Abstract and Keywords
This article examines materials analysis in archaeology. It explains that the ultimate goal of materials analysis in archaeology is to understand the human exploitation of the environment in terms of raw materials and processes used in the production of artefacts from stone, ceramic, metal, and glass as well as from the full range of organic materials. This goal is achieved through the investigation of the overall life-cycle or chaîne opératoire of surviving artefacts starting with their production and continuing through their distribution to their use, reuse, and ultimate discard. This article provides a series of case studies relating to the production technology of pottery, metals, and glass and discusses potential of and the problems associated with distribution studies.
The ultimate goal of materials analysis in archaeology is to understand human exploitation of the environment in terms of the raw materials and processes used in the production of artefacts from stone, ceramic, metal, and glass as well as from the full range of organic materials. Thus, an overview of man's interaction with the material world and technological development can be obtained (Henderson 2000; Tite 2001).
This goal is achieved through the investigation of the overall life-cycle or chaîne opératoire of surviving artefacts starting with their production and continuing through their distribution to their use, reuse, and ultimate discard. Thus, the first stage consists of the reconstruction of production, distribution, and use. The materials-analysis input into the reconstruction involves the determination of mineralogical, chemical, and isotopic compositions of the artefacts together with their macro- and microstructures. The subsequent second stage is then concerned with the interpretation of this reconstructed life-cycle in order to provide a better understanding of the behaviour of the people who produced, distributed, and used the resulting artefacts. This involves attempting to answer questions of the how and why type relating to the discovery and adoption of a new technology, the choice of a particular production technology or pattern of trade and exchange, and the reasons for change.
(p. 211) Following brief discussions of the basic principles behind and the scientific methods employed in the reconstruction of artefact life-cycles, and the aims and approaches to the interpretation of production and distribution, a series of case studies relating to the production technology of pottery, metals, and glass will be presented. Finally, the potential of and the problems associated with distribution studies will be discussed.
The reconstruction of the production technology involves identifying the raw materials used, together with the tools, energy sources, and techniques employed in procuring and processing these raw materials, and in fabricating and decorating the artefacts. The materials-analysis input to the reconstruction clearly depends on whether one is considering artefacts produced merely through mechanical modification of the raw materials (e.g. wood, bone, stone, pigments) or pyro-technologies in which the raw materials undergo chemical and microstructural change through the action of heat (e.g. plaster, pottery, metals, glass).
Whenever feasible, the reconstruction should start with archaeological fieldwork and excavation to locate and investigate both the raw-material sources, such as the quarries and mines from which stone and metals ores were obtained, and the industrial sites associated with the production of, for example, pottery, metals and glass. Next, the artefacts together with appropriate production debris (e.g. furnace fragments, crucibles, ores, slags, frits) are subjected to scientific examination using the analytical and microscopy techniques outlined below. The data on chemical composition and microstructure thus obtained then provide the basis for elucidating the raw materials used, and the sequence of processes involved in the production of the artefact. In interpreting these data, the possible changes in composition and microstructure during burial must be borne in mind.
In the reconstruction of the distribution of artefacts from their production centres (i.e. provenance studies; Wilson and Pollard 2001), fieldwork in an attempt to locate all possible sources of the relevant raw material in the predicted production region is again a crucial first step. The second step is to group together those artefacts made from raw material (e.g. flint, marble, clay, copper, iron) from the same source on the basis of the ‘fingerprints’ provided by their mineralogy, minor and trace-element compositions, or stable isotope compositions. Next, one tries to identify the actual raw-material source used by comparison of the compositional ‘fingerprints’ for the artefact groups with those for the sources located by fieldwork. Finally, one must try to establish the distribution patterns, and thus the trade (p. 212) and exchange, of the artefacts away from the various raw-material sources and production centres.
The success of the comparison between raw-material sources and artefact groups depends, of course, on the between-source variation exceeding the within-source variation, as specified in the so-called ‘provenance postulate’. Also, one needs to take into account possible changes in the compositional ‘fingerprint’ of the raw materials both during production of the finished artefact, a problem that is particularly severe in the case of metals, and during subsequent burial.
In establishing the use, as well as the reuse, to which artefacts were put in the past, the first step is a careful assessment of the archaeological context in which they were found. Secondly, one studies any surface wear resulting from use of the artefact, and analyses any residues, principally organic but also inorganic, surviving on or within the body of the artefact (Heron and Evershed 1993). Thirdly, one tries assess whether the performance characteristics of an artefact, as determined by its physical properties, are consistent with the inferred use. For example, is a stone axe sufficiently tough to survive impact, or does a pottery vessel have sufficient thermal-shock resistance for use in cooking?
Finally, the validity of both the proposed reconstruction of production technology and the hypothesized use of an artefact needs to be tested by experimental replication, in the Weld or laboratory as appropriate, of the industrial processes (e.g. pottery-firing, copper-smelting), artefact fabrication (e.g. pottery-forming, glassworking), or artefact use.
Scientific methods for reconstruction
The scientific examination of artefacts for the reconstruction of production, distribution, and use (Pollard and Heron 1996; Ciberto and Spoto 2000) starts with studying macrostructure using, for example, low-power binocular microscopy and X-ray radiography. Examination under a binocular microscope can reveal surface markings and joins related to how the artefact was fabricated, as well as evidence of surface wear resulting from use. X-ray radiography, by looking below the surface, provides valuable supplementary information on fabrication, revealing features otherwise concealed by, for example, corrosion or other weathering products (Lang and Middleton 1997).
The second stage is the investigation of microstructure, and the identification of mineral and other phases present, using a combination of optical microscopy, in (p. 213) both transmitted and reflected light, scanning electron microscopy, and, to a very limited extent, transmission electron microscopy. By examining thin sections prepared from stone artefacts in transmitted light (i.e. optical petrography), the rock types used can be identified on the basis of the minerals present together with their relative proportions, size, shape, and arrangement. This approach can be extended to the identification of the mineral and rock fragments present in or added to the clay used in the production of pottery. Similarly, by examining polished sections prepared from metal artefacts in reflected light, the mechanical and thermal treatments to which the metal was subjected during fabrication can be inferred on the basis of the observed grain structure.
The examination of polished sections using scanning electron microscopy (SEM) provides extremely valuable supplementary information to that provided by optical microscopy. This is, in part, because the SEM provides a higher magnification but, perhaps, more importantly because, with the addition of X-ray spectrometers, the analytical SEM or electron microprobe can be used to determine quantitatively the chemical compositions of the different phases or components present. Thus, in the case of pottery, the higher magnification provided reveals the microstructure of the clay matrix on the basis of which the Wring temperature employed in the production of the pottery can be estimated. Further, the chemical composition of a glass, for example, can be determined away from regions of alteration resulting from weathering during burial. In contrast, transmission electron microscopy, which provides even higher magnification, has had limited application in the study of archaeological artefacts, in part, because of the difficulties of sample preparation and image interpretation.
Further information on the mineral phases present can be obtained using X-ray diffraction, Fourier transform infrared spectroscopy (FTIR), and Raman spectroscopy. Of these techniques, Raman spectroscopy, because it is non-destructive and can operate as a microscope, is especially valuable for the identification of pigments and colorants associated with paintings, manuscripts, glass, and glazes. For the identification of organic compounds, a range of chromatography techniques have been used, of which gas chromatography with mass spectrometry is currently the most important.
