Dating the middle to late woodland transition in the Illinois Valley: Radiocarbon and thermoluminescence dates from the Baehr-Gust site

Midcontinental Journal of Archaeology, Spring 2003 by Holt, Julie Zimmermann, Feathers, James K

Since gamma radiation has a range of up to 30 cm in soil, TL dating normally requires measuring the radioactivity of the sherd's immediate context, and usually a sediment sample is collected adjacent to the sherd for that purpose. It is also useful if the sherd has been deposited at least 30 cm from boundaries (such as feature boundaries) to avoid potential complexities from differential gamma flux across the boundary.

None of the excavated sherds available from Baehr-Gust were collected with TL dating in mind. Consequently, sediment samples adjacent to the sherds were not obtained. Flotation samples were taken from most features, but these were already floated by the time this analysis was underway. The senior author therefore returned to the site and collected sediment samples from the general vicinity of the features, but it is not known how representative these are of the gamma radioactivity to which the sherds were actually exposed.

Without reliable means of measuring the external dose rate, absolute ratio scale dating is not possible. However, relative dating may be possible. Relative dating at an interval scale is possible for groups of sherds with the same provenience and thus the same external dose rate, but the external dose rate must still be known. The only advantage of interval scale dating is better precision, since the error on the external dose rate and other systematic errors can be treated as constant. If the external dose rate is not known, something close to interval scale dating may still be obtained for sherds from the same provenience. This is illustrated in the following discussion.

The age equation is written as follows:

Age = D^sub E^/(D^sub int^ D^sub ext^ (1)

where D^sub E^ is the equivalent dose in Gy (unit of absorbed dose) and D^sub int^ and D^sub ext^ are internal and external dose rates in Gy per unit time. The interval difference between two samples where D^sub ext^ is constant follows as:

Age (1) - Age (2) = {D^sub E^ (1)/[D^sub int^ (1) C]) - {D^sub E^ (2)/[D^sub int^ (2) C]} (2)

where C is the constant external dose rate and the parenthetic numbers are the sample numbers. By choosing a reasonable range of possible C values, one can determine if the difference in sample age is significant over any of the range. The exact difference can be determined only if C is known, but, in practice, a wide range of C values will seldom yield difference intervals that vary more than the error on that interval. This is in part because the external dose rate for fine-grained samples contributes only about 20-30 percent of the total dose rate and can tolerate relatively large uncertainties. If the difference is not significant, that is, age (1) equals age (2), one could determine C in theory by rearranging Equation 2:

C = {[D^sub E^ (2) * D^sub int^ (1)] - [D^sub E^(1) * D^sub int^ ((2)]}/[D^sub E^ (1) - D^sub E^ (2)] (3)

In practice, because of the subtraction, particularly of the two D^sub E^ terms, propagated errors tend to be quite high and relatively meaningless values obtained. A variant, called isochron dating, has been proposed by some authors (e.g., Aitkin and Valladas 1993) but has never been successfully applied to pottery.