Wilson et al.: Regional variation in the feeding cycle of |uvenile Theragra chalcogramma 
323 
are spawned during spring (Pinchuk et al., 2008) and 
then develop and grow during the subsequent summer 
(Pinchuk and Hopcroft, 2007). This can explain why 
the peak proportion of furciliae in fish stomachs pre- 
ceded that of juvenile and adult euphausiids. Euphau- 
siid population density appears to peak during autumn 
(Coyle and Pinchuk, 2005) or winter (Cooney, 1986) 
owing perhaps to prespawning aggregation behavior 
and sampling bias (e.g., Siegel, 2000). We contend that 
some fraction of these euphausiids were too big to be 
available as prey of late age-0 and early age-1 walleye 
pollock. Any starvation-related shrinkage in euphausiid 
body size (Pinchuk and Hopcroft, 2007) might therefore 
benefit juvenile walleye pollock. Sometime between the 
period when walleye pollock growth resumes in early 
spring (Wilson et al., in press) and the period when 
they consume large euphausiids in late summer (Wilson 
et al., 2009), yearling walleye pollock gain the ability 
to consume all sizes of the euphausiids within their 
foraging ambit. Because juvenile walleye pollock and 
euphausiid size compositions vary geographically (e.g., 
Wilson et al., 2009), we predict that regional variation 
occurs in the timing when juveniles gain full access to 
all sizes of euphausiids, which are a principal determi- 
nant of food-related habitat quality. 
Superior feeding conditions before and during winter 
in the Kodiak region can explain why resident juveniles 
were bigger and in better body condition than other 
juveniles (Buchheister et al., 2006). Superior feeding 
conditions were indicated by a euphausiid-rich diet 
and fuller stomachs. Euphausiid-rich diets may simply 
reflect the advantage of a body size that enabled the 
consumption of large euphausiids. However, recent re- 
search during late summer indicates that euphausiid 
populations in the Kodiak region can be substantially 
denser than in the Shumagin and Semidi regions (se- 
nior author, unpubl. data) and perhaps reflect different 
oceanographic characteristics. The Kodiak region is 
characterized by onshore flow into sea valleys, which 
were associated with euphausiid concentrations (Loger- 
well et al., 2010). In contrast, much of the shelf area in 
the Shumagin and Semidi regions is occupied by Semidi 
Bank, which causes offshore flow (Schumacher and 
Reed, 1986). Few euphausiids occur inshore of and over 
Semidi Bank, although concentrations are found over 
the adjacent continental slope and sea valleys (Wilson, 
2009). Euphausiids may also experience less grazing 
pressure in the Kodiak region than in the Shumagin 
and Semidi regions where juvenile walleye pollock can 
be more abundant (senior author, unpubl. data). 
During summer 2001, the best food-related habitat 
quality for juvenile walleye pollock appeared to have 
shifted to the Shumagin region. The Shumagin juve- 
niles experienced the strongest postwinter rebound in 
stomach fullness and no decline in the dietary propor- 
tion of euphausiids. This shift in habitat may reflect 
the relatively high abundance of small-size euphausi- 
ids observed later in the region (Wilson et al., 2009). 
In contrast, Semidi and Kodiak juveniles exhibited a 
relatively weak postwinter rebound in stomach fullness 
and relatively low euphausiid dietary proportions. The 
high proportion of epibenthic animals (e.g., cumaceans) 
among Kodiak fish during summer 2001 may be in- 
dicative of supplemental feeding during times of pelagic 
prey scarcity (Brodeur and Wilson, 1996). The better 
habitat quality in the Shumagin region may have been 
transitory as the diets of juveniles in the Semidi and 
Kodiak areas again became relatively euphausiid rich 
during late summer 2001, although the Shumagin mean 
SCW stayed relatively high. We acknowledge that our 
interpretation of apparent differences between Shuma- 
gin and Kodiak is tenuous given the paucity of Kodiak 
fish examined — a paucity that can bias dietary breadth 
(Ferry and Cailliet, 1996) and reduce estimate preci- 
sion, but our interpretation is consistent with season 
and region differences in body condition. 
Additional support of a postwinter shift in the re- 
gion of most-favorable feeding habitat is provided by 
regional differences in whole-body energy density. In 
concert with their relatively strong postwinter rebound 
in stomach content weight, Shumagin fish exhibited a 
stronger rebound in whole-body energy density than 
Semidi and Kodiak juveniles, and the Shumagin juve- 
niles had high energy densities during summer 2001 
(Buchheister et al., 2006). We therefore hypothesize 
that the area most favorable for juvenile walleye pol- 
lock feeding is determined by euphausiid availability 
and shifts from the Kodiak region during winter to 
the Shumagin region during summer. Whether this 
change extends to seasonal shifts in walleye pollock 
nursery location, as distinguished from less productive 
juvenile habitat (Dahlgren et al., 2006), depends on the 
impact that the variation in food habits has on growth 
and survival. 
In the field context of the present study, juvenile wall- 
eye pollock food habits and associated body condition 
have direct implications for growth rate. The relevance 
of food to growth was underscored by the similarity of 
our growth estimates from food-based modeling with 
those from independent otolith-based studies. Otolith 
growth of age-0 walleye pollock 5 days before capture 
in late summer 2000 indicated body growth rates of 
0.03 to 0.19 g/d (Mazur et al., 2007), which encompass 
our estimates of 0.03 and 0.04 g/d (see Table 3, mul- 
tiply the specific growth rate by body weight). During 
2001, the peak growth rate of yearling walleye pollock 
estimated from growth increments in otoliths was 0.59 
mm SL/d (Wilson et al., in press), which was slightly 
above our late summer 2001 estimates of 0.39 and 0.56 
mm SL/d. We found no estimates of growth rate during 
winter, but the average of our two new wintertime esti- 
mates (-0.03 mm SL/d) is consistent with the negligible 
change in body length of juveniles throughout the west- 
ern GOA (Brodeur and Wilson, 1996). The difference 
between our winter estimates reflects the difference 
in dietary proportion of euphausiids associated with 
differences in stomach content weight and diet energy 
density. We speculate that similar region-specific di- 
etary differences explain the previously observed faster 
growth (Bailey et al., 1996) and larger body length 
