Can functional traits explain phylogenetic signal in the composition of a plant community?

Phylogeny-based and functional trait-based analyses are two principle ways to study community assembly and underlying ecological processes. In principle, knowing all information about species traits would make phylogenetic information redundant, at least that component of phylogenetic signal in the distribution of species among communities that is caused by phylogenetically related species sharing similar traits. In reality, phylogenies may contain more information than a set of singular, discretely measured traits because we cannot measure all species traits and may misjudge which are most important. The extent to which functional trait information makes phylogenetic information redundant, however, has not been explicitly studied with empirical data in community ecology. Here, we use phylogenetic linear mixed models to analyze community assembly of 55 understory plant species in 30 forest sites in central Wisconsin. These communities show strong phylogenetic attraction, yet variation among sites in 20 environmental variables could not account for this pattern. Most of the 15 functional traits we measured had strong phylogenetic signal, but only three varied strongly among sites in ways that affected species abundances. These three traits explained only 19% of variation in phylogenetic patterns of species co-occurrence. Thus, phylogenies appear to provide considerably more information about community assembly than the functional traits measured in this study, demonstrating the value of phylogeny in studying of community assembly processes even with abundant functional traits.