African iron working has been the subject of Western ethnographies for well over a century. The earliest work was often undertaken by missionaries and colonial administrators. The aspect of this industry which provoked the most interest amongst Westerners was the role of magic and symbolism. The often tacit assumption was that African iron working was 'held back' by the ritual surrounding production. This formed part of a wide-ranging 'project' to portray Black African achievements as inferior compared to those of the West. Africans were assumed to be intellectually inferior, and the 'superstition' surrounding iron working was seen as a clear example of this. Some of the early ethnographies were rather perfunctory, with most emphasis on interviews rather than actual observations of smelting and smithing practice.

As traditional African iron working (in particular smelting) declined economically in the twentieth century, so ethnographies became less frequent. The topic has undergone something of a renaissance in the last few decades, however, due in part to the emergence of archaeometallurgy. In particular, the archaeological and scientific study of early iron working in Europe and elsewhere has benefited greatly from an examination of recent practice in Africa. Almost all iron production in Europe before the Middle Ages was carried out using the bloomery (or direct) process in which the iron was obtained from the ore as a solid (rather than as a liquid, as is modern practice). In Europe, the bloomery process declined after the development of the blast furnace. By the later twentieth century there were some metallurgists who doubted that iron could be regularly obtained by a direct process. Tylecote's (1965) survey of some of the early ethnographic accounts showed that the technique was widely used in Africa in the nineteenth century. The furnaces, metal and waste products described were all clearly produced by the direct process. Tylecote and his contemporaries, however, showed little or no interest in rituals associated with iron working. Attention was directed exclusively on technological aspects.

Archaeological interest in metal working also focused on the social organisation of metal working and the social standing of the smith. This derived in large part from Childe's influential model of European Bronze Age society (Childe 1942). Childe suggested that smiths in Bronze Age Europe were mobile; they had few social ties and so could travel between different social groups, selling their wares across wide areas. The idea that smiths could operate 'outside' society in such a way received a substantial blow with the publication of Rowlands (1971) survey of the social standing of smiths in African societies (taken from a survey of the then available literature).

The last two decades have seen an enormous increase in interest in traditional iron working in Africa. This recent interest differs from that seen previously in the ways that the research is carried out and how the subject is conceptualised. Despite the decline of traditional African iron working a number of researchers (e.g. Avery, Barndon, Celis, Childs, David, de Barros, de Maret, Echard, Fowler, Goucher, Herbert, Killick, McNaughton, Schmidt, and van der Merwe) have successfully persuaded indigenous groups to carry out iron smelting and smithing (in many different locations across sub-Saharan Africa). In many cases, smelting ceased only a generation or two ago, but by drawing on the knowledge of those who smelted in their youth, as well as the knowledge of younger smiths, it has been possible to smelt ore and produce bloomery iron and steel according to traditional methods. The core of recent field work on traditional iron smelting and smithing has been the direct observation of the processes. This has often been enriched by the co-operation of specialists from many different academic fields (especially ethnography and metallurgy). This ethnography is extremely important as much of the detailed technological and ritual knowledge of traditional African iron working (especially smelting) is dying out with its last practitioners. In a sense, this is 'rescue ethnography'. The ethnography which is carried out is sophisticated, however, and does not attempt to reconstruct iron working as a 'timeless' practice free from outside (especially Western) influence. Many of the recent ethnographic surveys of iron working have explicitly addressed the changes in metal working forced by changing social and economic conditions in recent centuries.

