Because many in that age class were independent of their radio-collared mothers, they were not recaptured or reobserved, and their fate was not known. Hence, these must be considered minimum survival values, and likely are below the actual values.

In the Beaufort Sea, survival of cubs was approximately 65% from den exit to the end of their first year of life. Yearlings fared much better, with 86% surviving to weaning (Amstrup and Durner 1995). Observations of the young of radio-collared females substantiate the observation from Hudson Bay that most cub mortality comes early in the period after emergence from the den (Amstrup and Durner 1995; Derocher and Stirling 1996), but depart radically from the very minimal yearling survivals observed there. Derocher and Stirling (1996) suggested that a heavy harvest accounted for much of the yearling mortality in Hudson Bay. Nonetheless, only 15% (44% × 35%) of the cubs produced were confirmed to survive through their second autumn. This contrasts with the survival of 56% (65% × 86%) of Beaufort Sea bears surviving until weaning, 5 months beyond their second autumn. If actual values are close to the minimums reported there, the differences in survival could more than compensate for the apparent reproductive differences between bears in the Beaufort Sea and Hudson Bay.

Tait (1980) hypothesized that brown bear females may choose to abandon a single cub on the chance that they might enhance fitness by breeding again and giving birth to twin or larger litters. That concept has resulted in much discussion and debate about parental care and investment in young. Whether or not it makes sense mathematically, such a strategy apparently does not prevail among polar bears. In the Hudson Bay area, single cubs may actually survive at a somewhat higher rate than cubs from larger litters (Derocher and Stirling 1996). Furthermore, deaths of dependent young in the Beaufort Sea were independent of litter size, and cubs were lost at similar rates whether as whole litters or portions of litters (Amstrup and Durner 1995). Parental investment in single polar bear cubs is not different from investment in litters of two or more. Single cubs are often much heavier than twin cubs (S. C. Amstrup, unpublished data), and survival of cubs appears to be heavily dependent on their weight (Derocher and Stirling 1992).

Estimating survival rates of independent polar bears has been an even greater challenge than estimating survival of dependent young. Eberhardt (1985) argued that survival of adult marine mammals must be in the high 90% range for their populations to be sustaining. However, early estimates of survival in polar bears derived by mark and recapture methods were much lower (Amstrup et al. 1986). More recent estimates derived from Hudson Bay, where the intensity of marking exceeds all other study areas, still have ranged between 0.86 and 0.90 (Derocher and Stirling 1995a; Lunn et al. 1997). Only by relying on radiotelemetry monitoring of individual animals have estimates in line with Eberhardt's (1985) theory been developed. Amstrup and Durner (1995) estimated that survival of adult females in the prime age groups may exceed 96%. Although that estimate fits well with population dynamics theory, the fact that it is much higher than estimates derived by other methods suggests added work on polar bear survival is necessary.