One of the important assumptions of compositional analysis is that the elemental composition of an artifct reJIects the source of the materials used to make it. Thus, pottery j o m a particular source will be chemically similar to the raw materials from that source. This '%ommonsense" assumption seem beyond dispute, but the fact that pottery is a mixture of clay, water, and often temper added by the potter, complicates the interpretation of compositional data j o m ceramics. This article examines the relationship between potters' behavior in obtaining and using raw materials, on the one hand, and the chemical composition of theirjnishedpottery, on the other, by comparing the elemental composition of ethnographic pottery and raw materials j o m contemporary pottery-making communities in the Valley of Guatemala. The results of this research show that the relationship between pottery and its constituent raw materials is not as obvious as was j r s t supposed. The article concludes with an alternative approach to compositional analysis that is more in line with the realities of real-world pottery production.EW ANALYTICAL APPROACHES HIGHLIGHT the problems of relating human behavior F to its material residues more than the compositional analysis of pottery. For archeologists, the problem of compositional analysis consists of relating the chemical composition of an artifact to the behavior of its producer. Unfortunately, the elemental composition of pottery (like that obtained from instrumental neutron activation analysis, inductively coupled plasma spectroscopy, or atomic absorption spectroscopy) is difficult to relate to potters' behavior because the relationship between the chemical elements in the pottery and the potters' behavior is not obvious.In order to relate the chemical elements in the pottery and potters' behavior, archeologists and "archeometrists" assume that the ceramic composition reflects the composition of the source raw materials (usually clay). This "commonsense" assumption works well with some natural materials unaltered by humans (e.g., obsidian), but the complex chemical and behavioral factors involved in pottery production can affect the simple relationship between the chemical elements in the pottery and the supposed source of its constituent raw materials. With ceramics, the concept of "source" is thus problematic because "source" can be thought of as a single mine, a single widespread clay stratum, all clays in a single drainage, a single community of potters, or perhaps even a group of such communities. One way to evaluate the assumption relating the elemental composition of pottery and behavior is to analyze materials from living communities of potters . mm L t Arnold, Neff, and Bishop] P O~E R Y "SOURCE" ANALYSIS 75 the region. As mentioned earlier, ceramic pastes will probably not match the composition Arnold, Neff, and Bishop] P O~E R Y "SOURCE" ANALYSIS 85