678 



Fishery Bulletin 89(4), 1991 



Discussion 



Our study indicates that the diets of cod and skate vary 

 according to season and bottom type. Although there 

 were significant differences (P<0.05) in the mean size 

 of cod between some samples which could create some 

 bias in the analyses, the changes in the diet of cod ap- 

 peared to follow the same general pattern between 

 season and bottom type independent of cod size in each 

 sample (Fig. 2). During spring, cod fed on large soft- 

 shelled male snow crab, while during summer and fall, 

 predation was directed more towards alternate prey 

 species, such as fish and other invertebrates. Seasonal 

 variations in cod diet could reflect the abundance and 

 availability of potential prey species. Such a hypothesis 

 is in accord with Lacroix and Marcotte (1962), Kinnear 

 and Livingstone (1979), Kohler and Fitzgerald (1969), 

 and Corbeil (1954), who argued that seasonal and 

 regional variations in the quantity of prey species eaten 

 by cod are probably due to the distribution of individual 

 prey species, and that prey choice in cod is influenced 

 by prey size, availability, abundance, and ease of cap- 

 ture. Thus, our results may reflect the relative densities 

 and size composition of each crab species present at 

 each sampling site at given times. We found that snow 

 crab were more numerous in cod stomachs from sand 

 and mud bottoms located offshore (0.59 crab/stomach) 

 than on inshore rocky and gravel bottom (0.05 crab/ 

 stomach). In contrast, larger numbers of toad crab (1.35 

 crab/stomach) were found on rocky and gravel bottoms 

 inshore than on mud and sand bottoms offshore (0.50 

 crab/stomach). We also determined that the two Hyas 

 species were present in cod stomachs at similar den- 

 sities (0.37 H. araneus and 0.42 H. coarctatusl cod 

 stomach), regardless of bottom type. 



Cod and skate appeared to have different feeding 

 patterns for crabs. On average, skate consumed five 

 times more snow crab (1.60 crabs/stomach) than toad 

 crab (0.33 crab/stomach). In contrast, cod captured in 

 the same tows as skate consumed equivalent numbers 

 of snow crab (0.30 crab/stomach) and toad crab (0.28 

 crab/stomach). Based on a beam trawl survey in the 

 same area during the same periods, Robichaud et al. 

 (1989) determined that snow crab density (11.4 crabs/ 

 1000 m 2 ) was four times higher than the density of 

 toad crabs (2.8 crabs/1000 m 2 ). If beam trawling is 

 considered to be an unbiased sampling method reflect- 

 ing the actual densities of both crab species in the area, 

 we can argue that cod feed more on toad crab than 

 snow crab. In contrast, skate would appear to be a more 

 opportunistic predator, feeding on snow crab and toad 

 crab in direct proportion to their actual occurrence. 



The stomach contents of cod and skate probably 

 reflect different foraging modes. Cod detect their 



prey by visual and olfactory cues, but are not capable 

 of detecting buried prey (Brawn 1969). Since snow crab 

 can bury themselves in sediment (Brunei 1960, Miller 

 1975), cod may experience difficulty in locating them. 

 In comparison, toad crab appear to depend on cam- 

 ouflage rather than burial for protection (D.A. Robi- 

 chaud, pers. obs.) and are usually found at the sediment 

 surface. As skate prefer foraging in sediment (McEach- 

 ran et al. 1976), it should be possible for them to detect 

 snow crab buried in mud and sand. In addition, they 

 would also, presumably, be able to locate toad crab on 

 the surface. However, skate may not be as good at 

 catching crabs on the surface. 



When attempting to elucidate the impact cod preda- 

 tion could have on snow crab recruitment, many eco- 

 logical factors need to be considered: distribution pat- 

 terns of the predator and prey in time and space, size 

 of the predator and prey, availability and vulnerabil- 

 ity of the prey, presence of alternative prey species; 

 and behavior of predator and prey. However, one other 

 important non-ecological factor which may influence 

 one's assessment of the relative importance of a prey 

 species in a predator's diet is the method of stomach 

 content analysis employed by the investigator. Differ- 

 ent methods of analysis can lead to markedly different 

 conclusions (Hyslop 1980). For example, in this study, 

 if we consider only weight as a measure of the impor- 

 tance of snow crab in cod diet, the greatest quantity 

 of crab present in cod stomachs was during the crab 

 molting period in spring. However, if we consider the 

 actual number of snow crab per stomach or their fre- 

 quency of occurrence, the number of crab eaten dur- 

 ing summer (0.90 crab/stomach) was almost three times 

 higher than spring (0.33 crab/stomach) and the fre- 

 quency of occurrence was almost twice the spring level 

 (44% vs. 27% in spring). During spring, cod ate fewer 

 but larger soft-shelled male snow crab. During sum- 

 mer and fall, even though the percentage by weight 

 of snow crab was much lower, cod consumed more 

 numerous small juvenile crabs. 



Although there are constraints to the present anal- 

 ysis (for example, sampling method, fish size, and the 

 small number of samples), the results provide insight 

 into the predatonprey relationships between fish and 

 crab. However, given the constraints and based on the 

 statistical advice provided, it was decided that any ad- 

 ditional effort to statistically isolate the influence of 

 single independent variables such as gear type, season, 

 predator size, depth, or substrate would not only be 

 inappropriate, but liable to create misleading conclu- 

 sions. For example, studies by Powles (1968), Brethes 

 et al. (1987), Coulombe et al. (1985), and Robichaud et 

 al. (1989) have shown that substrate is a more impor- 

 tant influence than depth on snow crab distribution. 

 We stress that depth, although it may be correlated 



