298 
Fishery Bulletin 96(2), 1998 
Some of the interspecific differences in the utiliza- 
tion of fish prey may also be attributed to differences 
in predator size. The mean size of juvenile coho and 
chinook salmon was greater than that of pink, chum, 
and sockeye salmon. Although the importance of size- 
related variability in prey consumption of juvenile 
salmon is recognized (e.g. Brodeur, 1991), we consid- 
ered that a detailed analysis by predator size was 
beyond the scope of this study. 
Increased fish prey in the diets of juvenile salmon 
in 1984 may reflect higher abundances of larval and 
juvenile stages of certain prey species compared with 
the previous year. Unfortunately, most fish prey spe- 
cies found in the salmon stomachs in our study could 
not be identified. However, we noted opposite trends 
in the CPUE for two potential teleost prey species 
from 1983 to 1984: the CPUE of juvenile walleye pol- 
lock in our seine catches increased 25-fold whereas 
that of juvenile herring declined 40-fold. In 1984, ju- 
venile walleye pollock were important in the diet of 
adult coho salmon (Fisk 3 ); the conditions that pro- 
longed offshore feeding by adult coho salmon on these 
fish in 1984 may also have favored increased 
piscivory by juveniles. 
An increase in available fish prey between years 
may also correlate with an increase in environmen- 
tal temperature. The average SST in 1984 ( x =14.5°C) 
was higher than in 1983 ( x =13.7°C). Increased SST 
in northern waters may have beneficial effects on 
the early life history of fish preyed upon by juvenile 
salmon (Bailey and Incze, 1985). Temperatures in 
outside waters in 1984 followed a long-term warm- 
ing trend related to the 1982-83 El Nino event. Tem- 
perature increased at depth as well as at the sur- 
face, and positive temperature anomalies persisted 
in northern waters beyond 1983 (Cannon et al., 1985; 
Royer, 1985). In coastal waters off Oregon and Wash- 
ington, oceanographic conditions varied greatly ow- 
ing to the 1982-83 El Nino event and affected prey 
species composition, but the proportion of fish bio- 
mass in juvenile salmon diets generally did not vary 
for the same months between years (Brodeur and 
Pearcy, 1990). 
Changes in salmonid diet patterns may reflect den- 
sity-dependent species interactions. Other workers 
have noted such changes with respect to an increase 
in the density of pink salmon, typically the most 
abundant species: 1) the diet of other salmonids be- 
came more diverse, particularly in chum salmon; and 
2) diet overlap among pink, chum, and sockeye 
salmon decreased (Birman, 1969; Andrievskaya, 
3 Fisk, G. 1985. Final report 1984 troll logbook program. 
Alaska Trollers Assoc., 130 Seward St., No. 213, Juneau, AK 
99801, 41 p. 
Table 6 
Estimated mean fullness (0=empty, 6=distended), degree 
of digestion ( l=fresh, 4=completely digested), and percent- 
age of empty stomachs for pooled habitats and time peri- 
ods for juvenile salmon collected in marine waters of south- 
eastern Alaska and northern British Columbia in 1983 and 
1984; n = sample size. 
Habitat 
n 
Stomach 
fullness 
(0-6) 
Degree of 
digestion 
(1-4) 
Empty 
stomachs 
(%) 
Pink salmon 
Inside inlet 
70 
4.0 
1.9 
0 
Inside passage 
223 
3.9 
2.3 
0.4 
Outer coast inlet 
68 
3.5 
2.7 
1.5 
Outside (0-37 km) 
361 
3.0 
2.2 
3.3 
Outside (>37 km) 
93 
4.0 
2.3 
0 
Chum salmon 
Inside inlet 
56 
4.2 
2.3 
0 
Inside passage 
125 
4.7 
2.4 
0 
Outer coast inlet 
62 
4.6 
2.2 
0 
Outside (0-37 km) 
169 
2.6 
2.7 
5.3 
Outside (>37 km) 
41 
3.3 
2.4 
0 
Sockeye salmon 
Inside inlet 
24 
4.3 
2.5 
0 
Inside passage 
58 
4.5 
3.0 
1.7 
Outer coast inlet 
0 
— 
— 
— 
Outside (0-37 km) 
227 
3.4 
2.3 
0.9 
Outside (>37 km) 
52 
3.6 
2.5 
0 
Coho salmon 
Inside inlet 
159 
4.2 
2.3 
1.3 
Inside passage 
205 
4.4 
2.3 
0.5 
Outer coast inlet 
52 
4.2 
2.5 
0 
Outside (0-37 km) 
115 
4.0 
2.4 
0 
Outside (>37 km) 
12 
4.3 
3.0 
0 
1970; Tadokoro et al., 1996). In the Sea of Okhotsk, 
diet overlap was lower in the coastal zone, where 
salmon density was greatest (Andrievskaya, 1970). 
We did not observe such density effects. We found 
significant diet overlap even in the four cases out of 
five where density effects could have been demon- 
strated among these species (see Jaenicke and 
Celewycz, 1994). The exception, when chum salmon 
diet did not overlap significantly with either pink or 
sockeye salmon diets, occurred during July 1984 in 
the outside waters of British Columbia. 
Density-dependent effects on diet may be reflected 
in a third dietary attribute, the amount of food con- 
sumed. In the eastern coastal zone of the Sea of 
Okhotsk, for example, 30% of juvenile pink salmon 
and 66% of juvenile chum salmon sampled with gill 
nets (time of day not presented) had empty stom- 
achs, a condition not found offshore, where juvenile 
salmon density was lower (Andrievskaya, 1970). 
Feeding conditions appeared to be much better for 
