Dispersal in a stream dwelling salmonid: Inferences from tagging and microsatellite studies

We used both direct (mark-recapture) and indirect (microsatellite analysis) methodologies to investigate dispersal between two putative populations of brook charr (Salvelinus fontinalis) in Freshwater River, Cape Race, Newfoundland, Canada. Over a 5-year study period, mark-recapture data revealed some movement by fish, but the proportion of recaptured fish migrating from one population area to another was low (0-4.1%). Additionally, during sampling periods in the spawning seasons, no fish was found in the alternate population area to that of its first capture. Despite this pattern of limited movement, microsatellite analysis based on sixteen polymorphic loci provided no evidence of genetic differentiation. Indirect estimates of dispersal parameters varied greatly between different methods of analysis. While use of a coalescent-based model yielded estimated migration rates congruent with the results of the mark-recapture study, other methods resulted in much higher estimates of migration between the populations. In particular, the lack of genetic differentiation coupled with likely violations of the assumed island model prevented generation of meaningful estimates of dispersal using F_st. The disparities between migration rates estimated from the mark-recapture work and from the different indirect methods highlight the difficulties of using indirect methods to estimate dispersal on an ecological timescale. However, mark-recapture methods can fail to detect historical or episodic movement that is important in an evolutionary context, and we therefore argue that a combination of direct and indirect methods can provide a more complete picture of dispersal than either approach alone.