Journal of Archaeology in the Low Countries 1-1 (May 2009)

Coin use in a dynamic frontier region. Late Iron Age coinages in the Lower Rhine area

Nico Roymans, Joris Aarts

Keywords: Archaeology, Celtic coinage, Lower Rhine region, social use of coins, relation with historically documented tribal dynamics.


The introduction of money in the form of standardised objects of value made of metal and bearing images marked a new phenomenon in the pre-Roman societies of Western and Central Europe. In the Late Iron Age, the Lower Rhine region formed part of the northern peripheral zone of the La Tène culture, whose influence in this region has emerged as stronger than was previously thought. This is reflected among other things in the large numbers of ‘Celtic’ coins from this region.[1] Although very little was known about these coins until about 1980, the number of coins in the archaeological record, as well as what we know about them, has increased dramatically in recent decades.

This study seeks to survey these earliest coinages in the Lower Rhine region. We start with a few introductory remarks about the development of Celtic numismatics, and follow with a discussion of the research potential of coins from the Lower Rhine region. We then survey the evolution of coinage and coin production in this area from the 2nd century BC. The most important coin groups are discussed, with an emphasis on their distribution, dating and possible attribution to a particular tribe. Finally, we address the implications of these coinages for some broader socio-cultural issues. Our aim is to provide answers to the following questions: Why and by whom were Celtic coins produced and in what types of context were they used? To what extent can we link patterns in the numismatic material to the historically documented formation of a series of new tribes in the decades immediately after Caesar’s conquest of Gaul? What factors attributed to the relatively slow, late start of coin use among Lower Rhine groups, and why did the use of coins stop at the Dutch/Belgian coastal plains and the area north of the Rhine delta? We shall argue that the study of coinage from the Lower Rhine region can make a significant and original contribution to our understanding of both Late Iron Age societies in the broader sense, and of their increasing integration into the Roman empire.

1 Celtic numismatics as a field of study

The field of Celtic numismatics developed in Europe from the 19th century onward, initially as a side branch of Antique numismatics. Originally, it concentrated on the coins struck by Western and Central European peoples in the centuries before the Roman conquest. These coinages began in the 4th century BC with the fabrication of faithful copies of Greek-Hellenistic coins. However, it wasn’t long before coins appeared with an iconography of their own, quite distinct from that of Mediterranean coin types.[2]

The initial focus of Celtic numismatics was on the typochronological organisation of the material, based on iconography, metrology, and the study of coin associations in hoards. But just as important was the attribution of coin groups to a specific tribe or people. For Gaul, this was based largely on the tribes mentioned by Caesar, which were –rather uncritically- assumed to have existed already much earlier, namely at the time the coinages were produced. This method of ethnic attribution is not without its problems, however, as we shall discover later. A key reference work for northern Gaul and the Rhineland is the synthesis of the Belgian numismatist Simone Scheers, which appeared in 1977.[3] Since the 1990s, the research focus has increasingly shifted to the processes of coin production, circulation and deposition, employing methods and interpretative frameworks popular among archaeologists. Researchers have come to realise that a coin’s information value lies not only in the object itself but also, and more specifically, in the context in which it is found. As a result, a recent in-depth analysis of archaeological contexts has prompted both a radical revision of the typochronology of northern Gallic coins and a study of the relationship between coins and ritual practices.[4]

These days archaeologists are increasingly engaged in the study of Celtic coins. Often their ultimate goal is to gain an appreciation of broader social phenomena, such as the nature and functioning of systems of exchange, ritual behaviour, the socio-political organisation of groups and how this changed over time.

A point of special interest are the functions of Celtic coins. Since the 1970s scholars have repeatedly pointed out the dangers of anachronistic interpretations, and have sought to arrive at more historically contextualized notions of coin use and exchange in ancient societies.[5] Inspired by Bohannan’s study of systems of exchange among the Tiv of pre-colonial Nigeria and by Lucas’ analysis of systems of exchange in medieval Ireland, one of the present authors developed a model for Late Iron Age societies in northern Gaul.[6] The model proposes the existence of a multicentric economy in which different, more or less mutually exclusive, spheres of exchange operated. Gold and silver coins functioned as items of exchange and as a means of payment in the prestige sphere, while the low-value bronze coins which emerged in the last phase of the Iron Age were used in everyday transactions in the subsistence sphere.[7] Many have associated the appearance of low-value coins with emerging markets and the beginnings of a monetary economy, but this remains a matter of conjecture.

At the same time, however, the frequent occurrence of large numbers of Celtic coins at cult places in Western Europe raises the question of the importance of ritual forms of exchange. Consequentially, a need has developed for models which focus on the interrelationships between the economic, social and religious dimensions of coin production and use. Relevant here is Bloch & Parry’s model showing the articulation of two kinds of exchange – that of a short-term sphere concerned with individual gain and competition, and a long-term sphere in which the reproduction of collective values, norms and cosmologies is paramount.[8]

Recent research has shown that the production of coins must be viewed in the light of the reproduction of the long-term sociocosmic order of the society which produced them. Appadurai’s notion of ‘tournaments of value’ is of particular value here: regularly recurring religious festivals in which the central values of the society were shaped anew and competition between elites played a key role.[9] Theuws employed this concept in his analysis of the production of gold coins in early medieval Maastricht, probably in the context of recurrent rituals and ceremonies on the feast of St. Servatius.[10] Cult places are favourite locations for such tournaments of value. Appadurai’s concept also appears relevant to coin production in Late Iron Age societies. We should not view the minting of coins (not just gold, but low-value coins as well) as a strictly profane affair, governed by economic and political considerations. The fact that sanctuaries tend to be the largest find spots for coins,[11] together with the frequent occurrence of religious symbols on coins, suggests a link with ritual and ceremonies addressing the central values and identity of communities.[12] Coins may have been minted under the patronage of deities.

