Table 1. -Comparison of diet characteristics of fish examined in net feeding investigation. F = 

 unmodified side of double trawl (see Figure 1); G = gilled in meshes of body of trawl; ^tests:* = 

 **•» = P<0.(X)5. 



FISHERY BULLETIN: VOL. 73. NO. 4 



'fish-catcher" side of double trawl; P = 

 = P<0.05, ** = P<0.025, *** = P<0.01, 



'Time of tow initiation. 



^Fish scales only; no other fish remains present. 



plankters counted in each subsample was 775 

 (range: 453-1,079). 



RESULTS 



Fish Size Distribution 



Because of possible relationships between 

 number of items in stomachs (also other diet 

 characteristics) and size of predator (e.g., Nesis 

 1965; Hopkins and Baird 1973), the mean length 

 and size range of fish from each trawl were com- 

 pared for each pair (Table 1). There were no sig- 



nificant differences in mean length (^-test, P>0.05) 

 in 18 of 19 pairs. For the single exception, Cera- 

 toscopeh(>< ivarniiiigi from tow 161, the mean size 

 of fish from the catcher size was larger (P<0.05; .f 

 (SL): 38 vs. 34 mm). However, since the distribu- 

 tions had considerable overlap, this set was 

 included in the study. 



Prey Abundance in Stomachs 



It wa s necess ary to apply a square root trans- 

 form ( /X-l-0.5 ) to the data on number of prey 

 per stomach since the high frequency of empty 



910 



