Rocha-Olivares: Age, growth, and mortality of Lutjanus pern 
567 
< / > 
I 
>. 
05 
Q. 
=> 
O 
O) 
® 
CT> 
< 
16 r 
14 
12 
10 
8 
6 
4 
2 
0 
16 r 
14 
12 
10 
8 
6 
4 
2 
0 4s 
L_ 
0 
6 
12 1 
3 
5 
1 2 
2 15 1 
S 2 
2 21 
7 
2 
2 
1 8 
1 
_L 
1 i 
2 1 
4 
1 1 
1 1 
1 
CO 
0 
5L 
CD 
1 
S 
zr 
o 
<D 
O 
O 
CO 
O 
B) 
(O 
CD 
Q. 
31 
® 
3 
~ (/> 
8 10 12 14 16 10 20 30 40 50 60 70 80 90 
Whole-otolith age (years) 
Total length (cm) 
Figure 4 
(A) Scatter plots of sectioned-otolith and scale ages against whole-otolith ages. Numbers represent 
the frequency of occurrence. (B) Mean age difference (± standard error of the mean) of scale and 
sectioned-otolith ages minus whole-otolith age against length of Lutjanus peru. 
and whole-otolith ages were as large as six years, where- 
as for sectioned otoliths, differences did not exceed three 
years (Fig. 4A). Scale ages were consistently lower than 
whole-otolith ages in fish larger than 50 cm TL, whereas 
the largest sectioned-otolith and whole-otolith age de- 
viations were found in fish above 40 cm TL (Fig. 4B). 
The mean sectioned-otolith and whole-otolith age de- 
viations did not exceed one year, although mean scale and 
whole-otolith age deviations often exceeded one year. 
Six independent variables with highly significant 
coefficients were included in the multiple regression 
model (Table 3). The model explained 96% of the vari- 
ability in the data with a standard error of less than 
one year. A linear regression of whole-otolith age on 
otolith weight alone accounted for 93% of the vari- 
ance. The residuals revealed a trend indicating a 
certain degree of heteroscedasticity but were nor- 
mally distributed (x 2 =10.41, P= 0.005). Whole-otolith 
ages determined with the multiple regression model 
for 28 fish with unreadable otoliths were very close 
to the VBGF (see “Discussion” section). 
Growth 
Mean observed length-at-age were not significantly 
different between sexes, ft-tests, 0.95>P>0.09), however 
Table 3 
Regression coefficients for the multiple regression model 
of whole-otolith age as a function of otolith and fish mea- 
surements of Pacific red snapper. SE = coefficient stan- 
dard error, SEE = standard error of the estimate. R = mul- 
tiple correlation coefficient; No. = number of observations. 
Variable 
Coefficients 
SE 
P 
Intercept (c) 
-6.183 
1.0940 
0.000 
Total length squared 
1.306E-03 
1.4910E-04 
0.000 
ln( Gutted weight) 
1.353E-05 
3.1700E-06 
0.000 
Otolith weight squared 
5.915 
0.7313 
0.000 
Otolith width squared 
-7.391E-01 
2.4400E-01 
0.000 
In (otolith ventral 
radius) 
2.153E-02 
7.6050E-03 
0.005 
Otolith weight 
-1.014E-02 
3.3360E-03 
0.003 
R = 
0.9797 
SEE = 
0.6469 
No. = 
283 
male and female VBGF parameters computed sepa- 
rately from individual data were significantly differ- 
ent (P 2 =609.5, P«0.001) (Fig. 5). The 99% Roy-Bose 
simultaneous confidence intervals indicated that the 
three VBGF parameters significantly contributed to the 
observed difference in predicted growth (6.091<L oo 