2 The Lower Rhine coin database and its research potential

In this study we focus on coin use in the Lower Rhine plain, a region corresponding to the present-day southern and central Netherlands, northern Belgium and the adjacent German Lower Rhine region. It is also the northern periphery of the area where Celtic coins were used. The past decade has seen a radical change in our perception of the social organisation of the societies that lived here in the Late Iron Age: they have been shown to be more complex than we thought. The settlement pattern is not simply one of individual farmsteads and small hamlets; it also features larger settlements with centre functions, particularly in the religious domain.[13] The practice of collective weapon depositions at cult places points to the emergence of a new social arena for elite competition in which martial values were cultivated. The contrast with the more southerly regions in Northern Gaul is less extreme than is often suggested.[14]


Fig. 1 Tribal map of the Lower Rhine region and neighbouring areas at the time of Caesar’s conquest.

So wherein lies the specific research potential of the study of Celtic coins in the Lower Rhine region? We would like to distinguish three elements:

a. The region’s location in the northernmost zone of coin-using societies on the European continent. A number of intriguing questions concern the process of coin introduction. How should we construe the late and sluggish start in the Lower Rhine zone? How should we interpret the substantial regional differences within this area? And how did societies which used coins differ from societies which did not?

b. The relationship between the emergence of new coin issues and the historically documented, dynamic political geography of the region in the second half of the 1st century BC. We are dealing here – in part as a consequence of Roman frontier policy – with the dynamics of disappearing tribes and the genesis of new ethnic groupings (see Figs. 1 and2). To what extent are these processes evident in the numismatic material?

c. The quality of the numismatic data. In the Netherlands, many coins found by amateur archaeologists using metal detectors are systematically recorded and inventoried. This has produced a meteoric rise in the number of finds (see Table 1). We now have access to fairly representative distribution patterns of the various coin types for the Netherlands, as well as to context information.


Table 1. Increase in the number of coins found in the Netherlands, with 1980 and 2005 as reference years.


Fig. 2 Historically documented migrations of Germanic groups to the western Lower Rhine region in the second half of the 1st century BC.

Figure 3 shows the processes behind the formation of the current database of Celtic coins from the study region. This diagram underpins a host of interpretations with regard to the production, circulation and deposition of coins in this study. Distribution maps are an important analytical tool, but they leave us with some fundamental questions: just how representative are they of coins actually present in the soil and to what extent do they present a distorted picture of ancient coin deposition and loss?[15]


Fig. 3 Diagram showing the formation processes of Late Iron Age coinages in the Lower Rhine region (after Roymans 2004, Fig. 4.9).

Distribution patterns are influenced by regional differences in the intensity of metal detector use, different national legislation in this area, and differences in soil type and in patterns of modern agrarian land use. For example, the number of find spots in the Netherlands has jumped spectacularly within a short time, by the sheer intensity of reported finds and by effective coin inventory programmes (see Tables 2 and 3). Conspicuous here is the high proportion (about 75%) of coins found in ordinary rural settlements. The Dutch data suggests that this category of find spots may be underrepresented in the data from France, Belgium and the German Rhineland, something we need to bear in mind when interpreting distribution maps. Nevertheless, this does not mean that distribution maps are useless: even within the Netherlands, they reveal significant patterns. We can also compare the distributional evidence with the composition of coin lists of special find spots, such as cult places, oppida and Roman camps in the different regions. This too may tell us something about regional coinage patterns.


Table 2. Distribution of silver and bronze triquetrum coins across different find spot categories in the Batavian region and the German Rhineland. (After Roymans 2004, table 6.1).


Table 3. Distribution of AVAVCIA coins over different find spot categories in the Netherlands, Belgium and Germany. (After Aarts & Roymans, in press).

3 Survey of the principal coin groups from the Lower Rhine region

Imported gold coins (2nd – mid-1st century BC)

As yet, no first-generation Celtic coins have been found in the Lower Rhine region. The first coins did not appear here until the 2nd century BC. Initially, they all came from the more southerly regions (Fig. 4), which points to intensive interregional contacts.

The earliest gold coins to appear here are several quarter staters of the Scheers 23 type, with a male head on the obverse and a winged horse on the reverse (Fig. 5). The coins with a Pegasus originate from the Middle Rhine region (Fig. 6) and can be placed in the 2nd century BC.[16] The gold hoards of Beringen (mid-2nd century BC) and Niederzier (first half 1st century BC), both of which comprise a combination of gold jewellery and imported coins, can be dated to the same early period.[17]


Fig. 4 Find spots of imported gold coins from the Late Iron Age in the Lower Rhine region. a. coin imported from the southwest; b. coin imported from the southeast; c. hoard (after Roymans 2004, Fig. 4.1.).


Fig. 5 Gold quarter staters of the Pegasus type (Scheers 23 type) from the 2nd century BC. a.-b. find spot unknown; c. Limburg province (NL) (Scale c. 3:1. Photo Geld- en Bankmuseum, Utrecht).


Fig. 6 Distribution of gold quarter staters of the Pegasus type (Scheers 23 type). Open circle: exact find spot unknown. Triangle: hoard (after Heinrichs/Rehren 1996, Fig. 12, updated).

