McQuinn: Year-class twinning in sympatric spawning populations of Clupea harengus 
135 
tion in several teleost species (Lett and Doubleday, 
1976; Holdway and Beamish, 1985; Rowe and Thorpe, 
1990), including Atlantic herring (Marti, 1959; Raitt, 
1961; Anthony and Fogarty, 1985; Haist and Stocker, 
1985). Several studies have concluded that length, 
rather than age, is the “critical” factor determining 
the onset of first maturation in herring (Burd, 1962; 
Beverton, 1963; Toreson; 1990). Variations in growth 
rates within a cohort, whether they are density-de- 
pendent or not, will therefore influence the age at 
which different components of a cohort will reach the 
critical length. 
We have seen in the present study that variable 
juvenile growth rates do influence the onset of first 
maturation in herring (the age of maturity) and thus 
affect which season is adopted for spawning. When 
the growth characteristics of a cohort were relatively 
uniform in early life, as represented by the unimodal 
length distribution of the immature 1982 cohort as 
3- year-olds, most of the cohort subsequently matured 
in synchrony and spawned in the spring of 1986 as 
4- year-olds. However, when the immature 3-year-old 
length-frequency distribution showed signs of differ- 
ential growth rates, as with the 1980 spring-hatched 
and 1986 autumn-hatched cohorts, maturation was 
asynchronous. Those individuals from the 1980 co- 
hort with an advanced length at age matured as au- 
tumn spawners in the fall of 1983 at age 3 years and 
4 months. A large proportion of this cohort subse- 
quently spawned the following spring at age 4, and 
the remainder took advantage of an additional 
growth season before spawning in the fall of 1984 as 
autumn spawners. The autumn-spawning individu- 
als of this cohort were therefore significantly longer 
at age than those of the same cohort that remained 
spring-spawning (Fig. 6A). Winter et al. (1986) also 
concluded that the faster-growing spring-hatched in- 
dividuals matured as autumn-spawners. Conversely, 
the 1979 autumn-hatched individuals that became 
spring spawners did so the previous spring at age 3 
years and 8 months and thus showed a shorter length 
at age than those that remained autumn spawners 
and matured at age 4 (Fig. 6B). We also observed 
that the adopted season was maintained after the 
initial spawning because this length difference per- 
sisted until at least age 6. 
This crossover also explains the observed pattern 
of twinning — that is to say a strong spring-spawn- 
ing year class matched with a strong autumn-spawn- 
ing year class from the previous year. The fact that 
twinning is seldom seen between spring- and autumn- 
spawning year classes of the same year is due to the 
ageing convention for herring (Hunt et al. 8 ), which 
does not consider the possibility of crossover between 
these populations. The present study has demon- 
strated that this crossover can occur in both direc- 
tions, i.e. spring-hatched herring can contribute to a 
autumn-spawning year class (1980) and vice versa 
(1986). It should also be mentioned that although 
the effects of crossover between sympatric herring 
populations is more striking when large year classes 
are involved, resulting in year-class twinning, the 
significant correlation found between subsequent 
autumn- and spring-spawning year classes in east- 
ern Newfoundland (Winters et al., 1986) indicates 
that crossover undoubtedly occurs to some extent 
with all year classes. 
The present study therefore does not support the 
concept of discrete sympatric seasonal-spawning 
populations in Atlantic herring. The data presented 
here suggest that the progeny of a given seasonal 
population do not necessarily recruit to the parental 
population but may indeed contribute to a local popu- 
lation of another reproductive season. Furthermore, 
the spawning season that is established at the time 
of first maturation is maintained for the remainder 
of adult life. 
Acknowledgments 
I wish to acknowledge the efforts of Celine Trudeau- 
Simard and Joanne Hamel for the analysis of the 
biological samples, including age and hatching-sea- 
son determinations. I also thank Yvan Lambert and 
two anonymous reviewers for their helpful comments 
on the manuscript. 
Literature cited 
Aneer, G. 
1985. Some speculations about the Baltic herring Clupea 
harengus membras in connection with the eutrophication 
of the Baltic Sea. Can. J. Fish. Aquat. Sci. 42 (suppl. 
1 ):83— 90. 
Anthony, V. C. 
1971. The density dependence of growth of the Atlantic 
herring in Maine. Rapp. P.-V. Reun. Cons. Int. Explor. 
Mer 160:197-205. 
Anthony, V. C., and M. J. Fogarty. 
1985. Environmental effects on recruitment, growth, and 
vulnerability of Atlantic herring Clupea harengus harengus 
in the Gulf of Maine region. Can. J. Fish. Aquat. Sci. 42 
(suppl. 1 ): 158—173. 
Baxter, I. G. 
1959. Fecundities of winter-spring and summer-autumn 
herring spawners. J. Cons. Int. Explor. Mer 25( 11:73-80. 
Beverton, R. J. H. 
1963. Maturation, growth and mortality of clupeid and 
engraulid stocks in relation to fishing. Rapp. P.-V. Reun. 
Cons. Int. Explor. Mer 154:44-67. 
