Comparative variation of morphological and molecular evolution through geologic time: 28S ribosomal RNA versus morphology in echinoids.
Résumé
The comparatively good fossil record of post-Palaeozoic echinoids allows rates of morphological change to be estimated over the past 260 million years and compared with rates of molecular evolution. Parsimony analysis of morphological data, based predominantly on skeletal characteristics, and parsimony, distance and maximum likelihood analyses of molecular data, from the first 380 bases from the 5' end of the 28S rRNA molecule, for 10 species of echinoid produce congruent phylogenies. The molecular sequence chosen is demonstrably far from saturation and sister groups have divergence times ranging from about 15 to 260 Ma. Parsimony analysis allows the great majority of molecular and morphological apomorphies to be placed in one of 18 independent geological time intervals, providing a direct measure of rates of evolution for periods in the geological past. Because most molecular fixed point mutations in our sequences cannot be polarized unambiguously by outgroup comparison (making the outgroup states effectively random), distance and parsimony analyses both tend spuriously to root the echinoid tree on the longest internal branch. A topology identical to that derived from morphological data is, however, obtained using Maximum Likelihood and also parsimony analysis where outgroup rooting is restricted to more conserved regions. This is taken as the correct topology for assessing rates of evolution. Overall, both morphological and molecular changes show a moderately strong correlation with time elapsed, but a weaker correlation with one another. Statistically significant differences in evolutionary rate are found between some, but not all, pair-wise comparisons of sister lineages for both molecular and morphological data. The molecular clock rate for echinaceans is three times faster than that for cidaroids and irregular echinoids. Spearman's rank correlation test, which requires only relative magnitude of changes to be known, suggests that morphological change has a slightly better correlation with time than does molecular change, averaged over all ten species. However, when just echinaceans are considered an extremely good correlation is found between the number of molecular changes and time elapsed, whereas morphological change remains poorly correlated. Thus, molecular rates approximate to a clocklike model within restricted echinoid clades, but vary significantly between clades. Averaging results over all echinoids produces a correlation that is no better than the correlation between morphological change and time elapsed.