Booth et al.: Age validation, growth, mortality, and demographic modeling of Tnakis megalopterus 
109 
Figure 4 
Observed length at age (dots) and von Bertalanffy pre- 
dicted growth (line) of captive and wild Triakis mega- 
lopterus (sexes combined) sampled from the southeast 
coast of South Africa. The dotted lines between the 
dots represent the growth in length between successive 
ages after validation with oxytetracycline hydrochloride. 
years for females) estimated by Kusher et al. (1992). In 
T. megalopterus, sexual maturation occurs at approxi- 
mately 79% and 83% of asymptotic length for males and 
females, respectively. In terms of maximum observed 
length, sexual maturation occurs at 86% and 70% of as- 
ymptotic length. These ratios were higher than those of 
T. semifasciata, which are 63% and 72% for males and 
females, respectively (Kusher et al., 1992). Although 
our estimates are high, and could possibly be improved 
with larger sample sizes, they indicate that T. megalop- 
terus is vulnerable to overexploitation and that manage- 
ment measures (particularly exemption of this species 
from commercial exploitation in South Africa) need to 
be rigorously enforced to protect this endemic species 
(Compagno et al., 1989). 
Unfortunately all captive specimens examined in 
this study exhibited retarded growth. In some teleosts, 
this has been attributed to an effect of OTC (Mona- 
ghan, 1993). In both Japanese wobbegong ( Orectolobus 
japonicas) and nurse sharks (Ginglymostoma cirratum), 
Tanaka (1990) and Gelsleichter et al. (1998) showed 
that OTC, at concentrations ranging from 20 to 80 
mg/kg, had little adverse effects on growth or health. 
It is therefore unlikely that OTC, administered at 50 
mg/kg, would have caused growth retardation, and 
there appears to be an alternative explanation for the 
negative growth observed — one that may be related to 
temperature stress. 
Sustained water temperature, if beyond the optimal 
range of the species, can have an adverse affect on 
growth. Selong et al. (2001) showed that if an upper 
temperature threshold is exceeded, the result is inhib- 
ited feeding and later, death. Summer temperatures in 
the Bayworld Aquarium exceeded the upper threshold of 
Fishing mortality rate (per year) 
Figure 5 
Isopleth of the conditional intrinsic rate of population 
increase (r) for Triakis megalopterus at different com- 
binations of age at 50% selectivity and fishing-induced 
mortality rate. The dashed line illustrates the current 
age at 50% selectivity of 11.13 years. The inset figure 
illustrates the conditional intrinsic rate of population 
increase as a function of fishing mortality at the cur- 
rent age at 50% selectivity. 
the natural temperature range tolerated by spotted gul- 
ly sharks in their natural habitat — occasionally spiking 
at 27°C — whereas the normal temperature range off the 
Eastern Cape coast is approximately 12°C to 21°C in 
waters of <20 m. All the captive sharks were observed 
to reduce their food consumption and were noticeably 
thinner during summer than in winter, indicating that 
the summer elevated temperatures were exerting physi- 
ological stress on these display specimens. Whereas 
wild sharks are able to respond to high temperatures 
by swimming to deeper or cooler waters, this option is 
not available to captive individuals and they may not be 
able to avoid temperature stress. The aquarium animals 
used in this study were removed from display and were 
euthanized when they began to show signs of lethargy 
and exhibited marked weight loss and heat stress. It 
should be noted that some individuals of this species 
were able to cope with the high temperatures and sur- 
vived multiple years, indicating that there is individual 
variation in susceptibility to heat stress. 
The growth of the specimens over a seven-year period 
followed the growth curve and supports the robustness 
of the von Bertalanffy growth model for this species. 
Meaningful estimates of length between capture and 
recapture events in slow growing, long-lived species 
can only be attained if the time interval between the 
two events is large. If the time interval is too small, 
measurement error would be high in relation to the 
