Model testing by analyses of molecular variance (AMOVA) and multivariate matrix regressions (MMR).

The test statistic reported is the proportion of variance in the data that was accounted for by each model. For AMOVA this may be the proportion of variance among populations (FST, in bold case, where #groups = 1) or among groups (FCT, normal case, where #groups>1). Conditional MMR tests assess the proportion of marginal variation remaining after geography is taken into account. Sequential MMR tests fit the best model to the data and the statistic reported is that proportion of the remaining variation that is attributable to the second model. Models (P, phenotype; B, biogeography; G, geography; M, molecular; Lt, latitude; and Lg, longitude) are listed in the sequential order in which they were fitted to the regression. Significance was denoted *** (p