84 



Fishery Bulletin 104(1) 



at the theoretical unexploited level. An increase in the 

 mean size at first capture to that which would maxi- 

 mize yield per recruit was predicted to be associated 

 with a small increase in biomass per recruit. Changes 

 in the exploitation rate were predicted to have a greater 

 impact on biomass per recruit, which was estimated to 

 be above SO^i: of the unexploited levels at the optimum 

 exploitation rate iE^^ ^) (Table 7). Estimates of precau- 

 tionary target and limit exploitation rates (£,,„, and 

 ^iimit' f°'" ^- nebulosus were 0.33 and 0.43, respectively. 

 For both D. Pictum and L. nebulosus, the exploitation 

 rates predicted from the yield per recruit function (Eg j 

 and Efnax' were equal to or in excess of values where 

 F=M (0.5) for all sizes at first capture. 



The range of fishing mortality rates estimated for D. 

 pictum (0.37-0.62/yr) was substantially greater than 

 both the target (F„p, = 0.07/yr) and limit (F|,„„j=0.09/yr) 

 biological reference points. The range of fishing mortal- 

 ity rates for L. nebulosus (0.15/yr to 0.57/yr) were also 

 in excess of the biological reference points for this spe- 

 cies (F„pj=0.10/yr and F,,„„,=0.13/yr). 



Discussion 



Age and growth 



The formation of alternating translucent and opaque 

 growth zones in fish otoliths has been associated with a 

 variety of factors, including seasonal variations in water 

 temperature, photoperiod, feeding, and reproduction 

 (Moe, 1969; Reay, 1972; Panella, 1980; Manickchand- 

 Heileman and Phillip, 2000). Although the mechanisms 

 of growth-increment formation are poorly understood, 

 the deposition of the opaque zone in tropical species 

 generally occurs in the spring and summer months 

 during periods of accelerated growth, whereas the trans- 

 lucent zone is formed when there is reduced metabolic 

 activity (Beckman and Wilson, 1995). The formation of 

 opaque and translucent zones in the sagittal otoliths 

 of D. pictum and L. nebulosus determined in our study 

 follows this generalized pattern. 



The southern Arabian Gulf exhibits marked sea- 

 sonal variability in oceanographic characteristics; sea 

 water temperatures can exceed 34°C in summer and 

 fall to 2rC in the winter (Sheppard et al,, 1992). The 

 close association of the formation of opaque zones with 

 increasing seawater temperature indicates that tem- 

 perature could be the principal environmental signal 

 stimulating the deposition of these growth zones. Other 

 allied variables, such as productivity and subsequent 

 food availability, may also be associated with seasonal 

 growth-rate oscillations and the formation of growth 

 increments. The validation of the annual periodicity 

 of increment formation adds to a growing body of evi- 

 dence (Fowler, 1995) and dismisses the misconception 

 that annuli do not form in the otoliths of reef fish due 

 to a lack of seasonality in the tropics (e.g., Sparre and 

 Venema, 1992). Nevertheless, the edge analysis method 

 used should ideally be conducted over a 2-year cycle and 



could have been more rigorous with the use of larger 

 sample sizes and by conducting the analyses for indi- 

 vidual ages. Furthermore, it is important to distinguish 

 between the validation of increment periodicity and 

 absolute age (Campana, 2001). Although our study has 

 provided empirical evidence for an annual pattern of 

 increment formation, the absolute age of the study spe- 

 cies remains to be validated. Validation of the absolute 

 age could be achieved through independent means such 

 as a mark-recapture study and the chemical marking of 

 juvenile fish of known age. 



The maximum number of opaque bands counted for 

 D. pictum in the present study (13) was considerably 

 less than the maximum age of 31 years estimated by 

 Loubens (1980) for this species in New Caledonia. Like- 

 wise, our longevity estimate of 14 years for L. nebulosus 

 was less than that of Mathews and Samuel (1991) (20 

 years) and of Edwards and Shaher (1991) (21 years) for 

 this species in the northern Arabian Gulf and Gulf of 

 Aden, respectively. Our longevity estimates therefore 

 were most likely to have been underestimated owing 

 to the absence of fish close to the maximum reported 

 sizes for these species. 



A method of validating growth parameters involves 

 the comparison of growth performance indices (<f>') in 

 terms of growth in length with other estimates obtained 

 for the same or a similar species (Gayanilo and Pauly, 

 1997). Values of <P for D. pictum available from other 

 studies have ranged from 2.88 (Loubens, 1980) to 3.24 

 (Baillon and Kulbicki, 1988), and an estimate of 3.07 

 has been obtained for this species in the Gulf of Aden 

 (Edwards et al., 1985). The estimate obtained in our 

 study (2.81) compares with the lower end of this range. 

 Values of <& for L. nebulosus have ranged from 2.55 

 (Kuo and Lee, 1986) to 3.41 (Baillon, 1991), and an es- 

 timate of 2.87 obtained for this species in the northern 

 Arabian Gulf in the waters off Kuwait (Baddar, 1987) 

 compares well with our estimate of 2.80. Although the 

 growth parameters in our study would appear to be of 

 the right order (by comparison), improvements in our es- 

 timates could have been made by the addition of larger 

 specimens close to the maximum size for both species. 



Despite its widespread use, the von Bertalanffy 

 growth function may not be suitable for hermaphro- 

 ditic populations (Appeldoorn, 1996). Growth analyses 

 have shown distinct differences in the sizes of equal-age 

 males and females of protogynous species (Moe, 1969; 

 Nagelkerken, 1979; Garratt et al., 1993), and experi- 

 ments have shown what are considered to be growth 

 accelerations leading to sex change (Ross et al., 1983). 

 Failure to account for growth spurts in yield models 

 can result in significant over estimation of both maxi- 

 mum yield and optimal effort (Bannerot et al., 1987). 

 Although L. nebulosus is a protogynous hermaphro- 

 dite (Young and Martin, 1982; Ebisawa, 1990), the 

 results of the analysis of residual sums of squares in 

 our study indicated that there are no differences in the 

 growth characteristics between sexes and that the use 

 of growth parameters from pooled data would therefore 

 be justified in yield-per-recruit analyses. 



