food item. Finally. Group IV consisted of small U. chuss and S. 

 chrysops and small-to-medium M. bilinearis, all of which fed 

 heavily on hyperiid amphipods while consuming copepods and 

 caridean decapods as well. 



Small R. erinacea overlapped incompletely with other amphipod 

 consumers, especially medium U. chuss. small 11. regia, and large 

 P. oblongus (Group I, Fig. 31). Medium-to-large R. erinacea also 

 showed a very high similarity in diet to these three species (Group 

 III. Fig. 31). but their prey (mainly C. irroratus) were quite differ- 

 ent at this size. Thus, interspecific diet overlap, rather than intra- 

 specific. was greater for these species. Small 5. chrysops were very 

 similar in diet to small U. chuss and M. bilinearis (Group IV), but 

 were different from larger 5. chrysops, which shared food items 

 with medium U. chuss and C. araifrons. Other examples of higher 

 interspecific vs. intraspecific diet overlap were also evident (Fig. 

 31). 



In winter, three major groups of similar feeders were classified 

 (Fig. 32). Group I was composed of those fishes (Subgroup I-A) 

 such as L. americanus and large M. bilinearis which fed largely on 

 fishes (primarily 11. chuss) and a few carideans {D. leptocerus): 

 and those fishes (Subgroup I-B) which fed on hyperiids and caride- 

 ans as well as fishes. As in fall, small M. bilinearis and U. chuss 

 had similar diets and constituted a group (Group II) with high simi- 

 larity. Group III (Fig. 32) was comprised of five species in several 

 size classes: all fed primarily on corophiid amphipods (U. irrorata 

 and E. rabricornis). Two subgroups were present: Subgroup I1I-A 

 was comprised of small P. oblongus. R. erinacea. C. arctifrons, 

 and large M. americanus. all of which had E. rubricornis as the 

 most abundant food item. Subgroup III-B consisted of larger P. 

 oblongus, U. chuss. C. arctifrons. R. erinacea. and small-to- 

 medium M. americanus. With the exception of the largest H. 

 oblonga. all size-classes of these species fed primarily on U. 

 irrorata (the second most important food for large H. oblonga). 

 Erichthonius rubricornis was the second most abundant food for 

 most of these entities. Further subgroups (1 and 2) were distin- 

 guished by a secondary preference for other amphipods or alterna- 

 tively for decapods. A multispecies group of large fishes which fed 

 mainly on C. irroratus was absent in winter. 



In winter, as in fall, different size classes within a species of preda- 

 tor were classified in different groups. Small R. erinacea fed mainly 

 on E. rubricornis and overlapped with M. americanus. small P. 

 oblongus, and small C. arctifrons. Medium R. erinacea fed mainly 

 on U. irrorata. E. rubricornis. D. sculpta. and B. serrata, overlap- 

 ping with large U. chuss. M. americanus. and again with larger P. 

 oblongus. The largest class of skates also fed heavily on U. irrorata. 

 Again, as in the fall, small U. chuss overlapped in diet most closely 

 with small M. bilinearis. However, larger U. chuss fed more on gam- 

 marideans and decapods and were classified with similar feeders 

 (Subgroup III-B). Large and medium M. bilinearis fed on fishes and 

 were grouped with other piscivores, but small individuals fed on 

 items similar to those taken by small U. chuss. All M. americanus 

 fed primarily on gammarideans and were included in Group III. 

 Small C. araifrons and H. oblonga were similar in diet (Subgroup 

 III-A). and larger individuals of these species were grouped with 

 other gammaridean feeders (Subgroup III-B). 



In spring (Fig. 33), the classification of predator entities resulted 

 in several small groups, each characterized by high intragroup diet 

 similarity. These groups were in turn joined together at lower levels 

 of similarity. An additional large multispecies group consisted of 

 several loosely joined entities (Group I). This group consisted 

 mainly of piscivores such as L. americanus, large U. chuss. and 

 small U. regia. These last two species also ate C. irroratus, in addi- 



tion to fishes, and were joined with other decapod consumers (M. 

 americanus, large U. regia, and large H. oblonga). The largest 

 sizes of M. bilinearis consumed cephalopods (/. illecebrosus) and 

 fishes (C. arctifrons) and were included in this rather dissimilar 

 group. 



