534 
Fishery Bulletin 95(3), 1997 
were made; we were unaware of fish length and any 
prior age determination during counting. Incomplete 
marginal increments were not counted. Furthermore, 
counts of right and left otoliths were conducted sepa- 
rately. In situations where the first two counts differed, 
a third independent count was performed. The assigned 
age corresponded to the two counts that were in agree- 
ment. If agreement could not be reached on two of the 
three counts, the otolith was considered unreadable and 
was not used. Otoliths in YOY juveniles (>100 mm SL) 
were counted according to the procedures used for YOY 
king mackerel in Collins et al. (1988). 
Nonlinear regression analysis was used to describe 
the relation between age and length. Statistical 
analyses were performed with SYSTAT software 
(Wilkinson, 1988) and Table Curve ( Jandel Scientific) 
and were based on a significance level of 0.05. 
Results 
Several features of increment deposition were ob- 
served to be consistent among the otoliths examined. 
Two diffuse and poorly defined increments (core in- 
crements) surrounded the primordium (Fig. 3). Mean 
core width was 11.4 mm and there was little varia- 
tion with fish length (SB=0.54 mm, n=40, length 
range=9. 0-300.1 mm). Although these increments 
were counted as daily, the nature of their deposition 
was clearly different from that of subsequent rings. 
This finding indicated that they were formed during 
a separate developmental stage. The absence of fish 
younger than 9 days precluded precise determina- 
tion of the time period represented by these two in- 
crements. Subsequent increments were clearly de- 
fined on most lapilli and were easily discernible in 
whole otoliths examined under a light microscope 
without any special preparation (grinding or polish- 
ing). Subdaily increments occurred, particularly in 
older juveniles, and were discernible from the daily 
increments (Fig. 4). 
Marginal increment analysis 
Of 165 fish examined for marginal increments, 13 
were not used in the final analysis owing to damage 
to the otoliths or to uncertainties in distinguishing 
the marginal or adjacent increments (or both). No 
significant difference in SMI was found between left 
and right lapilli (paired Ptest, P=0.191). Examina- 
tion of fish captured during the 1613-2330 h time 
period revealed an obvious split in the stage of mar- 
ginal increment completion (Table 1). Aunimodal dis- 
tribution of mean SMI, for fish captured over a 24-h 
period, was obtained if the mean SMI of those otoliths 
whose margin was bordered by a translucent zone 
