Predation drives morphological convergence in the Gambusia panuco species group among lotic and lentic habitats

Predation risk as a cost of reproduction in animals has recently received increased empirical and theoretical attention. Higher risk may be associated with all stages of reproduction. Examples of evolutionary responses to this increased risk include habitat choice, duration and timing of display and copulation, changes in brightness of breeding coloration, and changes in life history traits such as age of reproduction and reproductive effort. Copyright © 1991. Published by Elsevier Ltd.

Predation is heterogeneously distributed across space and time, and is presumed to represent a major source of evolutionary diversification. In fishes, fast-starts--sudden, high-energy swimming bursts--are often important in avoiding capture during a predator strike. Thus, in the presence of predators, we might expect evolution of morphological features that facilitate increased fast-start speed. We tested this hypothesis using populations of western mosquitofish (Gambusia affinis) that differed in level of predation by piscivorous fish. Body morphology of G. affinis males, females, and juveniles diverged in a consistent manner between predatory environments. Fish collected from predator populations exhibited a larger caudal region, smaller head, more elongate body, and a posterior, ventral position of the eye relative to fish from predator-free populations. Divergence in body shape largely matched a priori predictions based on biomechanical principles, and was evident across space (multiple populations) and time (multiple years). We measured maximum burst-swimming speed for male mosquitofish and found that individuals from predator populations produced faster bursts than fish from predator-free populations (about 20% faster). Biomechanical models of fish swimming and intrapopulation morphology-speed correlations suggested that body shape differences were largely responsible for enhanced locomotor performance in fish from predator populations. Morphological differences also persisted in offspring raised in a common laboratory environment, suggesting a heritable component to the observed morphological divergence. Taken together, these results strongly support the hypothesis that divergent selection between predator regimes has produced the observed phenotypic differences among populations of G. affinis. Based on biomechanical principles and recent findings in other species, it appears that the general ecomorphological model described in this paper will apply for many aquatic taxa, and provide insight into the role of predators in shaping the body form of prey organisms.

We investigated the interplay between natural selection and gene flow in the adaptive divergence of threespine stickleback (Gasterosteus aculeatus) that reside parapatrically in lakes and streams. Within the Misty Lake system (Vancouver Island, British Columbia), stickleback from the inlet stream (flowing into the lake) have fewer gill rakers and deeper bodies than stickleback from the lake--differences thought to facilitate foraging (benthic macroinvertebrates in the stream vs. zooplankton in the open water of the lake). Common-garden experiments demonstrated that these differences have a genetic basis. Reciprocal transplant enclosure experiments showed that lake and inlet stickleback grow best in their home environments (although differences were subtle and often not significant). Release-recapture experiments in the inlet showed that lake fish are less well-suited than inlet fish for life in the stream (higher mortality or emigration in lake fish). Morphological divergence in the wild and under common rearing was greater between the lake and the inlet than between the lake and the outlet. Genetic divergence (mitochondrial DNA and microsatellites) was greatest between the lake and the upper inlet (1.8 km upstream from the lake), intermediate between the lake and the lower inlet (0.9 km upstream), and least between the lake and the outlet stream (1.2 km downstream). Relative levels of gene flow estimated from genetic data showed the inverse pattern. The negative association between morphological divergence and gene flow is consistent with the expectation that gene flow can constrain adaptation. Estimated absolute levels of gene flow also implied a constraint on adaptation in the outlet but not the inlet. Our results suggest that natural selection promotes the adaptive divergence of lake and stream stickleback. but that the magnitude of divergence can be constrained by gene flow.