Testing macroevolutionary hypotheses: diversification and phylogenetic implications
MetadataShow full item record
This dissertation research tests macroevolutionary hypotheses in a large vertebrate group (Reptilia: Serpentes), and detects methodological problems of using molecular data to test macroevolutionary hypotheses. The first chapter examines lineage diversification of snakes. It is hypothesized that snake evolution underwent major lineage diversification after gaining the ability to swallow large prey and again in the Miocene during the so-called "age of snakes". I use two new methods and a time-tree to test these hypotheses. I find no evidence for a diversification rate shift following the evolution of large gape size. In contrast, Neogene snake diversification was significantly higher than over any other geologic time interval in snake evolutionary history. This elevated diversification appears to be driven by the evolutionary radiation of the colubroids, which exhibit diversification rates 2.5 times greater than other related snake lineages. The second chapter examines mimicry of head shapes between slug-eating, false-viper, and viper snakes. Viper mimicry is hypothesized to have produced fascinating mimics, which might have led to increased diversification. I provide the first quantitative test of viper mimicry. I use traditional tests, the latest morphological model-fitting approaches and a new time tree to test this hypothesis. I find that head shape has evolved towards similar selective optima, when specific viper and false-viper clades are included in the models. This suggests that the central adaptive hypothesis of mimicry between slug-eaters and viperids likely occurred. The third chapter surveys effects of using incomplete molecular data matrices to infer phylogenetic tree branches of a rapidly diversifying group. I simulate phylogenies with increasing missing-data levels that are deleted throughout the tree randomly or taxonomically. I compare simulated-tree branch lengths to true-tree branch lengths using a geometric trees-space method and a commonly used macroevolutionary measure that detects node distributions as a proxy for slowdowns in diversification rates. I find the amount of missing data has more impact in branch length estimation than presence of a fast radiating clade. This suggests our current use of molecular evolutionary models are sufficiently powerful to deal with intricacies of adaptive radiations and that missing data should always be accounted for in phylogenetic studies.