An appreciation of the scale and frequency of climatic oscillations in the past few million years is modifying our views on how evolution proceeds. Such major events caused extinction and repeated changes in the ranges of those taxa that survived. Their spatial effects depend on latitude and topography, with extensive extinction and recolonization in higher latitudes and altitudinal shifts and complex refugia nearer the tropics. The associated population dynamics varied with life history and geography, and the present genetic constitution of the populations and species carry attenuated signals of these past dynamics. Phylogeographic studies with DNA have burgeoned recently and studies are reviewed from the arctic, temperate and tropical regions, seeking commonalities of cause in the resulting genetic patterns. Arctic species show distinct shallow genetic clades with common geographical boundaries. Thus Beringia is distinct phylogeographically, but its role as a refugial source is complex. Arctic taxa do not show the common genetic pattern of southern richness and northern purity in north-temperate species. Temperate refugial regions in Europe and North America show relatively deep DNA divergence for many taxa, indicating their presence over several Ice Ages, and suggesting a mode of speciation by repeated allopatry. DNA evidence indicates temperate species in Europe had different patterns of postglacial colonization across the same area and different ones in previous oscillations, whereas the northwest region of North America was colonized from the north, east and south. Tropical montane regions contain deeply diverged lineages, often in a relatively small geographical area, suggesting their survival there from the Pliocene. Our poor understanding of refugial biodiversity would benefit from further combined fossil and genetic studies.