A miner's perspective on Pb isotope provenances in the Western and Central Mediterranean

The conventional approach to ore provenance studies of ancient silver coins and artifacts has been to first analyze and then try to match them to published data about mining districts, a difficult task given our incomplete knowledge of these. While literary sources are useful to identify possible provenances, they potentially bias interpretations proper because of a variety of limitations of their time. Archeological evidence in the form of mining shafts, galleries, spoil heaps, and tools also provides a tangible and reliable record of mining, but dating such activity can be problematic and the record is inconsistent. Here we propose a new approach driven by Pb isotopic data rather than numismatic groups. Statistical analysis of Pb isotopic data is used to identify ore-defined isotopic clusters. This new method is based on an algorithm that predicts the number of isotopic clusters necessary to fulfill the simple condition that variance within isotopic clusters is minimized whereas inter-cluster variance is maximized. Since each cluster reflects a discrete geological episode within a particular environment broadly datable to a specific Pb model age, it can be identified as a potential source exploited by ancient miners. We explore the potential of this method in two examples using data from coins and ores respectively. In the first example, Roman Republican silver coins form three ‘end-member’ clusters sourced in mining districts with Cenozoic, Mesozoic, and Paleozoic Pb model ages. The example demonstrates how sources of silver used to mint coinage of the Roman Republic shifted within 50 years of the end of the Second Punic War in 201 BCE. In the second example using Aegean galena samples, Pb isotopes distinguish components with model ages datable to the Hercynian basement, the recent Aegean tectonic province, and Cyprus, noting that significant silver mining districts may remain unidentified in either Spain or the Aegean world. We further clarify a number of potential analytical issues and advocate that users of Pb isotopes for tracing archeological artifacts measure all four lead isotopes and inspect the 12 proposed isotope combinations in order to select those that provide the best representation of the data. We also emphasize that full advantage should be taken of the geologically informed parameters (model age and Th/U/Pb relationships) to identify the geological context of metal sources.