822 



Fishery Bulletin 97(4), 1999 .^ 



Small sample sizes during the winter months (from 

 April to July), due in part to small numbers offish be- 

 ing taken in the fishery, made analysis of marginal in- 

 crements problematic, especially for scales because 

 samples were also not obtained in August-September. 

 With the exceptions of otoliths and possibly scales, our 

 data were not sufficient to confirm that only one zone 



otoliths 



fl2=0.67 

 SD=72 36 

 a=0 1322(0.0246) 

 y1=499(5) 

 y2=807 (8) 

 J I I I 



1.200 

 1.000 

 800 

 600 

 400 

 200 

 



8 9 10 11 12 



Scales 



H2=0.73 



SO=66 35 



6=2 408(0 203) 



y1=418 (9) 



y2=823 (8) 



I I I I I 



1 



8 9 10 11 12 



Vertebrae 



R2=0 70 



SD=64 50 



a=0 1844 (0 0243) 



y1=485(7) 



y2=788 (6) 



10 11 12 



Growth! curves 



4 5 6 7 8 9 

 Age (no, zones) 



10 11 12 



Figure 7 



Relationship between fork length and age for the different struc- 

 tures used to age kingfish (A-C) and (D) comparison of the 

 three growth curves. The growth curves were calculated by 

 using Schnute's growth model. /?-, overall standard deviation 

 (SD), and the parameters (standard error) describing the growth 

 model are shown. _V[ is size at age 1, y.^ is size at age .5, a and b 

 are parameters that describe the shape of the curve. The for- 

 mulae for the growth curves are found in Table 1. 



was formed per year. No single category of marginal 

 growth ever had more than 50% of fish for any struc- 

 ture. A previous study (Mitani and Sato, 1959) also 

 found that fish (Seriola quinqueradiata) collected in 

 any one time period showed a wide range of marginal 

 growth conditions. Growth zones on the edge of the 

 otoliths of S. dumerili were also found over five months 

 of the year (Manooch and Potts, 1997). The wide 

 range of marginal growth patterns in our study may 

 in part be due to grouping of samples into 

 two-month intervals or because fish from three age 

 classes (2—4 growth zones) were used in analyses. 

 Fish from different age classes have previously been 

 shown to lay down zones at slightly different times 

 of year (e.g. Francis et al., 1992). 



Analysis of marginal increments often provides 

 only partial validation of age estimates because 

 older fish with slower growth often do not show 

 seasonality in formation of zones. A method 

 should not be considered accurate until all re- 

 ported ages are validated (Beamish and McFar- 

 lane, 1983). Validation of age estimates in older 

 fish will require a mark-recapture study, but fur- 

 ther work on validation in younger fish may use 

 a variety of approaches (e.g. length-frequency 

 analyses if cohorts are easily recognized, mark- 

 recapture etc). In the present study, few tagged 

 fish were at large more than a year (Fig. 3B), 

 which suggests that many fish would have to be 

 tagged in a mark-recapture study to recover suf- 

 ficient fish to make this approach feasible. 



There have been few validation studies on Seriola 

 spp. and all have involved analysis of marginal in- 

 crements. Mitani and Sato (1959) have suggested 

 that one zone is laid down each winter in opercular 

 bones ofS. quinqueradiata and Baxter ( 1960) foimd 

 that zones are formed between November and 

 January in scales of S. lalandi (formerly dorsal is). 

 In otoliths of S. lalandi opaque zones appear to be 

 laid down in August or September Timing of zone 

 formation in scales and vertebrae was more vari- 

 able and may differ between structures because pro- 

 cesses involved in deposition vary among bone, 

 scales, and otoliths (Simkiss, 1974). 



Precision of aging estimates 



Exact agreement among age estimates for each 

 structure was generally poor; agreement within 

 one zone was reasonable for all structures. The 

 use of percent agreement (i.e. percent offish aged 

 alike between sets of multiple readings) has been 

 criticized because it fails to take into account the 

 range of year classes offish and therefore can be 

 used only for age-specific comparisons (PCimura 



