(Bradley, R. (forthcoming) in Fokkens, H. and Harding, A.F. (eds) Oxford Handbook of the European Bronze Age) 7.

Hoards and the deposition of metalwork

Richard Bradley

Introduction: The Archbishops treasure In 2009, divers working in the River Wear beside Durham Cathedral discovered a remarkable collection of metal artefacts. They could be identified as gifts presented to Archbishop Michael Ramsey who had lived in the vicinity after he retired as head of the Anglican Church. Their sources were very varied. They included a cross presented to him by the leader of the Russian Orthodox Church, a silver trowel given by the Bengal Coal Company when Ramsey laid the foundations of a church in India, a gold coin celebrating the renovation of the Grand Shrine of Ise in Japan, and a series of medals commemorating the second Vatican Council. These finds aroused considerable interest in the press, since it was difficult to decide why Ramseys property had ended in the river (The Guardian, 22 October 2009). Two schools of thought emerged, with widely diverging views. The version favoured by the cathedral authorities was that he had been the victim of a burglary which he had failed to report. The thieves had dumped these items as they would be easy to identify (www.durhamcathedral.co.uk/introduction/news/156). A second view, favoured by some of Ramseys colleagues, was that he was troubled by owning so many valuables and by the possibility that they would appear on the market after his death. Their special character

191

needed to be protected, and that was why he discarded them. He may even have regarded them as offerings to the deity, for he belonged to the section of the Church of England that approves of this procedure. There was also the argument, put forward by the divers themselves, that the objects had been deposited in small groups, in different places and probably on different occasions. That was not consistent with the behaviour of thieves. Ramsey died more than twenty years ago and the mystery is unlikely to be solved. It is understandable that deposits of Bronze Age metalwork some of them found in rivers have been still more difficult to interpret. Moreover, they form only one part of a much larger body of material which occurs in a series of other contexts. Among the most frequent are the artefacts from hoards of different kinds, but there were also single finds and grave goods. Although the hoards provide the main focus of this chapter, they cannot be discussed in isolation. Some of the problems raised by Archbishop Ramseys treasure are equally relevant to Bronze Age archaeology.

Hoards and the development of prehistoric archaeology Hoards can be defined as collections of buried objects which were apparently deposited together on the same occasion. They can be used to work out which kinds of artefacts were used concurrently. Seriation of the different groups also allows scholars to place them in sequence. This has been a concern of prehistorians for well over a hundred years. Where traces of organic hafts survive, individual items can be investigated by radiocarbon dating. This method has supported the framework built up by studies of these collections and has established an absolute chronology for Bronze Age metalwork. There are important studies of the hoards of different areas of Europe (examples include Von Brunn 1968 and Maraszek

192

2006) and some individual collections provide the subject matter for entire monographs (e.g. Coffyn, Gomez de Soto, and Mohen 1981; Needham 1990). At the same time, the study of hoards has led to the identification of a series of regional traditions of metalworking. This has shed a certain light on patterns of communication, and there have even been attempts to identify local styles of artefacts with specific communities in the past. Thus Colin Burgess (1980) has compared the distributions of particular metal types in Bronze Age Wales with the areas occupied by tribes recorded in the Roman period. Similarly, George Eogan (1974) finds echoes of the historic subdivision of Ireland in the distribution of different kinds of Late Bronze Age metalwork. Their approach recalls the analytical method followed by Gordon Childe, but instead of using the contents of hoards to identify regional cultures, he employed them to reconstruct the distinctive character of metal production. Following earlier writers, he identified particular combinations of artefacts with different stages in the manufacture and distribution of bronzes. He also emphasized the special status of the smith who, he considered, had been free to move between different communities (Childe 1958: chapter 10). If some hoards were associated with the activities of metalworkers, others were interpreted as personal property that had been concealed but not recovered. Peaks in the deposition of these collections were even identified as periods of crisis when valuables were hidden and lost. In this way the evidence of Bronze Age hoards could be used to write a kind of political history. That was especially true where their chronology matched that of hillfort building, since it was often treated as indirect evidence of warfare.

193

Problems of interpretation There are three main problems with traditional approaches of this kind. The first is selfevident. If hoards are regarded as collections of personal valuables, why are their contents virtually restricted to metalwork? The only exceptions appear to be the ceramic hoards of Central Europe (Harding 2000: 331-3). Why were other kinds of property excluded from the collections of metal artefacts? Why were other kinds of property excluded from these collections? And why did the deposition of hoards decline so rapidly when bronze was replaced by iron (Bradley 1998: 15960)? Perhaps the distinctive character of the raw material was considered to be particularly significant. That would not be surprising since the distributions of copper and tin are so restricted. The same applies to gold. The second problem is the difficulty of accepting that so many stores of valuables were concealed and never recovered. It seems unlikely in a period when graves were being robbed. There are regions in which the siting of hoards follows such a predictable pattern that these collections can be discovered by archaeologists today. Why was it more difficult in the past when the positions of some of the deposits appear to have been marked? Thirdly, it is hard to understand why relatively few collections of bronzes are associated with occupation sites which would have provided accessible and convenient locations for collections of valuables as they obviously did during the Roman period. Instead, the prehistoric metalwork may be found in isolation. Some was buried near settlements, yet there is little overlap between the finds associated with those sites and the objects discovered in hoards. It seems as if metalwork was made and used in the domestic

