Compatibility benefits of social and extra-pair mate choice in the zebra finch

Behavioural ecologists aim at providing insights into the evolutionary and ecological processes that shape animal behaviour. Mate choice is a decision faced by most animals that can strongly affect an individual’s reproductive success, an important fitness component. This behaviour has therefore the potential to show many adaptations which have been the subject of a vivid research interest over the last decades. Studies on mate choice have typically focused on female preferences for traits that increase a male's overall attractiveness, which supposedly reflects the male’s absolute quality. Preferences for such traits are expected to provide females with benefits such as good paternal care or ‘good genes’ for their offspring. Nevertheless, in some species, individuals actually show little consensus on what represents a high-quality partner. Such individually-specific preferences are poorly understood, but it has been hypothesized that they have evolved to optimise genetic compatibility (to increase offspring viability) or, although rarely mentioned, to optimise behavioural compatibility (to facilitate joint parental care). The zebra finch is a life-time monogamous passerine whose mating preferences for putative quality traits have been widely studied but where no such quality indicator has been repeatedly shown to influence mating decisions. In this species, preferences seem instead largely idiosyncratic. In this thesis, I investigated whether female zebra finches choose genetically and/or behaviourally compatible social mates (Chapters 2 and 3). More precisely, I measured, in a large-scale breeding design, the fitness consequences of free mate choice for compatibility while experimentally controlling for effects of overall quality (Chapter 3). In this system, genetic incompatibility results in embryo mortality and therefore hatching failure, while behavioural incompatibility results in offspring mortality. I therefore expected social mate choice for genetic and/or behavioural compatibility to optimise embryo and offspring survival (Chapter 3). Moreover, I tested whether siblings who grew up together avoid choosing each other as social partner; in other words, if they avoid inbreeding depression by choosing a genetically more compatible partner (Chapter 2). In addition, I performed a meta-analysis of published experiments on zebra finches that allow for different mechanisms of kin discrimination to take place (Chapter 2). Finally, I investigated whether females enhance their extra-pair behaviour when paired to an apparently genetically incompatible male to obtain compatible genes benefits (Chapter 1). Indeed, extra-pair behaviour is largely hypothesized to be an adaptive response that would allow females to compensate for a potentially sub-optimal social mate choice. Specifically, I tested whether female extra-pair mating evolved as a counter strategy when females experience low hatching success with their social partner, that is to say whether female extra-pair mate choice targets fertility benefits and/or compatible genes benefits (Chapter 1). Contrary to all expectations derived from optimality, zebra finches were unable to identify partners with whom they would minimise embryo mortality (Chapter 3), or to recognize unfamiliar kin on the basis of genetic similarity per se to minimize inbreeding depression (meta-analysis of Chapter 2), and did not adjust their extra-pair mating behaviour in response to repeated hatching failure (Chapter 1). This suggests that, in zebra finches, individuals have not evolved any ability (other than avoiding familiar kin, experiment of Chapter 2) to judge genetic compatibility of any kind, despite the adaptive value of such behaviour. Finally, allowing free social mate choice did enhance pair fitness due to direct compatibility benefits (Chapter 3). This thesis provides therefore the first evidence of mate choice for behavioural compatibility. Nevertheless, the adaptive significance of the underlying choosiness remains speculative (Chapter 3). A null model consisting of random non-adaptive genetic variation in sensory systems and variation in phenotypes that are better at stimulating some sensory systems than others might well account for the observed fitness differences between compatible and incompatible pairs. This study highlights that there are limits to adaptation