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

With these considerations in mind, the Baehr-Gust artifact catalogs were examined for pairs of sherds matching two criteria. First, each pair of sherds had to consist of one sherd classified as White Hall and a second sherd classified as Havana or Hopewell ware. Second, the pair of sherds had to come from the same provenience within a feature, with no recorded evidence of disturbance. Point proveniences were not available for most sherds from Baehr-Gust. It is therefore unknown if the pairs of sherds selected for dating were actually deposited adjacent to one another, or how close they may have been to feature walls. We cannot be sure each pair experienced exactly the same external dose rate, but they are assumed to be reasonably close. The uncertainty in the distance between the sherds will be considered later.

Of the thousands of sherds from several dozen features inventoried in the NYU catalog sheets, only five pairs from four features were located that met both criteria and were therefore considered suitable for TL dating. All other instances of co-occurring White Hall and Havana-Hopewell pottery were associated with clearly recorded evidence of disturbance within the feature. Of the five pairs that met the criteria, one was excluded because one of the sherds had been contaminated with glue.

Since TL dating is destructive, all selected sherds were photographed, and they were submitted to several Illinois Valley ceramic analysts to verify the classifications of the ceramics. Besides Winters, who originally classified the ceramics, these analysts included Eugene Boesch, Duane Esarey, Ken Farnsworth, and Mike Wiant.

There was a considerable amount of disagreement over how the sherds should be classified. In particular, there was disagreement over how undecorated body sherds should be classified (cf. Griffin 1952:121-122; Morse 1963:12). Many body sherds were apparently classified by Winters as White Hall because they had sand temper. However, because sand temper might occur naturally in local clays, the presence of sand temper is probably not an absolute criterion for classifying undecorated body sherds. Other body sherds were apparently classified by Winters as White Hall or Havana based on the thickness of the sherds (White Hall pottery generally has thinner walls than Havana pottery). Given that wall thickness can vary within a single vessel, this too is probably not an absolute criterion for classifying undecorated body sherds.

Because of the disagreement over classification of the sherds, only one of the four selected pairs was a good test, in that all five ceramicists who examined the sherds agreed that one of the paired sherds was White Hall and the second sherd was a Havana-Hopewell variant. This pair of sherds came from Feature 91-XVIII. The three pairs of sherds with classification problems came from Feature 1(1987), Feature 4(1988), and Feature 89-XVIII.

TL analysis was conducted at the University of Washington luminescence laboratory. Routine procedures were followed (Feathers 2000a) and specific details are reported in Feathers (2000b). Briefly, the polymineral fine-grain (1-8[mu]m) method was employed, which is advantageous here because the external dose rate plays a smaller role than it does for coarse-grain analyses (Aitken 1985). Equivalent dose was determined by combining additive dose and regeneration growth curves (which describe the relationship between luminescence and dose) in a method known as the "Australian slide" (Prescott et al. 1993). Radioactivity was measured by thick source alpha counting for U and Th contents and by flame photometry for total K. Cosmic radiation was calculated after Prescott and Hutton (1988) and, although part of the external dose, was treated as a known component. The b-value method was used to correct for lower alpha efficiency. Storage of irradiated discs for times ranging up to eight weeks provided an assessment of any nonthermal fading in the luminescence signal. No significant fading was determined. Equivalent dose was stable over more than 50[degrees] between 300-400[degrees]C for all samples. Scatter in the growth curves, which were generally linear, was relatively low but accounted for much of the error in age. Sensitivity changes by factors of 1.3 to 2 were the rule for regeneration. Various derived parameters used to determine the dates are given in Table 1.