The first Lower Rhine emission: Eburonean triskeles staters (Scheers 31 type, mid-1st century BC)

The earliest coins struck in the region of study are staters of the Scheers 31 type with a triskeles on the obverse and a horse on the reverse. They are made of poor-quality gold, and some even of bronze. This coin group can be dated to the mid-1st century BC on the basis of its low weight and gold content, and its association with other late coins in the hoard of Heers. The date of this coin issue, together with its distribution across the southeastern Netherlands and Central Belgium (Fig. 7), allows us to attribute the coins to the Eburones, a tribe mentioned by Caesar.[18] It is impossible to pinpoint the places where they were minted. The current distribution pattern shows concentrations in Central Belgium and the Dutch river region.


Fig. 7 Distribution of gold triskeles coins of the Scheers 31 type, attributed to the Eburones. a. 1-5 coins; b. >20 coins; c. hoard (after Roymans 2004, Fig. 4.3, with some additions).

Lower Rhine triquetrum staters and the ethnogenesis of the Batavi (c. 50-15 BC)

The silver and bronze ‘rainbow cups’, which feature a triquetrum on the obverse and a configuration of dots and circles on the reverse (Fig. 8), are also mainly of Lower Rhine origin. They are the youngest variants of a long-running coin series that began with gold staters in the first half of the 1st century BC. These staters came from the trans-Rhine region and were probably struck in the oppidum on the Dünsberg in Hessen. After about 50 BC, production of these coins largely shifted to the Lower Rhine area, where the series was continued in silver and, still later, in bronze. This was a gradual process; even the youngest bronze coins often contain a small quantity of gold or silver, or are sometimes gilded. This points to a conservative adherence to the old series and to the use of these coins in the same kind of transactions as the older gold coins. A salient feature is the presence of small additional marks on the reverse of many silver and bronze coins; in total no less than 20 different variants are now known. The coins still occur in the earliest Augustan forts in the Lower Rhine region, although it is not likely that new coins were minted in this period. All in all, this coin group seems to include a large number of issues spread over time from the period c. 50 to 15 BC. The coins are concentrated in the eastern part of the Dutch river region (Fig. 9), a fact which – together with the relatively late dating – would justify the attribution of most issues to the Batavi.[19] Their production coincides with the historically documented ethnogenesis of the Batavi. The coins may have played a role in the integration of native and migrant groups into the new Batavian community, perhaps dominated by an elite from north of the Rhine. Of interest here is the presence of large numbers of coins (more than 250 specimens) at the cult place of Empel.


Fig. 8 Hoard of silver triquetrum staters from Echt (L). (Photo Restaura, Haelen).


Fig. 9 Distribution of silver and bronze triquetrum coins. a. Roman camp; b. other find spots; c. hoard (after Roymans 2004, Fig. 6.1).

Rhineland silver quinarii with ‘dancing manikin’ (Scheers 57 type; c. 65 BC - 1 AD)

Another long-running coin series are the quinarii of the Scheers 57 type, with a figure of a ‘dancing man’ on the obverse and a horse on the reverse. Schulze-Forster and Heinrichs have recently discussed the development of this coin group.[20] It began as a trans-Rhine issue, with the oppidum on the Dünsberg in Hessen as the probable minting centre (Fig. 10). A younger group consists of silver with a high copper content and is concentrated in the western Lower Rhine region, where it also appears to have been minted.[21] Circulation was concentrated in the Cologne hinterland, and to a lesser extent in the Dutch river region. Partly on the basis of the dating of these coin issues, the cluster in the Cologne region may be associated with the Ubii.[22] There is a plausible link to historical accounts of Ubian migration from the east to the west bank of the Rhine during Agrippa’s first or second governorship (38/39 or 19/18 BC).[23] There are too few coins from the Dutch river region for the Batavi to have played any role in the more recent issues of this series.[24] Schulze-Forster dates the quinarii of the younger group to after 30 BC. Their presence in Augustan army camps suggests that these coins continued to circulate until about AD 10, although their production will have ceased somewhat earlier.


Fig. 10 Distribution of silver coins with ‘dancing man’ (Scheers 57 type). a. coins from the early series; b. coins from the late series; large symbols >10 coins (after Schulze-Forster 2005, Fig. 6; Heinrichs 2005, Fig. 6.0; with additions and corrections for the Netherlands).

Silver coins with the legend ANNAROVECI (Scheers 58 type) and the ethnogenesis of the Tungri (c. 50-25 BC)

Concentrated in central Belgium is a small, local coin group of silver quinarii with a high copper content featuring a right-facing head on the obverse, a horse on the reverse and the circular legend ANNAROVECI on both sides (Figs. 11 and12). Haselgrove dates the coins to between 60 and 20 BC.[25] However, the strong similarity between the horse on the reverse and the horse of the more recent AVAVCIA coins (Fig. 11) argues for a dating in the third quarter of the 1st century BC. This late dating renders Scheer’s proposed attribution of this coin type to the Eburones an anachronism.[26] More plausible is an association with the ethnogenesis of the Tungri. A leader by the name of Annaroveci may have played a prominent role here.


Fig. 11 Bronze AVAVCIA coin with legend (class I; left) and ANNAROVECI quinarius (right).


Fig. 12 Distribution of silver ANNAROVECI coins, Schemers 58 type (after Schemers 1996, Fig. 3, with additions).

