Meroles revisited: complementary systematic inference from additional mitochondrial genes and complete taxon sampling of southern Africa's desert lizards.

The desert lizards (Meroles: Lacertidae) form a small clade of ground-dwelling taxa inhabiting arid southwest Africa (Arnold, 1991). All seven species have some portion of their range in the Namib Desert, and four of the seven are Namib endemics. Their close geographic ties to the Namib are hypothesized to reflect an interesting pattern of speciation, one in which the lineage infiltrated progressively extreme desert habitats through successive rounds of adaptation (Arnold, 1981, 1990, 1991). Ecological competition plays a key role in Arnold s (1981) speciation model, establishing a sequence of displacement, subsequent adaptation, and species formation. In this iterative process, the lineage s inchoate species came to occupy increasingly xeric environments, culminating in successful entry to aeolian dunes of the Namib s vast ‘‘sand seas.’’ A phylogenetic tree reflecting this iterative mode of speciation should yield a pectinate topology, and, indeed, Arnold (1991) generated just such a tree for Meroles, one fully resolved with little character conflict, in a parsimony analysis of morphological data. To test Arnold s phylogenetic estimate and address independently its progression of morphological adaptation, Harris et al. (1998) conducted a molecular phylogenetic analysis of the desert lizards, using sequence data from the mitochondrial 12S and 16S ribosomal genes. Their molecular phylogeny corroborated the monophyly of Meroles, with its inclusion of the formerly monotypic Aporosaura anchietae [transferred to Meroles by Arnold