240 



Fishery Bulletin 97(2), 1999 



O 



60 



50 - 



40 



30  



E 

 o_ 



c 

 « 



E 



0) 



o 



.E 20 



% 



2 10 



-10 



A Poor Knights (n= 16) 

  Cook Strait (n= 57) 

 o Oamaru (n = 65) 





1,000 



2,000 

 Days at liberty 



3,000 



4,000 



Figure 12 



Relationship between growth increment and period at liberty for tagged 

 hapuku. n = sample size. 



Table S 



Comparison of hapuku annual growth increments at two 

 reference lengths (5.5 cm and 85 cmi based on length-at- 

 age (R^., ), length-frequency, and tag-recapture data. 



Annual growth increment (cm) 



Total 



length Tag- 



(cmi Length-at-age Length-frequency recapture 



The best GROTAG model fit e.stimated the mean 

 annual growth rate to be 4.25 cm/yrover the length 

 range 55-85 cm. That estimate falls between the 

 length-at-age estimates of annual growth for 55-cm 

 and 85-cm hapuku (Table 5). A nonlinear growth 

 model was not significantly better than the linear 

 model, probably because of the narrow length range 

 of the tag-recapture data, and the small sample size 

 (n = 138). 



We conclude that the growth curves and growth 

 rate estimates derived from length-frequency and 

 tag-recapture data are consistent with those from 

 otolith-based, length-at-age data for the length and 

 age ranges over which the data overlap. In conjunc- 

 tion with the results of the OTC injection experiment, 

 this agreement among growth-rate estimates pro- 

 vides strong support for our hypothesis that otolith 

 band pairs are deposited annually in hapuku. 



Our length-at-age growth curve for male hapuku 

 is similar to that of McDougall (1975), who made 

 band counts in broken and burnt otoliths from 

 hapuku collected in Cook Strait (Fig. 10). However, 

 McDougall reported a slightly faster growth rate for 

 females (Fig. 10). It is not possible to determine 

 whether the difference in female growth rates is real, 

 or a result of McDougall's small sainple size (n=72). 

 McDougall's largest hapuku was 120 cm and his old- 

 est fish was 26 years. 



A study of hapuku growth at Juan Fernandez Is- 

 land, based on scale annuli, reported much faster 

 growth rates and a maximum age of only 12 years 

 (Pavez and Oyarzun. 1985). Scale rings are typically 

 crowded to the point of being unresolvable in hapuku 

 older than 8 years (McDougall, 1975); it is likely 

 therefore that the ages of Pavez and Oyarziin's larger 

 fish were underestimated. 



Lack of precision and reader bias mean that otolith 

 ageing is unlikely to be useful when accurate hapuku 

 ages are required (e.g. in the estimation of year-class 

 strength from an age-frequency distribution ). How- 

 ever, the precision and accuracy problems experi- 

 enced in this study had little effect on the shape of 

 growth curves fitted to length-at-age data, and we 

 believe that the growth parameters reported here are 

 relatively robust. 



The large negative t^ values for all growth curves 

 indicate a lack of fit of the von Bertalanffy model to 

 the data for small hapuku. This may result from one 

 or more of the following: 1 ) lack offish less than three 



