Ronella botanica Gheerbrant, Codrea, Hosu,, 1999

GAFFNEY, EUGENE S, TONG, HAIYAN & MEYLAN, PETER A, 2006, EVOLUTION OF THE SIDE-NECKED TURTLES: THE FAMILIES BOTHREMYDIDAE, EURAXEMYDIDAE, AND ARARIPEMYDIDAE, Bulletin of the American Museum of Natural History 300 (300), pp. 1-698 : 31-41

Ronella botanica Gheerbrant, Codrea, Hosu,

Sen, Guernet, Lapparent de Broin, and Riveline, 1999

TYPE SPECIMEN: A partial plastron, JBB-21, University Babes-Bolyai, Cluj, Romania.

TYPE LOCALITY: JBB-21 of Gheerbrant et al. (1999: 519), botanical garden of Jibou, Romania.

HORIZON: Calcaires de Rona, Sparnacian/Thanetian ( Gheerbrant et al., 1999); Late Paleocene, Late Thanetian, Rona Limestone ( Lapparent de Broin et al., 2004).

DIAGNOSIS: See Lapparent de Broin et al. (2004).

ETYMOLOGY: For the botanical garden type locality ( Gheerbrant et al., 1999: 519).

REFERRED MATERIAL: 117 fragments belonging to at least six individuals ( Gheerbrant et al., 1999). More complete shells are described in Lapparent de Broin et al. (2004).

PREVIOUS WORK: Gheerbrant et al. (1999), Lapparent de Broin et al. (2004).

DISCUSSION: This series of partial shells and shell elements are pleurodiran, as shown by the pelvic scars on the plastron. A recent description of new material ( Lapparent de Broin et al., 2004) provided a nearly complete shell for this species, which substantiated it as a close relative of Dortoka .

NANORDER PLATYCHELIRA, NEW

TYPE GENUS: Platychelys Wagner, 1853 . INCLUDED TAXA: Family Platychelyidae . DIAGNOSIS: Same as family Platychelyidae .

FAMILY PLATYCHELYIDAE BRÄM, 1965 TYPE GENUS: Platychelys Wagner, 1853 .

INCLUDED GENERA: Platychelys Wagner, 1853 ; Notoemys Cattoi and Freiburg, 1961 (5 Caribemys Fuente and Itturalde-Vinent, 2001 ).

DIAGNOSIS: Pleurodires differing from all other pleurodires (except Chelus ) in having very wide costovertebral tunnel, differing from all other pleurodires in having an articulation tubercle on anterior edge of first thoracic rib, and carapace shape with anterior edge wide and straight, posterior sides tapering; other distinguishing characters: neurals alternating in size, as in Dortoka ; first thoracic rib nearly as large as second thoracic rib; thoracic vertebral centra flat ventrally, thoracic ribs flat and broad without ventral keel; first thoracic central articulation concave, wider than high; thoracic ribs 9, 10, and 11 forming sacrum and attaching to ilium; iliac scar on costals 7 and 8 and suprapygal.

DISCUSSION: Notoemys and Platychelys are united by Rueda and Gaffney (2005) to form the Platychelyidae . Platychelys oberndorferi has been described by Lang and Rütimeyer (1866), Rütimeyer (1873), Zittel (1877), Bräm (1965), and Lapparent de Broin (2001). Notoemys laticentralis has been described by Cattoi and Freiburg (1961), Fuente and Fernandez (1989), and Fernandez and Fuente (1994). Rueda and Gaffney (2005) have described another species, N. zapatocaensis , and have argued that ‘‘ Caribemys ’’ oxfordiensis belongs in Notoemys , a conclusion we accept. Notoemys laticentralis is known from the shell, some vertebrae, some appendicular elements, and a partial skull. Although we have been unable to examine the partial skull, we suspect that the bone anterolateral to the basisphenoid in Fernandez and Fuente (1994: fig. 2B) is the pterygoid rather than the quadrate, as in chelids and pelomedusids. Because this partial skull of Notoemys is incomplete, we have been able to code only a few characters for it, and, at least for the present, we consider Notoemys as another shell-only taxon.

