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

Because research activities were more frequent at NC than at NA, we considered whether human disturbance could have caused mortality or movements that influenced the NC decline (Mattlin 1978, Lunn and Boyd 1991). Pups were weighed and marked or tagged at 14-d intervals beginning in late December at NC and in mid-January at NA, but we observed no injury or mortality that could be attributed to this handling. At NC, 222 perinatal fur seals were handled for attaching instruments (Walker and Boveng 1995), and the seals' behavior was monitored after instrumentation; 93% stayed near the area in which they were captured at NC. Thus, even though NC was disturbed more frequently than NA, we observed no disturbance-related pup mortality or shifts in seals' preferences for breeding sites that would explain the population decline at NC. Predation by leopard seals, with a consequent drop in recruitment, remains the most likely cause of the decline.

We do not know when leopard seal predation and the decline in fur seal births began at NC, but our results indicate that they were already underway when we began this study [ILLUSTRATION FOR FIGURES 2 AND 3 OMITTED]. Hunt (1973) observed 16 fur seal pups on 13-14 December 1970 at Seal Island, all in one colony that presumably was NC. Assuming that he counted all the live pups and that half the pups had yet to be born, [approximately]32 pups would have been born that year. The colony must have grown substantially before beginning the present decline, perhaps in the early 1980s. (A forward projection from 1970/1971 with a growth rate of 25%/yr as we observed at NA, and a backward projection from the 4.2% decline in our observed counts from NC, would intersect at a value of 380 pups in 1981/1982.)

We have very few data from which to estimate the frequency of leopard seal predation at other fur seal colonies in the Elephant Island vicinity. However, on the two occasions when we were able to count pups twice in one season, the second count was lower by more than would be expected in the absence of predation: In 1989/1990, a second count of 132 fur seal pups at the Large Leap colony on 27 February indicated that daily pup mortality there had been [approximately]1.4% since the earlier count on 12 January (Table 3), over four times the average mortality rate at NA on Seal Island. At Cape Lindsey, counts obtained on 19 January and 21 February 1992 indicated a daily decline of [approximately]0.5%, slightly greater than our highest estimate for NA. The actual mortality was likely [greater than]0.5%, however, because the first count was obtained by observations at a greater distance than the second count and is likely to have underestimated the initial number of pups. The colonies at Large Leap Island and Cape Lindsey are adjacent to several tidal lagoons in which fur seal pups develop their swimming abilities; these lagoons appeared to be as accessible to leopard seals as the lagoon in which leopard seals hunt fur seal pups at NC on Seal Island (L. M. Hiruki et al., unpublished manuscript). It seems plausible, therefore, that leopard seals may cause substantial pup mortality at these colonies, as well. These possibilities, and our results describing the distribution of fur seal birth dates and onset of leopard seal predation at Seal Island, indicate that accurate monitoring of future population trends would be facilitated by pup counts obtained on or about 1 January, .just after births are complete but before significant predation occurs.