826 



Fishery Bulletin 88(4). 1990 



I- 



g 08 



W= 0.000031 SL' 

 n = 184 



10 15 



STANDARD LENGTH (MM) 



Figure 6 



Weights (W) and standard lengths (SL) cif 184 late-larval and early- 

 juvenile black skipjack tuna measured soon after capture. The 

 power function of the log-log fitted regression line and 95% predic- 

 tion belts of weight on length are included. The r- of the log-log fit 

 was 0.927, the 95% confidence limits of the regression coefficient 

 were 2.6641-2.8917 g/mni, and the standard error of the estimate 

 was 0.1790. 



those for fish held longer (Fig. 5d). The greatest aver- 

 age growth rates in weight (3.9-4.0 g/day) were re- 

 corded for fish surviving 60-73 days {n = 3). After that 

 time, average daily weight increases declined dras- 

 tically (ri = 4). 



Weight-length relationship 



Capture weights of the live black skipjack held for 

 growth experiments (Fig. 4) were estimated using a 

 weight-length regression equation based on fresh mea- 

 surements and weights of 184 other black skipjack 

 ranging from 5.6 to 19.7 mmSL and 0.004 to 0.132 g 

 (Fig. 6). The power function of the fitted regression 

 is lnU'= -10.3806 -H 2. 77791aSL. The standard er- 

 rors of parameters Ina and h are 0.1326 and 0.0577, 

 respectively. 



Tests for normality of the residuals (Filliben 1975), 

 independence or lack of autocorrelation of the residuals 

 (Durbin-Watson statistic), and a constant variance 

 (homoscedasticity) of the residuals (Wesolowski 1976) 

 failed to indicate any violation of the assumptions of 

 linear regression. 



Effect of temperature 



Sea-surface temperatures during each sampling trip 

 are shown in Figure 2. The data show a gradual decline 

 in temperatures commencing in November or Decem- 

 ber, reaching lows of about 23-25 °C in March, followed 

 by a gradual warming to maximum stable temperatures 

 of about 28-29°C from April or May until October or 

 November. Black skipjack spanning the entire size 

 range encountered (7.1-18.4 mmSL) were taken when 

 the temperature ranged between 28.0 and 29.0°C. Only 

 smaller individuals (7.7-12.2 mmSL) were caught when 

 sea-surface temperatures were below 27.0°C, but the 

 sample size was low (n = 4). 



Neither the growth rates nor final sizes attained in 

 captivity were significantly correlated with mean water 

 temperatures in the aquaria over the course of the ex- 

 periments. However, the final lengths (r= -0.559, 

 P< 0.001) and weights attained (r = -0.572, P< 0.001), 

 average daily growth in weight (/• = -0.503, O.OOK 

 P<0.002), and the number of days in captivity (r = 

 -0.568, P< 0.001) were negatively correlated with 

 minimum temperatures recorded in the laboratory. No 

 such relationships were observed with the maximum 

 temperatures recorded. 



Discussion 



Except for the black skipjack held in shipboard aquaria 

 by Clemens (1956), these experiments are the first in 

 whicli late-larval or early-juvenile tunas were collected 

 in the wild and reared in captivity for experimental 

 growth studies. Similar experiments on other scom- 

 l)rids are being conducted at the Achotines Laboratory. 

 To our knowledge, 167 days is the longest time any 

 scombrid has been held in captivity beginning at an 

 early life stage. Harada et al. (1973) held .4((.r('.s- tapei- 

 nosomn ( = A. rochei) larvae for 52 days, Harada et al. 

 (1974) reared Sarda orientalh larvae up to 99 days, 

 and Harada et al. (1980) grew Thunnus nlbacares 

 larvae for a maximum of 38 days. These larvae were 

 all hatched from artificially-fertilized eggs obtained 

 from ripe females, and fed unspecified rations of zoo- 

 plankton and fish larvae. Previous to our studies, young 

 E. lineatus were reared on only one occasion. Clemens 

 (1956) collected black skipjack by dipnet at night, and 

 held 10 individuals averaging about 27 mmSL for up 

 to 12 days in shipboard aquaria. Average growth in 

 length was 3.1-3.6 mm/day. The paper does not state 

 at what water temperature the fish were held nor how 

 much food was provided during the experiment. The 

 growth rates measured in Harada et al.'s (1973, 1974, 

 1980) experiments were 3.0, 3.0, and 1.3 mm/days, 

 respectively. The growth rates reported by Clemens 



