464 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



CaJSTACEANS 

 I 



craJSTACEA^e 



OTHER FOOD Z_ 



CRLSTACEANS 

 OTHER EBPO ' 



, , _^ 'OTHER 



-^t:::>-TvbLLUScs 



OTVIER 

 MOLLUSCS 



LDNGLINE 



tN-.l821 



TROLL 



GILL NET 



(N.e7i 



Figure 5. — Comparative Importance, by volume, of major food elements found in 348 albacore stomachs, according 



to method of capture. 



(Graham, 1959). Another possibility is that the 

 gill net-caught fish regurgitated food wliile strug- 

 gling to escape the net. A third possibility is 

 that albacore do feed at night, but at a reduced 

 rate. These suggestions are discussed more fully 

 in the section dealing with feeding related with 

 time of day. 



DifTerences among the aggregate total volumes 

 of major food items of albacore captured by troll- 

 ing compared with longline-caught and gill netted 

 fish are evident (fig. 5). Fish and squid formed 

 approximately equal portions of the food of long- 

 line-caught albacore, while fish comprised 79 per- 

 cent and squid comprised 11 percent of the diet of 

 troll-caught albacore. This agrees with McHugh's 

 data (1952), which showed fish as 68 percent and 

 squid as 12 percent of the food volume of troll- 

 caught albacore. In both cases saui-y, ColoJahls 

 sp., comprised the bulk of the fishes. Rcintjes and 

 King (1953) also found fishes to form a larger 

 portion of the diet of troll-caught yellowfin tuna 

 when compared with longline-cauglit yellowfin in 

 both the Line Islands and Phoenix Islands areas. 



Perhaps the trolling method, which employs a 

 lure skipping and plunging along a few feet below 

 the surface, may especially attract albacore pre- 

 viously conditioned by a diet of fish liaving the 

 gross characteristics of a trolling lure. If this is 

 true, and there were numbers of albacore in the 

 trolling area which had been feeding on organisms 

 (e.g. crustaceans) which do not have these char- 

 acteristics, a portion of the available albacoi'e 

 might not be efficiently exploited. Joubin and 



Roule (1918), liowever, found that amphipods 

 were the main food of troll-captured albacore in 

 the Gulf of Gascony. 



Yuen (1959) has pointed out that the feeding 

 behavior of skipjack tuna (Katsuuwnus pelanm) 

 may be conditioned by previous feeding. He hy- 

 pothesized that livebait fishing methods iised in 

 Hawaii take advantage of an already- existing 

 feeding excitement in the skijjjack. 



The frequency of occurrence of major food 

 groups is shown in figure 6. Differences between 

 the longline and gill net-caught fish are large, 

 but even if the gill net data are considered 

 atypical, there are still substantial differences be- 

 tween longline- and troll-caught fish. The spread 

 between values for these two groups for three 

 classes of food are as follows: squid, 35 percent; 

 fish, 27 percent ; crustaceans, 24 percent. It may 

 be simply that the larger, longlined fish require 

 more food than the smaller, troll-caught albacore 

 and thus would be apt to have more types of food 

 in their stomachs. Possibly the reason squid and 

 fish dominate in gill netted albacore is that their 

 hard parts (squid beaks and eye lenses and fish 

 vertebrae) remain in the stomachs after other or- 

 ganisms have been completely digested. 



Representatives of 32 fisli families and 11 in- 

 vertebrate orders were found in the food of long- 

 lined albacore, compared to 9 fish families and 10 

 invertebrate orders for troll-caught and 4 fish fam- 

 ilies and 5 invertebrate orders for gill net-caught 

 albacore. The most frequently occurring fish fam- 

 ilies in the longlined albacore were Gempylidae, 



