Neochelys. Considering the large number of common characters, this is a reasonable interpretation. There are differentiating characters, however, and as the other known Neochelys skulls (Jiménez Fuentes, 1988: fig. 3; Lapparent de Broin, 2001: 170–171; 2003a: table 1) have so far not been described or figured with sutures, it is hard to judge what the diagnostic skull characters for the European species of Neochelys might be. Therefore, we keep Papoulemys and find it to come out in a trichotomy with Neochelys and the tribe Stereogenyini. It is possible that a revision of Neochelys using the now available skull material (including N. fajumensis) would place laurenti in that genus.

Papoulemys laurenti Tong, 1998

Type Specimen

Musée des Dinosaures, Esperaza, France, T12.

Type Locality

Saint-Papoul, Aude, France.

Horizon

Cuisian, early Eocene.

Diagnosis

As for genus.

Previous Work

Tong (1998), Lapparent de Broin (2001).

Tribe Stereogenyini, new

Diagnosis

Members of the Podocnemididae; dorsal process of palatine reaches frontal in side wall of braincase; fossa precolumellaris absent; both foramina nervi hypoglossi combined and recessed in a short canal opening on occipital surface.

Included Taxa

Bairdemys Gaffney and Wood, 2002; Shweboemys Swinton, 1939; Stereogenys Andrews, 1901; Mogharemys, n. gen.; Cordichelys, n. gen.; Latentemys, n. gen.; Brontochelys, n. gen.; Lemurchelys, n. gen.

Discussion

There should be a coordinate tribe-level taxon for Stereogenyini, and according to the analysis presented here (fig. 98), it would be Neochelys + Papoulemys (or Erymnochelys + Peltocephalus if one were just considering the recent fauna). However, as discussed elsewhere (Phylogenetic Analysis) the resolution in this part of the magnatribe Erymnochelydand is not well supported for some taxa. Therefore, we are avoiding the formal recognition of a tribe for the sister taxon to Stereogenyini, but it's probably Neochelys.

Subtribe Mogharemydina, new

Diagnosis

As for genus Mogharemys.

Included Taxa

Mogharemys, n. gen.

Mogharemys, new genus

Type Species

Sternothaerus blanckenhorni Dacqué, 1912.

Included Species

Mogharemys blanckenhorni.

Distribution

Miocene of Egypt.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known only from the skull, uniquely possessing a median maxillary ridge on palatine smaller than lateral maxillary ridge; secondary palate and midline palatal cleft absent, medially expanded triturating surfaces not meeting on midline; orbits facing more anteriorly than laterally as in Bairdemys venezuelensis; eustachian tube confluent with fenestra postotica; parietal-pterygoid contact in septum orbitotemporale as in Tribe Stereogenyini and in contrast to Erymnochelys, Peltocephalus, and Podocnemis; dorsal process of palatine reaches ventral process of frontal as in Tribe Stereogenyini and in contrast to all other podocnemidids; posterior pocket of fossa orbitalis as found in Tribe Stereogenyini absent; fossa precolumellaris absent as in all Tribe Stereogenyini and in contrast to Erymnochelys, Peltocephalus, and Podocnemis; foramina nervi hypoglossi combined into one recessed opening posteriorly as in Tribe Stereogenyini.

Discussion

This new genus is based on a skull in the British Museum, BMNH R.8440 (figs. 48–51) that was described but not named, by 178179Williams (1954), who did not realize that the skull was the holotype of “Sternothaerus” blanckenhorni Dacqué, 1912. Williams listed characters showing BMNH R.8440 to be distinct from all other podocnemidid skulls, but refrained from naming it because he thought it was probably the skull of the shell-based taxon “Podocnemis” aegyptica Andrews, 1900, which Williams stated (1954: 5) “is on shell characters barely, if at all, distinguishable from Erymnochelys madagascariensis.” Williams hypothesized that the Moghara skull, BMNH R.8440, was “a structural intermediate between the Recent genera, Peltocephalus or Erymnochelys, and the Oligocene Dacquemys” (178179Williams, 1954: 6). Williams also suggested that BMNH R.8440 could be the skull of the other Moghara shell, Podocnemis bramlyi, but did not decide in favor of either hypothesis.

Williams discussed “Sternothaerus” blanckenhorni, the only other Moghara skull known at that time. He felt the sole figure of the specimen, a dorsal view (Dacqué, 1912: fig. 12) neither supported nor contradicted the assignment to Pelusios ( =  Sternothaerus). However, Williams was misled by the written description of Dacqué (1912: 321), that “Sternothaerus” blanckenhorni was characterized by a strongly developed secondary palate, which does not occur in BMNH R.8440. No other figures of “Sternothaerus” blanckenhorni, showing the ventral view, for example, have appeared or were available to Williams, although the cast, MB.R.2860, was in the Berlin Museum at this time. Williams was aware that the Egyptian Tertiary had podocnemidids, such as Stereogenys, with well-developed secondary palates and concluded that “Sternothaerus” blanckenhorni could be one of these, but, in any case, a secondary palate is not present in BMNH R.8440. BMNH R.8440 has an expanded triturating surface compared with the living podocnemidids, which was apparently all that Dacqué meant, but it was not expanded to the extent seen in Stereogenys.

A recent examination by E.S.G. of the cast of “Sternothaerus” blanckenhorni Dacqué, 1912, MB.R.2860, in the Berlin Museum shows that the original of the cast is actually BMNH R.8440. In ventral view the comparisons are exact, even though some of the cast does not preserve surface details. The actual specimen, BMNH R.8440, has lost the prefrontals and some chips from the edges of the skull, so the dorsal view is now significantly different from the 1912 figure. When Dacqué examined the specimen it was in the collection of the Geological Survey of Egypt (Dacqué, 1912: 321); exactly how it came to be at the BMNH is unknown. The reconstruction of the skull produced here (fig. 48) combines both the original specimen and the missing snout features from the cast.

Mogharemys blanckenhorni Dacqué (1912), new combination

Sternothaerus blanckenhorni Dacqué, 1912

Pelusios blanckenhorni 178179Williams, 1954

Type Specimen

BMNH R.8440, the holotype of “Sternothaerus” blanckenhorni Dacqué, 1912, is a partial skull in two pieces, not in contact. Dacqué did not specifically designate a type, and the only specimen he figured was the cast of this skull. However, it is apparent that he considered the skull the type specimen, and we affirm that. Dacqué (1912: 321) mentions shell fragments associated with the skull. What seem to be these specimens was examined by one of us (R.C.W.) and found to be too fragmentary to support Dacqué's original identifications. They were not figured.

Type Locality

“Wadi Moghara” label (see Miller, 1999).

Horizon

Unnamed formation, Miocene. “untermiocänen Eisensandstein von Moghara: die Knochenplatten aus derselben Stufe in Wadi Faregh” (Dacqué, 1912: 321).

Diagnosis

As for genus.

Etymology

For the locality, Wadi Moghara, Egypt, and emys, Greek for “freshwater tortoise.”

Referred Material

MB.R.2860, a cast of BMNH R.8440, made before damage resulting in the loss of the snout occurred, figured in Dacqué (1912: fig. 12).

Previous Work

See below.

Discussion

The species name blanckenhorni, has been used by two different authors for two different species with two different type specimens. Both specimens are podocnemidids from the same area, the Fayum of Egypt, but from slightly different ages and different localities. Professor M. Blanckenhorn was a German geologist who worked in Egypt for many years (Blanckenhorn, 1900, 1903, 1921) and was particularly active in collecting in the Moghara area (Andrews, 1906), and apparently popular.

The older blanckenhorni was named by von Reinach in 1903 (1903a) and consisted of the anterior half of a plastron, at that time housed in the Munich Museum, and which was probably destroyed in World War II (Crumly, 1984). Named “Podocnemis” blanckenhorni by Reinach (1903a: 460), Reinach also named it in Reinach (1903b: 27; the present authors do not know which publication was first, so we use the lettering from Kuhn, 1964). Andrews (1906: 292) suggested that “Podocnemis” blanckenhorni was a junior synonym of “Podocnemis” fajumensis Andrews, 1903, and we have accepted that.

In 1912, Dacqué used the species name for “Sternothaerus” blanckenhorni, the specimen named here Mogharemys blanckenhorni (Dacqué). It was Kuhn (1964) who confused blanckenhorni Reinach, 1903a and 1903b, with blanckenhorni Dacqué, 1912, and put them into synonymy. Lapparent de Broin (2000a) clearly differentiated the two species.

All of this may be moot, as the commonest shell in the Moghara unit according to Andrews (1900: 2), belongs to “Podocnemis” aegyptica Andrews, 1900 (fig. 87). Perhaps someday a skull-shell association may allow a new assessment of the taxonomy of BMNH R.8440.

Subtribe Stereogenyina, new

Diagnosis

Members of the Podocnemididae; secondary palate unique among turtles in being formed by maxillae and palatines that are separated on the midline by a narrow cleft; median maxillary ridge of other Podocnemididae absent; palate characterized by variably developed anterior convexity and posterior concavity; palatine makes up half or more of secondary palate; fossa orbitalis with extensive posterior pocket behind orbital rim enclosed by septum orbitotemporale.

Included Taxa

Bairdemys Gaffney and Wood, 2002; Shweboemys Swinton, 1939; Stereogenys Andrews, 1901; Cordichelys, n. gen.; Latentemys, n. gen.; Brontochelys, n. gen.; Lemurchelys, n. gen.

Infratribe Bairdemydita, new

Diagnosis

Members of the Subtribe Stereogenyina; cheek emargination present; labial ridge high and narrow; angle of front of skull is 90° or less, in contrast to the Infratribe Stereogenyita in which it exceeds 90°; pinched snout absent (slightly developed in Bairdemys hartsteini and approached by Bairdemys winkleri).

Included Taxa

Bairdemys Gaffney and Wood, 2002; Cordichelys, n. gen.; Latentemys, n. gen.

Discussion

The other clade coordinate with this one is the Infratribe Stereogenyita, which has no cheek emargination. There is, however, a serious problem with this character as only one specimen in this subtribe is determinable (a Stereogenys); the other three genera lack preserved cheeks. However, the two clades are also distinguishable by the labial ridge, which is low and thick in the Stereogenyita and high and thin in the Bairdemydita, the primitive condition. The angle of the snout is somewhat arbitrary, and the identification of what is a pinched snout and what is not has varied in the literature (Gaffney and Wood, 2002; Sánchez-Villagra and Winkler, 2006; Gaffney et al., 2008).

Cordichelys, new genus

Type Species

Podocnemis antiqua Andrews, 1903.

Included Species

Cordichelys antiqua.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known from the skull and shell; uniquely possessing among the Tribe Stereogenyini a cordiform (heart-shaped) carapace, shallow interorbital depression, pterygoids narrowly meeting on midline in contrast to either not meeting (Shweboemys and Stereogenys) or broadly meeting (all other Tribe Stereogenyini), a large ventral process of the postorbital in the septum orbitotemporale contacting palatine and preventing jugal-pterygoid contact; medial edges of palatal cleft curved in contrast to Lemurchelys, Shweboemys, and Stereogenys; low ventral convexity on triturating surface in contrast to high in Bairdemys; eustachian tube confluent with fenestra postoticum; antrum postoticum open as in Latentemys in contrast to nearly closed in Bairdemys; frontal and prefrontal flat in profile as in Shweboemys in contrast to rounded in Bairdemys.

