Migration among populations is widely thought to undermine adaptive divergence, assuming gene flow arises from random movement of individuals. If individuals instead differ in dispersal behavior, phenotype-dependent dispersal can reduce the effective rate of gene flow or even facilitate divergence. For example, parapatric populations of lake and stream stickleback tend to actively avoid dispersing into the adjoining habitat. However, the behavioral basis of this nonrandom dispersal was previously unknown. Here, we show that lake and stream stickleback exhibit divergent rheotactic responses (behavioral response to currents). During the breeding season, wild-caught inlet stream stickleback were better than lake fish at maintaining position in currents, faced upstream more, and spent more time in low-current areas. As a result, stream fish expended significantly less energy in currents than did lake fish. These divergent rheotactic responses likely contribute to divergent habitat use by lake and stream stickleback. Although rheotactic differences were absent in nonbreeding fish, divergent behavior of breeding-season fish may suffice for assortative mating by breeding location. The resulting reproductive isolation between lake and stream fish may explain the fine-scale evolutionary differentiation in parapatric stickleback populations.

Gene flow is widely thought to homogenize spatially separate populations, eroding effects of divergent selection. The resulting theory of 'migration-selection balance' is predicated on a common assumption that all genotypes are equally prone to dispersal. If instead certain genotypes are disproportionately likely to disperse, then migration can actually promote population divergence. For example, previous work has shown that threespine stickleback (Gasterosteus aculeatus) differ in their propensity to move up- or downstream ('rheotactic response'), which may facilitate genetic divergence between adjoining lake and stream populations of stickleback. Here, we demonstrate that intraspecific variation in a sensory system (superficial neuromast lines) contributes to this variation in swimming behaviour in stickleback. First, we show that intact neuromasts are necessary for a typical rheotactic response. Next, we showed that there is heritable variation in the number of neuromasts and that stickleback with more neuromasts are more likely to move downstream. Variation in pectoral fin shape contributes to additional variation in rheotactic response. These results illustrate how within-population quantitative variation in sensory and locomotor traits can influence dispersal behaviour, thereby biasing dispersal between habitats and favouring population divergence.

The combined use of otolith chemistry and tissue isotopes has the potential to reveal movements, habitat associations and food web interactions at a variety of spatial and temporal scales. Here, a combination of otolith Ba:Ca life‐history transects with muscle tissue δ13C and δ15N values has been used to assess habitat use and oligohaline residence in red drum Sciaenops ocellatus in subtropical estuaries in the north‐western Gulf of Mexico. Tissue isotopes were distinct among capture locations, particularly between bays with differing proximities to freshwater inflow sources. Otolith edge Ba:Ca values and tissue δ13C values were not correlated. These results indicated that fish were neither residing in nor feeding in oligohaline waters for significant periods of time within the tissue turnover window of several months prior to capture. Nonetheless, spatial differences in tissue isotope values indicated limited mixing among bays and relatively high site fidelity during estuarine occupancy. Lifetime otolith Ba:Ca transects revealed individual variability in the magnitude of residence in oligohaline waters. Using a mean oligohaline occupancy threshold, an estimated 82% of individuals used oligohaline waters at some point in their life. Nonetheless, 66% of individuals spent <20% of their life histories in oligohaline waters, suggesting intermittent and infrequent excursions into low salinity waters. Finally, a literature survey identified 56 peer‐reviewed publications using combinations of otolith chemistry and tissue stable isotope ratios with a wide range of marker pairings and study aims. The diversity of ecological questions that can be asked with the combined use of these two approaches will provide valuable insight into fish ecology.