seawater at a given level of oxygen saturation. 

 Swimming speed and survival time were mea- 

 sured. They found that survival time and swim- 

 ming speed were independent of oxygen levels in 

 excess of 4 ppm; below 4 ppm survival time was 

 directly and swimming speed inversely propor- 

 tional to dissolved oxygen amounts. So apparently 

 4 ppm is close to the incipient lower lethal limit for 

 skipjack tuna under the given experimental con- 

 ditions. For modeling distribution limits, Barkley 

 et al. (see footnote 1) used a more conservative 

 figure of 5 ppm. 



However, a physiological limit of 4 or 5 ppm is 

 not necessarily a behavioral limit; if the limit is 

 approached slowly under natural and otherwise 

 unstressful conditions, can a fish adaptively re- 

 spond? Whitmore et al. (1960) found that coho 

 salmon, Oncorhynchus kisutch, avoided water of 

 lowered oxygen levels yet which produced no res- 

 piratory distress. In contrast, kawakawa, Eu- 

 thynnus affinis, a species closely related to skip- 

 jack tuna, tolerated 2-ppm water for short periods 

 in order to get food (Chang and Dizon 3 ). 



In the present experiment, I tested the re- 

 sponses of free-swimming tunas — both skipjack 

 tuna and yellowfin tuna, Thunnus albacares — 

 encountering slowly changing oxygen levels. The 

 rate of change was comparable with that which a 

 tuna might encounter in nature. Yellowfin tuna 

 were tested for comparison because they are 

 abundant in the same areas of the eastern tropical 

 Pacific avoided by large skipjack tuna. Finally, 

 salinity fronts have been suggested as a factor 

 determining distribution, so responses to decreas- 

 ing salinity levels were also examined. 



Materials and Methods 



Eight skipjack tuna and three yellowfin tuna 

 were tested with decreasing oxygen levels, and 

 three skipjack tuna, and one yellowfin tuna were 

 tested with decreasing salinity levels. Fish were 

 chosen from stocks at the Kewalo Research Facil- 

 ity of the Southwest Fisheries Center, National 

 Marine Fisheries Service, NOAA, Honolulu, 

 Hawaii. Tuna stocks for this experiment were kept 

 in outdoor tanks (7.3 m diameter x 1.2 m deep) 

 until used; they were then removed by angling 



3 Chang, R. K. C, and A. E. Dizon. Low oxygen levels as 

 barriers to voluntary movements of tunas. Manusc. in prep. 

 Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., 

 NOAA, Honolulu, HI 96812. (Material presented at 26th Tuna 

 Conference, Lake Arrowhead, Calif., 29 Sept.-l Oct. 1975.) 



with a barbless hook and transferred to the swim 

 chamber in a plastic bag partially filled with wa- 

 ter. This is a good transfer technique since fish on 

 occasion have fed immediately after transfer. 



The responses of tunas to decreasing oxygen and 

 salinity levels were examined in a tank system 

 consisting of a swim chamber equipped with 

 photocells for monitoring and recording fish be- 

 havior. (For details see Dizon et al. 1977.) The 

 swim chamber was a 6.1 m diameter x 0.61 m deep 

 fiber glass tank fitted with a concentric inner wall 

 so the fish was constrained to swim in a 0.75-m 

 channel around the periphery. Six laps equaled 

 100 m. Water (24°C) was introduced and removed 

 from the swim channel through two pairs of con- 

 centric rings of polyvinyl chloride pipe. Entering 

 (or exiting) water divided equally into two inflow 

 (or outflow) pipes, each flowing countercurrent to 

 the other to provide minimum oxygen or salinity 

 asymmetry and horizontal transport of water 

 within the swim channel. Water was recirculated 

 through an outside circuit at 1,136 liters/min to 

 insure thorough mixing of any introduced new 

 water. New seawater was added to the tank at 

 38 liters/min. 



Oxygen was reduced in the tank by replacing 

 the 38 liters/min new seawater with 38 liters/min 

 anoxic seawater obtained at our well head before 

 aeration and introduced into the intake of the 

 1,136 liters/min recirculation pump. Oxygen de- 

 creased approximately exponentially within the 

 swim chamber — 0.06 ppm/min after 30 min and 

 0.03 ppm/min after 60 min. Salinity levels in the 

 swim chamber were reduced by introducing aer- 

 ated freshwater (38 liters/min) into the pump in- 

 take. Salinity decreased exponentially — 0.07%o/ 

 min after 30 min and 0.03%o/min after 60 min. 



Passage of the fish was sensed at four photocell 

 stations (six photocells/station) at 90° intervals 

 around the periphery of the swim channel. Infor- 

 mation from the photocells was translated into 

 swimming speed (minutes per lap), direction 

 (clockwise or counterclockwise), and frequency of 

 reversals or swimming direction by digital logic 

 equipment and printed on adding machine tape. 



Procedures were quite simple; tuna (starved for 

 1 day) were moved into the tank and allowed 100 

 min to habituate; swimming speeds were continu- 

 ously recorded to provide baseline data to compare 

 with behavior during periods of changing oxygen 

 or salinity. After 100 min, a test was started and 

 behavior was recorded as salinity or oxygen de- 

 creased. Oxygen and salinity levels were allowed 



650 



