Crocodylus madagascariensis, GRANDIDIER, 1872

Brochu, Christopher A., 2007, Morphology, relationships, and biogeographical significance of an extinct horned crocodile (Crocodylia, Crocodylidae) from the Quaternary of Madagascar, Zoological Journal of the Linnean Society 150 (4), pp. 835-863 : 836-839

CROCODYLUS MADAGASCARIENSIS GRANDIDIER, 1872

Grandidier (1872) and Gray (1874) independently based C. madagascariensis on recent material. Neither author specified a type. Grandidier (1872) did not include a figure, but Gray (1874) illustrated the material at his disposal, which is now housed at the Natural History Museum, London. Both descriptions suggest a rather gracile animal, with proportions much like those of C. niloticus .

Evaluation of the name is difficult because C. madagascariensis is a nomen nudum regardless of which description is used. Nevertheless, the animal figured by Gray (1874) is indistinguishable from C. niloticus currently found in the Malagasy Republic. The material described by Grandidier (1872) is also consistent with the living form. Boulenger (1889) and most subsequent authors regarded C. madagascariensis as a junior subjective synonym of C. niloticus .

Living crocodiles in Madagascar were later treated as a distinct subspecies ( C. niloticus madagascariensis ) based on scalation features said to distinguish Malagasy populations from those on the mainland ( Fuchs et al., 1974a). Recent molecular work has found minimal genetic divergence between Malagasy and eastern African populations of C. niloticus ( Schmitz et al., 2003; Hekkala, 2004), suggesting modern populations in Madagascar were established relatively recently.

CROCODYLUS ROBUSTUS GRANDIDIER & VAILLANT, 1872

Crocodylus robustus was established on the basis of fragmentary subfossil remains from Amboulisatre. No figures were included, and the material is probably now lost (see below). These authors described a rather broad-snouted crocodile with more ‘robust’ proportions than those of extant C. niloticus . In particular, they noted the wide angle of the mandibular symphysis relative to C. niloticus and the dorsoventral depth of the snout, which prevented extrusion of the dentary teeth through the premaxilla in mature specimens.

Grandidier & Vaillant (1872: 150) apparently did not have complete skulls or associated postcrania, but felt that some of the elements ‘came from the most important parts of the skeleton to permit a complete understanding of the animal’. They listed six dentaries (right and left), two premaxillae, two maxillae, and among ‘numerous other cranial bones’, two ‘mastoids’ (squamosals) and some frontals. They also had three ilia, a fragmentary ischium, ‘more than 40’ vertebrae, and numerous osteoderms. No associations between specific elements were implied, but they stated that ‘at least three individuals’ were represented.

The diagnostic characters listed by Grandidier & Vaillant (1872) relate to robust build. The teeth and alveoli were described as ‘enormous’, and the snout as ‘short’ with a symphyseal angle of 49° and a large distance between the lateral borders of the maxillae. References were made, albeit not directly, to wide premaxillae and exclusion of the nasals from the external naris. Finally, they stated that the premaxillae lacked holes for reception of the anterior dentary teeth, a character known to vary ontogenetically in most crocodylians ( Kälin, 1933).

The new crocodile was classified as Crocodylus on the basis of three characters. First, the enlarged fourth dentary tooth occluded in a notch between the premaxilla and maxilla, a character presently regarded as plesiomorphic at the level of Crocodylia (Willis, 1993; Brochu, 1999). Second, ‘the frontoparietal opening [supratemporal fenestra], judging from the curvature of the internal border of the squamosal, was largely open’. Because it is unclear what the authors meant, it is difficult to compare this statement with the markedly constricted supratemporal fenestrae of material described here (see below). Finally, there were 19 upper and 15 lower teeth, a formula consistent with Crocodylus ( Iordansky, 1973) .

The earliest illustration of a fossil referred to C. robustus was a skull figured by Barbour (1918: plate 1; Fig. 1 ). Mook (1921) republished the same plate and added a more detailed description. The specimen, MCZ 1006, is consistent with what Grandidier and Vaillant wrote: the bones of the rostrum (especially the premaxillae) are broad, and although the nasals contribute to the narial rim ( Mook, 1921; see below), their contribution is minimal, and it is easy to envision fragmentary material leading one to conclude that the premaxillae completely surround the nares.