The smaller, more similar groupings present in spring (e.g., 

 Group II) were monospecific in many cases, indicating more spe- 

 cialization in the diet within each species, and less interspecific 

 overlap in food in the spring. Thus, small and medium M. bili- 

 nearis (Group II) consumed predominantly S. elegans and were 

 grouped together. Larvaceans were the most abundant prey for all 

 sizes of C. arctifrons. and all sizes of this predator clustered 

 together. Most size-classes of M. americanus clustered with small 

 R. erinacea. Both species fed mainly on E. rubricornis and U. 

 irrorata. All sizes of U. chuss (except the two largest individuals) 

 were included in a single group of high similarity in spring. In fall 

 and winter, small U. chuss fed quite differently from large ones and 

 were classified separately with smaller individuals of other species, 

 such as C. arctifrons, S. chrysops. and M. bilinearis. In spring, 

 however, all 11. chuss except the two largest individuals (451 and 

 500 mm SL) formed a distinct group. This group of U. chuss was 

 joined with another group consisting of larger P. oblongus and R. 

 erinacea. for which U. irrorata was the most abundant prey but 

 which also fed heavily on decapods. 



In summer (Fig. 34) interspecific overlap in diet again increased. 

 A rather large group (Group I, Fig. 34) included the many species 

 which consumed fishes during the summer and those species that 

 led primarily on planktonic invertebrates. Small U. chuss and M. 

 bilinearis fed similarly, as in fall and winter, and were grouped 

 together (Subgroup I-A. Fig. 34). They had consumed primarily 5. 

 elegans. P. gaudichaudi . and some gammarideans. A single small 

 goosefish which had consumed chaetognaths (S. elegans) was 

 included. The remainder of Group I consisted of those entities 

 which had eaten fishes. Group II consisted of fishes for which C. 

 irroratus was the most abundant food, followed by amphipods, 

 other decapods, and fishes. This group consisted of large predators 

 of decapods, such as M. americanus. R. erinacea. P. oblongus. 

 and U. regia. 



Group III consisted of amphipod eaters. Erichthonius rubri- 

 cornis and U. irrorata were the two most abundant prey for all 

 fishes in Subgroup III-A. and these two amphipods were also abun- 

 dant in the diets of other Group III fishes. Other amphipods were 

 also taken by Group III fishes. 



In summer, as in most other seasons, different sizes of most pred- 

 ator species were included in different feeding groups. Thus, small 

 R. erinacea clustered with other amphipod feeders, and large skates 

 were included with larger individuals of other species which fed on 

 decapods (primarily brachyurans such as C. irroratus). Large U. 

 regia and P. oblongus fed on brachyuran decapods, whereas 

 smaller individuals of both species fed more on fishes and caridean 

 decapods. As in all other seasons except spring, small M. bilinearis 

 and 11. chuss were grouped together. However, larger M. bilinearis 

 (151-400 mm SL) were included in a single assemblage which fed 

 more on carideans and fishes. Urophycis chuss was associated with 

 three separate groups. The smallest (1-100 mm SL) red hake fed on 

 S. elegans and P. gaudichaudi and were associated with Group I- 

 A. Intermediate-sized fish (101-300 mm SL) fed on gammarideans 

 (U. irrorata and E. rubricornis). decapods, and 5. elegans. and 

 belonged to Group III. Large red hake (>300 mm SL) fed primar- 

 ily on fishes (C. arctifrons) and C. irroratus. Macrozoarces ameri- 

 canus was associated with two groups: Small ocean pout fed mainly 

 on amphipods (E. rubricornis, U. irrorata, and A. vadorum), and 



