Djourabus dabba, Peigné & Bonis & Likius & Mackaye & Vignaud & Brunet, 2008

Peigné, Stéphane, Bonis, Louis De, Likius, Andossa, Mackaye, Hassane Taïsso, Vignaud, Patrick & Brunet, Michel, 2008, Late Miocene Carnivora from Chad: Lutrinae (Mustelidae), Zoological Journal of the Linnean Society 152 (4), pp. 793-846 : 834-835

DJOURABUS DABBA SP. NOV.

Holotype: TM 293-01 - 006 , fragment of right hemimandible with m1, and TM 293-01 - 053 , anterior part of the mandible with left canine (broken), most probably belonging to the same individual ( Fig. 8 , Tables 1, 2).

Etymology: The specific name comes from ‘dabba’ meaning ‘thick’ in the Goran language, in reference to the great width of m1.

Referred material: Only the holotype.

Age and distribution: TM 293 site, Toros-Menalla, late Miocene, Chad.

Description

Mandible ( Fig. 8 , Table 1): We did not find an anatomical connection between the anterior fragment of the mandible and the fragment of the right corpus bearing m1 because a short piece of the right hemimandible is missing. The mandible is stout and very thick anteriorly; it is narrower posteriorly and very tall. The symphysis is tall and short, and has ventral processes as in Sivaonyx species. There is one large mental foramen on the left side below the p2-3 diastema and two slightly smaller ones on the right side in the same position.

Dentition ( Fig. 8 , Table 2): The canine tip is broken and/or strongly worn and the enamel is poorly preserved; the tooth section is elliptical. A 4.5-mm-long diastema separates the canine from the first cheek tooth. The latter is identified as a single-rooted p2. The presence of a p1 is rare in Lutrinae. When the tooth is present as may occur in species of Lutra , Aonyx and Sivaonyx , it is a minute tooth and its alveolus is very close to or coalescent with that of the canine. The alveolus of p2 is 8.5 mm long and about 7 mm in width. The p3 was double-rooted and obliquely set in the mandible, with the mesial root labial. The mandibular part bearing p4 is not preserved. The m1 is very broad and short; its mesiolabial face is obliquely orientated due to the transverse orientation of the paraconid. The trigonid cuspids are heavily worn and clearly bunodont. There are no distinct crests on the cuspids, which are inflated, especially the lingual face of the paraconid, metaconid and entoconid and the labial face of the protoconid and hypoconid. The trigonid cuspids form a fairly closed triangle and the trigonid basin is small, shallow and widely opened distally. The paraconid is very strong, much stronger than the metaconid, from which it is separated by a V-shaped valley, the bases of the cusps being closely applied to each other; its mesial face is nearly vertical in lingual view and its mesial cingulid is poorly developed. The protoconid is wide and short; it is orientated approximately sagittally; its distal face is strongly oblique. The metaconid is a well-individualized cusp, slightly distal to and widely separated from the protoconid, closer to the paraconid; the cross-section of the cusp is nearly rounded. The talonid is short, slightly narrower than the trigonid and has a distally squared-off occlusal outline. It is molarized, with each distal corner of the talonid being occupied by a low blunt cuspid, an entoconid lingually and a slightly larger hypoconid labially; distally, the two cuspids are separated by a shallow notch. The hypoconid crest is poorly developed and joins the distal face of the protoconid mesially. The talonid basin is shallow and in the centre of the talonid. The cingulid is present only around the mesial base of the paraconid.

Comparisons and discussion

Compared with that of Sivaonyx species , the mandible of TM 293-05-053 is stouter, thicker anteriorly and taller posteriorly ( Table 1). The m1 is proportionally much broader than in the species of Sivaonyx , especially the trigonid, which is broader than the talonid; the width of the lower carnassial in S. africanus is not known, however ( Table 2). In TM 293-05- 006, cuspids are more bunodont, in particular the metaconid, which is more distant from the protoconid and closer to the paraconid than in Sivaonyx spp. The paraconid and protoconid shelves are shorter and the paraconid is much more transversely orientated than in the carnassial of Sivaonyx . The talonid of m1 is more molarized, shorter and narrower, with a squared-off distal margin. The cingulid is much more reduced than in species of Sivaonyx , and present only on the mesial face of the paraconid; in S. africanus , where the labial part is not preserved, this character remains to be confirmed.

