BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES 



25- 



20- 



< 

 x 



o 





< 



10- 



5- 



0- 



.._./. 



T 



T 



T 



T 



8 12 16 20 



NUMBER OF POOLED STOMACHS 



24 



28 



FIGURE 2— Cumulative prey curves for the first 28 stomachs of each of the 5 rock- 

 fish species. 



in a food spectrum) when all items are in equal 

 proportion in the diet (Levins 1968). These values 

 were normalized as B n = BIR, which ranges from 

 (most uneven distribution) to 1 (totally even 

 distribution among the prey present). This index 

 assumes equal availabilities of the different prey 

 to all predators. 



Several indices of dietary overlap have been 

 proposed and tested with known distributions of 

 prey organisms (see Cailliet and Barry 1979 4 ; 

 Linton et al. 1981; Wallace 1981). The coefficient of 

 overlap described by Colwell and Futuyma (1974; 

 identical to Schoener's (1970) index but not ex- 

 pressed as a percentage) was chosen as it was 

 found to be realistic for a wide range of true over- 

 laps (Linton et al. 1981). This coefficient is as 

 follows: 



dh = 1.0 -0.5(2, 



Pij - phj I 



where p tJ and phj are the proportions of prey j 

 found in the diets of species i and h respectively. 

 This coefficient has a minimum of (no overlap 



4 Cailliet, G. M., and J. P. Barry. 1979. Comparison of food 

 array overlap measures useful in fish feeding habits analysis. 

 In S. J. Lipovsky and C. A. Simenstad (editors). Fish food habits 

 studies, p. 67-79. Proc. 2d Pac. Northwest Tech. Workshop, 

 Wash. Sea Grant. 



of prey) and a maximum of 1 (all items in equal 

 proportions). 



Analysis of Diet Variations 



The sample sizes of S. pinniger and S. flauidus 

 were sufficient to permit detailed analyses of their 

 food habits, including seasonal, latitudinal, diel, 

 and predator-size variations. 



The 368 specimens of S. pinniger and 264 of S. 

 flavidus were grouped into 10 mm length catego- 

 ries (Fig. 3). The distribution of S. pinniger 

 lengths from the two surveys was similar and no 

 significant differences in the means were found 

 (Student's t-test; P > 0.05). Specimens of S. 

 flavidus collected during the seasonal survey were 

 significantly larger (P < 0.001) than those of 

 the summer survey. Sebastes pinniger averaged 

 about 40 mm larger than S. flavidus for both 

 surveys combined. Corrections were made for this 

 difference where appropriate in the analyses. 



To simplify the analysis of dietary variation 

 in S. pinniger and S. flavidus, eight major types 

 of prey were selected for comparison, based on 

 their gravimetric importance or frequency of oc- 

 currence. Numerical abundances were not used 

 because of the great disparity in prey sizes en- 

 countered and the problem of making counts on 



273 



