824 



Fishery Bulletin 97(4), 1999 



Lyons, 1991]) suggesting that S. lalandi was a more 

 difficult species to age. For kingfish, scales (7. 69^) 

 had lower values of CV than did otoliths (12%) and 

 vertebrae (11%). The lowest CV values were for struc- 

 tures where no fish were in their first year (e.g. 

 scales), and the highest values were for structures 

 where fish were aged as in their first year (e.g. 

 otoliths and vertebrae). Similarly, Kimura and Lyons 

 (1991) found that the CV was highest for species in 

 wMch 1-year age classes were present and lowest 

 where the youngest age classes were 2 or 3 years. 

 Coefficient of variation therefore depends on the 

 minimum age offish and we suggest that its use be 

 limited to comparisons where fish have the same 

 range of ages. Such indices, if averaged over all fish, 

 are also thought to simplify the data because any 

 trends that might occur with age are ignored (see 

 Hoenig et al. , 1995). Precision of aging decreased with 

 age of fish for all structures, suggesting that esti- 

 mates of precision should not be averaged over all 

 age classes. We believe that the methods of Richards 

 et al. ( 1992) give a better indication of aging error, 

 and therefore precision, than did methods that sum- 

 marized over all age classes. 



Length at age 



Estimates of mean size-at-age calculated from growth 

 models suggested that the resulting growth curves 

 from otoliths, scales, and vertebrae were similar for 

 all but the youngest age class. Although other aging 

 studies on Seriola spp. have calculated average 

 length-at-age (e.g. Mitani, 1955; Mitani and Sato, 

 1959), few have fitted models to the data (but see 

 Baxter, 1960; Manooch and Potts, 1997). Growth 

 models from all aging structures were within the 

 range of growth oi^ Seriola spp. described elsewhere 

 (e.g. Mitani, 1955, 1958; Mitani and Sato, 1959; 

 Baxter, 1960; Murayama, 1992). 



Further research needs to focus on the position of 

 the first zone in all structures because this may con- 

 tribute to differences in the shape of the growth 

 curves between structures. No fish were found to be 

 in their first year (i.e. 0) when aged with scales, but 

 fish were found to be in their first year when aged 

 with otoliths and vertebrae. There are also differ- 

 ences in estimated mean size between scales and both 

 otoliths and vertebrae for fish with one growth zone, 

 which is also likely to be due to the detectability of 

 the first zone. Either scales lay down a false first 

 zone or the first zone is not detected in otoliths and 

 vertebrae, possibly because it is close to the edge of 

 the structure. Collections of fish from recruitment 

 over a 12-month period and sampling of all struc- 

 tures may help elucidate the position of the first zone. 



Although, it is well known that age estimates from 

 scales are inaccurate once growth becomes asymp- 

 totic (Beamish and McFarlane, 1987), scales did not 

 appear to underestimate ages of kingfish in relation 

 to otoliths in the present study but may have under- 

 estimated ages of older fish in relation to vertebrae 

 (e.g. Fig. 5). In a previous study on S. lalandi, spaces 

 between circuli became increasingly narrow after the 

 seventh year, so that it was impossible to differenti- 

 ate between one year's growth and the next (Baxter, 

 1960). Few fish, however, had greater than seven 

 growth zones in the current study. 



Many studies have found that sectioned otoliths 

 are preferable to whole otoliths for aging because 

 examination of whole otoliths does not take into ac- 

 count the asymmetric deposition of material (Cam- 

 pana, 1984). After a certain age, deposition may 

 thicken the otolith and growth along the anteropos- 

 terior axis may decline; therefore only transverse 

 sections would reveal recently formed growth zones 

 (Beamish, 1979). Growth zones on whole otoliths 

 were interpretable for aging kingfish, but zones could 

 rarely be interpreted on sectioned otoliths. This is 

 unusual because in many studies, sectioned otoliths 

 have been found to give more precise and accurate 

 estimates of age than whole otoliths (e.g. Beamish, 