NANORDER EUPLEURODIRA GAFFNEY

AND MEYLAN, 1988

DIAGNOSIS: Only procoelous caudal articulations; neural bones from number 2 posterior are hexagonal in shape, with ante-

TABLE 2

Major Groups of Pleurodira

Podocnemididae present no no more absent long 2 nd absent no

Bothremydidae absent no no more absent long? absent yes & no

Euraxemydidae absent no no more absent short nd 2 absent yes

2 Araripemydidae absent yes yes less absent short nd 2 absent yes

TABLE Continued

Pelomedusidae present no yes less absent short nd 2 absent no

Chelidae present no yes less present short 5 th present yes

.

large

interni suprapygal Zolhafah . Bothremydini contact Basioccipital-opisthotic of exoccipital part posterius carotici canalis prootic enclosure cavum of retroarticularis cervical reaching series neural and Taphrosphyini ventral. exposure. Kurmademy Taphrosphyini and Anteroventral Foramen formed by of Degree acustico-jugulare Splenial Processus Biconvex scale Cervical Complete a Except b Extensive c Except d in Only rolateral contacts shorter than posterolateral contacts; mesoplastra equidimensional and not meeting in midline except in Pelusios (mesoplastra absent in most but not all Chelidae ); axillary process of hyoplastron reaches third peripheral (also in Dortoka ); cervical postzygapophyses elevated on neural spine.

DISCUSSION: The Eupleurodira was named by Gaffney and Meylan (1988) for the living pleurodires, consisting of the family Chelidae and the (then) family Pelomedusidae . These authors used the absence of medially meeting mesoplastra and supramarginal scales as the diagnosis. In cladogram 2, the mesoplastron character is still diagnostic for Eupleurodira, but the absence of supramarginals is more ambiguous due to their presence in the Platychelyidae , which requires either a reversal in Platychelys or independent loss in Notoemys . Other postcranial characters are also useful to diagnose this group. The main ambiguity is Dortoka , a shell-only genus that falls outside the Eupleurodira in cladogram 2, but this placement requires a few reversals (see also table 1).

In cladogram 1, with deletion of the shell-only taxa, Eupleurodira is the same as Pleurodira and has 41 synapomorphies diagnosing it.

HYPERFAMILY CHELOIDES GRAY, 1825

TYPE GENUS: Chelus Duméril, 1806 . INCLUDED TAXA: Family Chelidae . DIAGNOSIS: Same as family Chelidae .

FAMILY CHELIDAE GRAY, 1825

TYPE GENUS: Chelus Duméril, 1806 .

INCLUDED GENERA: Chelus Duméril, 1806 ; Acanthochelys Gray, 1873 ; Chelodina Fitzinger, 1826 ; Elseya Gray, 1867 ; Emydura Bonaparte, 1836 ; Hydromedusa Wagler, 1830 ; Phrynops Wagler, 1830 ; Platemys Wagler, 1830 ; Pseudemydura Siebenrock, 1901 ; Rheodytes Legler and Cann, 1980 ; Elusor Cann and Legler, 1994 ; Bonapartemys Broin and Fuente, 2001 ; Lomalatachelys Broin and Fuente, 2001 ; Prochelidella Broin and Fuente, 2001 ; Palaeophrynops Broin and Fuente, 2001 ; Yaminuechelys Fuente, Lapparent de Broin, and Manera de Bianco, 2001 .

DISTRIBUTION: Early Cretaceous to Recent, South America and Australia (see King and Burke, 1989; Iverson, 1992; Broin and Fuente, 2001; Fuente et al., 2001).

DIAGNOSIS: Eupleurodires with the following unique characters: cheek emargination uniquely extensive, reaching parietal and usually squamosal; quadratojugal absent; cervical formula (2((3((4((5))6))7((8).

DISCUSSION: Monophyly of the Chelidae is well corroborated. Within-group relationships based on morphology ( Gaffney, 1977b; Gaffney and Meylan, 1988; Bona and Fuente, 2005) and molecules ( Seddon et al., 1997; Shaffer et al., 1997; Georges et al. 1998; Krenz et al., 2005) differ significantly, and the molecular phylogenies seem to be well supported. Nonetheless, we have not adopted a within-group cladogram and have therefore coded some characters as variable. Note that recent discoveries of Cretaceous chelids (Fuente et al., 2001; Lapparent de Broin and Fuente, 2001; Fuente, 2003; Bona and Fuente, 2005) support the morphology-based analyses. See King and Burke (1989), Iverson (1992), and Cann (1998) for literature on the living species. See also table 2 for characters of the Chelidae .

The chelid cervical formula of (2((3((4((5))6))7((8) has been suggested for the Platychelyidae and possibly Dortoka by Lapparent de Broin and Murelaga (1999), in which case this would become a synapomorphy at the level of Pleurodira or Megapleurodira . However, this is based on only a few cervicals, and our own examination of these taxa makes this generalization questionable.