194

arena, but was separated from other kinds of artefacts when it was placed in the ground (Fontijn 2003). The medieval English law of Treasure Trove recognized the difficulty of interpreting such finds. It distinguished between collections of valuables meant to be recovered later, and those which were intended to remain in the ground; often they were in graves. The legislation laid down that items which were expected to be retrieved should become the property of the state. The others belonged to the owner of the land in which they were found. The law applied to objects of gold and silver, but the same distinction has been used in Bronze Age studies. Until recently interpretations varied between different parts of Europe, even where the archaeological evidence took exactly the same form. Thus scholars in Germany and Scandinavia considered that most hoards should be identified as votive offerings that were never meant to be recovered. Archaeologists in Britain and France, on the other hand, were more prepared to countenance a practical interpretation in which property was buried for safe-keeping and eventually lost (Bradley 1998: 1517).

Contents and contexts It is important to consider the contexts of these deposits together with their contents. Again traditional approaches have been misleading. Hoards have been classified according to at least three separate criteria, which to some extent overlap. Firstly, they might be categorized according to the identity of the people who deposited the artefacts. Thus personal hoards could be recognized by their distinctive composition, so that individual deposits were attributed, without much discussion, to men or women, warriors, craft workers or ritual specialists. Secondly, those collections associated with smiths were also categorized

195

according to different stages in the production of bronze artefacts. Thus some collections were identified as stores of newly made objects awaiting distribution to the customer, whilst others were interpreted as accumulations of scrap metal that had been brought together for recycling. A good example is Stuart Needhams detailed analysis of the metalwork deposits from Petters Sports Field in the Thames Valley (Needham 1990). Merchants hoards might include multiples of the same type of object, some of them unfinished or unused. Founders hoards, on the other hand, contained a variety of different kinds of objects, some of which had been broken and were mixed with casting waste. Lastly, the hoards could also be categorized according to the kinds of artefacts being made, so that separate deposits were associated with tools, weapons and ornaments respectively. These distinctions are sometimes treated as evidence of craft specialisation.

Contexts Collections of metalwork were found in a variety of contexts. At first the most important distinction was between deposits which would have been easy to recover and those where this would have been difficult or impossible to achieve. A fundamental distinction, which also extended to single finds, was between discoveries in dry land and those from watery locations, especially rivers and bogs. It seemed as if artefacts buried in the soil could have been hidden or stored for later retrieval, whilst those deposited in wet places would have been hard to find and equally hard to recover. Such distinctions extended to the composition of the hoards themselves. For the most part tools and metalworking residues were buried on dry land, whilst intact weapons were common finds in rivers (Torbrgge 1971). Groups of ornaments did not conform to this simple scheme, although many were found in bogs. With

196

this exception, it seemed as though deposits of Bronze Age metalwork could be divided into two distinct groups: votive offerings or ritual hoards which were not meant to be recovered, and non-ritual or utilitarian hoards which would have remained accessible after they were deposited. The latter group provided most evidence of artefact production (Bradley 1998: 10 14). A similar distinction was proposed by Janet Levy (1982) in a study of hoards in Bronze Age Denmark, although she paid more attention to the places in which they were found and the character of the artefacts themselves. She also studied the composition of the different hoards, their organisation in the ground (where it was recorded), the treatment of the objects found in each collection, and the presence or absence of food remains. Like other writers, she distinguished between sacred and mundane deposits, but Levy placed more weight on ethnographic evidence for the character of votive offerings in traditional societies. This had the advantage of relating the archaeological evidence to thinking in anthropology, but by relying on cross-cultural generalisation she moved the discussion away from the specific cultural contexts in which these items were accumulated. In fact the details of such deposits are too easily overlooked. For example, the dry land hoards come from a variety of completely different locations and do not form a particularly useful category. They are found close to settlement sites, but they may also be associated with more remote places, including hilltops, caves, cliffs, rock fissures and passes (Wyss 1996). They can be associated with older monuments, including megalithic tombs, round barrows, cairns and stone circles (Bradley 2000: 1567). It is true that in some cases these groups of artefacts could have been recovered, but that was not always so. The collections of half-

197

melted artefacts dropped into rock fissures provide an obvious exception. On the other hand, there are examples in which the positions of Bronze Age hoards were marked by an outcrop, a boulder or a mound. In principle it should have been possible to find them. Their locations may not have been secret, nor were they necessarily forgotten. Perhaps the artefacts remained undisturbed because contact with them was forbidden. In the same way, the wetland locations are extremely diverse. They include fastflowing rivers, brackish water, bogs, lakes, pools, streams and springs. Others are on the coast, and collections of Bronze Age metalwork have even been recovered from the sea. They are usually interpreted as the contents of shipwrecks, but, as Alice Samson (2006) has pointed out, some of them, like those from Langdon Bay or Sotteville-sur Mer, have a similar composition to groups of river finds and dry land hoards. In fact the contents of the Bronze Age wreck from the Ra de Huelva in southern Spain have been reinterpreted as a votive deposit (Ruiz-Glvez 2000). Other finds of metalwork are associated with fords, wooden causeways or bridges like that at Berlin-Spandau (Schwanger 1997), and still more have been found close to heaps of burnt stones in Britain and Ireland, some of which were associated with troughs for heating water. Bronze artefacts are also discovered inside artificial ponds and wells (Yates and Bradley 2010a). The character of the metal finds reflects some of these distinctions. In central Germany, for instance, finds of weapons are associated with the main river channels, while smaller items, especially pins, are found in marshes (Kubach 1979). In the same way, in the English Fenland weapons were deposited in the principal rivers, whilst entire hoards, including groups of ornaments, were placed in still water. Often these deposits consisted of intact objects, but close to the burnt mounds along the edge of the wetlands the

