36 
Fishery Bulletin 95( I ), 1997 
fry and subyearling chinook salmon often use tidal 
marshes where they eat drift and emergent insects 
and epibenthic crustaceans (Kjelson et al., 1982; 
Simenstad et al., 1982; Levings et al., 1991; Shreffler 
et al., 1992), whereas, larger, yearling chinook salmon 
spend little time in salt marshes but quickly move to 
neritic habitats (Simenstad et al., 1982). When 
subyearling fish move to neritic habitats their diet 
shifts to fishes, decapod larvae, euphausiids, and drift 
insects (Simenstad et al., 1982). McCabe et al. (1986) 
observed that, in the Columbia River estuary, 
subyearling chinook salmon in pelagic areas were 
significantly larger than those caught in shallow in- 
tertidal habitats and that the prey of juvenile chinook 
salmon varied with season, habitat, and position in 
the estuary. Feeding behavior is also influenced by 
environmental factors, for example turbidity (Gre- 
gory and Northcote, 1993). 
Diets of juvenile chinook salmon in freshwater 
reaches of river systems often are dominated by lar- 
val, pupal, or adult insects that are captured mainly 
in the drift at the surface or in the water column 
(Becker, 1973; Craddock et al., 1976; Sagar and 
Glova, 1987, 1988; Rondorf et al., 1990; Healey, 1991; 
Levings and Lauzier, 1991; Smirnov et al., 1994). 
Depending on season and habitat, both terrestrial 
insects as well as different developmental stages of 
aquatic insects can be important prey for chinook 
salmon in rivers (Rondorf et al., 1990; Levings and 
Lauzier, 1991). Insects are also important constitu- 
ents of the diets of juvenile chinook salmon in many 
estuaries (Healey, 1980, a and b, 1982, 1991; Levings, 
1982; McCabe et al., 1986; Kask et al., 1988; this 
study), particularly in fresh or brackish water tidal 
marshes (Kjelson et al., 1982; Levings et al., 1991; 
Shreffler et al., 1992). 
Whereas insects are important prey in freshwater 
and upper estuaries, fishes are important prey of 
juvenile chinook salmon constituents in the lower 
reaches of estuaries as well as in marine, neritic or 
subtidal areas (Healey, 1980a; Myers, 1980; Kjelson 
et al., 1982; Simenstad et al., 1982; Argue et al., 1986; 
McCabe et al., 1986; Levings et al., 1991; Reimers et 
al. 3 ; Nicholas and Lorz 4 ). Fish prey are also predomi- 
nant in the diets of juvenile chinook salmon in ma- 
rine waters off Oregon and Washington (Peterson et 
3 Reimers, P. E., J. W. Nicholas, T. W. Downey, R. E. Halliburton, 
and J. D. Rogers. 1978. Fall chinook ecology project, AFC- 
76-2. Federal Aid Progress Reports, Fisheries. Oregon Dep. 
Fish and Wild]., 2501 S.W. First Ave., PO. Box 59, Portland, 
OR 97207. 
4 Nicholas, J. W., and H. V. Lorz. 1984. Stomach contents of 
juvenile wild chinook and juvenile hatchery coho salmon in sev- 
eral Oregon estuaries. Oregon Dep. Fish and Wildl., 2501 S.W. 
First Ave., P.O. Box 59, Portland, OR 97207. Progress Rep. 
84-2, 9 p. 
al., 1982; Emmett et al., 1986; Brodeur and Pearcy, 
1990, 1992; Brodeur et al., 1992), in the Gulf Islands 
area of the Strait of Georgia (Healey, 1980b), and in 
the Fraser River plume (St. John et al., 1992). 
In Coos Bay, the increase in importance of marine 
fish in the diets of juvenile fall and spring chinook 
salmon at the lower-bay stations may reflect an up- 
per-bay, lower-bay gradient in the abundance of fish 
prey. Juvenile osmerids, sandlance, and rockfish were 
the predominant fish prey of juvenile chinook salmon 
in Coos Bay. In Yaquina Bay, larval and juvenile 
stages of these species were present in peak abun- 
dances in plankton samples from the extreme lower- 
bay and offshore stations (Pearcy and Myers, 1974). 
Myers (1980) caught more species of fishes in the 
lower than in the upper section of Yaquina Bay and 
suggested that much of the food for juvenile chinook 
salmon residing in the bay was supplied by tidal ex- 
change with the ocean. She also suggested that high 
temperatures in the upper bay inhibited movement 
of predominantly marine species into the upper bay. 
A similar mechanism may be operating in Coos Bay. 
In our beach-seine samples large juvenile and adult 
surf smelt were much more abundant at lower than 
at mid-bay stations (average catch per set was 2,290, 
237, 108, 30, and 12 at stations 1, 2, 3, 4, and 5, re- 
spectively). If, as was the case in Yaquina Bay, smaller 
larval and juvenile smelt also are more abundant in 
lower Coos Bay, the increased consumption by juve- 
nile chinook salmon of these fish prey in the lower bay 
may be a consequence of their greater density there. 
Acknowledgments 
We thank John Chapman for assistance in the iden- 
tification of gammarid amphipods. Jim Digiulio of 
Aquatic Biology Associates identified the insect prey. 
Carl Brookins, Alton Chung, Matt Wilson, Ann Raich, 
and Karen Young assisted in the field, and Terrin 
Ricehill assisted in the laboratory; their help is 
greatly appreciated. Three anonymous reviewers 
provided very helpful comments on earlier drafts of 
this manuscript. This study was supported by NOAA, 
Northwest Fisheries Science Center, Grant 
NA47FE0182-02. 
Literature cited 
Argue, A. W., B. Hillaby, and C. D. Shepard. 
1986 . Distribution, timing, change in size, and stomach 
contents of juvenile chinook and coho salmon caught in 
Cowichan Estuary and Bay, 1973, 1975, 1976. Can. Tech. 
Rep. Fish. Aquat. Sci. 1431, 168 p. 