The third stage is the determination of the chemical composition of the artefact, for which a very wide range of analytical techniques are currently available. The factors determining the choice of method include the range of elements for analysis and their concentrations (i.e. major, minor, and trace), accuracy, sample size, sample throughput, and cost. Currently, the methods most frequently used are neutron activation analysis (particularly for pottery provenance studies), X-ray fluorescence spectroscopy, and inductively coupled plasma mass spectrometry (ICP-MS). Of these, ICP-MS probably has the greatest potential for the future. In addition to the wide range of elements that can be analysed, the very low detection limits, and the high sample throughput, its particular strength is the possibility of (p. 214) replacing acid dissolution of a sample removed from the artefact, by direct laser ablation of the surface. With this approach, damage to the artefact is minimal, and the potential inaccuracies arising from incomplete sample dissolution are avoided.
The final stage in the analysis of archaeological artefacts is the determination of their stable isotope compositions using mass spectrometry, with laser ablation now again being frequently employed as the method of sampling. Lead isotopes have been extensively used in an attempt to determine the geological sources of lead, silver, and copper ores exploited in antiquity. Carbon and oxygen isotopes have been similarly used in an attempt to determine the marble sources exploited in antiquity, and strontium is now starting to be used for glass-provenance studies.
In addition to scientific examination, it should be borne in mind that further help in the reconstruction of production technology and use can come from contemporary ancient writings and illustrations, and from ethnoarchaeological studies.
The effective interpretation of the life-cycle of surviving archaeological artefacts requires a holistic approach. This must take into account the fact that production, distribution, and use are firmly embedded within the overall situational context that includes environmental and technological constraints, the economic and subsistence base, the social and political organization, and the religious and belief systems of the people under consideration (Sillar and Tite 2000). Thus, one needs to consider, first, how the environmental, technological, economic, social, political, and ideological contexts impinge on (i.e. both constrain and drive) production, distribution, and use, and conversely, how these latter impact back on the contexts. Second, because of the strong interdependency between these different stages in the overall life-cycle, one needs to consider the interaction between production, distribution, and use.
A central problem of interpretation is the need to avoid ‘reading the present into the past’. Instead, one must try to adopt an emic approach in which one tries to understand or, at the very least, consider what the artisans thought they were achieving through their choice of a particular production sequence, and what the people using the artefacts thought were the properties of their artefacts. Therefore, ethnoarchaeological studies have a definite contribution to make by exposing us to other ways of thinking about the material world, and by reminding us that artefacts are used in creating and expressing social relationships and that rites, myths, and taboos can be associated with production, distribution, and use (Childs and Killick 1993; Barley 1994).
In the interpretation of production, the primary questions relate to technological discovery, adoption, choice, and change. In order to explain how a new technology (p. 215) was discovered, one needs to consider, first, how the necessary raw materials became available. Second, one looks for possible contacts with other technologies that could perhaps have provided the techniques necessary for production, and for the existence of a context in which either chance or deliberate experimentation could have occurred. In turn, to explain how these requirements for the discovery of a new technology might have been met, one needs to consider whether there was a change in the trade and exchange pattern that could have resulted in the acquisition of the new raw materials and/or the new techniques necessary for production. In addition, one looks for a change in the scale and organization of production (Costin 1991), and the status of the artisans that could have resulted in contact with other technologies as well the possibility for experimentation. Similarly, the adoption of a new technology is dependent on the demand for new materials or new types of artefact for distribution and use. Therefore, in explaining how these demands arose, one looks for the emergence of a new subsistence pattern, a new social elite, a new pattern of trade and exchange, or a new religion.
Again, in explaining technological choice and change (Schiffer and Skibo 1997), the factors that one needs to consider are, first, the availability and performance characteristics of the raw materials, tools, energy sources, and procurement, processing, fabrication, and decorative techniques used in production. Second, there are the more social and cultural influences such as the artisan's perception of the raw materials and techniques chosen (i.e. the technological style) and the ability these to express, for example, some aspect of social identity (Lemonnier 1986). Third, the pattern of trade and exchange can influence technological choice, for example, through providing or restricting access to raw materials, or by determining the scale and hence organization of production. Fourth, the intended uses of the artefacts can influence technological choice by determining the performance characteristics and physical properties that are required of the artefacts. In this context, it is important to avoid an overemphasis on the role of artefacts as purely utilitarian commodities, such as containers, tools, and weapons. Instead, one must also consider the social and ideological roles of artefacts in establishing bonds between social groups at all levels from families to states, or as exchange items and gifts used to accrue future benefits. Thus, the value of an artefact was not necessarily associated with its utilitarian performance characteristics. For example, at least in the early stages of metal production, the colour, reflectivity, ductility, or power of transformation via melting and recasting of metals could have been more important than their hardness and toughness. Alternatively, the value of an artefact could have been associated with its rarity, its past history, or even the ideological significance of the source from which the raw material was obtained.
In order to interpret artefact distribution patterns in terms of trade and exchange, it is important to quantify the trends in distribution away from the raw-material sources and centres of production. Thus, as discussed by Plog (1977), one needs to establish the scale or intensity, the directionality and symmetry, and the duration (p. 216) of the associated trade or exchange. Further, it should be emphasized that it is rarely satisfactory to consider the distribution of a single artefact type in isolation. Instead it is desirable, as far as is possible, to consider the full range of artefacts and subsistence products that are included in the trade and exchange system.
The fall-off in the quantity of a particular artefact type with increasing distance from source can, in principle, assist in distinguishing between the different modes of trade or exchange (Renfrew 1975). For example, the fall-off with distance from source will tend to be more rapid in the case of down-the-line exchange as compared to ‘middleman’ trading. Similarly, redistribution from a central place is likely to produce localized regions of higher concentration along the fall-off curve. However, in attempting to infer the mode of exchange from the fall-off data, it is necessary also to consider the means of transport available (i.e. land, river, or sea), since this can significantly affect the distance travelled by the artefact away from source. Further, artefacts used for social, political or ideological purposes will tend to have travelled greater distances than that used for utilitarian purposes. Therefore, it is rarely possible to identify unambiguously the mode of exchange or trade from the fall-off data.
The discovery or initial invention of pottery, which occurred independently in many parts of the world, can be readily explained in terms of the fire-hardened clay observed to result from the use of fires many thousands of years before the appearance of pottery. In contrast, the reasons for the adoption of pottery as a new material are more complex and tend to vary according to the part of the world under consideration (Barnett and Hoopes 1995; Rice 1999). Although the production of pottery was once thought to have been associated with the beginnings of agriculture, there are several instances where this is clearly not the case. For example, what is perhaps the earliest pottery was produced in Japan around 10,000 bc by Jomon culture fishermen. Similarly, in the Near East the introduction of pottery took place around 6000 bc, long after the inception of the Neolithic farming way of life, whereas in Mesolithic northern Europe pottery was utilized before the appearance of domesticates. Therefore, in considering the reasons for the introduction of pottery, one must take into account the use of pottery both for utilitarian purposes, such as the processing of food and the transport and storage of liquids and solids, and for social and ideological purposes, such as for display or gift exchange.
