Itoh et al.: Swimming depth, water tempeiature preference, and feeding of Thunnus orientalis 



539 



The tag temperature used in our comparison was the 

 temperature at m depth recorded in the summary file 

 because that file contained a much larger number of days 

 than that of the detail file. In support of this, we confirmed 

 that the average water temperature over all depths for a 

 day in the detail file had only slight differences of -0.1 

 ±0.7°C in average (range: -4.0-+4.3°C) from the tem- 

 perature at m recorded for the day in the summary file. 

 Because the fish swam near the surface, the temperature 

 recorded to represent m depth also represented well the 

 temperature at all depths where the fish actually swam. 

 Frequencies of days were summed from the data for all 

 individuals in one-degree temperature bins separated by 

 year, by month, and by area, such as the East China Sea, 

 the Sea of Japan, and the Pacific Ocean. 



The water temperature recorded by archival tags com- 

 monly ranged from 14° to 20°C (Fig. 4). When water of this 

 temperature range was located within an accessible area 

 (e.g. many months in the East China Sea and the Pacific 

 Ocean, and November in the Sea of Japan), almost all fish 

 were found in such water. Where water temperature was 

 higher (e.g. June 1996, June 1997, and between June and 

 October 1998 in the East China Sea) or lower (e.g. May 

 1996 and April 1997 in the Sea of Japan), fish tended to 

 choose water that was close to this range. However there 

 were a few cases in which fish stayed in water with a 

 temperature outside of this range (e.g. between July and 

 September 1998 in the Sea of Japan) even when water of 

 the 14-20°C range was accessible to them. 



Visceral temperature and feeding events 



Temperature of the fish viscera ranged from 13.0° to 

 30.9°C. It was usually higher than the water temperature 

 and the thermal excess for a given individual ranged from 

 1.3° to 4.6°C, and averaged 3.0° ±1.0°C. The visceral tem- 

 perature changed in parallel with the water temperature 

 all year (Fig. 5). 



In preliminary experiments, pen-held fish were fed com- 

 pletely thawed mackerel of 20-30 cm FL twice a day at 

 approximately 900 h and 1500 h. The following changes in 

 visceral temperature before and after feeding were generally 

 observed (Fig. 6). All figures given below regarding thermal 

 excess and timing after feeding are average values. The vis- 

 ceral temperature in a stable state just before feeding (n=5) 

 was 3.7°C higher than ambient water temperature. This 

 thermal excess decreased to 2.3°C at 22 minutes sifter feed- 

 ing, and then increased to 7.6°C at 7.7 hours after feeding. 

 Then it slowly decreased again and reached a stable thermal 

 excess of 3.1°C at 21.0 hours after feeding. When the fish 

 fed again before the visceral temperature stabilized (n=7), 

 with the thermal excess still as high as 6.8°C, the thermal 

 excess reached a high of 8.5°C but the increase after feeding 

 was 1.7°C, much smaller than that observed in the previous 

 case (3.9°C ). The time required to reach the highest visceral 

 temperature and the time to change back to a stable state 

 were similar to the previous case. When the fish were not fed, 

 because of rough sea conditions, the visceral temperature 

 stayed stable all day (e.g. 16 January on Fig. 6). 