198

same types occurred as fragments mixed with metalworking residues (Yates and Bradley 2010a). Among them was the largest hoard found in Britain, from Isleham. In this example the distinction between dry land and water finds breaks down. It does so in other cases. Much of the Late Bronze Age metalwork in Belgium comes from an underground river inside a cave, le Trou de Han, yet other items are found in caves where water is absent (Warmenbol 1996). Similarly, a detailed study of the findspots of metalwork hoards in south-east England shows they were closely associated with the courses of rivers and streams. Here the main association was with fresh rather than salt water, but the objects themselves were buried in the ground nearby. Their locations are particularly revealing. There are finds of metalwork outside settlements identified by field survey (Fig. 7.1). Groups of hoards were deposited along the spring-line, with individual deposits beside the source of the water (Yates and Bradley 2010b). Others were buried near burnt mounds whose distribution follows streams and rivers, or on slight promontories overlooking a confluence (Fig. 7.2). There is even a case in which a hoard from Ditchling Common was associated with a mineral spring, famous in the Victorian period for its medicinal properties. Such evidence suggests that the distinction between wetland and dry land hoards is much too simple. Both groups are exceptionally diverse and it is not clear that all the terrestrial finds shared the same chance of recovery. Some places were readily accessible, yet the deposits associated with them remained intact. Other hoards were in remote locations well outside the area settled all year round. In the same way, the collections of metalwork from the wetlands show considerable variation, and it seems likely that different kinds of deposit were

199

associated with different kinds of water. Moreover, the siting of some of the finds from dry land was influenced by the presence of water in the vicinity.

Contents If the contexts of these finds are more varied than is commonly supposed, the same is true of the treatment of the artefacts themselves. Some of the most striking patterns were overlooked because the hoards were investigated for evidence of style and chronology. Just as they were studied in museums with little regard for the distinctive character of the findspots, their components were catalogued with no concern for the biographies of the objects themselves. Only in recent years has more attention been paid to the traces of use-damage and repair on specific artefacts, or to the ways in which they were treated when they were taken out of circulation. On one level metal analysis provided abundant evidence for the recycling of raw materials. On another, hoards of scrap metal were identified in the ground, but there was little connection between these two kinds of research. Rather than investigating the ways in which these groups of artefacts were assembled, specialists lamented that so many of the objects were incomplete, as it made them difficult to classify. The artefacts found in hoards can be intact or in fragments, and certain collections contain a mixture of both. In some cases it seems as if these items were broken when they were brought together, but there are other instances in which incomplete artefacts were assembled from different sources; classic cases include the French hoard of Vnat (Coffyn, Gomez de Soto, and Mohen 1981). As mentioned earlier, there are numerous hoards which contain freshly made objects and some which had never been finished, but it is clear that these

200

collections also include tools, weapons and ornaments with a significant history of use and repair. It is important to distinguish between the damage caused in the course of their history and that inflicted when they entered the ground. This is especially important in the case of collections containing only one kind of artefact, for example sickles in Central and Eastern Europe (Sommerfeld 1994). Because the same objects are represented so many times, it is tempting to suggest that they were employed as units of metal which may have had prescribed values in exchange. This argument could apply to other artefacts, for example axes (Fontijn 2003: 250-1), but in each case it requires wear analysis to establish whether they had ever been functioning tools (Kienlin and Ottaway 1998). That work is sometimes combined with analysis of their composition, for some axe heads, like those from the latest hoards in North-West France, were probably too soft to be used. Their weights have also been studied. It is possible that common forms of artefact, including axes and sickles, seem to conform to a series of standard weights. The small bronze figurines of Late Bronze Age Scandinavia provide another illustration (Malmer 1992). The sizes and weights of entire hoards have also been considered (Maraszek 2006). A novel approach is to investigate the number of artefacts that were deposited in different collections, for sometimes there seem to be significant patterns (Brandherm 2007). Unfortunately, none of these approaches can show why the hoards were buried. The quantity and weight of the artefacts might have been calculated by the smiths before they made fresh items, but they could also have measured the amount of metal that had to be dedicated to the gods. The deposition of coins at Iron Age and Roman sanctuaries provides a possible analogy.

