Brodeur and Pearcy Trophic relations of juvenile salmon off Oregon and Washington 



619 



Once aboard, the juvenile salmon were quickly sep- 

 arated from the rest of the catch and anaesthetized in 

 MS-222 to prevent regurgitation during handling. Each 

 fish was tentatively identified, measured to the nearest 

 mm (fork length), individually labeled, and either pre- 

 served whole in formalin after slitting the body cavity 

 (1980 and part of 1981) or frozen whole in a'-20°C 

 freezer. 



In the laboratory ashore, species identifications were 

 verified and individual wet weights were recorded to 

 the nearest 0.1 g. Stomachs were then removed from 

 a random subsample of up to 10 individuals per species 

 for each collection, with the stipulation that the entire 

 size range of each species represented in the sample 

 be included. Stomachs were individually preserved in 

 a 10% formalin solution and then transferred to a 70% 

 ethanol solution prior to examination. Subyearling and 

 yearling juveniles of each species in their first year in 

 the ocean were distinguished from adult fish using scale 

 analysis (Fisher and Pearcy 1988). 



Stomach analysis 



Stomachs were opened and the relative fullness was 

 subjectively assessed on a scale from (empty) to 5 

 (fully distended). The entire stomach contents were 

 blotted on absorbent paper to remove excess moisture 

 and weighed to the nearest 1 mg. The contents were 

 identified under a dissecting microscope to the lowest 

 possible taxonomic level and life-history stage. During 



this analysis, each prey taxon was assigned a digestion 

 code ranging from (well digested) to 4 (fresh), and 

 a digestion level for the entire stomach was derived 

 from these codes based on the relative proportion 

 by weight of each taxon. Each prey taxon was then 

 enumerated, blotted to remove excess moisture, and 

 weighed to the nearest 1 mg. 



Statistical analyses 



Three measures were used to determine the impor- 

 tance of each prey taxon to a particular predator: the 

 percent frequency of occurrence in non-empty stom- 

 achs (F), the percent of total number of prey organisms 

 (N), and the percent of total weight of prey organisms 

 (W). These measures were combined into a single 

 number, the Index of Relative Importance (IRI = F 

 (N-i- W)), modified from that described by Pinkas et al. 

 (1971) using weight instead of volume, so that com- 

 parisons can easily be made between the relative prey 

 composition of different collections or species. Prey 

 items that were digested or taxa difficult to count, such 

 as gelatinous organisms, were not assigned IRI values. 

 The IRI values were then converted to percent of total 

 IRI for each predator species. 



Niche breadth, expressed as the scope of utilization 

 of food resources by each predator species, was 

 calculated using the Shannon-Weaver formula: 



H', = 



P|j (log2 P,j) 



j = i 



where P,, = the proportion by weight of a prey item 

 j in predator i (Petraitis 1979). This index is influenced 

 by both the number of species in the stomachs and the 

 evenness with which they are distributed among the 

 stomachs, and attains a maximum value (H'^a^) of 

 log2 (number of prey taxa). The ratio of H' to H'^ax 

 provides a measure of the evenness with which the 

 resources are distributed among the predators (Pielou 

 1977). 



Diet overlap was calculated among all species for the 

 entire data set and between coho and chinook by cruise 

 or collection where at least 10 stomachs of each species 

 were examined. Schoener's Percent Similarity Index 

 (PSI) was used since it was found to have the most 

 favorable properties within the range of normal overlap 

 values in the absence of prey availability data (Linton 

 et al. 1981, Wallace 1981) where: 



PSI = [1.0 - 0.5 (X I P,j - Phjl)] X 100 

 where P, , is the propoiiion by weight of food category 



