A geographic analysis of snowshoe hare population demography

Snowshoe hare (Lepus americanus) populations across northern Canada and Alaska undergo 8- to 11-year cycles in numbers, but population trends in southern Canada and the contiguous United States are apparently either weakly cyclic, irruptive, or largely stable. Although the demographic attributes (population density, reproductive rates, and survival rates) of northern and southern hare populations should differ considerably to account for such differential trends, to date limited rangewide analyses of hare demography have been undertaken. I reviewed hunter harvest estimates and basic demographic attributes for hare populations across North America, and assessed the effect of latitude, longitude, and latitude x longitude interaction on the magnitude and variation of such attributes. Harvest estimates tended to be synchronous across the Great Lakes area and along the St. Lawrence River, whereas they varied more dramatically along a westward gradient and in hare populations suspected of being cyclic. Hare densities tended to be higher among apparently noncyclic than cyclic populations at numerical lows, as well as among studies where hare numbers were apparently increasing. Populations from northern latitudes tended to breed later than those in the south, and females from western areas produced larger, but fewer litters, than those from eastern populations; total productivity was similar across geographic areas. Survival rates for both adult and juvenile hares were higher in increasing populations than in decreasing populations. Survival of adults also decreased along a northwest gradient, whereas that of juveniles decreased across a western gradient and with longitude, was lower in apparently noncyclic populations, and was also lower in populations in areas of high lynx (Lynx canadensis) densities. I conclude that, although disparity clearly exists in the trends of various hare populations, the absence of strong latitudinal gradients in demographic attributes fails to support the hypothesis that differential survival/predation is responsible for the regional differences in numerical trends.