354 
Fishery Bulletin 11 6(3-4) 
shown in Suppl. Fig. 5. The seasonal (oscil¬ 
lating) VBGF curve (Fig. 6) was also shown 
to fit data well (MSE=0.018), where the 
asymptotic length (L oo =20.94 cm TL), was 
found very close to the L max (20.7 cm TL). 
It is worth noting the high value of the C 
parameter (1.261), which clearly suggests 
a strong oscillating growth pattern with a 
prolonged no growth phase or a sustained 
period of no growth (Pauly and Gaschutz 3 ). 
In order to identify the season when 
change in growth occurs, or to examine the 
probable effect of sex change on the growth 
rate (Tsangridis and Filippousis, 1992; Vi- 
dalis and Tsimenidis, 1996), the observed 
mean lengths at age for each sex of picarel 
sampled were estimated during the life of 
the fish (Suppl. Fig. 6). 
Examining the growth rate, it appears 
that males exhibit a higher growth rate than 
females, and a strong oscillating growth 
pattern is noticed during the 2 nd , the 3 rd , 
and the 4 th year of life. More specifically, a 
decrease of the average length for each sex 
is marked from summer to autumn. Dur¬ 
ing the life of the species, the difference 
in length between males and females was 
found to be positively correlated with age 
(Pearson test: /?=34, P<0.05). 
Length and age at sex change and proportion 
of early maturing males 
The proportion of males was found to be 
positively correlated with length (Pearson 
test: n- 34, P<0.05) and with age (Pearson 
test: n- 57, P<0.05), whereas the inverse oc¬ 
curred with the proportion of females. The 
coefficients of the binomial sex change model 
were a=-10.105 (P<0.01), 6=0.654 (P<0.01) 
and the L 50 was estimated as: L 50 =15.34 cm 
TL (CI=15.17-15.51) (Fig. 7A). The length 
at sex change divided by the L max of the in¬ 
dividuals, was 0.74. Respectively, the coefficients ob¬ 
tained from the number of males in each length class 
were a- -4.711 (P<0.01), 6=0.135 (P<0.01), and the A 50 
was 36.4 months (CI=35.21-37.63 months) or 3.0 years 
(Fig. 7B). 
Using the logistic parameters obtained from the 
number of males in each length class at the 5 th percen¬ 
tile of the population length distribution, we estimated 
the proportion of EMMs at 2.4%. 
Discussion 
Despite the weak contrast between the opaque and hy¬ 
aline zones particularly after the 3 rd annulus, the an¬ 
nuli of picarel were easily readable and show a normal 
growth pattern. The use of monthly samples during the 
first year of picarel life was essential to track comple¬ 
tion of the 1 st annulus. 
The periodicity of annulus formation, shown by 
monthly examination of Mis, showed that these incre¬ 
ments were true annuli. It should be noted that differ¬ 
ent marks can be occasionally laid down on whole oto¬ 
liths, which are not age related but are due to events 
such as settlement, spawning, or maturity (Campana, 
2001). However, those marks do not show the normal 
periodicity and structure of annuli as happens with 
true annuli. In the present work, the date of annuli 
completion, based on the Mis study, was estimated to 
occur after June, when the percentage of otoliths with 
a hyaline area on the edge was the highest (see Suppl. 
Fig 3). 
The time of estimated annuli formation coincides 
with the end of spawning and is in general agreement 
