340 
-e GROTAG model 
-& Age-based model 
100 120 
80 
-e- GROTAG model 
-& Age-based model 
60 80 100 120 
Growth rate (mm FL/year) 
20 40 
0 
-4- Atlantic region 
- GOM 
fo) 
N 
= 
jo} 
is) 
= 
40 60 80 
20 
0 
Fork length (mm) 
Figure 4 
Growth rates of female bonnetheads (Sphyrna tiburo) estimated with the 
optimal GROTAG models, which are based on tag-recapture data, and mean 
growth rates calculated with the age-based (Francis, 1988a) von Bertalanffy 
growth function, with associated 95% confidence intervals (dotted lines). 
Growth rates are shown for bonnetheads tagged (A) in the northeastern Gulf 
of Mexico (GOM) and (B) in the estuarine waters of the Atlantic coast of the 
southeastern United States (Atlantic region). Also provided are (C) estimates 
of growth rates from the optimal GROTAG model for each of the 2 regions. 
Age-based models were generated by using original length-at-age data 
from Lombardi et al. (2007) for sharks in the GOM and from Frazier et al. 
(2014) for sharks in the Atlantic region. Data used in the GROTAG mod- 
els are from bonnetheads tagged and recaptured in the GOM during 1993-— 
2006 and in the Atlantic region during 1998-2019. Error bars indicate 95% 
confidence intervals generated by bootstrapping (with 5000 iterations). 
FL=fork length. 
Fishery Bulletin 118(4) 
Results from GROTAG models con- 
firm previously published differences 
in growth between the 2 regions. Both 
males and females in the GOM had 
faster average growth at smaller lengths 
and slower growth with a smaller L. 
than bonnetheads from the Atlantic 
region. Visual inspection of 95% CIs 
does not reveal significant differences 
between regions because of large uncer- 
tainty caused by low sample sizes, but 
results of likelihood-ratio tests confirm 
differences in growth. Significant differ- 
ences in growth have been detected for 
other coastal migratory sharks (Carlson 
and Baremore, 2003; Driggers et al., 
2004; Vinyard et al., 2019); however, the 
reasons for these regional growth differ- 
ences are unknown. 
Growth curves derived from GROTAG 
and age-based models are not directly 
comparable; instead, Francis (1988a) 
suggested that comparisons of growth 
rates between data types are more 
appropriate. All GROTAG models pre- 
dicted slower initial growth and faster 
growth at longer lengths compared with 
estimates from age-based models. The 
slower estimated initial growth rates 
likely result from a lack of tagged young- 
of-the-year bonnetheads in all 4 data 
sets, likely a combination of gear bias 
and variation in spatial distribution of 
young-of-the-year bonnetheads (Drig- 
gers et al., 2014). The GROTAG model for 
females in the GOM is based on the larg- 
est sample size of tagged and recaptured 
juvenile bonnetheads (including indi- 
viduals tagged as small as 400 mm FL) 
and, consequently, the best agreement 
between model growth rates at small 
lengths. However, growth rates are 
significantly different at long lengths. 
Estimates from the GROTAG model for 
females in the Atlantic region have excel- 
lent agreement with those from the age- 
based model, especially as lengths reach 
predicted L... Parameters of the VBGF 
are not significantly different between 
models, confirming estimated growth 
from the age-based model for females in 
the Atlantic region. Although the VBGF 
parameters are defined differently for 
the length-at-age data and tag-recapture data, the results 
of other studies indicate that models based on tag-recap- 
ture data can produce growth curves that are similar to 
those from age-based models (Natanson et al., 2002). 
The GROTAG models for males and females predicted 
significantly faster growth at longer lengths, larger L.,, 
variation may also be artifacts of sample size. Future 
efforts should focus on continued tagging of bonnetheads 
in these regions because a more robust data set would 
allow age-independent comparisons of growth and, there- 
fore, determination of latitudinal differences in growth 
within the GOM. 