HYPERFAMILY PELOMEDUSOIDES COPE,

1868, NEW RANK

TYPE GENUS: Pelomedusa Wagler, 1830 .

INCLUDED TAXA: Families Pelomedusidae , Araripemydidae , Euraxemydidae , Podocnemididae , Bothremydidae .

DIAGNOSIS: Pleurodires with the unique possession of prefrontals meeting on the midline and nasals absent; parietal-squamosal contact absent; splenial absent; cervical scale absent; cervical vertebrae procoelous with biconvex second vertebra; primitively, processus paroccipitalis of opisthotic projects posteriorly beyond squamosal, and prearticular-angular contact present; anterior plastral lobe usually reaches anterior carapace margin.

MAGNAFAMILY PELOMEDUSERA

COPE, 1868

TYPE GENUS: Pelomedusa Wagler, 1830 .

INCLUDED GENERA: Pelomedusa Wagler, 1830 ; Pelusios Wagler, 1830 ; Araripemys Price, 1973 .

DIAGNOSIS: Pelomedusoides with extensive temporal and cheek emargination (also in Kurmademydini and Teneremys ); axillary process not reaching costal one (not in Pelomedusa ).

DISCUSSION: This taxon is based on cladogram 2 (fig. 292), which shows the Pelomedusidae and Araripemys as sister taxa. However, if the high-missing-value shell-only taxa are deleted (cladogram 1, fig. 288), then Araripemys is outside the Pelomedusidae and there is no Pelomedusera. Even when the shell-only taxa are included, the few shared characters do not strongly support this group. However, the clade does exist in this analysis and is worth recognizing in regard to relationships of the Pelomedusidae (sensu stricto). This clade is also proposed by Kischlat (1996b). Another way to reflect this relationship would be to put Araripemys into the family Pelomedusidae .

The axillary process reaches the first costal in Pelomedusa , the preseumed primitive condition for the Pelomedusidae , because the process reduction in Pelusios seems to be related to the hinge mechanism unique to that genus.

FAMILY PELOMEDUSIDAE COPE, 1868 a

TYPE GENUS: Pelomedusa Wagler, 1830 .

INCLUDED GENERA: Pelomedusa Wagler, 1830 ; Pelusios Wagler, 1830 .

DIAGNOSIS: Pelomedusoides with condylus occipitalis formed only by exoccipitals; foramen caroticum laterale absent; neural series not complete to suprapygal; vomer absent; basioccipital-opisthotic contact present (also in Podocnemididae ); incisura columellae enclosing eustachian tube and stapes (also in Podocnemididae ).

DISCUSSION: Although the published fossil record for this family is very poor,

Lapparent de Broin (2000a) listed a number of new records extending back to the Paleocene. Unfortunately, the basis for these new identifications is not known. See King and Burke (1989) and Iverson (1992) for literature on the living species of Pelusios and Pelomedusa . See table 2 for a comparison of the Pelomedusidae to other families of Eupleurodira.

FAMILY ARARIPEMYDIDAE PRICE, 1973

TYPE GENUS: Araripemys Price, 1973 .

INCLUDED GENUS: Araripemys Price, 1973 (see below for Taquetochelys Broin, 1980 , discussion).

DIAGNOSIS: As for genus.

DISCUSSION: Taquetochelys decorata Broin, 1980 was described from the Aptian of Gadoufaoua, Niger, on the basis of a series of shell fragments ( Broin, 1980: pl. 2, figs. 1a, 1b, pl. 3, figs. 2–11). The fragments have a characteristic surface texture, similar to that of Araripemys . Broin (1980) placed the genus in the Araripemydidae on the basis of the texture. It has small, lateral mesoplastra, absent in Araripemys . At present, the material is inadequate to establish a proper diagnosis and is considered Pelomedusoides incertae sedis (see Dubious Taxa below); however, it is possible that better specimens from the type locality could make the taxon diagnosable.

Fuente and Lapparent de Broin (1997) described a partial carapace as being a new but unnamed taxon of Pelomedusoides too incomplete to place in a family, but similar in some characters to Araripemys . They discussed the inadequate nature of Taquetochelys and listed presumed occurrences of Araripemys .

See also table 2 for comparisons among families of Pelomedusoides.

Araripemys Price, 1973

TYPE AND ONLY INCLUDED SPECIES: Araripemys barretoi Price, 1973 .

