984 



Fishery Bulletin 97(4), 1999 



data on which the calculations for R. taylori were 

 based are likely to have met at least two of these 

 assumptions. Because the area from which samples 

 were collected was subject to little or no fishing, the 

 population size should have remained relatively con- 

 stant and thus fluctuations in recruitment would be 

 minimized. The necessity of excluding the first age 

 class from the catch curve analysis resulted from the 

 selectivity of the gill nets used to collect specimens. 

 For the age classes included in the analysis, the size 

 of individuals was similar because growth rapidly 

 reaches an asymptote, and therefore selectivity was 

 probably reasonably constant. The assumption that 

 mortality was constant was more difficult to test. 

 However, the small size of i?. taylori and the small 

 change in its size after the first year or two of growth 

 (Simpfendorfer, 1993) suggest that all age classes are 

 likely to be subject to similar levels of predation. 



The estimates of natural mortality for female R. 

 taylori from the catch curve and from the indirect 

 method of Hoenig ( 1983 ) were similar: 0..56 and 0.60, 

 respectively. Such a finding is contrary to that re- 

 ported by Cortes and Parsons ( 1996), who suggested 

 that the Hoenig ( 1983) method would produce lower 

 natural mortality estimates than those from catch 

 curves. The present study suggests that for some 

 species the Hoenig method will produce acceptable 

 results. However, Cortes and Parsons' (1996) caution 

 that demographic studies based solely on a natural 

 mortality estimate from the Hoenig (1983) method 

 should still be heeded because it is clear that the 

 accuracy of this method may vary between species. 



The most biologically plausible estimates of natu- 

 ral mortality for/?, taylori (0.698 for males and 0.561 

 for females) are significantly higher than those re- 

 ported so far for many other species of shark. Spe- 

 cies with mortality levels of a similar magnitude are 

 the bonnethead shark (Sphyrna tiburo) {0.625; Cortes 

 and Parsons, 1996) and the Atlantic sharpnose shark 

 {Rhizoprionodon terraenovae) (0.42; Cortes, 1995). 

 Each of these three species (/?. taylori, R. terraenovae, 

 and S. tiburo) are short-lived (6-12 yr). In longer- 

 lived shark species, estimates of natural mortality 

 are much lower. For example, Sminkey and Musick 



(1996) estimated an M of 0.105 for Carcharhinus 

 pliimbeus, which lives at least 30 years, and Smith and 

 Abramson (1990), Cailliet (1992), and Au and Smith 



( 1997 ) estimated that Triakis semifasciata, which lives 

 to 30 years, has an M between 0.139 and 0.15. 



The small size of i?. taylori probably results in it 

 being subject to high levels of predation throughout 

 life. The small size at birth (220-260 mm) means that 

 the first age class is particularly vulnerable to pre- 

 dation. The results of the sensitivity test, in which 

 natural mortality was doubled for the youngest age 



class, indicate that this technique, which has previ- 

 ously been used to account for increased predation 

 of young sharks, was not realistic fori?, taylori. This 

 is a result of the high value of natural mortality as- 

 sumed for older age classes. It was not possible to 

 estimate what, if any, increase in predation occurs 

 for the youngest age class of/? . taylori. However, 

 any increase in predation would result in reduced 

 estimates of r. 



Demographic analysis 



The most likely demographic results for R. taylori, 

 those based on natural mortality estimated from the 

 catch curve, indicate that the maximum value of r is 

 about 0.27. This value is one of the highest for a spe- 

 cies of shark. Cortes and Parsons (1996) estimated a 

 similar rate (0.272-0.283) for Sphyrna tiburo from 

 Florida with Hoenig's method to estimate natural 

 mortality. Rhizoprionodon taylori and S. tiburo have 

 similar life histories, with short lifespans, rapid 

 growth, early maturity, and high natural mortality. 

 All other published studies of shark demography 

 have estimated intrinsic rate of increase at less than 

 0.1 (e.g. Hoenig and Gruber. 1990; Cailliet, 1992; 

 Cailliet et al., 1992; Cortes, 1995; Sminkey and 

 Musick, 1996; Au and Smith, 1997; Cortes, 1998). In 

 most cases these species have been longer lived and 

 slower to mature and have exhibited lower natural 

 mortality than /?. taylori or S. tiburo. The exception 

 to this is /?. terraenovae, which grows larger than /?. 

 taylori and is relatively short-lived ( 10 yr), but has a 

 lower reproductive rate (Cortes, 1995). 



Demographic analysis of/?, taylori is sensitive to 

 the way age-specific reproductive rate is calculated 

 because of this species's short lifespan. Whether age- 

 specific natality is calculated on the basis of the pro- 

 portion of a population surviving at the beginning or 

 the end of an age class leads to a very large differ- 

 ence in r. This effect is magnified by the high natu- 

 ral mortality of/?, taylori. It is therefore important 

 to take account of gestation period in the calculation 

 of reproductive rate. This is less important in longer 

 lived species in which natural mortality is lower and 

 in which there is a relatively small difference between 

 the proportions of a population surviving until the 

 beginning and the end of an age class. 



This study indicates that accurate age data are 

 required for demographic analysis. The results for 

 /?. taylori showed a high level of sensitivity to changes 

 in age at maturity, but only limited sensitivity to 

 changes in maximum age. Given that the age data 

 on which this research were based have not been 

 validated, there remains some uncertainty about the 

 results. Further work to validate these data (espe- 