Species with a more bunodont dentition belong to the genera Enhydritherium and Enhydriodon . Compared with Enhydritherium terraenovae from the late Miocene – early Pliocene of North America, our material differs in its larger size (m1 length range in E. terraenovae is 15.8 to c. 18; Berta & Morgan, 1985), taller mandible, its m1, which is relatively broader due to the transverse orientation of the paraconid, trigonid that is broader than the talonid (the opposite is true in E. terraenovae ), lesser enlargement of the metaconid, which remains smaller than the protoconid in Djourabus dabba , and its reduced labial cingulid. We find, however, some similarities, such as the squared-off distal margin of the talonid and the bunodonty.

The comparison between our material and that of Enhydriodon (sensu Morales & Pickford, 2005) is more problematic. As Morales & Pickford (2005) pointed out, there has been much confusion about this genus, notably because the two known species (the type species E. sivalensis and E. falconeri ) are based on the upper dentition. The lower dentition of Enhydriodon remains mostly unknown. An isolated m1 from the Siwaliks was assigned to E. sp. cf. E. falconeri by Pilgrim (1931: pl. 2, fig. 5) then to E. falconeri ( Pilgrim, 1932: pl. 2, fig. 15); Pilgrim (1932: 85) also states that he mentioned this tooth under E. sp. aff. E. sivalensis in one of his previous papers, i.e. Pilgrim (1913: 282). This tooth ( BMNH-M 12350, cast of GSI D161) was more recently assigned to E. sivalensis on the basis of its large size ( Repenning, 1976) but more recently authors have assigned it either to E. sivalensis (e.g. Morales & Pickford, 2005: fig. 4) or E. cf. falconeri (e.g. Willemsen, 1999: table 3). Lydekker (1884: pl. 45, fig. 3) assigned to E. sivalensis a second fragment of mandible with the distal part of m1, which is housed in the Ipswich Museum. Both specimens undoubtedly differ from our material. The isolated m1 ( BMNH-M 12350) is markedly more elongated than TM 293-01-006; the cuspids are less bunodont, the paraconid is much narrower and in a less transverse orientation, and the metaconid is more widely separated from the paraconid and more closely connected to the protoconid by crests. Finally, on both BMNH-M 12350 and the Ipswich specimen the talonid is basined and its distal margin has a curved occlusal outline, which clearly contrasts with the squared-off distal margin of the poorly basined, molarized talonid of TM 293-05-006. Despite the clear distinction between our material and the lower teeth previously assigned to Enhydriodon , there is no guarantee that these m1s effectively belong to that genus. As pointed out by Pilgrim (1932), BMNH-M 12350 is not very different from the m1 of Sivaonyx bathygnathus . In addition, the stratigraphical origin of the material of Enhydriodon is unclear. The material assigned to the genus comes from ‘Siwalik Hills’, but the precise locality is unknown. BMNH-M 12350 comes from around Hasnot, in the Potwar Plateau ( Pakistan) and was considered to come from the ‘Dhok Pathan Stage’, which does not provide much stratigraphic information (see comparisons of Sivaonyx beyi with Sivaonyx bathygnathus ); in the Potwar sequence, the Dhok Pathan Formation spans nearly 7 Myr (10.1 to c. 3.5 Ma; Barry et al., 2002). Recent discoveries report the presence of Enhydriodon sivalensis from the late Pliocene of India ( Verma & Gupta, 1992); the material previously listed as Enhydriodon sp. from the Hexaprotodon biostratigraphic zone in the Potwar plateau ( Pilbeam et al., 1979) is probably not an otter at all (J. Barry, pers. comm.). With an age of c. 7 Ma ( Brunet et al., 2004), the material from Chad is markedly earlier than the single known, well-dated Enhydriodon material. This, and the differences mentioned above, support the generic distinction of TM 293-05-006 and -053. It cannot be ruled out, however, that this material may belong to a form close to Enhydriodon , which can only be tested when we find associated upper and lower teeth of this animal.

Figures

Figure 8. Djourabus dabba gen. nov. sp. nov., photographs of the dentition. TM 293-01-006, holotype: fragment of right hemimandible with m1 in A, labial view, B, dorsal view, and C, lingual view; anterior part of the mandible in D, right lateral view and E, dorsal view. Scale bar = 10 mm.