480 



Fishery Bulletin 98(3) 



L. argentimaculatus, L. gibbus, and L. bohar 

 (Fig. 7). There were no opaque zones visible past 

 the OTC mark on both axes of one L. gibbus, 

 and L. monostigma had no OTC mark visible on 

 the ventral axis. An opaque zone was coincident 

 with the OTC mark on the ventral axis, but not 

 the sulcal axis, of L. bohar. 



The ease of interpretation of opaque zones in 

 the otolith preparations was tested by comparing 

 the pairs of counts made by the two otolith 

 readers. Nonparametric, Wilcoxon signed rank 

 tests showed significant differences between 

 readers in estimates offish age along both reading 

 axes (ventral axis /;=82 S=285.0 P > | S | =0.0001; 

 sulcal axis n=S2 8=285.5 P >|S| =0.0001 ) but 

 not in interpretation of the number of opaque 

 zones past the OTC mark (ventral axis ;!=61 

 S=21.0 P > I S I =0.5034; sulcal axis «=62 S=15.0 

 P > I S I =0.6476). An age bias plot (Campana et 

 al., 1995) showed that the mean and standard 

 deviation of the CAF estimates were 2.9 ±0.5 

 for 2+ fish, 3.5 ±0.6 for 3+ fish and 4.2 ±0.4 for 

 4+ fish aged by the senior author We inferred 

 that there was similar interpretation amongst 

 readers of the area of interest outside the OTC 

 mark and that the definition of the first opaque 

 zone was a source of bias for youngest fish. 



Sections from Liitjaniis otoliths have been 

 read along the sulcal L3 axis in a protocol 

 developed by Newman et al. (1996), but opaque 

 zones may be closest together in this area and 

 measurements of growth past OTC marks might 

 best be made elsewhere. To justify our choice 

 of measurement axes for subsequent analyses 

 we compared the mean growth and variance 

 from the two axes in Table 2. Growth rates 

 along the ventral axis were consistently twice as 

 high as those along the sulcal axis; therefore we 

 judged that errors in measurement of zone radii 

 would be least, and interpretation best, along 

 the ventral axis. The coefficients of variation 

 were similar for both axes, with the exception of 

 those for L. setae. 



Lutjanus sebae showed the highest mean 

 growth rates and variability along both reading 

 axes (Table 2 ) which led us to investigate artifacts 

 of captivity. The growth rates of otoliths of L. 

 johnii and L. sebae in the field were much lower 

 than those of captives (Figs. 4 and 6). Captive 

 L. johnii had mean otolith growth rates of 0.78 

 mm/yr along the ventral axis — twice as high as 

 those tagged in the field ( 0.36 mm/yr). These differences were 

 highlysignificant(n=20ndf=lddf=18F=22.49P>P=0.0002), 

 but no significant difference in otolith weight was detected 

 when fish length was used as a covariate (/i=20, ndf=l 

 ddf=18 F=0.19 P>F=0.6704). This latter finding indicated 

 that faster otolith growth rates were correlated with the 

 faster somatic growth of captive fish. 



Further evidence of this correlation and the effects of 

 captivity on somatic growth were an increase in mean 



FL of captives by about 7 mm/month for L. johnii and 

 about 12 mm/month for L. sebae, compared with field 

 growth rates of 0.2-6.2 mm/month for L. johnii and only 

 2.6-3.9 mm/month for L. .se6ot'. One field-tagged L. johnii 

 (s6667. Fig. 6 1 showed both an apparent decline in FL and 

 no otolith growth outside the OTC mark on the ventral 

 axis. 



Plots of increment cycle width (IVi along the ventral 

 axis against cycle number (a) were made to examine 