The reconstruction of the production technology for pottery involves determining what raw materials were used and how they were prepared, how the pottery (p. 217) was formed, how it was decorated, and how it was fired (Rice 1987; Tite 1999). In terms of raw materials, the primary question concerns the balance between the clay minerals that provide plasticity and the non-plastic inclusions (e.g. sand, flint, limestone, and shell) that reduce the shrinkage during drying, and the extent to which the latter were indigeneous to the clay or were added as temper. The possible methods of forming include modelling by hand, building up the vessel using coils or slabs of clay, and throwing on a wheel. A high-gloss surface finish that reduces the permeability of the pottery to liquids can be achieved either by burnishing the partially dried surface with a pebble, or by the application of a fine-textured clay slip. In addition, powdered mineral pigments can be used to decorate the surface prior to firing or the vessel can be subjected to post-firing sooting. The pottery is then fired at temperatures typically in the range 600–1000°C, either by open Wring, such as in a bonfire, or in a closed structure, such as a kiln (Gosselain 1992). With the former the rise in temperature is extremely rapid, the final temperature tends to be at the lower end of the above range, and it is not possible to control the Wring atmosphere. In contrast, with the latter the rise is temperature is much slower (hours rather than minutes), the final temperature tends to be at the upper end of the above range, and the atmosphere can be controlled to achieve either an oxidizing or reducing Wring.
In reconstructing pottery production, thin-section petrography provides information on the non-plastic inclusions, and the forming method used can be inferred from careful visual examination, supplemented by X-ray radiography. Examination of polished sections in a SEM provides a powerful method for studying surface treatments, and bonfire- and kiln-firings can be tentatively distinguished on the basis of the Wring temperature, as estimated from the microstructure observed in a SEM or from the mineral phases identified by X-ray diffraction.
With the introduction of wheel-throwing and kiln-firing in the Near East by the fourth millennium bc, the majority of the techniques required for the production of unglazed earthenware were known. Therefore, in the interpretation of pottery technology, the primary question is why were different technological choices made in different regions at different periods (Tite 1999)?
In answering this question, the first factor to consider is what raw materials were available and what were their properties. Availability of suitable clays depends on both the local environment and possible political control of the clay sources. The properties of the available clay will determine how it needs to be processed before use. If the clay is too sticky, then temper will need to be added, but if it is too stiff the clay will need to be refined or dung added (Rye 1976). A second factor to consider is the scale of production and the extent of craft specialization. Thus, for small-scale household production involving part-time potters and local consumption, coil-forming of vessels and open Wring, as in a bonfire, would normally be used. Further, because of the very fast heating rate associated with bonfire Wrings, the clays used must be coarse-textured, otherwise the steam resulting from the loss (p. 218) of water during Wring cannot escape and the pottery will crack. In contrast, for large-scale workshop production involving full-time potters and long-distance trade, wheel-throwing and the use of a kiln-firing are generally more appropriate. In this case, the clay used must be fine-textured, otherwise the potters risk severe abrasion to their hands.
A third factor to consider are the uses to which to the pottery was to be put (Braun 1983). Thus, when water is being stored it needs to be kept cool. This requires vessel walls with a high permeability so that the water can slowly leak to the surface, where its evaporation results in cooling. Therefore, water-storage vessels must be made from coarse-textured clay and their surfaces must be unsealed. In contrast, the walls of cooking pots must have low permeability otherwise heat is wasted in evaporating liquid that leaks to the surface. The inner surface of cooking pots must therefore be sealed, either by the application of a resin coating or by the prior cooking of glutinous food that penetrates and seals the surface pores. A further requirement of cooking pots is that they must be able to survive the rapid changes in temperature associated with their placing on or removal from a fire. To achieve the necessary high thermal-shock resistance, the clays used should again be coarse-textured, since the inclusions serve to prevent or at least slow down crack propagation. However, in this context it should be remembered that bonfire Wring involves very rapid changes in temperature and, therefore, a vessel that is sufficiently coarse-textured to survive such a Wring should also be suitable for cooking (Woods 1986).
A final factor to consider is the potter's and consumer's perception of the raw materials and processing techniques used in the production of the pottery (Barley 1994). For example, ethnoarchaeological studies have established that in many parts of Africa pottery production is seen as being related to birth, with the potter often also being the midwife. In this context, the addition of grog temper (i.e. ground-up pottery sherds) to a clay is seen as an act of ‘rebirth’ through which ‘a reversal of time’ is achieved. Similarly, it has been argued that the temper used in pottery production in the Neolithic Orkneys was collected from sites (chambers tombs, earlier settlements) that were of ancestral significance to the potters. Thus, residing in the choice of temper as well as in the other stages of pottery production, one can envisage a technological style, which is analogous to decorative style and is specific to each potter or group of potters.
The story of the beginnings of metal production is much more complicated than that for pottery (Muhly 1988; Tylecote 1991; Killick 2001). Currently, the (p. 219) earliest known use of metal seems to have been in south-eastern Anatolia at the late eighth-millennium bc Neolithic site of Cäyonü Tepe, where more than fifty artefacts, produced from native copper by a combination of hammering and annealing, have been found. However, the earliest metal to have been smelted from its ores was probably lead, a necklace of lead beads dating to the early sixth millennium bc being found at Catal Hüyük in south-central Anatolia. Smelted copper was probably first produced in the Near East during the fifth millennium bc. Then, during the fourth millennium, copper alloys start to be produced, together the beginnings of gold and silver production. Finally, during the second half of the second millennium bc, we see the beginnings of iron production, again in Anatolia and the Near East (Maddin 1987).
Copper and its alloys
The first stage in the reconstruction of the production technology for copper-based artefacts involves the investigation of the extraction and processing of the ores (Craddock 1995). Extraction involved the removal of the ore from mineral veins running through the host rock in either open quarries or mines. Antler picks and stone hammers would have been used to extract the ore, aided where appropriate by fire-setting which made the ore more friable. Processing involved crushing the ore to aid the removal of waste rock, roasting to convert copper carbonates and sulphides to oxides, and finally smelting, in which a mixture of ore and charcoal was heated to a temperature in the range 1000–1200°C in a forced draught. During smelting, carbon monoxide formed from the charcoal reduced the copper oxide to copper metal. For relatively pure ores a simple bowl furnace would have been used and minimal slag produced. In contrast, for impure ores a shaft furnace would have been used and a flux, in the form of iron oxide, would have been added to the ore-and-charcoal mix. The flux reacted with the waste rock to form a molten slag which, in the more sophisticated procedures, separated out from the molten copper and could have been removed from the base of the shaft furnace as smelting proceeded.
Reconstruction of the extraction and processing of the ore involves surveying and excavating the quarries or mines and the smelting sites. The industrial debris from these sites, such as unused ores and their waste products, furnace remains, tuyeres, and slags, are then examined in the laboratory. From the chemical compositions and microstructure of the furnace remains and the slags, the operating conditions (Wring temperature, time, and atmosphere) of the furnace can be estimated. A possible final step is to test the validity of the proposed reconstruction of the smelting process by attempting to produce copper metal in a replicate furnace built and fired in the laboratory or field.
