FOOD OF YELLOWFIN TUNA 



95 



Each organism was identified as completely as 

 was practicable and the number of each species 

 or group of organisms was recorded. Each sub- 

 division or kind of food was then measured volu- 

 metrically by the displacement of water in a 

 graduated cylinder of appropriate size. Bait used 

 to capture the yellowfin was not considered in 

 this analysis. 



The methods used to identify food organisms 

 and organic remains varied to a great extent, and 

 depended on the group of animals under considera- 

 tion. The arthropods were identified to taxonomic 

 order from general exoskeletal characteristics. 

 Although Euphausiacea, Mysidacea, and De- 

 capoda were difficult to separate, the scarcity of 

 the first two orders made the problem of secondary 

 importance. Amphipoda, Isopoda, Stomatopoda, 

 and most Decapoda were easy to distinguish to 

 taxonomic order even with partial remains. Some 

 members of the amphipod families of Calliopiidae, 

 Hyperiidae, Oxycephalidae, and Gammaridae 

 found in this region were recognized by certain 

 cephalic characteristics and body shape. The 

 species of Stomatopoda, which even when badly 

 damaged were readily distinguished by certain 

 exoskeletal features, have been described by 

 Brooks (1886), Kemp (1913), Edmondson (1921), 

 Bigelow (1931), and Townsley (1950), who re- 

 ported on Pacific forms and constructed keys 

 for their separation. Such decapod crustaceans 

 as the postlarval Palinuridae and Nephropsidae 

 were readily identified because each possesses 

 unique familial traits. Identification of the latter 

 family was aided by the work of Holthius (1946). 



The shelled molluscs found in the stomachs of 

 tuna were readily identified as heteropods, ptero- 

 pods, and nautiloid cephalopods from distinguish- 

 able shell remains. The cephalopods were sepa- 

 rable to squid and octopods on the basis of general 

 body shape, number of arms, presence or absence 

 of a gladius (pen), and modification of suckers 

 into hooks. The presence of tentacles (fifth pair 

 of arms) was used as a distinguishing trait in all 

 squid except in the aberrant family of eight- 

 armed squid, Octopodoteuthidae, where the modi- 

 fication of suckers into hooks indicated a decapod 

 mollusc. Berry (1914) used the perforation of 

 the eyelid, arrangement of the suckers, and the 

 hectocotylus (modification of one tentacle into an 

 accessory sex character in the male) to distinguish 

 genera and species of cephalopods; however, all 



251078-53 2 



of these structures are susceptible to the destruc- 

 tive action of the digestive juices so that in most 

 instances organisms were identifiable only as 

 squid or octopods. 



Pelagic tunicates were seldom identifiable to 

 family because the soft body readily disintegrated 

 into gelatinous fragments in the tuna stomach. 

 These remains retained certain characteristics, 

 however, that distinguished them from the 

 coelenterates and pelagic molluscs with similar 

 gelatinous structure. 



The teleost fishes were readily recognized by 

 their skeletal remains. Further identification, 

 even to taxonomic order, was dependent on cer- 

 tain traits, many of which were readily lost. 

 Engulfment often separated the head from 

 the body, mutilated the fins, skin, and lateral line, 

 and removed scales, making identification diffi- 

 cult. Fishes with bony protuberances, carapace- 

 like integument, and other distinguishable hard 

 parts were the most easily identified. Familial 

 identification was often dependent upon singular 

 characteristics, such as bony scutes in the Ca- 

 rangidae and teeth and mandibles in the Tetrodon- 

 tidae, Diodontidae, Alepisauridae, Aulostomidae, 

 Belonidae, and Hemirhamphidae. In more gener- 

 alized groups possessing neither unique nor resist- 

 ant parts, identification could not be easily made. 

 Juvenile fishes often lack traits characteristic of 

 the adults; for these, identification to the family 

 frequently was impossible. It is estimated that 

 80 percent of the fish specimens could be identified 

 to family. 



The most useful references for the identification 

 of fishes were Fowler (1928, 1931, 1934, 1949), 

 Jordan and Evermann (1905), Gilbert (1905), 

 Weber and De Beaufort (1913-1936), De Beau- 

 fort (1940), Schultz (1943), Brock (1950), and 

 De Beaufort and Chapman (1951). Berg's (1947) 

 modification of Regan's system of classification 

 was used for the forage fishes. A reference collec- 

 tion of invertebrates and fishes maintained at our 

 laboratory in Honolulu was used intensively 

 during the study. 



EVALUATING FOOD COMPONENTS 



Three general systems of analysis and methods 

 of expressing results have evolved from the many 

 investigations of the stomach contents of fish, 

 birds, mammals, and other animals. These sys- 

 tems might be termed the "numerical," the 



