studies of the tag-estimation procedure were needed. This 

 seems to have been the first reahzation that unknown factors 

 (such as tag-induced mortaUty?) were causing bias in the an- 

 nual Petersen-type estimates. 



At various times over an 8-yr period, adult female reproduc- 

 tive tracts were collected on the snowy Pribilof Islands during 

 November and December. They were examined without suc- 

 cess for early embryos. Finally, on 3 December 1958, three out 

 of five adults taken on St. Paul Island contained implanted 

 embryos, the largest embryo being only 20 mm long (Scheffer 

 1960b; Abegglen et al. footnote 28, p. 20). A proper study of 

 fur seal embryology is still to be done. Material must be col- 

 lected at sea in winter, a time when pelagic sealing is difficult 

 and dangerous. 



David E. Sergeant (Fisheries Research Board of Canada) 

 drev,' plans in 1958 for transplanting Alaskan fur seals to cer- 

 tain islets off northeastern Newfoundland (in letter of 5 Oc- 

 tober 1964). He later decided that southern fur seals (Arc- 

 tocephalus sp.) would be more promising subjects since they 

 do not migrate far from their breeding grounds. 



1959 



The sealing season of 1959 brought a shock — the take of 

 males was only 30,195, the smallest since 1927. Was the new 

 policy of killing females responsible? Biologists argued two 

 points to dispel this idea: 1) Only 10,000 males of the class of 

 1956 appeared in drives in 1959 when 30,000 had been ex- 

 pected. In 1956 the first large killing of females took place. Of 

 27,599 killed, about 12,700 had born a pup that summer 

 (Abegglen and Roppel 1959, table 2). If all of the male pups, 

 or 6,350, had lived to enter the 1959 harvest, the harvest of 

 3-yr-olds would still have been 13,650 short of expectations. 2) 

 The count of dead pups on land in 1956 was the highest on 

 record. Were the survivors in poor body condition when they 

 set out to sea and, as a consequence, did few return to con- 

 tribute to the harvest of 1959? 



The season of 1959 brought widespread realization that sur- 

 vival of individual year classes in the Pribilof herd would fluc- 

 tuate widely as long as the herd remained close to the limiting 

 factors in its environment. Harvest management would be im- 

 proved if a method of forecasting the size of the harvest could 

 be developed. Ways of predicting the size of a 3-yr class were 

 discussed by Chapman (1959)'° among them the following: 1) 

 Counting dead pups and weighing live ones each summer. 

 Both would provide evidence of the general welfare — the start 

 in life — of the current year class. 2) Studying the kill of 2-yr- 

 olds. Over a period of years, a relationship between the kill of 

 2's in one year and 3's the next might be calibrated. A trial 

 prediction was made in 1960; the results were as follows: 



Kill of 3 -year males on St. Paul Island, 1961 



Predicted Actual 



Kill through 31 July 19,000 29,523 



Kill through 7 August 34,500 40,172 



3) Studying the early season kill of 3-yr-olds. The kill for the 

 first 15 d would indicate the probable take for the season. This 

 method was tried in 1959. The forecast for St. Paul Island was 

 9,469; the actual take was 10,494. Such a short-term forecast 

 can never be of much value; it nips too close to the heels of the 

 fact. 4) Studying the weather records for St. Paul Island. The 

 belief was held for a while that, when the mean annual 

 temperature for year X ending 30 June is high, the return of 

 3-yr males in year A" 4- 3 will be high (Abegglen et al. 1961:15"). 

 5) Studying the pelagic take of males, by year class, at ages 1, 

 2, and 3, in advance of the land take. For example, if the class 

 of 1960 contributed strongly to the pelagic take of yearlings in 

 1961, of 2-yr-olds in 1962, and of 3-yr-olds in early 1963, it 

 would presumably contribute strongly to the land take of 3-yr- 

 olds in later 1963 (Abegglen et al. footnote 31, p. 39-41). 



The low harvest of 1959 suggested that crowding in the fur 

 seal population causes increased fluctuation in survival rate of 

 the young. Chapman (in Abegglen et al. footnote 31, p. 71-73) 

 pointed to another demonstration. For the period of 1947-55 

 he showed that the proportion of returning bachelors tended 

 to vary inversely as the size of the year class. That is, when the 

 number of male pups estimated in September was high, the 

 fraction returning 4 yr later was low, around 10-11%. When 

 the number estimated was low, the percentage was high, 

 around 26-28%. (Not only the proportions, but the absolute 

 numbers, of returning 4-yr-olds varied with the pup estimates.) 

 Chapman's correlation has not been challenged in principle, 

 though some of the figures he used in 1959 have been adjusted 

 in the light of recent information. 



It is advisable at this point to discuss a paradox: 1) A high 

 count of dead pups in 1956 was used to explain a low return of 

 bachelors in 1959. 2) A high count of dead pups means poor 

 survival into September and consequently a low estimate of 

 living pups at time of tagging. 3) A low estimate, according to 

 Chapman's correlation, means a high return of bachelors. The 

 explanation: A high count of dead pups is important as an in- 

 dicator of a weak class — one which will experience severe mor- 

 tality later at sea. A high count of dead pups is relatively less 

 important as a factor in the September pup estimate, for dead 

 pups make up only a fraction of the total class. The relation- 

 ship under 1) above is therefore determining. 



Fur seal biologists for a decade or more have wondered "Is 

 mortality equal for the sexes up to age 3 or 4?" Today the 

 evidence points to a higher rate for males. The evidence began 

 to accumulate in the late 1950's when the herd-reduction pro- 

 gram first brought in hundreds of tagged adolescent females, 

 as well as males. As early as 1959, however, Karl Niggol had 

 made a "search for sex disparity in fetal death rates" 

 (Niggol 1960:428). He argued that, if the mortality during ges- 

 tation shifts measurably in favor of the survival of one sex, 

 then we might conclude that the trend persists in the yearlings 

 and 2-yr-olds, both of them classes which are difficult to sam- 

 ple. Niggol tabulated the sex ratios of 3,081 fetuses by lO-d 

 periods during the last 6 mo of gestation. The ratios were re- 



'"Charman, D. G. 1959. Preliminary report on forecasting the kill of male fur 

 seals on the Pribilofs. In C. E. Abegglen, A. Y. Roppel, and F. Wilke. Alaska 

 fur seal investigations, Pribilof Islands, Alaska. Report of field activities, June- 

 September 1959, p. 49-58. Unpubl. rep. Northwest and Alaska Fish. Cent., Natl. 

 Mar. Mammal Lab., Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE., 

 Seattle, WA 981 15. 



"Abegglen, C. E., A. Y. Roppel, A. M. Johnson, and F. Wilke. 1961. Fur 

 seal investigations, Pribilof Islands, Alaska. Report of field activities, June- 

 November 1961. Unpubl. rep., 148 p. Northwest and Alaska Fish. Cent., Natl. 

 Mar. Mammal Lab., Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE.. 

 Seattle, WA 981 15. 



45 



