FISHERY BULLETIN; VOL. 87, NO. 4, 1989 



computed from four or more observations. The 

 first three principal components of S. pinniger 

 variability explained 75% of the total variance 

 summed over the six growth anomaly time 

 series; loadings are shown in Figure 5A. Corre- 

 sponding principal component time series (Fig. 

 6), or "modes of variabihty," effectively filter the 

 data to draw attention to the dominant signals 

 discussed above (Fig. 3). By considering the 

 loading values (and thus the relative contribu- 

 tion of each age class) along with the trends in 

 the principal component time series, we de- 

 scribed the dominant signals in the time series of 

 the original data. The loadings of the first mode 

 were approximately the same for all six age 

 groups (Fig. 5A); this principal component rep- 

 resented the dominant mode of coherent vari- 

 ability with about the same amplitude in all age 

 groups. The corresponding principal component 

 time series (Fig. 6) is nearly averaged from 1940 

 to 1970, followed by an abrupt increase begin- 

 ning about 1972, and represented the high 

 growth rates for all age groups after 1972 (cf. 

 Fig. 3). This was the most pronounced signal in 

 the record, accounting for 36% of the variability 

 over the six age gi-oups. 



p 

 13 

 > 



u 



IX 



_3 

 (0 

 > 

 U 



a. 



12 3 4 5 6 



Age Group 



Figure 5. — Component (age gi-oup) loadings to the three 

 principal components (PC's) explaining the greatest amount 

 of variability in the gi'owth time series. A. Sebastes pin- 

 niger. B. S. diploproa. Circles represent the loadings of 

 the first PC, triangles the second, and squares the third. 



-2 



1940 



1950 



1960 



1970 



1980 



Figure 6. — Principal component time series for the three modes explaining 

 75.0% of the total variance in gi'owth of Sebasfes pinniger from the six time 

 series in Figure 3. The curve represents the double 5-yr running average. 



798 



