Log Bayes factors and model parameters from analysis of migration (software Migrate-n) in the riparian shrub <i>Myricaria germanica</i> collected from four catchments in Switzerland.

Bayes Factors were constructed in comparison with the model with the largest log likelihood (Bayes Factor zero). Model probability was calculated by dividing the marginal likelihood of a given model by the sum of the marginal likelihoods of all models. Model probabilities: *0.01<si<0.05; **0.05<si<0.10; ***0.95<si<1.00. NC, no convergence of model, thus excluded for calculation of Bayes Factors. Migration rate estimates and population sizes of transformed data are shown in the lower panel. Populations were sorted in downstream order with increasing number, i.e. Pop1 was the most upstream. M1→2 denotes the migration rate from population 1 to population 2.