BRANSTETTER ET AL.: AGE AND GROWTH OF TIGER SHARK 



Table 1 . — Back calculations by age class for the Virginia and Gulf of Mexico samples of tiger sfiarks, Galeocerdo cuvieri. Ages are based 

 on age at the formation of the winter mark. Lengths to nearest cm TL. Significantly different mean lengths at age between samples in- 

 dicated by asterisks ('* P < 0.001; * P < 0.01). 



procedure (Fig. 4) to adequately describe the growth 

 rate. However, young age classes in the Gulf of Mex- 

 ico (hereafter referred as Gulf) sample were slight- 

 ly larger at age than their Atlantic counterpart. 

 Independent von Bertalanffy curves for each data 

 set had different parameter estimates. Regressions 

 of the curves, linearized by log transformation, were 

 analyzed for covariance (SAS Institute 1985) and 

 were significantly different (P < 0.0001). Simul- 

 taneous nonlinear regression analysis of the two von 

 Bertalanffy curves derived from back-calculated 

 mean lengths at age produced parameter estimates 

 with nonoverlapping simultaneous confidence inter- 

 vals (Bernard 1981). 



Back calculations by age class for each sample 

 (Table 1) also showed the more rapid growth of the 

 Gulf juveniles. Mean lengths at the formation of the 

 winter annuli were significantly different (P < 0.01 

 or P < 0.001) for early age classes of the two 

 samples. Neonates in both samples increased near- 



ly 50 cm in length the first 6 months, and the Age 

 I Gulf tiger sharks continued to grow at 50 cm/year, 

 but Atlantic Age I individuals grew <40 cm/year. 

 Gulf tiger sharks continued to grow approximately 

 4 cm/year faster than the Atlantic population until 

 the fourth year. Growth rates then became similar; 

 Gulf tiger sharks were simply larger at age. 



Observed and back-calculated lengths at age 

 (Table 2) corresponded within each sample. Com- 

 parisons of observed and back-calculated lengths did 

 not indicate the occurrence of Lee's phenomenon. 

 Differences in observed and back-calculated lengths 

 at age were attributable to the fact that most speci- 

 mens in both samples were taken in summer; there- 

 fore, observed lengths at age were larger than back- 

 calculated lengths at age based on the winter formed 

 annuli. 



The growth rate estimated from centra was val- 

 idated with one tag-recapture. A female, tagged 3 

 November 1978, was estimated to be 230 cm, and 



273 



