I studied the social organization and foraging ecology of eastern coyotes (<i>Canis latrans</i>) in Nova Scotia, from 1992 to 1997. Breeding pairs formed the nucleus of coyote social groups and generally traveled with 1-3 other coyotes during winter. Mean winter traveling group size was similar for packs utilizing white-tailed deer (<i>Odocoileus virginianus</i>) or snowshoe hare (<i>Lepus americanus</i>) as a primary food item. Thus, increased use of large prey was insufficient to explain group formation and cohesion of coyotes. Home-range sizes decreased significantly with deer and hare densities. Coyotes used the same territories during winter and summer, and from year to year. During winter, coyotes used, and killed deer, in areas of low deer density proportionately more than expected, likely owing to the increased vulnerability of deer in these areas. Territoriality prevented coyotes from concentrating in deer wintering areas. Based on the analysis of 2,443 scats, deer and hare were the dominant food items. During winter, coyotes killed 76-86% of the deer they consumed. Most deer killed by coyotes were not malnourished. Deer killing rates were negatively correlated with hare density. Mean urinary urea nitrogen: creatinine ratios were correlated positively with hare density. Overall, high use of deer appeared to have been associated with increased vulnerability due to winter severity or, in the case of young fawns, inability to escape. During mild winters, coyotes are forced to focus their hunting efforts on prey other than deer, regardless of density, due to low vulnerability of deer. When severe winter conditions occur, coyotes switch to feeding mainly on deer. Coyotes exhibited a Type I numerical response to the total biomass of deer and hares. Survival of fawns during summer was positively correlated with hare density. However, the total response of coyotes to changes in deer and hare density indicated that the presence of hare increased predation on deer by supporting higher coyote densities. A model based on the total response of coyotes to changes in deer and hare abundance predicted that food competition rather than predation probably regulated deer populations. Extirpation due to intense predation was a possible outcome for deer at densities <0.6 deer/km2.