The above is a generalized view of the food habits of the yellowfin tuna of 

 the Celebes Sea, In what follows a brief record will be made of some of their pe- 

 culiarities „ 



On both cinises specimens of the sabahii. Chanos chanos (ForskSl) appeared 

 from time to time in the record, but it should not be thought that these are natural 

 food. The bait fish wliich the ShSnan Maru used in the experiment were salpahii reared 

 iu southern Formosao The specimens of this fish v^ich appeared in the stomachs of 

 the yellowfin should rightly be considered to be bait fish rtiich the tuna swallowed, 

 as is shown by the fact that theae fish were hardly digested at all., 



Small shore fishes like Monacanthus - Balistes. and Ostracion . which can hardly 

 be thought to be pelagic, appeared in large numberso One reason for this is prob- 

 ably that the tuna migrate in close to shore, however^ these fishes are often seen 

 accompanying drifting timbers and so forth and it is wondered whether this may not 

 be the explanation of thair appearing in such comparatively large numbers among the 

 natural foods of the tuna. At any rate it is interesting to note that fishes of 

 this k5nd are rather important as food for the yellowfin tuna. 



Among the crustacoans the Palinuridae, Squills sp., Scyllaridae, and a 

 tenagaebi ( Leander sp„ ) were seen, with the tenagaebi the most common. 



It is remarkable that a?nong the fishes an Apodes resembling the conger eel ap- 

 peared once and two pipefish were found. 



As stated above, the nsost comn.on mollusks mere squids, but it should oe noted 

 that many specimens of nautilus (?,' also appeared. The fish in which these speci- 

 mens were found were taken on grounds near the aforementioned Karakelong I^ 



The following is a discussion of seasonal differences. 



Since the experiments covered only one voyage in each season and the materials 

 available for study were in the numbers detailed above, it may be inappropriate to 

 use them to show seasonal changes, however, as was stated before, all but a part" of 

 the fishing stations were in exactly the same sea area and it is not therefore im- 

 possible to consider the marked differences between the results from the two cruises 

 as being due to seasonal variation. For this reason the marked differences between 

 the results from the two cruises will be here treated as seasonal variations. Of 

 course, any time a more detailed survey is made in the future there will be plenty 

 of room for corrections. 



To take up the squid first, on the 1933 cruise, which ?jas made during the 

 season when the northeast monsoon prevails, squid appeared in the stomachs of the 

 fish taken at all but one fishing station, but on the 193A cruise, made during the 

 southwest monsoon season, squid appeared at only 11 out of the 23 stations. This 

 is considering the problem by fishing stations, but if we look at it from the point 

 of view of the number of fish whose stomachs contained squid, in 1933 24^ out of 3A 

 tuna were eating squid, \B^ich is over 70;£ of the total, while in 1934, with 11 out 

 of 23 fish eating squid, this percentage dropped sharply to less than 4,8/5. 



In 1933 a considerable nurriber of anchovy ( Engraulis) appeared in the material 

 studied, but in 1934 they v?ere entirely absent and instead a herring ( Stolephorus ) 

 Trtiich htid not appeared in the previcis year's collections was found in some numbers. 



The trunkfish ( Ostracion ) was remarkably abundant in 1933, hit in 1934 it de- 

 creased markedly. At the same time there was a marked increase in the numbers of 



-4 



