Plains Earthlodges: Ethnographic and Archaeological Perspectives

7. This Old Earthlodge Village: How Long Were Sites of the Middle Missouri Tradition Occupied?

7 This Old Earthlodge Village How Long Were Sites of the Middle Missouri Tradition Occupied? Stephen C. Lensink While the importance of knowing the duration of site occupancy has been acknowledged, concerted attempts to apply methods of estimating occupancy duration to Plains villages are limited. Dennis Toom (1992a, 1992b) completed a study of Middle Missouri tradition sites in South Dakota and determined that villages were relocated approximately every 25 years, with the longest occupation of a village being 50 years. He concluded that the 25 Middle Missouri tradition components in the area represent only 2–4 contemporaneous villages with a total population of 500–1,000 persons. Likewise, Tiffany (1991a) and Dallman (1983:30) estimated occupation duration ranges for selected Middle Missouri tradition sites in Iowa, and Tiffany (1982:91) proposed that there would have been three contemporaneous villages in the Little Sioux locality of Iowa with a total population of 450–720. This chapter outlines multiple methods for estimating the duration of occupation of earthlodge villages and applies the methodology to the Initial variant of the Middle Missouri tradition (IMM), a.d. 1100–1250. The Initial Middle Missouri variant has been divided into two divisions, eastern (IMMe) and western (IMMw) (Figure 7.1) (Gibbon 1993; Henning 2001; Tiffany 1983). Since Toom (1992a, 1992b) has already analyzed the western division, the following discussion focuses on the eastern division, comprised of the Lower James, Cambria, Big Sioux, Little Sioux, and Brandon phases. The Big and Little Sioux phases in northwestern Iowa, represented by 33 habitation sites, will be emphasized speci¤cally (Figure 7.2). Finally, three applications of occupation duration information will be examined: (1) estimating population size in a locality, (2) evaluating hypotheses about climate change, and (3) developing models of cultural process. METHODS OF ESTIMATING SITE OCCUPATION DUR ATION There are at least six methods of estimating site occupation duration. These rely on using data about (1) 14 C date dispersion, (2) 14 C mean differences, (3) house rebuilding episodes, (4) rates of midden accumulation, (5) ceramic consumption rates, and (6) subsistence remains and caloric expenditures. This chapter considers the ¤rst ¤ve. Method 1: 14 C Date Dispersion One method of determining the time span of site occupation utilizes the dispersion of a set of 14 C dates obtained from a site. Some archaeologists believe that a series of tightly clustered radiocarbon dates from a single-component site re®ects a brief occupation span, while a set of dates with a larger dispersion statistic re®ects a longer occupancy. Asch and Brown (1990:181), in a study of the dates from the Late Prehistoric Oak Forest site in Illinois, however , conducted Monte Carlo simulations to reach the following conclusions: (1) “If the precision of dates is approximately 70 years, the dispersion of a dating series is a poor indicator of the duration of a single-component occupation ” and (2) “Because the power of dispersion statistics to discriminate between occupations of differing length will increase approximately as the square root of sample size, a moderate increase in the size of a sample series Figure 7.1. Initial variant of the Middle Missouri tradition. 134 Stephen C. Lensink above the six dates run for Oak Forest will not greatly increase the sensitivity of the statistic.” These results imply that 14 C date dispersion can be used to determine duration of occupation only if either the precision of the dates is dramatically improved, say by an order of magnitude, or the number of dates is greatly increased, again perhaps by at least an order of magnitude. For the Oak Forest case, this would mean standard errors of less than 10 years or a suite of dates greater than 60, or both—clearly impractical constraints given current technology and budgets. Method 2: 14 C Mean Differences It is possible to estimate the mean of a population with greater con¤dence than estimating its variance. Therefore, the difference in the means of two sets of dates—one from the bottom and the other from the top of a village midden—might provide an estimate of the site’s duration of occupation. The technique would only work on sites where (1) the occupations are relatively Figure 7.2. Sites of the Big and Little Sioux phases in northwest Iowa. These sites are part of the Mill Creek culture. This Old Earthlodge Village 135 KEY • Mill Creek sites 20 km " " '\ \ \ \ , I I I I...