Despite the evidence of population segregation from marking, survey, and radiotelemetry data, initial evaluations using genetic techniques suggested small differences among polar bears in different geographic regions. Such small differences might be expected under Pedersen's (1945) hypothesis of a globally wandering panmictic polar bear population, but not in light of current knowledge of movements. Using protein electrophoresis, Larsen et al. (1983b) found little variation in allozymes among polar bears. They concluded there was no reason to reject a hypothesis of one common polar bear population worldwide. Larsen et al. (1983b) assumed that high gene flow across the Arctic and strong common selective pressures reduced variation among polar bears. Supporting that concept, Durner and Amstrup (1995) recorded the movement of a radio-collared polar bear from near Prudhoe Bay, Alaska, to Greenland. That bear was only 1 of 106 equipped with platform transmitter terminals or satellite radio-collars between 1985 and 1992. Relatively few such movements, however, would be required to genetically homogenize polar bear populations worldwide (Paetkau et al. 1995).

Corroborating the work of Larsen et al. (1983b), Cronin et al. (1991) found little mitochondrial DNA (mtDNA) differentiation among bears of the Beaufort and Chukchi seas. Mitochondrial DNA and protein electrophoresis, however, may have a relatively low ability to resolve genetic variation among populations (Scribner et al. 1997). Therefore, the absence of differences in these markers may not be surprising.

More recent studies using highly variable nuclear genetic markers called microsatellites have resolved differences among polar bears living in different areas. Paetkau et al. (1995) found significant differences in genetic distances among members of four Canadian populations of polar bears. They concluded that the long-distance movements of polar bears have not resulted in complete genetic mixing of populations. Furthermore, Paetkau et al. (1999) reported significant correlations among movement data and genetic data from polar bear populations worldwide. They found greater differences among populations in the Canadian Arctic than among populations surrounding the polar basin. Such contrasts fit well with observed differences in movement patterns in these areas (Amstrup et al. 1986, 2000; Messier et al. 1992; Amstrup and Gardner 1994; Bethke et al. 1996; Scribner et al. 1997).

Genetic management units that correspond with boundaries defined by radiotelemetry have now been identified for most polar bear populations (Paetkau et al. 1999). The correspondence between movement data and recent genetic data allows managers to make better decisions about harvest or other human activities that could have population-level impacts.