Magnetic Sourcing of Obsidian Artifacts: Successes and Limitations

The eruptive conditions of obsidian flows vary widely, thereby producing a variety of distinctive compositions, microstructures, microlite abundances, and textures. These characteristic properties are frequently used in provenance studies that seek to establish a link between an archaeological artifact and a specific volcanic source. Compositional and geochronologic data are the most commonly used measurements in successful obsidian provenance studies; however, most of these approaches are destructive, time-consuming, and costly, or lack accuracy. The pioneering study of McDougall et al. (1983) showed that measurements of obsidian’s magnetic properties are a nondestructive, quick, and inexpensive alternative approach that can correctly distinguish groups of obsidian artifacts. This study examines whether an artifact’s provenance can be established by comparing its magnetic properties to a set of geological reference samples. Here we report the results of a comprehensive study of obsidian artifacts and geologic samples from SW Asia. Artifacts from Tell Mozan, Syria (n = 98) were analyzed for natural remanent magnetization (NRM), low-field susceptibility, frequency dependence of susceptibility, saturation magnetization, saturation remanence, bulk coercivity, and coercivity of remanence. The magnetic properties of the artifacts were then compared to a database of >500 obsidian samples collected across the same geographic region. As a control, the geochemistry of the artifacts and all the geological samples were measured using an electron microprobe. Our results show that magnetic measurements can successfully separate artifacts into different groupings. However, the distributions of magnetic susceptibility and NRM for the artifacts were distinctly lower than those of the geologic samples. We interpret these differences as a sign that the selection criteria of Neolithic and Bronze Age toolmakers were different from those of modern geologists. The toolmakers sought high-quality obsidian (e.g., isotropic fracture, low abundances of tiny mineral inclusions) and repeatedly visited the same outcrops ("workshops") on a rhyolitic flow where such obsidian was accessible. In contrast, geologists were trying to obtain good geographic coverage of an obsidian flow to capture its full compositional variation. The geochemical properties of the obsidian samples are largely unaffected by this difference in selection criteria, but the magnetic properties of obsidian change dramatically over a single flow. For this reason, our efforts to link an artifact to a specific source were unsuccessful. In short, the magnetic properties of obsidian are too sensitive to be used to characterize a single flow. For magnetic sourcing to be effective, geologic reference samples must come solely from outcrops that served as workshops for past toolmakers. Magnetic measurements may still be useful in situations where (1) potential obsidian sources have similar or identical geochemical compositions, or (2) a researcher wishes to distinguish between specific outcrops within a single obsidian flow.