(p. 220) The second stage in reconstruction involves the determination of the composition of the copper-based artefacts by chemical analysis of the bulk metal. For the earliest use of copper, an important question is whether the copper was native (i.e. copper metal found embedded in rock) or smelted from its ore. Although native copper tends to contain lower concentrations of impurities than smelted copper, this is not always true, and therefore an unambiguous answer to this question is often not possible. With the introduction of copper alloys, chemical analysis establishes with which other metals (e.g. arsenic, antimony, tin, lead, zinc) or combination of metals the copper was alloyed.
The final reconstruction stage involves the investigation of the processes used in the fabrication and decoration of copper-based artefacts. Other than in the early use of native copper, which would have been hammered into the required shape, the starting-point for the production of most copper-based artefacts would have been to melt the copper or its alloy in a crucible and then pour the molten metal into a mould. The mould, which could have been either open or closed, would have been made from clay, stone, or even metal itself. The resultant casting would have then been worked to its final shape by a sequence of hammering in a cold state followed by annealing at a few hundred degrees centigrade to remove the stresses, and resultant brittleness, that the hammering would have induced. For an artefact such as an axe, for which the casting was close to the final shape, the cold-working would have been minimal. However, for a bowl which would have typically been worked up from a cast disc of metal, repeated hammering and annealing would have been necessary. The primary method for the investigating the fabrication sequence is the study of the microstructure of the metal in a polished and etched section in reflected light under an optical microscope (Scott 1991). A dendritic or ‘tree-like’ microstructure is observed for an as-cast object. These dendrites are distorted by hammering, and are then replaced by equiaxial grains by subsequent annealing.
In addition to decoration by engraving or raising the surface, copper objects were often enhanced by applying a thin coating of precious metal (i.e. gold and silver) over the surface (La Niece and Craddock 1993). An early method of gilding was to hammer pure gold to form very thin leaf, which could have been applied directly to the surface of the copper object. Subsequently, during the Roman period, the method of mercury gilding, in which an amalgam of gold and mercury was applied to the surface and the whole object heated to expel the mercury, was invented. These two methods of gilding can usually be distinguished by a combination of surface examination to identify joins between adjacent sheets of gold leaf, and chemical analysis of the surface for traces of surviving mercury.
A fundamental question relating to the beginnings of copper metallurgy is whether it was invented in the Near East and from there diffused to all other parts of the Old World, or whether there was independent invention of metallurgy in several different parts of the Old World. On the basis of the chronology for its (p. 221) introduction, as provided by radiocarbon dating, independent invention of metallurgy, at least in the Balkans and the Iberian Peninsular, now seems possible (Renfrew 1969), and this hypothesis is reinforced by the very different smelting processes employed in these areas as compared to the Near East. However, the nature of the discovery whereby a hard rock was converted to a malleable metal is such that the possibility of some link, however indirect, between these different areas must remain. The case for the independent invention of copper metallurgy in China remains unresolved, because of uncertainties about the dating of the earliest metal objects. However, there seems little doubt that copper-smelting was independently invented in the New World in the mid-second millennium bc.
Whether or not copper metallurgy was independently invented in different parts of the Old World, it is clear that the reasons for its adoption, and its technological development, differed widely in different regions. In context of these different technological developments, a primary question relates to the choice of copper alloys used. In many parts of the Old World the sequence from pure copper, to arsenic- or antimony-rich copper, to bronze (i.e. an alloy of copper and typically about 10 per cent tin), to leaded bronze, and finally to brass (i.e. an alloy of copper and zinc) was followed (Craddock 1985; Northover 1988). An early explanation for this alloy sequence was that it reflected the layers within an ore deposit, from a superficial oxidized carbonate ore, to a sulphide ore enriched in copper, arsenic, and antimony, and finally to the primary sulphide deposit (Charles 1985). Although this ‘standard model’ is now seen as being a serious oversimplification both geologically and archaeologically, the type of ore locally available was clearly a factor in determining the choice of smelting method and alloy.
However, in attempting to explain the choice of alloys used, it is also necessary to consider the physical properties of the different alloys. Thus, arsenical copper (i.e. an alloy of copper plus a few per cent of arsenic) has a lower melting-point than pure copper and therefore is more easily cast. It is also harder and can be more readily further hardened by cold-working. Arsenical copper was therefore more useful for the production of tools and weapons. In this context, there is some evidence that higher arsenic alloys were chosen for producing knives, which needed to be hard and sharp, whereas lower arsenic alloys were chosen for axes, which needed to survive impact and therefore must not be too brittle. Finally, arsenical copper differs from pure copper in having a silvery colour.
The switch from arsenical copper to bronze could have, in part, reflected the exhaustion of the enriched sulphide ore. However, because tin would have been added as a separate component, the composition of the alloy could have been more readily controlled. Again, the use of tin results in a change to a more golden colour. Also, because tin is not volatile, its use was less hazardous for the ancient metal producers than exposure to the highly poisonous arsenic fumes. The addition of lead to a bronze further lowers the melting-point and also increases the mobility of the molten metal. Leaded bronze is therefore more easily cast. However, the (p. 222) addition of 2–3 per cent of lead is sufficient to achieve these improvements, whereas up to 30 per cent of lead was frequently added to bronzes. Since lead does not form a true alloy with copper, it is present as small globules and, at high concentrations, its addition would have seriously weakened the metal. Therefore, in these cases, lead which could have been obtained as a by-product of silver production was most probably added as a cheap substitute for copper and tin.
The initial adoption of copper metallurgy in the Near East and Europe was most probably as a high-status material for display and gift exchange. Thus, the early objects tended to be beads, pendants, and other ornaments rather than functional tools and weapons, and as a result, colour would have been an important parameter in the choice of alloys. However, subsequently the production of tools and weapons gained in importance, and therefore mechanical properties would have become a factor in determining alloy choice. In contrast, in the New World metal never replaced stone, bone, and pottery as the main material for the production of utilitarian objects. Therefore, colour continued throughout as a primary factor in the choice of alloy, with copper being extensively alloyed with gold and silver, as well as with arsenic and tin (Hosler 1994).
Other than the occasional use of meteoric iron, the first iron objects were probably produced as a by-product of copper-smelting in that, if an excess of iron oxide was added as a flux to facilitate the formation of slag, iron metal mixed with copper could have resulted. The subsequent production of metallic iron direct from iron ore would have involved a similar smelting process to that employed for the production of copper, the principal difference being that the furnace atmosphere needed to be more reducing for iron production. The product of iron smelting was a spongy mass of metallic iron, referred to as the bloom, in which slag prills and charcoal remained. Because the melting-point of bloomery iron (as opposed to the cast iron produced in China) is about 1500°C, casting of objects from molten iron was not possible in antiquity, and instead the blacksmithing technique of forging was employed. This involved reheating the metal to red heat and hammering it on an anvil while still hot.
The first step in the production of iron artefacts was to consolidate the bloom by extended forging, first into a billet and then into what are sometimes referred to as currency bars. The resultant iron, which typically contains less than 0.5 per cent of carbon, is soft but tough (i.e. not brittle), and is referred to as wrought iron. The wrought iron could have been converted (i.e. carburized) into steel, which has a carbon content of between 0.5–2 per cent, by heating it in the presence of charcoal. The result is a harder and therefore more useful metal. Further processes that were used to modify the mechanical properties of the steel were quenching, which (p. 223) involved rapidly cooling the metal from about 700°C and resulted in a very hard but brittle metal, and tempering, which involved reheating to 250–450°C in order to re-toughen the metal. As for copper artefacts, the main method for determining the processing to which an iron artefact has been subjected is to study its microstructure in polished section in reflected light, carburization, quenching, and tempering each resulting in a distinctive microstructure. However, the microscopic examination of iron objects tends to be more difficult than for copper because iron is much more susceptible to corrosion.
