382 
Fishery Bulletin 95(2), 1 997 
and more sunshine in March, resulted in the success 
of these year classes. A continuation of a more sum- 
mer-like (southwesterly) wind pattern through Sep- 
tember may result in better larval herring survival 
in the Gulf of Maine during these years. Average fre- 
quencies of southwest winds off Nova Scotia can vary 
between 10% and 39% for the months of July-Sep- 
tember, 1955-1980 (Hudon, 1994), and summer 
(June-August) wind stress over the eastern conti- 
nental shelf is generally toward the northeast and 
about 0.25 dyn/cm; whereas in fall (September-No- 
vember), wind stress shifts toward the southeast and 
can be twice as strong (Saunders, 1977). 
Effects of turbulent mixing on food encounter rates 
must be considered because southwesterly winds are 
lower in velocity in the Gulf of Maine during the early 
(August-Becember) larval phase. The effects of tur- 
bulence on the availability of zooplankton prey for 
larvae are related to those biological processes (pri- 
mary production) that are disturbed by physical pro- 
cesses, i.e. turbulence generated by wind mixing 
(Rothschild and Osborn, 1988; Sundby et al., 1989; 
MacKenzie and Leggett, 1991). Recently, the overall 
probability of larval feeding has been described as a 
dome-shaped function of turbulent velocity with 
maximum feeding, depending on turbulence level and 
behavioral characteristics of predator and prey 
(MacKenzie et al., 1994). Calmer wind conditions 
through September when most first-feeding larvae 
are present, could enhance larval survival and re- 
sult in the success of these year classes. Strong re- 
cruitment to walleye pollock, Theragramma chalco- 
grarnma, stocks in the Gulf of Alaska (Megrey et al., 
1994) and Bering Sea (Bailey et al., 1986) has been 
linked to initially calm wind conditions and is asso- 
ciated with calm periods preceded and succeeded by 
periods of stronger mixing (Bailey and Macklin, 
1994). For Atlantic herring in the Gulf of Maine, these 
conditions would result from periods of calm south- 
westerly winds in conjunction with stronger south- 
easterly winds. However, strong mixing can disrupt 
layers of prey (Lasker, 1975; Wroblewski and 
Richman, 1987; Owen 1989) and has also been linked 
to reduced growth of Atlantic herring larvae (Heath, 
1989). 
Three other environmental variables, the amount 
of March sunlight, August-December southeasterly 
winds, and November southeasterly storms, were 
related to herring year-class size. Dispersal associ- 
ated with the latter two southeasterly wind factors 
could result in a positive effect on recruitment by 
transporting larvae spawned in the western Gulf of 
Maine and on Jeffreys Ledge toward inshore larval 
overwintering and juvenile nursery areas along the 
Maine coast (Lazzari and Stevenson, 1992). Because 
herring larvae feed on zooplankton, spring phy- 
toplankton production and, ultimately, sunshine 
should be positively related to larval survival. The 
timing of plankton blooms in the Gulf of Maine was 
highly influenced by the amount of sunshine avail- 
able early in the year (Townsend and Spinrad, 1986). 
Therefore reduced sunshine in March would have a 
detrimental effect on the spring bloom, resulting in 
fewer food resources for herring larvae and reducing 
recruitment as seen in Ezzy’s (1988) model using first 
quarter sunshine. 
Recruitment of animals with planktonic stages is 
a complex process; we would not expect any single 
factor affecting the early larval stage to dominate 
the entire survival process (Wooster and Bailey, 1989; 
Campbell and Graham, 1991 ). In our study, although 
more days of southerly winds were generally associ- 
ated with higher than expected age-2 recruitment, 
this was not always the case. For example, in two of 
the six years (1976 and 1984, Fig. 3) when south- 
western winds averaged > 25% higher than normal, 
strong year classes were not produced. Other factors 
may have reduced year class size (e.g. predation on 
larvae or age-1 juveniles) during these periods of 
lower abundance, or some other conditions may not 
have been suitable for prey production or feeding. 
We would have been surprised if the relation of any 
environmental factor and recruitment had always 
been consistent, because a high larval survival rate 
appears to be a necessary, but not sufficient, condi- 
tion for strong recruitment. Year-class strength can 
instead be determined by conditions that prevail 
during the juvenile life stage in some years (Campbell 
and Graham, 1991; Bailey and Spring, 1992). The 
environment does not act alone in affecting recruit- 
ment success; biotic effects, competitive interaction 
between species, and the removal of adults caused 
by fishing mortality, should be considered (Drink- 
water, 1987). The results of our analyses to date are 
interesting and worth expanding, with more research 
and analyses, to other Atlantic herring stocks to de- 
termine the effects of the environment, particularly 
the wind-driven transport of larvae, on their recruit- 
ment variability 
Acknowledgments 
The authors thank several individuals who main- 
tained the Environmental Monitoring Project at the 
Maine Marine Resources Laboratory including L. 
Churchill, D. Smith, and W. Welch. Special thanks 
to D. Libby for invaluable computer assistance and 
to K. Friedland, National Marine Fisheries Service 
who collaborated with us on the stock assessment. 
