Individual assignment test reveals differential restriction to dispersal between two salmonids despite no increase of genetic differences with distance.

In many species genes move over limited distances, such that genetic differences among populations or individuals are expected to increase as a function of geographical distance. In other species, however, genes may move any distance over a single generation time, such that no increase of genetic differences is expected to occur with distance. Patterns of gene dispersal have been assessed typically using this theoretical property. In this study, this classical approach based on a Mantel test was compared to a new method using individual assignment to reveal contrasts in dispersal patterns between 15 populations of brook charr Salvelinus fontinalis and 10 populations of Atlantic salmon Salmo salar sampled in eastern Canada, where both species co-occur naturally. Based on the Mantel test, we found evidence for neither an increase of genetic differences with distance in either species nor a significant contrast between them. The individual-based method, in contrast, revealed that individual assignment in both species was non random, being significantly biased toward geographically proximate locations. Furthermore, brook charr were on average assigned to a closer river than were salmon, according to a priori expectations based on the dispersal behaviour of the two species. We thus propose that individual assignment methods might be a promising and more powerful alternative to Mantel tests when isolation by distance cannot be postulated a priori.