FISHERY BULLETIN: VOL. 75, NO. 2 



opercles were not readable, especially those 

 sampled from yellowtail rockfish, where 32 of 35 

 could not be used for age determination. 



Consistency of otolith and scale readings subse- 

 quently was compared in a larger sample. A chi- 

 square test for paired data corrected for continuity 

 revealed that exact agreement between otolith 

 readings was significantly greater than exact 

 agreement between scale readings for yellowtail 

 (P<0.05) 5 and black (P<0.005) rockfishes (Table 

 3). No significant difference occurred between 

 readings of otoliths and scales for canary rockfish 

 (P>0.90). Percent agreement between first read- 

 ings of both structures for all three species was 

 low. 



TABLE 3. — Percent agreement in estimates of age between first 

 and second readings of the same structure and between first 

 readings of different structures (otoliths and scales) sampled 

 from the same yellowtail rockfish, canary rockfish, or black 

 rockfish caught off Oregon, 1974-75. 



I Frrsr reading 

 ] Second reading 



In terms of consistency of readings, the otolith is 

 the best structure of those examined for age 

 determination of yellowtail, canary, and black 

 rockfishes; yet, even this method is questionable. 

 Deviations of readings of yellowtail rockfish 

 otoliths by two readers generally increased with 

 age of the fish (Figure 1). For canary rockfish 

 otoliths read twice by the same person, deviations 

 of readings initially increased and then stabilized 

 with increasing age (Figure 2). Deviations of 

 readings of black rockfish otoliths read twice by 

 the same person also increased with age of the fish 

 (Figure 3). The distribution of deviations is 

 skewed considerably in the positive direction, 

 indicating that the second reading was substan- 

 tially lower than the first. For our largest sample 

 of 322 yellowtail rockfish, 481 canary rockfish, 

 and 357 black rockfish, respectively, 71, 76, and 

 76% of the two readings deviated by no more than 

 ±1 assumed annulus. In a study on Pacific ocean 

 perch by Westrheim (1973), 85% of two otolith 

 readings by different people deviated by no more 



FIGURE l.— Age composition of 322 yellowtail rockfish obtained 

 by two independent readings of their otoliths; specimens were 

 collected from fish processing plants in Astoria and Coos Bay, 

 Oreg., 1973-74. 



 First reaemff 

 ^ Stcond raodine 



FIGURE 2. — Age composition of 353 canary rockfish obtained by 

 two independent readings of their otoliths; specimens were 

 collected from fish processing plants in Astoria and Coos Bay, 

 Oreg., 1974. 



I First reading 



J Second reading 



Probability of a greater chi-square value. 



408 



FIGURE 3.— Age composition of 242 black rockfish obtained by 

 two independent readings of their otoliths; specimens were 

 collected from fish processing plants in Astoria and Coos Bay, 

 Oreg., 1974. 