DISTRIBUTION: Albian Late Aptian (Early Cretaceous), Ceará, Brazil.

ETYMOLOGY: For the Araripe Chapada, the region where the type was found ( Price, 1973).

REVISED DIAGNOSIS: A member of the Pelomedusoides with extensive temporal and cheek emargination as in Pelomedusidae , but in contrast to the roofed condition of Podocnemididae , Bothremydidae , and Euraxemydidae, and to the extreme cheek emargination of Chelidae ; basisphenoid very long, nearly reaching palatines, with posterior expansion resulting in a shape unique among Pelomedusoides; narrow triturating surface without accessory ridges and with thin labial ridge, in contrast to Bothremydidae and Podocnemididae ; incisura columellae auris not enclosed by bone; fossa precolumellaris deep, as in Chelidae and Pelomedusidae ; quadrate-basioccipital contact absent, as in Chelidae and Pelomedusidae ; ventral exposure of exoccipital extensive, in contrast to all other Pelomedusoides; prootic widely exposed ventrally and forming foramen posterius canalis carotici interni, in contrast to Euraxemydidae , Bothremydidae , and Podocnemididae ; processus interfenestralis of opisthotic widely exposed ventrally, in contrast to Euraxemydidae , Bothremydidae , and Podocnemididae .

Very flat, sculptured carapace in which first costals reach shell margin between nuchal and first peripherals, very long neck, reduced plastron lacking mesoplastra and gular scutes, inverted V-shaped entoplastron, J-shaped epiplastra forming a sharp point anteriorly, three midplastral fontanelles, postzygapophyses joined forming a single articular surface in cervical vertebrae 2–8, first thoracic strongly sutured to nuchal, and medial and lateral centralia absent ( Meylan, 1996).

DISCUSSION: The cranial part of the diagnosis has been revised on the basis of the new skull material and reexamination of previous material. The postcranial part of the diagnosis is essentially from Meylan (1996).

Araripemys barretoi Price, 1973

Araripemys arturi Fielding, Martill, and Naish, 2005 .

TYPE SPECIMEN: Divisao de Geologia e Mineralogia, Departamento Nacional de Produçao Mineral , Rio de Janeiro, DGM-DNPM 756 -R ; shell lacking anterior margin, part and counterpart ( Price, 1973: figs. 1–5).

TYPE LOCALITY: Gypsum mine, ca. 2 km NE Santana do Cariri, Ceara, Brazil ( Price, 1973) (fig. 13).

HORIZON: The Romualdo Member of the Santana Formation lies between the Crato Member of the Santana Formation (below) and the Exu Formation, which overlies the Santana Formation. Both of the latter horizons are most likely Albian in age. This suggests an Albian age (Early Cretaceous ±110 mya) for the Romualdo Member of the Santana Formation and for nearly all specimens of Araripemys ( Maisey, 1990, 1991). Fielding et al. (2005) described an Araripemys from the Crato Formation that probably extends the taxon back to the Late Aptian.

DEPOSITIONAL ENVIRONMENT: Near-shore marine ( Maisey, 1990) with probable freshwater episodes ( Maisey, 2000). Also occurs with the chelonioid Santanachelys ( Hirayama, 1998) , the bothremydid Cearachelys , and the euraxemydid Euraxemys . Maisey (2000) made the argument that the Crato has significant freshwater deposits with saline influxes. Araripemys is rare in the Crato and its relation to freshwater/saline deposits is ambiguous. Maisey (2000) also showed that the Santana Formation proper also has freshwater episodes. Considering that Araripemys is much more common in this unit, it may also be fresh water in habit. The life environment of the Santana turtles is still unclear.

DIAGNOSIS: As for genus.

ETYMOLOGY: In recognition of Abel Barreto, the discoverer of the type ( Price, 1973).

