Figure 2.

Rate of gain and rate of loss in the two‐locus gain–loss model with one weakly diverged locus ($D_{\mathrm{w}}=0.2$) and a more strongly diverged one ($D_{\mathrm{s}}=0.4$). Shown are the rates as a function of distance between the loci (measured in units of recombination rate, r). Panels (A) and (B): rate of gain at the weakly diverged locus (circles) and at the more strongly diverged locus (squares) for weak selection (A), and for strong selection (B). Dashed lines indicate the mutation rate μ (this rate corresponds to the rate at which a neutral mutation lands and fixates at a neutral locus in a diploid population of size N). Panels (C) and (D): corresponding rates of loss for the parameters in (A) and (B), respectively. Circles and squares overlap in (D). Other parameter values: selection parameter $\sigma =4$ in (A) and (C) or $\sigma =2.5$ in (B) and (D); population size in each deme, $N=1000$; migration rate, $m=0.1$; mutation‐effect size, $\epsilon =0.05$; recombination rate, $r=0.0005$; mutation rate, $\mu =2\times {10}^{-5}$. Number of simulations used are as follows: 2 × 10⁶ in (A), 5 × 10⁵ in (B), 10³ in (C), and 200 in (D).