the jth fish; the Xj is the mean age calculated for 

 the jth fish. The third method used the mean coeffi- 

 cient of variation (V) (Chang 1982). This was an alter- 

 native index to APE that provided a statistical test 

 of reproducibility between agings. The coefficient 

 was a measure of the standard deviation between 

 all ages of fish divided by the mean age. The sum 

 of the coefficients was then divided by the number 

 of fish aged. PA, APE, and V are all expressed as 

 percents. 



The difference between PA and APE or V is that 

 the former measures the actual agreement between 

 age sets and the latter measures the amount of varia- 

 tion between age sets. TWo treatments (a treatment 

 was the comparison of two or more age sets) that 

 had the same PA values might have differing degrees 

 of variation and therefore different values of APE 

 or V. The smaller the APE or V values, the less varia- 

 tion there was between age determinations. 



Values for PA, APE, and V were computed for 10 

 treatments: scales vs. otoliths from R^; scales vs. 

 otoliths from Rg; scales from Rj vs. R2; otoliths 

 from Rj vs. R2; scales from Rj and R2 vs. true age; 

 otoliths from R^ and Rg vs. true age; male scales 

 from Rj vs. R2; female scales from R^ vs. R2; male 

 otoliths from R^ vs. R2; and female otoliths from Rj 

 vs. R2. Treatments were also compared as to the 

 relative sizes of the PA, APE, and V values. 



Results 



Figure 2 compares the physical appearance of 

 otoliths and scales for ages 3 through 7. The photo- 

 graph of the age-3 otolith shows its orientation as 

 it lies in the sacculus. The dorsal side, including the 

 anti-rostrum, shows the best contrast between the 

 light opaque (summer growth) regions and the dark 

 hyaline (winter growth) regions. The annular rings 



6yr. 328mm 9^ 



7yr. 326 mm % 



in the sacculus. Lines drawn on the otoliths and scales show annular growth. 



699 



