Stequert et al.: Age and growth of Thunnus albacares 



127 





of 3 counts) corresponding to our 

 sample (Fig. 7). For acetate rep- 

 lica readings, the standard error 

 of the counts is quite small (Table 

 1), and for counts on transverse 

 sections, the coefficient of varia- 

 tion of Chang ( 1982 ) is acceptable 

 (CV=2.802%, n=151).. 



Transverse sections observed 

 with the SEM clearly show that 

 some increments overlap inside 

 the otolith (Fig. 4). These incre- 

 ments cannot be seen on the ex- 

 ternal face even after acid etching 

 for acetate replication. This over- 

 lap does not seem to appear on 

 oblique sections where increments 

 are more evenly spaced. In addi- 

 tion, the plane of growth of the 

 otolith changed when viewed in a 

 transverse plane (Fig. 8). From 

 this plane (Fig. 3), the ventral sur- 

 face of the otolith grows at an 

 angle of about 120° to the original plane of growth 

 which contains the primordium. The direction of 

 otolith growth changes before the fish reaches 28 cm 

 fork length. For all individuals, the distance between 

 the primordium and the top of this angle (Dl, Figs. 3 

 and 8) has a stable value (D 1=644 ;/m [±4 fixa\). Then, 

 the otolith growth depends only on the plane of 

 growth of D2 (Figs. 3 and 8). This change in growth 

 direction occurs, on average, 50 (±5) days after the 

 date of hatching (?? = 151). Because of this change, the 

 use of back-calculation can be complicated. Because 

 ages obtained with the acetate replica technique were 

 similar to those based on transverse sections, we used 

 the ages based on transverse sections for this study. 

 All the observations described above (comparisons 

 between oblique and transverse sections, replicas, 

 light microscope, and SEM) show that the observa- 

 tion of microstructures on transverse sections of yel- 

 lowfin tuna otoliths with a light microscope are suit- 

 able to estimate age in days. 



Growth 



The estimations of parameters were calculated by 

 using the nonlinear regression procedure and are 

 summarized in Table 2. The fits of each of these mod- 

 els are highly significant (P<0.01). For the three 

 models, F-ratios of ANOVA (respectively 14,698, 

 14,799, and 10,960) and r 2 values (0.974 for each of 

 them) are very similar to one another (Table 2). We 

 used the von Bertalanffy growth curve to express our 

 results and to compare with results obtained by other 



It 



Dorsal 



edge 



Figure 3 



Transverse section (A-B) of a right yellowfin tuna, Thunnus albacares, sagitta 

 (length of the section D = Dl + D2). 



Reading area 



Surface of 



thi- otolith 



ov . 



Figure 4 



Daily microincrements on a transverse section of yellow- 

 fin tuna, Thunnus albacares, otolith (SEM, magnification 

 = 800x, scale bar = 38 ^m). ov = overlapping of increments. 



investigators because this model is the most widely 

 used. 



The model was also applied to males (rc=63) and to 

 females (n=6l) separately. Males and females were 

 similar in size at any given age (Table 3 ). ANCOVA 



