Slatkin's Paradox: when direct observation and realized gene flow disagree. A case study in Ficus.

In flying insects, there is frequently a lack of congruence between empirical estimates of local demographic parameters and the prediction that differentiation between populations should decrease with increasing dispersal, a puzzling phenomenon known as Slatkin's Paradox. Here, we generalize Slatkin's Paradox to other taxa, drawing from available information on dispersal to predict the relative importance of pollen vs. seed migration in structuring broad-scale patterns of genetic variation in Ficus hirta, a dioecious fig whose pollen is dispersed by minute, species-specific fig wasps and whose seeds are disseminated by a variety of vagile vertebrates (especially bats and birds). Local-scale observational and genetic studies of dioecious understory figs suggest comparable rates of pollen and seed migration. In contrast, we found unusually low nuclear differentiation (F(ST)  =0.037, R(ST)  =0.074) and high chloroplast differentiation (G(ST)  =0.729, N(ST)  =0.798) among populations separated by up to 2850km, leading us to reject the hypothesis of equal pollen and seed migration rates and to obtain an equilibrium estimate of the ratio of pollen to seed migration of r=16.2-36.3. We reconcile this example of Slatkin's Paradox with previously published data for dioecious figs and relate it to the picture of exceptionally long-distance wasp-mediated pollen dispersal that is emerging for large monoecious fig trees. More generally, we argue that Slatkin's Paradox is a general phenomenon and suggest it may be common in plants and animals.