Sky islands are ideal systems for determining the effects of climatic oscillations on species distributions and genetic structure. Our study focused on montane stonefly populations in the Great Basin of western North America. We used niche-based distribution modelling, phylogeography and traditional species-based biogeography to test several hypotheses as follows: (i) genetic differentiation among Doroneuria baumanni populations will be independent of hydrologic connectivity (headwater model); (ii) Sky islands were colonized when habitat was more continuous and populations likely experienced multiple expansions and contractions; (iii) Colonization events were coincident with the late Pleistocene and Holocene; and (iv) Shared topography and climate history will result in concordant patterns of genetic differentiation in D. baumanni and occurrences of 32 stonefly species across the region. Overall, ΦST’s and coalescent-based estimates of migration were consistent with the headwater model. Maximum likelihood and Bayesian gene trees identified three major nonoverlapping east-west clades. Distribution modelling indicated more suitable habitat in the Great Basin during the Last Glacial Maximum than at present, but none during the last interglacial period. Demographic analyses showed evidence of population expansion in one of the three major east–west clades. Intra-clade divergence times (60 000–183 000 ybp) were well within the late Pleistocene while among-clade divergence times (499 000–719 000 ybp) were deeper. Genetic differentiation in D. baumanni and distributions of stonefly species were significantly concordant. These results imply that climatic oscillations have played major roles in shaping the genetic structure and distributions of Great Basin stoneflies, but that divergence among clades occurred much earlier than our late Pleistocence/early Holocene predictions.