REFERRED MATERIAL: THUg 1357 , skull (figs. 28, 29), shell, cervicals, some limb elements ; near Crato (or Jardim?), Ceara, Brazil (label), purchased from ‘‘ J. Karl 9/11/ 1990 (Rhg 9)’’, Romualdo Member, Santana Formation ; THUg 1907 , skull (figs. 32, 33) with articulated cervicals, associated shell fragments ; Santana do Cariri ?, Ceara, Brazil, Romualdo Member, Santana Formation, purchased from ‘‘ Schwickert (Kranz)’’, Germany, 6/6/1994 ; AMNH 22550 , fully prepared carapace and plastron ; carapace complete, plastron complete anteriorly but lacking both xiphiplastra; both scapulae and coracoids, proximal left and complete right humeri, both femora, left tibia and fibula, right pubis, ischium and ilium, left ilium, left astragalus, eighth cervical and numerous caudal vertebrae; AMNH 22556 , fully prepared, partial shell with most proximal parts of postcranial skeleton, plastron complete except for lateral parts of bridge, central part of the carapace, thoracics, proximal caudals, and last two cervical (seven and eight) vertebrae, pectoral girdles and humeri, and pelvis with both femora ; AMNH 24452 , partially prepared, small, badly fractured shell, left otic region of skull preserved in epoxy adjacent to shell, right coracoid, both femora, posterior caudals, cervicals seven and eight, and left humerus ; AMNH 24453 , large shell with plastron and carapace complete, complete cervical series folded over carapace ; partial skull, including right half of basicranium, right otic region, parts of both maxillae and dentaries; both scapulae, humeri, femora, pelvis, articulated right carpus (fixed to carapace), and disarticulated elements from both feet; AMNH 24454 , anterior quarter of shell with complete cervical series, complete right front limb, skull (figs. 30, 31, 282), lower jaws, hyoid, posterior parts of shell preserved but unprepared ; AMNH 22551 , unprepared shell with cervical vertebrae exposed ; AMNH 24455 , partially prepared shell with both hands and feet in articulation ; AMNH 24456 , unprepared shell and partial skull ; AMNH 24457–24461 , five unprepared shells ; Staatliches Museum für Naturkunde, Karlsruhe, PAL 3979, lateral edge of shell exposed ventrally .

PREVIOUS WORK: Price (1973) described Araripemys barretoi on the basis of a partial shell ( Price, 1973: figs. 1–5; Meylan and Gaffney, 1991: photo on p. 329). Price erected the family Araripemydidae for Araripemys and placed it in the superfamily Pleurosternoidea in the suborder Amphichelydia . Gaffney (1972a, 1975b) argued that the Amphichelydia was a paraphyletic ‘‘wastebasket’’ taxon, largely consisting of cryptodires and pleurodires lacking more derived features. All authors subsequent to Price (1973) have placed Araripemys in the Pleurodira ( Broin, 1980, 1988; Schleich, 1990; Kischlat and Campos, 1990; Hirayama, 1991; Meylan and Gaffney, 1991; Kischlat, 1996a, 1996b; Meylan, 1996).

Broin (1980) was the first to refer Araripemys to the Pleurodira . Along with a new genus, Tacquetochelys, based on shell fragments, she suggested the family had affinities to chelids. In 1988, Broin moved the Araripemydidae to the Pelomedusoides. Schleich (1990) described the first new shell material of Araripemys since Price (1973) and followed Broin’s identification of it as a pleurodire. Kischlat and Campos (1990) described the first vertebral and limb material of Araripemys , and concluded that it was ‘‘nearer to the chelids ancestors than to pelomedusids (sensu latissimo) ancestors [sic]’’ ( Kischlat and Campos, 1990: 387). In an abstract, Kischlat (1996a: 45 A) mentioned Araripemys as having ‘‘uncertain affinities’’, but in another abstract (1996b) he said it ‘‘is a sister-group of Pelomedusa + Pelusios ’’, a conclusion in agreement with our analysis that would support the magnafamily Pelomedusera clade.

Kischlat and Campos (1990) figured a number of limb and vertebral elements, as well as a plastron. Schleich (1990) figured shell material. Meylan and Gaffney (1991) have a limited description and photographs. The most extensive description of Araripemys to date is Meylan (1996). This described and figured nearly the complete osteology: shell, limbs, and vertebrae. The skull restoration ( Meylan, 1996: fig. 4) is modified here by the discovery of new material, but the detailed stereophotographs of a partial skull, AMNH 24453, are still important.

Meylan and Gaffney (1991) presented a list of characters and figures of the first skull known of Araripemys (a restoration altered by new material described here). They concluded that Araripemys is not a chelid but a ‘‘pelomedusid’’ (5 Pelomedusoides) based on cervical and cranial characters. The only detailed phylogenetic analysis of Araripemys published to date is Meylan (1996), who presented a character analysis and dataset. He concluded that Araripemys and Euraxemys (at that time, ‘‘FR 4922’’) were sister taxa and the sister group to all other Pelomedusoides, a conclusion that we now dispute.

