On large catches of some dominant species, random subsamples (at 

 least 30 specimens) were dissected. Each stomach was labeled, 

 individually wrapped in cheesecloth, and fixed in 10% seawater 

 Formalin. 



After proper fixation, stomachs were soaked in water and trans- 

 ferred to either 40% isopropanol or 70% ethanol. For analysis, 

 each stomach was cut open and its contents sorted by taxon and 

 counted. Fragments such as crustacean parts, polychaete setae, or 

 fish bones were counted as one animal, unless abundance could be 

 estimated by counting pairs of eyes (crustaceans), otoliths (fishes), 

 or other parts. 



Volume displacement of food items was measured by using either 

 a graduated cylinder (Windell 1971) or a calibrated vial and buret 

 (McEachran et al. 1976). Displacement of small species was esti- 

 mated by using aO. 1 cm 2 grid (Windell 1971). 



Since methods of food habits analysis are variously biased 

 (Hynes 1950; Pinkaset al. 1971: Windell 1971). the relative contri- 

 bution of different food items to the total diet was determined using 

 three methods: 1) The number of stomachs in which a food item 

 occurred was expressed as a percentage of the total number of stom- 

 achs of a series containing food (percent frequency of occurrence): 

 2) the number of individuals of each type of food was expressed as a 

 percentage of the total number of food items from all stomachs lor a 

 series (percent numerical abundance): 3) the volume displacement 

 of food items was expressed as a percentage of the total volume of 

 food from all stomachs examined of a series (percent volume dis- 

 placement). 



From these three measurements an index of relative importance. 

 IR1 (Pinkas et al. 1971). was calculated for each pres species and 

 each higher taxon as follows: 



IR1 = {N+V)F 



where: IR1 = index of relative importance. 

 N= numerical percentage. 

 V= volumetric percentage, and 

 F= frequency of occurrence percentage. 



This index has been useful in evaluating the relative importance of 

 different food items found in fish stomachs (Pinkas et al. 1971: 

 McEachran et al. 1976; Sedberry and Musick 1978). The 1R1 was 

 used in the present study to describe the food habits of each species 

 and to determine seasonal and predator size differences in the rela- 

 tive importance of food items. 



Overlap in diet among dominant predators was measured using 

 cluster analysis. Stomachs of predators were treated as collections 

 and were subjected to normal cluster analysis on the basis of prey 

 similarity, using percent standardized numerical abundance (Clif- 

 ford and Stephenson 1975). because sample sizes were unequal. 

 Flexible sorting (Lance and Williams 1967; Clifford and Stephen- 

 son 1975). with (3 = -0.25, was used, based on resemblance mea- 

 sured by the Bray-Curtis similarity index (Bray and Curtis 1957). 

 expressed as follows: 



RESULTS 



Food Habits Analysis 



S,.= l 



E .Y. + AJ 

 i 



where 5 , is the similarity in diet between the predator speciesy and 

 k;X is the abundance of the /'th prc> species for predator/; and X, is 



the abundance of the ith pre> species for predator k. 



Over 12.500 individual fish were dissected in the field for stom- 

 ach analysis. A total of 6.087 stomachs representing the nine domi- 

 nant species was examined in the laboratory. Initially, only seven 

 species were to be examined: Raja erinacea Mitchill, Lophius 

 americanus Valenciennes, Urophycis chuss (Walbaum), Merluc- 

 cius bilinearis (Mitchill), Stenotomus chrysops (Linnaeus), 

 Citharichthys arctifrons Goode, and Paralichthys oblongus (Mit- 

 chill). Urophycis regia (Walbaum) and Marcrozoarces americanus 

 (Schneider) were added to this list because they were dominant spe- 

 cies in the catches in summer 1977 (Table 1). 



Raja erinacea. — The little skate was abundant in the study area 

 at all times of the year (see Table 1), feeding mainly on amphipods, 

 decapods, cumaceans. and polychaetes (Fig. 2. Table 2). Pelecy- 

 pods, fishes, and isopods were also consumed. The relative impor- 

 tance of these major taxa of food remained fairly constant 

 seasonally, although juvenile fishes were somewhat more impor- 

 tant and cumaceans less important in fall samples. Juvenile fishes 

 were also important in the diet of R. erinacea in the summer. 

 Although the relative importance of the major taxa was nearly con- 

 stant seasonally, the species composition of these taxa in the diet 

 changed somewhat seasonally (Table 2). 



The food habits of Raja erinacea varied greatly with size (Fig. 

 3). Smaller skates fed more on amphipods and cumaceans. whereas 

 larger skates fed more on decapods and polychaetes. The smallest 

 skates i 1-100 mm DW) fed on numerous small food items (Fig. 4), 

 and fed on these in increasing numbers up to 200 mm DW. At about 

 200 mm DW. R. erinacea showed the most pronounced change in 

 food habits and feeding strategies, switching to fewer, larger food 

 items (primarily decapods). 



Lophius americanus— The goosefish. though not as abundant 

 as many other fishes on the outer shelf, was common and composed 

 a considerable portion of the biomass of fishes in the study areas 

 because of its large size (Table 1). Goosefish fed mainly on fishes 

 and to a lesser extent on benthic invertebrates during all seasons 

 (Fig. 5. Table 3). Decapods and cephalopods were less important 

 as food, and polychaetes. amphipods, asteroids, and chaetognaths 

 were only occasionally found in stomachs. Fishes were the most 

 important food for all sizes of L. americanus (Fig. 6), although 

 larger fish were eaten by larger/., americanus (Fig. 7). 



L'rophycis chuss— Although red hake were abundant during this 

 stud), size composition of the population varied seasonally. Juve- 

 niles predominated in fall (,v SL = 49 mm) and summer (x SL= 147 

 mm), but were rare in winter and spring, when larger fish moved 

 into the area (.v SL = 251 and 238 mm. respectively). This was 

 reflected in the much smaller contribution of this species to the bio- 

 mass of fishes during the fall and summer (Table 1). 



Red hake fed primarily on amphipods. which were important as 

 food at all times of the year, especially in fall. During this season 

 they made up most of the diet (Fig. 8. Table 4). Decapods and poly- 

 chaetes were also dominant prey taxa. and copepods were impor- 

 tant in fall and winter. Chaetognaths, absent from the diet in fall and 

 winter, were commonly consumed during the spring and summer. 

 Much seasonal variation was observed in the diet at the species 

 level within these higher taxa (Table 4). particularly within the 

 Amphipoda. 



