Rapidly evolving traits and the comparative method: how important is testing for phylogenetic signal?

The indiscriminate application of phylogenetically based comparative methods in cross-species correlated evolution analyses has been questioned. It has been argued that traits are not always significantly correlated to their phylogenetic history, and that correcting for phylogeny in these cases may be unnecessary and may even introduce statistical error. Statistical diagnostics to test for phylogenetic signal have recently been presented. Here, we investigate the relationship between habitat and vocal characters within a tropical bird community. We show that the bird song parameters investigated are not correlated to their phylogenetic history, indicating that certain aspects of bird song can be subject to rates of evolution that are much more rapid than speciation events. In contrast, we show that the habitat parameter is significantly correlated to its phylogenetic history. Previous comparative studies, which have not taken phylogenetic signal into account, show a significant association between habitat and song. With respect to our continuous data, we demonstrate that analyses which fail to correct for phylogeny in traits that show phylogenetic signal (i.e. habitat), or those that correct for phylogeny in traits that are phylogenetically independent (i.e. bird song parameters), support an association between habitat and song. Analyses that incorporate tests for phylogenetic signal, however, reject this association, and thus call for a re-evaluation of the evidence on correlated evolution of habitat and bird song. Therefore, using tests for phylogenetic signal before and after conducting a cross-species correlated evolution analysis is crucial to the outcome of a comparative study when analysing rapidly evolving traits.