64 



Fishery Bulletin 93(1). 1995 



terraenovae first enter these fisheries, as well as the 

 relative proportions of each age and size group rep- 

 resented, preclude a more detailed analysis at this 

 time. However, the demographic analysis represent- 

 ing the best case scenario indicated that under the 

 present fishing level R. terraenovae should not enter 

 the fishery until individuals reach about 97 cm TL or 

 almost 6 years of age if the population was managed to 

 just replace itself. There is evidence that smaller ani- 

 mals are being caught in the various fisheries, but the 

 proportions of each age class are unknown. 



The biological parameters incorporated in sce- 

 narios 1 through 3 represent the best, most reliable 

 information available. Data on age and growth were 

 taken from a tetracycline-validated laboratory study 

 (Branstetter, 1987) which indicated that females 

 mature entering their fifth year of life (age-4) and 

 that maximum age is between 8 and 10 years. In 

 another study (Parsons, 1985), female maturity was 

 estimated at between 2.4 to 3.9 years. However, this 

 study used only males and mean lengths for age 

 classes, which, as pointed out by Branstetter (1987), 

 affected the von Bertalanffy parameters. The possi- 

 bility of earlier female age at maturity and even 

 longer lifespan was incorporated in several of the 

 demographic analyses (scenarios 5, 6, 7, 8, 13, and 14), 

 which evidently yielded more liberal results on which 

 more risk-prone management decisions could be based. 



The unpublished information on fertility at age was 

 derived from a study on the reproductive biology of 

 R. terraenovae (Parsons, 1983) and relates female to- 

 tal length to number of uterine eggs or embryos for 78 

 specimens. Parsons ( 1983) also noted that tropical popu- 

 lations of R. terraenovae had been reported to have as 

 many as 12 embryos. This possibility was taken into 

 account by doubling fertility at age in several analyses 

 (scenarios 4, 7, 8, and 14), which again produced more 

 optimistic estimates of population parameters. 



The age, growth, and reproduction information 

 used in this study was based on animals collected in 

 the northern central and western Gulf of Mexico. The 

 extent to which this information is applicable to the 

 entire population or whether there are different 

 stocks in the Gulf with different age, growth, and 

 reproductive capabilities is not known. For example, 

 I recently examined an 82-cm-TL pregnant female 

 with 3 embryos, measurements which fit nicely the 

 regression equation of Parsons (1983), but which 

 would result in a back-calculated age of 3 years with 

 the von Bertalanffy growth function, although the 

 female could have been older, e.g. age 4, owing to 

 variability in size at age, which is not uncommon in 

 sharks (Kusher et al., 1992, and references therein). 



The most important and also the most difficult 

 parameter to estimate is natural mortality (M). The 



value of M used in this study was taken from Hoenig's 

 ( 1983) relationship between longevity and total mor- 

 tality for virgin or lightly exploited stocks. The as- 

 sumption that Z could be approximated to M, or that 

 no fishing mortality occurred during the period for 

 which growth parameters for this species were de- 

 rived, may have been violated. However, the possi- 

 bility of lower natural mortality values was incorpo- 

 rated in several analyses (scenarios 9 through 14). 

 While Hoenig's equation represents a shortcut and 

 obvious simplification of reality, the lack of catch and 

 effort data, or age or size composition for stocks of 

 this species precludes calculation of any other esti- 

 mates of M at this time. Lack or inappropriateness 

 of both fishery and biological data may explain why 

 several other researchers have used the same ap- 

 proach to estimate natural mortality in shark popu- 

 lation studies. Except for age-0 Negaprion brevir- 

 ostris (Manire and Gruber, 1993), no actual age-spe- 

 cific estimates of natural mortality are available for 

 any shark species. 



The value derived for M (0.42) in this study is 

 equivalent to an annual survivorship of 0.66, which 

 is low when compared with survival estimates for 

 other species of sharks. Values derived from Hoenig's 

 (1983) regression equation include 0.82 for the an- 

 gel shark, Squatina californica, (Cailliet et al., 1992); 

 0.85 for N. brevirostris (Hoenig and Gruber, 1990); 

 0.87 for Triakis semifasciata (Smith and Abramson, 

 1990; Cailliet, 1992); and 0.90 for Carcharhinus 

 plumbeus (Hoff, 1990). Grant et al. (1979) derived a 

 value of 0.90 for the Australian school shark, 

 Galeorhinus australis, using cohort analysis, and 

 Walker ( 1992) used a value of 0.82 in a dynamic pool 

 fishery simulation model of the gummy shark, 

 Mustelus antarcticus, which was also obtained 

 through cohort analysis. The lower survivorship 

 value for R. terraenovae may be due to the smaller 

 size of this species which would make it more sus- 

 ceptible to predation by other sharks, especially at 

 early ages, since pups are born at about only 30 cm 

 TL in coastal waters about 10 m deep (Castro, 1993). 



The very high estimate of F (0.428) used in this 

 study was derived from a shark stock assessment 

 that is the basis for the recently implemented (26 

 April 1993) FMP for sharks of the Atlantic Ocean. 

 However, the accuracy of this estimate, based on a 

 4-year catch-and-effort time series, is uncertain, and 

 the demographic analyses undertaken in this study 

 indicate that R. terraenovae is vulnerable to high 

 removal levels in the early years of life. 



It is also possible that the age, growth, and repro- 

 ductive data used in this study are only representa- 

 tive of the population at a time when fishing pres- 

 sure was not as high as it is at present. Potential 



