136 
Fishery Bulletin 1 10(1 ) 
the increased retention on the bank were also consid- 
ered to be contributing factors to lower egg and larval 
mortality that year. 
Although faster growing larvae generally experience 
greater survival, their timing in the seasonal cycle ap- 
pears to be critical. Buckley et al. (2010) have taken a 
more detailed and theoretical approach to the seasonal 
trends in mortality and growth of cod and haddock 
larvae from the GLOBEC years. The ratio of mortality 
to growth (Ml G) was examined as an index of seasonal 
changes in cohort biomass that provides the greatest 
window of opportunity for survival. Variability in MIG 
can be related to photoperiod, temperature, and prey 
and predator abundance. Although the critical factors 
affecting M and G are tied to the fixed seasonal cycle, 
their deviations can have complex pathways. Predation 
is thought to be the most likely critical factor affecting 
M. Although the temperature range experienced by 
most larvae from a minimum in March to May is only 
a few degrees (4-7°C), there was an extremely large 
increase in M (0.03 to 0.15) — greater than can be ac- 
counted for by metabolic processes alone. Predators also 
increase with the seasonal cycle and their consumption 
rates increase with temperature and day length. Howev- 
er, there was no direct correlation between temperature 
and M; the two warmest years, 1999 and 1995, had low 
M and high M, respectively. 
Planktivorous fish such as Atlantic herring ( Clupea 
harengus ) and Atlantic mackerel (Scomber scombrus ) 
could have a major impact on larvae and pelagic ju- 
veniles as they migrate northward in the spring and 
overlap with patches of larvae on the southern flank 
of Georges Bank (Garrison et ah, 2000). Potentially 
high mortality can occur at the time of settlement dur- 
ing July-August where there is limited gravel-complex 
habitat on eastern Georges Bank to provide protection 
from predation (Lough, 2010). Demersal juvenile fish 
are preyed upon by many piscivorous fish such as spiny 
dogfish ( Squalus acanthias), silver hake (Merluccius 
bilinearis), larger Atlantic cod, and longhorn sculpin 
( Myoxocephalus octodecemspinosus ) (Link and Garrison, 
2002). Tsou and Collie (2001) applied an eight-species 
multispecies virtual population analysis (MSVPA) on 
six important fish species on Georges Bank for 1978-92 
to estimate predation mortalities of age-0 and age-1 
fish. Silver hake accounted for more than 40% of pre- 
dation on age-0 cod, and cannibalism and predation by 
spiny dogfish contributed another 30%. Cannibalism 
was high before 1983, increasing slightly with increas- 
ing stock biomass. Haddock also experienced high pre- 
dation at age-0 with its main predator being cod. We 
reanalyzed a time series (1974-2002) of the Georges 
Bank piscivorous biomass data (silver hake, spiny dog- 
fish, cod) published by Steele et al. (2007, fig. 5b); how- 
ever, no significant correlations were found for cod and 
haddock age-1 recruits, pelagic juvenile abundance, or 
demersal juvenile mortality rates. 
Cod and haddock appear to have different controls 
leading to recruitment. Populations in the low latitudes 
of Georges Bank respond differently than other stocks. 
In this study cod recruitment appears to be set during 
the larval stage during years of low stock abundance, 
but during the juvenile stage when the previous year’s 
recruits are high, indicating a density-dependent pre- 
dation of age-1 fish preying on age-0 group juveniles. 
Young haddock, however, have different morphology, 
physiology, and behavior that allow them to grow rap- 
idly and survive when prey are abundant, and there- 
fore environmental effects, especially during the larval 
stage, can have the greatest control. It is still possible 
for haddock to be controlled during the juvenile stage 
when cod stocks are high. Friedland et al. (2008) exam- 
ined the correlation between the Georges Bank haddock 
1962-2004 survivor ratio (log 10 R/SSB) and the egg- 
larval retention, feeding conditions, juvenile size, and 
estimated hatch date, timing, and magnitude of spring 
and fall phytoplankton blooms and only found a signifi- 
cant relationship for the magnitude of the fall bloom 
and exceptional year classes during the 1998-2004 
period, primarily due to the strength of the 1998, 2000, 
and 2003 year classes. They hypothesized that the in- 
creased production provided increased benthic prey for 
the maturing adults. Their analysis did not rule out 
the contribution of survivors during early life stages 
because the available data were limited and highly 
variable. Also, their March-April zooplankton prey 
data still show a positive trend with the survivor ratio 
and the correlations are only marginally nonsignificant. 
Mountain and Kane (2010) showed that the survivor- 
ship of Georges Bank cod and haddock larvae changed 
between the 1980s and 1990s, coincident with change 
in the zooplankton from a dominant Calanus fimarchi- 
cus community to smaller species such as Pseudocala- 
nus spp. The smaller species may have increased the 
growth and survival of haddock larvae in the 1990s 
because haddock prefer smaller prey for a longer du- 
ration, whereas the larger Calanus community in the 
1980s may have favored cod, which are more selective 
pursuit predators. 
Comparison with other gadoid populations 
For Bering Sea walleye pollock (Theragra chalcogramma ) 
recruitment, it has been suggested that the relative dom- 
inance of activating or constraining factors is responsible 
for the transition from larvae to juveniles (Bailey et 
al., 2005). During periods of changing climate, such as 
observed for the North Pacific, recruitment control for 
pollock could shift between larvae and juveniles owing 
to the dominant mechanisms controlling survival (Cian- 
nelli et al., 2005). A more recent study (Coyle et al., 2011) 
showed that the change from cold to warm periods was 
associated with a shift from large to small copepods and 
that the low survival of age-0 pollock was due to poor 
feeding and increased predation by large predators. 
For Atlantic cod in the high latitude Barents Sea, 
higher recruitment at age-3 tended to be associated 
with warmer years when the spawning sites moved 
northward along the west coast of Norway (Sundby and 
Nakken, 2008). Svendsen et al. (2007) modeled the ef- 
