Red Earth, Green Glass, and Compositional Data: A New Procedure for Solid-State Elemental Characterization, Source Discrimination, and Provenience Analysis of Ochres

Ochres are a diverse category of naturally occurring iron-enriched earths and rocks, as well as iron oxide minerals, that derive their color from iron-containing chromophores and are suitable for use as pigments. Over the last two decades, provenience studies of archaeological ochres have grown from a rarity largely of interest only to specialists to an accepted and expected part of the archaeological science panoply. The most effective approach to distinguishing among sources of ochre and assigning archaeological pigments to their origin is multi-elemental characterization or elemental fingerprinting. In this study, we coupled a sample preparation method not previously used in ochre archaeometry with elemental fingerprinting by Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS) and Electron Probe MicroAnalysis (EPMA). We present a procedure for lithium borate (LiBo) fusion of samples for solid-state analysis, optimized for use with ochres, and designed for the budget and laboratory equipment constraints faced by many professional and student archaeologists. This method development research is part of the broader project OLKARIA: Ochre Landscapes of Kenya - Anthropological Research and Iron-oxide Archaeometry, which seeks in part to characterize the elemental composition of all known geologic ochre sources in the Kenya Rift Valley. Using a subset of project OLKARIA samples prepared by LiBo fusion and measured with LA-ICP-MS and EPMA, we successfully distinguished among six geologic ochre sources and a sample of commercially available iron oxide pigment. Our ability to uphold the Provenience Postulate for this data set compared favorably with source discrimination analyses done using data from Neutron Activation Analysis (NAA) of whole ochre for the same samples. LiBo fusion presents potential solutions to some criticisms of solid-state analysis of ochre using beam techniques, including issues arising from mineralogical heterogeneity, variable surface topography, the presence of free and chemically bound water, and the lack of matrix-matched standard reference materials. We also address the challenges of applying compositional data analysis best practices to ochre with an emphasis on the issues of rounded zero replacement and multivariate normality and highlight the work that remains to be done in this area.