390 



Fishery Bulletin 88(2). 1990 



Figure 1 



Tuna and billfish grounds off Virginia. 



To help evaluate the relationship of the various food 

 items found in stomachs we employed an index of 

 relative importance (IRI) (Pinkas 1971): 



IRI = {N + V) F, 



where N = numerical percentage, V = volumetric per- 

 centage, and F = frequency of occurrence percentage. 

 Percent IRI consisted of the IRI value of each prey 

 category (unidentified fish and cephalopods excluded) 

 divided by the sum of the IRI values (unidentified fish 

 and cephalopods excluded). To determine if the quan- 

 tity of the key trophic group differed by area of capture 

 (Fig. 1), displacement volume was compared against 

 three areas sampled with a one-way ANOVA model 



(with displacement volume as the dependent variable 

 and area of capture as the independent factor). The key 

 trophic group was composed of pooled volumetric con- 

 tributions of both identified and unidentified teleost re- 

 mains. The three areas were: (1) the "Hot Dog," (2) 

 '"Fish hook" and "S.E. Lumps," and (3) "21 Mile Hill." 

 The "Fish hook" and "S.E. Lumps" areas were pooled 

 because boat captains generally fished both areas dur- 

 ing the same trip. The remaining areas, "26 Mile Hill" 

 and "V-Buoy," were both eliminated from the hypoth- 

 esis test because of low sample sizes (N = 2 for both). 

 Significant differences were contrasted by a Student- 

 Neuman-Kuels (SNK) multiple range test set at an 

 experiment-wise error rate (EWER) of 0.05 (Zar 1984). 

 Tests for normality and equality of variance (Zar 1984) 



