Isolating evolutionary phenomena in analyses of disparity

Abstract

Understanding how the diversity of organismal form came to be is a fundamental outstanding challenge in evolutionary biology. Analyses of this morphological variety (disparity) typically seek to characterise how it changes between groups and through time. However, interpretations of the patterns they produce are confounded by artifacts of sampling, data loss, and experimental design, although how exactly they manifest in analyses of disparity is unclear. Given the tendency for the results of such analyses to be invoked as evidence of major evolutionary events and macroevolutionary rules, it is crucial that this uncertainty be resolved. Simulations offer a means by which the effects of different confounding factors can be isolated in a static evolutionary environment. Despite their obvious utility, simulation-based approaches have seen limited application in studies of disparity to date. This thesis leverages the power of simulations to shed light on the most prevalent sources of systematic error in analyses of disparity. I analyse simulated and empirical data using a variety of distance metrics, ordination methodologies, and quantitative indices to estimate the degree to which cladistic sampling practices bias estimates of disparity. Building on these results, I then systematically degrade empirical and simulated datasets to replicate the effects of biostratinomic processes on morphological data and characterise the resultant changes in disparity using a variety of methods. Finally, I test the generalisability of hypotheses of metazoan morphological evolution by conducting the first analysis of fungal disparity using a bespoke dataset. Not only do these analyses support the idea of a common mode of multicellular body plan evolution across animals and fungi, but they also demonstrate that sampling at different taxonomic levels biases analyses towards different patterns in disparity. Accounting for all potential sources of systematic error is essential if the evolution of organismal disparity is to be understood.

Date of Award	6 Dec 2022
Original language	English
Awarding Institution	University of Bristol
Supervisor	Davide Pisani (Supervisor) & Philip C J Donoghue (Supervisor)