Heteropods and pteropods from this collection were 

 removed and reported on by Russell (1960). The bulk of 

 the material was then transferred to the National Marine 

 Fisheries Service Systematics Laboratory at the U.S. 

 National Museum of Natural History. To facilitate iden- 

 tification, the stomach contents were divided into broad 

 categories: fishes, cephalopods, other mollusks, 

 crustaceans, jellyfish, and miscellaneous. D. M. 

 Damkaer identified the crustaceans and, under the 

 guidance of Clyde Roper (a specialist in cephalopod tax- 

 onomy), the cephalopods and the remainder of the in- 

 vertebrates. Leslie Knapp (Smithsonian Oceanographic 

 Sorting Center (SOSC) ) supervised the identification of 

 the fishes with the assistance of several specialists on the 

 SOSC list. Representative specimens of forage species in 

 good or fair condition have been deposited in the relevant 

 collections of the National Museum of Natural History 

 (worms, mollusks, crustaceans, and fishes). 



The identification of forage specimens depended on 

 the degree to which digestion had damaged the organism 

 and on its stage of- maturity, particularly in the 

 crustaceans and cephalopods. Fishes were more com- 

 pletely identified than were other groups, usually to 

 family and often to species. A large portion of the in- 

 vertebrate forage could be identified only to order or 

 superfamily. Therefore, the number of taxa from any one 

 category of forage organisms, while giving a picture of the 

 diversity of forage, should not be considered quan- 

 titatively. 



The relative importance of a forage category can con- 

 veniently be thought of in terms of the amount of energy 

 it affords its predator, measured by number and size of a 

 particular organism eaten. The frequency with which an 

 organism is eaten, furthermore, often gives a rough es- 

 timate of its general availability to a predator. To 

 evaluate the composition of forage, therefore, we used 

 these three measurements: 1) Frequency was calculated 

 as percent frequency of occurrence, e.g., in what percent- 

 age of the total number of stomachs examined a par- 

 ticular organism was found. 2) The number of in- 

 dividuals per stomach from a particular forage group was 

 compared numerically. 3) The size of a forage organism 

 was expressed as length: crustaceans as mm total length, 

 cephalopods as mm mantle length, and fishes as mm 

 standard length (SL, tip of snout to the base of the 

 caudal fin). Accuracy of measurement was dependent 

 upon the degree to which the forage specimen had been 

 digested; approximations were often made in cases where 

 skeletons were incomplete, but suggestive of the 

 organism's full size. 



We considered a forage organism to be of greatest im- 

 portance when it was eaten with relatively high frequen- 

 cy, was comparatively larger than other forage organisms 

 eaten, and was consumed by an individual predator in 

 larger numbers than the mean number of other forage 

 groups present in that stomach. The less important a 

 forage organism appeared in these three aspects, relative 

 to the other organisms eaten, the less important it was 

 considered to the predator's diet. No statistical tests 

 were applied. 



RESULTS 



Taxonomic lists of forage components of Thunnus and 

 Alepisaurus are presented in Appendix Table 1. Forage 

 organisms which are additions to taxa listed for the 

 Atlantic Ocean by Dragovich (1969) for Thunnus or by 

 Haedrich (1964) for Alepisaurus are noted. Fishes, 

 crustaceans, and mollusks made up the bulk of the 

 collection. Medusae, salps, Sargassum, and parasitic 

 nematodes and trematodes, which are not included in 

 the forage lists, were also present. The forage composi- 

 tion for all Thunnus species consisted of, by volume, 

 about 45% fishes, 35% cephalopods, 15% crustaceans, 

 and 5% miscellaneous items. 



Vertebrate Forage (Fishes) 



Fishes occurred with a frequency of 66-100%, 

 depending on the species of predator. Those consumed 

 by tunas ranged in length 9-360 mm SL (x 65 mm) and 

 represented 88 genera in 58 families. Fishes eaten by 

 Alepisaurus varied over a greater size range (8-846 mm 

 SL, x 98 mm) and represented 40 genera in 34 families. 

 The majority of forage fishes in Thunnus and 

 Alepisaurus were immature forms of midwater fishes and 

 epipelagic post-larvae and juveniles of shore fishes. 



Fish were present in 95% (266 stomachs) of the T. 

 albacares stomachs examined. Evaluation of fish forage 

 composition is based on 209 stomachs, however, in which 

 fishes were identifiable. Of the 48 T. alalunga examined, 

 79% (38 stomachs) contained fish. Evaluation is based on 

 14 stomachs containing identifiable specimens. Seventy- 

 three percent (38 stomachs) of the T. thynnus consumed 

 fish. Forage composition analysis is based on 24 

 stomachs containing identifiable specimens. Forage fish 

 were present in all 14 of T. obesus examined, but could 

 be identified further in only eight. 



The contents of Alepisaurus stomachs were generally 

 in better condition than those from the tunas. Ninety- 

 one percent (81 stomachs) of the Alepisaurus specimens 

 examined had consumed fish, most of which were iden- 

 tifiable at least to family. The percent identifiable from 

 Alepisaurus stomachs may be larger than from tunas 

 because identification was easier, perhaps because diges- 

 tion in Alepisaurus takes place mainly in the intestine, 

 the stomach being used only for storage (Rofen 1966). 



Thunnus albacares. — Forty families of fishes were 

 found in the stomachs of yellowfin (Appendix Table 2). 

 Those which occurred with at least 1% frequency are in- 

 cluded in Figure 2. 



Ten families occurred with 10% frequency or more: 

 Balistidae, 41%; Carangidae, 31%; Bramidae, 24%; 

 Chiasmodontidae, 21%; Syngnathidae, 18%; 

 Priacanthidae, 13%; Tetraodontidae, 13%; Holocen- 

 tridae, 12%; Acanthuridae, 11%; and Scombridae, 10%. 

 Except for the Scombridae, none was outstanding in 

 either size or number. In most families, one genus or one 

 species was chiefly responsible for its high frequency: 

 Monacanthus spp., Caranx sp., Pterycombus sp., 



