604 



Fishery Bulletin 98(3) 



dorsal lobe. The percentage of otoliths with a stained band 

 at the growing tip was calculated for each month. Although 

 an extended data set was available for edge analysis, as a 

 result of the seasonal nature of the Seychelles and Mauri- 

 tius fisheries, sample sizes were either small or not avail- 

 able for certain months. 



Results 



Based on the assumption that the increments seen in the 

 otoliths were annual, the length and age range of individ- 

 uals from each species examined in each method is pre- 

 sented in Table 2. The sample size is also shown. 



A greater number of samples were examined during 

 edge analysis. This method was less intensive than that 

 required for either back-calculation or marginal increment 

 analysis, allowing more samples to be examined during 



the period of study and increased the range of lengths and 

 ages examined. 



Back-calculation 



A linear relationship existed between the fork length and 

 otolith radius offish in all three species, to which the geo- 

 metric mean regression (GMR) of fork length from otolith 

 radius was fitted (Figs. 4—6). 



The von Bertalanffy growth parameters derived from 

 mean back-calculated length-at-age data from A. virescens, 

 P. filamentosus and L. mahsena are shown in Table 3. Initial 

 L^ estimates for L. mahsena were low compared with the 

 maximum lengths seen in length-frequency distributions 

 from the region (52 cm). The asymptotic length, L^^ was 

 therefore constrained to levels in keeping with the maxi- 

 mum length found in the distributions, and the remaining 

 von Bertalanffy growth parameters were re-estimated. 



For the snapper species, both L^ and K estimates derived 

 from the back-calculated leng^h-at-age data were lower 

 than those historical estimates derived with length-based 

 methods. ForL. mahsena, constra.\mr\gL,^ to a level in keep- 

 ing with the maximum length found in catches resulted 

 in estimates of X' comparable to those estimated by Bautil 

 and Samboo (1988). 



Marginal increment and edge analysis 



The number of individuals examined in marginal incre- 

 ment analysis are presented in Table 4 by species and time 

 period. 



Because the region between translucent and opaque 

 zones took up the stain, the presence of a stained area at 

 the otolith edge during edge analysis indicated the start of 

 opaque band formation. 