Bronze AVAVCIA coins and articulation with the Roman monetary system (Scheers 217 type, c. 25 BC-AD 10)

The most recent, comprehensive and widely distributed Lower Rhine coin group is that of bronze ‘AVAVCIA’ coins, the obverse of which shows a swastika and the reverse a horse (Fig. 11). Because the coins occur on such a vast scale in the earliest Roman camps and civil centres (Fig. 13), it is widely assumed that they were fully integrated into the Roman monetary system. It has been suggested that they were minted to make up for a shortage of small-denomination Roman coins. This raises the fundamental question of whether it is still possible to speak of a tribal issue, or whether they are better regarded as provincial Roman coins. A recent study of this coin type identified a fundamental distinction between distribution patterns for class I coins with the circular legend AVAVCIA on the reverse and class II/III coins without a legend.[27] Coins from the former group are concentrated in central Belgium (Fig. 14), with only small numbers known to be from Roman military and urban centres in the Rhineland. We suggest that class I coins be viewed as a tribal emission, struck before the first Roman forts were established in the Rhine zone under Drusus from c. 15 BC onwards. This issue can probably be ascribed to the Tungri. Further evidence to support this, apart from the distribution pattern, is the fact that the horse on the reverse is virtually a copy of the one on the ANNAROVECI coins discussed above. This brings us to the question of whether the functional interpretation of these class I coins as ‘small change’ is still tenable if we accept that this was a pre-Roman tribal issue. After all, a monetized market exchange system can hardly have existed in the Tongres area in the period before Drusus’ arrival. It makes more sense, as with the ANNAROVECI coins, to associate the production to the formative phase of the Tungri tribal confederation.[28] It is not inconceivable, as with Annaroveci above, that this involved a leader by the name of Avaucia.


Fig. 13 Distribution of bronze AVAVCIA coins (Schemers 217 type, all classes). a. 1-5 coins; b. 6-15 coins; c. 16-100 coins; d. >100 coins ; e. Roman camp (after Aarts & Roymans, in press, Fig. 3).


Fig. 14 Distribution of bronze AVAVCIA coins, variant with legend (class I). a. 1-5 coins; b. 6-15 coins; c. 16-100 coins; d. Roman camp (after Aarts & Roymans, in press, Fig. 50).

The coins from class II/III are uninscribed. Their distribution is clearly concentrated in the Roman camps in the Lower Rhine region, where they occur in vast numbers (Fig. 15). The coins appear to have been struck in the camps themselves, as evidenced by the identification of local variants at Haltern and the discovery of a flan mould fragment in the Roman camp on the Kops Plateau in Nijmegen.[29] The excavators associate this find with the local production of AVAVCIA coins.[30] What we can say is that the class II/III coins were struck after 15 BC by order or with consent of the Roman authorities with the aim to satisfy the enormous demand for small denomination coins in the Roman camps.


Fig. 15 Distribution of bronze AVAVCIA coins, variants with no legend (class II-III). a 1-5 coins; b 6-15 coins; c 16-100 coins; d >100 coins; e Roman camp (after Aarts & Roymans, in press, Fig. 6).

4 Conclusions and discussion

The present data gives us a clear picture of the development of Celtic coinage in the Lower Rhine region. Coinage began here in the 2nd century BC with imported gold coins. It was not until about the mid-1st century BC that coin production – starting off with gold Eburonean triskeles staters –came into full swing, reaching a peak in the second half of the 1st century with issues of low-value coins. Distribution patterns also shed light on the considerable regional differences in intensity of coin use in the study region. Coins are virtually absent from the coastal zone and the area immediately north of the Rhine delta, while the density of find spots is so high in regions like the eastern part of the Dutch river delta and the adjacent zones upriver along the Lower Rhine and the Meuse that we may safely assume that the population of almost every rural settelement was familiar with coin use.

The above data suggests several leads for further discussion. The first concerns possible links between the numismatic material and the historically documented ethnic dynamics of the Lower Rhine frontier region, where the latter half of the 1st century BC was pre-eminently a period of ethnogenesis, or, more specifically, the formation of new tribal confederations (Fig. 2). The Batavi, Ubii and Tungri were new entities, and we see this reflected in the numismatic material, each group minting its own issues. Also of interest is the fact that the coins produced in the territory of the Batavi and Ubii – triquetrum staters and quinarii featuring the ‘dancing man’ respectively – were continuations of already existing coin series from beyond the Rhine, with the oppidum on the Dünsberg in Hessen as the probable centre of production. Here we see special coin movements that may relate to historically documented group migrations.[31] This topic also confronts us with an important methodological issue. The Lower Rhine example, with its intense tribal dynamics and discontinuities, highlights the dangers of projecting tribes further back in time on the basis of numismatic data.

A second topic are the reasons behind the minting of large quantities of low-value bronze or poor-quality silver coins in the study region in the second half of the 1st century BC. In contrast to older gold coins, the relatively late, low-value coins in Gaul and the Rhineland are generally regarded as small change and are associated with emerging markets and the monetized exchange of goods.[32] However, there is no serious evidence in the Lower Rhine region for the existence of larger market centres, which argues against a strictly economic interpretation of native coin issues in the period before the advent of Roman camps under Drusus in about 15 BC. It is therefore in traditional tribal contexts that we must seek an explanation. Coins were probably minted in connection with public ceremonies and rituals (tournaments of value, to use Appadurai’s term), which were designed to produce and reproduce central ideas and values of collective identities. As such, coinage contributed to the symbolic construction and social cohesion of groups. It is perhaps no coincidence that the coin production boom in the Lower Rhine region coincided with the formation of new tribal entities. The production of ANNAROVECI and class I AVAVCIA coins ran parallel with the ethnogenesis of the Tungri. We can associate the minting of the bulk of Lower Rhine triquetrum coins with the ethnogenesis of the Batavi, while the production of Lower Rhine variants of the quinarii featuring the ‘dancing manikin’ may be linked to the formation of the Ubii.[33] These were new political groupings (arising out of Roman frontier politics), which also sought to profile themselves as ethnic entities. From this perspective then, the low-value coins were not primarily struck to satisfy an economic need for small change. The first coin series to which this applied were the class II/III AVAVCIA coins, which appear to have been deliberately minted for use within a Roman monetary system, targeting the camps and camp villages.