The Culture and Technology of African Iron Production (hereafter Culture and Technology) contains a series of papers on African iron working: some are taken from a conference in 1988, while other are reprints or updates of important papers published elsewhere. Schmidt's introduction places the various papers within a wider context which ranges over changes in how iron working in Africa has been conceptualised by Western scholars, including the role of technology and ritual. De Maret and Thiery review the available evidence for 'How old is the Iron Age in Central Africa?'. The increased excavation of archaeological sites in Africa and the availability of radiocarbon dating has meant that data relevant to this question have burgeoned in recent years. When very little data were available the answer to this question was straight forward. As more data become available the picture is (at least initially) less clear. It seems that iron working was underway in the Gulf of Guinea and the interlacustrine area by the middle of the first millennium BC. There are a few earlier dates (second millennium BC) but these are regarded sceptically at the moment. The next three papers describe iron smelting in two different African societies. In each case local people are persuaded to reconstruct iron smelting technologies which have been abandoned in the last generation or two. Such rescue ethnography may involve imperfectly remembered practices, but it is still important, as it may be impossible to reconstruct in a few generations. Goucher and Herbert present a study of Bassari smelting which was on a sufficient scale to be described as 'proto-industrial' on the eve of colonial conquest. Research on Bassari iron smelting has been relatively intense (e.g. two recent University of California at Los Angeles doctorates). The technology and rituals of iron smelting amongst the Bassari were adapted to achieve high levels of production (e.g. natural draught and a relaxation of many of the taboos against female participation). Barndon describes how the part-time iron smelters of Ufipa in western Tanzania used a complex technology involving tall forced-draught furnaces for the initial reduction, with a much smaller furnace for bloom purification and consolidation. Barndon pays particular attention to how the technological solutions to Fipa iron smelting are embedded within their culture. The explanations of the symbolic power of rituals draw on associations and relationships found in other contexts of the society. Schmidt's chapter on the Barongo provides a fascinating study of a group who seem to have emerged as a result of social and economic upheaval in the wake of the slave trade. The Barongo survived as iron smelters in western Tanzania longer than many other smelting groups because of the abundant local resources (especially ore and wood) and their remoteness from imported scrap iron and steel. Schmidt argues that many aspects of Barongo iron smelting technology and ritual are drawn from a variety of different sources, reflecting the varied origins of the Barongo themselves. Schmidt employs the concept of bricolage to help explain how Barongo iron smelters select solutions from a wide range of technological and ritual options (a theme which is explored in more depth in Iron Working in East Africa).

The chapter by David and Robertson offers a welcome examination of iron smithing rather than smelting. The ways in which Montagnard and Muslim smiths in northern Cameroon have responded to changes in the economic landscape (in particular the introduction of cheap imported scrap iron and steel) are explored in great detail. Childs and Dewey examine the ways in which iron was smithed in ancient and modern Zaire and Zimbabwe. Blooms were extensively hammered to produce a series of distinctively shaped axes, some of which were utilitarian while others were invested with considerable symbolic meaning (in particular political power).

A series of papers follows which discusses the extent to which a distinctive and innovative iron smelting technology developed in north-western Tanzania. The idea that placing the greater part of tuyères inside the furnace allowed the air introduced to the furnace to be pre-heated (and so increasing the temperature reached inside the furnace) was first put forward by Schmidt and Avery in a 1978 article in Science. This is reprinted here with a few extra clarifications. This is followed by an updated version of a 1985 article by Schmidt and Childs on the excavation of early iron working sites along the western shores of Lake Victoria (originally published in African Archaeological Review, 1985). A critique of the 'pre-heating hypothesis' by Rehder and a reply by Avery and Schmidt from the Journal of Field Archaeology (1985) are reprinted here as a single chapter.

This is followed by further criticisms by Killick and defence by Avery and Schmidt. Schmidt and Avery propose that a technologically advanced mode of iron smelting arose in north-western Tanzania in the first millennium BC. Archaeological evidence from the early Iron Age sites at Rugomora Mahe and Kemondo Bay indicate that iron smelting slags were formed at temperatures of at least 1350°-1400° C. This is at least 100° C higher than that implied for European bloomery furnaces. The appearance of the tuyères from these sites (reduced and vitrified on the outer surfaces) suggested that the greater part of the tuyères had been placed inside the furnace and thus allowed the air being forced in to be pre-heated. Pre-heating allowed higher temperatures to be attained, which ensured more efficient reduction with less use of fuel, as well as the reduction of relatively poor ores. A bloom was also excavated from one of the Kemondo Bay furnaces. It was placed in what must have been a ritual context in a small pit dug into the base of the furnace. Metallography showed that this bloom was of steel rather than iron. The pre-heating hypothesis put forward by Schmidt and Avery explained the nature of the used tuyères, how the slags were formed and why steel rather than iron was formed. Schmidt and Avery were aware that their findings had considerable ramifications for the way in which iron and steel production in Africa was viewed: they demonstrated that Africans were capable of considerable technological achievement.

