THE ADAPTIVE SIGNIFICANCE OF EGG-SIZE VARIATION WITHIN AND AMONG POPULATIONS OF ATLANTIC SALMON

Abstract

This thesis focuses on the classic problem of investment per offspring. It is an attempt to (i) reconcile theoretical research with empirical methods that can be used to test theory, (ii) test a fundamental prediction that arises from classic theory, and (iii) test one of the more recent theoretical developments. We use Atlantic salmon (Salmo salar) as a model organism. Drawing from the classic Smith–Fretwell model, we provide defensible definitions of offspring fitness that can be used in empirical studies, and we show using simulation that the Weibull-1 statistical model provides the best estimates of optimal investment patterns. Next, we apply these methods to mark-recapture data collected for juvenile Atlantic salmon. This experiment supports the prediction that parental reproductive success is maximized by increasing investment per offspring when environmental conditions become unfavourable. Having verified this prediction, we test a general extension of classic theory which broadly suggests that large-bodied females decrease the quality of the offspring environment, such that larger females in a population ought to invest relatively heavily in investment per offspring. This might occur, for example, when larger females have a greater fecundity and if optimal investment per offspring increases with sibling competition among non-dispersive offspring. The results of this experiment generally do not support the idea that large females decrease the quality of the offspring environment in Atlantic salmon. Finally, we also provide evidence against a verbal hypothesis that attempts to explain inter-population variation in egg size of salmonids as an adaptation to population-specific spawning substrates. We conclude that the classic model of egg-size optimization can be a useful tool for understanding patterns of reproductive allocation in nature, but that investment per offspring is an extremely complex trait that cannot be fully understood by invoking a simple optimality model. Variation in investment per offspring, especially that which occurs within populations, is most parsimoniously attributed to the physiological factors (e.g., variation in testosterone levels), morphological constraints (e.g., the size of the pelvic aperture) and genetic factors (e.g., genetic correlations arising from pleiotropic genes) that affect this phenotype and that constrain adaptive evolution of this trait.