Models of Mating System Evolution: Gamete Competition, Hermaphroditism and Sexual Selection

Abstract

To understand how mating systems evolve, we depend on both (i) theoretical explanations and predictions, supported by mathematical modelling, and (ii) quantitative tools to test predictions rigorously. This thesis is divided equally between these two aims. The first three papers explore the evolution of mating systems using analytic and simulation models. I begin by considering a long-standing puzzle in marine invertebrate systems: the widespread association between a species' body size, whether it is hermaphroditic or has separate sexes, and its mode of fertilisation (i.e. whether eggs and sperm are released into the water or retained by the adult until fertilisation). I argue that local competition among eggs for fertilisation can explain these patterns, which arise in taxa as diverse as sea stars, corals and polychaete worms (Paper 1). I then turn to egg trading – the alternating exchange of egg parcels during mating by simultaneous hermaphrodites – which is one of the best-supported cases of reciprocity between non-relatives. I show that egg trading is under positive frequency-dependent selection and should evolve most easily when potential mates are encountered frequently (Paper 2). Once evolved, egg trading selects for female-biased sex allocation. I explain how this bias allows simultaneous hermaphroditism to persist stably, even in motile species living at high population densities, where simultaneous hermaphroditism is otherwise predicted to be unstable (Paper 3). These three papers contribute to the resolution of Williams' paradox – the mismatch between the empirical distribution of hermaphroditism and our theoretical expectations – by providing finer-grained predictions for when hermaphroditism should occur. The last three papers provide quantitative tools for the measurement of natural (and particularly sexual) selection. I argue that pre-mating sexual selection should be understood as a two-step causal process: traits affect mating success, which in turn affects reproductive success. Most previous work has focussed too narrowly on one or the other of these steps, leading to confusion about what sexual selection is and how we should measure it. I provide a new statistical framework that integrates both steps in the sexual selection pathway, and includes a third path whereby traits directly affect reproductive success. This leads to better estimates of the strength of sexual selection on traits (Paper 4). I then compare various indices of sexual selection, often used in comparative work, that do not rely on trait measurements. I show that the recently defined Jones index outperforms all others in predicting the actual strength of sexual selection, because it accounts for both steps in the sexual selection pathway (Paper 5). Lastly, I provide a new way to quantify the total selection acting on a trait, including both directional and non-directional selection (Paper 6). Show more