Rockshelters, Rockart and Grinding Activity: A Preliminary Assessment of Relationships in Picket Wire Canyonlands, Comanche National Grasslands

The Pinon Canyon Maneuver Site (PCMS) has undergone extensive inventory over the last two decades. Although several thousand Native American sites have been recorded within the PCMS boundaries there are 662 that can be used for this analysis. In these cases, the observations were recorded in a consistent fashion. This limited database contains both nominal and categorical variables or fields. Within this data set there are 171 sites that possess observations with at least one of the following variables: rochshel, rockwarc, groundstone, rockart, or architecture.
We asked one fundamental question of the PCMS data set: To what extent is the presence of prehistoric grinding or milling activity, rock-art, and architectural remains spatially associated with the human use of a rockshelter? Analysis suggests that rockshelters, both modified and unmodified, show a weak association with evidence of grinding activities (Table 1). The presence of rock art is much less associated with modified rockshelters and devoid of any demonstrable association with unmodified rockshelters. Log-linear analysis (binary logit SYSTAT v.10.2) indicates however that the presence of grinding activities is the best predictor of rockshelter occupation (Table 2). That is, controlling for rock art, the probability of groundstone being present at an unmodified rockshelter is 16.5 times greater than evidence of these remains being absent. Similarly, at modified rockshelters, the probability of groundstone being present at the site is 5.75 times greater than its presence not being observed.