820 



Fishery Bulletin 97(4), 1999 



A more complex aging error model, with para- 

 meters that accounted for effects due to aging struc- 

 tures showed that fish in their first year (i.e. growth 

 zones) were likely to be overestimated with scales 

 and vertebrae than with otoliths. Otoliths underes- 

 timated the age of fish in their first year, although 

 this is due to the relative nature of the model be- 

 cause vertebrae and scales overestimated the age of 

 these fish (Fig. 6B). Scales showed less bias in rela- 

 tion to the other structures for fish aged 1^ (Fig. 6B). 



Estimates of rates of growth 



Patterns in growth of kingfish from NSW were ob- 

 tained from Schnute's (1981) growth model fitted to 

 the estimated most probable ages of fish calculated 



from the different aging structures (Fig. 7). Size-age 

 data were best fitted by three-parameter models for 

 otoliths, vertebrae (both case 2, see Table 1) and 

 scales (case 3). Using the best fitting model for each 

 structure, we estimated that the average lengths (± 

 standard error) offish with one growth zone were 499 

 (±5), 418 (±9), and 485 (+7) mm FL for otoliths, scales, 

 and vertebrae respectively. At age 5, fish were 807 (±8 ), 

 823 (±8), and 788 (±6) mm FL for otoliths, scales, and 

 vertebrae respectively. The maximum age offish was 9 

 yr for otoliths and scales, compared to 1 1 yr for verte- 

 brae (Fig. 7). Vertebrae tended to produce higher age 

 estimates than did otoliths and scales, with 5% offish 

 showing greater than five growth zones in vertebrae 

 compared with only 27( for scales and otoliths. Using a 

 single age for each fish, we estimated an age-length 

 key for fish sampled during the current study 

 (Table 3). Fish in all length classes were found in 

 more than one age class (Table 3). 



The different aging structures showed signifi- 

 cant differences in mean size-at-age for some 

 age classes. Scales showed significantly differ- 

 ent mean size-at-age to otoliths and vertebrae 

 for fish with one growth zone (^-tests, P<0.05). 

 Otoliths and vertebrae gave similar estimates 

 of mean size-at-age for fish with less than eight 

 growth zones (<-tests, P<0.05). With the excep- 

 tion of fish with one growth zone, scales gave 

 similar estimates of mean size-at-age to otoliths 

 for all other age classes (t-tests, P<0.05). Al- 

 though differences between scales and vertebrae 

 were statistically significant for some age classes 

 (fish with 4—6 growth zones), these differences 

 were not likely to be biologically important. 



Kingfish that were measured at both tagging 

 and recapture were at large for between days 

 (i.e. recaptured the same day that they were 

 tagged) and 5 years (Fig. 3B). Growth of recap- 

 tured fish ranged from a decrease of 350 mm to 

 an increase of 800 mm (Fig. 3A). The frequency 

 distribution of fish at large <30 days showed 

 that fishermen were just as likely to underesti- 

 mate the length offish as they were to overes- 

 timate the length of fish (Fig. 3A). The mean 

 difference in size of fish between tagging and 

 recapture for fish recaptured within 30 days was 

 11.6 mm (±0.4, SE), suggesting that there was 

 little bias in measurements between tagging and 

 recapture. Size offish at tagging ranged from 220 

 to 1200 mm, although the majority of fish were 

 between 400 and 600 mm TL (Fig. 3C). 



The best fit to the complete tag-recapture 

 data set (model 1 in Table 4) showed a high pro- 

 portion of outliers (P=0.04). Twenty-two fish (or 

 2.7^^) that had absolute standardized residu- 



