NOTE Laurenson et al : Growth curves of two estuanne populations of Cnidoglanis macrocephaius 



887 



asterisci in C. macrocephaius from the Swan Estu- 

 ary. The importance of confirming that each succes- 

 sive translucent zone is formed annually is demon- 

 strated by the results obtained by Hyndes et al. 

 (1992) for whole sagittae of Platycephat 'us specula- 

 tor in Wilson Inlet. In that species, mean monthly 

 marginal increments showed a very clear seasonal 

 trend when individual marginal increments on all 

 unsectioned otoliths were pooled, irrespective of the 

 number of translucent zones. However, they did not 

 show conspicuous trends when the data for the mar- 

 ginal increments on unsectioned otoliths with two, 

 three, four, and five or more translucent zones were 

 each plotted separately. In other words, when mar- 

 ginal increment data for all otoliths were pooled, the 

 pronounced seasonality exhibited by the mean mar- 

 ginal increments on otoliths with one translucent 

 zone of P. speculator had an overwhelming influence 

 on the data set. 



The von Bertalanffy growth curve did not suffi- 

 ciently describe the growth of C. macrocephaius from 

 Wilson Inlet; the lengths were consistently greater 

 than the mean length at ages 7 and above and showed 

 increasing divergence with age. This was far less of 

 a problem in Swan Estuary where older fish were 

 less abundant. The generalized growth curve pro- 

 vided better fits to the data than the von Bertanlanffy 

 curve for males and females in both systems, when 

 both lengths at age and back-calculated lengths were 

 used. Furthermore, likelihood ratio tests showed that 

 this improvement was significant in three of the four 

 cases for the population in Wilson Inlet. Such im- 

 provement is consistent with the observation that 

 when there is an acceleration of growth early in life, 

 the von Bertalanffy growth curve does not provide 

 as adequate a fit as the Schnute, Gompertz, or 

 Richard's curves (Schnute, 1981; Campana and 

 Jones, 1992). 



While the presence among younger fish of smaller 

 back-calculated lengths than mean lengths at age 

 (Table 1 ) would be consistent with Lee's phenomenon 

 (Ricker, 1975), it could also have been brought about 

 by the low numbers of younger fish in the samples. 

 The fits of the common curves constructed for each 

 sex in Wilson Inlet from lengths at age and back- 

 calculated lengths were improved by only 0.2% for 

 females and 0.8% for males when separate curves 

 were used. However, this was not the case for fish in 

 the Swan Estuary, where the sum of squares was 

 improved by 1.6% for females and by as much as 4.2% 

 for the males. The differences in improvement in fit 

 in the two systems probably reflects the fact that, 

 while the 0+ age class in the Swan Estuary was 

 caught in greater numbers, it tended to be represented 

 in samples by the larger members of this age class. 



The improvement of fit obtained by using separate 

 growth curves was small, both in comparisons be- 

 tween males and females in Wilson Inlet and Swan 

 Estuary and in comparisons between corresponding 

 sexes in the two systems. This applied to curves con- 

 structed both from lengths at age and back-calcu- 

 lated lengths. In none of these cases was the sum of 

 squares improved by more than 0.6%. However, al- 

 though the differences between the curves for each 

 sex in each system and for the corresponding sexes 

 in the two populations were small, and even though 

 the lengths varied considerably at a given age, the 

 curves were still statistically different with a likeli- 

 hood ratio test (usually P<0.001). These differences 

 probably reflect the influence of the large number of 

 data points used to construct the growth curves. 



The small magnitude of the differences between 

 these growth curves is demonstrated by the fact that 

 at age 4, the lengths of males and females in Wilson 

 Inlet and the Swan Estuary, predicted from the gen- 

 eralized growth curve, generally differed by less than 

 3%, irrespective of whether the curve was constructed 

 from lengths at age or back-calculated lengths. Thus, 

 although there were usually highly statistically sig- 

 nificant differences between curves, the actual dif- 

 ferences between the curves for the two sexes in each 

 population and between the corresponding sexes in 

 those populations are almost certainly of limited bio- 

 logical significance. 



In conclusion, the growth of C. macrocephaius in 

 Wilson Inlet was similar to that in the Swan Estu- 

 ary. This similarity occurred despite the fact that 

 water temperatures in the latter system were over 

 5°C higher in the summer (c.f. Loneragan and Pot- 

 ter, 1990; Potter et al., 1993). Wilson Inlet is eutrophic 

 and therefore more productive (Lukatelich et al., 

 1987) and consequently contains a greater abundance 

 of the large deposit-feeding benthic invertebrates 1 

 that make a major contribution to the diet of C. 

 macrocephaius (Nel et al., 1985; Laurenson, 1992). 

 Therefore the similarity between the growth rate of 

 C. macrocephaius in Wilson Inlet and the Swan Es- 

 tuary may reflect a compensation for lower water 

 temperatures by greater prey abundance. 



Acknowledgments 



Our thanks are expressed to F. Baronie, D. Gaughan, 

 P. Geijsel, P. Humphries, G. Hyndes, and F. Neira 

 for assistance with sampling. Gratitude is also ex- 



1 Platoll, M. School of Biological and Environmental Sciences, 

 Murdoch Univ., Murdoch, Western Australia 6150, Australia. 

 Personal commun., 1991. 



