samples of bone fragments were submitted for radiocarbon dating
to the Center for Accelerator Mass Spectrometry (CAMS), Lawrence
Livermore National Laboratory.
In addition, three carbon samples from the Salmon Beds
were submitted for conventional radiocarbon assay to the Brock
University Earth Sciences Radiocarbon Laboratory (BGS).
The results are shown in Table 6.
carbon samples were chosen as they were among the largest
samples from the site, because they were all from adjacent
units and because they provided a series from different depths
L6 (Level 6, 25-30 cm), L8 (Level 8, 35-40 cm) and L11 (Level
11, 50-55 cm).
All three resulting dates from Brock University overlap
within one sigma range of one another.
Two samples (BGS 2161, 2163) produced almost identical
dates while BGS 2162 yielded a slightly later date.
different explanations can be offered for the dating anomalies:
three dates were deposited within a short time span between
approximately AD. 1000 and 1220 years.
BGS 2161 may have been contaminated in its origin - for
example, it could have been derived from an older tree
or washed in wood that was burned at a later date.
BGS 2162 may have been contaminated, possible during excavation
or post excavation handling.
The lab calculated AD date yielded 3 probable dates
(1163, 1173, 1180) suggesting some uncertainty for this
three bone collagen dates were selected because they were
derived from adjacent units and were from three upper levels
at the site-Level 2 (5-10 cm), Level 4 (15-20 cm) and Level
6 (25-30 cm). In
addition, the Level 6 date from the two laboratories could
potentially verify one another. The dates provided by CAMS of 610 +/- 40, 400 +/- 40, and 710
+/- 40 yrs B.P. are all within a 300 year period.
They are however, not in the expected sequence because
the sample closest to the current surface should be the most
recent and there should be a progression in ages from Level
2 through to Level 11.
three earlier dates were likely deposited within a short time
between approximately AD 1000 and 1220 years.
Because of the relative depths of the three samples,
it is most probable that BGS 2161 is too early and should
date younger that BGS 2162. Level 11 (Sample BGS 2163) is
near the base of the cultural materials which indicates that
deposition of this terrace began slightly earlier than 1000
years ago. The
terrace has then built up gradually to its modern level.
the three more recent dates, Level 2 is near the surface and,
as such, it should have the most recent date.
The date provided by the Level 2 sample (CAMS 60322)
of 610 +/- 40 could result from mixing, either as a result
of river action, the excavation of a cultural pit or the action
of an animal such as a beaver or a muskrat.
Another possiblity is that some mixing could have occurred
due to dredging of the river channel in the early twentieth
dredging was conducted along sections of the Columbia River
to permit paddle wheelers to navigate the Columbia River.
Center for Accelerator
Mass Spectrometry, Lawrence Livermore National Laboratory
Brock University Earth Sciences Radiocarbon Laboratory
final conflict in the radiocarbon dates occurs between two dates
from Level 6 (25-30 cm). These two samples were taken from adjacent units at this level.
A date derived from a bone sample yielded 710 +/- 40 B.P.
(CAMS 60324) while a carbon sample yielded a date of 989 +/- 50
years B.P. (BGS 2161). No
explanation can be provided for this apparent difference except
as suggested earlier that the date on charcoal may be derived from
wood that died sometime earlier than when it was used.
the date from Level 2 (CAMS 60322) and the date from carbon from
Level 6 (BGS 2161) are eliminated, the remaining dates provide a
logical progression from the base to the top of the site as follows:
988 +/- 50 (Level 11), 880 +/- 50 (Level 8), 710 +/- 40 (Level
6) and 400 +/- 40 (Level 4) years B.P. (Figure 30).
inexactness of radiocarbon dating is a recognized phenomenon.
All of the dates are plotted on Figure 30.
These dates were used to determine a Pearson Product-Moment
Correlation. If deposition
had occurred at a consistent rate the correlation should have a
value of 1.0. The Pearson Product Moment Correlation for the six dates yielded
a value of 0.73 indicating that the correlation of these dates to
depth is acceptable.