The primary questions in the context of the beginnings of iron metallurgy are why was iron first adopted as a new material, and subsequently, what factors determined whether artefacts continued to be made from bronze or were instead made from iron? In attempting to answer these questions, social, economic, and technological factors must each be considered (Haarer 2001).
The prestige of meteoric iron or ‘metal from heaven’ and subsequently iron produced as a by-product of copper-smelting was probably the initial inspiration for the deliberate production of iron that is thought to have first occurred in Anatolia or Iran during the second half of the second millennium bc. However, a further factor could have been a disruption in, or even exhaustion of, the supply of tin or copper. Because iron production was a new technology, requiring new skills that would initially have been in short supply, the ironworkers would have been under the control, and thus have contributed to the status, of the elite class. From the Near East iron production fairly rapidly spread across the Mediterranean into western Europe (recent excavations at Hartshill in Berkshire suggest that iron was being smelted in England as early as the tenth century BC), most probably by some process of diffusion rather than as a result of independent invention. Once the technology became more widespread, the wide availability of iron ores, as compared to copper ores, would have made the control of iron production by the elite class more difficult.
Furthermore, as iron became less of a high-status material and began to be more extensively used for the production of functional objects, economic factors and the mechanically properties of iron became more important in determining whether or not it was used in preference to bronze (Salter 1989). In addition to the wide availability of extensive iron-ore deposits, a further potential economic factor was that, because the percentage of iron in iron ores is significantly higher than that of copper in copper ores, iron production would have been more economical in the use of fuel. In terms of mechanical properties, iron was initially inferior in many ways to bronze, which is harder than wrought iron, comparable in hardness to steel (i.e. carburized iron), and, although less hard than quenched steel, is tougher (i.e. less brittle). However, a significant advantage of iron over bronze was the control that was possible over hardness and toughness by a combination of carburization, quenching, and tempering. In addition, by welding together alternate strips of wrought iron that provided toughness, and steel that provided hardness, a composite (p. 224) material combining both toughness and hardness was produced. This ‘piling’ process was also extended to produce elaborate pattern-welded swords, such as that found with the Anglo-Saxon ship burial at Sutton Hoo in Suffolk, which combined excellent mechanical properties with distinctive, status-enhancing decoration.
Vitreous materials (i.e. glasses, glazes, and enamels) are amorphous solids that differ from crystalline materials in that their constituent atoms form a lattice that lacks long-range order. As a result, glasses gradually soften rather than having a well-defined melting temperature, they are generally transparent, and their colour can be readily modified by the addition of different metal ions.
The first vitreous materials were glazed stone (mainly quartz and steatite), and faience, which consists of a ground quartz body coated with a glaze that initially was coloured blue by the addition of copper (Tite et al. 1998). Small glazed objects such as beads, scarabs, and seals were first produced in the Near East and Egypt during the fourth millennium bc. The discovery of the process for producing a blue glaze most probably occurred during the smelting of copper. Thus, a coloured glaze could have been formed on a sandstone furnace wall as a result of the reaction between the plant ash from the fuel, copper from the ore, and silica from the sandstone. The appeal and prestige of the early glazed objects was probably associated with their brightly coloured, smooth, and shiny surfaces that were similar in appearance to semi-precious stones such as turquoise and lapis lazuli. Subsequently, occasional small glass objects are found in both the Near East and Egypt dating from the late third millennium bc onwards. However, it was not until about 1500 bc that significant quantities of glass, including glass vessels, began to be produced. At about the same time the range of colorants used in both faience and glass was extended from the previously dominant copper to include also cobalt, calcium antimonate, and lead antimonate.
The primary question in the reconstruction of the production technology of vitreous materials is the identification of the raw materials used (Sayre and Smith 1961, 1974; Henderson 1989). The great majority of ancient glasses are based on silica (60–70 per cent) as the network former with the addition of an alkali flux (15–20 per cent) to reduce the softening-point of the glass, lime, and magnesia (5–10 per cent) as stabilizers that limit the solubility and hence susceptibility to weathering of the glass, and a range of different colorants (0.5–3 per cent). The two possible sources of silica are ground quartz pebbles, which contain very few (p. 225) impurities, and quartz sand, which can introduce significant amounts of lime and magnesia. The two possible sources of alkali are natron from the natural evaporitic deposit of Wadi Natrun in Egypt, and the ashes obtained from burning coastal and desert plants such as those of the Salsola genus. The former consists predominantly of sodium carbonate and bicarbonate, and again contains few impurities, whereas the plant ash can contribute significant amounts of soda, potash, magnesia, and lime. Rather than being added as a separate component, the lime and magnesia stabilizers were most probably introduced either with the quartz sand or with the plant ash. The colorants used were either associated with metal ions (copper for blue, cobalt for dark blue, iron for pale yellow or green, manganese for purple) incorporated in the glass lattice, or opacifiers (calcium antimonate for white, lead antimonate for yellow) present as small particles that are dispersed through the glass and scatter the light. Information on the raw materials used can generally be inferred from the bulk chemical composition of the glass, although there can be problems in establishing with which of the raw materials particular impurities are associated.
Beginnings of glass production
The initial discovery of glass most probably arose through poor compositional or temperature control (i.e. excess alkali or heat) during the production of faience (Peltenburg 1987). The question then arises as to why there was a delay of more than 2,000 years between the production of glazed stones and faience and that of glass. One possible explanation is that glazed stone and faience production involved cold-working (carving or moulding to shape), whereas the production of glass vessels that were formed round a clay-based core involved the manipulation of hot, viscous fluids, a process that was more akin to metalworking. It has therefore been argued that the transition to glass production would have required input from metalworkers, and that this input could have occurred through greater contact between faience- and metalworkers, resulting from the changing control over and organization of artisans following the political upheavals in the Near East and Egypt during the sixteenth century bc. Alternatively, the adoption of glass could have been driven by an increased demand for a substitute for semi-precious stones, such as lapis lazuli and turquoise, at a time when supplies from the Indus valley were reduced as a result of the decline in the Harappan civilization.
It is now generally accepted that glass was first produced in any significant quantity in Mesopotamia from about 1500 bc, using ground quartz pebbles as the source of silica, and a soda-rich plant ash as the source of both the flux and stabilizers (Shortland and Tite 2000; Tite et al. 2002). The glass industry was subsequently introduced into Egypt during the reign of Tuthmosis III (1479–1425 bc) through a combination of glass objects and ingots being imported as tribute, (p. 226) and the bringing back of captive Mesopotamian glassworkers. The question that follows is whether Egypt subsequently relied on imported raw materials, in the form of ingots and cullet, or whether glass was being produced in Egypt rather than merely worked.
