AGE DETERMINATION OF ALBACORE 



357 



Results of ring counts 



After deciding upon the criterion to be followed, 

 we made a series of independent ring counts and 

 ring-radii measurements. A total of 9 series of 

 counts and measurements resulted in agreements 

 between any 2 series of as low as 8 percent and 

 as high as 45 percent. There was poor agreement 

 not only between counts of the two readers but 

 also between counts made by the same person. 

 Table 1 shows the best agreement attained be- 



Table 1. — Best agreement in ring counts between two readers 

 and between any two series of the same reader 



[The readers are designated as A and B; the numerals denote separate read- 

 ings; for example, A2 represents the second reading made by reader .4.] 



tween readers (44.8 percent) and between sets of 

 readings made by the same reader (38.9 percent 

 for reader A and 43.9 percent for reader B). 

 While most of the disagreement was by 1 ring 

 (31 to 50 percent), discrepancies of as many as 

 4 rings occurred in many of the comparisons. 

 Consistency was often not achieved even between 

 counts of the anterior and posterior cones of the 

 same vertebra. 



It was observed that much of the inconsistency 

 in the counts could be attributed to the difficulty 

 in distinguishing rings near the vertex of the cone 

 and also at its margin. This does not mean, how- 

 ever, that the rings on the central portion of the 

 cones were clear and easily distinguishable, be- 

 cause here also there were many inconsistencies. 



Samples from small fish 



Age determination by means of hard parts fre- 

 quently becomes more difficult and often breaks 

 down completely in older fish. Therefore, at- 

 tempts were made to age fish less than 90 cen- 

 timeters in length in order to determine whether 

 or not the use of vertebrae from larger, older fish 

 was the cause of the poor results. In one trial 

 consisting of two series of counts, the agreement 

 was 56 percent for fish less than 90 centimeters 

 as compared with 30 percent for larger fish. The 

 higher agreement with smaller fish is, of course, 

 to be expected because of the fewer rings. Fur- 



ther trials with samples of small fish disclosed 

 that the counts were still far too variable to be 

 acceptable as indicators of age. 



The vertebrae were examined under natural and 

 incandescent light and also under polarized and 

 various colored lights with no noticeable improve- 

 ment in the appearance of the rings. Various 

 types of stains were also tried, but here again the 

 results were disappointing from the standpoint of 

 improving the readability of the rings. Thus, we 

 found that we could not satisfactorily duplicate 

 ring counts and therefore could not meet the 

 first prerequisite to the use of vertebrae for aging 

 albacore. Our findings are in contrast to Partlo's, 

 who reported that "counts of rings were repro- 

 ducible with high consistency." 



Determination of time of ring formation 



In order to validate the vertebral method for 

 aging albacore, it is necessary that some direct 

 evidence be obtained to show that the rings on 

 the centra are true year marks, which are laid 

 down annually. It appears reasonable to assume 

 that if these rings are laid down regularly once 

 each year, rather than randomly, a study of the 

 margins of vertebrae collected periodically through- 

 out a year should indicate the approximate period 

 of ring formation. The width of the margin from 

 the last ring to the edge of the centrum (the in- 

 crement of the last ring) should be at a maximum 

 immediately before, and at a minimum just at or 

 after the formation of the last ring. A descrip- 

 tion of such a method of analysis for the deter- 

 mination of time of ring formation on tuna scales 

 is given by Nose and others (1955). 



An attempt was made to obtain some indication 

 of the time of ring formation, despite the fact 

 that any such study would probably be futile 

 unless the rings or annuli could first be read with 

 assurance. Measurements were made from the 

 mid-point of the centrum to the margin of the 

 vertebra (vertebral radius T) and to the last ring 

 (ring radius R n ) . Although the last ring was care- 

 fully selected in each instance, there was a possi- 

 bility of error because rings on the margins were 

 often indistinct. The difference between the ver- 

 tebral and last-ring radii was considered the incre- 

 ment. Figure 5 shows a plot of the relative 

 increment (absolute increment/vertebral radius) 

 plotted against the month of capture of the fish. 

 The data show no indication of regularity in time 



