Population growth of Antarctic fur seals: limitation by a top predator, the leopard seal?

Nine of the 10 model estimates of total births at both colonies were greater than the corresponding peak live pup counts and the estimates of total births were very close to the peak count of live pups plus the number of dead pups observed prior to the peak count (Table 1).

The estimates of mortality rates at NA from 1989/1990 to 1994/1995 indicated that [approximately]0.3% of pups died each day (Table 1), with an approximate standard error of 0.1%. Estimates of cumulative pup mortality at NA, from birth to [approximately]75 d of age, ranged from 16% of the cohort in 1993/1994, to 29% in 1990/1991. At NC, mortality from nonpredatory causes was similar (1630%; Table 2), partly because the daily rate was fixed at the NA rate. The estimated proportion of pups taken by leopard seals during the same period ranged from 32 to 38% of total births. Total pup mortality ranged from 51 to 66% (Table 2) and was probably even higher in 1989/1990 ([approximately]79%), given that the estimated number of pups taken by leopard seals in that year was twice the highest number seen in the other years (Table 2).

Fur seal abundance and trends in the Elephant Island area

The number of pups counted at the four main breeding sites in the Elephant Island area (Seal Island, Large Leap Island, Cape Lindsey, and Cape Valentine; [ILLUSTRATION FOR FIGURE 1 OMITTED]) increased from 772 pups in 1986/1987 to 1010 pups in 1993/1994 (Table 3), indicating an annual growth rate of 3.8 [+ or -] 0.2% (mean [+ or -] 1 SE, [R.sup.2] = 0.997). In 1993/1994, two other small colonies were discovered in the northwestern islets of the Seal Islands archipelago, as well as one on the southern point of Cape Lindsey, Elephant Island (Table 3). There were at least 1096 fur seal pups born in the Elephant Island vicinity in 1993/1994 (Table 3), allowing for the fact that most of the colonies were counted after the expected peak of seasonal abundance. That number of pups would correspond to a total population of [approximately]4500 fur seals if the age structure was similar to that at South Georgia in the early 1970s (Payne 1979).

[TABULAR DATA FOR TABLE 1 OMITTED]

DISCUSSION

Fur seal pup production and mortality at Seal Island

Assessment of model assumptions. - The model for estimating breeding chronology and pup mortality was based upon three primary assumptions: (1) that birth dates are normally distributed, (2) that there is no leopard seal predation of fur seal pups at NA, and (3) that the daily rate of pup mortality from nonpredatory causes is the same at NA and NC and is constant throughout the breeding season.

Results from other studies of Antarctic fur seals support the first assumption. Although there is some variation attributable to seal age and food availability (Boyd et al. 1990, Duck 1990, Lunn and Boyd 1993a, b, Boyd 1996), individual females tend to give birth at the same time of the season each year (Boyd 1996). The duration of pregnancy in Antarctic fur seals appears to be normally distributed, and is highly correlated [TABULAR DATA FOR TABLE 2 OMITTED] with birth date (Boyd 1996). Thus, it seems reasonable to assume that birth dates are also normally distributed.