Postcrania with six neurals extending to costals six; first neural four sided; pectoral scales do not contact mesoplastra, but do contact entoplastron and epiplastra.

Distribution

Egypt, late Eocene.

Included Species

Cordichelys antiqua.

Etymology

Cordi-, from Latin cor, “heart,” in allusion to heart-shaped carapace; chelys, Greek “turtle.”

Discussion

Cordichelys resolves as the sister taxon to the remaining members of the Infratribe Bairdemydita, but the number of supporting characters is not great. Cordichelys and Latentemys are relatively generalized within the Tribe Stereogenyini.

Cordichelys antiqua (Andrews, 1903), n. gen.

Podocnemis antiqua Andrews, 1903

Podocnemis stromeri Reinach, 1903a

Shweboemys antiqua Wood, 1970

Stereogenys cromeri Andrews, 1906 (in part)

“Pod. Stromeri var. major,” Reinach, 1903a

Type Specimen

CGM 10038 (Andrews, 1903: pl. 8, figs. A, B; more detailed description with same figure in Andrews, 1906), a nearly complete shell so heavily coated with gypsum that none of the scute sulci and only some of the bone sutures can be recognized.

Type Locality

“north of Birket el-Qurun” Andrews (1906: 289).

Type Horizon

“Qasr el-Sagha beds (Middle Eocene)….” Andrews (1906: 289); now regarded as late Eocene (Holroyd et al., 1996; Seiffert, 2006).

Diagnosis

Genus is monotypic.

Referred Specimen

YPM 7457 (formerly YPM 6205; field number 67-502, accession 7152; skull drawings published by Gaffney, 1979: fig. 136 as YPM 6205), associated skull and shell (figs. 53–57); “Eocene, Qasr el-Sagha Fm., Zeuglodon Valley, Egypt, collected Lloyd Tanner, 1966–67” (label and YPM database, original accession data states only “Zeuglodon Valley”). According to P. Holroyd (personal commun.) the Qasr el-Sagha Fm. does not outcrop in Zeuglodon Valley and the source is more likely the Birket Qarun Fm. However, Moustafa (1974: fig. 4) shows a “Podocnemis antiqua” partial shell in situ in the Qasr el-Sagha Fm. that may be YPM 7457 (P. Holroyd, personal commun.).

Discussion

Considering the large number of commonly shared characters evident in the shells of Cordichelys antiqua and “Podocnemis stromeri,” i.e., the pointed xiphiplastra, straight outer borders of the posterior plastral lobe, short anterior plastral lobe, indentation at the front end of the carapace, same number of neurals, and in general the similar configurations of bone sutures and scute sulci; it seems reasonable to conclude that all these specimens belong to a single species. Therefore, we have synonymized “Podocnemis stromeri” Reinach, 1903a, including “Pod. Stromeri var. major,” with Cordichelys antiqua.

Description

Although the type specimen of “Podocnemis” antiqua (CGM 10038) consists of a fairly complete shell, it is so heavily coated with gypsum that none of the scute sulci and only some of the bone sutures can be recognized. The Yale specimen (YPM 7457, figs. 56, 57), which forms the basis for the following description, now permits a much fuller account to be given of the shell of this poorly known species as well as of its skull.

Andrews (1906: 289) commented regarding the type of “Podocnemis” antiqua that: “The carapace in this species is flatter posteriorly than in front, where it seems to have been highly arched. The nuchal border is straight, without a trace of any emargination.” Crushed bone may have misled Andrews in reconstructing the only specimen he had available for study. Inspection of the Yale specimen confirms only part of this description. The entire carapace was apparently rather flat and more or less low arched in cross section. No evidence of a more pronounced arching toward the front can be detected in the Yale specimen. In addition, a definite emargination in the anterior edge of the nuchal is clearly discernible, producing an unusual cordiform shell.

Eleven peripherals that extend along either side of the carapace from the nuchal back to the pygal are preserved on one side or the other of YPM 7457. Contrary to Andrew's observation (1906: 290) that “The last marginals [peripheral bones] and pygals are somewhat enlarged,” these elements appear to be roughly equivalent in size to the anterior peripherals, and it would appear that the middle peripherals in the series are actually the largest. Numbers seven, eight, and nine are larger than the others, being not only longer but also wider than the others. The nuchal is much wider than long and, as noted above, is indented along the anterior margin. Andrews (1906: 289) mentioned that “at the point of junction of the nuchal, first neural, and first costal there is a small opening through the shell.” Apparently he regarded this as a natural fenestration. No similar structure is evident on the Yale specimen, and in view of the poor state of preservation of the type it is reasonable to assume that the holes that Andrews observed can be attributed to distortion caused by crushing. Six neurals are present, the first one abutting directly against the posterior end of the nuchal. The first neural is fusiform in shape, while the following four are hexagonal and elongate anteroposteriorly. The last neural is pentagonal in outline and as wide as it is long. Intervening between the last neural and the suprapygal are three costals (numbers six, seven, and eight) that meet in the midline. Of the eight costals, the first is the largest and there is a progressive decrease in size posteriorly. The suprapygal is pentagonal, its borders abutting against the last pair of costals, the 11th pair of peripherals, and the pygal. In typical pleurodiran fashion, the pelvic girdle was fused to the carapace and plastron. However, because of severe weathering in the area of the iliac attachment, it is not possible to determine the outline of the scars representing the point of fusion between pelvis and carapace.

There are five vertebral scales, all roughly hexagonal in outline. The first is slightly wider at its widest point than the nuchal bone underlying it. It converges anteriorly and is narrower than the nuchal where it reaches the anterior margin of the shell. This species is unique among Pelomedusoides in that the first vertebral extended forward to the anterior margin of the carapace, preventing the first pair of marginal scales from meeting in the midline (fig. 56). Four pairs of pleural scales are arranged along either side of the vertebrals. Extending around the fringe of the carapace are 12 marginal scales. The first pair do not meet at the front of the shell, as noted above, because of the interposition of the first vertebral scale. The sulci between the pleurals and marginals are situated well down on the peripheral bones, more or less midway between their distal and proximal ends.

In the plastron of the type, the scale sulci are unrecognizable, and the pattern of bone sutures on the outer surface is obscured. What Andrews (1906: 290) described was actually the bone suture pattern on the inner surface of the plastron. The Yale shell (YPM 7457) confirms his general description, however. The posterior lobe of the plastron is narrower than the anterior one, and it extends much farther from the bridge than does the anterior lobe. The exterior surface of the plastron is completely flat except for a slight upward curve at the posterior tip of the acutely pointed xiphiplastra. Situated rather more anteriorly than in most other pleurodires, the diamond-shaped entoplastron permits only a narrow contact between the epiplastra at the anterior margin of the plastron. From the outer angles of the entoplastron, sutures separating the epiplastra from the hyoplastra extend to the outer border on a line approximately perpendicular to the long axis of the shell. The complex pattern of these sutures, which Andrews described, is characteristic only of the inner surface of the plastron. The sutures between the hyoplastra and hypoplastra bisect the middle of the bridge and terminate laterally in a junction with the mesoplastral sutures. No mesoplastra have been preserved, but from the notch between the lateral borders of the hyo- and hypoplastron it is possible to determine that they were relatively small and probably hexagonal. The suture between the hypoplastra and the xiphiplastra is situated about one-third of the distance from the posterior end of the bridge to the posterior end of the plastron. A V-shaped notch is excavated into the posterior border of the xiphiplastra. The relative position and orientation of the ischial and pubic scars are not very different from those of recent specimens of Pelusios, Pelomedusa, and Podocnemis. Possibly, however, in Cordichelys antiqua the ischial scar may be relatively larger than the pubic scar as compared to the living forms. The pelvis itself has not been preserved.

Interposed between the gulars is a narrow triangular intergular that extends posteriorly onto the entoplastron and separates the small triangular gulars at the midline. Along the posterior part of their midline borders, the humerals meet each other, but anteriorly they too are separated by the intergular. For the most part, the lateral portions of the humero-pectoral sulcus coincide with the epi-hyoplastral suture. The pectoral-abdominal sulcus does not cross the mesoplastra; the position of this and the remaining plastral sulci are indicated in figure 56.

Latentemys, new genus

Type Species

Latentemys plowdeni, new species.

Included Species

Latentemys plowdeni.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known only from the skull; uniquely possessing large, oval orbits, with widest dimension dorsally; medial edges of palatal cleft curved in contrast to Lemurchelys, Shweboemys, and Stereogenys; low ventral convexity on triturating surface in contrast to high in Bairdemys; eustachian tube separated from rest of fenestra postotica as in Bairdemys and in contrast to Cordichelys; antrum postoticum open in contrast to nearly closed as in Bairdemys; frontal and prefrontal slightly curved in profile as in Brontochelys and in contrast to Shweboemys; jugal-pterygoid contact present due to absence of ventral process of postorbital in septum orbitotemporale in contrast to Cordichelys; basisphenoid relatively short in contrast to other Infratribe Bairdemydita; large cavum pterygoidei prevents quadrate-pterygoid contact posteriorly in contrast to Cordichelys and Bairdemys.

Distribution

?Egypt.

Etymology

Latent, from Latin latens, -entis, “hidden,” in allusion to the indeterminant locality data; emys, Greek “freshwater tortoise.”

Latentemys plowdeni, new species

Type Specimen

BMNH R.11998, a nearly complete skull (figs. 58–Fig. 59Fig. 6061). Collected by W. Plowden.

Type Locality and Horizon

Probably Egypt, probably Tertiary, possibly Miocene. The specimen was donated to the BMNH by W. Plowden, who obtained it in Egypt during military service. The dark brown, loose matrix type of preservation suggests the Moghara Fm. rather than Eocene-Oligocene sediments.

Diagnosis

Genus is monotypic.

Etymology

In recognition of the presumed collector, W. Plowden.

Discussion

Latentemys is the sister taxon to Bairdemys in our resolution of the characters. But the position of both Latentemys and Cordichelys are relatively poorly supported in this cladogram (s. 98, 99).

Bairdemys Gaffney and Wood, 2002

Type Species

Bairdemys hartsteini Gaffney and Wood, 2002.

Included Species

Bairdemys hartsteini, B. venezuelensis, B. sanchezi, B. winklerae. AMNH 30000, an unnamed skull from the Castillo Formation, is considered as another species of Bairdemys (see below).