This broader cultural interpretation helps to explain several features of these low-value coin issues:

a. Their presence in ritual or ceremonial contexts, especially cult places

b. Their depiction of religious symbols, especially the triskeles and horse in the Lower Rhine region

c. The fact that some low-value coin series (particularly bronze triquetrum staters) gradually evolved out of gold prototypes and were still enriched with a small quantity of gold and/or silver right until the end. This suggests that they were not struck as low-value coins, but should be seen as continuing the old stater tradition.[34]

A third topic is the extent to which the introduction of coinage implies a new kind of society. It is generally accepted that coin use was characteristic of more hierarchical tribal formations dominated by elites. After all, coin use – and certainly coin production – presupposes leaders operating at the tribal or subtribal level who were in a direct or indirect way responsible for the coin issues. These leaders used coinage to create social networks and to strengthen the collective identity and social cohesion of communities. The frequent occurrence of coins in ordinary rural settlements shows that individuals were highly integrated into supralocal networks dominated by elites. As such, Late Iron Age societies which minted coins appear to have had a different structure than societies from earlier phases of the Iron Age.

A fourth topic is why coin use was not adopted by societies north of the Rhine delta or in the Belgian-Dutch coastal region.[35] A combination of factors may have been responsible for this phenomenon. Certainly relevant was the fairly limited agrarian potential of coastal landscapes and the low intensity of exchange networks with regions further to the south.[36] The isolated geographical position of the coastal zone was exacerbated still further by the lack of overseas contact of the Rhine/Meuse delta with England during the Iron Age. The relatively isolated position of coastal communities remained a constant throughout late prehistory and protohistory. In the late Bronze Age and early Iron Age, these communities appear to have had no access to prestigious goods like swords, bronze vessels and horse gear.[37] In the Roman period, handmade native pottery predominated for much longer there, and urban development came about very slowly and laboriously. At the same time, socio-cultural factors may also explain why coin usage failed to be taken up by these communities. We could, for instance, envisage societies with a strong inward-looking identity and a highly segmented tribal structure in which there were no authorities to mint coins. In short, there was no real need for a coinage.


Fig. 16 Model for the production, circulation and deposition of Late Iron Age coinages.

A final topic is our changed perception of the production, circulation and deposition of Celtic coins in the light of the above. In summary, we can devise the following model (Fig. 16). The primary minting of coins occurred in a ritualised context, emphasising the central values of the community that struck the coins. Evidence for this can be found in the imagery on the coins, which refers to their sacral origin. This appears to be a universal phenomenon, however, and not one which is confined to Late Iron Age societies.[38] Further evidence could be the fact that many coins have been found on the sites of central cult places. This requires some explanation. For most coins, we only know where they ultimately ended up in the soil as a result of deliberate deposition or loss. A concentration of coins at a cult place can be explained in two ways. The first assumes that the coins were produced there during tournaments of value by people who were the key players. They were part and parcel of the gift exchange between these central figures and their followers that took place in this ceremonial context. The coins may have been offered at the cult place itself, or have been lost there, or a combination of the two. Most coins, however, will have been taken at the close of the ceremony to the homes of the participants. This led to a rapid primary distribution of coins across the settlements within the core region where the tournament of value was held. The money may then have been kept for later use within the settlements, or offered in a ritual context in the settlement itself. Coins unearthed by archaeologists excavating rural settlements will thus have been part of ‘unsuccessful’ savings hoards or of ritual depositions. Coins from the savings hoards might then have begun a ‘secondary’ circulation through use in various kinds of payment, especially in the social-ritual sphere. As stated above, the use of coins for market exchange seems less likely because our region lacked any significant market centres. But even if they had existed, this does not mean that money was actually used for payments in the subsistence sphere. A monetary market can exist only if it is supplied regularly with new coins. But the coins were not minted for this purpose, and it is unlikely that the secondary circulation referred to above possessed enough volume to make possible a monetized market exchange.[39] It is conceivable that coins were offered during this secondary circulation and that many of the coins found at cult places are the material result of a final ritual deposition. This is the second explanation for the concentration of coin finds in central cult places. Coin finds from sanctuaries can thus be interpreted as a kind of palimpsest representing different stages in the circulation of money: the initial stage during the production and issue of coins, and the final stage, following a series of payments, when the coins once again became part of a ritual deposition.[40] Once included in such a deposition, it is unlikely that coins later embarked on a third circulation phase. The re-use of offered money (conversions from the long to the short term in Bloch & Parry’s terms) carries a taboo in many societies, and there are indications that this was also the case in the Late Iron Age and the Roman period.[41]

Analysing distribution maps is a complex matter because we often do not know which moment during circulation a coin deposition reflects. Our task is made even more difficult by post-depositional processes that colour our perception. Nevertheless, we can in some instances link the distribution of particular coin groups to the model described above. For example, the earliest AVAVCIA issues can be associated with the ethnogenesis of the Tungri, and a significant portion of the silver and bronze triquetrum staters with that of the Batavi. Clearly, there was a greater need to affirm or create value systems in periods when new group identities were being forged. This could have boosted the numbers of tournaments of value in the Lower Rhine region from about 50 BC until the early Augustan period. The minting of coins may have played a significant role here. We have already shown that the many additional marks on triquetrum staters indicate a series of issues spread over time.[42] These marks can probably be interpreted as markers for a particular occasion when the coins entered into circulation – in other words, a tournament of value.

