130 
Fishery Bulletin 95(1 ). 1997 
1979 Autumn-hatched 1980 Spring-hatched 
Figure 2 
Schematic representation of crossover between spring- and autumn-spawning her- 
ring populations. The relative amount of crossover between the 1979 autumn-hatched 
and 1980 spring-hatched cohorts is contrasted with that between the 1981 autumn- 
hatched and the 1982 spring-hatched cohorts. 
crossover was much less important in absolute terms 
in the latter (Fig. 2). This pattern is different for the 
1986 and 1987 autumn- and spring-hatched cohorts 
(Table 1). There appears to have been a larger net 
migration (31%) from the autumn-hatched cohort to- 
wards the spring-spawners at age 4 (age=no. of rings 
+ 1). Furthermore, the 1979 autumn-spawning year 
class showed a trend of a decreasing proportion of 
spring-hatched individuals from age 4 to 6 as more 
autumn-hatched individuals matured and recruited 
to the year class (Table 1). Conversely, the 1981 au- 
tumn-spawning year class showed an increasing per- 
centage of spring-hatched individuals with age ow- 
ing to the overwhelming dominance of the large 1982 
spring-hatched cohort compared with the small 1981 
autumn-hatched cohort (Fig. 2). 
I have also summarized the mean lengths and stan- 
dard deviations of the immature 1980 and 1982 
spring-hatched cohorts to compare their average 
growth characteristics prior to their first spawning 
(Table 2). Although the means are similar at similar 
ages, the standard deviations are quite different, 
those for the 1980 cohort being 44% to 79% greater 
than the 1982 cohort from age 2 to age 3. The differ- 
ence between the two cohorts is even more obvious 
when their length-frequency distributions at age 3 
are compared (Fig. 3, A and B). The length distribu- 
tion of the 1980 cohort is wider and bimodal. The 
