Founder effects drive high mitochondrial dN/dS in island rails

Insular (island-limited) populations typically show signatures of weak purifying selection, indicating high genetic load and reduced fitness compared with mainland populations. However, the source of this pattern is often unclear-it may reflect residual signatures from a temporary period of small effective population size (Ne) associated with island colonization (founder effects), persistently small Ne due to the lower carrying capacity of islands (range limitations), or relaxed selective constraints unrelated to Ne. Here, we disentangle these hypotheses by analyzing the drivers of variation in evolutionary rates of nonsynonymous (dN) and synonymous (dS) sites in nine mitochondrial genes (8001 bp) from 40 rail species (Aves: Rallidae). We find that insular species with short terminal branches (indicating recent island colonization) have highly elevated mitochondrial dN/dS across multiple mitochondrial genes. In contrast, rails representing more ancient island colonizations have dN/dS ratios that are indistinguishable from mainland/widespread species. Furthermore, we find that island size is unrelated to dN/dS among island species. These results indicate that insular rails suffer a high initial cost of island colonization and undergo a period of inefficient selection due to founder effects, but that there is little impact from longer-term range limitations or relaxed selection.