DISCUSSION 



Planktonic and nektonic organisms comprising the 

 diets of the tunas and lancetfishes ranged widely in size 

 and form, suggesting that these predators feed oppor- 

 tunistically. Other data on the food of tunas show, world- 

 wide, a large degree of consistency in forage composition, 

 probably reflecting their association with a distinctive 

 global epipelagic community (Haedrich and Nielsen 

 1966; Parin et al. 1969; Brodulina 1974). Fishes were con- 

 sistently the largest and most frequently occurring tax- 

 onomic group and, where volumes were measured, con- 

 stituted the largest forage volume. Cephalopods generally 

 ranked second to fishes in all three measurements. Other 

 invertebrates were generally much smaller but were often 

 eaten in large numbers, particularly amphipods and 

 larval crustaceans, probably encountered in swarms. 



Despite such overall consistency, definite differences 

 in the relative importance of certain forage categories do 

 indicate some degree of feeding selectivity, either 

 anatomical or behavioral. Some interspecific and in- 

 traspecific feeding differences can be traced to the 

 anatomy of the predator. The size of the gill raker gap 

 determines the minimum size of prey that can be cap- 

 tured and retained (Magnuson and Heitz 1971) and the 

 maximum prey size is determined by the greatest disten- 

 sibility of the predator's mouth and esophagus. Indeed, 

 the small-mouthed tunas consumed generally smaller 

 prey (mean forage fish length, 98 mm SL) than did the 

 large-mouthed lancetfishes (mean forage fish length. 240 

 mm SL) . 



A similar trend in forage utilization can be seen within 

 a single tuna species. Thunnus albacares were divided 

 into three size groups (60.0-99.9 cm FL, 100.0-119.9 cm 

 FL, and 120.0-169.9 cm FL) to facilitate forage com- 

 parison. The largest fishes eaten (Gempylidae, x length 

 216 mm SL; Coryphaenidae, x length 141 mm SL; 

 Alepisauridae. x length 120 mm SL; Paralepididae, x 

 length 105 mm SL; Scombridae, x length 96 mm SL) 

 were consumed most frequently by the largest sized 

 tunas; among the prey families only the scombrids 

 appeared in the smallest sized tunas. Conversely, the 

 frequency of occurrence of the smallest forage organisms 

 generally decreased with increasing tuna size, as shown 

 by the overall mean frequency of occurrence for larval 

 decapods in Appendix Table 6 (.t 12.2°? frequency of oc- 

 currence in 60.0-99.9 cm FL tunas; x 10.39 in 100.0-119.9 

 cm FL tunas; x 3.339 in 120.0-169.9 cm FL tunas). 

 Similar trends were not apparent in the other tuna 

 species (Appendix Tables 7-8). 



A second limitation of forage composition is deter- 

 mined by the fish's predatory ability, particularly swim- 

 ming speed. While the tunas are powerful, efficient 

 swimmers, Alepisaurus is comparatively slow: this 

 difference is reflected in the type of organism most fre- 

 quently consumed. Swift-moving muscular squids (such 

 as Gonatus and Onychoteuthis; Roper and Young 1975) 

 were eaten much more frequently by tunas than by 

 Alepisaurus; both capture squids more efficiently. 



however, than does the 10-foot Isaacs-Kidd midwater 

 trawl. The predatory ability of tunas is further reflected 

 in the frequency with which they consume such fast- 

 swimming fishes as Belonidae, Scomberesocidae, Ex- 

 ocoetidae, Carangidae, and Coryphaenidae. The absence 

 of these families from the Alepisaurus stomachs, 

 however, may be due to the fact that they are found only 

 near the surface, where Alepisaurus probably does not 

 feed, rather than because Alepisaurus is unable to catch 

 them. 



Differences in depths at which predators feed probably 

 constitute another mode of forage selectivity. In order to 

 compare differences in feeding depth, we divided forage 

 components somewhat arbitrarily into three categories 

 according to their position in the water column: 

 Sargassum -associates (strictly surface organisms 

 associated with pelagic Sargassum in the upper 5 m of 

 water); near-surface (the upper 20 m); and midwater 

 (both vertical migrators and those which remain at 

 depths). 



Thunnus albacares evidently fed mainly at the surface, 

 particularly on Sargassum associates. Bits of Sargassum 

 present in stomachs were probably accidentally ingested 

 in the process of capturing associated fauna, of which 

 Portunus sayi, the sargassum crab, and large numbers of 

 epipelagic decapod larvae were most frequently con- 

 sumed (see Fig. 4). The most frequently occurring fishes 

 in T. albacares stomachs were juveniles of reef and shore 

 fishes, also associates of the Sargassum community. 



The Sargassum community plays a similar role in the 

 nutrition of other peak oceanic predators such as the 

 dolphins Coryphaena hippurus and C. equiselis (Gibbs 

 and Collette 1959). This utilization is reasonable to ex- 

 pect, as the Sargassum community provides the major 

 concentration of forage organisms in the otherwise barren 

 oceanic surface waters (see Fine 1970; Dooley 1972). 



In contrast with the surface-oriented T. albacares, the 

 diets of T. alalunga, T. thynnus, and T. obesus, com- 

 posed mainly of midwater organisms, suggest that these 

 tunas feed at somewhat greater depths. Forage com- 

 positions observed may be incomplete, however, because 

 the samples examined were comparatively small. 



The relatively small representation of epipelagic 

 organisms and high frequency of occurrence of midwater 

 fishes and invertebrates in the diet of Alepisaurus in- 

 dicates that this predator feeds in the midwater layers 

 (Figs. 3, 4). It might be noted in particular that 

 Alepisaurus consumed the deeper-occurring gelatinous 

 squids (Histioteuthidae, Bathyteuthidae, and 

 Cranchiidae) and the gelatinous octopod, Alloposus 

 mollis (family Alloposidae), much more frequently than 

 did any of the tunas (except for T. thynnus feeding on 

 Alloposus mollis). These cephalopods are probably not as 

 abundant in the shallower levels, where tunas feed. 



Relative feeding depths discussed above generally cor- 

 respond to hooking depth records for tunas (Shomura 

 and Murphy 1955; Yabe et al. 1963; Osipov 1968; Legand 

 and Grandperrin 1973; Saito 1975) and lancetfishes. 

 Thunnus albacares is generally the most shallow- 

 occurring species; T. alalunga and T thynnus occur 