201

Many deposits are associated with the by-products of metalworking: casting jets, moulds, crucibles, slag and ingots. For that reason they have been interpreted by Levy (1982) as non-ritual hoards. This approach makes the unwarranted assumption that the artefacts brought together in these deposits were assembled for practical reasons. They were meant to be melted down in order to make new objects, but it never happened. At first sight the argument is plausible, as the evidence of metal analysis shows that recycling was very common in the Bronze Age. On the other hand, the information provided by the hoards can be taken too literally. This is acknowledged by Claude Mordant (2007) in a new study of the French hoard from Villethierry which he and his colleagues had originally interpreted in purely functional terms. Now its position on the boundary of two cultural traditions suggests that it played a specialized role. There are three main problems that need to be addressed. The first is that the evidence of scrap hoards does not correspond with the information provided by metal analysis. It is clear that copper and bronze artefacts were made of recycled material from an early stage, but scrap hoards are most common in the Late Bronze Age. It was not necessary to store these collections in the ground in order to undertake the process, and in any case it still remains uncertain why so much metal was buried and not recovered. On one reading of the evidence scrap metal was used to make new objects throughout the Bronze Age, but only in certain phases did much of this material escape the melting pot. Why it did so has still to be discussed. Secondly, there is evidence that bronze artefacts were not merely broken to reduce them to a manageable size or weight. In some cases they were reduced to fragments using

202

extraordinary force. It was not easy to break all these objects, and there was little need to do so. Louis Nebelsick (2000), who has drawn attention to this phenomenon, observes that the level of violence with which it was accomplished seems to have varied according to the locations where the hoards are found. Thus in southern Germany metalwork from the domestic landscape received less drastic treatment than artefacts from what he calls wild places, such as mountains or cliffs. Thirdly, a purely practical interpretation of the scrap hoards would not account for some striking regularities in the composition of the collections themselves. They vary along regional lines across different parts of North-West Europe, but there is no reason to believe that these were the only combinations of artefact types that were available for reuse. Rather, the composition of the hoards even those containing metalworking residues seems to have been governed by local conventions, so that all the raw material assembled by the smith cannot have been considered on equal terms (Maraszek 1998; Turner 2010). Clearly, certain combinations of different types were required in hoards. Some were considered appropriate and others evidently were not. This argument faces the problem that only the artefacts that escaped recycling can be studied today, but what appears to be a difficulty may provide a vital clue. Perhaps a specific selection of the raw material had to be deposited in the ground when other artefacts were melted down. The objects that were actually reused might have been much more diverse. The contents of these hoards need not provide a representative sample of the types in circulation. That is clear from the discovery of moulds for making artefacts of forms that have rarely been

203

found. In the same way, the moulds for making bronze artefacts can appear in quite different frequencies from the objects themselves (Rassmann 1996).

The role of the smith The conventional interpretation of metal hoards was beset by another problem that is being addressed by new research. The argument that certain hoards are non-ritual or utilitarian makes certain assumptions about the roles of the smiths. One reason why this has happened is the influence of Gordon Childe (1958) who thought of them as free agents moving between different communities and marketing their skills on a commercial basis. That conception may be anachronistic and is not consistent with accounts of metalworking in non-Western societies which suggest that the entire process is ritualized and governed by specific protocols (Budd and Taylor 1995). Smiths can enjoy a special position in the community and their unusual skills take on something of the qualities of magic. It may be wrong to think in terms of contemporary notions of manufacture and trade. A new interpretation of the role of the smith comes from archaeological research in South Scandinavia. A recent study by Joakim Goldhahn (2007) draws attention to the association between cremation burials and evidence of bronze working. Both can be associated with the same deposits of burnt stone and with a series of specialized buildings that are usually described as cult houses. During the Late Bronze Age human bodies are known to have been burnt to an exceptionally high temperature. According to Goldhahn, this could only have been achieved in a furnace. It might explain why cremation and metal production were sometimes carried out together. For that reason he suggests that in Northern Europe the smith was a ritual specialist. Perhaps the strongest evidence comes from Hallunda where a

204

cult house has been identified in association with cremation burials and the largest concentration of metalworking furnaces in Bronze Age Sweden (Goldhahn 2007: 293-306). A comparable but less detailed argument has been put forward by Joanna Brck (2006) who suggests that in Britain and Ireland the fragmentation of the metalwork in Late Bronze Age hoards reflects the treatment of human remains during the same period. It is difficult to take the discussion further, but in southern Germany ingots, weights, copper cake and unfinished objects have been found in Late Bronze Age burials (Winghart 2000). Such arguments raise the possibility that the transformation of raw metal involved rituals as well as technological procedures. It was a dangerous process in both practical and social terms, so it was often undertaken in special places. Perhaps it was necessary to deposit a prescribed portion of the material as a votive offering. That may be why certain selections of objects were brought together and treated in such a distinctive manner. It may also be why they were buried and never recovered. In this interpretation the link with metalworking remains, but its utilitarian character is questioned. The non-ritual hoards defined by Janet Levy (1982) could have had a more specialized character than is normally supposed. In fact they may result from rituals undertaken by the smiths who were obliged to return to the earth a portion of the material obtained from it. The social anthropologist Mary Helms suggests this interpretation in a recent paper which draws on evidence from the European Bronze Age: Depositional objects like metals ... derived from raw materials ... believed to be redolent with the cosmological power of the earth that originally generated them. I ... posit that Bronze Age people believed, as many traditional cultures have, that if human society

205

takes living, energized material from the world for human use, then human society has the obligation to return life energy to the earth to replenish the store, so that ... the renewal and regeneration of all living things may continue (2009: 155). The distinctive character of Bronze Age hoards has no exact counterpart in the ethnographic record. Helmss interpretation might account for the unusual character of these deposits, but it depends on the same methodology as Janet Levys scheme. In the end the strongest arguments will always be based on archaeological observations.

Some wider implications If Bronze Age hoards represent a particular kind of votive offering, how were they related to other deposits of metalwork? Should they be considered separately from single finds, and how, if at all, do they compare with the objects in graves? Re-assessment of the social and ritual importance of smiths raises some new possibilities, but it brings fresh problems, too. Once it is accepted that the transformation of metals was not just a technological process, it becomes clear that other aspects of hoards and hoarding also need to be considered.

