Evolution and community assembly in the genus Protea: Insights from genetics and morphology

The evolution and maintenance of the biological diversity of the Cape Floristic Region has long intrigued biologists. Much its diversity is due to evolutionary radiations, but the extent to which those radiations are adaptive has not been explored. Here I investigate the role that adaptation has played in the diversification of the white proteas (Protea section Exsertae), a recently radiating clade from two angles: population genetics and morphological diversity. I first used population genetic techniques to determine if the radiation demonstrates the burst of speciation associated with adaptive radiations. Inferences based on 10 microsatellite loci reveal that while the white proteas radiated recently they did not radiate as rapidly as expected if adaptive differentiation is driving speciation. In addition, I found evidence that there is little gene flow between extant populations of most species. ^ Second, I investigated the degree to which the species were adaptively differentiated from each other or whether local adaptation was driving differentiation. Using a common garden experiment I found associations between both species mean environment and the local environment and plant traits. These results suggest that while plant traits are related to the species mean environment, there are more meaningful relationships with the population level environment. The genetics and morphology together demonstrate that within the white proteas the processes underlying the radiation are different from those envisioned by current models of adaptive radiation and that geographic isolation has played a large role in the diversification of the group.^ Lastly, I investigated the ecological processes that structure Protea communities. I estimated the influence of similarity in both functional traits and phylogenetic distance on the co-occurrence of species and repeated this analysis using larger regions and species pools. I found that similarity in some traits is positively correlated with species co-occurrence while similarity in others is negatively correlated with species co-occurrence, and closely related species were more likely to co-occur than we expect. This suggests that both habitat filtering and niche partitioning are acting, but on different traits. We found that increasing the size of the species pool had no effect on the patterns that we detected.