Phylogeny and evolution of Lerista (Lygosominae, Scincidae, Squamata).

In this thesis, I investigate the phylogeny and evolution of Lerista, a clade of more than 75 species of scincid lizards, distributed in arid, semi-arid, and seasonally dry habitats throughout Australia. Among extant tetrapods, Lerista is exceptional in comprising a large number of closely-related species displaying prodigious variability of body form; several species possessing well-developed, pentadactyl limbs resemble typical non-fossorial scincids in body proportions, while many other species exhibit varying degrees of limb reduction and body elongation, including two that are highly elongate and entirely limbless. The extensive variation in limb morphology observed among species, incorporating at least 20 distinct phalangeal configurations, has prompted some authors to identify Lerista as the best available model for studying limb reduction in squamates. Nonetheless, lack of a well-resolved phylogeny has impeded investigation of the pattern and mode of limb reduction and loss within the clade. The primary goal of my research was to furnish a comprehensive phylogenetic hypothesis for Lerista, enabling more sophisticated study of the evolution of limb morphology and body form in this clade than has previously been possible. A recent phylogenetic analysis of mitochondrial DNA sequences for a series of Australian Sphenomorphus group scincids (including two species of Lerista) recovered several well-supported, major clades, although these were generally separated by relatively short branches associated with low support values. Applying a recently described methodology for inferring lineage-level polytomies, I use ATP synthetase-β subunit intron sequences and the existing mitochondrial DNA data set (with sequences for additional taxa) to assess the hypothesis that the poorly resolved basal relationships within the Australian Sphenomorphus group are a consequence of the major clades having originated essentially simultaneously. Phylogenetic analyses of the separate mitochondrial DNA and intron sequence data reveal a number of congruent clades, however, the relationships among these clades indicated by the two data sets are generally incongruent. Although this may be partly ascribed in to error in estimating phylogenetic relationships due to insufficient data, some incongruence is evident when uncertainty in inferred relationships is allowed for. Moreover, the congruent clades are typically separated by very short branches, several having a length insignificantly different from zero. These results suggest that initial diversification of Australian Sphenomorphus group scincids was rapid relative to the substitution rates of the mitochondrial DNA and intron fragments considered, if not essentially simultaneous. The pattern and rate of limb reduction in Lerista are investigated, employing a nearly complete phylogeny inferred from nucleotide sequences for a nuclear intron and six mitochondrial genes. Ancestral digit configurations reconstructed assuming this phylogeny indicate at least ten independent reductions in the number of digits from a pentadactyl condition, including four independent losses of all digits, three from pentadactyl or tetradactyl conditions. At the highest rate, complete loss of digits from a pentadactyl condition is estimated to have occurred within no more than 3.6 million years. Patterns of digit loss for the manus and pes are consistent with selection for preserving hindlimb utility as the limbs are reduced, and suggest that intermediate digit configurations exhibited by extant species do not represent transitory stages in a continuing process of limb reduction. An increase in the relative length of the body is demonstrated to precede digit loss in lineages experiencing substantial reduction of the limbs, supporting the hypothesis that limb reduction and loss is a consequence of the adoption of lateral undulation as a significant locomotory mode. However, less extensive limb reduction may proceed in the absence of body elongation, perhaps due to a decrease in absolute body size. The exceptionally high frequency and rate of limb reduction in Lerista emphasise the potential for rapid and dramatic evolutionary transformation of body form in squamates. The substantial divergence of relative limb and body length evident within Lerista is more readily explained by the correlated progression model of phenotypic transformation than the independent blocks model. At each step in the attainment of a limb-reduced, elongate body form, alterations to the relative length of the limbs are accompanied by changes in relative snout-vent length (or vice versa) enabling the maintenance of locomotory ability. Nonetheless, some dissociation of hindlimb reduction and body elongation is possible, emphasising the potentially variable intensity of functional constraints and, accordingly, that the independent blocks model and correlated progression are extremes of a continuum of models (each invoking a different degree of functional integration) and do not describe discrete categories of phenotypic change. An increase in the extent of seasonally dry and arid habitats coincident with the origination of Lerista would have facilitated limb reduction and body elongation by furnishing an environment conducive to the adoption of fossorial habits, however, trends toward a limbless, highly elongate body form may be attributed primarily to the very low probability of re-elaborating reduced limbs. Such asymmetry in the probabilities of possible phenotypic changes may be a significant cause of evolutionary trends resulting in the emergence of higher taxa.Thesis (Ph.D.) -- University of Adelaide, School of Earth and Environmental Sciences, 200