There is solid theoretical reasoning but few statistical approaches available to test the hypothesis that phenotypic traits and phenotypic evolutionary rates could be associated to rates of taxonomic diversification. A positive relationship between phenotypic and taxonomic evolution is expected to occur during adaptive radiations, or when the acquisition of key innovations allows rapid expansion into new adaptive zones. Such patterns of trait-associated diversification are intensely studied under both neontological and paleontological approaches. However, albeit the inclusion of fossil information is welcome, paleontological studies often lack the phylogenetic dimension, which makes the results potentially depending on untested phylogenetic effects. We devised and applied a novel method seeking to find instances of trait-associated diversification, which is specifically meant to make full use of fossil phenotypes and phylogenies. We performed extensive simulations to test the power of the method and to assess the incidence of false positives (Type I error) and applied the method to test the hypotheses that body size relates to taxonomic diversification in mammals and dinosaurs.

We demonstrated that the new method is powerful, accurate and produces low Type I error. Consistently with previous reports, we found statistical evidence that hadrosaurid dinosaurs experienced a significant positive shift in diversification rates, coupled to low rates of morphological evolution. In mammals, body size is linked to taxonomic diversification by a weak negative relationship.

R codes, and additional results