Fielding et al. (2005) named a new species, A. ‘‘ arturi ’’, considered here a synonym of A. barretoi , discussed below.

DISCUSSION: Two new specimens, THUg 1357 and THUg 1907, and further preparation of the two skulls available to Meylan (1996) allow a reinterpretation of the skull morphology of Araripemys and a reassessment of its relationships. The interpretation of Araripemys as the sister group of the Pelomedusoides or the alternative of Araripemys as the sister group to the Pelomedusidae , does not rest on a large number of characters (see discussion above). A multichotomy with Pelomedusidae and all other Pelomedusoides, however, is well supported.

Broin (1980) identified ‘‘? Araripemys sp. ’’ from the Aptian of Gadoufaoua, Niger, on the basis of a group of shell fragments, one of which is figured ( Broin, 1980: pl. 3, fig. 1). The surface texture is similar to Araripemys , and there is an eighth costal showing the iliac suture, so it is a pleurodire. However, the shell texture is not unique to Araripemys , and it alone is insufficient to extend the range of Araripemys to Africa. This specimen must be considered Pleurodira incertae sedis.

Fielding et al. (2005) described a new species of Araripemys from the Upper Aptian/Lower Albian Crato Formation south of Nova Olinda in the Araripe Basin, northeast Brazil (using Maisey, 1991, terminology; Fielding et al. [2005] used Crato and Santana as distinct formations). They used a slightly different peripheral shape, shell outline ( Fielding et al., 2005: fig. 7), and difference in the shape of the terminal phalanges to differentiate their new species, A. ‘‘ arturi ’’, from A. barretoi . These characters are insufficient to diagnose a new species. The type of A. ‘‘ arturi ’’ is a small individual, about 180 mm carapace length, and these supposed differences could be a result of minor ontogenetic variation, sexual dimorphism, or individual variation. The shell shape of A. ‘‘ arturii ’’ was determined ( Fielding et al., 2005) by extrapolation of a portion of one side lacking the midline, and the restoration could easily be flawed due to slight preservational bias. The preservation of the type appears to be inadequate to be certain of the shape of the terminal phalanges. The type is in a block that has been split, and it appears that there has been some loss of bone from the area of the specimen from which the terminal phalanges are known. Notched or ‘‘arrowshaped’’ unguals are related to claw development, which is variable within recent turtle species. Examination of Araripemys barretoi specimens (AMNH 24453–24455) also shows that all unguals are not arrow-shaped. Given the weakness of these diagnostic characters, we place A. ‘‘ arturi ’’ in the synonymy of A. barretoi until a stronger argument can be made for its recognition as a separate species. Nonetheless, the specimen does belong to Araripemys and its discovery in the Crato Formation makes it the oldest described Araripemys . Although the age of the unit has some ambiguity (see also Berthou, 1994; Maisey, 2000), at least some of the Crato seems to be Aptian. The type horizon of Araripemys barretoi is the Albian Romualdo Member of the Santana Formation.

MAGNAFAMILY PODOCNEMIDERA

COPE, 1868

TYPE GENUS: Podocnemis Wagler, 1830 .

INCLUDED TAXA: Families Euraxemydidae , Podocnemididae , Bothremydidae ; Teneremys lapparenti Broin, 1980 .

DIAGNOSIS: Pelomedusoides with these unique characters: at least half of prootic covered ventrally by quadrate and basisphenoid; processus interfenestralis of opisthotic covered ventrally; foramen posterius canalis carotici interni formed at least in part by pterygoid (formed by pterygoid and basisphenoid primitively); inguinal buttress extends to center of fifth costal (not known for Euraxemydidae ).

DISCUSSION: The monophyly of the magnafamily Podocnemidera, consisting of the Euraxemydidae plus the epifamily Podocnemidinura ( Podocnemididae and near relatives) plus the Bothremydidae , seems relatively robust using the above basicranial characters. The poorly known Teneremys lapparenti de Broin, 1980 might be better resolved when skull material is described.

MAGNAFAMILY PODOCNEMIDERA, INCERTAE SEDIS

Teneremys de Broin, 1980

TYPE AND ONLY INCLUDED SPECIES: Teneremys lapparenti de Broin, 1980 .

DISTRIBUTION: Aptian, Niger ( Broin, 1980) .

ETYMOLOGY: From Ténéré Desert, Niger, type locality.

DIAGNOSIS: Same as species.

Teneremys lapparenti Broin, 1980

TYPE SPECIMEN: MNHN GDF 820, a partial carapace and skull.

