Abstract/Summary

Computers have fundamentally advanced all fields of biology over the past 50 years. Within evolutionary biology, for example, researchers can now leverage computational approaches to detect macroevolutionary patterns with unprecedented rigour and objectivity using large phylogenies of species. But evolution is a historical process, and macroevolutionary hypotheses demand direct evidence from the deep past. The work presented in this thesis applies the latest phylogenetic comparative methods to fossil and simulated datasets and highlights how fossils from deep time inform our understanding of evolution. Sampling is central to all science, including comparative analyses, and can shape how we detect patterns and processes. Chapter 1 shows that a large effective (evolutionary) sample size of independent changes is crucial for accurately inferring rates of evolution and differentiating support among evolutionary hypotheses. Chapter 2 shows how disproportionate geo�palaeontological sampling affects inferences of dispersal rates and ancestral geographic locations. I offer an innovative approach for assessing geographic sampling biases in the fossil record. The fossil record is also a window into diverse ecosystems and a wide range of global conditions. It thus offers unique and independent data sources for testing evolutionary hypotheses. Chapter 3 describes the utility of fossils for assessing general ecological principles based on extant taxa and modern climates. The study shows that Bergmann's rule does not extend to Mesozoic dinosaurs and mammals and that their poleward dispersal did not drive increases in body size. Chapter 4 describes an approach for directly quantifying change in functional equations along phylogenetic branches and applies it to the locomotor evolution of dinosaurs. The study reveals a close connection between the rate of locomotor evolution and speciation. Together, these works highlight the utility of the fossil record for informing evolutionary models and our understanding of evolution.