308 



Fishery Bulletin 105(2) 



formation occurred daily. This is in agreement with 

 Nyman and Conover (1988) who used tetracycline to vali- 

 date that rings were formed daily in juvenile bluefish. 

 This finding indicates that the tetracycline mark was 

 formed directly after the injection and that the number 

 of rings, after tetracycline marking, corresponded with 

 the number of days in the study period. 



The Dahl-Lea equation provided the most accurate 

 estimate of the initial size of the bluefish at the start 

 of the experiment. This is the simplest equation and is 

 based solely on the linearity of the otolith-size-body- 

 size relationship. The accuracy of the estimation may 

 be due to bluefish being a fast growing species. Wright 

 et al. (1990) found that faster growing salmon smolts 

 fit a linear growth model more favorably than slower 

 growing conspecifics. It is possible that daily otolith 

 growth is conservative, to a certain extent, as found 

 by Panfili and Tomas (2001), but that the width of the 

 daily growth increments may be a function of metabo- 

 lism as postulated by Wright et al. (1990) or a function 

 of temperature (Mosegaard et al., 1988). Therefore, 

 species with high metabolisms, such as bluefish. may 

 be expected to show a more direct relationship between 

 otolith growth and somatic growth than a fish with a 

 slower metabolism. 



The Fraser-Lee equation underestimated the true 

 initial lengths of the fish by a mean of 2.8%. This un- 

 derestimate may be due the length-intercept constant c 

 in the to the Fraser-Lee equation. When this constant 

 is given a biological interpretation of zero, the equa- 

 tion is the same as the Dahl-Lea equation; therefore 

 the interpretation of c is important in determining the 

 accuracy of the Fraser-Lee and BPH equations. The 

 c calculated by the linear regression was a negative 



value, -31.41; the underestimation was likely caused 

 by changes in the otolith-size-body-size relationship 

 during different life stages. 



Hare and Cowen (1997) found significant differences 

 between different ontogenetic stages during the bluefish 

 larval period. These differences would explain why the 

 regression of the otolith radius on fork length data 

 did not predict a biologically reasonable y-intercept, 

 although the data were strongly linear: the ratios dur- 

 ing the juvenile and larval stages were different. When 

 applying BCFs a biological intercept may be more use- 

 ful and could make the Fraser-Lee BCF the more ac- 

 curate equation. However, because otolith formation 

 occurs during the early egg stage (0-24 hours) (Hare 

 and Cowen, 1994), it would be problematic to get an 

 accurate mean length at time of formation. Given that 

 this measurement is likely to be less than the 2.0-2.4 

 mm hatching size (Klein-MacPhee, 2002) and that the 

 otolith growth to somatic growth relationship is linear, 

 forcing an intercept of zero should provide reasonably 

 accurate length estimations. 



In summary, this study was designed to validate and 

 compare back-calculation methods for a fast growing 

 juvenile bluefish, rather than to characterize bluefish 

 growth. All four back-calculation formulae did result 

 in close estimates (s8.6%) of the true initial length 

 regardless of growth rate (0.4 to 3.0 mm FL/day). Al- 

 though the sample sizes were not adequate to properly 

 explore the observed variability in growth rate the 

 fact that the precision of the BCFs was not affected 

 by growth rate illustrates a link between the rate of 

 otolith growth and body growth, at least in the short 

 term. In the case of bluefish, the results of this study 

 indicate that the relationship between otolith growth 



