512 



Fishery Bulletin 100(3) 



Point, divers swam around the reef area, covering different 

 habitat types (ledges, vertical walls, cracks, etc.), and 

 attempted to catch every lobster encountered. 



The catch was maintained in a bin of seawater until the 

 sex, shell state (sensu Pollock, 1973) and carapace length 

 (CD of each animal could be recorded. Hard old and soft old 

 shell states were considered to represent premolt animals, 

 soft new and hard new to represent postmolts, and those 

 animals in the hard shell state were considered to be in the 

 intermolt phase of the molt cycle. Postmolt and intermolt 

 specimens were released, whereas premolt animals were 

 measured to the nearest 0.1 mm and transported to the 

 laboratory. Whenever relatively few premolt specimens of 

 a particular size-category were captured during routine 

 searches, additional dive time was spent collecting these. 

 In the laboratory, captive specimens were transferred to 

 perforated plastic jars (either 250 or 500 cm-'' in volume, 

 with square perforations of approximately 25 mm-), which 

 were placed in a 2-m'' holding tank. To mimic natural 

 conditions (Table 1), aquarium water temperature was 

 controlled between 11° and 14°C during holding, with a 

 mean of 12.6°C (SD=0.48°C), and salinity was maintained 

 at 35-36%f by periodic addition of freshwater or partial 

 replacement of seawater. Stocking density did not exceed 

 three specimens per jar, and larger specimens were held 

 singly or in pairs. By using a combination of jar number, 

 sex, premolt carapace length, and the pattern of missing 

 limbs, it was possible to identify each individual. Because 

 J. lalandii are unable to feed shortly before molting 



(Zou-tendyk, 1988), it was assumed that no feeding was 

 required. Jars were checked for molting individuals at 

 least every second day. Cast exoskeletons were examined 

 to identify those molting, and each new carapace length, 

 as well as any limb regeneration, was recorded once the 

 shell had hardened. Increment data recorded for December 

 1996 were discarded because of aquarium failure: raised 

 temperatures and lower oxygen levels resulted in high 

 mortality during molting. Because increments from spe- 

 cimens maintained in the laboratory for more than 18 

 days were considered unreliable, only animals that molted 

 within this period were included in analyses. 



Tests of the hypothesis that somatic growth rates of 

 juveniles on the harbor wall had not declined since 

 Pollock's (1973) study were complicated by the absence 

 of his original data. All that remain are mean molt in- 

 crements for male and for female specimens categorized 

 into 5-mm-CL size classes, their standard deviations, 

 and their sample sizes. Calculation of combined mean 

 increments for males and females of a given size is simple, 

 but their standard deviations had to be approximated 

 by adding corresponding sums of squares for males and 

 females and dividing this value by the overall degrees of 

 freedom within the size class (Table 2). It had been our 

 intention to compare these reconstructed data with those 

 from our study by using a two-way AN OVA (site x size 

 class), but this was not possible, given the limitations of 

 Pollocks ( 1973) data. Instead, the data were subjected to a 

 one-way ANOVA and subsequent Student-Newman-Keuls 



