Meeting Abstract

57.5  Tuesday, Jan. 6  Multivariate comparative analysis using OUCH BUTLER, MA*; KING, AA; University of Hawaii; University of Michigan mbutler@hawaii.edu

Correlated evolution, the hypothesis that evolutionary forces are acting on characters jointly, is a major feature of evolutionary theory. For example, whether genome size evolves in correlation with basal metabolic rate is a question which has interested biologists for decades. Adaptation can often be even more complex, involving suites of characters. For example, birds and bats which fly in close quarters amongst the trees have shorter wingspan and lower aspect ratio, which reflects their need for greater maneuverability. Recently, methods have been developed for adaptive evolution which model the evolutionary response of a continouous phenotype in response to multiple (hypothesized) adaptive regimes. These methods often greatly improve the fit of comparative data over pure Brownian motion models. However, they have not been fully extended to the multivariate case. Here we generalize the model-based approach to adaptive evolution (Hansen 1997, Butler and King, 2004) for the bivariate or multivariate case. We illustrate how one may model adaptive evolution by hypothesizing the operation of different selective regimes on various branches of the phylogeny. We illustrate the method by reanalyzing the evolution of genome size and basal metabolic rate in archosaurs, and male and female body size in Caribbean Anolis lizards, and comparing with a univariate analyses.