What then are the possibilities for future research? Our knowledge of Celtic coinage in the study region can be substantially enhanced by the detailed study of dies, supplemented by metal analysis. This would increase our understanding of issue size and the location of mints. The ongoing meticulous registration of new coin finds from excavations and of coins in private collections will of course continue to be of critical importance. As a result, we can expect coin distribution and context-differentiation patterns to be more representative of ancient deposition than they are now. At a higher level of analysis, there is a need for a new conceptual framework to interpret the entire repertoire of use for Celtic coins. This will entail a study of exchange systems in the broader sense, one which incorporates the socio-political, economic and religious dimensions of exchange. Celtic coins offer excellent opportunities in this regard because of their religiously inspired iconography, their occurrence in a wide range of contexts and their link to political authorities.[43]


We wish to thank Colin Haselgrove, David Wigg, Arent Pol and Paul Beliën for their comments on an earlier version of this text, published as the Van Gelder Lezing (Roymans 2008). Unless otherwise stated, the illustrations have been prepared by Bert Brouwenstijn (Archaeological Centre VU University Amsterdam). Annette Visser translated the text into English.

Nico Roymans, Joris Aarts

Universiteit van Amsterdam


1. The term ‘Celtic’ is not used here in a strictly ethnic sense. It has emerged that in the Lower Rhine region Celtic coins were struck by tribes whom the Greco-Roman written sources describe as ‘Germanic’.

2. For a general introduction to Celtic numismatics, see Allen & Nash 1980.

3. Scheers 1977.

4. See Haselgrove 1999 and the bundle of articles collected in Haselgrove & Wigg 2005 respectively.

5. For example, it is important to point to substantial differences from coin use in pre-modern and modern city states and the Roman empire, all of which had a series of different coin denominations that functioned within a partially monetary economy dominated by markets.

6. Bohannan 1955; Lucas 1989; Roymans 1996, 45-47; Creighton 2005, 71-76.

7. Roymans 1990, 131 ff.; idem 1996, 44-49. See also Creighton 2005, 71-76, and Nick 2006, 111-112.

8. Bloch & Parry 1989, 24.

9. Appadurai 1986; Theuws 2004.

10. Theuws 2004, 128.

11. Haselgrove & Wigg-Wolf 2005, 12. In Picardy, there is even evidence that coins were minted at sanctuaries. Cf. Delestrée 2005. We should also remember that ritual coin depositions occurred at non-religious sites as well, such as oppida or rural settlements.

12. On the religious symbolism on coin images, cf. Haselgrove & Wigg-Wolf 2005, 13; Creighton 2000, 42 ff. In the Lower Rhine region, the triskeles and horse were particularly popular religious symbols.

13. We can point to the pre-Roman cult places of Empel, Kessel and Elst, which had a supra-local relevance. Cf. Roymans 2004, 12 ff.

14. For recent analyses of Late Iron Age societies in the Lower Rhine region, see Roymans 2004, chapter 2; Gerritsen & Roymans 2006; Roymans 2007.

15. For this discussion, see Roymans 2004, 45 ff., 84-87 and Bazelmans’ (2003) critical analysis of the distribution patterns of Roman coins in Friesland.

16. Haselgrove 1999, 128.

17. Beringen: Van Impe et al. 1997/1998. Niederzier: Göbl et al. 1991.

18. See Roymans 2004, chapter 4, for a recent synthesis of this coin type, with further references.

19. For a comprehensive analysis of this coin group, see Roymans (2004) chapter 6. An association with the Batavi is also apparent from the distribution of most variants featuring additional marks. Heinrichs (2003) initially argued for an Ubian emission, but later qualified his position, adding that the Batavi – as well as the Ubii – could have been responsible: Heinrichs 2005, 184. See also Nick 2006, 45, who believes it is too early to draw conclusions from present distribution patterns.

20. Schulze-Forster 2005; Heinrichs 2005. See also Nick 2006, 71.

21. Schulze-Forster 2005, 164ff. The earliest group corresponds to his I-IIIA series and the younger to his IIIB-C series. Most coins from the area west of the Rhine belong to the younger group.

22. Heinrichs 2005; Schulze-Forster 2005, 172 ff, is always cautious with tribal associations.

23. Recently, Eck (2004, 46-55) emphatically opted for a link to Agrippa’s second governorship.

24. Fifteen coins from 12 find spots are now known from the Batavian river region. Significantly, the cult place of Empel has yielded only three specimens (one from the early series) from a total of more than 850 Celtic coins: Roymans & Aarts 2005, table 1. See also Van den Brandhof 2005.

25. Haselgrove 1999, 157.

26. Scheers 1996, 11-12. For a discussion of the relation of this issue to the ethnogenesis of the Tungri, see also Aarts & Roymans, in press.

27. Aarts & Roymans, in press.

28. Aarts & Roymans, in press.

29. Haltern: Scheers 1996, 22; Ilisch 1999, 285. Flan mould, Nijmegen: Van Enckevort & Joosten 2002.

30. An alternative interpretation would be that the blanks made in this mould were used for striking imitations of low value Roman coins.

31. See the discussions in Roymans 2004, 82-89; Schulze-Forster 2005, 172 ff; Heinrichs 2005.

32. Cf. for example Van Heesch 1998, 45 ff.; Heinrichs 2003, 299; Nick 2006, 126 ff, 187.

33. According to historical sources, the Ubii moved in 39/38 or 19/18 BC under the governorship of Agrippa from the east to the west bank of the Rhine and settled in the Cologne hinterland. Cf. Eck 2004, 46-55. However, this is undoubtedly a highly simplified representation of events. As with the Batavi, we should think in terms of a new ethnogenesis resulting from a blending of the dominant migrant group (Ubii) with residual native groups.

