496 
Fishery Bulletin 99(3) 
collected suggests that the fish fed in a different area or at 
a different time. 
Comparison of allopatric and sympatric 
aggregations 
If competition occurs between sympatric species, one would 
expect that, given similar prey fields, the quantity or qual- 
ity of prey consumed would improve when fish are allo- 
patric. If such changes are prolonged, growth or survival 
differences are probable. We found intraspecific changes in 
diet composition and food quantity for both species from 
allopatric to sympatric aggregations, and weak interspe- 
cific differences in feeding between sympatric fish. Both 
species of fish avoided small calanoids whether a competi- 
tor was present or not, but the proportional use of ener- 
getically advantageous taxa (large calanoid, euphausiid, 
and hyperiid prey) and the selection for large calanoids 
were lower in the presence of competitors. Comparisons 
of sympatric fish showed that pollock ate proportionally 
more high energy-producing prey (euphausiid biomass, 
hyperiid frequency, large calanoid numbers) and selected 
Metridia spp. more strongly than herring. Overall, how- 
ever, high diet overlap between allopatric and sympatric 
fish of either species in autumn indicated little change 
in diet composition owing to sympatry. Because both spe- 
cies of sympatric fish ate less total prey than allopatric 
fish, quantity, rather than quality of prey decreased in 
the presence of competitors or perhaps as a density- 
dependent response. Drastic diet shifts were demonstrated 
in one study involving fish removal; planktivorous spe- 
cies became benthivorous when their competitors were 
removed from a lake (Persson and Hansson, 1999). In 
our study, however, although the intraspecific comparisons 
(allopatric-sympatric) were within month, they differed 
by depth, time, or region; the interspecific comparisons 
of allopatric fish were a month apart in different years. 
Many authors have found differences in zooplankton prey 
fields on these scales 10 (e.g. Springer et al., 1989; Celewycz 
and Wertheimer, 1996), making it difficult in our study to 
separate the effects of sampling differences from those due 
to competition on diet and feeding. 
Seasonal feeding 
We observed the highest frequency of empty stomachs for 
both species in the November 1994 samples; empty stom- 
achs, however, were more frequent for herring than for 
pollock. Similarly, in other seasonal studies, the propor- 
tion of empty stomachs in YOY herring peaked in winter 
(Foy and Norcross, 1999a; Foy and Paul, 1999), but empty 
stomachs were never observed among YOY pollock, even 
though stomachs were least full in December. 12 As zoo- 
plankton biomass declined between the winter months of 
October and February, the feeding response and whole 
body energy content of herring also declined (Foy and Paul, 
1999). The herring relied on stored energy to overwinter 
(Paul et ah, 1998a; Paul and Paul, 1998), but pollock are 
thought to allocate energy from year-round feeding for 
somatic growth (Paul et al., 1998b). 
We also observed declines in zooplankton coincident with 
feeding declines in autumn; conversely, a greater food sup- 
ply of small calanoids supported trends toward more in- 
tensive feeding in summer. Feeding on small calanoids in 
summer could have been density dependent because this 
taxon was found to be most prominent in the diets when it 
was more abundant in zooplankton samples. Juvenile Bal- 
tic herring exhibited a similar functional response to great- 
er food densities (Arrhenius and Hansson, 1999). However, 
decreased feeding on small calanoids is more likely related 
to fish size and energy requirements. Both species (espe- 
cially pollock in November 1994) were larger in autumn 
when they ceased consuming small calanoids, and both 
species avoided them more often. Larger prey are more effi- 
cient sources of energy for larger fish, if available. The larg- 
er autumn fish were probably better able to prey on late 
stages of macrozooplankton than the summer fish (Merati 
and Brodeur, 1996; Haegele, 1997; Kamba, 1977), in syn- 
chronism with a seasonal difference in their abundance 
and availability 13 (Bollens et al., 1992a; Stone and Jessop, 
1994; Incze et al., 1997). A shift to larger prey is consis- 
tent with a decrease in prey numbers without a change in 
prey weight because fewer large prey are needed for equiv- 
alent weight. The biomass percentages of large prey con- 
sumed (large calanoids and euphausiids) and the numeric 
percentages of small prey consumed (larvaceans) both in- 
creased in autumn compared with summer (see also Foy 
and Norcross, 1999a). Ontogenetic changes and size-selec- 
tive feeding have been reported by others, as well, for juve- 
nile herring (e.g. Raid, 1985; Munk, 1992; Arrhenius, 1996) 
and juvenile pollock (Grover, 1991; Brodeur, 1998). 
In conjunction with the seasonal diet transition, total 
prey number and most other measures of quantity were 
lower in autumn than in summer, particularly for her- 
ring. Autumn declines in juvenile pollock feeding rates 
were also observed in a study in Southeast Alaska in con- 
junction with a switch from small prey to larger prey. 12 
Similarly, feeding declined seasonally for Atlantic herring 
(de Silva, 1973), Baltic herring (Arrhenius and Hansson, 
1999), and Pacific herring (Foy and Norcross, 1999a). In 
our study, pollock consumed well above maintenance ra- 
tion (0.30 %BW at 7.5°C; Smith et al., 1986) in all time 
periods, but, the low prey %BW of herring in autumn 1994 
could indicate starvation, given a ration of 1.3-3. 6 %BW 
at 6.2-8.7°C (de Silva and Balbontin, 1974; Arrhenius and 
Hansson, 1994). Continued feeding and better fish condi- 
tion late in the year are advantageous for survival through 
the extreme conditions of winter, and relate to the species 
different strategies for overwintering. 
Declines in feeding with season were also indicated by 
increased diet overlap between sympatric species in No- 
vember compared with overlap in October, which suggests 
a density-dependent convergence of feeding with declining 
prey resources in late autumn. For sympatric fish in au- 
13 Mooney, J. R. 1999. Distribution, energetics, and parasites 
of euphausiids in Prince William Sound, Alaska. Unpubl. 
MS thesis, Juneau Center School of Fisheries and Ocean Sci- 
ence, Univ. Alaska Fairbanks, 11122 Glacier Hwy., Juneau, AK 
99801, 172 p. 
