Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragro chalcogramma and Clupea pallasi 
497 
tumn, lower diet overlap occurred when pollock were more 
selective for large calanoids and ate proportionally more 
euphausiids, and the herring selected larvaceans (October 
1995); greater dietary overlap in autumn occurred when 
the species ate similar percentages of these taxa and small 
calanoids (November 1994). 
Despite larger size in autumn, energy requirements of 
fish are lower during this period of less abundant food 
due to environmental changes. In PWS, surface temper- 
atures declined from approximately 12°C in summer, to 
10°C in October 1995, and 7-8°C in November 1994. 2 Tem- 
peratures below the thermocline were stable at 5-6°C, 
but thermocline depth varied seasonally, between approx- 
imately 40-60 m in summer and 50-90 m in either au- 
tumn. 34 Additional seasonal changes in the environment 
have been reported by Stokesbury et al. ( 1999). From sum- 
mer to winter in PWS, surface waters inside bays changed 
from being colder to being warmer than surface waters 
outside bays. Such environmental temperature changes in 
both the vertical and horizontal planes may influence the 
distribution of pollock and herring, their energy budgets 
(Smith and Paul, 1986), and the seasonal distribution and 
abundance of their prey. 
Fish can conserve energy during times of reduced prey 
by altering behaviors to maximize prey searching or to de- 
crease metabolic costs. For example, fish can alter school 
cohesiveness (Brodeur and Wilson, 1996b; Stokesbury et 
al., 1999), restrict their movement, or shift in residence to 
deeper water with colder temperatures (Sogard and Olla, 
1996; Ciannelli et al., 1998). However, they may do so at 
the risk of increased predation or decreased light for feed- 
ing. Herring are primarily visual feeders, requiring mini- 
mum light levels to feed (Blaxter and Hunter, 1982) but 
may also filter feed in low light (Batty et al., 1986; Stone 
and Jessop, 1994). They can respond to prey distributions 
correlated with thermocline depth (Fossum and Johannes- 
sen, 1979) and may shift feeding during the day to depths 
where light is not restrictive even if prey are concentrat- 
ed elsewhere (Munk et al., 1989). Similarly, the vertical 
distribution of juvenile pollock is affected by the relative 
availability of food and by numerous other factors that af- 
fect feeding conditions: predator presence, light, turbidity, 
pressure, temperature, size, and metabolic requirements 
(Bailey, 1989; Olla et al., 1996; Sogard and Olla, 1996; Ci- 
annelli et al., 1998). Juvenile pollock avoided light more 
and avoided cold water less with growth, especially under 
conditions of low zooplankton (Olla et al., 1996). Both ju- 
venile pollock (Smith et al., 1986) and YOY herring (de 
Silva and Balbontin, 1974; Arrhenius and Hansson, 1994) 
decreased prey consumption at colder temperatures; pol- 
lock also had lower maintenance rations and grew more 
rapidly under conditions of low food with colder temper- 
atures (Smith et al., 1986). Competition with pollock for 
food could compromise the winter survival of YOY herring 
by limiting the accumulation of their energy stores. 
Research showing major differences in the nutritional 
quality of forage species consumed by piscivores has 
sparked interest in their trophic interactions. Pollock lipid 
content was low compared with that of herring, but un- 
like herring, was not correlated with size (Anthony and 
Roby, 1997; Payne et al., 1999; Anthony et al., 2000). How- 
ever, only a few studies have examined the energetic con- 
sequences of feeding on different prey for juvenile pollock 
(e.g. Smith et al., 1986; Davis and Olla, 1992; Ciannelli 
et al., 1998) and herring (e.g. de Silva and Balbontin, 
1974; Arrhenius and Hansson, 1999). Because prey differ 
in nutritional composition, caloric density (Ikeda, 1972; 
Lee, 1974; Deibel et al., 1992; Davis et al., 1998), size, mo- 
bility, and behavior, their relative abundance is not the 
only factor of importance. For example, larvaceans are a 
highly visible taxon (Bailey et al., 1975) with caloric val- 
ue per unit weight similar to that of crustacean zooplank- 
ton even though they are gelatinous (Davis et al., 1998), 
but many more larvaceans must be consumed to accumu- 
late the equivalent calories obtained from the crustaceans. 
Davis and Olla (1992) showed in a controlled experiment 
that larval pollock growth, behavior, and lipid concentra- 
tion were affected by the nutritional quality of prey. In a 
field study, herring diets had the highest energy density 
of all in May, when large calanoids were the most im- 
portant taxon (Foy and Norcross, 1999a); however, they 
were not examined from late fall — a period for which our 
study showed that euphausiids were the prominent prey 
and others have shown they contain higher energy den- 
sity compared with earlier times of the year. 13 If sympatry 
induces feeding on prey of lesser nutritional quality for 
extended periods because of interference competition, fish 
growth and survival could be affected. 
For sympatry to occur, the distribution of juvenile wall- 
eye pollock and Pacific herring must overlap in three di- 
mensions: time (seasonal and diel), and both horizontal 
and vertical space. These species have different life his- 
tories (Smith, 1981; Lassuy, 1989) and patterns of move- 
ment change ontogenetically, suggesting that spatial over- 
lap is likely to vary. YOY herring generally school near the 
bottom along shore during the day, then move up to the 
surface at dusk and disperse (Blaxter and Hunter, 1982; 
Lassuy, 1989; Haegele, 1997). Early YOY pollock stay prin- 
cipally in surface water above the thermocline, perform a 
diel vertical migration (DVM), and disperse or move in- 
shore at night; depth distribution increases from summer 
to autumn 12 and with ontogeny (e.g. Nakatani, 1988; Karp 
and Walters, 1994; Olla et al., 1996). Stokesbury et al. 
(1999) found that herring and pollock were generally depth 
stratified but that both were aggregated in bays in July 
and October. Rather than having a strong species associa- 
tion, they may simply have an affinity for the same habitats 
at some points in their life histories (Brodeur and Wilson, 
1996a). Data collected monthly in PWS in 1994 (Willette et 
al., unpubl. data) showed that, after May, >50% of juvenile 
herring sets also caught juvenile pollock, and after July, 
>50% of juvenile pollock sets also caught juvenile herring. 
Willette et al. (1997) noted that diet overlap was more than 
twice as great for pollock and herring from sympatric sites 
than for these species from allopatric sites in late summer. 
These patterns suggest that sympatry and feeding competi- 
tion increase from spring through summer. 
Our study is the first to examine the feeding interac- 
tions of juvenile walleye pollock and Pacific herring. Its 
limited scope, because it was designed for other primary 
