Kasteile et a!.: Age validation of Microstomas pacificus by means of bomb radiocarbon 
381 
Table 2 
Sum of squared residuals ( SSR ) between A u C in category-1 Gulf of Alaska Dover sole ( Microstomus pacificus) otoliths and loess- 
smoothed curve of the Pacific halibut ( Hippoglossus stenolepis) reference chronology (Finer and Wischniowski, 2004). Results 
are tabled by unstandardized (No) or standardized (Yes) A 14 C values. The u and u are coefficients in the linear standardiza- 
tion defined for the series as {v | -j- i =(v y y^+ p)/ o) , where p and o were estimated by a least squares fit to the loess-smoothed 
curve of the reference chronology data set, or set to 0 and 1 respectively when unstandardized. When standardized, SSR was 
minimized with respect to p and a. Age bias was applied to each final age estimate such that -4 represents younger ages and 
+4 represents older ages. 
V ■(.; Standardized 
to reference 
Parameter 
-4 
-2 
Age bias (years) 
-1 
0 
+1 
+2 
+4 
No 
Yes 
p 
o 
SSR 
p 
a 
SSR 
0 
1 
40,429 
28.47 
1.10 
15,544 
0 
1 
13,179 
9.59 
0.94 
7969 
0 
1 
11,651 
1.63 
0.91 
9533 
0 
1 
20,140 
-6.17 
0.90 
16,834 
0 
1 
38,281 
-13.97 
0.92 
29,191 
0 
1 
66,696 
-22.80 
0.95 
45,706 
0 
1 
153,202 
-45.09 
1.11 
84,612 
Figure 4 
Series of seven plots for A 14 C %o in otolith cores of Dover sole ( Microstomus 
pacificus) category-1 (#) where the A 14 C is standardized and each plot, (A) 
through (G), corresponds to a purposeful age bias of -4 to +4 years, respec- 
tively, with a loess-smoothed curve ( ) of the Pacific halibut ( Hippoglossus 
stenolepis) reference chronology (O) (Finer and Wischniowski, 2004) and 99% 
simultaneous confidence intervals around the mean smooth. 
values parallels those seen in the 
Pacific halibut reference chronology 
but the values appear to be shifted 
earlier in time by 1 or 2 years. Also, 
from 1969 on, the Dover sole values 
are mostly lower than those in the 
reference chronology (Fig. 3). 
In the analysis of the category- 
1 ages, a purposeful bias of -1 or 
-2 years provided the lowest SSRs. 
In the unstandardized procedure, 
where the ages were purposely bi- 
ased -1 years, the SSR was small- 
est at 11,651 (Table 2). In the stan- 
dardized procedure when the ages 
were purposely biased -2 years, the 
SSR was smallest at 7969, where 
p-9.59 and a=0.94 (Table 2). For 
the standardized analysis, a se- 
ries of plots is presented, one plot 
for each of the seven sets of pur- 
posely biased ages (Fig. 4). Both 
the overall time shift and the low 
bias after 1969 were removed by the 
standardization procedure when the 
purposeful age bias was -2 years 
(Fig. 4B). Also, at this purposeful 
age bias and standardization, the 
Dover sole validation specimens 
were all within the 99% confidence 
intervals around the loess-smoothed 
reference chronology. 
The general difficulty in aging 
category-2 specimens is evident in the A i4 C results that 
are not synchronous with those from the category-1 
specimens. The range of possible ages for category-2 
specimens (Table 1) is shown by horizontal bars in 
Figure 5. Two of the three younger category-2 speci- 
mens were likely over-aged because only the right end 
of the horizontal bar is close to the values for the other 
specimens or the values of the loess-smoothed curve of 
category-1 specimens. Therefore, for these three speci- 
mens the choice of a younger age is more accurate. In 
