How reliable are our published archaeometric analyses? Effects of analytical techniques through time on the elemental analysis of obsidians R.G.V. Hancock a, b, * , Tristan Carter b a Department of Medical Physics and Applied Radiation Sciences, McMaster University, Hamilton, Canada, Ontario L8S 4K1 b Department of Anthropology, McMaster University, Hamilton, Canada, Ontario L8S 4L9 article info Article history: Received 29 June 2009 Received in revised form 21 September 2009 Accepted 3 October 2009 Keywords: Analytical accuracy Analytical precision Obsidian sourcing Group percentage error Instrumental comparison Anatolia Cappadocia East Go ¨ llu ¨ Dag ˘ Nenezi Dag ˘ abstract To assess the analytical accuracies and precisions of archaeometric elemental analyses by different techniques, a relatively homogeneous material such as obsidian must be studied. An assessment of published elemental concentration data from two Anatolian obsidian sources shows that while in most cases analytical accuracy is as high as is commonly expected, in some cases it is not. It also shows that the dispersions of elemental concentration data (indicators of analytical precisions) coming from modern analytical procedures are akin to the estimated homogeneity of the obsidian. Based on this latter observation, if one has element dispersion data from a single analytical technique, with a single source of obsidian as a control, data sets that contain multiple, but similar sources of obsidian may be differentiated. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Readers of scientific papers tend to assume that elemental analyses are as accurate and precise as is possible. But, all such analyses are dependent on combinations of technique-, instru- ment-, and human-related factors that can lead occasionally to low quality data (e.g. Fluegel, 2009). To avoid data problems, scientists have produced standard reference materials that may be used to help guarantee the accu- racies of analyses from different laboratories (e.g. Glascock, 1999). This approach has allowed data generated by a variety of analytical techniques to be exchanged as seamlessly as possible. However, a potential problem with the analysis of ‘‘standard’’ reference materials is that much more care and effort may be applied to their analysis than to the analyses of the archaeological or geological materials that are the foci of these research projects. To assess the accuracies of elemental analysis data in the archaeometric literature, it is not appropriate to focus on materials such as ancient and historic glasses and metals that have minor and trace element homogeneities of around 10  20%, or more, much like medium to coarse-ware ceramics. For these materials it is difficult to separate the material sources of data dispersion from the analytical sources. But, if one compares data from a homogeneous material such as obsidian there is a better chance of confirming the accuracies and precisions of analyses, especially when a specific obsidian source, or flow, is studied. While obsidian, or more specifically the products of a single flow of obsidian, is usually described as a relatively homogeneous material (e.g. Pollard and Heron, 2008; Kilikoglou et al., 1997), the actual degree of homogeneity is usually not defined, although it is noted that the partitioning of trace elements between the glassy and crystalline phases may increase their heterogeneity (e.g. Pollard and Heron, 2008). This means that on a fine scale, as is the case for the distribution of Au particles in sediments or ceramics, the concentrations of some trace elements may be much more variable than expected, especially when small samples are analysed. The homogeneity of major, minor and trace elements in obsidian is probably in the vicinity of 1% through to 6  10% at most (M.D. Glascock, University of Missouri, personal communi- cation 2007). This level of homogeneity is confirmed by Hughes (1994) who reported results from the XRF analyses of 200 * Corresponding author. Department of Medical Physics and Applied Radiation Sciences, McMaster University, Hamilton, Ontario L8S 4K1, Canada. E-mail addresses: ronhancock@ca.inter.net (R.G.V. Hancock), stringy@mcmaster. ca (T. Carter). Contents lists available at ScienceDirect Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas 0305-4403/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2009.10.004 Journal of Archaeological Science 37 (2010) 243–250