714 



Fishery Bulletin 102(4) 



white, indicating that most organic material had 

 been removed. A three-step elemental separation 

 procedure was used to remove calcium and bari- 

 um, elements that interfere with the detection of 

 radium in the TIMS process. This involved pass- 

 ing the samples through three cation exchange 

 columns, two containing a slurry of BioRad AG E 

 50W-X8 resin (first and second column), and one 

 containing EiChroM Sr® resin (third column). 

 The samples were introduced to a highly acidic 

 medium within the columns, which separated the 

 elements according to elution characteristics (An- 

 drews et al., 1999b). 



Radiometric age for each group was determined 

 by inserting the measured 210 Pb and 22B Ra activi- 

 ties into the secular equilibrium model (Smith et 

 al., 1991) and correcting for the elapsed time be- 

 tween capture and autodeposition. Because these 

 activities were measured from the same sample, 

 the calculation was independent of sample mass 

 (Andrews et al., 1999a, 1999b). Propagated uncer- 

 tainty associated with the final 210 Pb activity was 

 based on count statistics, and procedural error 

 and uncertainty for the final 226 Ra activity was 

 based on procedural error and an instrumental 

 TIMS analysis routine (Wang et al., 1975; An- 

 drews et al.. 1999b). The combined errors were 

 used to calculate high and low radiometric ages. 



Accuracy of age estimates 



Measured 210 Pb: 226 Ra activity ratios for each age 

 group, along with their total sample age (predicted 

 age + time since capture), were plotted with the 

 expected 210 Pb: 226 Ra growth curve. Each age group 

 range was widened by multiplying the minimum 

 and maximum age in the range by the age estimate 

 CV, which was determined from the variability in 

 age estimates among three readers. Agreement 

 between the measured ratio with respect to esti- 

 mated age and the expected ratio (ingrowth curve) 

 provided an indication of the age estimate accuracy. 

 Radiometric age was compared to the average pre- 

 dicted age for each group by using two tests: 1) a 

 paired sample £-test to determine if a significant 

 difference existed between the two age estimates 

 for the groups and 2) predicted age was plotted 

 against radiometric age and the correlation was 

 compared to a hypothetical agreement line (slope of 

 1) by using r-tests for slope and elevation. 



Results 



Figure 1 



Three images of a blackgill rockfish (Sebastes melariostomus) 

 otolith section viewed with transmitted light at 25x magni- 

 fication (top), 40x (center), and 80x magnification (bottom). 

 This section was aged most consistently as 90 years under a 

 microscope, but because of finer digital resolution and contrast, 

 the section pictured can be aged as high as 102 years. 



Estimation of age and growth 



Growth zones observed within otolith sections of most 

 blackgill rockfish were difficult to interpret. Distinction 

 of the first annulus was often ambiguous, and the band- 

 ing pattern during the first several years (1 to -10) of 



growth was, in some sections, wide and inconsistent. 

 After approximately 8 to 12 growth zones, the zone width 

 transitioned to a narrower zone, which became extremely 

 compressed after 20-40 growth zones. In some sections, 

 these older zones were beyond optical resolution, whereas 

 in others they were remarkably clear (Fig. 1). 