Schmidt and Avery chose to test their pre-heating hypothesis in an authentic environment by observing traditional iron smelting in north-western Tanzania. By observing Haya iron smelting, Avery and Schmidt hoped that many technological aspects of iron smelting may have remained unchanged in the area over two millennia or more. The Haya proceeded to smelt with a furnace, into which were placed the tuyères. Avery and Schmidt measured high temperatures inside the furnace -- in excess of 1820° C at one point. Schmidt and Avery's pre-heating hypothesis is controversial; it overturns many pre-existing and deep-seated notions about the development (or lack of development) of African technology. Rehder's criticism of the pre-heating hypothesis is based primarily on a mathematical model using thermodynamic theory and data. Rehder argues that Haya pre-heating could amount to little more than an extra 10° C, which would have little or no effect on the slags or metal produced. Avery, Schmidt, Rehder, and Killick argue the pre-heating hypothesis backward and forward. It is clear that there are serious technical and practical impediments to the accurate measurement of air temperatures inside tuyères and furnaces and to the production of thermodynamic models of the chemical and physical reactions which take place. In particular, thermocouples placed inside tuyères may be heated more easily than the air flowing past them and so may indicate a higher degree of 'pre-heating' than is actually the case. Attempts to model pre-heating thermodynamically (both by Rehder and by Avery and Schmidt) are less than satisfactory as the valveless bellows lead to an unsteady and turbulent forced draught. On balance, the nature of the slag and the tuyères indicate that pre-heating did occur. Rehder's and Killick's criticisms deserve careful consideration but are not powerful enough to disprove the pre-heating hypothesis.

In the final chapter in the book, Childs presents the detailed scientific examination of ores, slag, and metal from ancient and recent iron smelting in north-western Tanzania. The study of the microstructure of the ores and slags provides support for Schmidt and Avery's interpretation of Haya and earlier smelting procedures. Childs also recognises that, while many blooms are made from steel (containing significant levels of phosphorous and carbon), finished artefacts are usually made from phosphoric iron. She suggests that the difficulties of working phosphoric steel led smiths to decarburise the blooms regularly (but not exclusively) before working.

The second book reviewed here, Iron Technology in East Africa; Symbolism, Science, and Archaeology, covers some of the same topics discussed in Culture and Technology, but the approach is somewhat different. I knew this book was rather special as soon as I got it; a quick flick through showed that it managed to do justice to complex features of both the technology and the symbolism of African iron working. This is exemplified by the use of clear photomicrographs and the use of terms such as bricolage to explain the wide variety of sources drawn upon in the development of ritual.

Iron Technology in East Africa represents several decades of research by Schmidt and others amongst the Haya of north-western Tanzania. The book provides a detailed account of the 'experimental ethnoarchaeology' conducted to determine how the Haya smelted iron and specifically tests the pre-heating hypothesis. This book is about much more as well. Schmidt had already used archaeological and ethnohistorical methods to reconstruct Iron Age settlement and history around Lake Victoria. In Iron Technology in East Africa, Schmidt draws on his considerable knowledge of the local cultures to place iron smelting in its specific milieu. In doing so, Schmidt provides a particularly 'thick' (Geertz 1973) account of iron smelting as a technological, social, and symbolic process.

The first three chapters introduce the subject and provide a discussion of Haya history and the history of Western study of African iron smelting. Schmidt is at pains to stress that he wants to use a critical methodology to 'deconstruct Western representations about African iron technology' (4). The fourth chapter 'Ethnoarchaeology and bricolage' provides a detailed ethnography of Haya iron smelting based on the re-enactment of iron smelting. The Haya had conducted iron smelting regularly until the 1950s, and Schmidt managed to persuade some elders who had smelted in their youth, as well as some younger iron smiths, to conduct smelts for him. Schmidt admits that the results were 'interactive performances conducted in an environment far removed from experience, routine, and memory' (12), but shows sufficient sympathy for technological, social, and symbolic aspects of the process to yield a detailed and reflexive account of iron smelting. Some of the difficulties which Schmidt had to negotiate illustrate how easily one might form a simplistic model of 'traditional' iron smelting. Since iron smelting had been abandoned more than a generation before, there was initially, at least, very little local interest in the project. The smelters were reticent about rituals associated with smelting, primarily because they were associated with traditional religions which were disapproved of by the local Christian church.

