Archaeomagnetic research in the United States midcontinent

This dissertation combines archaeomagnetic and independent chronometric data from 240 archaeological features to develop a regional secular variation curve for the U.S. midcontinent. These data were obtained from features located between 31.5-40.5° N latitude and 82.5-93.5° W longitude that have been dated to between 60 and 10,700 cal BP. The archaeomagnetic samples were collected from 41 sites within this region over the past 35 years under the direction of four different researchers: Robert DuBois (University of Oklahoma), Daniel Wolfman (University of Arkansas and New Mexico State Museum), Wulf Gose (University of Texas at Austin), and myself.

In this project, the data are initially smoothed through the moving windows method to form the first approximation of the curve. Outlier analyses and pairwise statistical comparisons are utilized to refine the smoothed curve, and the results are compared to other Holocene-aged secular variation records from North America. These analyses indicate that the final curve should be treated as three distinct segments with different precision and use recommendations. First, the 850-75 cal BP segment can be used to date archaeomagnetic sample from the project area with expected temporal precision of 100-200 years. Second, the 2528-850 cal BP segment can be used cautiously to date archaeomagnetic samples with an expected temporal precision of 200-300 years. Third, the 9755-4650 cal BP segment should be used for contextual dating purposes only, in that an undated sample can be put into a regional context through comparison with the segment's constituent samples.

Finally, three archaeological problems are addressed through the archaeomagnetic data. First, archaeomagnetic data are used to resolve the temporal conflict between an eastern Tennessee structure's morphology and a much earlier radiocarbon date obtained for the structure. Then, archaeomagnetic data are used to address a number of internal chronology questions regarding three Powers phase sites in eastern Missouri. Finally, the sequencing of several protohistoric and historic sites in eastern Tennessee is examined through a series of archaeomagnetic data.