FISHERY BULLETIN: VOL. 84, NO. 2 



to late July. Assuming a similar interval between 

 arrival and mating for nonpregnant females, most 

 nonpregnant females would be mated by mid- to late 

 August. Jordan and Clark (1898) stated that young 

 females were impregnated in early August, after old 

 females, and Abegglen et al. (1958) observed that 

 females ages 3 and 4 bred after the harems dis- 

 banded. Also, Craig (1964) reported females 

 ovulated for the first time in late August or Septem- 

 ber. The only evidence that I could find of late 

 mating in a nonpregnant cow was by Osgood et al. 

 (1915), who observed a harem bull mating a female 

 that was "not very young" on 21 August. 



A comparison of the age-specific arrival times for 

 each sex on St. Paul Island (Table 4) largely con- 

 firms the comments by Kenyon and Wilke (1953) and 

 Fiscus (1978) that arrival began progressively earlier 

 with increasing age. However, the current study in- 

 dicated that this phenomenon was obvious only for 

 young ages. It was seen in nonpregnant females 

 ages 1-3 and in males ages 1-6. Although no dif- 

 ferences in arrival times were shown for older males 

 and nonpregnant females, differences could exist, 

 but would be small. The differences in arrival times 

 became progressively less with age for males be- 

 tween 1 and 6 yr and apparently for females 

 between 1 and 3 yr. 



A comparison of the relative numbers returning 

 to St. Paul Island (Table 4) suggests that progres- 

 sively more males and females returned between 

 ages 1 and 3. The cumulative effect of the kill on 

 males of 2 and 3 yr prevented comparisons of abun- 

 dance with males >4 yr. For females, the number 

 of 4-yr-olds returning was probably not greater than 

 3-yr-olds, as suggested by the similarity in the 

 number of 3- and 4-yr-olds killed on hauling grounds 

 by mid-August (Figs. 3, 4). However, pregnancies 

 complicate comparisons of abundance on hauling 

 grounds between females 3 yr and older. Between 

 ages 4 and 10, an increasing proportion of females 

 become pregnant (Lander 1981) and thus go to 

 rookeries rather than hauling grounds. 



The data collected in this study suggest that, with 

 age, young seals of both sexes arrive progressively 

 earlier, and in progressively larger numbers. The 

 reason for these changes in arrival schedules lies in 

 an understanding of the mechanism that controls 

 the migration schedule. However, little is known 

 about this mechamism in the northern fur seal. The 

 mechanism, if it is like that of other vertebrates (see 

 Gauthreaux 1980; Baker 1978), is probably complex. 

 It could involve selective factors, such as food supply 

 and climate, and numerous environmental and 

 physiological factors, such as photoperiod, reproduc- 



tive hormones, and endogenous rhythms. For north- 

 ern fur seals, learned and innate components are 

 likely to be involved. There are several examples of 

 where learning has been suggested to be involved 

 in migration. When the species leaves the Pribilof 

 Islands for the southern migration, juveniles tend 

 to disperse widely in the North Pacific Ocean, preg- 

 nant females tend to travel to the coastal waters off 

 California, and adult males generally remain in the 

 northern Gulf of Alaska (Baker et al. 1970; Fiscus 

 1978). Baker (1978) has suggested that the juvenile 

 northern fur seals may explore the habitat, and, with 

 age, eventually learn the best wintering areas. Also, 

 an increasing proportion of immature seals return 

 to their natal sites on Pribilof Islands with age (Ken- 

 yon and Wilke 1953), although sometimes the natal 

 site is abandoned and a new colony is established, 

 such as at San Miguel Island, CA (Peterson et al. 

 1968). Baker (1978) has proposed that site recog- 

 nition may be learned shortly after birth, and with 

 time, the site is usually relocated. However, other 

 components of migration may be innate. For exam- 

 ple, the annual timing of arrival for pregnant 

 females on St. Paul Island is remarkably precise. 

 Peterson (1968) calculated the mean arrival date to 

 be 30 June for each of 3 years. Such precision seems 

 unlikely to be the result of only learning. Keyes et 

 al. (1971) examined the pineal gland of this species 

 for seasonal variations in hydroxy-indole levels for 

 various ages of males and females, and postulated 

 photoperiodic regulation of the reproductive cycle. 



A physiological event in the lives of young males 

 and females which coincides with the cessation of 

 arriving earlier and returning in greater numbers 

 is the attainment of sexual maturity. Baker (1978) 

 pointed out that sexual maturation controls the ini- 

 tiation of migration in many vertebrates. While a 

 few male northern fur seals begin to produce sperm 

 at 3 yr, most do not do so until about 5 yr (Kenyon 

 et al. 1954; Murphy 1969, 1970). The average female 

 conceives for the first time on her 5th birthday, 

 although typically ovulates for the first time on her 

 4th (Craig 1964; York 1983). Thus, it was during the 

 years of immaturity that young seals gradually syn- 

 chronized their arrival schedules with that of the 

 adults. Perhaps the gradual process of gonad 

 maturation in both sexes over several years plays 

 a role in inducing a cohort to migrate progressively 

 earlier in the year and in causing a greater propor- 

 tion to return to breeding sites. 



A relationship between sexual maturity and 

 changes in arrival times on St. Paul Island could ex- 

 plain two other arrival phenomena noted in this 

 study. In the first case, considerable annual varia- 



392 



