Phylogenetic interpretations of macroevolution in deep-time

The fossil record yields information on macroevolutionary patterns that remains inaccessible from the study of extant organisms alone, presenting a natural laboratory for us to test hypotheses about the long-term drivers and processes of evolution. Fossil data are therefore increasingly incorporated into evolutionary analyses, both on their own and in combination with neontological data. Phylogeny (an explicit hypothesis of the evolutionary relationships between taxa) can be used as a framework to enable direct comparison of results of comparative methods across many different timescales and taxa, and is now commonly used in investigations of fossil data. This represents an important step towards a unified approach, however, it is not yet fully understood what the effect of using fossil data is on the results of downstream phylogenetic comparative methods, which were originally developed with only living taxa in mind. In this thesis I explore the validity of phylogenetic interpretations of fossil record data. I begin with only taxonomic classification and show that this can in some cases substitute for a cladistically inferred phylogeny in phylogenetic comparative methods, without biasing results. Moving on to scenarios where a timescaled phylogeny is available I investigate the relationship between phylogeny and extinction in the geological past, show that phylogenetic clustering of extinction was common in tetrapods, and present a summary of the ways in which fossil data biases this measurement. Finally, with timescaled phylogenies and a detailed continuous trait dataset available, I interrogate the fossil record of Sauropterygia to uncover the processes of evolutionary change in this highly labile clade. By comparing the results of a suite of phylogenetic comparative methods I demonstrate that neck length evolved through changing vertebral counts rather than somite growth; that the clade experienced a release in evolutionary constraint at the Triassic-Jurassic boundary; and that evidence does not support evolution towards a stationary adaptive peak as a suitable model for phenotypic change in the clade