FISHERY BULLETIN: VOL. 85, NO. 2 



of numbers of northern fur seal pups born on St. 

 Paul Island since 1970 are presented; estimates for 

 1970-79 are based on data from Lander (1980). We 

 computed estimates for those years in which cen- 

 suses were made on all rookeries or for which data 

 were available to compute estimates according to 

 the methods developed in this paper. Regressions 

 of logarithms of numbers of pups born versus time 

 indicated a statistically significant decrease during 

 1975-81— a decrease of 7.5% per year with a stan- 

 dard error of 2%. During 1981-86, there is no statis- 

 tically significant decreasing or increasing trend; the 

 estimate of the slope is - 1.8% with a standard er- 

 ror of 1.8%. This slope is statistically different from 

 the - 7.5% slope calculated for 1975-81 (P > 0.90). 



DISCUSSION 



Our study indicates that we can obtain reasonable 

 estimates of the total number of northern fur seal 

 pups born from subsampling as few as four rook- 

 eries of St. Paul Island if estimates of numbers of 

 pups and breeding males are available for the sam- 

 ple rookeries and if a total bull count is available for 

 the island. Subsampling is successful because within 

 a given year, pup production is predictably propor- 

 tional to numbers of breeding males. Some refine- 

 ments in the reduction of bias and variance can be 

 made by restricting the subsamples to stratified 

 designs over large, medium, and small rookeries. 



The advantages of subsampling rookeries for cen- 

 susing northern fur seal numbers are considerable. 

 Most important is the reduction of total disturbance 

 on the northern fur seal population on the island. 

 Our sampling schedule over several years attempts 

 to apportion disturbance approximately equally so 

 that rookeries are neither under- or oversampled 

 through time. This is an important aspect of the 

 sampling design, since it is not known how great 

 the long-term impact of disturbance is. In addition, 

 subsampling requires a smaller crew for the shear- 

 ing and less time for resampling, resulting in con- 

 siderable savings of resources. 



Ratios of numbers of males to pups, and conse- 

 quently breeding females, vary considerably over 

 time, even in successive years. It is difficult to inter- 

 pret the meaning of these changes. During the 

 period covered by Figure 1 (1912-22), numbers of 

 pups born on St. Paul Island were increasing rather 

 rapidly. Since males begin to breed at an older age 

 than females, part of the increase in the ratio of 

 breeding males to pups may be explained by the 

 number of breeding males lagging a few years 

 behind the number of breeding females. Signifi- 



cantly different ratios from one year to the next 

 could also be due to differences in counting methods 

 or abilities among individual counters, or to differ- 

 ent survival rates among separate cohorts (e.g., 

 harvest rates). Figures 1 and 2 also imply a certain 

 consistency and a rather uniform rate of usage of 

 rookeries by breeding males and females, in that, 

 if a rookery accounts for 10% of breeding males 

 within a year, it will account for approximately 10% 

 of the total pup production within the same year. 

 A rookery's relative contribution to both these 

 populations may change but the correlation between 

 them does not appear to change. 



The recent history of the population of numbers 

 of pups on St. Paul Island in Figure 4 shows a de- 

 crease of about 7.5% per year during 1975-81. No 

 significant trend is detectable after 1981, although 

 the number born in 1982 was significantly higher 

 than in 1981 or 1983-86. The causes of the decline 

 are unknown. There is no evidence that pregnancy 

 rates have changed significantly since the 1950's 

 (Goebel and Gentry 1984"*). Thus, considerable at- 

 tention has centered on potential causes of increased 

 mortality of northern fur seals: entanglement in 

 debris (e.g.. Fowler 1985), effects of weather (Trites 

 1984; York 1985^), and direct effects on food avail- 

 ability from competition with fisheries in the North 

 Pacific Ocean (York and Hartley 1981; Swartzman 

 and Harr 1983; Kajimura 1984; Loughlin and 

 Livingston 1985^). One may also speculate that the 

 pattern of decline and possible stabilization in num- 

 bers of pups born resulted from a new disease which 

 abated or was controlled by an immune response of 

 the population (c.f., Geraci et al. 1982). 



Of the aforementioned explanations for the de- 

 cline, only entanglement has been cited as a major 

 contributing factor with an attributed mortality of 



«Goebel, M. E., and R. L. Gentry. 1984. The use of longitu- 

 dinal records of tagged females to estimate fur seals survival and 

 pregnancy rates. Unpubl. manuscr. National Marine Mammal 

 Laboratory, Northwest and Alaska Fisheries Center, National 

 Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., 

 Seattle, WA 98115. (Background paper submitted to the 27th 

 Annual Meeting of the Standing Scientific Committee of the North 

 Pacific Fur Sea! Commission, March-April 1984. Moscow, 

 U.S.S.R.) 



^York, A. E. 1985. Forecast of the 1985 harvest on St. Paul 

 Island. Unpubl. manuscr. National Marine Mammal Laboratory, 

 Northwest and Alaska Fisheries Center, National Marine Fisheries 

 Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. 

 (Background paper submitted to the 28th Annual Meeting of the 

 Standing Scientific Subcommittee of the North Pacific Fur Seal 

 Commission, March-April 1985, Tokyo, Japan.) 



*Loughlin, T. R., and P. A. Livingston (editors). 1986. Sum- 

 mary of joint research on the diets of northern fur seals and fish 

 in the Bering Sea during 1985. NWAFC Processed Report 86-19, 

 92 p. Northwest and Alaska Fisheries Center, National Marine 

 Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 

 98115. 



374 



