320 
Fishery Bulletin 109(3) 
sizes were from opportunistic sampling, which occurred 
during WinOl, SumOl, and in the Kodiak region. 
Frequently encountered prey types tended to have 
low individual weights, except for juvenile and adult 
euphausiids (Table 2). Copepods, larvaceans, and ptero- 
pods (Thecosomata) each occurred in >20% of the stom- 
achs and had mean individual prey weights <0.8 mg. 
Fish (Osteichthyes) were relatively uncommon (4.3% 
FO) and had the highest mean individual 
weight (626 mg). In contrast, euphausiid 
juveniles and adults occurred in half the 
stomachs and each individual weighed on 
average 19.9 mg. 
The percentage of small-size prey abun- 
dance in juvenile walleye pollock diets de- 
clined with seasonal progression (Fig. 4A) 
and with increased predator length (Fig. 4C) 
and was primarily due to declining percent- 
ages of small copepods and larvaceans and 
increasing percentages of large copepods and 
euphausiids. For euphausiids, the proportion 
of furciliae peaked in SumOl, whereas the 
proportion of juvenile and adult euphausiids 
peaked later, during LSumOl. 
Seasonal change in diet composition by 
weight was most influenced by the pro- 
portion of juvenile and adult euphausiids 
(Fig. 4B), which overall composed 61% of 
the SCW. In the Shumagin region, the pro- 
portional weight of euphausiids increased 
from 36% in LSumOO to 64% in SumOl; 
subsequently, in LSumOl, 51% of the stom- 
ach contents were euphausiid juveniles and 
adults and 40% were fish. Greater seasonal- 
ity was observed in the Semidi and Kodiak 
fish diets. In the Semidi region, euphausi- 
ids were more important during late sum- 
mer (>60%) than during WinOl and SumOl 
(<20%) when diets were dominated by large 
copepods. In the Kodiak region, euphausiids 
represented >60% of the diet in all seasons 
except during SumOl when most (84%) of 
the diet bulk was fish and other epibenthos 
(mostly cumaceans). The low number of Ko- 
diak fish examined during SumOl (20 fish) 
and LSumOl (7 fish) reduced confidence 
in the respective diet compositions. Thus, 
seasonality in diet composition was mostly 
attributable to euphausiids and was less 
evident in the Shumagin region than in the 
Semidi and perhaps Kodiak regions. 
Growth rate 
16 
1.4 
1.2 - 
1 0 
0.8 -| 
0.6 
04 
02 
374 Shumagin Semidi 
Kodiak 
o.o 
Season and year 
Figure 2 
Mean stomach content weight (% body weight [BW] ±1 standard 
error) of juvenile walleye pollock ( Theragra chalcogramma ) col- 
lected with trawl nets in three regions (Shumagin, Semidi, and 
Kodiak) of the western Gulf of Alaska during late summer 2000 
(LsumOO), winter 2001 (WinOl), summer 2001 (SumOl), and late 
summer 2001 (LsumOl). The numbers of fish examined are indicated 
next to each point. 
<L> 
Q. 
£■ 
16 
10 
0 1 1 
0 10 20 30 40 50 60 70 B0 90 100 
Number of stomachs 
Figure 3 
Number of prey types (•) (identified in stomachs of juvenile walleye 
pollock (Theragra chalcogramma) collected in trawl nets in three 
regions of the western Gulf of Alaska during four seasons (August 
2000-September 2001). The x-axis is truncated at 100 stomachs. 
Growth rate estimates ranged from -0.38 
mm SL/d during WinOl to 0.56 mm SL/d 
during LSumOl, but were not estimated for 
all season-region combinations (Table 3). 
For six season-region combinations, cover- 
age of diel periodicity in feeding (e.g., Merati 
and Brodeur, 1996) was deemed inadequate 
because <5 of the eight 3-h time bins that 
compose the diel feeding cycle were repre- 
sented. Consequently, growth rates were 
