It is often assumed that species which exhibit a greater propensity for dispersal are less susceptible to the impacts of habitat fragmentation; however, a growing body of literature suggests that such generalizations should be carefully evaluated as not all species appear to be equally sensitive to fragmentation. In this issue of Molecular Ecology, Callens et al. (2011) take an innovative approach to compare contemporary estimates of dispersal from an extensive mark-recapture and patch occupancy data set with historical estimates derived from multilocus population genetic models for seven sympatric forest-dependent species in the Taita Hills, Africa. As has been observed for forest-dependent species from the Amazon, populations of sedentary species were more strongly differentiated and clustered when compared to those of more dispersive taxa. The most intriguing result recovered though, was that the five species with similar historical estimates of gene flow (dispersal) differed substantially in their contemporary dispersal rates, suggesting that for some species the propensity for dispersal has decreased over time. As a consequence, the authors suggest that post-fragmentation estimates of dispersal on their own may not be the best predictors of how habitat fragmentation could affect forest-dependent animal communities.This work significantly advances our understanding of the dynamics of habitat fragmentation and makes a strong case for the need to integrate data on historical processes with contemporary data.
© 2011 Blackwell Publishing Ltd.