TYPE LOCALITY: Gadoufaoua, Niger ( Broin, 1980) .

HORIZON: ‘‘GAD 5’’ Aptian ( Broin, 1980).

DIAGNOSIS: Podocnemidera with these distinguishing characters: extensive temporal emargination; basisphenoid very elongate, in contrast to all other Pelomedusoides (except Araripemys ); foramen posterius canalis carotici interni formed in anterior part of basisphenoid-pterygoid suture, in contrast to all other Pelomedusoides; nuchal embayment present (also in some Bothremydidae ). See Broin (1980) for the original diagnosis.

ETYMOLOGY: For Albert de Lapparent, uncle of F. Lapparent de Broin.

REFERRED MATERIAL: MNHN GDF 819, 821–829.

PREVIOUS WORK: Only the type description ( Broin, 1980).

DISCUSSION: Teneremys is represented by a series of skulls and associated shells and skeletons from the Aptian of Niger, not well prepared, partially described by Broin (1980). Although key areas of the quadrate and basicranium are undescribed, enough skull characters are visible to place Teneremys in the magnafamily Podocnemidera, the group formed by Euraxemydidae + epifamily Podocnemidinura ( Podocnemididae + close relatives) + Bothremydidae . Because the material is still not well known, it seems best to place Teneremys as incertae sedis within this group. The analysis presented here is in contrast to Lapparent de Broin’s assertion (2000a: 67) that Teneremys is a close relative of the Pelomedusidae .

The shell material of Teneremys shows a carapace that seems to lack the cervical scale, and a neural series that may or may not reach the suprapygal and allow medial contact of costals 7 and 8. The plastron has the intergular scale completely separating the gulars and extending onto the entoplastron. The pectoral scale extends onto the mesoplastron.

The skull material of Teneremys is very limited at present. Although at least three entire skulls are available, poor preservation and preparation leave many characters undeterminable. The skull roof is clearly emarginate, to the extent seen in Pelomedusidae and Araripemys . The ventral view is obscured by the lower jaws, incomplete preparation, and poor preservation. The processus trochlearis pterygoidei is present and similar to that in Pelomedusidae . The basisphenoid is very elongate, almost completely separating the pterygoids, as in Araripemys . There are what appear to be two small foramina at the anterior end of the basisphenoid-pterygoid suture. It is possible that these are for the carotids, but if so, they are in a unique position for pleurodires. Unfortunately, the key area of the prootic and medial process of the quadrate is badly damaged and lacking preserved bone.

Teneremys has 56 % missing data and resolves as the sister taxon to ‘‘ Platycheloides ’’ cf. nyasae, GDF 801 (below). Because Teneremys is represented by a number of skulls and skeletons, after further preparation, hopefully it will become a significant taxon in future character analyses.

‘‘ Platycheloides ’’ cf. nyasae Haughton, 1928

SPECIMEN: MNHN GDF 801, shell, figured and described in Broin (1980); GDF 800, carapace; a third uncataloged carapace ( Broin, 1980).

LOCALITY: Gadoufaoua, Niger ( Broin, 1980) .

HORIZON: ‘‘GAD 5’’ Aptian ( Broin, 1980).

DISCUSSION: A nearly complete shell of a small pleurodire was figured and described by Broin (1980: pl. 1). The type and only specimen of ‘‘ Platycheloides ’’ nyasae Haughton, 1928 is too poorly preserved to usefully diagnose, and, as suggested by Broin (1980), MNHN GDF 800 is probably a different taxon anyway. It is listed in Lapparent de Broin (2000a) as ‘‘ Pelomedusidae Cope 1868 , Plesions to still extant genera of Pelomedusidae’’ ( Lapparent de Broin, 2000a: 67). The type of ‘‘ Platycheloides ’’ nyasae is too incomplete to be resolved in the dataset (and is considered here a nomen dubium), but MNHN GDF 800 is a nearly complete shell that resolves as the sister taxon to Teneremys . These two taxa together are the sister group to the superfamily Podocnemidoidea , but they are placed as incertae sedis within the magnafamily Podocnemidera due to extensive missing data and the ease with which they move around the cladogram with a few changes to the character and taxon list.

SUPERFAMILY EURAXEMYDOIDEA , NEW

TYPE GENUS: Euraxemys , n. gen.

INCLUDED GENERA: Euraxemys , n. gen.; Dirqadim , n. gen.

DIAGNOSIS: Same as family Euraxemydidae .

