NOTE Francis et al.: Uncoupling otolith and somatic growth in Pagrus auratus 



161 



A second set of four ANCOVAs tested the effect of the 

 seven sampling periods on sagitta size. In each case, the 

 slopes of the regression lines differed significantly among 

 sampling periods (Table 2). Slope coefficients declined mark- 

 edly between Periods 4 and 5 (Table 3); consequently, a 

 third set of four ANCOVAs was limited to data for Periods 

 1-4. Whereas slopes did not differ significantly for sagitta 

 length, height, or sulcal width, the intercepts did (Table 2). 

 The three size-variables increased relative to FL between 

 time-periods, i.e., snapper sampled later in the year had 

 larger sagittae than those sampled earlier (Fig. 2Bl. 

 The only sample-pairs that did not differ were Periods 1 

 and 2 for sagitta height and sulcal width measurements 



(Table 2). In the ANCOVA of sagitta weight vs. 

 FL, slopes differed significantly among the four 

 periods; thus the intercepts could not be tested 

 (Table 2). However, sagitta weight followed the 

 same trend as the other size-variables, being 

 greater in snapper sampled later in the year than 

 in those sampled earlier (Fig. 2A). 



Periods 2 and 3 data were used independently 

 to investigate the effect of growth rate on size- 

 variables within sampling periods. The data rep- 

 resent juveniles with estimated post-metamorphic 

 ages of 53.5-136.0 d, and lengths of 43-96 mm FL. 

 Estimated growth rates, averaged over the whole 

 juvenile life, ranged from 0.54 to 0.93 mm/d. Re- 

 siduals from regressions of Period-2 size-variables 

 vs. FL were negatively correlated with somatic 

 growth rate (r=-0.87, -0.70, -0.74, and -0.54 for 

 sagitta weight, length, height, and sulcal width, 

 respectively; p<0.01 in all cases). Therefore, 

 sagittae were heavier and larger (in all dimen- 

 sions) in slow-growing than in fast-growing snap- 

 per. Sagitta weight residuals are plotted against 

 growth rate in Fig. 3. 



Residuals from regressions of Period-3 size- 

 variables vs. FL were also negatively correlated 

 with somatic growth rate (r=0.45, -0.39, -0.13, 

 and -0.26 for sagitta weight, length, height, 

 and sulcal width, respectively). However, only 

 the sagitta weight correlation was significant 

 (p<0.05). 



To determine whether differences in somatic 

 growth rate might explain the observed differences 

 in sagitta size variables between sample periods 

 (Fig. 2), an analysis of variance was performed 

 on growth-rate estimates for Period-2 and -3 snap- 

 per. Variances for the two periods were homoge- 

 neous (F 4660 =1.15, p>0.05. Period-2 snapper had 

 significantly higher growth rates (.f 0.81 mm/d, 

 range 0.68-0.94 mm/d) than Period-3 snapper 

 (x 0.66 mm/d, range 0.55-0.82 mm/d, F u06 = 128.0, 

 p<0.001). 



Discussion 



Residuals analysis showed that in Period 2 and 

 over the length range 43-90 mmFL, slow-growing 

 snapper had larger sagittae (relative to FL) 

 than fast-growing snapper. In Period 3, a similar 

 but weaker pattern was found. This study, there- 

 fore, demonstrates uncoupling of sagitta and so- 

 matic growth in wild age-0+ snapper, and extends 

 a previous report of such uncoupling in reared 

 Japanese snapper up to 30mmSL (Secor et al. 

 1989). 



