Margulies et a\ Spawning and early development of captive Thunnus albacares 



253 



analyzed egg parameters possibly affected by size of 

 females (standardized egg production, egg size) only 

 during the period from June 1997 through July 1999. 

 During this period only the original broodstock group 

 was spawning, and we estimated that the majority of 

 females in the tank were actively spawning (i.e., ^20 kg 

 each). This eliminated any confounding effects due to 

 newly introduced immature females. Statistical analyses 

 of the data included linear regression, correlation, and 

 multiple regression, and followed the methods of Zar 

 (1984). Statistical programs were run in Microsoft Excel 

 and S-Plus 6.0 (Mathsoft, Inc., Seattle, WA). 



Results 



The spawning patterns and subsequent egg and larval 

 development are described only for the fish in the main 

 broodstock tank in the first 11 subsections. A short syn- 

 opsis of the spawning patterns of the fish in the reserve 

 tank is presented in the last subsection. 



Broodstock fish 



A total of 55 yellowfin tuna were initially stocked in 

 the main broodstock tank in June and September of 

 1996. At stocking, the fish ranged in length from 51 to 

 78 cm fork length (FL) (mean of 62 cm FL) and weighed 

 between 3 and 8 kg (mean of 5 kg). The sex composition 

 of the original 55 fish (determined later at death) was 

 54% female and 46% male. Spawning first occurred in 

 the main broodstock tank on 8 October 1996. At that 

 time, there were 24 females (ranging from 6 to 16 kg 

 and 65 to 93 cm FL) and 20 males (ranging from 6 to 14 

 kg and 66 to 86 cm FL) in the tank. From observations 

 of courtship behavior during the first 2 to 4 months of 

 spawning, it appeared that only a small group of larger 

 individuals (generally >80 cm FL) was spawning. This 

 pattern changed over time, as more of the broodstock 

 fish attained reproductive size, and by mid-1997 it 

 appeared from the courtship behavior that most of the 

 broodstock fish were participating in the spawning. By 

 mid-1997, there were 19 females (ranging from 13 to 

 30 kg and 87 to 108 cm FL, averaging 20 kg and 98 cm 

 FL) and 16 males (ranging from 15 to 30 kg and 89 to 

 112 cm FL, averaging 22 kg and 99 cm FL) in the main 

 tank. Using the estimated ages of wild yellowfin tuna 

 in the eastern Pacific Ocean (Wild, 1986) and applying 

 them to the lengths and weights of our broodstock fish 

 (Wexler et al., 2003), we estimated that the average age 

 of the broodstock fish in mid-1997 was approximately 

 2 years. 



The broodstock population ranged in number from 

 55 (in September 1996) to five (in July 1999). In mid- 

 August of 1999, we added 14 fish (ranging from 8 to 15 

 kg and 74 to 89 cm FL) to the main tank to supplement 

 the spawning group. In February of 2000, we added 

 another group of six fish (ranging from 4 to 14 kg and 

 58 to 71 cm FL) to the tank. The longest time that any 

 individual fish from these broodstock groups survived 



in captivity was 4.7 years, although average survival 

 in captivity was 1.9 years. Estimated ages of brood- 

 stock fish at death averaged 3.1 years and ranged from 

 1.4 to 6.1 years. Mortalities occurred gradually over 

 time — most often caused by strikes against the tank 

 wall — during the early morning hours before feeding. 

 The average size of the broodstock fish increased over 

 time, and reached a maximum in mid-August 1999. At 

 that time, there were four females (ranging from 46 to 

 77 kg and 141 to 150 cm FL, averaging 67 kg and 147 

 cm FL) and one male (50 kg and 133 cm FL) in the 

 main tank. Several of these fish attained sizes >85 kg 

 and 155 cm FL at the time of their deaths in March and 

 October 2000. Estimated growth rates in length for the 

 broodstock fish ranged from 0.9 to 4.0 cm/month and 

 decreased with increasing lengths of the fish. Estimated 

 growth rates in weight ranged from 0.8 to 1.6 kg/ month 

 for fish <19 kg and 1.7 to 1.9 kg/month for sizes >19 kg 

 (Wexler et al., 2003). 



Water temperature, spawning, and egg incubation 



The mean daily water temperature in the main brood- 

 stock tank fluctuated with the ocean temperatures and 

 ranged from 20.1° to 29.7°C, averaging 27.3'C (SD = 1.5) 

 (Fig. 2A). Periods of reduced temperatures occurred each 

 year during the coastal upwelling season from January 

 through March. Spawning occurred when the daily mean 

 tank temperature was 23.3° to 29.7°C, and averaged 

 27.7°C (SD = 1.1) at the time of spawning (Fig. 2A). The 

 minimum spawning temperature during most years and 

 seasons was about 24.0°C. The daily mean water tem- 

 perature decreased to 24.0°C or below during three of 

 the four years of the study; however, only two spawnings 

 occurred at mean temperatures below 24.0°C. During 

 most years, spawning occurred daily and continuously 

 for several months, but spawning ceased when tank 

 temperatures decreased below 24°C (March 1997, March 



1999, and February 2000) or decreased by at least 0.5°C 

 for at least one week (December 1997, October 1998, and 

 August 1999). The cessation of spawning in October 

 1998 extended for 5.5 months. During this period, water 

 temperatures were sufficient for spawning, although they 

 were steadily decreasing (from 28°C to <24°C) during 

 most of the period. 



The mean water temperatures in the egg incubation 

 tanks ranged from 21.9° to 29.8°C, averaging 27.6°C 

 (SD = 1.0) (Fig. 2B). The pattern of incubation tem- 

 peratures closely paralleled the pattern of mean daily 

 water temperatures for the main broodstock tank, 

 although the incubation temperatures did not decrease 

 as much as those of the main broodstock tank dur- 

 ing the upwelling period in early 1997 (because of 

 the influence of higher air temperatures on the in- 

 cubation tanks). From January 1997 through March 



2000, incubation temperatures and the broodstock 

 tank temperatures were highly correlated (/• = 0.74, 

 df=886, P<0.001), and we considered the two data 

 sets as equally representative of water temperatures 

 observed in the laboratory. 



