FISHERY BULLETIN: VOL. 81, NO. 2 



were eaten by L. hosei. We take the presence of den- 

 tary, cleithral, and opercular bones, however, as 

 evidence of direct ingestion, since these structures 

 are more subject to dissolution by digestive action 

 than are otoliths. The absence of remains from inter- 

 mediate predators (e.g., scombroids, trichiuroids) 

 suggests that no further trophic-level steps were 

 involved. 



The fishes which are most abundantly represented 

 in the three dolphin stomachs can be separated into 

 three groups based on morphological and ecological 

 similarities. Diretmus, Argyropelecus, and Ichthy- 

 ococcus are all silvery, deep-bodied, large-eyed forms 

 which inhabit upper mesopelagic depths between 

 250 and 450 m in the eastern tropical Pacific 

 (Robison 1973) and commonly occur in loose 

 aggregations. They possess gas-filled swim bladders 

 which undoubtedly make excellent echolocation 

 targets. These fishes do not undertake regular, ex- 

 tensive, diel vertical migrations. They eat copepods, 

 euphausiids, and small fishes which are associated 

 with vertically mobile sound scattering layers 

 (SSL). 



Lampanyctus, Lampadena, and Scopelogadus are 

 dark, thick-bodied fishes with medium-sized eyes 

 and regressed or fat-filled swim bladders as adults. 

 They occupy lower mesopelagic depths between 500 

 and 750 m (Robison 1973), and while smaller in- 

 dividuals may be significant components of SSL's, 

 specimens of the size range ingested by Lageno- 

 delphis hosei are not known to be regular vertical 

 migrators. Their food consists primarily of SSL crus- 

 taceans and fishes. 



Chauliodus, Gonostoma, Scopelarchus, the ever- 

 mannellids, and paralepidids are solitary, slender, 

 fast- swimming predators which prey upon micronek- 

 tonic (ca. 10-60 mm) fishes and crustaceans. These 

 fishes exhibit wide mesopelagic depth ranges be- 

 tween 305 and 1,250 m (Robison 1973), they do not 

 have swim bladders, and they undertake varying 

 degrees of vertical migration which are probably 

 related to the movements of their prey. 



Tobayama et al. (1973) found otoliths from Ich- 

 thyococcus elongatus , Polyipnus asteroides (Sternop- 

 tychidae), and Diaphus elucens? to be most numerous 

 in the stomach of a specimen of L. hosei collected off 

 Japan. These authors concluded that the fishes were 

 eaten at a relatively shallow depth at night. However, 

 it is likely that only Polyipnus could have been taken 

 near the surface. The two other species were prob- 

 ably taken no shallower than 300 m, day or night. 



Furthermore, the inclusion of Coryphaenoididae in 

 their listing of prey means that deep feeding must 

 have occurred, although the use of this name is mis- 



applied. Coryphaenoididae is an obsolete name for 

 the deep- living fishes of the family Macrouridae (it is 

 unlikely that they meant Coryphaenidae, since the 

 latter fishes do not possess otoliths). It is also pos- 

 sible that the otoliths in question are from Melanonus, 

 since the sagittae of these fishes are easily mistaken 

 for those of macrourids. The report of P. asteroides by 

 these authors may also be in error, since Baird (1 97 1) 

 stated that this species is known only from the west- 

 ern North Atlantic. 



As a collector of midwater fishes, L. hosei provides a 

 distinctly different perspective on the composition of 

 the mesopelagic fauna than that obtained by conven- 

 tional sampling methods. Many of the ingested fishes 

 were as large or larger than the maximum size of 

 specimens that have been collected by nets. In addi- 

 tion, fishes such as Diretmus, which are rare in trawl 

 collections, were shown to be surprisingly abun- 

 dant. 



The relative abundances of the fish species found as 

 dolphin food do not reflect the relative abundances of 

 midwater fish species in the eastern tropical Pacific 

 as determined by trawling or larval surveys (Ahl- 

 strom and Counts 1958; Ahlstrom 1971, 1972; 

 Robison 1973). Vinciguerria lucetia, which is one of 

 the most common fishes in trawl hauls, is represented 

 in the stomachs by only six otoliths. Only 12 otoliths 

 were found from Diogenichthys laternatus', which is 

 the most abundant species in larval surveys. 



These small, vertically migratory species have been 

 shown to be important in the diet of the spinner por- 

 poise, Stenella longirostris, in the eastern tropical 

 Pacific (Perrin et al. 1973b). The spotted porpoise, S. 

 attenuata, which cooccurs with the spinner, was 

 shown to feed primarily on epipelagic fish and squid 

 in the same study. While the number of L. hosei 

 stomachs we examined was too small for a valid com- 

 parison of feeding with S. longirostris and S. at- 

 tenuata, the low degree of similarity between prey 

 fish types suggests that each of the three cetacean 

 species has a different feeding strategy. Additional 

 support for this conclusion comes from the evidence 

 that only L. hosei consumes crustaceans (Perrin et 

 al. 1973b). 



Based on our understanding of the fishes whose 

 remains we examined, we conclude that the Fraser's 

 dolphins had been feeding selectively, by depth and 

 by prey size. Location and ingestion took place in 

 near or total darkness, regardless of time. Two depth 

 horizons were hunted, each containing a different 

 variety of mesopelagic fishes. The shallowest level 

 was not less than 250 m, and the deepest was not less 

 than about 500 m. The Fraser's dolphins fed with 

 similar success at each depth. 



288 



