Abstract
Parallel divergence and speciation provide evidence for the role of divergent selection in generating biological diversity. Recent studies indicate that parallel phenotypic divergence may not have the same genetic basis in different geographical locations – ‘outlier loci’ (loci potentially affected by divergent selection) are often not shared among parallel instances of phenotypic divergence. However, limited sharing may be due, in part, to technical issues if false‐positive outliers occur. Here, we test this idea in the marine snail Littorina saxatilis , which has evolved two partly isolated ecotypes (adapted to crab predation vs. wave action) in multiple locations independently. We argue that if the low extent of sharing observed in earlier studies in this system is due to sampling effects, we expect outliers not to show elevated F ST when sequenced in new samples from the original locations and also not to follow predictable geographical patterns of elevated F ST . Following a hierarchical sampling design (within vs. between country), we applied capture sequencing, targeting outliers from earlier studies and control loci. We found that outliers again showed elevated levels of F ST in their original location, suggesting they were not generated by sampling effects. Outliers were also likely to show increased F ST in geographically close locations, which may be explained by higher levels of gene flow or shared ancestral genetic variation compared with more distant locations. However, in contrast to earlier findings, we also found some outlier types to show elevated F ST in geographically distant locations. We discuss possible explanations for this unexpected result.