This paper reports on two experimental studies designed to assess how Tuscumbia limestone from northwestern Alabama responds to heating. Experimentally derived fire-cracked rock was examined petrographically in order to develop physical correlates for thermal alteration. The experimental samples were then compared to archaeological samples taken from the Late Paleoindian through Middle Archaic site of Dust Cave, Alabama, to determine whether or not the limestone found there deserves a designation as fire-cracked rock. Results suggest that Dust Cave's occupants made use of locally available Tuscumbia limestone as part of their cooking and processing technologies rather than transporting more distant Tennessee River cobbles to the site. Results further suggest that the cave's occupants preferentially selected massively bedded varieties of Tuscumbia limestone over more thinly bedded varieties due to its higher tolerance for thermal stress and overall greater resistance to degradation during use.
Fire-cracked rock (FCR) servedA an important and flexible purpose in prehistoric technologies. The archaeological and ethnographic literature is replete with examples of rock use as heating elements in roasting ovens, heat "radiators" for drying and processing food, sources of steam in sweat lodges, and boiling implements in cooking (Dering 1999; House and Smith 1975; Rapp et al. 1999; Sassaman 1993; Thorns 1989; Wandsnider 1997). Unfortunately, FCR has historically been an understudied-and arguably undervalued-artifact class compared to flaked and ground stone tools (Wilson and DeLyria 1999). And because of its shear abundance and cumbersome nature, FCR is often cursed for its space consumption rather than embraced for its interpretive potential.
As a result, the archaeological literature offers comparatively few robust discussions of the variability in FCR and only a few address the nuanced perspective that FCR research can impart to archaeologists' understanding of human behavior (Birk 1994; Hester 1991; Wilson and DeLyria 1999). Yet where such studies have been undertaken, particularly in southcentral Texas and the Pacific Northwest, they add greatly to our knowledge of prehistoric cooking and food-processing technologies (e.g., Birk 1994; Dering 1999; Hunziker 2006; Jackson 1998; Kritzer 1995; McDowell-Loudan 1983; Wendt 1988; Wilson and DeLyria 1999). For example, by studying the forms and rate of decomposition of FCR, Wilson and DeLyria (1999:87) concluded that significant quantities of labor and expertise were required to manage the raw materials necessary for large · scale camas roasting. Their experiments also allowed for the identification of stages of FCR use, from new rock, to recycled rock, to abandoned stones fractured beyond use. Other researchers such as Birk (1994) and Wendt (1988) have shown that preceramic populations had clear preferences for the rocks they chose as heating elements; understanding the selection criteria of FCR not only demonstrates that prehistoric cooks had a thorough knowledge of the performance characteristics of the rocks they used but also that stone cooking technologies may have required more foresight than initially meets the eye. As Wilson and DeLyria (1999:81) note, the "selection and management of rocks . . . were not trivial concerns for prehistoric households."
FCR is typically identified by one or more thermal alteration features, including irregular fracture surfaces, potlidding, and reddening (House and Smith 1995; Wilson and DeLyria 1999). These attributes have been well documented for igneous and metamorphic cobbles (e.g., quartzite), which frequently exhibit all three (House and Smith 1995; Rapp 1999). As such, they have become "iconic" characteristics of FCR. However, as this paper will show, sedimentary rocks such as sandstone and limestone do not necessarily display these same physical attributes (Bearden and Gallagher 1980; Rapp et al. 1999). Experimentation may perhaps be the most useful method for identifying and investigating FCR, for it has been shown that heated rocks may vary widely in appearance from region to region, depending on the raw material used (Wilson and DeLyria 1999). …