Diagnosis

A podocnemidid of the Tribe Stereogenyini known from the skull and shell; uniquely possessing among the Tribe Stereogenyini an extremely small and slitlike postotic antrum, and a ventral vertical flange on the squamosal; medial edges of palatal cleft curved in contrast to Lemurchelys, Shweboemys, and Stereogenys; eustachian tube separated by bone from rest of fenestra postotica as in Latentemys but in contrast to all other Tribe Stereogenyini; jugal-pterygoid contact present in contrast to Cordichelys, Shweboemys, and Stereogenys.

Postcrania with saddle-shaped cervicals (known only for B. hartsteini); carapace (known only for B. venezuelensis) lacking neurals, costals meeting on midline; anterior plastral lobe short in contrast to nearly all other Podocnemididae; pectoral-abdominal sulcus not crossing mesoplastron; pectoral-humeral sulcus in anterior half of entoplastron; intergular scales barely extending onto entoplastron.

Distribution

Miocene of Puerto Rico and Venezuela. Weems (2009) has described the shell and lower jaws of a Bairdemys from the Oligocene of South Carolina, thus extending the range of this genus. Weems also hypothesized that partial shell remains from the Miocene of Cuba (MacPhee et al., 2003) and the Miocene of Maryland (Collins and Lynn, 1936) are Bairdemys, but this material is too incomplete to confirm.

Bairdemys hartsteini Gaffney and Wood, 2002

Type Specimen

AMNH 27222, a skull without associated mandible figured by Gaffney and Wood (2002).

Type Locality

North side of Highway No. 2, west of Bayamon, Puerto Rico (see Gaffney and Wood, 2002: 4).

Horizon

Cibao Formation, middle Miocene (see Gaffney and Wood, 2002: 4).

Diagnosis

A species of Bairdemys differing from all other species in having a higher skull, a straight rather than convex or concave labial ridge in ventral view, and a wider distance between orbits; also differs from B. venezuelensis in having a smaller skull, a premaxillary notch, lower triturating surface convexity, shallower triturating surface concavity, and narrower posterior triturating surface width; also differs from B. sanchezi in having a slightly wider skull, a higher triturating surface convexity, a deeper triturating surface concavity, a larger and broader basisphenoid, and a less extensive cheek and temporal emargination; also differs from B. winklerae in having a broader, shorter snout, slightly wider skull, wider triturating surface, a premaxillary notch, and a narrower basisphenoid-quadrate contact.

Referred Material

None.

Previous Work

Gaffney and Wood (2002), Sánchez-Villagra and Winkler (2006), and Gaffney et al. (2008).

Discussion

This species is not well differentiated from B. venezuelensis, and another option for grouping the Bairdemys species would be to synonymize B. hartsteini with B. venezuelensis but to keep B. sanchezi and B. winklerae, which are more distinct. Although the size range of Urumaco skulls of B. venezuelensis has been expanded, this species continues to differ from B. venezuelensis in its smaller size.

Bairdemys venezuelensis (Wood and Díaz de Gamero, 1971)

Type Specimen

Laboratorio de Paleontología, Universidad Central de Venezuela (Caracas), VF 1176, a complete carapace figured by Wood and Díaz de Gamero (1971: pls. 1, 2, 4).

Type Locality

“North of Campo El Mamón, state of Falcón, Venezuela” (Wood and Díaz de Gamero, 1971).

Horizon

Late Miocene, Urumaco Formation.

Diagnosis

A species of Bairdemys differing from all other species in having a higher anterior triturating surface convexity, a deeper posterior triturating surface concavity, and a short midline contact of the pterygoids; also differs from B. hartsteini in having a larger skull, a convex labial ridge in ventral view, no premaxillary notch, and a wider posterior triturating surface; also differs from B. sanchezi in having a wider skull, a wider posterior triturating surface, a larger basisphenoid, no premaxillary notch, and a less extensive cheek and temporal emargination; also differs from B. winklerae in having a broad, convex, and short snout, a slightly wider skull, a wider posterior triturating surface, and a narrower basisphenoid-quadrate contact.

Etymology

For the country of Venezuela.

Referred Material

See list in Gaffney and Wood (2002).

Previous Work

Wood and Díaz de Gamero (1971), Gaffney and Wood (2002), Sánchez-Villagra and Winkler (2006), Gaffney et al. (2008), Sánchez-Villagra and Scheyer (2010). Winkler and Sánchez-Villagra (2006) describe eggs from a likely Bairdemys venezuelensis nesting site, indicating that it was a colonial nester laying its eggs in beaches and living in a marine or near shore marine environment. Sánchez-Villagra and Scheyer (2010) provide a useful overview of the Venezuelan fossil turtles.

Discussion

Although this is the oldest named species in the currently construed genus Bairdemys, the type species is hartsteini, named by Gaffney and Wood (2002) so that the genus had a skull as its type species, in case some or all of the shell features of B. venezuelensis prove to be taxonomically widespread, as is in fact the case according to Gaffney et al. (2006). Fortunately, the MCZ-collected specimens consist of skulls and shells from the same quarry (Gaffney and Wood, 2002). The large size, broad snout, and straight labial ridge (in anterior view) are characteristic of this species.

Bairdemys winklerae Gaffney et al., 2008

Type Specimen

AMU-CURS 98 (Gaffney et al., 2008).

Type Locality

North of El Picache, Urumaco Formation, state of Falcón, Venezuela.

Horizon

Late Miocene, Urumaco Formation.

Diagnosis

A species of Bairdemys differing from all other Bairdemys species in having an elongate, narrow snout with the labial ridge concave in ventral view, and a wide basisphenoid-quadrate contact; also differs from B. hartsteini in having a larger skull and no premaxillary notch; also differs from B. venezuelensis in having a narrower skull, a wider posterior triturating surface, no premaxillary notch, and a longer midline pterygoid contact; also differs from B. sanchezi in having a larger skull, a larger, anteriorly wider, basisphenoid, and a less extensive temporal and cheek emargination.

Referred Material

Probably AMU-CURS 96, AMU-CURS 97, and AMU-CURS 99, belong to this species, as these specimens have elongate snouts. The specimens come from the same stratigraphic level (quarry level) at the El Picache locality (11°14′ 40″N, 70°14′″4″E; top of the upper member in the Urumaco Formation). This locality represents the eastern equivalent of the “capa de Tortuga” located in Corralito (also top of the upper member).

Previous Work

Sánchez-Villagra and Winkler (2006).

Discussion

This species is described and figured in Sánchez-Villagra and Winkler (2006) and Gaffney et al. (2008). Its large size, elongate snout, and concave labial ridges are characteristic.

Bairdemys sanchezi Gaffney et al., 2008

Type Specimen

Skull, jaws, and anterior plastral fragment (Gaffney et al., 2008). AMU-CURS 186; collector: Orangel Aguilera, Alcaldía del Municipio Urumaco, Colección Rodolfo Sánchez.

Type Locality

Tío Gregorio, Falcón State, Venezuela (Aguilera, 2004). Locality coordinates: 11°14′43″N, 70°18′19″W. The Tío Gregorio locality represents the top of the upper member in the Urumaco Formation, and may be on the same stratigraphic level as the MCZ quarry in Tío Gregorio.

Horizon

Upper Member of the Urumaco Formation.

Diagnosis

A species of Bairdemys differing from all other Bairdemys species in having a low triturating surface convexity and shallow triturating surface concavity, a smaller basisphenoid acutely pointed anteriorly, a small skull, and an extensive cheek and temporal emargination; also differs from B. hartsteini in having a convex rather than straight labial ridge in ventral view; also differs from B. venezuelensis in having a long midline pterygoid contact and a premaxillary notch, and also differs from B. winklerae in having a short, broad snout with a convex labial ridge in ventral view, a premaxillary notch, and a triturating surface that is narrower anteriorly and wider posteriorly.

Referred Material

None.

Previous Work

Gaffney et al. (2008).

Discussion

This specimen is described and figured in Gaffney et al. (2008). The small size, flatter triturating surface, and extensive temporal and cheek emargination provide ready distinction of this species from the other species of Bairdemys. The shell fragment consists of the entoplastron and most of both epiplastra. The scale pattern is very similar to that of Bairdemys venezuelensis figured in Gaffney et al. (2006: fig. 275).

Bairdemys, species indeterminate

Specimen

AMNH 30000, a poorly preserved skull and partial carapace (Gaffney et al., 2008).

Locality and Horizon

Castillo Formation, Venezuela; see Sánchez-Villagra et al. (2000), Sánchez-Villagra and Winkler (2006).

Discussion

This skull has been figured in Sánchez-Villagra et al. (2000), Sánchez-Villagra and Winkler (2006), and discussed in Sánchez-Villagra and Scheyer (2010). Sánchez-Villagra and Winkler (2006) have most recently argued that it is a distinct species within Bairdemys. We agree, but a skull without sutures that is as poorly preserved as AMNH 30000 should not be named. It can be differentiated from the other four species recognized here by the posterior position of the condylus articularis, slightly wider anterior portion of the triturating surfaces, and the more horizontal orientation of the orbits.

Infratribe Stereogenyita, new

Diagnosis

Members of the Subtribe Stereogenyina; labial ridge low and blunt; pinched snout (not known for Brontochelys); cheek emargination does not reach level of orbit; eustachian tube separated by bone from rest of fenestra postotica (not in Cordichelys); angle of front of skull exceeds 90° in contrast to the Infratribe Bairdemydita, in which it is 90° or less.

Included Taxa

Shweboemys Swinton, 1939; Stereogenys Andrews, 1901; Brontochelys, n. gen.; Lemurchelys, n. gen.

Discussion

The low and blunt labial ridge is unique to this group among podocnemidids. The absence of cheek emargination, however, is found throughout the Infrafamily Erymnochelyodda and is known only for one specimen of Stereogenys among the Infratribe Bairdemydita.

Brontochelys, new genus

Type Species

Shweboemys gaffneyi Wood, 1970.

Included Species

Brontochelys gaffneyi.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known from the skull; uniquely possessing large orbits that face forward to a greater degree than in other Tribe Stereogenyini and frontal that forms more of the dorsal orbital margin than in other Tribe Stereogenyini; medial edges of palatal cleft curved in contrast to the other Infratribe Stereogenyita (Lemurchelys, Shweboemys and Stereogenys); triturating surfaces relatively flat as in Shweboemys and in contrast to most Tribe Stereogenyini; jugal-pterygoid contact prevents palatine-parietal contact; basisphenoid very wide, wider than basioccipital, in contrast to most other Tribe Stereogenyini in which it is narrower.

Distribution

Pakistan, Early Miocene.

Etymology

From Bronte, Greek “thunder”; chelys, Greek “turtle,” in allusion to the large size of the skull.

Brontochelys gaffneyi (Wood, 1970), new combination

Shweboemys gaffneyi Wood, 1970

Type Specimen

BMNH R.8570 (figs. 62–66), a nearly complete skull, lacking some of the preorbital area, most of the lateral and posterior skull roof, and most of the quadrates. Also figured and described in Wood (1970: figs. 1, 2; pls. 2B, 3B, 4B).

Type Locality and Horizon

Unknown, but Wood (1970, q.v. for discussion) concluded that the type specimen was from Early Miocene sediments in or near the Bugti Hills, Baluchistan, Pakistan.

