Not at random: Evidence for a regionalised coin supply?

Fleur Kemmers

Introduction In discussions on regions and regionality, the emphasis is usually on either the inhabitant’s point of view (to which community do I belong?) or on the strictly legislative aspect of which settlement, micro-region or pottery type belongs to which Roman administrative unit. This paper aims to explore another aspect of regionality, mainly in the military sphere, focusing on what the Roman authorities regarded as meaningful regional entities. The means to achieve this end is the analysis of the coin finds from several sites in the North-western Roman provinces, most importantly the coin finds from the Roman canabae legionis at Nijmegen, The Netherlands. Since the Roman state had the exclusive right to issue coins, and was responsible for their distribution to the troops all over the empire, a comparison of coins from different contemporaneous sites might show whether there was a regional diversity in the supply of coins, and if so, why.

First of all a short overview of previous research on the subject of coin supply will be given, in order to establish the current state of our knowledge. Secondly two remarkable aspects of the coin assemblage from the canabae legionis at Nijmegen will be discussed. Then the coins from several other sites will be compared to those from Nijmegen, to find out whether this site is exceptional or that a pattern can be observed. Thereafter an explanation for this pattern will be presented, followed by several final statements on regionality, coin supply and the Roman government.

Previous research Research on Roman coin finds on a more than purely descriptive level was until recently an almost strictly British and German affair. As a result, our knowledge of the Roman monetary policy and distribution of coin finds is almost solely based on coin finds from British and German sites. Nevertheless important observations were made. For Britain the work was mainly done by Reece (e.g. Reece 1984, Reece 2002) and Walker (Walker 1988), and more recently Hobley (Hobley 1998). With the exception of Walker’s research, such studies do not usually focus on individual sites, but on entire provinces or other large clusters of sites. Their most important conclusions, if restricted to the first century A.D., can be summarized as follows: the first decades after the conquest, there was no regular supply of coinage to Britain. While some years saw a massive influx of new coins, the next couple of years might be deprived of coins altogether. Until the end of Vespasian’s reign the years of an abundance of new coins coincided with activity at the mint at Lyon, which was only operative for intermittent periods. This last aspect seems quite logical, and is therefore usually taken for granted, since Lyon is considerably closer to Britain than Rome (the only other mint in the West in the first century), causing the costs and troubles involved in the transport of coins to be much smaller, and therefore a preferable option, if possible. One might ask however, why the mint at Lyon was not constantly operative, but minting coins in a limited range of denominations (mostly bronze coinage) in specific years only. Furthermore for the Domitianic period and beyond large differences were found between provinces, coins of year X, abundant in province Y, are a rarity in province Z and vice versa.

On the basis of coin finds from Germany scholars (Kraay 1956; Schulzki 1989) established that bronze coins, contrarily to precious metal coinage, show a very low mobility: once they were supplied to a region, they usually did not spread outside this region afterwards. Secondly a lot of research has been done (Wigg 1999; van Heesch 2000) on the coin supply to the Rhine and Lippe-areas in the Augustan period, the coin assemblages (as regarding the presence or absence of specific coin series) differing enormously per site. This turned out to have been caused solely by chronological factors however, each new campaign into Germany being accompanied by the distribution of new coin series.

New opportunities Thanks to a recent large scale excavation in the Netherlands the debate on coin supply has gained a new impetus. Between the years 1987 and 1997 part of the canabae legionis of the Legio Decima Gemina, stationed at Nijmegen, was excavated (Haalebos 1995; Haalebos 1999). On the basis of historical sources, and the excavations confirmed this knowledge, the site is known to have been inhabited only between approximately 70 and 105 A.D. (apart from a very short lived Augustan fortress on the same site over 80 years earlier (Haalebos 2002; Kemmers in press). Since all objects found during the excavations of the canabae legionis were supplied to or manufactured at the settlement in the Flavian period and discarded not later than 105, it is justified to regard it as a Flavian type-site. Apart from a wealth of small finds and pottery, this excavation has yielded over 2500 coins from stratified contexts. The importance of this assemblage does not only lie in the sheer number of the coin finds, but above all in the confined habitation period they represent. With the exception of the assemblages from large urban sites such as Rome and Pompeii, there are no –published– sites with such a large number of Flavian coins. It will be quite clear, that the coin finds from the canabae legionis at Nijmegen provide unique opportunities for studying in detail aspects of monetary policy and coin supply in the Flavian period.

Peculiar aspects of the coin assemblage from the Nijmegen canabae legionis In trying to reconstruct the supply of coins to Nijmegen, and the motives lying behind this, two aspects of the coin assemblage might be used. First of all it is striking that over 40 percent of the Flavian coins found in the canabae, had been issued in four years only (Fig. 2). Given the fact that the site was inhabited continuously for thirty years, by a more or less constant number of people, who received a fixed amount of pay each year, one would expect the same number of new coins to have been supplied to the site each year. Furthermore the coins supplied to the site in those peak-years are very homogenous in their reverse-types. While dozens of reversetypes were being minted at both Rome and Lyons, eight reverse types account for over sixty percent of the coins of those four years (see below). Is this merely coincidence, or can other explanations be sought?

