640 



Fisheiy Bulletin 88(4), 1990 



0.5 1,0 1.5 2.0 2.5 



GROWTH RATE (mm/day) 



3.0 



Figure 3 



Rate of circulus formation vs. fish growth rate scattergrams and 

 GM regression lines for age-0 smolts held in saltwater tanks (O, 



— ), for CWT jacks returning to Coos Bay in 1983 (D. ) and 



1985 (C>, --), and for CWT juvenile fish caught in the ocean (A, 

 ). 



fish held in saltwater tanks appeared to be nonlinear 

 (see footnote. Table 3). In general, circulus spacing vs. 

 growth rate values in the three groups that had reared 

 in the ocean were fairly close to the extrapolated linear 

 regression line for the smaller fish held in saltwater 

 tanks, although the variability about this regression line 

 was very large (Fig. 4). Mean growth rates of jacks 

 returning in 1985 and juveniles caught in the ocean 

 (1.49 mm/d and 1.53 mm/d, respectively) were signif- 

 icantly greater (i -tests, jo<0.01) than the mean growth 

 rate of jacks returning in 1983 (1.17 mm/d). Mean cir- 

 culus spacings of the two faster growing groups (3.86 

 and 4.04 mm at 88 x for jacks returning in 1985 and 

 for juveniles caught in the ocean, respectively) were 

 also both greater (;)<0.01) than mean circulus spacing 

 of the slower-growing jacks returning in 1983 (3.61). 

 In each group of fish some of the large variability in 

 the relationship between circulus spacing or circulus 

 deposition rate and fish growth rate was caused by 

 variability in the relationship between fish length and 

 scale radius. To remove this component of variation, 

 we compared circulus spacing and circulus deposition 

 rates with scale growth rates (Tables 4 and 5, respec- 

 tively). Correlation coefficients for the relationships of 

 circulus spacing and deposition rate with scale growth 

 rate were considerably higher than for the correspond- 

 ing relationships with fish growth rate, especially for 

 jacks returning to Coos Bay in 1983 and 1985. (Com- 

 pare Tables 2 with 4, and 3 with 5.) Circulus spacing 

 was positively and significantly correlated with scale 

 growth rate in all groups, whereas the relationship 

 between circulus spacing and fish growth rate was 



Table 3 



Circulus spacing (CSP, mm at 88 x ) vs. fish growth rate (GR, 

 mm/d): Correlation coefficients (r). probability that correla- 

 tion coefficient = 0.0 (/)) and geometric mean regression (GM) 

 forage-0 fish held in saltwater tanks (A). CWT jacks (age 1.0) 

 returning in 1983 (B) and 1985 (C), and CWT juvenile fish (both 

 ages 0.0 and 1.0) caught in the ocean (D). 



' For this group of fish the correlation cuefficieiit for ,-i tliinl- 

 order relationship, CSP = 0.91 + 5.35(GR) - 6.44(GR)- 

 + 3.04 (GR)l r = 0.83, was higher than for the linear 

 relationship. 



5.0 



4.0 



o 3.0-- 



Q- 



(/I 



_l 

 3 

 U 



o 



2.0 



0.0 



0.5 1.0 1.5 2.0 



GROWTH RATE (mm/day) 



2.5 



Figure 4 



Circulus spacing vs. growth rate scattergrams and GM regression 

 lines for age-0 smolts held in saltwater tanks (O, — ), for CWT jacks 

 returning to Coos Bay in 1983 (D, regression n.s.) and 1985 (0, 

 --) and for CWT fish caught in the ocean (▲, -•■-). 



significant for all groups but jacks returning in 1983. 

 Thus, the correlations of circulus spacing or rate of cir- 

 culus formation with scale growth rate were stronger 

 than the correlations with the underlying fish growth 

 rate. 



Discussion 



A positive correlation between rate of circulus forma- 

 tion and fish growth rate appears to be a common 

 feature among fishes. We found in young coho salmon 

 that rate of circulus deposition was positively and 