Diagnosis

Genus is monotypic.

Previous Work

Wood (1970).

Discussion

The distinction of this genus relies on the unique anteriorly facing orbits and large frontal bone, as well as other palatal features not occurring in the infratribe Stereogenyita.

Lemurchelys, new genus

Type Species

Lemurchelys diasphax, new species.

Included Species

Lemurchelys diasphax, new species.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known only from the skull; uniquely possessing a very deep crista supraoccipitalis base and adjoining parietal, deeper than in any other Tribe Stereogenyini, and the combination of a parallel-sided palatal cleft and broadly meeting pterygoids due to a short basisphenoid; snout shape wide as in Stereogenys and Shweboemys not narrow as in Latentemys and Cordichelys; ventral edge of orbital rim a curved surface not an acute edge as in all other Tribe Stereogenyini except Brontochelys; palatal concavity and convexity very low in contrast to Bairdemys; frontal and prefrontal flat in profile in contrast to Bairdemys; septum orbitotemporale with large postorbital and small palatine and parietal-palatine contact in contrast to Shweboemys and Stereogenys; base of crista supraoccipitalis and adjoining parietal very deep, deeper than in any other Tribe Stereogenyini.

Distribution

Miocene of Egypt.

Etymology

In allusion to its discovery in the Duke Primate Center, surrounded by a collection of live lemurs, many of which took a prolonged interest in our work.

Lemurchelys diasphax, n. gen. et sp.

Type Specimen

DPC 6425 (figs. 67–70), a nearly complete skull lacking the left otic region, the temporal roof covering, and the cheeks; DPC field number 86-1226.

Type Locality

MG-L-16, Moghara, Egypt.

Horizon

Moghara Fm., early Miocene (18–17 Ma; Miller, 1999).

Diagnosis

As for genus.

Etymology

Greek diasphax, meaning “cleft” in allusion to the cleft in the secondary palate characteristic of this group.

Previous Work

None.

Discussion

The distinction of this skull-based genus relies on the deep crista supraoccipitalis and the unique combination of palatal characters.

Shweboemys Swinton, 1939

Type Species

Shweboemys pilgrimi Swinton, 1939: 548.

Included Species

Shweboemys pilgrimi Swinton, 1939.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known from the skull; uniquely possessing an elongate basisphenoid that has very little palatine contact and very narrowly separates the pterygoids, which are almost in contact, in contrast to the longer basisphenoid, which broadly contacts the palatines in Stereogenys; pinched snout in contrast to Infratribe Bairdemydita; medial edges of palatal cleft parallel in contrast to the Infratribe Bairdemydita; basioccipital shorter than in all other Tribe Stereogenyini except Stereogenys; palatine with dorsal process in septum orbitotemporale that contacts parietal as in Stereogenys and in contrast to all other Tribe Stereogenyini.

Distribution

Burma, Pliocene or Pleistocene.

Discussion

This genus was erected by Swinton in 1939 on the basis of a single skull from Burma. Swinton recognized that it was a “pelomedusid” sensu lato and compared it with Podocnemis and Stereogenys. He was the first to determine the unique nature of the palatal cleft and recognize that this unique cleft also occurred in Stereogenys.

Wood in 1970 identified a second skull of Shweboemys pilgrimi (BMNH R.8432) using Swinton's description and new photographs of the type specimen from the Geological Survey of India. Wood (1970) also moved Andrew's “Podocnemis” antiqua to Shweboemys on the basis of the secondary palate and described a new species of Shweboemys, S. gaffneyi. At that point, the genus (essentially equal to our tribe Stereogenyini) consisted of three species: pilgrimi, antiqua, and gaffneyi.

In 1977, Jain described a new species of Pelomedusoides, pisdurensis, from the Late Cretaceous of India, first (Jain, 1977) assigning it to “Carteremys” then to Shweboemys (Jain, 1986). The change in generic assignment by Jain was prompted by the suggestion of one of us (E.S.G.) based on the published figures in Jain (1977). We still have not had an opportunity to examine the specimens, but, again based solely on the published figures, pisdurensis does not belong in Shweboemys and probably does not belong in the Tribe Stereogenyini. The figures of the ventral surface of the skull (supplemented by other photographs sent to E.S.G. by Jain) do show a broad palate but one that lacks the distinctive cleft that we argue is diagnostic for this group. From the photographs it is also difficult to be sure that the skull has the cavum pterygoidei diagnostic of the Podocnemididae. Therefore, we provisionally exclude pisdurensis from Shweboemys and the Tribe Stereogenyini and place it incertae sedis without generic assignment, within the Pelomedusoides, pending further study.

In the present study we remove Shweboemys gaffneyi Wood to a new genus, Brontochelys. We also remove Shweboemys (Podocnemis) antiqua (Andrews) to another new genus, Cordichelys. As we currently recognize Shweboemys it consists only of the type species, Shweboemys pilgrimi Swinton. The increase in taxonomic diversity in the podocnemidids with a secondary palate developed in this paper is consistent with current practice in Recent turtles, particularly Pelomedusoides.

Shweboemys pilgrimi Swinton, 1939

Type Specimen

Geological Survey of India number 17255, a partial skull consisting of the anterior portion with palate, lacking otic chambers and posterior and lateral areas of skull roof (not seen by present authors). Figured in Swinton (1939: figs. 1, 2) and Wood (1970: pl. 1).

Type Locality

“One mile NNE of Mauktet, Shwebo District, Burma” (Wood, 1970: 9).

Type Horizon

Irrawaddy Beds, Pliocene or Pleistocene, fide discussion in Wood (1970), but a Miocene age cannot now be excluded.

Diagnosis

Genus is monotypic.

Referred Specimen

BMNH R.8432, a partial skull (figs. 71–74) consisting of the anterior portion with palate, lacking right otic chamber, some of left otic chamber, and lateral and posterior skull roof. Figured and described in Wood (1970: pls. I, II, III; 2A, 3A, 4A). Only data is “Burma” (Wood, 1970).

Previous Work

Swinton (1939), Wood (1970).

Stereogenys Andrews, 1901

Type Species

Stereogenys cromeri Andrews, 1901.

Included Species

At the present time, we are including only the skull-based Stereogenys cromeri in the genus. We realize that a number of other taxa have been named as species of Stereogenys that are shell based. These species are listed below, but we do not know what generic allocation would be likely for them.

Diagnosis

A podocnemidid of the Tribe Stereogenyini known from the skull; uniquely possessing a secondary palate longer than in any other Tribe Stereogenyini, having a triturating surface of premaxilla with accessory ridge medial to labial ridge, and a basisphenoid that widely contacts the palatines and widely separates the pterygoids; pinched snout in contrast to Infratribe Bairdemydita; medial edges of palatal cleft parallel in contrast to the Infratribe Bairdemydita; basioccipital shortest of all Tribe Stereogenyini; palatine with large dorsal process in septum orbitotemporale that contacts parietal as in Shweboemys in contrast to all other Tribe Stereogenyini; triturating surface of premaxilla with accessory ridge medial to labial ridge in contrast to all other Stereogenyini.

Distribution

Late Eocene, Egypt.

Discussion

Andrews (1906: 298) assigned shells to Stereogenys by the following reasoning:

Commonality of occurrence can be used to associate skulls and shells in turtles only when there are a large number of specimens. There are rarely enough specimens (and that is certainly the case here) to make a statistical argument, and there are too many exceptions known. The “Podocnemis” referred to above was P. antiqua, here considered as Cordichelys antiqua, another tribe Stereogenyini member, closely related to Stereogenys, so that part of the argument is also problematic. To date there have been no skull-shell associations for Stereogenys. The shell-based species “Stereogenys” libyca Andrews, 1903, and “Stereogenys” podocnemoides Reinach, 1903b, are not assignable to genus on the basis of the shell, and they do not have skull material associated with them.

Nonetheless, the shells claimed by Andrews (1903: 119) to belong to the skull-based taxon Stereogenys cromeri, do have a distinctive set of characters: carapace oval and depressed, seven neural bones, the first one pentagonal and not meeting the nuchal bone but separated from it by midline contact of first pair of costal bones, buttresses relatively small, intergular scale large, separating both gulars and humerals, gular and humeral scales entirely on epiplastra. These shells could in fact belong to Stereogenys cromeri, as this shell type has not yet been associated with any skull. Although we describe this shell taxon below, there is no evidence linking it with Stereogenys.

Stereogenys cromeri Andrews, 1901

Type Specimen

CGM 10027, apparently now lost (C. Walker, personal commun., BMNH), nearly complete skull and jaws, partially crushed dorsoventrally, surface badly pitted and eroded. Figured in Andrews (1901: fig. 4) and Andrews (1906: fig. 95; pl. 25, figs. 2, 3). Represented by BMNH R.3007, a cast.

Type Locality

“North of Birket el-Qurun,” Andrews (1906: 296).

Type Horizon

“Qasr el-Sagha beds (Middle Eocene),” Andrews (1906: 296); now regarded as Late Eocene (Holroyd et al., 1996; Seiffert, 2006; see also Moustafa, 1974, for general discussion and map).

Diagnosis

Genus is monotypic.

Referred Specimens

DPC 4120 (figs. 77, 78), locality 83-L-37, from the Dir Abu Lifa Member, Qasr el-Sagha Fm. (see Holroyd et al., 1996: fig. 1 for stratigraphic position), field number 83-1118, well-preserved skull, undistorted and completely free of matrix, lacking left otic chamber, skull roofing elements, and premaxillae.

All of following specimens have “Qasr el-Sagha, Fayum” as locality data, and it is likely that at least some came from the “bone beds” (Beadnell, 1901), as Beadnell explicitly noted the turtles in his report. These “bone beds” are now recognized as part of the Dir Abu Lifa Member (Holroyd et al., 1996). CGM 10031, cast is BMNH R.8442, original presumed lost (C. Walker, personal commun., BMNH), partial skull crushed dorsoventrally; BMNH R.3190, a partial skull Andrews (1906: fig. 95), has most complete cheek of Stereogenys specimens, showing that there was not an extensive emargination; BMNH R.3189, laterally crushed skull lacking roofing elements, Andrews (1906: pl. 25, fig. 1), completely closed palate is an artifact of crushing, and sutures that are so convincingly shown are drawn in ink, a number of them incorrectly; BMNH R.3191 (figs. 79, 80), a braincase and partial palate with endocast, identified as Stereogenys on basis of extent of secondary palate, and basisphenoid contacting palatines and separating pterygoids; BMNH R.3906, consists of a poorly preserved partial skull and lower jaws, the lower jaws are very similar to the type lower jaws, CGM 10027, skull, however, is from a much smaller individual and does not belong with the jaws, although the ventral surface is poorly preserved, the posterior margins of the secondary palate are present, and they are more posteriorly placed than in CGM 10027, so it may not even be Stereogenys.

A lower jaw, University of Michigan 161 (fig. 81), Wadi Hitan, Fayum, can possibly be attributed to this species, although it is not associated with a skull and has some differences from the Andrews figured jaw (1906: pl. 25, fig. 3).