34. See Roymans 2004, 89-90.

35. For the rare occurrence of Celtic coins in the Belgian coastal region west of the Schelde, see Van Heesch 1998, 39, 43, Figs. 14 and 19.

36. Apart from coins, a range of other material groups from the repertoire of forms of the later La Tène culture (glass bracelets, swords, horse gear, bronze vessels, belt hooks, fibulae, etc.) occur rarely or not at all in the Belgian-Dutch coastal zone. On this subject, see Roymans 2007. The volume and the social impact of the production and exchange of sea salt by coastal communities remains a subject for debate. In the Iron Age, this salt probably circulated in a different sphere of exchange from that of coins.

37. Cf. the distribution maps for bronze swords from the late Bronze Age/Early Iron Age and of southern imports from the Hallstatt C phase in the Low Countries in Roymans 1991, Figs. 6, 9, 13 and 16.

38. We could think here of the depiction of gods on Greek and Roman coins or the portrayal of the emperor as a deity, but also of the fact that coins were often minted on or near the site of sanctuaries. For example, the mint in Rome was located near the temple of Juno Moneta, and it is believed that the two buildings were linked by a corridor (Meadows 2001, 27). But we also encounter references to the divine on medieval and modern coins, such as the motto ‘God zij met ons’ [may God be with us] on the former Dutch guilder.

39. This does not mean that money was not used in market exchange. However, barter and other forms of exchange would always have played an important role alongside money, and conversely, market exchange can only account for part of the coinage in Late Iron Age societies. This applies to both low-value bronze coins and to gold and silver coins.

40. Looking at the degree of wear and tear on coins from cult places gives us some idea here. For example, gold coins from the temple at Empel are very obviously worn, which means that they circulated for a long time or very intensively after being issued and before being offered again (Roymans & Aarts 2005).

41. Aarts 2005. In certain exceptional circumstances (e.g. famine), conversion from the long to the short-term sphere of exchange was indeed possible. Also, plundering by a group of outsiders could result in coins once again ending up in the sphere of short-term exchange. Caesar’s plunder of gold hoards from Gallic cult places is a good example of this.

42. Roymans 2004, 91.

43. For the cult place at Empel (where Hercules Magusanus was worshipped as principal deity), for example, the present authors have attempted to interpret various deposition practices – including that of coins – using a life-cyle model of warriors or soldiers. Cf. Roymans & Aarts 2005.


Aarts J. 2003: Monetisation and army recruitment in the Dutch river area in the early 1st century AD,. In T. Grünewald & S. Seibel (eds) Kontinuität und Diskontinuität. Germania inferior am Beginn und am Ende der römischen Herrschaft, Berlin/New York, 163-180.

Aarts, J. & N. Roymans in press, Tribal emission or imperial coinage? Ideas about the production and circulation of the so-called AVAVCIA-coinages in the Rhineland. In J. van Heesch (ed.) Liber amicorum S. Scheers.

Allen, D.F. & D. Nash 1980, The coins of the ancient Celts, Edinburgh.

Appadurai, A. 1986, Introduction: commodities and the politics of value. In: A. Appadurai (ed.) The social life of things. Commodities in a cultural perspective, Cambridge, 3-63.

Bazelmans, J. 2003, De Romeinse muntvondsten uit de drie noordelijke provincies. Methodische kanttekeningen bij een nieuwe periodisering der relaties, Utrecht (Van Gelder-lezing, 2).

Bohannan, P. 1955, Some principles of exchange and investment among the Tiv, American Anthropologist 57, 60-69.

Bloch, M. & J. Parry 1989, Introduction: money and the morality of exchange. In M. Bloch & J. Parry (eds) Money and the mora­lity of exchange, Cambridge, 1-32.

Creighton, J. 2000, Coins and power in Late Iron Age Britain, Cambridge (New Studies in Archaeology).

Creighton, J. 2005, Gold, ritual and kingship. In C. Haselgrove & D. Wigg (eds) Iron Age coinage and ritual practices, Mainz (Studien zu Fundmünzen der Antike 20), 69-83.

Delestrée, L.-P. 2005, Les offrandes monétaires gauloises sur les sanctuaires du Belgium. In C. Haselgrove & D. Wigg (eds) Iron Age coinage and ritual practices, Mainz (Studien zu Fundmünzen der Antike 20), 321-335.

Eck, W. 2004, Köln in römischer Zeit. Geschichte einer Stadt im Rahmen des Imperium Romanum, Cologne.

Gerritsen, F. & N. Roymans, 2006, Central places and the construction of tribal identities. The case of the Late Iron Age Lower Rhine region. In C. Haselgrove (ed.) Les mutations de la fin de l’Âge du Fer, Glux-en-Glenne (Collection Bibracte 12/4), 251-266.

Göbl, J., A. Hartmann, H.E. Joachim & V. Zedelius 1991, Der spätkeltische Goldschatz von Niederzier, Bonner Jahrbücher 191, 27-84.

Haselgrove, C. 1999, The development of Iron Age coinage in Belgic Gaul, The Numismatic Chronicle 159, 111-168.

Haselgrove, C. & D. Wigg-Wolf 2005, Introduction: Iron Age coinage and ritual practices. In C. Haselgrove & D. Wigg-Wolf (eds), Iron Age coinage and ritual practices, Mainz (Studien zu Fundmünzen der Antike 20), 9-23.

Heinrichs, J. & Th. Rehren 1996, Keltische Münzschrötlingsfor­men aus Bonn. Naturwissenschaftlich-numismatischer Vorbericht und historische Interpretation, Metalla 3.2, 83-100.

