Andrews et al.: Radiometric age validation of Megalops atlantlcus 
395 
Figure 3 
Transverse otolith sections from three Atlantic tarpon ( Megalops atlanticus) illustrating the inherent difficulty in counting annuli. 
Annulus-derived ages were estimated by counting visible growth increments for each section shown (Crabtree et al. 1995). One 
sagitta was used for counting annuli and the other for radiometric aging. (A) An otolith section for which the number of growth 
increments clearly exceeds the radiometric age estimate. The annulus-derived age estimate was 37 yr and the radiometric age 
estimate was 24.5 yr. (B) An otolith section with fewer growth increments than the radiometric age estimate, but which could have 
had a higher annulus-derived age estimate if the reader had taken a more aggressive approach. The annulus-derived age estimate 
was 17 yr and the radiometric age estimate was 28.0 yr. (C) An otolith section with a well-developed sulcal region and numerous 
growth increments that rapidly become compressed toward the margin at the lower left. ( D ) Close up of the region in the same 
section (C) where growth increments are compressed and narrowly spaced. Note that growth increments in the region are partially 
obscured by what appears to be a proteinaceous inclusion. The annulus-derived age estimate was 55 yr and the radiometric age 
estimate was 78.0 yr. 
derived age estimates that were higher than radiometric 
age. Poorly defined annual growth zones, or growth zones 
that become too small or compressed to be quantified, 
may explain annulus-derived ages that were lower than 
radiometric ages (Fig. 3). An additional consideration was 
that the age groups were composed of otoliths pooled to- 
gether on the basis of annulus-derived age. This means 
that the radiometric age determined for these groups was 
the average age of the otoliths in the sample. Hence, a 
mixture of otoliths of differing ages would create an unac- 
countable error in the radiometric age of the age groups. 
This variability, however, was minimized by using oto- 
liths of similar weight. 
Von Bertalanffy growth functions fitted to the radio- 
metric ages had growth parameters that were similar to 
the annulus-derived growth parameters (Crabtree et al., 
1995). This determination, however, is subjective because 
1) there were a low number of radiometric samples, 2) 
the confidence intervals were large, and 3) three out of 
seven male samples and three out of eight female sam- 
