Natanson et al.: Growth of Galeocerdo cuvier m the western North Atlantic 



949 



estimate. However, the tiger shark has been reported 

 to 469 cm FL (Castro, 1983 ). The L„ value estimated 

 from the Fabens (1956) method (293 cm FL), how- 

 ever, is lower than all reported values. In addition, 

 the Fabens (1965) method value for k is high (Table 

 2). We therefore concluded that the Gulland and Holt 

 (1959) VBGF is the more appropriate model to use 

 for the tiger shark. Cailliet et al. (1992) and Van 

 Dykhuizen and Mollet (1992) also preferred the 

 Gulland and Holt (1959) model over that of Fabens 

 (1965) for the angel and sevengill sharks, respectively. 



Data from this study indicated that neither tagging 

 nor tagging combined with OTC injection appears to 

 retard the growth rate in neonate and juvenile sharks 

 up to 150 cm FL (Fig. 3). To the contrary, the growth 

 rates from tagged sharks were higher than estimates 

 obtained from vertebral gi-owth bands (Table 3). The 

 growth rates for the first two years of life for the tiger 

 shark estimated from monthly length frequencies, tag- 

 recapture (Gulland and Holt. 1959), and tag-recapture 

 with OTC were all similar (Fig. 3). Tanaka ( 1990) found 

 that growth rates of OTC-injected Japanese wobbe- 

 gongs, Orectolobusjaponicits, and neonate swell sharks, 

 Cephaloscyllium umbratile, were not significantly dif- 

 ferent from controls. More recently, Gelsleichter et al. 

 ( 1998 ) evaluated the toxicity of OTC on gi-owth rates of 

 captive nurse sharks, Ginglymostoma cirratum, and 

 concluded that there were no adverse effects of OTC 

 on growth rate. These data support the use of OTC as 

 an effective method for determining vertebral band 

 periodicity without inten-upting normal growth pat- 

 terns (Tanaka, 1990; Gelsleichter et al., 1998). 



Our data do not verify nor refute age estimates for 

 the tiger shark previously obtained from vertebral band 

 counts (Branstetter et al., 1987 )( Table 2; Fig. 2). The k 

 value from our study is higher than Branstetter et al.'s 

 (1987) Atlantic value and our L is lower These differ- 

 ences are to be expected when comparing VBGF pa- 

 rameters from a conventional age-length (vertebral 

 methods) study with those derived from a growth in- 

 crement (tagging) study (Sainsbury, 1980; Francis, 

 1988). These types of curves are not directly compa- 

 rable because the parameters are derived differently 

 and, therefore, have different meanings (Francis, 1988). 

 However, comparison of the estimates obtained in these 

 studies to known values, such as size at birth and maxi- 

 mum size, can provide insight into the fit of the curves. 

 This information allows us to determine which curve 

 is best suited to be used for age at maturity and maxi- 

 mum age estimates. We believe that the current tag- 

 recapture values are more accurate on the basis of veri- 

 fication available from the monthly length-frequency 

 analysis and the consistency of the estimates to mea- 

 surable parameters such as size at birth and maximum 

 size. The L^ calculated by Branstetter et al. ( 1987) is 



lower than maximum reported sizes, and the size at 

 birth (73 cm FL), based on the von Bertalanffy curve 

 from that study, was high as related to known param- 

 eters (60-65 cm FL^ ). 



Statistically significant differences between age es- 

 timates from the Gulf of Mexico and North Atlantic 

 populations of tiger shark found by Branstetter et 

 al. (1987) may not be biologically significant (Yoccoz, 

 1991). The differences in age at maturity obtained 

 between these two areas is only 2-3 years. Consid- 

 ering the relatively slow growth and large overlap of 

 size at age for this species, a 2-3 year difference could 

 be included in the realm of measurement error. In 

 addition, we found statistically significant differences 

 in growth by year in first year tiger sharks obtained 

 from the same region over a period of five years. The 

 differences in growth between years indicate that 

 tiger shark growth is quite variable and probably 

 dependent on fluctuations of many parameters in- 



