Gillanders et aL; Aging methods for Seriola lalandi 



819 



ber and January (scales; Table 2). No data were, how- 

 ever, obtained for scales in August-September and 

 sample sizes were small between April and July (n = \- 

 7 fish per month) for all structures. No clear pattern of 

 marginal zones was observed in vertebrae (Table 2). 



Precision within and among structures 



Comparisons of two independent counts of zones in 

 a structure resulted in a relatively low level of agree- 

 ment (50-66'?^^ ). Depending on the structure, between 

 92% and 96% of readings agreed within one zone 

 (Fig. 4). Differences in counts of zones varied by up to 

 four zones (Fig. 4). Mean coefficients of variation among 

 counts ranged from 7.6% (scales) to 12% (otoliths). 



Comparison of readings between structures 

 showed a large amount of variation, with differences 

 between structures varying by up to six growth zones 

 (Fig. 5). Agreement between any two methods de- 

 creased with age, but otoliths and vertebrae had the 

 greatest concordance in fish aged 4 and over. Fish 

 were never assigned an age of when aged with 

 scales; fish assigned or 1 with other structures were 

 assigned an age of 1 with scales. Readings between 

 structures agreed within one zone between 88% (be- 



300 



200 



100 



400 



i 300 



o 



I 200 



E 



z 100 





 300 

 200- 



-4-3-2-101234 

 Difference (number of zones) 



Figure 4 



Differences between repeated counts of zones for three 

 structures used to age kingfish. Each comparison repre- 

 sents independent counts from a single reader 



tween vertebrae and otoliths) and 91% (between 

 scales and otoliths) of the time. 



Estimates of aging error 



Analysis of aging precision with the methods of 

 Richards et al. (1992) showed that the estimates of 

 assigned age were more precise for young fish 

 (Fig. 6A). For example, with otoliths, 98% offish with 

 an estimated most probable age of were likely to 

 be aged as 0, whereas only 74% offish with a most 

 probable age of 4 were likely to be aged as 4. This 

 precision can be compared with those obtained from 

 other structures (e.g. scales and vertebrae; Fig. 6A). 

 Fish aged with scales had a higher probability of 

 being consistently assigned ages 1 and 2. Fish aged 

 with vertebrae or otoliths, however, had a higher 

 probability of being consistently assigned age 4, 

 whereas there was little difference between the three 

 structures in assigning fish to age 3 (Fig. 6A). 



