220 



Fishery Bulletin 94(2), 1996 



ranged from 1.8-20.0 cm (|i=9.1, o=4.1). Squid size 

 in the fall collections was more variable than in the 

 summer collections and less variable than in the 

 winter and spring collections. 



Patterns in size at age 



The average monthly growth rate in ML and weight 

 suggested that individual growth of L. pealei was 

 associated with maturity stage (Table 3) because 

 higher growth rates were apparent for more ad- 

 vanced maturity stages. Relatively large coefficients 

 of variation for the monthly growth rate indicated 

 that the pattern of individual growth was highly 

 variable, especially in terms of weight. Monthly 

 growth rates of indeterminate-sex squid averaged 8.8 

 mm per month in ML (4.6-15.4 mm/month) and av- 

 eraged 0.9 g per month in weight (0.1-3.0 g/month). 

 Monthly growth rates of females averaged 18.2 mm 

 per month in ML (8.8-36.6 mm/month) and averaged 

 9.5 g per month in weight (1.0-40.2 g/month). 

 Monthly growth rates of males averaged 25.9 mm 

 per month in ML (9.6-64.0 mm/month (and averaged 

 21.4 g per month in weight ( 1.5-98.2 g/month I. The 

 average growth rates in length of females and males 

 were more than twice the average rates for indeter- 

 minate-sex squid. Similarly, the average growth rates 

 in weight of females and males were more than 10 

 times the average rate for indeterminate-sex squid. 

 For males, the average growth rates by length and 

 by weight were 1.4 and 2.2 times greater than for 

 females. Overall, these data indicated that individual 

 growth was highly variable and was related to ma- 

 turity stage and sex. 



Average growth rates were tested for significant 

 differences by sex and maturity stage by using un- 

 planned multiple comparisons procedures. First, 

 squid were categorized as indeterminate-sex, female, 

 or male. With respect to maturity stage, squid were 

 categorized as mature if they were maturing or ma- 

 ture; otherwise they were categorized as immature 

 (Table 2). Bartlett's homogeneity of variance test was 

 then applied to the sex and maturity stage groups. 

 The natural logarithmic transformation was applied 

 to the growth-rate data to stabilize variance prior to 

 testing for differences. The null hypothesis of homo- 

 geneous variances for growth rate in length was re- 

 jected for the samples grouped by sex (X 2 - 

 40.42»X 2 05[2 |=5.99) but was accepted for the 

 samples grouped by maturity stage (X* 2 =0.40 < 

 X 2 005|1 ]=3.84). The null hypothesis of homogeneous 

 variances for growth rate in weight was rejected for 

 the samples grouped by sex (X 2 =9.35»X 2 05l2 i=5.99 ) 

 and by maturity stage (X 2 =26.46»5C 2 05 | , ,=3.84 ). Be- 

 cause variances were inherently heteroscedastic and 

 sample sizes were unequal, the unplanned compari- 

 son test of Games and Howell ( 1976) (Day and Quinn, 

 1989 ) was applied to test for differences among group 

 means of log-transformed length-at-age data catego- 

 rized by sex. The group means of indeterminate-sex 

 and female squid* I Y n Y F \ =0.71 >MSD 005[IF] =0.U), 

 indeterminate-sex and male squid ( I Y f -Y M I =1.00 > 

 MSD u5 |/ v/ |=0. 16), and female and male squid ( I Y F - 

 Y M I =0.30 > MSD 0Aj5lF M] =0.16) were significantly dif- 

 ferent at the 5% level. For the length-at-age data 

 categorized by maturity stage, the Tukey-Kramer 

 comparison test (Sokal and Rohlf, 1981) was used 

 because variances were homogeneous. The immature 



