Brodziak and Macy: Growth of Loligo pealei 



229 



- 



-f*^* 



Age (days) 



Figure 6 



Growth curves of indeterminate-sex Loligo pealei in length (A) and in 

 weight (B). 



tively exponential, which is consistent with the find- 

 ings of Forsythe and Van Heukelem (1987). 



It is important to account for individual heterogene- 

 ity in size at age when determining a growth curve 

 because the use of mean values for size classes can ob- 

 scure the pattern of individual growth ( Alford and Jack- 

 son, 1993). In this study, growth curves were estimated 

 by using size-at-age data for individual squid. We found 

 that the use of a multiplicative lognormal error term 

 provided an adequate statistical framework for esti- 

 mating growth curves despite substantial heterogene- 

 ity in squid size at age. In general, a growth curve with 

 a multiplicative lognormal error term will be more ap- 

 propriate than an additive normal error term when size- 



at-age distributions are skewed because prediction er- 

 ror scales with size at age under the lognormal assump- 

 tion. Regardless, it is recommended that researchers 

 use residual patterns to evaluate the adequacy of an 

 assumed error structure for estimating a growth curve. 

 The fact that some L. pealei were hatched during 

 December-April was unexpected because the pro- 

 tracted spawning season of L. pealei in the north- 

 west Atlantic had previously been considered to ex- 

 tend from roughly April to November (Lange, 1982). 

 McMahon and Summers (1971) found that embry- 

 onic development of L. pealei was related to water 

 temperature. In their study, the average development 

 time required before hatching was 26.7 days for wa- 



