Quantitative tests of general models for the evolution of development.

Comparative developmental biologists have proposed models to describe patterns of conserved features in vertebrate ontogeny. The hourglass model suggests evolutionary change is most difficult at an intermediate "phylotypic" stage, the adaptive penetrance model suggests change is easiest at an intermediate stage, and the early conservation model suggests change is easier later in ontogeny. Although versions of some of these models have been discussed since the nineteenth century, quantitative approaches have been proposed only recently. Here we present quantitative phylogenetic approaches to evaluating trends in the evolution of ontogeny. We apply these approaches to the proposed models and demonstrate that an existing approach to assessing these models is biased. We show that the hourglass, adaptive penetrance, and early conservation models are unnecessarily complex explanations of the patterns observed in developmental event data for 14 species of vertebrates. Rather, a simpler model that postulates that evolutionary change is easier between ontogenetically adjacent events is adequate.