Previous Work

Andrews (1901, 1906), Wood (1970).

Type Species

Cambaremys langertoni França and Langer, 2005.

Distribution

Late Cretaceous, Brazil.

Diagnosis

A medium-size Pelomedusoides pleurodire (230 mm straight carapace length) known only from the shell and associated postcranial elements differing from other South American Cretaceous and Early Tertiary Pelomedusoides in having vertebral scales 2–4 relatively wide and hexagonal, a relatively long narrow nuchal bone (also long and narrow in Peirópolis A), first neural four sided, axillary buttress not extending onto second costal but reaching second peripheral anteriorly, suture for axillary buttress broad and curving with parallel sides, bridge peripherals unguttered, iliac attachment area on seventh costal with concave anterior margin that crosses from the eighth onto the seventh costal both medially and laterally; internal gutter of posterior peripherals and pygal absent.

Cambaremys langertoni França and Langer, 2005

Type Specimen

CPP 0252, partial carapace, mesoplastron, xiphiplastra, limb and vertebral elements (França and Langer, 2005).

Type Locality

“Quarry 2 of Price (see Campos and Kellner, 1999), at 900 m high, in the Serra do Veadinho area, municipality of Uberaba, Minas Gerais, Brazil; about 2.5 km to the N-NW of the village of Peirópolis,” 19°43′12″S, 47°45′12″W (França and Langer, 2005: 395). This locality is close to, but not the same as, the site of the large-scale excavations at Caiera Quarry, called “outcrop 1” in Carvalho et al. (2004: 978), Novas et al. (2008: 628), and shown in the quarry map in Kellner et al. (2005: 533), which have yielded the large collection of turtle specimens including Peirópolis A, Peirópolis B, and the types of Pricemys and Peiropemys. Novas et al. (2008: fig. 2) show a map with “outcrop 1” and “outcrop 2” which are presumed to be these two localities. Apparently (ibid.) the stratigraphic levels of the two sites are about the same, and we presume that Cambaremys, Pricemys, and Peiropemys were contemporaneous with each other and with the shell taxa Peirópolis A and Peirópolis B.

Horizon

Serra da Galga Member, Marília Fm., Late Cretaceous, probably Maastrichtian (França and Langer, 2005: 395).

Diagnosis

As for genus.

Referred Material

Known only from the type. We know of no additional specimens that could be referred to this taxon.

Previous Work

França and Langer (2005, 2006), Oliveira and Romano (2007), Romano et al. (2009).

Discussion

Our understanding of this taxon is based on the type description (França and Langer, 2005), as well as the dissertation in which the type (CPP-0252) is described in detail (França, 2004). We agree with Broin (1991) and França and Langer (2006) that there are multiple taxa represented in the material from late Cretaceous horizons in the vicinity of the town of Peirópolis. We have identified three shell morphologies: Cambaremys (França and Langer, 2005), Peirópolis A, Peirópolis B, and two types of skulls, Pricemys caiera and Peiropemys mezzalirai. To our knowledge, there are no skull-shell associations for the Peirópolis material (see also Broin, 1991), so at this time we name only the skulls with the understanding that if and when skull-shell associations are found, the nomenclature may need to be revised.

Romano et al. (2009) have suggested that the incompletely known Cambaremys langertoni is probably a juvenile Roxochelys wanderleyi. We disagree with this assessment. We have studied the type of Roxochelys wanderleyi, which is a plastron and partial carapace, and our independent assessment agrees with Romano et al. (2009: figs. 2B, 3C) that two additional complete shells (DGM LE 307, DGM Mezzalira) can be assigned to Roxochelys wanderleyi (see discussion in Systematics). This additional material allows a relatively complete understanding of the shell of this taxon. Although the type and only known specimen of Cambaremys langertoni is a carapace with only a small portion of the plastron, it is clear that these taxa differ in several significant features (table 4). The nuchal of Cambaremys is longer than wide while that of Roxochelys is the widest relative to length among the Cretaceous and early Tertiary South American podocnemidids we have studied (fig. 94). Furthermore, the nuchal of Cambaremys is narrower anteriorly than in Roxochelys, which is the widest in this set of taxa. It addition, the axillary buttress scar is parallel sided in Cambaremys as is shown for Peirópolis A (fig. 93A) and not tapering laterally as in Roxochelys (and Peirópolis B, fig. 93B), and the second costal is not thickened posterior to the axillary buttress in Cambaremys as it is in Roxochelys. Therefore, we conclude that although Cambaremys is poorly known and incertae sedis, it is not Roxochelys. In some ways it is more similar to Peirópolis A, but we have refrained from referring Peirópolis A to Cambaremys because the type of Cambaremys lacks important autapomophic features of Peirópolis A (see table 4 and below).

TABLE 4

Comparison of Shells of Six South American Late Cretaceous and Early Tertiary Podocnemidid Turtles

França and Langer (2005: 408–409) discuss and reject the possibility that Cambaremys is a juvenile “Podocnemis” brasiliensis. But Oliveira and Romano (2007) and we consider the latter taxon to be a nomen dubium.

Cerrejonemys Cadena, Bloch, and Jaramillo, 2010

Type Species

Cerrejonemys wayuunaiki Cadena, Bloch, and Jaramillo, 2010.

Distribution

Paleocene of northern Colombia.

Etymology

See Cadena et al. (2010).

Diagnosis

As for species.

Cerrejonemys wayuunaiki Cadena, Bloch, and Jaramillo, 2010

Type Specimen

“University of Florida, Florida Museum of Natural History Vertebrate Paleontology Collections, Gainesville, Florida/Museo Geológico, at the Instituto Nacional de Investigaciones en Geosciences, Minería y Quimica, Bogotá, Colombia” UF/IGM 33, Cadena et al. (2010: 368). The presence of the holotype and only known specimen in two widely separated institutions is unexplained in Cadena et al. (2010).

Type Locality

Cerrejón Coal Mine, Guajira Peninsula, Colombia. See Cadena et al. (2010) for further information.

Horizon

Cerrejón Fm., mid to late Paleocene, see Cadena et al. (2010) for further information.

Diagnosis

Cadena et al. (2010: 369). We do not dispute the conclusion that this taxon is diagnosable and separate from previously known taxa.

Etymology

See Cadena et al. (2010).

Referred Material

None.

Previous Work

Cadena et al. (2010).

Discussion

Cadena et al. (2010) produce a character matrix and phylogenetic analysis concluding that Cerrejonemys is the sister taxon to the genus Podocnemis. The principal character supporting this resolution is the relatively small postorbital and the jugal-parietal contact in Cerrejonemys. This may prove to be the case, but in our admittedly limited examination of the specimen, which is badly crushed, these sutures seem ambiguous, as they are in the published photographs. Cerrejonemys has a cavum pterygoidei, so it is a podocnemidid, but it apparently lacks the Podocnemis-like triturating surface ridges, and the temporal emargination is more extensive, comparable to that seen in Lapparentemys rather than Podocnemis. So, until better material is available, we feel that Cerrejonemys is best considered Podocnemididae incertae sedis.

Kenyemys Wood, 1983

Type Species

Kenyemys williamsi Wood, 1983.

Distribution

Pliocene of Kenya.

Diagnosis

From Wood (1983: 74):

Discussion

The best statement about the relationships of yet another shell-only species is “The relationship of Kenyemys to other African pelomedusids [ =  Pelomedusoides] is at present obscure” (Wood, 1983: 79).

Kenyemys williamsi Wood, 1983

Type Specimen

National Museum of Kenya NMK LT 127, a nearly complete but somewhat crushed, shell, figured in Wood (1983: figs. 1–5).

Type Locality

Lothagam Hill, southwest Turkana District, Kenya (Wood, 1983).

Horizon

Lothagam 1, Pliocene (Wood, 1983: 79).

Peirópolis A, unnamed shell taxon

Synonymy

It is likely that this shell-based form is the same taxon as one of the two skulls named from this locality, Pricemys or Peiropemys. Due to the disassociated nature of the DGM collections from Peirópolis (see below), it is not possible to assign this postcranial material to one of these two new skull-based taxa or to recognize other taxa based on postcrania that may be present. The shell material of Peirópolis A is larger in size than the postcranials of Peirópolis B, and the skull of Pricemys is larger than the skull of Peiropemys, consistent with the speculation that Peirópolis A is the shell of Pricemys. However, it should be kept in mind that at least one other shell taxon, Cambaremys langertoni, although distinctly smaller than Peirópolis A, is present at a locality close to and at the same level as the Caiera Quarry (see discussions under Cambaremys and Peiropemys).

Distribution

Caiera Quarry, near village of Peirópolis, near city of Uberaba, Minas Gerais State, Brazil (see França and Langer, 2005, for map). Collected from the quarry at base of the same hill (locality 1) near the lime plant that yielded the material referred to Peirópolis B, below, and the type skulls of Pricemys caiera and Peiropemys mezzalirai. See locality discussion under Peiropemys for more information.

Horizon

Serra da Galga Member, Marília Fm. (see Candeiro, 2009, and references).

Referred Specimens

Unnumbered complete carapace and plastron on top of DNPM/DGM cabinets and a DGM specimen labeled “Peirópolis 321.” Other disarticulated and uncataloged material shows autapomorphies of these two specimens and probably represent the same taxon.

Diagnosis

This large pelomedusoides pleurodire (to 700 mm straight carapace length) differs from other South American Cretaceous and early Tertiary Pelomedusoides in having no distinct surface sculpture; vertebral scales 2–4 relatively wide and hexagonal; nuchal bone relatively long and narrow; first or second neural four sided; axillary buttress not extending onto second costal and reaching to third peripheral anteriorly; suture for axillary buttress broad and curving with parallel sides; second costal not thickened to support axillary buttress; bridge peripherals guttered dorsally; iliac attachment area on seventh costal with convex anterior margin; costal 8 wide, contacting peripherals 9, 10, and 11; costal 7 narrow allowing peripheral 9 to contact costals 6, 7, and 8; internal gutter of posterior peripherals and pygal strongly developed.

Plastron with epiplastron having thickened anterior edge with rounded lip, and lacking single, large posteromedial process; xiphiplastron with C-shaped anal notch (except in MCT 322-R); gular scales restricted to epiplastra; intergular scale narrow; long humeral scale contact on midline; pectoral scales contact entoplastron and mesoplastra but not epiplastra.

Previous Work

Mentioned in Broin (1991).

Discussion

Based on its large size, one could associate this material with the large skull of Pricemys.

Peirópolis B, unnamed shell taxon

Synonymy

It is possible that this shell-based form is the same taxon as one of the two skulls named from this locality, Pricemys or Peiropemys. Due to the disassociated nature of the DGM collections from Caiera Quarry, Peirópolis (see below), it is not possible to assign this postcranial material to one of these two new skull-based taxa. But if one wishes to speculate, the shell material of Peirópolis B is smaller in size than the shell of Peirópolis A, and the skull of Peiropemys is smaller than the skull of Pricemys, consistent with the speculation that Peirópolis B is the shell of Peiropemys. The other Peirópolis area shell taxon (but not from Caiera Quarry), Cambaremys langertoni, is also distinctly smaller than Peirópolis A, and could be interpreted as another contender for the shell of Peiropemys.

