Assemblage first: Using provenance methods to understand 38,000 years of ochre use at Gledswood Shelter 1, Woolgar Country (northwest Queensland), Australia

Like stone artefacts, ochres (Earth mineral pigments) are durable, surviving from deep time archaeological contexts across the globe, leaving lasting records of the lifeways of those people who gathered and used them. However, unlike stone tools, variation between ochres is not always obvious. Ochres that look the same in colour and texture may have been gathered from distinct or disparate locations. Scientific analyses (such as trace element chemistry) are therefore required to be able to differentiate ochre sources, providing insights into the interactions of past peoples with their landscapes and each other. To date, most ochre provenance investigations have extrapolated archaeological patterns from the physicochemical analysis of few artefacts. This is especially true in Australia. Here, we describe patterns of ochre use through the 38,000-year occupation sequence at Gledswood Shelter 1 in Woolgar Country (what is now northwest Queensland). Using an analytic mainstay of sourcing studies, neutron activation analysis, we were able to geochemically characterise all suitable ochre artefacts, analysing 61 % of the assemblage to define patterns in procurement and use from prior to and throughout the Last Glacial Maximum, up until the recent past. Our findings demonstrate that valuable, otherwise unattainable, archaeological insights are generated through the application of provenance methods to archaeological ochre assemblages, regardless of their comparison to known or potential raw source materials. Ochre procurement and use at Gledswood Shelter 1 are discussed in the context of models of Pleistocene human population dispersal and Holocene social reorganisation in semi-arid, tropical northern Australia. • We present a revised chronology for GS1, incorporating OSL age determinations and recalibrated radiocarbon ages. • Contra to longstanding models of Pleistocene refugia, ochre use at GS1 was more intense between ca 38 ka and 18 ka. • The same local ochres are found throughout the assemblage, with a more mixed proportion and variety in the Holocene. • The GS1 ochre record aligns with recent re-assessments of LGM paleoenvironments in northern Australia. • We interpret GS1 as a reliable campsite along routes for Pleistocene human occupation and population dispersal.