224 



Fishery Bulletin 94(2), 1996 



for female samples hatched during November-May 

 was the case-IV curve which was selected over the 

 case-II curve because the hypothesis a-0 was ac- 

 cepted f/=2.68 < 3.98=F 005 ( 1,67)). The best curve for 

 male samples hatched during June-October was the 

 case-II curve which was selected over the case-IV 

 curve because the hypothesis a=0 was rejected 

 (/=5.46 > 3.97=F (M)5 (1,73>>. For each of the 7 other 

 subsets of length-at-age data, case IV yielded the best 

 growth curve (Table 8). 



Case-IV curves were the best growth models for 

 length at age of pooled-sex samples (Fig. 4A) and 

 pooled-sex samples hatched during November-May 

 (Fig. 5A). In general, the case-IV growth curves are 

 power functions that represent unbounded acceler- 

 ating growth with a theoretical minimum size. In 

 particular, case-IV curves are what Schnute termed 

 "T 111 power" growth curves, where size (G) at age (T) 

 is proportional to some positive constant K>1 raised 

 to the T h power. That is, G oc K 7 = exp[T InCfiO], so 

 that size increases exponentially as age increases. 

 In contrast, the case-II curve was chosen for pooled- 

 sex samples hatched during June-October ( Fig. 5A). 

 For these samples, growth in length appeared nearly 

 linear over the range of data, although there was 



some indication of a slight decrease in growth rate 

 past 250 days of age. 



For the length-at-age data partitioned by sex, case- 

 rV curves were also the best growth models, with 

 one exception. Exponential growth in length was 

 evident for indeterminate-sex (Fig. 6A), female (Fig. 

 7A), and male (Fig. 7A) samples. Similarly, growth 

 in length was exponential for female samples hatched 

 during June-October and during November-May 

 (Fig. 8A) and for male samples hatched during No- 

 vember-May (Fig. 9A). In contrast, growth in length 

 for males hatched during June-October was not ex- 

 ponential (Fig. 9A) and growth rate decreased be- 

 tween 200 and 250 days of age. 



For the subsets of the weight-at-age growth data, 

 most of the case-II and case-III curves were rejected 

 because their parameters were not significant and 

 there were only 4 instances where comparison of RSS 

 was used to select the best curve. The best curve for 

 pooled-sex samples was the case II curve which was 

 selected over the case-IV curve because the hypoth- 

 esis a=0 was rejected f/=4.30 > 3.87=F 1)5 < 1,350)). The 

 best curve for indeterminate-sex samples was the 

 case-II curve which was selected over case-IV curve 

 because the hypothesis «=0 was rejected (/~=5.61 > 