Distribution

Locality #1, Caiera Quarry, near village of Peirópolis, same quarry as Peirópolis A, Pricemys caiera, and Peiropemys mezzalirai. See locality discussion under Peiropemys for more information.

Horizon

Serra da Galga Member, Marília Fm. (see Candeiro, 2009, and references).

Referred Specimens

Peirópolis B is the smallest of the three shell taxa from this locality and the one for which there is the most limited amount of material. The material is all disarticulated but some of it is associated. At present we have the nuchal, costals 1, 5, 7, and 8, peripherals 1, 8 to 11, the suprapygal, and the pygal. We can only speculatively assign plastron material to this taxon on the basis of size.

Diagnosis

Shell relatively small (about 125 mm straight carapace length); shell texture with weak concentric rings, as in Bauruemys elegans; nuchal bone relatively short and wide; second neural bone may have been six sided; bridge peripherals smooth, not guttered; axillary buttress attachment area on first costal narrows anterolaterally; first costal–second costal contact area not uniform, thickened near axillary attachment scar; inguinal buttress attachment area on fifth costal relatively narrow and more extensive medially in contrast to Peirópolis A; iliac attachment area on seventh costal with concave anterior margin; ventral surface of peripheral 11 and pygal without gutter or ventral flange.

Possible plastral elements have epiplastron without the thickened anterior margin seen in Peirópolis A, and with a single, large posteromedial process; xiphiplastron with V-shaped anal notch.

Previous Work

Mentioned in Broin (1991).

Discussion

Based on characters in common, Peirópolis B could be identified as Bauruemys. However, the speculative association of plastral and some of the carapacial elements with the only associated material, DNPM MCT uncataloged (consisting of nuchal, costals 1, 5, 7, 8, peripherals 1, 8–11, suprapygal, pygal), is based solely on size. Furthermore, there is the definite possibility that at least some of this material belongs to the skull taxon, Peiropemys, which differs significantly from Bauruemys. We are also aware of the strongly conservative nature of pelomedusoides shells, which typically have easily differentiated skulls and very similar shells. So we refrain from identifying this as an unnamed species of Bauruemys, or as the skull of Peiropemys.

We should emphasize again that the individually disarticulated bones that we identify as Peirópolis A and Peirópolis B are, with a few stated exceptions, not associated with other elements. We informally associate them here on the basis of size, differences from Cambaremys, and in the case of some Peirópolis A material, with morphology of an articulated shell. It is possible that there are more than three species in these collections. Our speculative association of the smaller Peirópolis B shell elements with the smaller skull of Peiropemys and the larger Peirópolis A shell elements with the larger skull of Pricemys is based only on size. All of these assumptions are weak at best, and that is why we have not formalized or used them in the phylogenetic analysis.

“Podocnemis” argentinensis Cattoi and Freiburg, 1958

Type Specimen

MACN 17.988, partial plastron lacking anterior and posterior portions, figured by Cattoi and Freiburg (1958: fig. 1 and unnumbered plate).

Type Locality and Horizon

“Parte media de las ‘Margas Multicolores,’ Quebrada Queñoal, en el flanco occidental del espolón Esquina Blanca, Q. de Humahuaca, Provincia de Jujuy, Argentina” (Cattoi and Freiburg, 1958: 60).

Diagnosis

See Cattoi and Freiburg (1958: 60, 61) for diagnosis, but we have been unable to find any characters in it that differentiate this species from most species of podocnemidids. The most variable, although not necessarily most informative, parts of the podocnemidid plastron are the anterior plastral lobe with its intergular-gular-humeral scale arrangements and the posterior anal notch and xiphiplastral eminences. These areas are unfortunately missing in the type specimen of “Podocnemis” argentinensis Cattoi and Freiburg (1958: fig. 1 and unnumbered plate). A left epiplastron (Cattoi and Freiburg, 1958: unnumbered plate, lower), MACN 16.553, does show a small gular scale completely separated on the midline and a humeral scale partially separated on the midline. However, none of these features serve to distinguish this species.

Referred Material

Specimens in AMNH, see below.

Previous Work

Cattoi and Freiburg (1958); Cattoi and Freiburg (1961, which is an abstract announcing new specimens with no taxonomic novelties), Broin (1991: 515).

Discussion

The reality of this species is actually established by material in the AMNH from Jujuy Province, Argentina, presumed to come from or near the type locality of “Podocnemis” argentinensis (label), which includes more than a dozen partial shells and a nearly complete skull and jaws. The skull, now under study by Marcelo de la Fuente (Museo de La Plata) shows a complete cavum pterygoidei with a relatively short skull and wide triturating surfaces bearing accessory ridges. This skull may allow argentinensis to be substantiated and objectively diagnosed. A brief description of this material can also be found in Broin (1991: 515).

“Podocnemis” geologorum Simpson, 1943

Type Specimen

Originally AMNH 6781 but sent to Venezuela (Museo de Ciencias Naturales, Caracas; catalog number not known) before 1965, right posterior portion of carapace and plastron figured by Simpson (1943: figs. 1, 2).

Type Locality and Horizon

“Serie de Zaraza, Mioceno. Pozo Rendivú, en la Quebrada Honda, cerca de Zaraza, Estado Guárico, Venezuela” (Simpson, 1943: 56). Sánchez-Villagra and Scheyer (2010: 184) gave more information on the age of the unit, as “late Early–early Middle Miocene.”

Diagnosis

See Simpson (1943: 56). The diagnosis is not very comparative and mentions only Podocnemis expansa. The figures show a pygal notch, which, if actually present (the specimen is not well preserved in this area and there are no photographs published), would be unique among podocnemidids. The anal notch is narrow, but this is not unique.

Referred Material

None.

Previous Work

Wood and Díaz de Gamero (1971) mentioned this taxon in passing, suggesting that it is so poorly preserved that it could even be a chelid, but that it was not Podocnemis, based on the pygal shape. Sánchez-Villagra and Scheyer (2010) gave a new figure of the shell (ibid.: fig. 9.3), but were unable to confirm the presence of mesoplastra.

Discussion

The inadequacy of the material prevents a comparative diagnosis and phylogenetic analysis.

“Podocnemis” medemi Wood, 1997

Type Specimen

IGM 1863, plastron and partial carapace, figured by Wood (1997: figs. 9.3, 9.4B).

Type Locality

Cuenca de Melgar, Carmen de Apicalá, upper Magdalena River valley, Columbia (see Kay, et al., 1997, for locality data and references).

Horizon

Honda Group (Wood, 1997).

Diagnosis

“Differs from all other pelomedusids referred to Podocnemis in having a relatively short, rounded anterior and posterior plastral lobes; extremely narrow, V-shaped anal notch; proportionately small, squarish first pair of marginal scutes; enormous size (estimated midline carapace length  =  101.5 cm); and position and shape of ischial scar on xiphiplastron (triangular in shape, expanded laterally, and moderately withdrawn from margin of anal notch)” (Wood, 1997: 159).

Referred Material

See Wood (1997).

Previous Work

Wood (1997).

Discussion

Although this is another shell species readily differentiated from other shell species, its phylogenetic relationships are characterized by Wood (1997: 168) “The relationship of Podocnemis medemi to other members of the genus is unclear.” As with a number of these incertae sedis species, the absence of a good shell character diagnosis for the living Podocnemis is a problem preventing identification of these taxa as Podocnemis.

“Podocnemis” negrii Carvalho, Bocquentin, and Lapparent de Broin, 2002

Type Specimen

UFAC-PV 4441, partial carapace and partial plastron figured by Carvalho et al. (2002: figs. 1, 2).

Type Locality

Five km south of Rio Branco, Acre, Brazil (Carvalho et al., 2002).

Horizon

Solimões Fm., Miocene (Carvalho et al., 2002).

Diagnosis

Differs from other Podocnemis in having a midline crest formed in the area of the second and third vertebral scales, with an apex on the sulcus between the scales; lacks plastral tubercles; anterior plastral lobe is relatively long and the gular and intergular scales are relatively short; most similar to Podocnemis sextuberculata (Carvalho et al., 2002).

Referred Material

Only the holotype.

Previous Work

Carvalho et al. (2002).

Discussion

Although Carvalho et al. (2002) argue that this species is the sister taxon to Podocnemis sextuberculata, they do not produce diagnostic characters arguing that it belongs to the genus Podocnemis. Even though this shell is similar to P. sextuberculata, the characters in common are not unique among the family and do not securely identify the shell as a member of the genus Podocnemis. The fact that this is one of the oldest records for this genus is therefore in doubt. The oldest secure record for Podocnemis is P. bassleri, which may be a synonym for P. expansa.

“Podocnemis” pritchardi Wood, 1997

Type Specimen

UCMP 63782, shell, figured by Wood (1997: figs. 9.1, 9.4A).

Type Locality

UCMP locality V4531, upper Magdalena River valley, Columbia (see Kay et al., 1997, for locality data and references).

Horizon

La Victoria Fm., Miocene (see Kay et al., 1997, for geology).

Diagnosis

“Differs from all other South American pelomedusids referred to Podocnemis in having very narrow, almost rectangular, laterally placed mesoplastra; outer surfaces of the bridges curving downward and outward; extreme dorsoventral flattening of the shell; only six rather than seven neural bones; ischial scar on xiphiplastron very narrow along its entire length and situated directly adjacent to anal notch” (Wood, 1997: 156).

Referred Material

See Wood (1997: 156).

Previous Work

Wood (1997).

Discussion

Some of the diagnostic features might be questioned: the “extreme” flattening may be due to postmortem crushing and laboratory reconstruction as the buttresses show signs of damage, and the number of neurals varies considerably at six vs. seven among other Podocnemis species. The remaining characters do distinguish this species. However, the absence of a good diagnosis for the living Podocnemis is a problem recognized by the author: “Although it is possible that all the pelomedusids [ =  Pelomedusoides] described here [medemi and pritchardi] may eventually prove to be genuine representatives of Podocnemis, it is by no means certain at present” (Wood, 1997: 156).

Roxochelys Price, 1953

Type Species

Roxochelys wanderleyi Price, 1953.

Distribution

Late Cretaceous of Brazil.

Diagnosis

Referral of MCT 1722-R and DGM uncataloged “Mezzalira” specimen (figured in Romano et al., 2009: figs. 2B, 3C; see also Mezzalira, 1959, 1966) to R. wanderleyi allow for an expanded diagnosis. This medium-size Pelomedusoides pleurodire (about 300 mm straight carapace length) differs from other South American Cretaceous and Early Tertiary Pelomedusoides in having fine surface sculpture consisting of small polygons; vertebral scales 2–4 relatively wide and hexagonal; a relatively short, wide nuchal bone that is wide anteriorly; first neural four sided; axillary buttress extending onto second costal and reaching to third peripheral anteriorly, suture for axillary buttress broad medially and narrow laterally, second costal thickened to support axillary buttress, bridge peripherals unguttered; iliac attachment area on seventh costal with concave anterior that crosses from the eighth onto the seventh costal both medially and laterally; internal gutter of posterior peripherals and pygal absent; gular scales restricted to epiplastra; intergular scale wide; long humeral contact on midline; pectoral scales contact entoplastron but not epiplastra or mesoplastra.

