Minerals and Elements: Using Petrography to Reconsider the Findings of Neutron Activation in the Compositional Analysis of Ceramics from Pinson Mounds, Tennessee
Stoltman, James B., Mainfort, Robert C., Jr., Midcontinental Journal of Archaeology, MCJA
Of 164 sherds and soil samples from Pinson Mounds and vicinity previously analyzed by neutron activation analysis (NAA), 39 specimens (plus 13 other "local" sherds) were subjected to petrographic analysis. The goat of this study is to assess the differential effectiveness of the two techniques to discriminate locally produced from imported ceramic vessels at prehistoric sites. The conclusion of the NAA that "none of the analyzed sherds with nonlocal surface treatments was manufactured from nonlocal clays" is contradicted in several cases, thus calling into question the effectiveness of NAA to address reliably issues of ceramic production and exchange.
Ceramic composition can provide archaeologists with valuable information about many aspects of human behavior, including resource procurement, technology, organization of production, and cultural interaction (including exchange). The compositions of archaeological ceramics are normally characterized either chemically or physically. Various techniques have been employed to determine the chemical composition of ceramics, but neutron activation analysis (NAA) is used most commonly. The leading technique for determining the physical composition of ceramics is petrography. The two approaches characterize ceramic compositions in distinctly different ways-the former in terms of chemical elements, the latter in terms of minerals and rocks-and are generally believed to provide complementary information. The present study provides an empirical test of that supposition by subjecting to petrographic analysis a series of sherds from Pinson Mounds and vicinity in western Tennessee whose compositions have previously been determined by NAA.
In several previous studies, neutron activation and petrography have been employed jointly in the analysis of archaeological ceramics (e.g., Rands and Bishop 1980; Strazicich 1998; Triadan 1997; Zedeno 1994). The present study differs, however, in that petrography is utilized as an independent test of the findings of NAA rather than serving primarily "in a supportive and interpretive role" to NAA (Rands and Bishop 1980:19). The petrographic analysis of a uniquely large number of samples previously analyzed by NAA-35 sherds and four soils-substantially contradicts the findings of the earlier analysis, thus raising serious concerns about the complementarity of the two approaches for addressing questions of production and exchange of prehistoric ceramics.
Pinson Mounds (40MD1) is located in the West Tennessee Uplands physiographic province in Madison County, Tennessee, a region underlain by unconsolidated sediments of Cretaceous and Tertiary ages (Floyd 1965:9, 15; Figure 1). The site includes at least 12 mounds, an earthen enclosure, and several specialized activity loci extending over an area of approximately 160 ha. Long considered a Mississippian site based on the presence of five large, flat-topped, rectangular mounds, it was not until the 1970s and 1980s that intensive excavations confirmed its Middle Woodland affiliations (e.g., Mainfort 1980, 1988, 1996; Mainfort and Walling 1992). Pinson Mounds is now considered the largest Middle Woodland site in the Southeast (Bense 1994:154; Mainfort 1988:168, 1996:384).
The closest cultural affinities of Pinson are generally regarded to be with the Miller culture of northern Mississippi (cf. Jenkins 1979:172, 1982:82-83; McNutt 1996:215; Mainfort and Walling 1992). Like Miller culture ceramics, Pinson ceramics are sand tempered, consisting mainly of the types Furrs Cordmarked, Baldwin Plain, and Saltillo Fabric Impressed (e.g., Mainfort 1986; Mainfort and Chapman 1994; Mainfort and Walling 1992:122-123). Unlike "classic" Miller I-II assemblages, however, some grog- and mixed grog-and-sand-tempered ceramics occur at Pinson Mounds in Middle Woodland contexts (Mainfort and Walling 1992:122-123). In part, this probably reflects regional variation in ceramic technology. …