FISHERY BULLETIN: VOL. 84, NO. 2 



mainly on an analysis of seasonal changes in the 

 number of animals killed of each age during 

 harvests. I examine the evidence for arrival times 

 by order of decreasing age within each sex, and com- 

 pare the relative numbers returning for each age 

 of young seals. The published and unpublished 

 literature on northern fur seals is reviewed for in- 

 formation on arrival times and abundance. The rela- 

 tionship between arrival schedules, relative number 

 returning, and onset of sexual maturity is discussed. 



METHODS 



The kill data from St. Paul Island used in this 

 study were collected during 1956-82 by the National 

 Marine Mammal Laboratory, National Marine Fish- 

 eries Service, Seattle. Most data up to 1979 were 

 listed by Lander (1980), who noted the method of 

 data collection and the number killed by age, sex, 

 date, and location. Kozloff (1981 2 , 1982, 1983) listed 

 the data collected during 1980-82. Abegglen et al. 

 (1956, 1957, 1958, 1959) determined the age-specific 

 pregnancy rates of females killed during 1956-59. 

 These authors considered a female to be pregnant 

 when parous (carrying a term fetus), or recently 

 postpartum (lactating or uterus involuting). They did 

 not separate females into these two categories, or 

 determine whether postpartum females were carry- 

 ing a new conceptus. 



Almost all males and females were killed on haul- 

 ing grounds rather than on rookeries. No commer- 

 cial kills for males took place on rookeries, and only 

 a few took place for females. Typically, the kill of 

 both sexes on hauling grounds was made between 

 late June and mid- August. It consisted of a series 

 of consecutive 5-d circuits, or rounds, of all hauling 

 ground sites. During each round, a crew undertook 

 one killing operation at each site, and killed all seals 

 present of a particular sex and length. The body 

 length limits for harvesting were set in inches (in) 

 from nose to tip of tail, or from nose to base of tail. 

 I converted all lengths to cm and standard length, 

 using 1 in for tail length. Lander (1980) and the 

 North Pacific Fur Seal Commission (1984b) noted 

 the annual changes in management practices on St. 

 Paul Island. The changes included variations in body 

 length limits, kill dates, quotas, kill locations, and 

 special kills for sex and age. I used only data that 

 were collected under comparable management 

 restrictions. 



2 P. Kozloff (editor). 1981. Fur seal investigations, 1980. 

 NWAFC Processed Rep., 96 p. National Marine Mammal Lab- 

 oratory, National Marine Fisheries Service, NOAA, Seattle. 



Probit plots of age-specific cumulative length fre- 

 quencies were used to determine the percentage of 

 males and females of each age present in the kill for 

 each set of length limits. Sufficient age-length data 

 were not available for the plots from kills made on 

 St. Paul Island, but were available from samples col- 

 lected pelagically for research purposes by the 

 United States and Canada under the terms of the 

 North Pacific Fur Seal Commission. These data are 

 on file at the Pacific Biological Station, Nanaimo, 

 and at the National Marine Mammal Laboratory, 

 Seattle. The age-length data used were from seals 

 collected near St. Paul Island during June-August 

 1958-74. The lengths of females used were those of 

 postpartum and nonpregnant seals, the main cat- 

 egories of females killed on land. 



I assumed that seals were arriving on St. Paul 

 Island when the number killed increased in suc- 

 cessive rounds and that arrival was completed when 

 the number killed reached an asymptote. These 

 assumptions were valid only under certain circum- 

 stances. One was that all seals encountered of a 

 designated sex and length were killed, which was 

 the case. Another was that the number of seals 

 hauled out, and thus available for killing, was a con- 

 stant proportion of the number alive through the 

 harvest season. The assumption seems reasonable 

 in that Gentry (1981) estimated an average of about 

 19% of marked juvenile males were ashore at any 

 one time on St. George Island. Finally, the propor- 

 tion of a particular age and sex killed during each 

 year must have been sufficiently small so as not to 

 have substantially reduced cohort size, and thus 

 altered the trend in numbers killed by round. This 

 qualification was probably true for all ages, except 

 perhaps for 4-yr-old males. Lander (1981) estimated 

 the harvest utilization rate of males on St. Paul 

 Island to be only 2.8% for age 2, 40.3% for age 3, 

 14.7% for age 5, but 57.3% for age 4. Escapement 

 rates of females from the commercial harvest were 

 not calculated, but were probably high. The females 

 killed were mainly of ages 3 and 4 with the largest 

 annual take for age 3 in the years studied being 

 9,700, and for age 4 being 6,300. These figures com- 

 pared with about 55,000 and 48,000, respectively, 

 for females present in the whole population, based 

 on Lander's (1981) life table for the species. 



The number killed of each age up to the last day 

 of each round for each year was plotted to describe 

 the seasonal change in numbers killed. For males, 

 the most common last-day dates for each round were 

 in the series of 5-d rounds ending between 1 July 

 and 5 August. For years in which the dates for last- 

 day rounds differed from this series, the number of 



384 



