Interpopulation Gene Flow by Pollen in Wild Radish, Raphanus sativus

Abstract
Although gene flow can act as an important evolutionary force, direct measurements of interpopulation gene flow remain sparse. A paternity exclusion procedure was used to compare multilocus progeny genotypes with those of their maternal parents and with all possible local paternal parents in three populations of wild radish, Raphanus sativus, isolated by 100 m to 1 km from other conspecifics. The comparison permitted identification of those progeny sired by paternal parents in other populations. Over all the seed sampled, the rate of interpopulation gene flow by pollen varied twofold, from 8.2% to 17.9% per population per generation. Heterogeneity among plants varied considerably, from 0% to 44% of the seed set by an individual. The most important spatial correlate of successful interpopulation mating was the number of neighbors within 1 m, which had a significantly negative relationship with this parameter. The levels of gene flow by pollen reported here are sufficient to counterbalance selection at levels commonly reported in plant populations. Since interpopulation gene-flow rates should vary with population size, self-incompatibility, and pollen vector, we argue that the range reported for these populations of R. sativus is a small subset of those possible for other populations within this species. Given the probably idiosyncratic nature of interpopulation gene-flow rates, the importance of gene flow as a cohesive evolutionary force may range from trivial to critical, for the populations of both R. sativus and any other plant species.