Late Holocene forager-fisher and pastoralist interactions along the Lake Victoria shores, Kenya: Perspectives from portable XRF of obsidian artifacts Ellery Frahm a,b,c, ⁎, Steven T. Goldstein d , Christian A. Tryon b a Yale Center for the Study of Ancient Pyrotechnology, Department of Anthropology, Yale University, 10 Sachem Street, New Haven, CT 06511, United States b Department of Anthropology, Harvard University, Peabody Museum, 11 Divinity Avenue, Cambridge, MA 02138, United States c Department of Anthropology, University of Minnesota, Humphrey Center #395, 301 19th Avenue S, Minneapolis, MN 55455, United States d Department of Anthropology, Washington University in St. Louis, CB 1114 - 1 Brookings Drive, St. Louis, MO 63130, United States abstract article info Article history: Received 2 June 2016 Received in revised form 15 December 2016 Accepted 7 January 2017 Available online xxxx The East African Rift system created one of the world's most obsidian-rich landscapes, where this volcanic glass has been used to make tools for nearly two million years. In Kenya alone, there are N 80 chemically distinct obsid- ians along a 800-km north-south transect. Recently Brown et al. (2013) published their Kenyan obsidian database assembled since the 1980s. Specifically, they report elemental data measured by EMPA, ICP-MS, and WDXRF, pro- viding a rich basis for future sourcing studies. Here we report our use of portable XRF (pXRF), calibrated specif- ically and directly to the database in Brown et al. (2013), to examine interactions between Later Stone Age forager-fishers and pastoralists near Lake Victoria. Regarding our calibration to the WDXRF and EMPA datasets of Brown et al. (2013), the elements of interest have very high correlations (R 2 = 0.96–0.99) to our pXRF values, which show, on average, only a 2–5% relative difference from the published values. Use of pXRF data specifically calibrated to the datasets from Brown et al. (2013) greatly expands the impact of their work over three decades to catalog and characterize a multitude of Kenyan obsidians. Our focus here is investigating social contacts and ex- change between late Holocene populations that included Kansyore forager-fishers and Elmenteitan pastoralists. Similarities and differences in their obsidian access provide new insights into long-term interactions between foragers and food producers in eastern Africa. We report new sourcing results for obsidian artifacts from six late Holocene rock shelters along the Winam Gulf of Lake Victoria. The patterns in obsidian access are consistent with changing interaction spheres that are relevant to understanding forager-fisher social identities and subsis- tence strategies during periods of economic and demographic change. © 2017 Elsevier Ltd. All rights reserved. Keywords: Later Stone Age Eastern Africa Lake Victoria Forager-pastoralist interactions Obsidian artifact sourcing Portable XRF (pXRF) Analytical cross-calibration 1. Introduction Volcanism of the East African Rift, where the African tectonic plate is splitting in two, created one of the world's most obsidian-rich land- scapes, where this volcanic glass has been used, perhaps continuously, to make tools for almost two million years (Leakey, 1971; Clark and Kurashina, 1981; Brandt et al., 1996; Brandt and Weedman, 1997; Piperno et al., 2009). Many dozens of obsidian sources lie between Eri- trea in the north and Tanzania in the south. There are, by some tallies (Brown et al., 2013), N 80 chemically distinct obsidians in Kenya alone. It is unsurprising, therefore, that archaeologists have long been interest- ed to trace the distribution of obsidians across the region (e.g., Merrick and Brown, 1984a, b; Merrick et al., 1988, 1994; Negash and Shackley, 2006; Coleman et al., 2008; Nash et al., 2011; Ndiema et al., 2011; Ambrose, 2012). Before the successes of chemical obsidian sourcing (Cann and Renfrew, 1964), researchers sought to match obsidian arti- facts to these sources using density, refractive index, and similar physi- cal properties (Lucas, 1942, 1947; Leakey, 1945). Surveys of the southern Kenya Rift (Bower et al., 1977) revealed a greater number of obsidians than anticipated, complicating the potential for sourcing re- search. To conduct obsidian sourcing in Kenya, for example, one must first sample the multitude of obsidian sources that occur along a 800- km stretch between Lake Turkana on the Ethiopian border and Lake Na- tron on the Tanzanian border (Fig. 1). Alternatively, one could utilize a coherent, published database of Kenyan obsidians – that is, if the mea- surements of obsidian artifacts are sufficiently compatible with an existing source database. Recently Brown et al. (2013) published their full Kenyan obsidian source database, assembled since the 1980s (Merrick and Brown, Journal of Archaeological Science: Reports 11 (2017) 717–742 ⁎ Corresponding author at: Yale Center for the Study of Ancient Pyrotechnology, Department of Anthropology, Yale University, 10 Sachem Street, New Haven, CT 06511, United States. E-mail addresses: elleryfrahm@gmail.com, frah0010@umn.edu (E. Frahm). http://dx.doi.org/10.1016/j.jasrep.2017.01.001 2352-409X/© 2017 Elsevier Ltd. All rights reserved. Contents lists available at ScienceDirect Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep