HOPKINS and BAIRD: NET FEEDING IN MESOPELAGIC FISHES 



a horizontal tow in which a discrete depth was 

 maintained throughout. 



DISCUSSION 



While studies of the behavior of mesopelagic 

 fishes in small mid-water trawls used for research 

 are nonexistent, there is considerable information 

 available on fish behavior in larger commercial 

 trawls of many kinds (e.g., Ben-Tuvia and Dickson 

 1968). Generally, fish move in front of or away 

 from the walls of the trawl until they are 

 exhausted and are collected in the cod end. 

 Mesopelagic fishes from trawl cod ends often show 

 signs of abrasion (Harrisson 1967); consequently, 

 the likelihood of active and extensive net feeding 

 would appear low. Several authors (e.g., Collard 

 1970; Hopkins and Baird 1973) have suggested that 

 trauma induced by stress conditions in the trawl 

 environment would operate against active feeding 

 behavior. Reflexive gulping or "pseudo" feeding 

 behavior, however, resulting in the ingestion of 

 significant amounts of prey from the plankton rich 

 cod end is a potential mechanism whereby stomach 

 contents could be biased by net feeding. Several 

 studies have revealed diel periodicity in feeding in 

 mid-water fishes which indicates that at certain 

 times, at least, net feeding cannot be extensive 

 (Holton 1969; DeWitt and Cailliet 1972; Baird et al. 

 1975). 



The possibility of fishes foraging in front of the 

 cod end or fish-catcher can also be evaluated. At 

 standard trawling speeds (3.7-4.6 km/h) the trawl 

 moves at a rate of 1.0 to 1.3 m/s. For those 

 epipelagic species which have been examined, 

 foraging and cruising speeds range from about 1 

 to 4 body lengths per second and maximum burst 

 speeds are on the order of 10 to 30 body lengths per 

 second (e.g., Blaxter 1969; Baird et al. 1975). 

 Assuming similar swimming capabilities for mid- 

 water fishes and a fish size of 3 inches (76 mm), a 

 conservative estimate of foraging speeds should 

 be less than 0.3 m/s and burst rates of 0.8 to 2.3 

 m/s. All of the species examined here were less 

 than 76 mm in length except G. elongatum which 

 may have somewhat limited swimming capabili- 

 ties (Marshall 1971). In view of the swimming 

 speeds required, extensive foraging in front of the 

 cod end appears remote. In addition, the data from 

 L. yuentheri (tow 98), where prey of individuals 

 gilled in the net were compared with those from 

 the cod end, failed to reveal indications of net 

 feeding. 



The present results support the contention that 

 if net feeding does occur, it is not extensive in the 

 relatively small fragile fishes typical of the oceanic 

 mesopelagic enrivonment. Only the data on prey 

 abundance in stomachs of L. alatus could be con- 

 strued as statistical evidence of net feeding. In 

 both of these collections, however, mean size of 

 prey items and taxonomic composition of diet 

 were very similar in both sets of fish, while the diet 

 showed little agreement in terms of principal 

 taxonomic components (Table 2) with plankton in 

 the cod end, as might be expected from net feed- 

 ing. In three collections (tows 144, 145, 167), G. 

 elongatum (2 sets) and C. ivarmingi (1 set) from 

 the "control" side contained more food items than 

 fish from the cod end. Here again comparisons of 

 mean prey size, taxonomic composition of diet, and 

 major taxa in diet with that in cod end plankton 

 samples failed to reveal evidence of net feeding. 

 Furthermore, there were often substantial 

 differences between principal taxonomic com- 

 ponents of diet and plankton from cod end catches 

 from the same haul. Mean prey size (with two ex- 

 ceptions) and composition of principal taxa of 

 diets were nearly identical for all sets of com- 

 parison which further indicate the limited nature 

 of net feeding in this study. 



The use of fish scales as a criterion for net feed- 

 ing poses a number of difficult problems. Our 

 hydrocasts reveal, for instance, that fish scales oc- 

 cur naturally in the water column. Further, several 

 studies of both marine and freshwater teleost 

 fishes have shown that scales (probably also the 

 covering mucous and epidermis) can serve as a 

 major component of the diet, appear to be easily 

 digested, and may possibly have considerable nu- 

 tritive value (e.g., Roberts 1970, 1973; Carr and 

 Adams 1972, 1973). Scales were relatively rare in 

 stomachs examined here but did occur in fishes 

 from both sides of the trawl. Since scales are 

 present in the natural environment, may have nu- 

 tritive value, and are possibly easily seen, cap- 

 tured, and eaten, they could serve as a natural food 

 source or provide appropriate stimuli to elicit 

 ingestion. Until more evidence is obtained con- 

 cerning the role of scales in the natural diets of 

 fishes and their abundance in oceanic environ- 

 ments, the presence of scales in the stomachs of 

 mid-water fishes cannot be used with assurance as 

 an indicator of net feeding. 



Because of the difficulty of replicating trawl 

 conditions and obtaining sufficient material for 

 analysis, the variability in distributions of mid- 



913 



