Welsford and Lyle: Estimates ol growth of Notolabrus fuacola from length- and age-based models 



707 



>! - 



• LB Males 



LB Females 





B 



04 



g (cm/yr) 



• LB Males 



LB Females 



-0 4 



m (cm) 



Figure 5 



GROTAG bootstrap parameter estimates for Notolabrus fucicnla from Lord's Bluff, by sex: (A)g 20 versus g 28 , mean annual 

 growth at initial length 20 and 28 cm; (Bl u versus w, magnitude and timing of seasonal growth and (C) m versus s, 

 mean and standard deviation of measurement error. Contrasting crosses show the location of parameter estimates based 

 on the original data set (+ = males, x = females). 



nor LRTs indicated significant difference in any of the 

 model parameters, and bootstrap plots showed large 

 regions of overlap (not shown). 



Discussion 



Comparisons of models 



In this study, two methods, based on mathematically 

 different concepts, produced similar conclusions, namely 

 that growth in N. fucicola was faster at Lord's Bluff than 



at Point Bailey. The results of length-based and age- 

 based models also produced similar conclusions regard- 

 ing the methods most suitable for robust comparisons of 

 models and parameter estimates for different groups of 

 fish. Confidence intervals were only reliable indicators of 

 difference in cases where parameters showed low levels 

 of correlation between estimates and where highly sig- 

 nificant differences existed, such as in site comparisons 

 of the reparameterized VBGF parameters, and hence 

 were of limited utility. 



Likelihood ratio tests provided a robust method of 

 testing differences between models. However, we believe 