Heinrichs, J. 2003, Ubier, Chatten, Bataver. Mittel- und Niederrhein ca. 70-71 v.Chr. anhand germanischer Münzen. In Th. Grünewald & S. Seibel (eds), Kontinuität und Diskontinuität. Germania inferior am Beginn und am Ende der römischen Herrschaft, Berlin/New York, 266-344.

Heinrichs, J. 2005, Ubische Quinare im Lippegebiet: ein Modell. In J. Metzler & D. Wigg-Wolf (eds) Die Kelten und Rom. Neue numismatische Forschungen, Mainz (Studien zu Fundmünzen der Antike, 19), 183-228.

Ilisch, P. 1999, Die Münzen aus den römischen Militäranlagen in Westfalen. In W. Schlüter & R. Wiegels (eds) Rom, Germanien und die Ausgrabungen in Kalkriese, Osnabrück, 279-291.

Lucas, A.T. 1989, Cattle in Ancient Ireland, Kilkenny (Studies in Irish Archaeology and History).

Meadows, A. & J. Williams 2001, Moneta and the monuments. Coinage and politics in Republican Rome, Journal of Roman Studies 91, 27-49.

Nick, M. 2006, Gabe, Opfer, Zahlungsmittel. Strukturen keltischen Münzgebrauchs im westlichen Mitteleuropa, Rahden.

Roymans, N. & W. van der Sanden 1980, Celtic coins in the Netherlands and their archaeological context, Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 30, 172-258.

Roymans, N. 1990, Tribal societies in Northern Gaul. An anthropological perspective, Amsterdam (Cingula 12).

Roymans, N. 1991, Late Urnfield societies in the Northwest European Plain and the expanding networks of Central European Hallstatt groups. In N. Roymans & F. Theuws (eds), Images of the Past. Studies on Ancient Societies in Northwestern Europe, Amsterdam (Studies in Prae- and Protohistorie 7), 9-89.

Roymans, N. 1996, The sword or the plough. Regional dynamics in the romanisation of Belgic Gaul and the Rhineland area. In N. Roymans (ed.) From the sword to the plough. Three studies on the earliest romanisation of northern Gaul, Amsterdam (Amsterdam Archaeological Studies 1), 9-126.

Roymans, N. 2001, The Lower Rhine triquetrum coinages and the ethnogenesis of the Batavi. In Th. Grünewald (ed.), Germania inferior. Besiedlung, Gesellschaft und Wirtschaft an der Grenze der römisch-Germanischen Welt, Berlin/New York, 93-145.

Roymans, N. 2004, Ethnic identity and imperial power. The Batavians in the Early Roman Empire, Amsterdam (Amsterdam Archaeological Studies, 10).

Roymans, N. & J. Aarts, 2005, Coins, soldiers and the Batavian Hercules cult. Coin deposition in the sanctuary of Empel in the Lower Rhine region. In C. Haselgrove & D. Wigg (eds) Iron Age coinage and ritual practices, Mainz (Studien zu Fundmünzen der Antike 20), 337-359.

Roymans, N. & H. Hiddink, 2006, Ein Hortfund spätlatènezeitlicher „Regenbogenschüsselchen“ der Nordgruppe aus Echt (Niederlande), Archäologisches Korrespondenzblatt 36, 225-236.

Roymans, N. 2007a, On the latènisation of Late Iron Age material culture in the Lower Rhine/Meuse area. In S. Möllers, W. Schlüter & S. Sievers (eds), Keltische Einflüsse im nördlichen Mitteleuropa während der mittleren und jüngeren vorrömischen Eisenzeit, Bonn, 311-325.

Roymans, N. 2007b, Understanding social change in the Late Iron Age Lower Rhine region. In C. Haselgrove & T. Moore (eds), The Later Iron Age in Britain and beyond, Oxford, 478-491.

Roymans, N. 2008, Muntgebruik in een dynamisch grensgebied. Keltische munten in de Nederrijnse region, Utrecht (Van Gelder-lezing)

Scheers, S. 1977, Traité de numismatique celtique II. La Gaule belgique, Paris.

Scheers, S. 1996, Frappe et circulation monétaire sur le territoire de la future civitas Tungrorum, Revue belge de numismatique et de sigillo­graphie 142, 5-51.

Schulze-Forster, J. 2002, Die latènezeitliche Funde vom Dünsberg, Marburg (Ph.D. dissertation Universität Marburg).

Schulze-Forster, J. 2005, Der Dünsberg und die jüngsten keltischen Münzen in Hessen. In J. Metzler & D. Wigg-Wolf (eds): Die Kelten und Rom. Neue numismatische Forschungen, Mainz (Studien zu Fundmünzen der Antike, 19), 159-181.

Theuws, F. 2004, Exchange, religion, identity and central places in the early Middle Ages, Archaeological Dialogues 10, 121-138.

Van den Brandhof, W. 2005, Ubische of Bataafse quinarii. Typologie en datering, Detector Magazine 82, 13-17.

Van Enckevort, H. & I. Joosten 2002, Eine Keltische Tüpfelplatte vom Kops Plateau in Nijmegen. In M.R. Alföldi, H.-M. von Kaenel & J.P.A. van der Vin, Die Fundmünzen der römischen Zeit in den Niederlanden, Band III,1: Nijmegen-Kops Plateau, Mainz, 37-39.

Van Heesch, J. 1998, De muntcirculatie tijdens de Romeinse tijd in het noordwesten van Gallia Belgica, Brussels.

Van Impe, L., et al. 1997/1998, De Keltische goudschat van Beringen (prov. Limburg), Archeologie in Vlaanderen 6, 1-124.