524 



Fishery Bulletin 88(3), 1990 



and preserved in Gilson's fluid (Simpson 1951). and 

 fecundity determined using a volumetric method (Bage- 

 nal and Braum 1971). Condition factor (CF) was deter- 

 mined from the formula CF = lOOW/TL^* (Hile 1936). 



Age was determined using otoliths (sagittae) from 

 a total of 252 fish. Longitudinal sections of otoliths 

 were prepared following the technique described by 

 Holden and Raitt (1974). Sections were wet with water 

 and viewed against a black background using reflected 

 light. Distinct alternating broad opaque and narrow 

 translucent rings were present and the latter were 

 assumed to be annuli. Rings were determined to be 

 annual by using the marginal increment technique (e.g., 

 Moe 1969, Turner et al. 1983). Otoliths were reread by 

 the same worker one month after the initial reading 

 and a 90% agreement was found, while a 92% agree- 

 ment was found with another worker's readings. Mar- 

 ginal increment and otolith radius were measured using 

 an ocular micrometer (1 micrometer unit = 0.05 mm). 



Attempts were also made to determine age by ana- 

 lyzing the length-frequency distribution using cumula- 

 tive probability paper. Total length was plotted against 

 percentage cumulative frequency, resulting in a Cassie 

 curve (Cassie 1954). Age groups were separated using 

 points of inflexion which were chosen by eye. Only fish 

 caught during March-May 1978 were used in this 

 analysis. 



The von Bertalanffy (1938) model was used to 

 describe growth: 



L, = L^{1 - exp[-;^(f-f„)]} 



where L, is length at time t, K is the growth coeffi- 

 cient, L^ is the asymptotic length, and /„ is the hypo- 

 thetical age at zero length. These parameters were 

 estimated utilizing the Length Based Fish Stock As- 

 sessment (LFSA) Package of BASIC computer pro- 

 grams (Sparre 1987). 



Total annual mortality rate (Z) was estimated from 

 a length-converted catch curve (Pauly 1984). Natural 

 mortality rate (M) was determined from the empirical 

 relationship derived by Pauly (1980): 



which was determined by mesh selectivity experiments 

 (Sylvester 1986), and ^. (age at recruitment). Exam- 

 ination of length-frequency distributions of catches 

 indicated that croakers were fully recruited at a total 

 length of 25 cm, or 1 year of age. The average values 

 of K and t^^ for both sexes were used. Y/R analysis was 

 done for a range of Af values to determine their effects 

 on conclusions drawn. 



Results 



Size at maturity 



The percentage of mature males and females in each 

 3-cm size group is shown in Table 1 . The size at which 

 50% of the fish examined were mature was about 28 

 cm for males and 32 cm for females. 



log M = 0.0066 - 0.279 log L^ + 0.6543 log A' 



-I- 0.4634 T, 



where T is the mean water temperature which was 

 26.2°C in this area. Fishing mortality (F) was calcu- 

 lated as the difference between Z and M (Ricker 1975). 

 Yield per recruit (Y/R) in grams was determined 

 from the model of Beverton and Holt (1957). Analyses 

 were carried out using the LFSA package. The input 

 parameters were W^ (determined from the formula 

 W = a L'4^), K, /„, M, t,. (age at first capture of 2 years). 



Spawning pattern 



Examination of the monthly frequency of croakers in 

 each developmental stage showed the presence of all 

 stages throughout the year (Table 2). A higher percent- 

 age of ripe fish was found from February to August, 

 indicating an intensification of spawning during this 

 time. 



Sex ratio 



The overall sex ratio of 852 croakers was 1:1 .3, male: 

 female. This was significantly different from 1:1 (P 