This taxon differs from C. niloticus in numerous ways, the most obvious being the prominent squamosal ‘horns’ ( Fig. 2 ). The squamosals generally become upturned late in ontogeny in most species of Crocodylus , most notably in C. rhombifer from Cuba and in C. siamensis from south-eastern Asia and Indonesia, but they are unusually tall in the extinct Madagascar form and are present in all known mature skulls. The only other crocodylian known to show such horns is the Palaeocene alligatoroid Ceratosuchus burdoshi ( Schmidt, 1938; Bartels, 1984). Grandidier & Vaillant (1872) had isolated squamosals at their disposal, but they did not describe the presence of horns.

The skull of the extinct crocodile is broader and more robust than most C. niloticus and comes closest, in overall shape, to C. palustris . Snout shape varies widely within crocodylian species (e.g. Kälin, 1933, 1936; Hall, 1985; Hall & Portier, 1994), so much so that taxonomic judgements should be based on more than skull proportion. The skull of the extinct form is also dorsoventrally deep ( Fig. 2 ) to a greater extent than in any living species of Crocodylus , but similar to that in Osteolaemus .

Designation of a neotype: At the present time, C. robustus Grandidier & Vaillant, 1872 is a nomen nudum. The name has been used in various contexts (e.g. Bartels, 1984; Burness, Diamond & Flannery, 2001; Meers, 2002), and whatever its status from the standpoint of the ICZN, it has consistently referred to a recently extinct horned crocodile from Madagascar for the past several decades. The name has achieved a stable meaning. Conservation of C. robustus through designation of a neotype would preserve this stability.

A lectotype cannot be designated because the original material either no longer exists or cannot be identified from existing collections. Subfossil crocodylian specimens from Amboulisatre currently housed at the MNHN in Paris are congruent with the material described by Grandidier and Vaillant (C. A. Brochu, pers. observ.), but were collected after 1900. The original material might have been deposited in the Malagasy Academy in Antananarivo, but the building housing most of the Academy’s collection was destroyed by fire in 1995, along with most of the collections (S. Goodman, pers. commun.) .

Conservation of the name would promote clarification of the taxonomy of Quaternary Malagasy crocodiles. C. robustus has frequently been considered synonymous with C. niloticus (e.g. Boettger, 1913; Blanc, 1972, 1984; Fuchs et al., 1974a; Paulian, 1984). Reports of living C. robustus ( Vaillant, 1883; Vaillant & Grandidier, 1910) that later turned out to be very old C. niloticus ( Barbour, 1918) may have reinforced this view. It may also reflect confusion between C. robustus and C. madagascariensis .

The skull figured by Barbour (1918) and Mook (1921), MCZ 1006, is a good candidate for a neotype. It is clearly not C. niloticus – the prominent squamosal horns and anterior extent of the nasals are very different from C. niloticus , and the fact that other Malagasy subfossil skulls show this same combination of features argues against pathology. Concordance between what Grandidier & Vaillant (1872) described and what Barbour (1918) and Mook (1921) figured strongly suggests that they are conspecific.

The only conflict between the original description and the material described here concerns the supratemporal fenestra. It is unclear whether Grandidier & Vaillant (1872) simply meant that the fenestrae were open or that they were relatively large or wide; the former is common to all crocodylians except mature caimans and some Osteolaemus , and fossils later described as C. robustus actually have comparatively small fenestrae with caiman-like constricted dorsal rims. Comparatively large supratemporal fenestrae are more characteristic of longirostrine crocodyliforms ( Langston, 1973), and C. robustus was not longirostrine. Constriction of the fenestrae may not have been apparent from isolated squamosals – the anteromedial corner of each squamosal is concave, forming part of the margin of the fenestra, which might have led Grandidier and Vaillant to conclude that the Malagasy form had open supratemporal fenestrae in comparison with the much more constricted features in Osteolaemus .

Figures

Figure 1. MCZ 1006, neotype, Voay robustus, dorsal (A, C) and ventral (B, D) views. Scale = 5 cm.

Figure 2. AMNH 3101, Voay robustus, skull, right lateral view. Scale = 5 cm.