A distinctive feature of early glass from Egypt, as well as contemporary faience, was the extensive use of a dark-blue cobalt colorant. These cobalt blue vitreous materials are characterized by higher alumina and magnesia contents than the equivalent copper blue materials, together with the trace amounts of nickel, manganese, and zinc. On the basis of these data, it has been established that the most probably source of the cobalt colorant was the cobalt-bearing alums from the Dakhla and Kharga oases in the Western Desert of Egypt (Kaczmarczyk 1986). The discovery that the cobalt blue colorant in glass found in Egypt originated from an Egyptian source is very important, in that it strongly suggests that glass was being produced in Egypt as early as the mid-fifteenth century bc. The alternative explanation, that the cobalt-rich alum was exported to Mesopotamia and the glass thus produced was reimported to Egypt, seems less likely, especially in view of the fact that cobalt blue glasses are very much rarer in Mesopotamia.
From the mid-eleventh century bc onwards there appears to have been a decline in glass production in Egypt, which continued through until about the fourth century bc. Furthermore, when glass production was reactivated in the first millennium bc, the focus of this production appears to have been the eastern Mediterranean, and at the same time there had been a major change in the raw materials used. Quartz sand had replaced ground quartz pebbles as the source of the silica, as well as providing the lime stabilizer, and natron had replaced plant ash as the source of alkali; and it is this combination of raw materials that was used in the production of Roman, Byzantine, and Islamic glass through until the ninth century ad.
Roman, Byzantine, and Islamic glass production
The major technological development in glass production achieved by the Romans was the introduction of glass-blowing (Israeli 1991). As a result, the scale of glass production increased dramatically, and glass changed from being a luxury item to one of common everyday use.
An important question relating to Roman glass is that of where this glass was produced. The uniformity in composition of Roman glass suggests that there was only a limited number of production centres (Freestone et al. 2000, 2002). The source of the natron flux was almost certainly the Wadi Natrun in Egypt, and evidence for glass production, in the form of massive tank furnaces measuring up to 4m × 2m × 0.75m (Gorin-Rosen 2000), is at present confined to Egypt and the Levant. Therefore, one hypothesis is that all Roman glass was produced in Egypt and the Levant, where access to the natron would have been relatively easy, and that (p. 227) the resulting raw glass was broken up and exported to glassworking sites throughout the Roman Empire. One problem with this hypothesis is that the furnaces so far located date only from the fourth century ad onwards. However, Pliny does refer to the earlier use of sand from the Belus river in the Levant for glass production. Also, the alternative hypothesis would involve the transport of natron across the Mediterranean to Italy, something which, because of its solubility, would be more hazardous that its transport within Egypt or to the Levant.
Byzantine and early Islamic glass continued to be produced using a combination of quartz sand and natron. However, in the Levant in the seventh century ad there appears to have been a reduction in the soda contents, from 15–16 per cent in Byzantine glass to around 12 per cent in the early Islamic glass, suggesting that natron was in short supply (Freestone et al. 2000). The pressure on the supply of natron seems to have culminated in the ninth century ad, when natron ceased to be used as the flux in glass production in the Islamic Near East. The Islamic world then reverted to the use of a soda-rich plant ash as the flux for glass production, and at more or less the same time, a potash-rich plant ash started to be used in western Europe. The question arises as to whether the demise in the use of natron was due to environmental or political factors. For example, a period of reduced temperature and/or increased rainfall could have decreased the rate of evaporation from the Wadi Natrun lakes so that less natron was precipitated. However, there are no records of obvious climatic changes during the ninth century ad. In contrast, documentary evidence indicates that there was very considerable political disruption within the Delta and adjacent regions of northern Egypt throughout the period from the seventh to the ninth century ad, which could have made both access to the lakes for the collection of natron and the subsequent distribution of natron more difficult (Whitehouse 2002).
The potential of, and problems associated with, distribution or, as they are frequently referred to, provenance studies depends very much on the material under consideration.
Stone is perhaps the ideal material for distribution studies, in that stone artefacts were produced entirely by mechanical modification so that there was no change in (p. 228) the compositional ‘fingerprint’ between the source and the finished artefact. Furthermore, stone sources in the form of quarries or mines (for flint) can be readily located, to provide control material against which the artefacts can be compared.
Because of their distinctive mineralogy, the difference sources of igneous and metamorphic rocks can frequently be distinguished using thin-section petrography. For example, an extensive study of Neolithic stone axes from Britain has distinguished some six major axe groups, with the actual ‘factories’ at which the axes were produced being identified at Penmaenmawr in North Wales and Langdale in Cumbria (Bradley and Edmonds 1993). In the case of the Penmaenmawr source, the concentration of its axes falls off progressively with distance from the source, suggesting down-the-line exchange. In contrast, for the Langdale source the highest concentration of axes is in east Lincolnshire, from where there is again a progressive fall-off in concentration. One interpretation of this pattern is that axe material was collected from Langdale and taken in bulk to Lincolnshire, from where it was redistributed.
In the case of flint, which is a very fine-textured cryptocrystalline quartz, or obsidian, which is a volcanic glass, thin-section petrography is of only limited use, and instead the different sources must be distinguished on the basis of their minor- and trace-element compositions. An early classic study of obsidian found at Early Neolithic sites in the Near East showed that obsidian sources in central Anatolia supplied the Levant and those in Armenia supplied Mesopotamia, and that the fall-off in the concentration of obsidian artefacts suggested down-the-line exchange over distances of more than a thousand kilometres (Renfrew, Dixon, and Cann 1966).
Compared to stone, pottery distribution studies are more difficult, since the clays used will normally have been processed either by refining or by the addition of temper. Furthermore, there are very large numbers of possible clay sources, and those exploited in antiquity can only very rarely be located. Therefore, control material, against which pottery from archaeological sites can be compared, is only available when pottery production sites with kilns and associated debris have been found.
Thin-section petrography can again be used to identify the non-plastic inclusions in coarse-textured pottery, and in the case of igneous and metamorphic rock fragments, the actual source of the clay, if the inclusions are indigeneous, or the temper, if the inclusions are added, can sometimes be identified. For example, it has been shown that a high proportion of Neolithic, Bronze, and Iron Age pottery found in Cornwall was produced using a clay containing the constituent minerals of the igneous rock, gabbro, which is only found on the (p. 229) Lizard peninsula (Peacock 1969). Detailed fieldwork has further established that the clay for the most abundant petrographic group of the gabbroic pottery was obtained from a small area on the Lizard, a kilometre or so across (Harrad 2004). The continued use of this particular clay over such an extended period, even though other suitable clays were available in Cornwall, suggests that the Lizard source acquired some form of sacred and/or ancestral significance.
For fine-textured pottery, the different clay sources are normally distinguished on the basis of their minor- and trace-element compositions, and it is this approach that has dominated pottery distribution studies. Overall, there have been many hundreds of publications detailing such studies, spanning all parts of the world and all periods from the Neolithic up to more or less the present day, and revealing trade and exchange over distances ranging from a few to several thousands of kilometres.
