Analysis of estimates of population size of ringed s,eals (Phoca hispida) and polar bears (Ursus maritimus) from several areas indicated that estimates of one predicted the range of expected population size of the other in areas where ringed seals constitute the primary prey. In some areas, the closeness of this relationship indicates where estimates of either seals or bears may be inaccurate. The number of seals required to support a population of polar bears of predetermined size was estimated independently using both behavioral and energetic data. Behavioral estimates of the number of seals killed may overestimate energetic requirements and vice versa. Predation and energy matrices indicated that high levels of predation on seals are sustainable only if most animals killed are young-of-the-year. The field metabolic rate of the polar bear appears to be about twice the basal metabolic rate. Densities of seals vary in response to overall productivity of the ecosystem in different areas, and fluctuations in their numbers and reproductive rates between years can be used to monitor changes in productivity of the ecosystem. These changes also cause variation in productivity of bears, which indicates the sensitivity, at the population level, of the relationship between ringed seals and polar bears.