Schmidt's most distinctive insight into the Haya smelting process is the way in which the smelters had a variety of technological and ritual solutions, or 'protocols', which could be deployed when the process was not as successful as expected. Schmidt identifies the process whereby smelters chose some protocols and not others as bricolage (expanding on Lévi-Strauss and Lemonnier). The technological and the ritual protocols were often so closely bound together that they could not be separated easily. Altering ritual aspects of a smelt often altered the technology of the smelt. In this way, ritual may actually have had a liberating effect, as opposed to the traditional notion that African iron smelting (and technology in general) was held back by 'superstition'.

Following this are three chapters which provide a detailed scientific analysis of the smelting process, its waste materials and its primary product, as practised by the Haya and as seen from the excavation of early Iron Age sites in the area. This is based in large part on the metallographic work of Terry Childs with some input from Donald Avery. The construction and performance of the tuyères (seen as crucial in the 'pre-heating hypothesis') are discussed at some length.

Chapter eight draws on an analysis of the spatial organisation of Haya smelting and smithing to develop a 'middle-range theory' for the interpretation of archaeological features on early Iron Age smelting and smithing sites. Chapter nine explores structuring principals in Haya song and myth in order to aid in understanding the symbolism of iron and iron working. Iron has long been associated with power and fertility in Bantu-speaking Africa. By drawing on his already considerable knowledge of Haya history, Schmidt is able to construct a detailed analysis of the relationships between iron, iron production, human sexual reproduction, social reproduction, and social power. Such relationships have recently been explored by Herbert (1993), but Schmidt's account benefits from examining a restricted geographical and social space, and thus does not 'smooth out' regional differences, and attempts to see how symbolic relationships are manipulated over time, rather than construct some idealised ethnographic present. In chapter 10, Schmidt examines some of the more comprehensive ethnographic accounts of iron smelting from elsewhere in sub-Saharan Africa. This and the last chapter help to produce a coherent picture of the employment of technological and ritual protocols in smelting, which appear to have a considerable history in the region. In particular, ritual offerings are sometimes placed in a tiny pit within both Haya and early Iron Age smelting pits. The rituals associated with iron working can now be seen as an essential part of the search for technological solutions and as integrated into wider attempts to explain the relationships between people and their social and natural environment.

These two books together provide an excellent insight into African iron working. It is clear that the emerging picture of African iron working is a complex one. This is achieved by employing sympathetic ethnographic accounts based on actual practice and indigenous explanations of that practice, as well as rigorous scientific analysis of the practice and its products. Such approaches show that the technology employed was sophisticated and also place that technology within a wider symbolic and social context. We can also now see that African iron working is subject to significant regional and chronological variation. The scope of these two books is almost breath-taking. If they represent the current state of archaeometallurgy and in particular its relationship with ethnography, then the discipline has definitely come of age. I have no hesitation in recommending both of these books to anyone interested in archaeometallurgy, ethnography, the history of technology, African archaeology, and related disciplines.

Works cited
Childe, V.G. 1942. What Happen in History. Harmondsworth: Penguin Books.

Geertz, C. 1973. The Interpretation of Cultures. London: Fontana Press.

Herbert, E.W. 1993. Iron, Gender and Power. Rituals of Transformation in African Societies. Bloomington and Indianapolis: Indiana University Press.

Rowlands, M.J. 1971. The archaeological interpretation of prehistoric metalworking. World Archaeology, 3: 210-223.

Tylecote, R.F. 1965. Iron smelting in pre-industrial communities. Journal of the Iron and Steel Institute, 203: 340-348.

About the reviewer
Dr David Dungworth is a Teaching Fellow in Glass and Metals at the Department of Archaeology and Prehistory, University of Sheffield. He completed a degree in Ancient History and Archaeology at the University of Birmingham in 1991 and was awarded a PhD by the University of Durham in 1995 for research on Iron Age and Roman copper alloy use. This research has now been published online in issue 2 of Internet Archaeology. He may be reached by e. mail at the following address: <D.B.Dungworth@sheffield.ac.uk>.

Copyright © D. Dungworth 1998