FAMILY EURAXEMYDIDAE , NEW

TYPE GENUS: Euraxemys , n. gen.

INCLUDED GENERA: Euraxemys , n. gen.; Dirqadim , n. gen.

DISTRIBUTION: Albian of Brazil, Cenomanian of Morocco.

DIAGNOSIS: A member of the hyperfamily Pelomedusoides characterized by the unique possession of a medial process of the quadrate partially covering the prootic and narrowly contacting a ventral process of the exoccipital, in contrast to other pleurodires, which have either complete exposure or complete covering of the prootic ventrally; ventral process of exoccipital, which is exposed at lateral margin of basioccipital in an elongate foot (fig. 44), in contrast to nearly all pleurodires, which lack a ventral process or which have broad exoccipital exposure ( Araripemys ); other characters include an accessory ridge on mandibular triturating surface; fossa precolumellaris present but shallow, in contrast to absent (bothremydids) or deep (all other pleurodires except some Podocnemididae ); foramen posterius canalis carotici interni formed by basisphenoid and pterygoid (also prootic in Dirqadim ); quadratojugal-parietal contact present, quadratojugal large; foramen jugulare posterius partially closed (also in some Bothremydidae and in Brasilemys ); sutured dentary symphysis (not known for Dirqadim ); cervical articulations procoelous with biconvex second (not known for Dirqadim ); complete neural series reaching suprapygal (not known for Dirqadim ); abdominal scale midline length less than anal scale length (not known for Dirqadim ).

DISCUSSION: This group of two taxa is strongly corroborated. The present analysis argues that it is the sister group to the superfamily Podocnemidoidea ( Bothremydidae + Podocnemididae ). See also tables 2 and 3.

Euraxemys , new genus

‘‘FR 4922’’ Gaffney and Meylan, 1991; Meylan, 1996.

TYPE AND ONLY INCLUDED SPECIES: Euraxemys essweini , n. gen. et sp.

DISTRIBUTION: Early Cretaceous of Brazil.

ETYMOLOGY: Eurax, Greek for sideways; emys, Greek for turtle, in allusion to its pleurodiran nature.

DIAGNOSIS: A member of the family Euraxemydidae differentiated from Dirqadim by parietal-squamosal contact absent; temporal emargination more extensive; quadratojugal exposed in temporal emargination; skull longer and narrower; labial ridge thinner and straight in lateral view; accessory ridge on premaxilla absent; median concavity on premaxilla absent; triturating surface more expanded anteriorly; accessory ridge on maxilla weakly developed; antrum postoticum larger; prootic-opisthotic contact much shorter; median pterygoid contact longer; no ventral opening into canalis cavernosus; foramen posterius canalis carotici interni formed by pterygoid and basisphenoid without prootic contribution; foramen nervi vidiani exposed lateral to foramen posterius canalis carotici interni.

DISCUSSION: See table 3 for generic comparison.

Euraxemys essweini , new species

TYPE SPECIMEN: FR 4922, a nearly complete skeleton (figs. 39–46, 231, 232, 252–254, 281), also figured in Gaffney and Meylan (1991). A cast of the carapace made before embedding in plastic is AMNH 30568.

TYPE LOCALITY: Araripe Basin, Brazil (fig. 13).

HORIZON: Santana Formation, Albian, Early Cretaceous ( Maisey, 1990, 1991).

TABLE 3 Genera of Euraxemydidae

DEPOSITIONAL ENVIRONMENT: Near-shore marine ( Maisey, 1990) with freshwater episodes ( Maisey, 2000). See Araripemys for more discussion.

DIAGNOSIS: Same as genus.

ETYMOLOGY: In honor of Stephan Esswein, a student of pleurodire development at Tübingen University, who passed away prematurely in 1993.

REFERRED MATERIAL: None.

PREVIOUS WORK: Euraxemys has appeared in print as ‘‘primitive pelomedusid turtle’’ in Gaffney and Meylan (1991) and as ‘‘FR 4922’’ in Meylan (1996). The shell as illustrated (Gaffney and Meylan, 1991: 326) has errors that are corrected in the figure presented here (fig. 254). Meylan (1996) had FR 4922 in his pleurodire dataset and concluded that it was the sister group to Araripemys , together making up the Araripemydidae .

DISCUSSION: The present analysis argues that Euraxemys and its sister taxon, Dirqadim , are related to the superfamily Podocnemidoidea , not to Araripemys , as proposed by Meylan (1996).