Estimating gene flow in island populations.

A new method is presented for estimating the rate of gene flow into island populations using the distribution of alleles in samples from a number of islands. The pseudo maximum likelihood estimator (PMLE) that we derive may be applied to species with either discrete or continuous generation times. For Wright's discrete-generation island model, the method provides an estimate of theta = 2Nm where N is the (haploid) population size on each island and m is the fraction of individuals replaced by immigrants in each generation. For a continuous-generation island model, the corresponding parameter theta is the ratio of the immigration rate phi to the individual birth rate lambda. Monte Carlo simulations are used to compare the statistical properties of the PMLE with those of two alternative estimators of theta derived from Wright's F-statistics. The PMLE is shown to have greatest efficiency (least mean square error) in most cases for a wide range of sample sizes and parameter values. The PMLE is applied to estimate theta using mtDNA haplotypes and allozymes for subdivided populations of African elephants and Channel Island foxes.