by Thomas F. Hansen
The comparative method can be used to test hypotheses of adaptation by comparing groups of species that meet different adaptive challenges. This requires attention to phylogenetic correlations and to historical lags in achieving adaptation. The modern phylogenetic comparative method has provided some partial solutions to these problems, but the field has also suffered from a systemic lack of demand for biological justifications of its statistical procedures. Consequently, assumptions have been made for statistical convenience and are often inconsistent with the relevant biology. I argue that common comparative tests of adaptation, including Brownian-motion-based phylogenetic linear models and inferred-changes methods based on reconstructing ancestral states, violate essential characteristics of adaptation as a biological process. I discuss the requirements for biologically consistent comparative analysis of adaptation, and I review work towards this goal.