570 
Fishery Bulletin 96(3), 1998 
A number of studies have compared whole and sec- 
tioned otoliths as alternative structures for age de- 
termination. The magnitude of the difference be- 
tween whole-otolith and sectioned-otolith age varies 
from species to species, but the general trend is an 
underestimation of whole-otolith ages in older fishes. 
Wilson and Boehlert (1990) found that sectioned- 
otolith ages produced smaller estimates than 
whole-otolith ages, but growth rates were similar for 
both in the canary rockfish, Sebastes pinniger (Gill). 
In the long-lived orange roughy, Hoplostethus 
atlanticus Collett, Smith et al. (1995) found that sec- 
tioned-otolith ages exceeded whole-otolith ages af- 
ter the age of reproductive maturity. Ferreira and 
Russ (1994) found in the Great Barrier reef coral 
trout, Plectropomus leopardus (Lacepede), that the 
mean difference between sectioned-otolith and whole- 
otolith readings remains within one year in fishes 
up to 12 years, but increases abruptly to almost four 
years for age groups 13 and 14. Pearson et al. (1991) 
found that whole-otolith ages tend to be 
as much as eight years less than bro- 
ken-and-burnt otolith ages in the short- 
belly rockfish, S. jordani Gilbert. In L. 
peru, discrepancies between whole-oto- 
lith and sectioned-otolith ages were not 
extreme, but whole-otolith age determi- 
nations were not reliable beyond the 
15th annulus. It would be useful to ex- 
tend these observations to other popu- 
lations of L. peru to assess the variabil- 
ity of this pattern. 
The coefficient of determination of the 
multiple regression model is affected by 
the degree of colinearity of the indepen- 
dent variables. Although not all the as- 
sumptions required for the use of a 
multiple regression model are met by 
the data set, including homoscedasti- 
city, this does not preclude its results 
to serve as a first-order age approxima- 
tion. Otolith weight appears to be a very 
good predictor of age in this species, as 
it has been observed in P. leopardus 
(Ferreira and Russ, 1994). The model 
was used to estimate the age of those 
otoliths that had been discarded as un- 
readable and for which all variables 
were available, placing the predicted 
ages very close to the VBGF curve (Fig. 
7B). The success of this method suggests 
that it could be implemented for rou- 
tine age determination of this species. 
Growth 
The growth of L. peru was satisfacto- 
rily described by the generalized VBGF. 
The absence of a significant difference 
in mean length-at-age between sexes 
and the significantly different growth 
parameters might seem contradictory. 
However, the graphic representation of 
the data revealed that both sexes have 
very similar growth schedules during 
100 r 
0 i 1 “i 1 i 1 i 1 i _l_ i r i ' i 1 r -1 i r i 1 i 1 i 1 i 1 i 1 i 1 i 
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 
Age (years) 
Figure 7 
(A) Scatter plot of all length-at-age data used for VBGF parameter estima- 
tion. Length-at-whole-otolith age (circles), length-at-sectioned-otolith age 
(triangles), and VBGF curve obtained from this study (solid line). (B) VBGF 
curve from this study (solid line) and lengths at age reported in other stud- 
ies of age and growth of Lutjanus peru. Mean observed lengths at age (open 
circles) and mean back-calculated (dotted open circles) from Ruiz et al. ( 1985); 
mean observed length at age from Aguilar Salazar (1986) (squares), mean 
observed length at age from Castro (1981) (triangles). Diamonds joined by a 
dotted line indicate the von Bertalanffy growth curve reported by Gutierrez 
Vargas (1990). Filled circles represent age estimates obtained with the mul- 
tiple regression model for “unreadable” otoliths. 
