FISHERY BULLETIN: VOL. 87. NO. 1 



edge of the otolith, through the longest axis. Three 

 counts were made of each sagitta, and the mean in- 

 crement count from a pair of sagittae was used in 

 the analysis. Otoliths were rejected if incremental 

 counts within or between pairs of sagittae differed 

 by more than two. 



Validation of Ageing Technique 



To determine if the increments observed in the 

 otoliths of H. castelnaui were deposited daily, lar- 

 vae collected from the marina were treated with 

 tetracycline. Tetracycline is an antibiotic that is in- 

 corporated into calcium structures of fish during 

 growth. This can be restricted to a single day's in- 

 crement on the otolith (Tsukamoto 1985), and thus, 

 the date of treatment can be accurately identified. 

 This technique has become widely used in the vali- 

 dation of ageing techniques (e.g., Campana and 

 Neilson 1982; Schmitt 1984; Kingsford and Milicich 

 1987). 



Larvae were collected from Breakwater Marina 

 on 14th of October 1987 and were transported to 

 the laboratory at the Australian Institute of Marine 

 Science (AIMS). The fish were kept in ambient 

 photoperiod and temperature regimes for two days 

 to allow time for acclimation. They were fed twice 

 daily on wild zooplankton captured with a 15 fim 

 mesh plankton net from Chunda Bay, adjacent to 

 AIMS. 



Ten fish were kept overnight in a 4 L tank treated 

 with a 0.25 g/L tetracycline hydrochloride solution 

 (Schmitt 1984). Four larvae died during exposure 

 to the tetracycline. The remaining six larvae were 

 returned to a 120 L tank and fed as before for 10 

 nights and 11 days before being sacrificed. Sagit- 

 tae were dissected out of the remaining larvae and 

 viewed under fluorescent UV and natural light with 

 a compound microscope. Under fluorescent light an 

 ocular marker was aligned with the fluorescent band 

 in the otolith. The otolith was then examined under 

 natural light, and the number of increments between 

 the marker and the otolith margin counted. Both 

 sagittae for each fish were analyzed, and three 

 counts were made of each otolith. 



Statistical Procedures 



Laird-Gompertz and von Bertalanffy growth 

 models were fitted to the length-at-age data. Both 

 models have been shown to provide adequate fits 

 to length-age data of -i- fish in different situations 

 (e.g., Ralston 1976; Laroche et al. 1982). Zweifel and 

 Lasker (1976) presented a detailed discussion of the 



Laird-Gompertz function. The generalized equation 

 of the model is 



L, = Lo exp[Aola,{l - e-")] 



where L, = length (mm) at age t; Lq = length at 

 i = 0; 4q = specific growth rate at < = 0; and a = 

 rate of exponential decay. Gallucci and Quinn's 

 (1979) version of the von Bertalanffy equation was 

 used, where the generalized equation of the model is 



L, = oj/A- {1 - exp[-fc(< - f,,)]) 



where A: = growth constant; L„ = maximum lar- 

 val size obtained; oj = kL^; and ^q = x-axis 

 intercept. 



The BMDP P3R'' nonlinear least-squares regres- 

 sion program employing a modified Gauss-Newton 

 algorithm was used to fit both models. A measure 

 of goodness-of-fit was provided by calculating an r^ 

 value from residual and explained sums of squares 

 derived from the least-squares regression. Good- 

 ness-of-fit can also be assessed by examination of 

 standard errors and approximate 95% confidence 

 intervals of parameter estimates. 



Spawning frequency of H. castelnaui during the 

 sampling period was estimated by ageing larvae and 

 then back-calculating birthdates from the time of 

 capture. Periodicity in spawning was analyzed using 

 the SYSTAT SERIES^ program, employing an 

 autoregressive moving average (ARIMA) model 

 (Box and Jenkins 1976). Autocorrelation of each 

 value in a series with every other value will define 

 relationships between all points in the series. A plot 

 of partial autocorrelations will detect dependencies 

 in the data, and identify the period of any depen- 

 dency. 



RESULTS 



Otolith Morphology 



Growth increments were clearly visible in sagit- 

 tae of larval H. castelnaui. No marked changes in 

 increment morphology was evident, although in 

 some otoliths a narrowing and subsequent widen- 

 ing of increments occurred between increments 15 

 and 25 (Fig. 2). Counts of growth increments were 

 obtained from 378 larvae ranging from 5.6 to 22.5 



'BMDP 3R. 1983. BMDP statistical software. Univ. Calif. 

 Press. Berkeley, CA 94720. 



'SYSTAT SERIES. 1986. SYSTAT: The system for statis- 

 tics. Evanston, IL 50201. 



76 