Since no evidence of the original ore mineralogy survives in metal artefacts, distribution studies for metals start with the comparison of the minor- and trace-element compositions of the artefacts and the possible ore sources. However, in making these comparisons there are the problems of loss of volatile components during smelting, and the partitioning of impurities between metal and slag. Also, alloys are composite materials, each component of which contributes to the minor- and trace-element composition. Furthermore, metals are frequently re-melted and reused, with the possibility of mixing together metal from more than one original ore source. Therefore, where applicable, lead isotope analysis is favoured over minor- and trace-element analysis, since although the concentration of lead in the finished artefact is dependent on the method of smelting, the ratios of the four stable lead isotopes (Pb204, Pb206, Pb207, and Pb208) are unchanged from those in the ore. The problem of reuse and possible mixing of metal from more than one source remains. However, compared to clay, non-ferrous metal ore sources are limited in number, and the associated mines and/or smelting debris are readily located.
In addition to lead metal itself, lead isotope analysis can be applied to silver which is produced from the lead sulphide, galena, and copper, the ores of which contain trace amounts of lead. An important application of lead isotope analysis has been the investigation of Late Bronze Age oxhide ingots that have been found throughout the Mediterranean from the Levant in the east to Sardinia in the west, as well as in large numbers on the Uluburun and Cape Gelidonya shipwrecks off the Anatolian coast (Gale and Stos-Gale 1999; Stos-Gale 2000). The analyses show that ingots found on Cyprus, the Greek mainland, Sardinia, and on the two shipwrecks were almost certainly made from Cypriot copper. However, copper-based artefacts found on the Greek mainland were made from copper from Laurion in Attica, and those (p. 230) found on Sardinia were made from local Sardinian copper. Therefore, although the oxhide ingots represent a major component in the trade in copper, their use in the production of artefacts was more limited. To understand this, one must remember that, on the evidence from the shipwrecks, oxhide ingots were just one component of traded goods that included tin, glass, ebony, and ivory, as well as pottery vessels containing resins, spices, and foodstuff. Therefore, it is possible that the ingots were used as a standard of value, providing the means of balancing a complex network of exchange, and were seen as a commodity for conspicuous consumption rather than for use in the production of artefacts.
As in the case of metals, glass is again a composite material which can be re-melted and reused. However, at least in the beginnings of glass production, the number of production centres was limited, and each centre or region tended to maintain its own tradition in the choice of raw materials. Thus, the colorant for cobalt blue glass produced in Egypt during the second half of the second millennium bc was cobalt-rich alum, so that such glasses can be readily identified by their high alumina, nickel, manganese, and zinc contents. Using these criteria, it has been possible to show that the cobalt blue glass ingots found in the Uluburun shipwreck originated in Egypt, and that the cobalt blue glass beads and plaques from Mycenean Greece were most probably produced locally using glass ingots imported from Egypt (Tite et al. 2005).
Similarly, the flux used in the production of glass in the eastern Mediterranean during the first millennium bc was the natural evaporite natron, and as a result such glass can be distinguished by its low potash content. Therefore, on the basis of their low potash contents, glass beads found in France dating to between the ninth and the seventh centuries bc are thought to have been imported from the eastern Mediterranean (Gratuze and Picon 2005). Further evidence for the eastern Mediterranean origin of these beads is provided by the fact that those beads coloured by cobalt have the high alumina, nickel, manganese, and zinc contents characteristic of the use of cobalt-rich alum colorant from Egypt.
Barley, N. (1994). Smashing Pots: Feats of Clay from Africa. London: British Museum Press.Find this resource:
Barnett, W. K. and Hoopes, J. H., eds. (1995). The Emergence of Pottery. Washington, DC: Smithsonian Institution Press.Find this resource:
Bradley, R. and Edmonds, M. (1993). Interpreting the Axe Trade. Cambridge: Cambridge University Press.Find this resource:
(p. 231) Braun, D. (1983). Pots as tools. In J. A. Moore and A. S. Keene (eds.), Archaeological Hammers and Theories. New York: Academic Press: 107–34.Find this resource:
Charles, J. A. (1985). Origins of metallurgy In P. T. Craddock and M. J. Hughes (eds.), Furnaces and Smelting Technology in Antiquity. British Museum Occasional Paper, no. 48. London, 21–8.Find this resource:
Childs, S. T. and Killick, D. (1993). Indigenous African metallurgy: nature and culture. Annual Review of Anthropology, 22: 317–37.Find this resource:
Ciberto, E. and Spoto, G. eds. (2000). Modern Analytical Methods in Art and Archaeology. Chemical Analysis Monographs, vol. 155. New York: Wiley-Interscience.Find this resource:
Costin, C. L. (1991). Craft specialization: issues in defining, documenting, and explaining the organization of production. In M. B. Schiffer (ed.), Archaeological Method and Theory, Vol. 3. New York: Academic Press, 1–56.Find this resource:
Craddock, P. T. (1985). Three thousand years of copper alloys: from the Bronze Age to the Industrial Revolution. In P. A. England and L. van Zelst (eds.), Application of Science in Examination of Works of Art. Boston: Museum of Fine Arts, 59–67.Find this resource:
—— (1995) Early Metal Mining and Production. Edinburgh: Edinburgh University Press.Find this resource:
Freestone, I. C., Gorin-Rosen, Y., and Hughes, M. J. (2000). Primary glass from Israel and the production of glass in Late Antiquity and the early Islamic period. In M.-D. Nenna (ed.), La Route du verre. Lyon: Maison de l'Orient, TMO 33, 65–83.Find this resource:
—— Greenwood, R., and Gorin-Rosen, Y. (2002). Byzantine and early Islamic glassmaking in the eastern Mediterranean: production and distribution of primary glass. In G. Kordas (ed.), Hyalos Vitrum Glass. Athens: Glasnet Publications, 167–74.Find this resource:
Gale, N. H. and Stos-Gale, Z. A. (1999). Copper oxhide ingots and the Aegean metals trade: new perspectives. Aegaeum, 20: 267–77.Find this resource:
Gorin-Rosen, Y. (2000). The ancient glass industry in Israel. In M.-D. Nenna (ed.), La Route du verre. Lyon: Maison de l'Orient, TMO 33, 49–63.Find this resource:
Gosselain, O. P. (1992), The bonfire of enquiries—pottery Wring temperature in archaeology: what for? Journal of Archaeological Science, 19: 243–59.Find this resource:
Gratuze, B. and Picon, M. (2005). Utilisation par l'industrie verrieres des sels d'aluns des oasiségyptiennes au début du premier millénaire avant notre ère. In P. Borgard and M. Picon, J. P. Brun (eds.), Colloque International—L'alun de Méditerranée. Naples: Institut Français de Naples, 269–76.Find this resource:
