Xu and Mohammed An alternative approach to estimating growth parameters 



153 



suggested that ELEFAN I underestimates the true 

 value of K, and Rosenberg and Beddington (1987) 

 concluded that ELEFAN I estimated A' reliably only 

 if the true value of K was known within about 25%. 

 ELEFAN I also tends to have multiple maxima of 

 the score function (Rosenberg and Beddington, 1987; 

 Isaac, 1990), as found for the SLCA method (Basson 

 et al., 1988; Isaac, 1990), which makes it difficult to 

 select the best set of growth parameters from the 

 multiple solutions. The proposed method has the 

 advantage over ELEFAN I in that it applies a non- 

 linear least-squares technique with the Gaussian it- 

 erative method (SAS, 1992) for solution searching, 

 which is more sensitive in defining the best set of 

 growth parameters and which estimates variances 

 and covariances for the growth parameters so that 

 the growth between cohort and sexes can be com- 

 pared statistically. 



Castro and Erzini ( 1988) studied the effect of dif- 

 ferent recruitment patterns on ELEFAN I (Pauly and 

 David, 1980) and modal progression analysis imple- 

 mented in LFSA (Sparre, 1987b) with simulated 

 length-frequency data. The results were generally 

 encouraging, and ELEFAN I produced better esti- 

 mates than did modal progression analysis for the 

 case of multiple recruitment per year. However, for 

 both methods, multiple recruitment makes it diffi- 

 cult to estimate growth parameters (Castro and 

 Erzini, 1988). The recruitment pattern of green ti- 

 ger prawns in Kuwait waters is relatively simple and 

 the degree of overlap between cohorts is low. It would 

 be worthwhile to test the performance of the proposed 

 method for species with more complicated recruit- 



ment patterns and high overlapping cohorts with real 

 or simulated data. 



Implication of the variations in growth 



The highly significant differences in growth between 

 sexes of green tiger prawns is not surprising because 

 female shrimp grow to be much larger than males. 

 Female and male green tiger prawns in Kuwait wa- 

 ters are distributed in the same area and experience 

 the same environmental changes, but the annual 

 variations in growth curves between cohorts are more 

 obvious for males than for females (Table 6). These 

 variations suggest that the responses of the female 

 and male shrimp to environmental factors might be 

 different, and that the growth of males may be more 

 vulnerable to environmental changes. A possible rea- 

 son that females are less sensitive to environmental 

 change is that they have more energy reserves that 

 can be utilized to maintain homeostasis by shifting 

 from anabolism to catabolism (Pickering, 1981). Fe- 

 male shrimp grow faster and become larger than 

 males, have lower natural mortality, but with a 

 higher market demand suffer higher fishing mortal- 

 ity (Xu et al., 1995) compared with males. All these 

 differences between female and male shrimp should 

 be taken into consideration when monitoring the 

 shrimp fisheries, estimating the population abundance 

 and biomass, and formulating management policy. 



Acknowledgment 



This research was part of the Shrimp Fisheries Man- 

 agement Project sponsored by Kuwait Institute for 

 Scientific Research, the United Fisheries of Kuwait, 

 and the Public Authority for Agriculture and Fisher- 

 ies. The contributions of the Shrimp Fisheries Man- 

 agement Project staff to the continuous four-year 

 research-vessel surveys formed the basis for this re- 

 search. G. R. Morgan, J. M. Bishop, and two anony- 

 mous reviewers provided very helpful comments on 

 the manuscript. 



Literature cited 



Abramson, N. J. 



1971. Computer programs for fish stock assessment. FAO 

 Fish. Tech. Rep. 101, 96 p. 

 Basson, M., A. A. Rosenberg, and J. R. Beddington. 



1988. The accuracy and reliability of two new methods for 

 estimating growth parameters from length-frequency 

 data. J. CIEM 44:277-285. 

 Bernard, D. R. 



1981. Multivariate analysis as a means of comparing 

 growth in fish. Can. J. Fish. Aquat. Sci. 38:233-236. 