Hoards and single finds The distinction between hoards and single finds reflects the way in which Bronze Age studies have developed. Research on metalwork hoards showed which kinds of artefacts were made and used together, and so this work had priority. By definition, single finds did not lend themselves to the same approach and for that reason they played a secondary role. On the other hand, the objects in these two classes have so much in common that in other respects they can be considered together. Some of those classified as single finds may have formed

206

part of larger collections, but many of them really were isolated discoveries. They share a number of common properties with the contents of hoards. For example, certain of these objects have been broken in exactly the same ways as the artefacts found in those collections; the fragments resemble one another despite their different contexts (Bradley 2005: chapter 5). There is evidence that a few of the swords deposited in rivers had been burnt. It is not clear whether this happened on a cremation pyre or whether the weapons had been damaged by a smith. What is obvious is that they had been treated rather like the contents of a scrap hoard. Sometimes it is possible to distinguish between two different stages in the histories of these objects. For example, most of the weapons deposited in the River Thames show signs of duelling scars, re-sharpening and repair, but when these artefacts were discarded they were disabled with exceptional force. The proportion of artefacts that was deliberately damaged actually increased over time (York 2002). Again their treatment was comparable to the damage inflicted on objects in dry land hoards (Nebelsick 2000). There are also cases in which the same types are represented in hoards in one phase and as single finds in the next, suggesting that both kinds of deposit were equivalent to one another (Jensen 1972).

Single finds, hoards and grave finds Single finds of weapons are especially common in rivers and bogs, but complete and fragmentary examples are also represented in hoards (Fig. 7.3). Swords often occur in graves, but, as Torbrgge noted forty years ago, the two kinds of deposit rarely occur together. Either the contexts in which the artefacts were deposited changed over the course of time, or different practices were followed in different parts of Europe (Torbrgge 1971). As a result, grave finds and single finds can have mutually exclusive distributions (Fig. 7.4). This striking

207

pattern has encouraged the idea that weapons were deposited in water in the course of funeral ceremonies. Dry land hoards containing swords and spears may have been deposited in similar circumstances (Coombs 1975). A good example is the collection from Penard in south Wales. It is an attractive argument, especially in regions where elaborately furnished burials are rare or absent, but this interpretation presents certain problems. Only a few groups of river finds are accompanied by human remains which can be shown to be contemporary with the metalwork itself. Even then they may not represent a cross section of the Bronze Age population. The remains from the Thames, for example, are mainly skulls, and it is not clear whether entire bodies had once been present or whether unfleshed bones, or even heads, had been deposited in the water after preliminary treatment elsewhere (Bradley and Gordon 1988). At all events such finds have not been recorded from most of the rivers containing Bronze Age artefacts. Thus they represent the exception rather than the norm. Ornament hoards pose a similar problem, although they are more common in bogs. They have also been found on dry land. In some cases they could represent the personal equipment of one individual. Comparison with grave assemblages suggests that they were probably worn by women (Maraszek 2006: 13651). Different artefacts show different amounts of wear, implying that they were acquired at separate stages of life, even though they were deposited together on the same occasion. Such ornaments can occur as single finds, but most of them seem to have been discarded as complete sets. Indeed, there are hoards in South Scandinavia which contain the artefacts associated with as many as five separate people (Levy 1982). A good example is the collection from Skjdstrup in Jutland (Levy 1982: Plate 11 and

208

152). Indeed, there are hoards in South Scandinavia which contain the artefacts associated with as many as five separate people. Similar combinations of objects can be recognized from burials, but again the two groups of material are found in different areas. Although ornament hoards have occasionally been discovered together with bones, it is not always known whether the remains were human. Both weapons and ornaments can be represented by hoards and single finds in one region and were buried with the dead in another. They could also change their contexts over time, so that they might have been associated with the deceased in one phase and separated from them during a different period Bradley 1998: chapter 3). In both cases it is obvious that some of these objects had a lengthy history. They may have enjoyed a special significance, too. The high quality and distinctive decoration of the ornaments, which could be made of bronze or gold, suggest that they were not available to everyone; the same may be true of the weapons. It is uncertain whether the more elaborate artefacts were personal possessions or insignia, but their special character would have been threatened if they were treated in the same ways as other valuables when their period of use came to an end. That provides one reason why they were withdrawn from circulation. Perhaps their role as votive offerings protected their special character (Fontijn 2003). The same idea might explain why the gifts presented to Archbishop Ramsey as the head of the Church of England were deposited in the River Wear.