Haarer, P. (2001). Problematising the transition from bronze to iron. In A. J. Shortland (ed.), The Social Context of Technological Change: Egypt and the Near East, 1650–1550 BC. Oxford: Oxbow Books, 255–73.Find this resource:
Harrad, L. (2004). Gabbroic clay sources in Cornwall: a petrographic study of prehistoric pottery and clay samples. Oxford Journal of Archaeology, 23: 271–86.Find this resource:
Henderson, J. (1989). Scientific analysis of ancient glass. In J. Henderson (ed.), Scientific Analysis in Archaeology. Oxford University Committee for Archaeology Monograph 19. Oxford, 30–60.Find this resource:
—— (2000). The Science and Archaeology of Materials. London: Routledge.Find this resource:
Heron, C. and Evershed, R. P. (1993). The analysis of organic residues and the study of pottery use. In M. B. Schiffer (ed.), Archaeological Method and Theory, Vol. 5. New York: Academic Press, 247–84.Find this resource:
Hosler, D. (1994). The Sounds and Colors of Power. Cambridge, Mass.: MIT Press.Find this resource:
(p. 232) Israeli, Y. (1991). The invention of blowing. In M. Newby and K. Painter (eds.), Roman Glass: Two Centuries of Art and Innovation. Society of Antiquaries of London Occasional Papers no 13. London, 46–55.Find this resource:
Kaczmarczyk, A. (1986). The source of cobalt in ancient Egyptian pigments. In J. S. Olin and M. J. Blackman (eds.), Proceedings of the 24th International Archaeometry Symposium. Washington, DC: Smithsonian Institution Press, 369–76.Find this resource:
Killick, D. (2001). Science, speculation and the origins of extractive metallurgy. In D. R. Brothwell and A. M. Pollard (eds.), Handbook of Archaeological Sciences. Chichester: Wiley, 483–92.Find this resource:
Lang, J. and Middleton, A. eds. (1997). Radiography of Cultural Material. Oxford: Butter-worth-Heinemann.Find this resource:
La Niece, S. and Craddock, P. eds. (1993). Metal Plating and Patination. Oxford: Butter-worth-Heinemann.Find this resource:
Lemonnier, P. (1986). The study of material culture today: towards an anthropology of technical systems. Journal of Anthropological Archaeology, 5: 147–86.Find this resource:
Maddin, R. (1987). The early blacksmith. In B. G. Scott and H. Cleere (eds.), The Crafts of the Blacksmith. Belfast: Ulster Museum (UISPP), 7–17.Find this resource:
Muhly, J. D. (1988). The beginnings of metallurgy in the Old World. In R. Maddin (ed.), The Beginnings of the Use of Metals and Alloys. Cambridge, Mass.: MIT Press, 2–20.Find this resource:
Northover, P. (1988). Alloy design in the Bronze Age. In J. E. Jones (ed.), Aspects of Ancient Mining and Metallurgy. Bangor: University College North Wales, 44–54.Find this resource:
Peacock, D. P. S. (1969). Neolithic pottery production in Cornwall. Antiquity, 43: 145–9.Find this resource:
Peltenburg, E. J. (1987). Early faience: recent studies, origins and relations with glass. In M. Bimson and I. C. Freestone (eds.), Early Vitreous Materials. British Museum Occasional Papers no. 56. London, 5–30.Find this resource:
Plog, F. (1977). Modeling economic exchange. In T. K. Earle and J. E. Ericson (eds.), Exchange Systems in Prehistory. New York: Academic Press, 127–40.Find this resource:
Pollard, M. and Heron, C. (1996). Archaeological Chemistry. Cambridge: The Royal Society of Chemistry.Find this resource:
Renfrew, C. (1969). The automony of the south-east European Copper Age. Proceedings of the Prehistoric Society, 35: 12–47.Find this resource:
—— (1975). Trade and interaction. In J. A. Sabloff and C. C. Lamberg-Karlovsky (eds.), Ancient Civilizations and Trade, Albuqurque: University of New Mexico Press, 3–59.Find this resource:
—— Dixon, J. E., and Cann, J. R. (1966). Obsidian and early cultural contact in the Near East. Proceedings of the Prehistoric Society, 32: 30–72.Find this resource:
Rice, P. M. (1987). Pottery Analysis—A Sourcebook. Chicago: University of Chicago Press.Find this resource:
—— (1999). On the origins of pottery. Journal of Archaeological Method and Theory, 6: 1–54.Find this resource:
Rye, O. S. (1976). Keeping your temper under control: materials and the manufacture of Papuan pottery. Archaeology and Physical Anthropology in Oceania, 11: 106–37.Find this resource:
Salter, C. (1989). Iron industry in Iron Age Britian. In J. Henderson (ed.), Scientific Analysis in Archaeology. Oxford University Committee for Archaeology Monograph 19. Oxford, 250–73.Find this resource:
Sayre, E. V. and Smith, R. W. (1961). Compositional categories of ancient glass. Science, 133: 1824–6.Find this resource:
—— —— (1974). Analytical studies of ancient Egyptian glass. In A. Bishay (ed.), Recent Advances in the Science and Technology of Materials, Vol. 3. New York: Plenum Press, 47–70.Find this resource:
(p. 233) Schiffer, M. B. and Skibo, M. (1997). The explanation of artifact variability. American Antiquity, 62: 27–50.Find this resource:
Scott, D. A. (1991). Metallography and Microstructures of Ancient and Historic Metals. Los Angeles: J. Paul Getty Trust.Find this resource:
Shortland, A. J. and Tite, M. S. (2000). Raw materials of glass from Amarna and implications for the origins of Egyptian glass. Archaeometry, 42: 141–51.Find this resource:
Sillar, B. and Tite, M. S. (2000). The challenge of ‘technological choices’ for materials science approaches in archaeology. Archaeometry, 42: 2–20.Find this resource:
Stos-Gale, S. (2000). Trade in metals in the Bronze Age Mediterranean: an overview of lead isotope data for provenance studies. In C. F. E. Pare (ed.), Metals Make the World Go Round. Oxford: Oxbow Books, 56–69.Find this resource:
Tite, M. S. (1999). Pottery production, distribution and consumption—the contribution of the physical sciences. Journal of Archaeological Method and Theory, 6: 181–233.Find this resource:
—— (2001). Overview—materials study in archaeology. In D. R. Brothwell and A. M. Pollard (eds.), Handbook of Archaeological Sciences. Chichester: Wiley, 443–8.Find this resource:
—— Shortland, A. J., Nicholson, P. T., and Jackson, C. M. (1998). The use of copper and cobalt colorants in vitreous materials in ancient Egypt. In S. Colinart and M. Menu (eds.), La Couleur dans la peinture et l'emaillage de l'Egypte ancienne. Ravello: Centro Universitario Europeo, 111–20.Find this resource:
—— —— and Paynter, S. (2002). The beginnings of vitreous materials in the Near East and Egypt. Accounts of Chemical Research. 35: 585–93.Find this resource:
—— Hatton, G. D., Shortland, A. J., Maniatis, Y., Kavoussanaki, D., and Panagiotaki, M. (2005), Raw materials used to produce Aegean Bronze Age glass and related vitreous materials. Annales of 16e Congrès de l'Association Internationale pour L'Histoire du Verre. London, 10–13.Find this resource:
Tylecote, R. F. (1991). A History of Metallurgy. London: Institute of Metals.Find this resource:
Whitehouse, D. (2002). The transition from natron to plant ash in the Levant. Journal of Glass Studies, 44: 193–6.Find this resource:
Wilson, L. and Pollard, A. M. (2001). The provenance hypothesis. In D. R. Brothwell and A. M. Pollard (eds.), Handbook of Archaeological Sciences. Chichester: Wiley, 507–17.Find this resource:
Woods, A. J. (1986). Form, and function: some observations on the cooking pot in antiquity. In W. D. Kingery (ed.), Ceramics and Civilization Vol. 2. Columbus, Ohio: American Ceramic Society, 157–72. (p. 234) Find this resource: