820 



Fishery Bulletin 98(4) 



The observed growth variability in juvenile C. obscurits 

 may in part have been due to seasonal gi-owth variation, 

 migrations between warmer waters on the west coast of 

 Western Australia, and cooler waters on the south coast. 

 Either of these factors could possibly have produced sub- 

 stantial variation in growth rates. Further research on the 

 influence of temperature on growth of C. obscuriis would 

 prove useful in isolating the causes of growth variability 

 in this species. 



The estimated measurement error parameters from the 

 Francis (1988) method were similar but slightly lower 

 than those from the differences in length for individuals 

 at liberty less than 30 days. The close agreement between 

 these results indicates that the Francis (1988) method 

 provides accurate estimates of measurement error and 

 that the confounding of measurement error and growth 

 variability reported by Francis and Mulligan (1998) was 

 not a problem in the current study. The magnitude of the 

 mean measurement error was small (0-0.94 cm) in rela- 

 tion to the length of the sharks examined (55-140 cm). 

 Previous studies using the Francis (1988) method have 

 indicated that including ;?; in the model does not improve 

 the result, whereas values of s have included 1.2 cm (C 

 niilii. Francis, 1997), 1.5 cm IP. o.xygeneios, Francis et al., 

 1999), 1.6 cm iMustelus lenticulatus, Francis and Francis, 

 1992) and 7.2 cm (G. galeus. Francis and Mulligan, 1998). 



The length-at-age analysis indicated that male juvenile 

 C. obsciirus have significantly higher growth rates than 

 females. This result, however, was not supported by the 

 results from the Francis 1 1988) method that showed over- 

 lap in the 95% confidence intervals derived from boot- 

 strapping. Differences in growth rates do occur between 

 the sexes in sharks, but these differences are most com- 

 monly observed at ages close to maturity (approximately 

 20 years in C. obscuriis) (e.g. Simpfendorfer, 1993; Simp- 

 fendorfer et al., 2000). Thus the lack of significant differ- 

 ences in the growth rate between young male and female 

 C. obscuriis was not unexpected. The difference in the 

 results between the two methods may be due to the simple 

 linear model fitted to the length-at-age data not being 

 able to account for growth variability and measurement 

 error This demonstrates the improvement in understand- 

 ing that the more complex Francis (1988) model can pro- 

 vide over more simplistic approaches. 



The results of the length-at-age analysis in the current 

 study indicate that male C. obscurus injected with OTC 

 have significantly higher growth rates than noninjected 

 males. However, females showed no significant difference 

 between injection and noninjection with OTC. The parame- 

 ter estimates from the Francis ( 1988) model indicated that 

 injected animals grow 2.2 cm/year faster than noninjected 

 animals. However the 95'^i confidence intervals overlapped 

 slightly I injected, 9.68-11.63; noninjected, 7.43-10.01) 

 indicating that lliere were no significant differences in 

 growth rates. Although the significance of growth rate dif- 

 ferences varied between analysis methods and sexes, sug- 

 gesting that differences may have been the result of biased 

 data, there is sufficient evidence to warrant further inves- 

 tigation of the effect of OTC injection on juvenile C. obscu- 

 rus. Previous studies of elasmobranchs have not found 



significant differences in gi'owth rates between injected 

 and noninjected animals (Tanaka, 1990; Gelsleichter et 

 al, 1998; Natanson et al. , 1999). However, only one of 

 these studies (Natanson et al., 1999) was carried out on 

 sharks in the wild. This study was based on four OTC- 

 injected G. cuvier specimens and no statistical comparison 

 of growth rates between injected and noninjected individ- 

 uals was undertaken. 



If growth differences between OTC-injected and non- 

 injected C. obscurus do occur, they may result from the 

 antibiotic properties of OTC. It is likely that injections of 

 OTC would not increase growth rates; rather, the growth 

 of noninjected animals may be lower after tagging. The 

 capture and tagging process is likely to present a signifi- 

 cant source of stress for sharks — one that may increase 

 their chances of microbial infection and slow their growth 

 (e.g. Olsen, 1953; Davies and Joubcrt, 1967; Manire and 

 Gruber, 1991). Animals injected with OTC would be more 

 likely to overcome these infections quickly and thus grow 

 at rates indicative of untagged individuals. If this scenario 

 can be proven, then the faster growth rates predicted by 

 the Francis (1988) method for the injected juvenile C. 

 obscurus (10.84 cm/year) are likely to be most representa- 

 tive for this species. 



This study indicates that although a range of techniques 

 are available to estimate growth rates and von Bertalanffy 

 gi'owth parameters from tag-recapture data, they can pro- 

 duce variable results. One method, the Gulland and Holt 

 (1959) method, was not able to produce results for juvenile 

 C. obscu)-us. Another, the Fabens (1965) method, produced 

 unrealistic von Bertalanffy gi-owth parameters. Two meth- 

 ods, the Francis ( 1988) and length-at-age methods, indicated 

 that linear growth functions provided better predictive 

 power for juvenile C. ohscui-us than the nonlinear von Berta- 

 lanffy gi-owth function. The use of the Francis ( 1988 ) method 

 allowed for a more detailed investigation of the growth of 

 C. obscurus. including the investigation of measurement 

 error and growth variability. Estimation of these additional 

 parameters is particularly useful in tag-recapture studies 

 where commercial and recreational fishermen provide the 

 majority of data. Thus, although the more traditional meth- 

 ods are still commonly used for elasmobranchs, researchers 

 should consider the use of more detailed approaches, such 

 as that of Francis ( 1988), to extract the maximum possible 

 information from their data. 



Acknowledgments 



Norm Hall provided valuable advice on the implementa- 

 tion of the Francis method. Phil Unsworth, Tony Paust, 

 Justin Chidlow, Adrian Kitchingman, Justin Bellanger, 

 and Rory McAuley provided technical assistance in the 

 field and laboratory. I am grateful to Rod Lenanton for his 

 guidance and support during this project. Malcolm Francis 

 provided valuable comments on the manuscript and data 

 analysis. This project would not have been possible with- 

 out the commercial gillnet fishermen in southern Western 

 Australia who provided assistance in tagging sharks and 

 who returned many of the tags. This project was funded 



