Abstract
Amphidromy is a widespread migratory behavioural syndrome exhibited by fish (and some aquatic invertebrates) that spawn in fresh water and whose larvae migrate to pelagic marine (or lentic) habitats for a period of early growth, followed by a return migration to adult freshwater habitats. The fitness advantage of amphidromy has been the subject of prolonged debate, and we examined the hypothesis that amphidromy mainly increases fecundity through the production of small pelagic larvae. We compared egg size (a proxy for larval size) of closely related non-migratory and amphidromous fish species in the families Cottidae, Galaxiidae, Eleotridae and Gobiidae. To examine how egg size changes in relation to body size within a taxonomic group, we also compared egg size and maximum body size across most species of New Zealand non-migratory and amphidromous galaxiids. Non-migratory species generally have relatively larger eggs than their amphidromous confamilial species. This particular trait has evolved independently several times in each of the four families of amphidromous fish that have given rise to significant freshwater radiations. Amongst the New Zealand galaxiids, mean egg diameter increased with maximum body length for both non-migratory and amphidromous species; however, despite the considerably smaller relative body size of the non-diadromous species, the rate of increase in egg diameter relative to the increase in body size is considerably higher in the non-migratory fish. We propose that amphidromous fish maintain a high level of fecundity by producing small pelagic larvae. In contrast, the relatively large eggs and well-developed larvae of non-migratory species increase larval survival in what are often relatively harsh and unproductive freshwater habitats. Consequently, amphidromous species are likely to have a competitive advantage over their non-migratory relatives when close to a pelagic habitat in which their larvae can grow and develop and then migrate upstream, releasing them from recruitment limitation and giving them a local reproductive advantage over their less fecund non-migratory relatives. We argue that the persistence and distribution of both life-history strategies across the landscape depends on the relative difference in the net reproductive return for each strategy in relation to distance from a pelagic larval habitat, as mediated by the relative costs of migration and egg size/fecundity relationships.