Ecological processes clearly contribute to population divergence, yet how they interact over complex life cycles remains poorly understood. Notably, the evolutionary consequences of migration between breeding and nonbreeding areas have received limited attention. We provide evidence for a negative association between interpopulation differences in migration (between breeding and feeding areas, as well as within each) and the amount of gene flow (m) among three brook charr (Salvelinus fontinalis) populations inhabiting Mistassini Lake, Quebec, Canada. Individuals (n ϭ 1166) captured throughout lake feeding areas over two consecutive sampling years were genotyped (10 microsatellites) and assigned to one of the three populations. Interpopulation differences in migration were compared based on spatial distribution overlap, habitat selection, migration distance within feeding areas, and morphology. We observed a temporally stable, heterogeneous spatial distribution within feeding areas among populations, with the extent of spatial segregation related to differential habitat selection (represented by littoral zone substrate). Spatial segregation was lowest and gene flow highest (m ϭ 0.015) between two populations breeding in separate lake inflows. Segregation was highest and gene flow was lowest (mean m ϭ 0.007) between inflow populations and a third population breeding in the outflow. Compared to outflow migrants, inflow migrants showed longer migration distances within feeding areas (64-70 km vs. 22 km). After entering natal rivers to breed, inflow migrants also migrated longer distances (35-75 km) and at greater elevations (50-150 m) to breeding areas than outflow migrants (0-15 km; Ϫ10-0 m). Accordingly, inflow migrants were more streamlined with longer caudal regions, traits known to improve swimming efficiency. There was no association between the geographic distance separating population pairs and the amount of gene flow they exchanged. Collectively, our results are consistent with the hypothesis that reduced gene flow between these brook charr populations results from divergent natural selection leading to interpopulation differences in migration. They also illustrate how phenotypic and genetic differentiation may arise over complex migratory life cycles.