Natanson et al.: Growth of Galeocerdo cower in the western North Atlantic 



947 



the first year, the young grew from an average birth 

 size of 65 cm FL to 100-105 cm FL with a growth 

 rate of 40-45 cm/yr. (Fig. 4). More hmited data on 

 second year growth indicated a rate of 35-45 cm/yr 

 Even with this variation, growth in all years paral- 

 leled the tag-recapture growth curve and the indi- 

 vidual growth rates of the OTC specimens (Fig. 3). 



Discussion 



It is evident that traditional methods for aging te- 

 leosts do not always work well for sharks. Length- 

 frequency analysis is difficult owing to the slow 

 growth exhibited by most elasmobranch species. 

 Hard part analysis requires validation of the period- 

 icity of band formation, which is often difficult to 

 obtain for shark species. Vertebral age estimates for 

 the sandbar shark, Carcharhinus plumbeus, which 

 was one of the six species considered validated, have 

 been revised since Cailliet's (1990) paper by using 

 tag and recapture evidence (Casey and Natanson, 

 1992). The new data indicate that although the ver- 

 tebral bands may be formed annually in the young 

 shark, as validated ( Branstetter, 1987b), they are not 

 formed annually throughout the life of the shark and, 

 therefore, band counts severely underestimate age. 

 This type of revision highlights the need for valida- 

 tion of all size classes. It is advisable to use several 

 methods for aging to provide verification for the cho- 

 sen growth curve, particularly if vertebral band counts 

 are used without direct validation. 



Tag-recapture data can be a useful tool for age and 

 growth determination if accurate measurements are 

 taken at both tagging and recapture and if individu- 

 als are at liberty for a sufficient time for growth to 

 occur. However, problems are associated with this 

 method as well. For example, most length measure- 

 ments are estimated by recreational and commer- 

 cial fishermen. For slow growing sharks, it is impera- 

 tive to obtain accurate length measurements, par- 

 ticularly in large fish. This can prove difficult as well 

 as dangerous at tagging. Therefore, data on large 

 individuals is sometimes lacking and thus will bias 

 results. Additionally, some researchers have shown 

 that tagging with "M" type tags in small sharks, such 

 as the lemon shark, may retard growth (Manire and 

 Gruber, 1991 ). Analysis of the data can also be prob- 

 lematic. The Fabens (1965) method can lead to bi- 

 ased estimates because its basic premise, that tagged 

 individuals are at large for equal time periods, is of- 

 ten violated with sharks. Estimates from that method 

 lead to low values of L^ and high values of ^ (Chien 

 and Condrey, 1987). The Gulland and Holt (1959) 

 method, which allows for unequal times at liberty. 



