380 
Fishery Bulletin 106(4) 
; -Z< 
SSR - > ( v yu]~^u]) 
: X [((V ^'] + ^ ,/C7) - / -V[7] ]2 - 
An iterative process to estimate p and o can be 
found in Kastelle et al. (2008). Again, to evalu- 
ate for aging error, we purposely biased the ages 
by 0, ±1, ±2, and ±4. For each of these seven sets 
of ages, we estimated p and o, standardized the 
validation sample A U C observations with {v y[j] l, 
and calculated the SSR. As before, the smallest 
SSR indicated the best fit to the reference chronol- 
ogy and indicated if any overall aging bias (error) 
existed. It should be noted that if p and a are not 
estimated, but instead are set to p = 0 and o= 1, 
this process becomes the unstandardized proce- 
dure described earlier. 
The third and final method we employed was 
estimation of confidence intervals around the 
loess-smoothed reference chronology (Kastelle et 
al., 2008). For the estimated confidence intervals, 
simultaneous Bonferonni statistical inference was 
used (Miller, 1966) that calculates simultaneous 
(a=0.01) confidence intervals whose width is dependent 
on the number of distinct years at which comparisons 
are made between the reference chronology and valida- 
tion samples, and on the variability in the reference 
chronology. As an aid in our comparison between the 
standardized Dover sole otolith 4 14 C values and the 
Pacific halibut reference chronology, not only were the 
SSRs evaluated, the standardized 4 U C values were also 
viewed graphically when plotted against the posited 
birth years. This comparison generated seven plots, 
one for each purposeful bias, and included the loess- 
smoothed Pacific halibut reference chronology with con- 
fidence intervals. 
The category-2 otoliths were re-examined again after 
the 4 14 C results were known. The goal of this re-aging 
process was to use the z\ 14 C results as a guide for re- 
fining current otolith aging criteria. To facilitate this 
process, the 4 14 C in the category-2 otoliths was plotted 
with a loess-smoothed curve of category-1 results. This 
process allowed the age reader to learn how difficult- 
to-interpret otoliths are best aged. In this last re-aging 
process a minimum and maximum age were estimated 
for the category-2 otoliths. 
The 3-year core in Dover sole otoliths needed to be 
taken into consideration when analyzing the results. 
We assumed a mid-point of approximately 1.5 years for 
core deposition because the core represents material 
from the first 3 years of life. This means that we as- 
sumed linear otolith growth during the first three years 
of life. Linear otolith growth may not be accurate, but 
was assumed for simplicity and because any error from 
this assumption is trivial. The Pacific halibut reference 
chronology is based on material from only the first year 
of life; therefore, we assumed a mid-point of 0.5 years 
1950 1960 1970 1980 
Birth year 
Figure 3 
4 14 C %c in otolith cores plotted against birth year of Dover 
sole (Microstomus pacificus), by category 1 (•) and category 
2 (A), with a 95% confidence interval shown for each point. 
Two category-! specimens are overlapping at -108 h 14 C and 
birth year 1953. The line ( ) is a loess-smoothed curve of the 
Pacific halibut ( Hippoglossus stenolepis) reference chronology 
(Piner and Wischniowski, 2004). 
for core deposition. These assumptions mandated an 
approximate allowance of a (1.5-0. 5 =) 1.0-year shift in 
the comparison of the Dover sole results with the Pacific 
halibut reference chronology. 
Results 
Otolith selection and coring procedures 
The selection process generated specimens for which 
final ages agreed with initial ages and was followed 
by successful coring. In all but three of the specimens 
in category 1, the agreement between the initial age 
and final age was within 3 years. In this category, the 
maximum final age was 47 years, and the minimum age 
was 8 years (Table 1). In category-2 specimens, four out 
of five had discrepancies of over 3 years between the 
initial age and final age. The percentage CV (Kimura 
and Anderl, 2005) between initial ages and final ages 
in both categories was 4.21%. The average core weight 
across all categories was 5.4 mg (±0.2 mg standard 
error) and the average size was 1.93x2.97x0.49 mm, 
which was smaller than the guide provided by the 3- 
year-olds described earlier. 
,4 C analysis 
The 4 14 C in Dover sole otolith cores from category 1 
followed the expected general pattern of initial low 14 C 
before atmospheric testing followed by an increase syn- 
chronous with testing. It displayed a rise in about 1955 
(from below -100 %c) and peaked in 1966 at over 85%o 
(Table 1, Fig. 3). This trend in Dover sole radiocarbon 
