Figure 6.

The effect of a bout of forced hybridization on the fraction of populations evolving RI. Proportional change (y‐axis) was calculated as the difference in the number of populations evolving RI when simulations were initiated with F₁s versus randomly mating parental individuals over the number of populations that evolved RI when simulations were initiated with randomly mating parental individuals. Symmetrical incompatibilities, in general, evolved more frequently in simulations initiated with a hybrid deme composed of F₁ hybrid individuals compared to when initiated with equal proportions of randomly mating parental genotypes. Results are shown for each of the three simulated genetic architectures (panel rows) under the directional selection (a), diversifying selection (b), and selection‐for‐admixture (c) models of selection acting on “adaptive” loci. Recombination rates and migration were held at 0.2 and 0.001, respectively. In instances when there was a greater than 10‐fold increase in the proportion of populations that evolved RI, values were rounded down to 10. Missing points occur for parameter combinations where no populations evolved RI across simulations initiated under either initial condition