Reilly et al : Diets of pelagic juvenile Sebastes off central California 



51 1 



a single prey type each year, followed by several prey 

 types with indices at much lower values (Fig. 2). There 

 were few instances in which the two most important 

 prey types were similar in ranking index, e.g., yellow- 

 tail rockfish consuming euphausiid eggs (21.7) and 

 juvenile copepods (20.7) in 1985. This result suggests 

 that each year these species, to a large extent, spe- 

 cialize on foods that are intermittently abundant. Also, 

 the ranking index data, together with information on 

 frequency of occurrence and percent by number, in- 

 dicate the major prey items of pelagic juvenile rockfish 

 were various life stages of copepods and subadult 

 euphausiids. 



Dietary overlap 



The extent of interspecific dietary similarity was quan- 

 tified by comparing dietary overlaps among ali possible 

 pairs of species within each year (1984-87). Ten species 

 pairs were possible, but not all pairs were observed 

 each year since all five species were not always col- 

 lected (Table 6). Overlap indices are sensitive to the tax- 

 onomic level to which prey items are categorized; thus, 

 statistical tests of significance concerning the data are 

 arbitrary. Therefore, the convention established by 

 Langton (1982) and Brodeur and Pearcy (1984) was in- 

 voked. Overlap index values of 0.00-0.29 were con- 

 sidered low, values of 0.30-0.60 were considered 

 medium, and values >0.60 were considered high. 



Using these criteria, annual comparisons of the dis- 

 tribution of overlap indices for 1984 indicate that 60% 

 of all comparisons were classified as medium and 40% 

 were classified as high. Results from 1985 and 1986 

 indicate that 67% of the scores were medium and 33% 

 were high. In contrast, index values during 1987 

 generally had the lowest amount of overlap: 10% low, 

 70% medium, and 20% high. Based on these findings 

 we conclude that, although overall patterns of dietary 

 overlap do vary from one year to the next, variations 

 are relatively modest (only in 1987 was any low overlap 

 observed). Moreover, in this study most within-year 

 species pairings showed >30% overlap. The principal 

 exception to this generalization was for yellowtail and 

 shortbelly rockfish sampled in 1987. Their diets were 

 quite dissimilar. 



Overlap indices were also calculated for all possible 

 interannual intraspecific combinations. These calcula- 

 tions allow an assessment of the temporal stability of 

 the diet relative to the amount of interspecific dietary 

 overlap displayed during a given year. The frequency 

 distribution of dietary overlap values derived from self- 

 pairing of rockfish species from different years is 

 shifted well to the left (toward zero) of the distribution 

 of interspecific scores obtained within a year (Fig. 3). 

 These findings show that in any particular year the dif- 



20 



15- 



o 



c 



0) 



NWN Interspecific (within year) 

 ^H Interannual (within species) 



10- I 



- I al ll 



11 



0.00 0.20 0.40 0.60 0.80 



Dietary Overlap 



1.00 



Figure 3 



Frequency of dietary overlap indices among all interspecific 

 Sebastes pairs within years, compared with frequency of 

 overlap indices calculated for each Sebastes species self-paired 

 across years. 



ferent species of rockfish are opportunistic feeders that 

 utilize relatively similar prey items, but substantial 

 dietary change can occur from year to year. 



Latitude and depth effects 



Dietary variation with respect to station latitude (north 

 or south of lat. 37°20'N), station depth (deeper or 

 shallower than 100 m), and the interaction of these 

 variables were analyzed using MANOVA (Table 7). The 

 statistical significance of each analysis depended on the 

 particular combination of species and year examined. 

 In 1987, highly significant (P< 0.001) diet variations oc- 

 curred with depth for shortbelly and widow rockfish. 