Roxochelys wanderleyi Price, 1953

Type Specimen

MCT 216-R, a partial carapace and partial plastron. Not lost, contrary to Candeiro et al. (2006: 927) (G.R. Oliveira, personal commun.).

Type Locality

Araçatuba, Jubia (municipality of Mirandópolis, São Paulo State, Brazil (Price, 1953; Candeiro et al., 2006).

Horizon

Adamantina Fm., Turonian-Santonian (Candeiro et al., 2006).

Diagnosis

As for genus.

Referred Material

MCT 1722-R; DGM uncataloged “Mezzalira” shell (figured in Romano et al., 2009: figs. 2B, 3C; Mezzalira, 1959: 5, 6, locality 14).

Previous Work

Mezzalira (1959, 1966, 1981, 1989), Broin (1991), Oliveira and Romano (2007). Oliveira and Romano (2007) correctly considered this species a nomen dubium, but with the addition of complete shells to the hypodigm, we feel that the species has risen to the level of diagnosable but incertae sedis.

Discussion

The name Roxochelys wanderleyi Price, 1953, was applied to DGM 216-R a partial carapace and partial plastron. We agree that a shell figured by Romano et al. (2009: figs. 2B, 3C), collected by the geologist S. Mezzalira (the “Mezzalira” shell), can be identified as Roxochelys wanderleyi (see below). Furthermore, after additional study of MCT 1722-R (LE 307) it is apparent that this specimen also should be referred to Roxochelys wanderleyi.

Broin (1991) suggests that Roxochelys wanderleyi be considered a junior synonym of R. harrisi on the basis of similar ornamentation. However, this suggestion has not been followed in the recent literature (França and Langer, 2006; Oliveira and Romano, 2007; but see Candeiro et al., 2006). We concur with Oliveira and Romano (2007) that both “Podocnemis” harrisi Pacheco, 1913, and “Podocnemis” brasiliensis Staesche, 1937, should be considered nomina dubia.

“Stereogenys” libyca Andrews, 1903

Type Specimen

BMNH R.3039, a carapace and plastron lacking part of the left side. This specimen was returned to the CGM in the 1960s, but the authors have been unable to locate it or determine its CGM number, so we use the original BMNH number.

Type Locality

North of Lake Qarun, Fayum Depression, Egypt (Andrews, 1903).

Horizon

Jebel el-Qatrani Fm., Early Oligocene (Andrews, 1903).

Diagnosis

Known material supposedly differs from shells attributed to “Stereogenys” cromeri by having a broader and shorter nuchal bone, a U-shaped anal notch, a pentagonal intergular, and triangular humeral scales, but see discussion below.

Referred Material

BMNH R.3100, anterior plastral lobe; BMNH R.3120, anterior plastral lobe; AMNH 5089, weathered shell.

Previous Work

Reinach (1903b), Dacqué (1912), Williams (1954b), Lapparent de Broin (2000a).

Discussion

The type specimen, as based on figures in Andrews (1903) and his description, is a relatively well-preserved and complete shell, allowing a reasonable description and reconstruction (fig. 85, based on Andrews, 1903: pl. 7). Although this shell may be lost, we summarize its morphology below because it remains as one of the better-known shell taxa from the Fayum region. In any case, there are no skull associations with this material, so it is not identifiable as Stereogenys, and there is no character set that allows a shell-based generic identification. The supposed differences between the relatively well-preserved shell material of “Stereogenys” libyca and the poorly preserved shell attributed to “S.” cromeri are minor and could easily be individual variation or preservational (R.C.W.).

Description

The carapace is low arched and oval, slightly expanded posteriorly. A strong lateral carina crosses the bridge peripherals, and the shell bulges outward in the area covered by the last vertebral scale. The trapezoidal nuchal bone is slightly indented anteriorly, similar to Cordichelys antiqua, and is about one and a half times wider than long. The first pair of costals meet on the midline behind the nuchal bone separating it from the neurals. There are seven neurals, the anterior ones longer than wide, the posterior ones wider than long. The first neural is pentagonal, the rest are hexagonal. The midline meeting of the eighth costals separates the seventh neural from the suprapygal. The suprapygal is a broadly rounded diamond with the long axis perpendicular to the midline. There is a small notch at the posterior end of the rectangular pygal bone, marking the sulcus separating the last two marginal scales. Eleven pairs of peripherals border the eight costal bones. Peripherals four to seven form the bridge, the fifth and sixth contact the mesoplastra. The posterior peripherals are more expanded than the anterior. According to Andrews (1906: 304) “both buttresses are weaker and less developed than in most pleurodires, the inguinal being the stronger of the two.” The first of the five vertebral scales is relatively small and pentagonal. In contrast to the commoner condition in pleurodires, the first vertebral is narrower than the nuchal bone underlying it. The remaining vertebrals are hexagonal and much larger than the first. Marginal scales 1, 5, 7, 9, and 11 are pentagonal; the others are all quadrilateral.

The posterior lobe of the plastron is longer than the anterior lobe. The front margin of the anterior lobe is squared off and slightly concave. Andrews noted (1906: 304) that “The upper surface in this region is also gently concave from side to side, the concavity being bounded by thickened ridges borne on the epiplastrals and anterior part of the hyoplastrals.” Differences in the shape of the entoplastron were cited by Andrews (1903: 119) as a means of distinguishing between cromeri and libyca. Although the entoplastron of libyca is smaller than that of cromeri it is the same shape. In view of the variability in the shape of this bone in some pleurodires (e.g., fajumensis Dacqué, 1912: fig. 7), the taxonomic significance of the entoplastron shape is questionable. The posterior extremity of the entoplastron extends past the level of the axillary notch. Small, hexagonal mesoplastra, longer than broad, are wedged between the distal extremities of the hyo- and hypoplastra. The anal notch is shallow and U-shaped. Parallel to the outer border of the posterior lobe, there is a ridge on the inner surface, which fades out midway along the xiphiplastron (Reinach, 1903b: pl. 13, fig. 9). The lateral surface of this ridge slopes gently outward, tapering to a thin edge. The scales of the plastron overlap onto the dorsal plastral surface to a greater extent than in the living podocnemidids. Wide separation of the small gular and humeral scales is related to the large, pentagonal intergular, extending posteriorly to the middle of the entoplastron. A sulcus separating the sixth marginal from the abdominal scale slants across the mesoplastron, but the pectoral scale does not reach the mesoplastron.

Description of Shell Attributed by Andrews (1906) to Stereogenys cromeri

Although we do not feel that there is any evidence to associate this shell with the skull of Stereogenys cromeri, or anything else, it is worth discussing the interpretation that this shell is the same taxon as the shell-based “S.” libyca.

The shell material of this presumed taxon is poorly preserved, and there is also ambiguity in the various Andrews descriptions as to which specimen he is describing. The carapace is oval, its greatest lateral expansion slightly anterior to the inguinal notch. There may have been an indentation along the anterior carapace margin. The nuchal illustrated (Andrews, 1906: fig. 96) is slightly broader than long. The total number of neurals cannot be determined, but seven is most likely, based on the reconstruction and the close similarity to the shells of “S.” libyca. Instead of contacting the nuchal bone, the first neural is separated by the midline contact of the first costal bones. The first neural is pentagonal, the only other preserved one, probably the fifth, is hexagonal, as in the better preserved “S.” libyca. The pygal is trapezoidal, and there are the usual eight pairs of costals, but the number of peripherals is indeterminate.

The plastron of “Stereogenys” cromeri has a bridge longer than the posterior lobe that is longer than the anterior lobe. The anterior lobe is more angular than rounded and its anterior edge is slightly concave. The epiplastra are broader laterally than at the midline giving the anterior lobe a squared-off shape. “In some specimens the upper surface of the epiplastrals is raised into a sort of boss near their posterior angle, and from this a slight ridge is continued backwards to the axillary buttress” (Andrews, 1906: 300). The external outline of the entoplastron, although poorly preserved, is variable in the available specimens. Being rhomboidal to escutcheon shaped in specimens described by Andrews (ibid.). Mesoplastron sutures are absent. “The posterior border of the plastron seems to have been notched in the middle line as in St. libyca, but in no specimen is this region well preserved” (Andrews, 1906: 300). And yet the anal notch shape was used to distinguish cromeri from libyca. Later work by one of us (R.C.W.) suggests that a specimen containing a left xiphiplastron (BMNH R.3200) shows that the anal notch is small and broadly V-shaped in contrast to “S.” cromeri. Little of the plastral scale pattern is preserved in the attributed material. There is a large, apparently, seven-sided intergular scale, although the actual difference between this and the large five-sided intergular of “S.” libyca is a matter of interpretation. The intergular scale extends posteriorly to the middle of the entoplastron and completely separates the humerals and the gulars. The gulars are small, triangular scales and the humerals are slightly larger and trapezoidal. Both are entirely on the epiplastra. No other scale markings are apparent in the available material.

Stupendemys Wood, 1976

“Stupendemys” souzai Bocquentin and Melo, 2006

Type Species

Stupendemys geographicus Wood, 1976.

Distribution

Miocene of northern South America.

Diagnosis (from Wood, 1976: 2)

“ Shell gigantic; carapace depressed, with irregular nodular contours on external surface and deep median notch at front; anterior border of nuchal bone thickened and moderately to strongly upturned; posterior peripheral bones moderately scalloped along margins; neurals arranged in uninterrupted sequence, numbers two through six hexagonal, the seventh pentagonal. Mesoplastra hexagonal to subcircular, largely confined to bridge; lateral ends of pectoral-abdominal scute sulci terminating just anterior to axial notches of shell.” For further diagnostic characters of vertebrae and limbs, see Wood (1976: 2, 3).

Discussion

Despite Bocquentin and Melo (2006), we do not recognize other species of Stupendemys fide Meylan et al. (2009). This is another well-diagnosed, distinct podocnemidid taxon that lacks a skull and has insufficient characters to be resolved in our character analysis. See discussion under Caninemys for why this shell is probably not that genus and in any case, there are no skull-shell associations.

Stupendemys geographicus Wood, 1976

Type Specimen

MCNC 244, medial portion of carapace, left femur, parts of scapulocoracoid, cervical vertebra (eighth?), Wood (1976: figs. 1, 3, 9).

Type Locality

West of Quebrado Tio Gregorio, near town of Urumaco, Venezuela (Wood, 1976: 3).

Horizon

“Upper member of Urumaco Formation” (Wood, 1976: 3). See also Gaffney and Wood (2002: 4), Sánchez-Villagra and Aguilera (2008), as well as entire number of Paläontologische Zeitschrift 82 (2) for more recent information on Urumaco Formation and its fauna.

