Hazell el al : Somatic growth rate of Jasus lalandii 



513 



multiple range post hoc tests (SigmaStat, version 2.03, 

 [SPSS Inc., 1997]). Of greatest interest in this analysis 

 were the results of post hoc tests in which means of the 

 corresponding size classes of the two sample periods were 

 compared. 



For comparison of length-related growth increments 

 at the two sites in our study, individual molt increments 

 were plotted against premolt carapace lengths, for each 

 combination of sex and site, and their linear regressions 

 were compared by using a two-way ANCOVA (Zar, 1984; 

 Norman and Streiner, 1994). 



Investigating relative intermolt periods from field 

 samples requires identifying those specimens that are 

 considered to be molting. Rock lobster juveniles in the 

 hard old shell state isensii Pollock 11973]) were easily 

 identified as such by visual inspection, and we confirmed 

 this by breaking the tip off an antenna and inspecting the 

 condition of the underlying integument. Furthermore, on 

 returning these specimens to the laboratory, they generally 

 molted within 20 days. Because field samples were taken 

 at intervals at least this long (Table 1), the proportion of 

 each sample found to be in the hard old shell state was a 

 reasonable indication of the relative number of juveniles 

 molting during that sample month. This proportion, in 

 turn, was used as a proxy for intermolt period under 

 the assumption that shorter intermolt periods would 

 result in relatively higher numbers of juveniles that were 

 molting. The numbers of specimens with soft old or soft 

 new shells did not accurately reflect the relative rate of 

 molting because, as a result of behavioral modifications, 

 rock lobster in these conditions may have had a lower 

 catchability than those with hard shells. 



Using this index of relative intermolt period, we per- 

 formed a log-linear analysis (Zar, 1984; Norman and Strei- 

 ner, 1994) to assess the interdependence of the factors site, 

 size-category, sample month, and molt state with regard 

 to the frequency of observations in each subcategory. Only 



size classes smaller than 40 mm carapace length were used 

 in our analysis because the decreasing molt frequency 

 with increasing carapace length means that larger sample 

 sizes are required to accurately estimate the proportion of 

 animals of the size class in which rock lobsters are molting 

 in the population. However, the length distribution of ju- 

 venile lobsters meant that sample sizes decreased above 

 40 mm CL. 



Results 



An average of 323 juvenile J. lalandii was collected per 

 sample at the harbor wall (256-464), and 291 at Mouille 

 Point (239-365). After directed collections for additional 

 premolt specimens of various sizes, a total of 375 juveniles 

 from the harbor wall and 383 from Mouille Point were 

 maintained in the laboratory for analysis of growth rates. 

 Mortality rates were low (<1%) among these specimens, 

 except during December 1996, when aquarium malfunc- 

 tion resulted in extensive fatalities during molting. Can- 

 nibalism was not observed during laboratory trials. 



Individual molt increments amongst the captive spec- 

 imens were highly variable, ranging from 0.4 to 5.5 mm. 

 However, mean molt increments for each of seven con- 

 secutive 5-mm-CL size-intervals ( Table 2 ) were consistently 

 and significantly smaller in our study (1996-97 season) 

 than those measured in 1971-72 (Table 3, Fig. 2). Thus, the 

 hypothesis that somatic growth rates of juveniles on the 

 harbor wall had not declined since Pollock's (1973) study 

 could be rejected at one level. Unfortunately, there were 

 no comparative data to provide accurate information on 

 corresponding intermolt periods. 



In terms of our samples from Mouille Point and the 

 harbor wall, significant, positive linear relationships were 

 found between molt increment and premolt carapace 

 length for both males and females at both sites (Fig. 2, 



