372 
Fishery Bulletin 115(3) 
Original and traditional von Bertalanffy Francis 
Fixed-Lo and Mooij Gompertz Logistic 
Age (years) 
Figure 6 
Growth curves for (A) females, (B) males, and (C) the entire sample 
of narrownose smooth-hound ( Mustelus schmitti) captured in Anegada 
Bay, Argentina, in 2008. Black dots indicate data points for females and 
males. Note that curves for the original and traditional VBGF models 
and for the fixed-L 0 variant and the model with the Mooij parameteriza- 
tion are identical; therefore, each pair has only one type of line. 
values of the means from the obser- 
vational data, back-calculated lengths 
were higher than the observed lengths 
(i.e., a slight positive Lee phenome- 
non occurred in the data). This trend 
was more evident for females than for 
males. 
The 7 growth models used to deter- 
mine the growth parameters of nar- 
rownose smooth-hound fitted the age 
data adequately (Fig. 6, A and B). 
For females, the original and tradi- 
tional VBGF, Francis parameterization 
(L m = 888.93, calculated from L 1; L 2 , and 
L 3 ), and Gompertz models provided 
somewhat similar estimates for L„. 
In comparison with results from those 
models, L„ from the fixed-L 0 variant 
and the model with the Mooij function 
were lower, and the logistic model pro- 
vided an intermediate estimate of L„. 
For males, the situation was analogous; 
however, from the logistic model was 
similar to the values of L„ estimated 
by the original and traditional VBGF, 
Francis parameterization (L«=748.05, 
calculated from L lt L 2 , and L 3 ), and 
Gompertz models. These parameter es- 
timates are shown in Table 3. Back-cal- 
culated data fitted to the growth func- 
tions that were employed produced re- 
sults similar to those obtained from the 
observational data. It is worth noting 
that, for females, the estimates for L„ 
were lower than those obtained from 
the observational data, but, for males, 
this pattern was not observed (Table 3). 
For females, the original and tradi- 
tional VBGF variants and the model 
with the Francis parameterization of 
the VBGF produced the smallest AIC 
and BIG values and the highest Akaike 
weight of evidence ( w , the conditional 
probability of each model); when com- 
bined, these 3 growth models repre- 
sented more than 65% of the weight 
of evidence (Table 4). The Gompertz 
model ranked second with a w of 0.17 
for females. The fixed-L 0 VBGF variant 
and the model with the Mooij parame- 
terization of the VBGF produced higher 
values of AIC and BIG and a low w for 
females. For males, the fixed-L 0 vari- 
ant and the model with the Mooij pa- 
rameterization had the lowest AIC and 
BIG values and high w. Although the 
original and traditional variants and 
the model with the Francis parameter- 
ization ranked second, they account 
for much of the combined w. All model 