Space and time There are difficulties with all these interpretations. If certain objects played such a specialized role that they had to be sacrificed once their use was over, how could the same types of

209

objects have been accumulated and melted down by smiths? And how could ornaments, weapons and tools have been mixed together as a source of raw material? One possibility is that this process removed these items from circulation as effectively as their deposition in a river. On the other hand, it could suggest that they lost their distinctive character when their outward appearance was transformed (Fig. 7.5). Another possibility is that these items could be reused precisely because they no longer had their original associations (Mordant 1998; Fontijn 2008). This could happen in at least two ways. Either they had been brought together outside the area in which they carried special connotations, or they were made in styles that were becoming obsolete. Both possibilities are suggested by existing studies of hoards, but they are seldom discussed in these terms. The distribution of metalwork hoards shows two distinctive patterns which have been identified in different parts of Europe. The first concerns the distinction between mixed hoards whose contents combine the conventional categories of tools, weapons and ornaments, and those in which only one of those classes of artefact is represented. If particular kinds of object had prescribed roles in Bronze Age society they might have been kept separate in the area where they had a special significance. Similarly, those types would be more likely to appear together outside that region. Such a pattern sometimes can be identified. For example, in the English Middle Bronze Age the main concentration of single category hoards is bounded by that of mixed deposits (Bradley 1998: 1223). A more striking spatial pattern involves hoards and single finds. If the metalwork found in rivers played a specialized role, would it have happened where the same kinds of artefacts were included in hoards of scrap metal? The two types of deposit often have

210

complementary distributions so that the regions with concentrations of isolated weapons or groups of ornaments are effectively enclosed by deposits in which these objects had been reduced to fragments and mixed with the by-products of metalworking. This phenomenon is clearly documented in Britain, Germany and Denmark (Bradley 1998: 1237; Mordant 1998). It may be that objects which had travelled outside the areas in which they played a restricted role could be treated as a source of raw material. In exactly the same way objects of unfamiliar types entering a region from outside might have melted down so that they could be replaced by more appropriate forms. Again that happened towards the edge of their overall distribution. It might explain why so many scrap hoards occur near the coast, for this is where foreign artefacts were most likely to be encountered (Fontijn 2008). There is yet another possibility to consider. Burgess and Coombs (1979) suggested that metalwork hoards in the British Isles were most frequent towards the ends of the phases to which they are assigned. It is difficult to assess their argument which is based on artefact typology, but it does raise an interesting possibility. If they are correct and if their hypothesis needs to be verified in other regions it might suggest that some of the material consigned to the ground was already going of fashion. That is not to endorse their idea that bronze was dumped because it lost its value as iron came into use: an explanation which would not explain the intensity of hoarding in Central Europe during Hallstatt A1 (Harding 2000: 356). It seems more likely that certain objects could be recycled because they no longer possessed their original significance. When it happened, a proportion of the material was retained as a votive offering.

211

This discussion has emphasized three significant points. The first is that hoards should not be investigated without paying equal attention to single finds and the artefacts deposited with the dead. The relationship between these categories helps to identify a number of important issues. Secondly, it is essential to study hoarding in its wider geographical and chronological contexts, as the same forms of artefacts can appear as both hoard finds and single finds. The relationships between them need to be defined in more detail than is possible here. Lastly, it is not enough to catalogue the contents of these deposits, for the treatment afforded to the different artefacts provides a vital source of information. Each of these comments has the same implication. The hoard is an artefact of a particular kind of archaeological analysis. In the past these collections of artefacts did not possess any unifying feature. It is the task of modern research to take this category apart and to investigate its separate components. The collections of artefacts that prehistorians have grouped together as hoards need to be investigated as part of a wider enquiry into the deposition of Bronze Age metalwork.

Unfinished business It is too soon to summarize every new development in the study of metalwork deposits, and this account has made no attempt to do so. Instead it has highlighted some of the approaches that appear to be most productive. At the same time, it may be helpful to mention some topics that require more investigation. It is important to emphasize that that there is much still to learn. The final section of this chapter suggests a number of ways in which this can be achieved. One set of problems concerns the use of raw materials, and the other relates to the places in which artefacts are found.

212

A first priority is to consider the roles of hoards and single finds over a longer period than the Bronze Age. The debates that have preoccupied Bronze Age specialists have also taken place in the archaeology of the Iron Age, Roman and Early Medieval periods, and similar issues have been discussed with little reference to a longer history or a more general body of ideas. It is especially important to consider the earliest deposits of artefacts in wetlands. In Northern Europe they predate the adoption of farming, but similar practices extended throughout the Neolithic period (Karsten 1994). It would be useful to know more about the ways in which this tradition was modified when metal became available, for copper and bronze must have been thought to possess unusual properties. As mentioned earlier, it is also necessary to investigate the ways in which ritual behaviour changed as bronze was replaced by iron. If the source of the raw material was an important factor, then it is a priority to investigate hoards and single finds of goldwork in more detail than has happened so far (Morteani and Northover 1993). There are some deposits in which both bronze and gold are found, but it was unusual to use these materials together. To what extent do the patterns suggested for deposits of bronze artefacts apply to those of gold? Were they circulated and consumed according to the same procedures? Were the same people responsible for working both kinds of metal, and were the artefacts themselves deposited in similar places? In many regions that has still to be established. Objects made of both materials are often represented by fragments. Why did this happen? It implies that the missing parts must have been treated separately, whether or not they were melted down. It seems possible that the circulation of broken objects was a

213

common practice and that it played a role in creating and maintaining relationships between people in the past. In principle, different parts of the same object could have had separate histories, and there may be cases in which the pieces were reunited after an interval. Alternatively, they might have been deposited individually. These questions have been investigated by John Chapman in Eastern Europe (Chapman 2000: 11221), but specialists on the metalwork of other regions show little awareness of this important research. If these questions need to engage scholars who study metalwork, other observations are more relevant for field archaeologists. There is every reason to believe that certain deposits were appropriate in certain kinds of places. The distinction between river finds and dry land hoards is only one example of a more general process. It seems as though particular objects or combinations of objects were usually discarded in quite specific locations. Although these relationships could change from one region to another, in some instances they were maintained over a considerable period of time. David Fontijns work in the southern Netherlands provides the best example of these patterns, for in this case sickles and pins were associated with houses in the settlements; axes, sickles and spears were deposited in streams and marshes further away; and swords and more elaborate ornaments were confined to the major rivers (Fontijn 2003). In other regions hoards are associated with particular topographical features. In principle it should be possible to predict the future pattern of discovery and to take it into account in the management of the cultural heritage. That has already been achieved in a study of votive offerings in one area near Amsterdam (Kok 2008), and the same methods could be applied in other parts of Europe.

