of decapod larvae in the diet of many fishes (R. erinacea, U. chuss. 

 M. bilinearis, M. americanus) in the summer may also reflect this 

 phenomenon. 



Predator Size and Feeding Strategy 



The food habits of dominant shelf fishes changed considerably 

 with size, as noted in other fishes (Tyler 1972; Ross 1978; Werner 

 1979). For most predators this change was a switch to different, 

 larger, prey taxa. Many predators (R. erinacea. U. chuss, U. regia, 

 and M. bilinearis) fed on increasing numbers of similarly sized, 

 small food items, up to a certain length. At this point, there was a 

 rapid increase in mean prey size for larger predators, with a con- 

 comitant decrease in the number of prey consumed. Total volume 

 of food increased with increasing fish length. Ross ( 1978) noted a 

 similar progression in food habits with increasing size and sug- 

 gested this strategy should maximize energy intake at the onset of 

 reproduction, a time of increased energy demand. 



Schoener (1971) predicted, from optimal foraging models, that 

 food size should decrease with decreasing predator size, and should 

 do so asymptotically. Numerous examples demonstrate the trend of 

 his prediction, but evidence for an asymptote has been sparse, and 

 one study indicated it did not occur (Schoener 1971). Most shelf 

 fishes studied (Figs. 4. 7. 10. 13. 16. 25, 28) demonstrate this phe- 

 nomenon, but there are exceptions. Macrozaarces americanus 

 (Fig. 19) fed heavily on small food items throughout the size-range 

 examined, with larger fish retaining small prey in the diet, while 

 broadening their feeding to include larger prey items. Slenotomus 

 chrysops underwent a marked change in prey-size preference 

 between 100 and 150 mm standard length (Fig. 22). but then prey 

 size remained relatively constant. There is no asymptote at the 

 low er end of the length range for 5. chrysops. 



Larger predators should take a greater size range of food, and 

 food diversity (i.e.. number of prey types or species) should be 

 greater in large animals, unless available small prey are sufficiently 

 more diverse (Schoener 1971 ). In those predators for which benthic 

 prey dominated (all except L. americanus and M. bilinearis). such 

 a relationship is evident for prey types. Although amphipods, the 

 dominant prey for smaller predators, decrease in abundance, they 

 remain relatively common in the diet of large fish even as other 

 larger prey items are added. However, large prey items include the 

 much less diverse decapods and larger polychaetes. The high diver- 

 sity of available small prey (amphipods. isopods. cumaceans) result 

 in smaller fishes having a more diverse diet at the species level. 



Overlap in Diet 



Most predator species were selective on the macrobenthos, par- 

 ticularly on corophiid and ampeliscid amphipods and decapods. 

 These crustaceans were important food for these predators, result- 

 ing in considerable overlap in diet. 



Cluster analysis of predator species and size-classes based on 

 prey similarity indicates that intraspecific and interspecific diet- 

 overlap relationships change considerably with season and with 

 Fish size. Although there was considerable interspecific overlap in 

 diet, there is evidence for intraspecific food-resource partitioning. 

 Small fishes overlapped in diet intraspecifically as well as with 

 small fishes of other species. The larger fishes of these species also 

 exhibited interspecific dietary overlap, but fed quite differently 

 from the juveniles. These differences in diet overlap with size were 

 correlated with changes in feeding strategy with increased fish 

 length. For example, in fall (Fig. 31), all R. erinacea ranging 



between 51 and 250 mm DW fed similarly and were grouped 

 together within a larger group of similar feeders. However, skates 

 larger than 250 mm DW fed differently and were grouped together 

 with other large decapod feeders. Raja erinacea demonstrated a 

 marked change in food habits at 250 mm DW (Fig. 4) where this 

 shift in food-overlap relationships occurred. The other species that 

 grouped with R. erinacea also demonstrated a parallel change in 

 feeding strategy with increased size. Thus, although intraspecific 

 changes in diet with increased size may prevent intraspecific over- 

 lap in diet, considerable interspecific overlap occurred. Although 

 intraspecific and interspecific diet-overlap relationships changed 

 seasonally, it is apparent that intraspecific differences in feeding are 

 as important as interspecific differences in structuring the predator 

 community. 



Several reasons may account for the considerable amount of 

 interspecific overlap in diet exhibited by shelf fishes. Optimal for- 

 aging theory predicts that as food becomes scarce, predators will 

 take a wide variety of food and similar predators occupying the 

 same habitat will converge in diet (Pyke et al. 1977). Alternatively, 

 some authors have hypothesized that as food density lowers, coex- 

 isting predators will specialize on different prey and food overlap 

 will decrease. Considerable food overlap would only be expected if 

 food were abundant (Jones 1978). Some field studies support this 

 latter hypothesis, although this may be due to a lack of measure- 

 ment of actual resource availability. Thus. Keast (1965) and Zaret 

 and Rand ( 1971) found that fishes specialized in diet and that inter- 

 specific overlap was at a minimum during the food-impoverished 

 season. Maximum food overlap occurred when food levels were 

 high [see also Ross (1977) and Townsend and Hildrew (1979)]. 

 Tyler (1972) reported little overlap in the diets of northern marine 

 demersal fishes and concluded that food limitation led to speciali- 

 zation and food-resource partitioning. The present results indicate 

 that shelf fishes are selective in their feeding, but that considerable 

 interspecific overlap occurs in diet. The question remains whether 

 this overlap is due to a food shortage (Pianka 1976; Pyke et al. 

 1977) or a food abundance (Zaret and Rand 1971; Ross 1977: Jones 

 1978). Boesch et al. (1977) and Boesch ( 1978) reported that density 

 and abundance of macrobenthos on the outer shelf were generally 

 high and persistent year-round. Walsh et al. (1978) reported an 

 increase in plankton productivity in the early spring and suggested 

 most of this productivity was transferred to the bottom. This could - 

 lead to a superabundance of food near and on the bottom in the 

 spring. It is noteworthy that food overlap among shelf fishes was 

 lowest in spring, and that some of this was due to normally benthic 

 predators (e.g., U. chuss and C. arctifrons) switching to planktonic 

 prey. It appears that minimal overlap in diet of shelf fishes in the 

 present study is associated with a superabundance of prey in the 

 spring, supporting the hypothesis of optimal foraging (Pianka 

 1976; Pyke etal. 1977). 



The question also remains to be answered as to whether there is 

 competition for food among shelf fishes. Although there was much 

 interspecific diet overlap among shelf fishes, overlap need not nec- 

 essarily lead to competition unless resources are in short supply 

 (Pianka 1976). Extensive niche overlap may actually be correlated 

 with reduced competition (Pianka 1974, 1976; Jones 1978). Most 

 shelf fishes exhibited extensive overlap in habitat and food, but it is 

 not known if these resources are in short supply. Predator exclusion 

 experiments on the outer shelf indicate that the macrobenthic com- 

 munity is, in part, predator controlled (Boesch 1978) and that popu- 

 lations of certain species, including those important as prey to 

 fishes (e.g., corophiid amphipods). may be kept below earning 

 capacity by fish predation. Whether this predation pressure keeps 



