catches and in stomachs of both tuna species in 

 most areas in somewhat similar percentages, al- 

 though those in net micronekton were generally 

 higher than those in tuna stomachs. Adult squil- 

 lids were a significant item only in areas 4 and 5. 

 There they occurred comparably in net catches 

 and stomachs of yellowfin tuna; the skipjack ma- 

 terial, in which they did not occur, was scanty 

 for those areas. 



Percentages of port.unids were less similar; they 

 were generally much higher for yellowfin tima 

 stomachs than for net catches or skipjack tuna 

 stomachs. Percentages of unidentified crab mega- 

 lopa were variable, but generally low except in 

 stomachs of skipjack tuna from area 5. Sergestids 

 occurred in the net micronekton of some areas, 

 sparsely in the yellowfin tuna stomachs of areas 

 11 and 12, and not at all in skipjack tuna. Other 

 cnistacean families and groups in tables 8 to 14 

 (Pasiphaeidae, Penaeidae, Phronimidae, Oxyceph- 

 alidae, Phrosinidae, and Palinuridae) appeared 

 in the net catches only and on the whole sparsely, 

 in the areas to which the tables refer. 



Possible reasons for the diiferences in percentage 

 composition of the micronekton as caught by nets 

 and by tunas are given in the following para- 

 graphs. These reasons are in addition to the sparse 

 and probably unrepresentative sampling for net 

 micronekton and skipjack tuna stomachs in some 

 areas, to which reference has already been made. 

 It is also possible that tuna concentrate for feed- 

 ing upon aggregations of certain species of prey ; 

 such species might, therefore, be better represented 

 in tuna stomachs than in net catches. 



(1) The nets were used at night, but most tunas 

 from the surface-hook and surface-net fisheries are 

 caught in the daytime. Because diurnal vertical 

 movement causes many kinds of micronekton to be 

 more plentiful in near-surface waters by night than 

 by day, the nets and tunas proba^bly sample differ- 

 ent assemblages of micronekton. This jioint was 

 discussed by King and Iversen (1962), who found 

 some of the same differences between net-caught 

 and tuna-caught micronekton as those described 

 above. 



(2) The nets fisli tlirough a layer of water some 

 90 to 95 m. thick, but the range of depth at which 

 surface-caught yellowfin tuna and skipjack tuna 

 feed is unknown. If, as seems likely, the tuna feed 

 closer to the surface than the average depth of the 



net, epipelagic organisms would probably be bet- 

 ter represented in tuna stomachs than in net 

 catches. 



(3) Tuna may fail to catch organisms which are 

 available for capture, for lack of stimuli which re- 

 lease feeding behavior. Such stimuli can be visual 

 or chemical (Magnuson, 1963, and references 

 there). Visual stimuli could be lacking in the 

 presence of suitable prey by night, or by day if the 

 organisms are translucent. 



(4) Nets, even at a speed of 5 knots, may fail to 

 catch certain kinds of alert or strong micronekton 

 tliat could be captured by tunas, which are credited 

 witli swimming speeds up to 40 knots in short 

 bursts (Walters and Fierstine, 1964). 



(5) Tuna stomachs frequently contain a signifi- 

 cant proportion of semidigested material which is 

 not identifiable to family. This was so in Alver- 

 son's material. Some families of micronekton 

 might, therefore, be eaten by tunas despite their 

 absence in the appropriate columns of tables 8 to 

 14. 



A combination of tlie first and third points (net- 

 ting by night, but tuna feeding by day) could ex- 

 plain the much better representation of mycto- 

 phids, sternoptychids, nemiclithyids, stomiatids, 

 idiacanthids (and perhaps other mesopelagic fish 

 families mentioned above, and sergestids), in the 

 net-caught than in the tuna-caught micronekton. 

 These groups are known to move closer to the sea 

 surface by night (when tunas probably cannot see 

 them well) than by day (wlien they are probably 

 too deep, generally, for the tunas to catch). Alver- 

 son (1961) showed that tunas will catch the mycto- 

 phid Benfhosema pterota when it is available at 

 the surface in daylight. It is not clear that this 

 explanation could apply to leptocephali, but these 

 animals might not be taken by tunas because of 

 their translucence; the same might apply to the 

 phronimids, oxycephalids, phrosinids, and palinu- 

 rids (the last is represented in the net catches only 

 as translucent larvae) . 



The second and fourth points (the net catching 

 at greater depth and slower speed than the tunas) 

 could explain the greater occurrence of active epi- 

 pelagic fish (scombrids, scomberesocids, thunnids, 

 exocoetids, coryphaenids, polynemids, carangids, 

 and perhaps others) in tuna-caught than in net- 

 caught micronekton. The Exocoetidae (flying- 

 fishes) are a good example : only one small speci- 



106 



U.S. FISH AND WILDLIFE SERVICE 



