The scincid lizard genus Mabuya currently contains approximately 104 described and valid species (pers. obs.). It ranges from southeast Asia west through Asia, the Seychelles, Madagascar, Africa, the Cape Verde Islands, and into tropical America. In many areas, the species are a common and conspicuous component of the herpetofauna. However, the species taxonomy of Mabuya can be difficult. For example, keys often come down to partially overlapping characters such as the number of longitudinal scale rows at midbody, the number of subdigital lamellae, the number or the degree of development of scale keels, and subtle aspects of color pattern. Few unequivocal characters of squamation are available. Given the difficult alpha taxonomy of the group, it is not surprising that there has been no attempt to analyze the species of Mabuya phylogenetically. Therefore, any character that helps to distinguish unequivocally groups within Mabuya is potentially important from the point of view both of species taxonomy and phylogenetics. This note describes such a character in the squamation of the temporal region.

Materials and Methods

The squamation in the temporal region was examined in specimens of species of Mabuya in collections of the Australian Museum, Sydney, and the University of Michigan, Ann Arbor. In addition, the literature was searched for good illustrations clearly showing the temporal region. Character-state polarities were determined by reference to the morphologically primitive scincid genus Eumeces (pers. obs.), especially the subgenus Pariocela. Institutional abbreviations are as listed in Leviton et al. (1985).

Results and Discussion

In most skinks, including Eumeces, the lateral edge of the parietal overlaps the upper edge of the upper anterior temporal scale lying just posterior to the two pretemporals (upper and lower). In Mabuya, this pattern (Fig. 1A) appears to occur in most, if not all, species from the New World, Cape Verdes, and the Seychelles, and in most species from mainland Africa (Appendix). In contrast, in some species of Mabuya, the parietal is itself overlapped by the upper edge of the upper anterior temporal scale (Fig. 1B). This derived pattern occurs in most, if not all, southeast Asian species, in at least one south Asian species, in all endemic Madagascan species, and in a few mainland African species (Appendix).

View Figure

• Download as Powerpoint
• Download Figure

One feature that can make the homology of the upper anterior temporal scale initially confusing is a tendency for the lower pretemporal to be reduced in size and/or displaced slightly ventrally below the upper pretemporal, thereby losing its contact with the parietal (Fig. 1C). When this occurs, the upper anterior temporal could be misinterpreted as a posteriorly displaced lower pretemporal (Fig. 1D). However, this interpretation seems unlikely, because it would require the precise replacement of the upper anterior temporal by the lower pretemporal because there is no evidence of a posteriorly displaced upper anterior temporal. Furthermore, the posterior displacement of the lower pretemporal and the loss of the upper anterior temporal is a less parsimonious interpretation than the reduction in size and perhaps slight ventral displacement of the lower pretemporal only. In addition, there is no evidence in any species of Mabuya of a morphocline indicating a posterior displacement of the lower pretemporal, but there is evidence of a morphocline for the reduction and ventral displacement for the lower pretemporal (e.g., M. carinata and an unidentified species 1 in Appendix) from Sri Lanka. Parenthetically, some species of Mabuya seem to have the lower pretemporal consistently squeezed down to a position directly below the upper pretemporal and separated from the parietal, for example M. boulengeri (n = 5; Fig. 1C) and possibly Sri Lankan species 2 (n = 3). Therefore, this character may also be useful in Mabuya (below).

A few species of Mabuya show variation with regard to the two character states in the parietal/anterior upper temporal overlap pattern. The two states can vary both among and within specimens (Table 1). The two states can even vary within one scale, that is, the overlap pattern can change along the common suture between the parietal and the upper anterior temporal, for example, M. maculilabris (one side in AMS R 106885), and M. multifasciata (one side in AMS R 92626). Indeed, a “cross-over” pattern appears to be common, if not typical, in M. longicaudata (Fig. 1E). In nine specimens, the cross-over pattern occurs bilaterally in seven and unilaterally in one with the other side showing the primitive condition; the remaining specimen has the derived condition bilaterally. In all cases among the three above-mentioned species showing the cross-over pattern, the primitive overlap pattern lies anterior to the cross-over point, and the derived pattern lies posterior. In general, in those species showing any variation in this character, there is usually a very clear modal condition (Table 1), and this condition has been used to characterize the species (Appendix). We include M. longicaudata among the group showing the derived condition.

Table 1. Intraspecific Variation in the Parietal–Upper Anterior Temporal Overlay Pattern in Species of Mabuya. The number of specimens in which the primitive or derived condition occurs on both sides or in which one condition occurs on one side and the other condition occurs on the other side (“mixed”)

View Fullscreen

There are two taxonomic/zoogeographic problems in which this new character may be of immediate use. The first problem is the identity of Mabuya betsileana described by Mocquard (1906). This species is known only from the holotype said to have come from Betafo, province du Betsiléo, Madagascar. However, it has been suggested that the holotype was a misidentified specimen of M. perrotetii from mainland Africa (see Brygoo, 1983). In that all the endemic Madagascan species of Mabuya have the derived state of the upper anterior temporal overlapping the parietal, the condition of the supposed endemic M. betsileana is of interest. In fact, it has the primitive condition of the parietal overlapping the upper anterior temporal, as does M. perrotetii. This observation supports the possibility that M. betsileana is a junior synonym of M. perrotetii instead of being a valid Madagascan endemic.

The second problem in which the character may be useful is the relationships of Mabuya ferrarai, known from two specimens from Somalia (Lanza, 1978). This species was originally compared with only two other species, M. boulengeri from southeastern Africa and M. maculilabris from central Africa. It was thought to be most closely related to the latter. However, M. ferrarai shares the derived anterior upper temporal overlap pattern with M. boulengeri as well as a ventrally displaced lower pretemporal (Lanza, 1978:fig. 3B, the small unlabelled scale just below the scale labeled “s,” the upper pretemporal). Both features are unique among species of Mabuya on the African mainland. This suggests that M. boulengeri and M. ferrarai may be each other's closest living relatives.

In summary, the parietal/anterior upper temporal overlap pattern described here promises to be helpful both in identifying species of Mabuya and in preparing regional and general keys. Furthermore, the broad geographic concordance between the states of the character and geography (see Appendix) suggests that it will also be useful in the future phylogenetic analysis of this large and widespread genus.