Wolves (Canis lupus) and arctic foxes (Alopex lagopus) are the only canid species found throughout the mainland tundra and arctic islands of North America. Contrasting evolutionary histories, and the contemporary ecology of each species, have combined to produce their divergent population genetic characteristics. Arctic foxes are more variable than wolves, and both island and mainland fox populations possess similarly high microsatellite variation. These differences result from larger effective population sizes in arctic foxes, and the fact that, unlike wolves, foxes were not isolated in discrete refugia during the Pleistocene. Despite the large physical distances and distinct ecotypes represented, a single, panmictic population of arctic foxes was found which spans the Svalbard Archipelago and the North American range of the species. This pattern likely reflects both the absence of historical population bottlenecks and current, high levels of gene flow following frequent long-distance foraging movements. In contrast, genetic structure in wolves correlates strongly to transitions in habitat type, and is probably determined by natal habitat-biased dispersal. Nonrandom dispersal may be cued by relative levels of vegetation cover between tundra and forest habitats, but especially by wolf prey specialization on ungulate species of familiar type and behaviour (sedentary or migratory). Results presented here suggest that, through its influence on sea ice, vegetation, prey dynamics and distribution, continued arctic climate change may have effects as dramatic as those of the Pleistocene on the genetic structure of arctic canid species.

Wolves in the Canadian Arctic Archipelago face several challenges to persistence: a harsh habitat, an unstable prey base, and potentially significant anthropogenic influences. These external factors, if combined with genetic constraints common to island populations, could be particularly difficult to withstand. To determine the genetic status of Arctic Island wolves, we used 14 microsatellite loci to estimate population variation and the extent of interisland and island-mainland gene flow. All island populations were significantly less variable than mainland wolves; although inbreeding is currently insignificant, the two least variable populations, Banks and the High Arctic (Ellesmere and Devon Islands), showed genetic signatures of recent population declines. Recovery after a bottleneck appears to result, in large part, via recolonization from other islands. These extinction-recolonization dynamics, and the degree of similarity among island wolves revealed by Bayesian clustering, suggest that Arctic Island wolves function as a metapopulation. Persistence of the metapopulation may be supported by periodic migration from mainland populations, occurring primarily through two corridors: Baffin Island in the Eastern Arctic, and Victoria Island in the Western Arctic. This gene flow could be compromised or eliminated by loss-due to climatic warming or increased human activity-of sea ice in the Northwest Passage.

Habitat loss, fragmentation, overharvest, and other anthropogenic factors have resulted in population and distribution declines for North American wolverines (Gulo gulo). Currently, wolverines east of the Hudson