Itoh et al.: Swimming depth, water temperature preference, and feeding of Thunnus onentalis 



537 



12 16 

 Time of day (h) 



Figure 1 



Sample records of swimming depth (thick solid line), water temperature ( lower thin solid line), and temperature of viscera 

 (upper thin solid line) for Pacific bluefin tuna from the detail files of two archival tags. Upper panel: records of a fish from 

 23 to 25 December 1995 in the East China Sea. Lower panel: records of another fish from 15 to 17 June 1997 in the East 

 China Sea. Shadows indicate nighttime. 



to near the sea surface (upper panel in Fig. 1, and Fig. 3). 

 At dusk, after a rapid descent from near the sea surface to 

 the maximum depth of 89 ±34 m, fish slowly and steadily 

 ascended for about 40 minutes. The time that maximum 

 depth was recorded was most frequently at four minutes 

 before (or after) the time when the archival tag sensed the 

 first (or final) light at dawn (dusk) for 29% (37%) of cases 

 where this behavior was observed. The end (onset) of rapid 

 ascents (descents ) occurred most frequently at the time when 

 the tag sensed the first (final) light, that is to say for 68% 

 (49%) of all cases. The light detection threshold for the tags 

 used in our study corresponded to an intensity in the blue- 

 green transmission window of seawater (470 nanometers 

 (nm) center, 60 nm width of the filter) of about 3 x 10"'' times 

 the surface noon solar intensity on a clear day in that same 

 spectral band. Because the light-sensitive region of the 

 measurement stalk was located below the animal's body, 

 we assumed that the eyes encountered a somewhat higher 

 intensity than did the light sensor. Quick examination of 

 tags in air in the early October revealed that the time when 

 the archival tag sensed the first light at dawn or the final 

 light at dusk was about 40 minutes before sunrise or about 

 40 minutes after sunset, respectively. Some combination of 

 these vertical swimming behaviors was observed in 1081 

 of 1452 days (74%) during which the detail file data were 

 available for both dawn and dusk. They occurred at both 

 dawn and dusk in 679 days (47%), only at dawn in 77 days 

 (5%), and only at dusk in 325 days (22%). 



Occurrences of these behaviors differed by area and sea- 

 son (Table 1). The area was determined by using fish loca- 



tions estimated as described in Itoh et al. ( 2003 ). The average 

 occurrence of these depth excursions at dawn and dusk was 

 as high as 87-88% in the East China Sea from November 

 to January but decreased to 15-49% from February to June 

 and October. In the Sea of Japan, the average occurrence 

 was 4-39% in April, May, and September to November. In 

 the Pacific Ocean, the behaviors were observed in 75-90% of 

 days from May to July, in contrast to the relatively low occur- 

 rences within the East China Sea in the same season. 



Water temperature 



Water temperatures recorded in archival tags ranged from 

 8.3° to 28.4°C at m depth, and 1.4° to 28.4°C when all 

 depths were combined. A range of water temperatures 

 that appeared to be preferred by young Pacific bluefin tuna 

 was estimated. A simple frequency distribution of recorded 

 water temperatures was inappropriate because fish could 

 be forced to tolerate water out of their preferred tempera- 

 ture range because of a lack of water with a more favorable 

 temperature within the geographical range accessible to the 

 fish. Instead we compared the range of water temperature 

 within the geographical area accessible to the fish with the 

 frequency distribution of actual recorded water tempera- 

 tures, grouped by year, month, and sea. We assumed that 

 the accessible areas in each sea were those areas that a 

 tag reported as its position. These include areas along the 

 Japanese coast between 35° and 45°N in the Sea of Japan, 

 30-45°N in the western Pacific Ocean west of 160°E, and 

 25-45°N in the eastern Pacific Ocean east of 160°E. In 