Diagnosis

As for genus.

Referred Material

Specimens described in Wood (1976) from type locality: MCZ 4376, shell and associated elements (Wood, 1976: pl. 1; Meylan et al., 2009: fig. 6, lower); MCNC 245, plastron and associated elements (Wood, 1976: fig. 2); MCZ 4377, cervical (Wood, 1976: fig. 3); MCZ 4378, humerus (Wood, 1976: fig. 8).

Specimens from late Miocene of southwestern Amazonia described in Gaffney et al. (1998): LACM 131946, nuchal (Gaffney et al., 1998: fig. 3); LACM 138028, right scapula (Gaffney et al., 1998: fig. 4); LACM 137948, cervical.

Other specimens identified as Stupendemys but disputed by Meylan et al. (2009) are described in Bocquentin and Negri (1993), Lapparent de Broin et al. (1993), Bocquentin and Melo (2006).

Previous Work

Bocquentin and Negri (1993), Lapparent de Broin et al. (1993), Gaffney et al. (1998), Bocquentin and Melo (2006), Scheyer and Sánchez-Villagra (2007), Meylan et al. (2009), and Riff et al. (2010). Sánchez-Villagra and Scheyer (2010) provide an overview of the genus and nice restoration.

Type Specimen

Faculty of Sciences of the University of Salamanca, Spain (Sala de las tortugas de la Universidad de Salamanca, Jiménez Fuentes, 1988), FCUS 326, a plastron.

Type Locality

15 kms west of Salamanca, Spain.

Horizon

Ludien (Eocene).

Diagnosis

See Jiménez Fuentes (1975). We do not note any diagnostic characters.

Discussion

This taxon is not recognized in Lapparent de Broin's (2001) detailed compilation of European turtles, supporting our interpretation that the taxon is not diagnosable. The type consists of a plastron with a pair of intergular scales that divide the gulars and partially divide the humerals, a common variation in Neochelys arenarum (Broin, 1977: fig. 12). We see no features allowing this specimen to be differentiated from Neochelys. The specimen may demonstrate the presence of Neochelys arenarum in the Spanish Eocene.

Latisternon microsulcae Auffenberg, 1981

Type Specimen

Kenya National Museum F3532, a left epiplastron (figured Auffenberg, 1981: fig. 2).

Type Locality

Olduvai Gorge, Tanzania (see Auffenberg, 1981: 512 for more information).

Horizon

Pleistocene, Olduvai Group (see Auffenberg, 1981: 512 and Lapparent de Broin, 2000a: 68 for more discussion).

Diagnosis

See Auffenberg (1981: 511, 512) for characters differentiating this epiplastron from some other epiplastra. The association of a nuchal bone (Kenya National Museum F4729; figured Auffenberg, 1981; fig. 3) is not certain.

Discussion

This taxon is recognized by Lapparent de Broin (2000a: 68) as “Pelomedusoides indet. incertae familiae, ?Pelomedusidae” without comment. The type is inadequate for a reasonable diagnosis.

“Podocnemis” aegyptiaca Andrews, 1900

Type Specimen

CGM unnumbered, complete plastron and partial carapace (Andrews, 1900: pl. 1). Not seen.

Type Locality

Moghara Oasis, eastern end of Qattara Depression, Egypt.

Horizon

Moghara Fm., early Miocene (18–17 Ma; Miller, 1999).

Diagnosis

Differs from all other podocnemidids in the shape of the second and third vertebral scales, which have lateral protrusions anterior to the pleural sulci and a sinuous sulcus with the adjacent pleural scale. The intergular scale is very small, barely separating the large gular scales, which constrict the midline contact of the pectoral scales, virtually identical to that found in Erymnochelys madagascariensis.

Referred Material

A number of specimens have been referred to this species, although the basis for identifying some of the material that does not overlap morphologically with the type specimen is uncertain. BMNH R.2960, plastral fragment; BMNH R.2961, plastral and carapace fragments; CGM unnumbered, complete shell figured by Fourteau (1920: fig. 21) not seen by R.C.W. at CGM. Nonetheless, the partially inaccurate figure of Fourteau is important in that it apparently shows another individual with the vertebral scale character unique to aegyptiaca suggesting that the vertebral pattern is not an individual variation. Unfortunately, both this specimen and the type seem to be missing.

Previous Work

Andrews (1900), Dacqué (1912), Fourteau (1920), Williams (1954c).

Discussion

This species (fig. 87, based on Andrews, 1900, and Fourteau, 1920) seems to be adequately diagnosed in the literature, although the only two good specimens seem to be missing. The nearly identical scale pattern of the plastron in aegyptiaca and Erymnochelys led Williams (1954c) to place aegyptiaca in that genus. The frequently conservative nature of Pelomedusoides shells makes this conclusion difficult to confirm without associated skull material.

“Podocnemis” bramlyi Fourteau, 1920

Type Specimen

An unnumbered plastron with parts of right side of carapace attached (Fourteau, 1920: fig. 23). According to Fourteau (1920: ix), the specimen is in the CGM collection, but it could not be found by one of us (R.C.W.) in the 1960s.

Type Locality

Moghara Oasis, Egypt (Fourteau, 1920).

Horizon

Moghara Fm., early Miocene (Fourteau, 1920; Miller, 1999).

Diagnosis

Again, the only character for this taxon lies in the scale pattern of the anterior plastral lobe. In this case, the intergular scale separates the small, triangular gular scales and contacts and partly separates the humerals. The intergular is entirely on the entoplastron.

Discussion

The scale pattern in bramlyi is common in South American podocnemidids (e.g., Peltocephalus, Podocnemis erythrocephala) and led Williams (1954c) to suggest bramlyi as the ancestor of Peltocephalus. However, the scale pattern on the anterior lobe of podocnemidid shells is inadequate as the sole source of characters to determine phylogenetic relationships in this group due to intraspecific variability.

Another question, however, is whether or not bramlyi could be the shell for the skull-based Mogharemys, as they are from the same unit. There is no suggestion of physical association, although the locality data for both known specimens of bramlyi and Mogharemys is so inadequate that they could be from the same place. The only character present in bramlyi allies it with the magnatribe Eymnochelydand, which also includes Mogharemys, so it is possible. At present, however, bramlyi remains a nomen dubium.

“Podocnemis” brasiliensis Staesche, 1937

Bauruemys brasiliensis (Staesche, 1937) Kischlat, 1994

Type Specimen

DGM 214-R in Price (1953: 12) but DGM 2980 in Candeiro et al. (2006: 927). Correct number is MCT 214-R (G.R. Oliveira, personal commun.). Partial plastron figured in Staesche (1937: fig. 1 [restored], pls. 16, 17) and Price (1953: figs. 1, 2.).

Type Locality

Araçatuba Jupia, Mirandópolis, São Paulo State, Brazil.

Horizon

Adamantina Fm. fide Candeiro et al. (2006).

Diagnosis

Although diagnoses are given by both Staesche (1937) and Price (1953), these do not actually distinguish this taxon from many others known today.

Referred Material

None.

Previous Work and Discussion

The type is inadequate for an objective diagnosis, and Oliveira and Romano (2007) correctly considered it a nomen dubium.

“Podocnemis” freibergi Agnolin, 2004

Type Specimen

MACN PV 18243, fragment of anterior carapace.

Type Locality

Near Andalhuala, Santa María, Catamarca Province, Argentina (Agnolin, 2004).

Horizon

Andalhuala Fm., Late Miocene (Agnolin, 2004).

Diagnosis

See Agnolin (2004) for inadequate diagnosis.

Referred Material

None.

Previous Work

None.

Discussion

The diagnosis and description do not distinguish this carapace fragment from a number of possible pelomedusoides. It is a nomen dubium.

“Podocnemis” harrisi Pacheco, 1913

Roxochelys harrisi (Pacheco, 1913) Broin, 1991

Roxochelys harrisi (Pacheco, 1913) Candeiro et al., 2006

Type Specimen

DGM 287, right xiphiplastron and three peripherals, Pacheco (1913: text fig. 6, pl. 3, fig. 6a–6e), “all of which appear to be lost at present” (Candeiro et al., 2006: 927).

Type Locality

Colina, São Paulo State, Brazil.

Horizon

Adamantina Fm. fide Candeiro et al. (2006).

Referred Material

None.

Previous Work and Discussion

Price (1953) raised the issue of similarity between harrisi and wanderleyi, which was formalized by Broin (1988, 1991) and Kischlat (1994), both of whom referred these species to Roxochelys. The species harrisi itself has never had reasonable material added to the type specimens, which are undiagnosable by any objective criteria. Oliveira and Romano (2007) considered it a nomen dubium.

“Stereogenys” podocnemoides Reinach, 1903b

Podocnemis podocnemoides (Reinach) Schmidt, 1940

“Erymnochelyinae indet. Genus indet. (Neochelys group?: ‘Stereogenys’) podocnemoides” Lapparent de Broin, 2000a

Type Specimen

An unnumbered plastron (Reinach, 1903b: pl. 10) in the Bayerische Staatssammlung für Paläontologie und historische Geologie in Munich. Not seen, apparently lost in World War II (Crumly, 1984).

Type Locality

Northwest of Lake Qarun, near “Dime” (Reinach, 1903b)

Horizon

“Obere Mitteleocän,” probably Qasr el-Sagha Fm., Late Eocene.

Diagnosis

The only diagnostic character likely to be useful at this point is the anterior lobe of the plastron, which shows a long, narrow intergular scale separating both gular and humeral scales. The carapace, not part of the type but associated with it by Reinach (1903b: pl. 11, fig. 1), does not show the supposedly diagnostic character of the shell attributed to “Stereogenys,” the separation of the first neural from the nuchal.

Discussion

This species must be considered to consist only of the type plastron as there is no evidence that it was part of the same individual as the referred carapace. The plastron apparently comes from the Eocene Qasr el-Sagha Fm. and is a possible candidate for the shell specimens in the DPC from the same unit, rather than the Oligocene fajumensis that we have chosen. It is possible that the type specimen of “Stereogenys” podocnemoides Reinach is in fact the same species as the DPC material, but difficult to demonstrate with the present material. In the first place, the type specimen, even if it still existed, is inadequate to diagnose a podocnemidid species in our opinion. Secondly, there is variation in the known Oligocene specimens of fajumensis, including an articulated set of epiplastra and entoplastron (AMNH 5093) that has a long intergular scale separating the gulars and reaching (but not separating as in podocnemoides) to the humeral scales. Thirdly, there are a number of Eocene Neochelys from Europe that have this exact pattern (Broin, 1977: figs. 12, 22). Nonetheless, there are no known specimens of Oligocene fajumensis or of DPC specimens here identified as fajumensis that actually have the scale pattern shown in the type of podocnemoides. But it is still probably Neochelys.

It would be best to recognize “Stereogenys podocnemoides” Reinach, 1903b, as a nomen dubium.