620 



Fishery Bulletin 88(4), 1990 



Figure 2 



Hierarchy of data group- 

 ings used in the analysis of 

 dietary variations in juve- 

 nile Pacific salmon. 



j in the diet of species i, and Pi,j is the proportion by 

 weight of food category j in the diet of species h. This 

 index ranges from when two predators have no prey 

 in common to 100 when they have identical diets. 

 Overlap values greater than 60 are generally consid- 

 ered significant. Food availability was considered to be 

 the same for all predators in any one comparison. 



The detailed stomach data were then truncated to 

 major taxonomic categories so that general trends in 

 feeding by month, year, and area could be elucidated. 

 Stomach samples of each salmon species were grouped 

 into smaller subsets in a number of different ways to 

 examine diet variability (Fig. 2). The diets of coho and 

 Chinook salmon were first examined by cruise so that 

 temporal variability between cruises could be assessed. 

 Data were then grouped by month into four time- 

 periods: May (three cruises), June (six cruises), July- 

 August (three cruises), and September (three cruises). 

 Seasonal changes in diets were then analyzed for these 

 time-periods, regardless of the year or area of collec- 

 tion. Similarly, diets were analyzed by year of collec- 

 tion for those years (1981-84) when multiple cruises 

 exist. North-south geographic variations were exam- 

 ined for each of three subareas within the total sam- 

 pling area. The collections were stratified latitudinal- 

 ly (Fig. 1) into three areas: (A) north of 46°40'N, (B) 

 between 46°40'N and 45°20'N, and (C) south of 

 45°20'N (Brodeur et al. 1987a). Food habits were ex- 

 amined for each of these areas for all months and years 

 combined. Finally, diets were examined within each 

 year/month/area subset or by collection so that smaller- 

 scale variations in the food and dietary overlap could 

 be assessed. 



Another potential source of geographic variability 

 was that associated with inshore-offshore variations in 

 prey availability and abundance. This source of varia- 

 tion was examined for several cruises where a large 

 number of collections of coho or chinook salmon were 

 obtained far offshore. For these cruises, collections 



were divided into inner shelf (<18 km from shore), 

 middle shelf (18-37 km), and outer shelf (>37 km) sta- 

 tions. Diets of chum and sockeye salmon were also 

 examined for variation with respect to the different 

 factors when sample sizes were adequate. 



The relative importance of the interannual, seasonal, 

 and geographic variations seen in the diets of coho and 

 chinook juveniles was tested by comparing the presence 

 and absence of a particular major prey category using 

 a variance test of binominally distributed data (Sne- 

 decor and Cochran 1967) for each factor individually 

 irrespective of the others. When a value exceeded the 

 tabulated 0.05 chi-square percentage, the null hypo- 

 thesis that the diets were similar was rejected. 



Results 



Coho salmon 



General food habits Juvenile coho salmon had a 

 relatively diverse diet with many different prey cate- 

 gories represented in the stomachs examined for all 

 years combined (Appendix Table 1). A high percentage 

 (95.2%) of the stomachs contained food, and the overall 

 stomach fullness (i 3.0) and digestion (x 2.3) codes 

 were high indicating that many of the juvenile coho had 

 fed prior to capture. 



The primary food groups consumed by juvenile coho 

 salmon were fishes, decapod larvae, amphipods, eu- 

 phausiids, pteropods, and copepods (Appendix Table 

 1). Larval and juvenile fishes were the most important 

 prey making up 72% of the total weight and 60% of 

 the total IRI. Rockfishes Sebastes spp., northern an- 

 chovies Engraulis mordax, and Pacific sand lance 

 Ammodytes hexapterus were the dominant fish taxa. 

 Although other fish families were represented in the 

 diet, notably the Osmeridae, Cottidae, Hexagrammi- 

 dae, and Pleuronectidae, each of these families made 

 up less than 1% of the total diet by percent IRI. 



