Pastoralist iron production on the Laikipia Plateau, Kenya: wider implications for archaeometallurgical studies

Recent research conducted on the Laikipia Plateau in Kenya highlighted several shortfalls in the existing models for understanding and interpreting past iron production technologies. The research was undertaken in order to assess variability in past pastoralist iron production techniques – the first such archaeometallurgical exploration in the region. In all, seven furnaces and one iron production refuse area were excavated across two discrete sites separated by only 3 km: Mili Sita and Cattle Dip, both dating to the second half of the second millennium AD. The recovered archaeometallurgical materials – slag, furnace wall, and tuyeres – were analysed using a combination of optical microscopy, SEM-EDS and ED-XRF. The results revealed that the technologies in question were complex and sophisticated, and their reconstruction and interpretation required some special considerations. The exploitation of titania-rich black sand was observed to bear a major influence on the resulting slag compositions at both sites, and use of these specific materials resulted in consistently efficient smelting episodes, demonstrating the technical competence and skill of the past smelters. However, although the slag compositions and microstructures of all the samples were markedly similar between the two sites, striking stylistic variation was apparent in the furnace design. This raised questions not only about the organisation and identity of the smelters who were working in this area at this time, but also of the reliability of assuming that similar slag compositions – peculiar as these might be – reflect similar technological (or sociotechnical) systems (cf. Pfaffenberger, 1992 ). A further problem encountered was that the consequently ulvite-rich slag was comprised of five, not three, major constituents – titania and lime in addition to iron oxide, alumina and silica. This meant that the operating parameters of these smelting episodes could not be easily interpreted using the usual models, which tend to be based around the assumption of fayalitic slag. As such, this paper will discuss to what extent consistency in technological style, in a broad, sociocultural sense, can be inferred through slag analysis, as well as the need for the formulation of wider reaching models that encompass the full scope of technical variation in iron production.