Pacific Obsidian Sourcing by Portable XRF

Article excerpt


The use of portable XRF analysis to characterise the geochemistry of Pacific obsidians is reported. Obsidian source samples from New Britain, the Admiralties, Fergusson Island and the Banks Islands were successfully characterised and then used to source 966 samples of obsidian from three (SE=SZ-8, SE-RF-2, SE-RF-6) Reef/Santa Cruz Lapita sites. The analysis demonstrates the capabilities of portable XRF and confirms the previous analysis of this material by combined density and PIXE-PIGME techniques. Our conclusions also indicate the importance of the Bao source region at Talasea and support the earlier suggestion that during the Lapita period an important passage crossed the Willaumez Peninsula in the vicinity of the Talasea airstrip and the Bao sources.

Keywords: Obsidian, Lapita, geochemistry, XRF


The earliest settlers of the Western Pacific carried along with them distinctive elaborately decorated pottery and obsidian from their immediate homeland region in the Bismarck Archipelago, where obsidian had been widely used since the Pleistocene (Fredericksen 1997; Summerhayes et al. 1998; Torrence et al. 2004; Torrence et al. 1992). The makers of this Lapita archaeological tradition moved quickly out into the Pacific at circa 3200 BP, transporting significant quantities of obsidian (Specht 2002) southeast over 2000 km across the first major water gap into Remote Oceania to the Reef/Santa Cruz Group (Green 1987; Sheppard 1993) and then on to Vanuatu (Galipaud and Swete-Kelly 2007). Small amounts of what was probably heirloom material was ultimately transported as far away as New Caledonia (Sand and Sheppard 2000) and Fiji (Best 1987), a straight line distance of over 4000 km.

Characterisation and sourcing of this obsidian has played a significant role in the examination of origins of Lapita cultural patterns, changing patterns of regional interaction in the Bismarck Archipelago, long-distance transport and interaction in the Lapita period, and chronology of resource use. Characterisation research has been popular since the late 1960s (Ambrose and Green 1972; Green 1998; Key 1968) and has made use of a wide variety of analytical methods, including emission spectroscopy (Key 1968), WD-XRF (Smith et al. 1977), INAA (Leach and Warren 1981), density (Green 1987; Torrence and Victor 1995), PIXE-PIGME (Bird et al. 1981; Bird et al. 1997; Summerhayes et al. 1998) and most recently ICP-MS (Ambrose et al. 2009) and Raman Spectroscopy (Carter et al. 2009). Since the development in the 1980s of the use of PIXE-PIGME for the non-destructive characterisation of obsidian by Roger Bird at the Lucas Heights Research Facility in Sydney (ANSTO), most geochemical characterisation of Pacific obsidians has been conducted using that technique.

Although the PIXE-PIGME at Lucas Heights has produced quality results, access is limited and costly if you wish to analyse many hundreds of samples and there are sample size limitations. As with all archaeological geochemical characterisation, there is a need for a cheap, fast, non-destructive and portable system which allows adequate source characterisation and discrimination. Developments in recent years in the area of portable energy dispersive XRF systems have made archaeologically-effective systems available at reasonable cost (Craig et al. 2007). In this paper we report our use of an Innov-X portable XRF system to characterise the Lapita obsidian sources and then source the obsidians recovered from the Reef/Santa Cruz Lapita sites (SE-RF-6, SE-RF-2, SE-SZ-8) excavated by Roger Green (1976; Sheppard 1993). Our work demonstrates the effectiveness of this system in rapidly and effectively characterising many hundreds of samples with a system based in our archaeological laboratories at the University of Auckland.

Obsidian sources

All reported source regions and sources used in the Lapita period are studied in this report (Appendix), with the exception of a source in Tonga which was used only by the Lapita settlers of Tonga. …