The role of habitat shift in the evolution of lizard morphology: evidence from tropical Tropidurus

Abstract

We compared morphology of two geographically close populations of the tropical lizard Tropidurus hispidus to test the hypothesis that habitat structure influences the evolution of morphology and ecology at the population level. T. hispidus isolated on a rock outcrop surrounded by tropical forest use rock crevices for refuge and appear dorsoventrally compressed compared with those in open savanna. A principal components analysis revealed that the populations were differentially distributed along an axis representing primarily three components of shape: body width, body height, and hind-leg length. Morphological divergence was supported by a principal components analysis of size-free morphological variables. Mitochondrial DNA sequences of ATPase 6 indicate that these populations are closely related relative to other T. hispidus, the rock outcrop morphology and ecology are derived within T. hispidus, and morphological and ecological divergence has occurred more rapidly than genetic divergence. This suggests that natural selection can rapidly adjust morphology and ecology in response to a recent history of exposure to habitats differing in structure, a result heretofore implied from comparative studies among lizard species.

Figure 1

Patterns of microhabitat use by two populations of T. hispidus. The number of observations is indicated by n.

Figure 2

Bivariate plot of PCA on morphological characteristics of two T. hispidus populations occupying structurally different habitats showing the first two principal components (PC) of log-transformed morphological variables. R and S designate Roraima rock outcrop and savanna populations, respectively.

Figure 3

Aligned ATPase 6 sequences for T. montanus and four T. hispidus haplotypes. R and S designate Roraima rock outcrop and savanna populations, respectively. Sequences are deposited in the GenBank database (accession nos. U83494–U83498U83494U83495U83496U83497U83498).

Figure 4

Most parsimonious relationships among T. montanus and T. hispidus haplotypes (total number of mutational steps, 127; consistency index for characters informative under parsimony, 0.704). Numbers above internal branches are the proportion of times the branch appeared in 1,000 bootstrap replicates. Numbers below each branch are the minimum and maximum number of mutations occurring along each branch under all possible optimizations. R and S designate Roraima rock outcrop and savanna populations, respectively.