214

Such work is bound to take place at an extensive geographical scale, but there is scope for other research on the immediate surroundings of hoards and single finds. Too few of these collections are examined while they remain in the ground, and the investigations that do take place can be conducted on too small a scale. In some regions it is clear that hoards were deposited outside settlements of the same date, but only rarely have both been treated together. It is a relationship that can be investigated by sample excavation, fieldwalking and various forms of remote sensing, and one that should be studied more intensively in the future. That may be easier to achieve now that large areas of ground are investigated by developer-funded fieldwork.

Coda The discovery of Michael Ramseys treasures in the River Wear posed a difficult problem for the authorities of Durham Cathedral. Had the gifts presented to him during his time as Archbishop of Canterbury been stolen and discarded by the thieves, or had they been deposited by Ramsey himself? Were they taken out of circulation to protect their special character, or could he have intended them as gifts to God? It is quite possible that they fulfilled both roles at the same time. Deposits of bronze artefacts pose the same problem (among many others), and the ways in which they have been interpreted are equally diverse. Again there has been a debate between those who see such collections as votive offerings, and scholars who believe that they can be understood in practical terms. If opinion is divided on the character of a deposit which is less than thirty years old, it is hardly surprising that the existence of prehistoric hoards should be so difficult to explain.

215

Bibliography Bradley, R. (1998). The passage of arms (2nd edn). Oxford: Oxbow. (2000). An archaeology of natural places. London: Routledge. (2005). Ritual and domestic life in prehistoric Europe. London: Routledge. and Gordon, K. (1988). Human skulls from the River Thames, their dating and significance. Antiquity, 62: 5039. Brandherm, D. (2007). Swords by numbers, in C. Burgess, P. Topping, and F. Lynch (eds.), Beyond Stonehenge. Essays in Honour of Colin Burgess. Oxford: Oxbow, 288300. Brck, J. (2006). Fragmentation, personhood and the social construction of technology in Middle and Late Bronze Age Britain. Cambridge Archaeological Journal, 16: 297315. Brunn, W. von (1968). Mitteldeutsche Hortfunde der jngeren Bronzezeit. Berlin: de Gruyter. Budd, P. and Taylor, T. (1995). The faerie smith meets the bronze industry: magic versus science in the interpretation of prehistoric metal making. World Archaeology, 27: 13343. Burgess, C. (1980). The Bronze Age in Wales, in J. Taylor (ed.), Culture and Environment in Prehistoric Wales. Oxford: British Archaeological Reports (British Series), 76: 24386. and Coombs, D. (1979). Preface, in C. Burgess and D. Coombs (eds.), Bronze Age Hoards. Some Finds Old and New. Oxford: British Archaeological Reports (British Series), 67: ivii. Chapman, J. (2000). Fragmentation in Archaeology. London: Routledge. Childe, V. G. (1958). The Prehistory of European Society. Harmondsworth: Penguin.

216

Coffyn, A., Gomez de Soto, J., and Mohen, J.- P. (1981). Lapoge du bronze atlantique (le dpt de Vnat). Paris: Picard. Coombs, D. (1975). Bronze Age weapon hoards in Britain. Archaeologia Atlantica, 1: 49 81. Eogan, G. (1974). Regionale Gruppierungen in der Sptbronzezeit Irlands. Archologisches Korrespondenzblatt, 4: 31927. Erbach-Schnberg, M.-C. zu (1985). Bemerkungen zu urnenfelderzeitlichen Deponierungen in Obersterrreich. Archologisches Korrespondenzblatt, 15: 163-78. Fontijn (2003). Sacrificial landscapes. Cultural biographies of persons, objects and natural places in the Bronze Age of the southern Netherlands c. 2500 600 BC. Analecta Praehistorica Leidensia, 33-34: 1392. (2008). Tradershoards. Revisiting the relationshipbetween trade and permanent deposition: the case of the Dutch Voerhout hoard, in C. Hamon and B. Quilliec (eds.), Hoards from the Neolithic to the Metal Ages. Technical and Codified Practice. Oxford: British Archaeological Reports (International Series), 1758: 517. Goldhahn, J. (2007). Ddens hand en ess om brons- och hllsmed. Gothenburg: Gothenburg University. Harding, A.F. (2000). European Societies in the Bronze Age. Cambridge: Cambridge University Press.

