628 



Fishery Bulletin 88(4), 1990 



regions (Brodeur and Pearcy 1986), which could result 

 in different feeding conditions for juvenile salmonids 

 within each region. The Columbia River region may be 

 quite different hydrographically and biologically from 

 the upwelling regions off Washington and Oregon due 

 to the presence of a warm, low-salinity plume extend- 

 ing over much of the Columbia River area during the 

 summer. Oceanic species (e.g., Sebastes spp., E. paci- 

 fica) rarely occurred in the stomachs collected from the 

 Columbia region, whereas £■. niordax, a species whose 

 northern subpopulation spawns within the relatively 

 stable Columbia River plume (Richardson 1980), was 

 generally well represented in salmon stomachs col- 

 lected in this region. 



The relatively minor cross-shelf variations in the 

 major prey taxa consumed by coho and chinook were 

 not expected, considering that many studies have found 

 substantial variations in cross-shelf species distribu- 

 tions (Peterson and Miller 1976, Richardson and Pearcy 

 1977, Richardson et al. 1980, Shenker 1988). However, 

 the frequency of occurrence of major prey may not be 

 a representative measure of the diet from a particular 

 area. The larvae of many meroplanktonic taxa (e.g., 

 Sebastes spp., Hemilepidotus spp., and Cancer spp.) are 

 generally found offshore and progressively migrate or 

 are transported inshore as they grow prior to settling 

 to an inshore benthic habitat as juveniles (Richardson 

 et al. 1980, Shenker 1988). Euphausiids consumed 

 in our inshore study area were mainly T. spinifer-a, 

 whereas E. pacifica were consumed in the offshore 

 area. These euphausiid species have little overlap in 

 distribution (Hebard 1966). Hyperiid amphipods were 

 also represented by an inshore species, Hyperoche 

 medusarum, and an offshore species, T. pacifica (Lorz 

 and Pearcy 1975). These and other cross-shelf species 

 differences were not detectable when analyzing stom- 

 ach contents at higher taxonomic levels. However, 

 species-level distinctions may be irrelevant to a forag- 

 ing predator, if the size, energy content, and behavior 

 of both prey species are similar. 



The high prey-diversity and generally high niche- 

 breadth values agree with previous studies which 

 indicate that many of these salmon species are not 

 specialists in their oceanic feeding modes, but rather 

 consume any available prey within the proper size- 

 range. Many of the same geographic, interannual, and 

 seasonal patterns found in the feeding habits of the 

 adult salmonids and pelagic nonsalmonid species 

 (Brodeur et al. 1987a) were found in our study. These 

 similar patterns suggest that the zooplankton and 

 ichthyoplankton population cycles, which are intricately 

 coupled to seasonal production cycles, may be impor- 

 tant determinants of the feeding ecology of these 

 salmon species. The ability to switch to alternate prey. 



when preferred prey are limiting, may be an important 

 factor in the marine survival of salmon. 



The high overlaps between coho and chinook juve- 

 niles seen for some craises or individual collections may 

 signify that some competition for prey may be occur- 

 ring. This interaction could be particularly acute since 

 coho and chinook exhibit a high degree of spatial 

 overlap in their distributions (Pearcy and Fisher In 

 press). This would be conceivable only if prey resources 

 were limiting. 



Because of the highly opportunistic feeding mode of 

 most salmonids and the substantial heterogeneity in 

 the physical and biological environment, a large-scale 

 study over extended time-periods may be necessary to 

 adequately describe the feeding dynamics of juvenile 

 salmon. Major departures from the long-term mean 

 oceanographic conditions, as exemplified by an El Nino 

 event, can strongly affect the feeding ecology of many 

 pelagic planktivores. Fulton and LeBrasseur (1985) 

 have hypothesized that a northward shifting of the 

 Subarctic Boundary from its normal position intersec- 

 ting the coast off Oregon or northern California to well 

 above Vancouver Island, British Columbia, as occurs 

 in El Niiio conditions, may expose salmon and other 

 pelagic predators to a novel suite of available prey, with 

 a corresponding downward shift in prey size. Grover 

 and 011a (1987) found that a smaller mean size of 

 copepod was consumed by larval sah\eT\shAnoplopo)fin 

 fimbria, during the El Nino year of 1983 than during 

 1980, a year of relatively normal oceanographic 

 conditions. 



Our study also demonstrated anomalies in species 

 composition in the diet during 1983 and early 1984 

 compared with other years; relatively large northern 

 euphausiids were replaced by much smaller decapod 

 larvae (i.e., Graspidae, Porcellanidae, and Pinnotheri- 

 dae) and euphausiids (Nycti phones simplex) of southern 

 origin (Brodeur 1986). Similar interannual shifts were 

 seen in the fish prey-size spectrum consumed by coho 

 and chinook salmon juveniles with generally smaller 

 prey consumed during 1983 and 1984 (Brodeur In 

 press). Consumption of smaller prey must be balanced 

 by consumption of a greater number of prey of equi- 

 valent caloric content in order to maintain the similar 

 growth rates seen for coho salmon during the early 

 summers of 1983 and 1984 as non-Nino years (Fisher 

 and Pearcy 1988). Unless prey are more aggregated 

 during El Nino years, smaller prey would require a 

 substantial increase in time and energy spent forag- 

 ing, relative to time spent avoiding predators. This in- 

 creased foraging time, at the expense of predator 

 avoidance, may have led to the low coho and chinook 

 salmon survival in the ocean during the El Nifio years 

 (Johnson 1988). 



