Trait lability as a predictor of diversification dynamics in flowering plants

Rates of diversification differ between angiosperm lineages. To date, attempts to explain this heterogeneity have focused on the potential correlation between speciation and extinction rates and particular key traits. However, an often-overlooked explanation is that evolutionary lability, here defined as the rates of trait change, may be a better predictor of speciation and extinction rate heterogeneity than the observed traits themselves. Here, we show how this can be tested by using hidden Markov models (HMMs), which allow for several rate classes associated with speciation, extinction, and transition between trait states across a phylogeny. Using a phylogenetic dataset of 13 angiosperm clades including 10,474 species, we show that higher rates of change between open and closed-canopy biomes is consistently associated with higher lineage turnover rates (speciation + extinction rates) across clades. We demonstrate how HMMs can be leveraged in ways that go beyond their conventional use as null models in diversification analyses, and that comparing different rate classes can unveil novel patterns of biological interest. These patterns result in a shift in focus from static traits to dynamic evolutionary processes and may provide a more comprehensive understanding into how biodiversity is generated and maintained, in angiosperms and other organisms.