608 



Fishery Bulletin 98(3) 



100% 



(U 



0) 

 Q. 



75%- 



S 50%, ^ 



0) 



25% 



0% 



Jan Feb Mar Apr May 



Figure 8 



Proportion of L. mahsena otoliths with stained edges on the gi'owing tip of 

 the dorsal lobe, by month. Numbers denote sample size. 



Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 

 Month 



Figure 9 



Marginal increment plot by month for A. virescens. Mar- 

 ginal increment measured in VID units (pixels). Data 

 segregated as in Table 1. Error bars denote one stand- 

 ard error 



increments on scales in some species (Chey, 1933, in Patt- 

 ella, 1980). If this species is sensitive to relatively small 

 changes in temperature, currents, food supply or food qual- 

 ity, relatively short periods of fluctuation in these factors 

 may result in frequent "indeterminate increments." Even 

 if validation had been achieved. P. filamentosiis otoliths 

 appear unsuitable for routine aging procedures owing to 

 the difficulties in differentiating periodic increments from 

 other increments not related to a regular time scale. 



These difficulties highlight the dangers in assuming the 

 periodicity of increments in otoliths on the basis of valida- 

 tion of that species or a similar species in another loca- 

 tion (e.g. Smith and Kostlan, 1991) and also identify the 

 need for thorough validation studies. For example, our 

 study indicated that the assumption of annual, and pos- 

 sibly daily periodicity for increments in the otoliths for P. 



filamentosi/s in Seychelles, based on published data, would 

 have been incorrect. 



Published work (e.g. Sparre and Venema, 1998) suggests 

 that consistent periodic increments do not form in oto- 

 liths of tropical species; this is particularly so in the case 

 of low-latitude species. The validation results obtained 

 by our study contrast sharply with this view; indeed, our 

 results add to the expanding number of tropical species 

 for which the validation of annual increments has been 

 successful. 



The results of models used in fisheries management, e.g. 

 analytical yield per recruit models (Beverton and Holt, 

 1957), are sensitive to uncertainty in the estimates of 

 input parameters such as the von Bertalanffy growth para- 

 meters (Mees and Rousseau, 1997). Current estimates of 

 growth and mortality in fish in Seychelles and Mauritius 



