158 



Fishery Bulletin 97(1), 1999 



found that the majority of the vertebrae had an 

 opaque band as the first band after the angle change 

 (Table 4). Only samples where the nature of the first 

 band was the same in all three counts were consid- 

 ered for our analysis. 



A statistically significant linear relation was found 

 between centrum diameter and PCL (Fig. 3). There 

 was no significant difference between the sexes in 

 the slopes (P>0.5), but a significant difference be- 

 tween the elevations (P<0.001). Because the inter- 

 cepts differed only by 0.02, no visual difference in 

 the predicted values of females and males was per- 

 ceptible, and the common intercept was used to plot 

 a regression for the combined sexes. The intercept 

 was close to zero; therefore no correction, such as 

 the Fraser-Lee method (Carlander, 1969; Branstetter, 

 1987), was used. 



One of the 16 white sharks injected with OTC 

 (BT433) was recaptured after the completion of the 

 study. The shark was tagged during a fishing com- 

 petition on 30 October 1994; it measured 140 cm and 

 weighed 46 kg. On recapture 942 days later, on 28 

 May 1997, it measured 209 cm ( 150 kg). The squares 

 in Figure 3 indicate the length and centrum diameter 

 of this shark at tagging, i.e. at the OTC marker and 

 at recapture. 



Because of the similarity of the above regressions 

 for females and males, back-calculations were per- 

 formed on combined sex data. Mean back-calculated 

 lengths were lower than observed values (Table 5). 

 Lee's phenomenon, a tendency for back-calculated 

 lengths of older fish in the earlier years of life to be 



systematically lower than those of younger fish at 

 the same age (Carlander, 1969; Smith, 1983), was 

 evident, but not consistent. In sharks with 7 and 8 

 GRs, for example, the back-calculated values for each 

 GR class were higher than those for sharks with 6 

 GRs. 



A multiplicative relationship was found between 

 centrum diameter and mass and no significant dif- 

 ference was found between the sexes in the slopes 

 (P>0.5), but a significant difference between the el- 

 evations (P<0.001) (Fig. 4). Again, because of the 

 similarity of the regressions for females and males, 

 back-calculations were performed on combined sex 

 data. Back-calculated mass was lower than observed 

 values (Table 6). Lee's phenomenon was evident, but 

 not consistent, and showed a similar trend to that of 

 back-calculated lengths. Because of the use of mean 

 GRs, the number of observed lengths (or masses) per 

 GR class, plus the number of back-calculated lengths 

 (or masses) in the next GR class, did not always add 

 up to the total number of back-calculated lengths (or 

 masses) of the previous GR class (Tables 5 and 6). 



The observed ratio of translucent to opaque last 

 bands differed significantly from the expected ratio 

 (X^ test, P<0.001, /!=33), irrespective of whether 

 opaque band deposition was assumed to occur in sum- 

 mer or in winter. For this analysis only vertebrae 

 were used where the nature of the last band was the 

 same in all nine counts (i.e. in all three counts of all 

 three methods). The exclusion of borderline cases, 

 i.e. only summer months (22 Dec-20 Mar) and win- 

 ter months (21 Jun-22 Sepj, did not alter the result 