217

Helms, M. (2009). The master(y) of hard materials: thoughts on technology, materiality and ideology occasioned by the Dover boat, in P. Clark (ed.), Bronze Age Connections. Cultural Contact in Prehistoric Europe. Oxford: Oxbow, 14958. Jensen, J. (1972). Ein neues Hallstattschwert aus Dnemark. Beitrag zur Problematik der jungbronzezeitlichen Votivfunde. Acta Archaeologica, 43: 11564. Karsten, P. (1994). Att kasta yxan in sjn. Lund: Acta Archaeologica Lundensia. Kienlin, T. and Ottaway, B. (1998). Flanged axes of the North Alpine region: an assessment of the posibilities of wear analysis on metal artefacts, in C. Mordant, M. Pernot, and V. Rychner (eds.), Latelier du bronzier en Europe. Production, circulation and comsummation du bronze. Paris: ditions du Comit des Travaux historiques et scientifiques: 27186. Kok, M. (2008). The Homecoming of Religious Practice. Doctoral thesis, University of Amsterdam. Kubach, W. (1979). Deponierung in Mooren der sdhessischen Oberrheinebene. Jahresbericht fr Vorgeschichte der Universitt Frankfurt a. M. 1979: 198310. Levy, J. (1982). Social and religious organisation in Bronze Age Denmark. Oxford: British Archaeological Reports (International Series), 124. Malmer, M. (1992). Weight systems in the Scandinavian Bronze Age. Antiquity, 66: 377 88. Maraszek, R. (1998), Sptbronzezeitliche Hortfunde entlang der Oder. Bonn: R. Habelt.

218

 (2006). Sptbronzezeitliche Hortfundlandschaften in atlantischer und nordischer Metalltradition. Halle: Landesamt fr Denkmalpflege und Archologie Sachsen-Anhalt. Mordant, C. (1998). Dpts de bronze et territories lage du Bronze, in C. Mordant, M. Pernot, and V. Rychner (eds.), Latelier du bronzier en Europe. Production, circulation and comsummation du bronze, Paris: ditions du Comit des Travaux historiques et scientifiques: 185210. (2007). Le dpt de Villethierry (Yonne). Un relecture des donnes, in C. Burgess, P. Topping, and F. Lynch (eds.), Beyond Stonehenge. Essays in Honour of Colin Burgess. Oxford: Oxbow, 33543. Morteani, G. and Northover, J. P. (eds.) (1993). Prehistoric gold in Europe. Mines, metallurgy, manufacture. Dordrecht: Kluwer Academic Press. Nebelsick, L. (2000). Rent asunder: ritual violence in Late Bronze Age hoards, in C. Pare (ed.), Metals make the world go round. Oxford: Oxbow, 16075. Needham, S. (1990). The Petters Late Bronze Age Metalwork. London: British Museum. Rassmann, K. (1996). Untersuchungen zu sptbronzezeitlichen Hortfunden im nrdlichen Schwarzmeergebiet, in C. Huth (ed.), Archologische Forschungen zum Kultgeschehen in der jngeren Bronzezeit und frhen Eisenzeit Alteuropas. Bonn: Habelt, 53555. Ruiz-Glvez, M. (2000). Ritos de paso y puntas de paso. La Ra de Huelva en el mundo del Bronze final Europeo. Madrid: Universidad Complutense. Samson, A. (2006). Offshore finds from the Bronze Age in north-western Europe: the shipwreck scenario revisited. Oxford Journal of Archaeology, 25: 37188.

219

Schwanger, S (1997). Wanderer kommst Du nach Spa - Der Opferplatz von BerlinSpandau. Ein Heiligtum fr Krieger, Hndler und Reisende, in A. Hnsel and B. Hnsel (eds.), Gaben an die Gtter. Berlin: Staatliche Museum zu Berlin - Preussischer Kulturbesitz, 616. Sommerfeld, C. (1994). Gertegeld Sichel. Studien zur monetren Struktur bronzezeitlichen Horte im nrdlichen Mitteleuropa. Berlin: de Gruyter. Torbrgge, W. (1971). Vor- und frhgeschichtliche Flussfunde. Bericht der RmischGermanischen Kommission, 52: 1146. Turner, L. (2010). A Re-interpretation of the Later Bronze Age Metalwork Hoards of Essex and Kent. Oxford: British Archaeological Reports (British Series 507). Warmenbol, E. (1996). Lor, la mort et les Hyperborens. Le Bouche des Enfers ou le Trou de Han Han-sur-Lesse, in C. Huth (ed.), Archologische Forschungen zum Kultgeschehen in der jngeren Bronzezeit und frhen Eisenzeit Alteuropas. Bonn: Habelt, 2334. Winghart, S. (2000). Mining, processing and distribution of bronze: reflections on the organisation of metal supply between the northern Alps and the Danube region, in C. Pare (ed.), Metals Make the World Go Round. Oxford: Oxbow, 1519. Wyss, R. (1996). Funde von Pssen, Hhlen, aus Quellen und Gewssern der Zentral- und Westalpen, in C. Huth (ed.), Archologische Forschungen zum Kultgeschehen in der jngeren Bronzezeit und frhen Eisenzeit Alteuropas Bonn: Habelt, 41728. Yates, D. and Bradley, R. (2010a). Still water, hidden depths. The deposition of Bronze Age metalwork in the English Fenland. Antiquity, 84: 40515.

220

 (2010b) The siting of metalwork hoards in the bronze age of south-east England. The Antiquaries Journal 90: 1-32. York, J. (2002). The life cycle of Bronze Age metalwork from the Thames. Oxford Journal of Archaeology, 21: 7792.

221