Figure 1: The distribution pattern of the quadrans RIC 436, issued by Domitian. 1: Valkenburg; 2: Alphen aan den Rijn; 3: Rossum; 4: Vechten; 5: Nijmegen; 6: Neuss; 7: Köln-Altenburg; 8: Heidenburg; 9: Mainz; 10: Zugmantel; 11: Hofheim; 12: Frankfurt-Heddernheim; 13: Rottweil; 14: Mirebeau; 15: Augst; 16: Vindonissa; 17: Avenches (Kemmers 2003).

A second remarkable aspect of the coin assemblage from the Nijmegen canabae legionis is the extraordinarily high amount of quadrantes, the smallest denomination in the Roman monetary system. Over 400 of them were found, 98% of the identifiable ones (316) being of the same type, issued by Domitian somewhere between A.D. 82 and 83 (Carradice 1983: 124). In a previous paper (Kemmers 2003: 33–34) I argued that the quadrantes are not an anomaly, but a continuation of a need for small change, felt since the Augustan era in the Rhine-area, being caused by a low-price level in this frontier region. What concerns us here however, is not the purely economic aspect, but why nearly all the quadrantes are of the same type. This implies that a consignment of those specific coins was sent to Nijmegen, but why, when, and what other places did it reach?

Regionality? To start with the quadrantes first, an inventory of the distribution pattern of the Domitianic quadrans-type found at Nijmegen (RIC 436) was made (Kemmers 2003: 19–27). As shown in Fig. 1 the type has a rather restricted distribution pattern: while it has not been found in Britain, large parts of Belgium and France and the Mediterranean (with the exception of Rome, not on the map), there are large concentrations of the type along the Rhine. Even more important however, is the fact that nearly all of them were found in military settlements, clustering around the legionary fortresses of Nijmegen, Neuss, Mainz, Vindonissa and Mirebeau. The unifying aspect of those sites, is that those five sites were the headquarters of the legions of the Rhine-army in the Domitianic era (coin finds from Bonn, also a legionary fortress at that time were not available). Apparently only those legions received a special consignment of quadrantes.

A solution presents itself in the war against the Chatti, a transrhenane tribe in A.D. 83 (Strobel 1987). For the first time since Caligula had led his troops to the Rhine in A.D. 39, an emperor himself, together with his imperial staff, came to lead the legions of both the Upperand Lower Rhine army into war. Furthermore it has been proven that the quadrantes were struck around A.D. 83 (before the imperial salutation of Domitian) and that there was a need for small change in the region. Those three pieces of evidence combined seem to lead to the conclusion that the imperial authorities, being confronted themselves with the situation, took measures to provide in the need for small change of those units involved in the war. Though small change might have been needed in other regions, quadrantes were not supplied there, most likely because the authorities simply did not know or did not care.

Regional differences in the supply of bronze coins As has been argued above, while some regions were supplied with coinage in specific years, others were not. It is not within the scope of this paper to discuss why this might have been the case (with the exception of the extraordinary case of the quadrantes), but rather to focus on several aspects of the supply. Since precious metal coins are rare as site finds, we will restrict ourselves to bronze coinage. To compare the situation at Nijmegen with that elsewhere, two other sites with a reasonable number of coin finds were selected in different regions, Richborough (Reece 1981), in Britain and Hofheim (Gorecki 1994), in Germany (near Mainz). All three are military sites during the Flavian period. In Fig. 2 the percentage of the Vespasianic and Domitianic coins from those sites per year is shown. While the three sites share their peaks in the years 71, 72 and 77–78 (though the emphasis differs), the peaks in Domitian’s reign are different for Britain. This has been observed before (Hobley 1998: 22), so we will not go into further detail here.

0 5 10 15 20 25 30 35 40 45 70 71 72 72-73 73 74 75 76 77-78 79 81 82 84 85 86 87 88-89 90-91 92-94 95-96 Year Percentage Nijmegen n = 758 Hofheim n = 159 Richborough n = 226 Figure 2: Bronze coin percentages per year for A.D. 70–79 and A.D. 81–96 at Nijmegen (author’s data), Hofheim (Gorecki 1994) and Richborough (Reece 1981).

Though on first sight the peaks in the Vespasianic period are very similar for all three sites, which could lead one to assume one homogeneous bulk of coins being supplied to the armies in the west, they differ in an important aspect: their reverse-types. While a wealth of reverse types were minted at Lyons (the provenance of the majority of the coins on the three sites) in A.D. 71, on each of the three sites, two or three reverse types represent over fifty percent of the coins for that year. However, for each of those sites the dominant coin types are different. As shown in Fig. 3, 4 and 5 at Nijmegen the types of Victoria (Victory), Securitas (Security) and Aequitas (Equity, Stability) dominate, while at Hofheim the former two and the reverse-type depicting the goddess Rome are in the majority. At Richborough by far the most common type is the one depicting the Roman eagle on the globe (the Aquila), followed by Securitas, Victoria and Providentia (Providence).

A.D. 71 Nijmegen Aequitas 20% Securitas 16% Victoria 27% Other 37% n = 142 Figure 3: Ratios of the different reverse-types on bronze coins of the year A.D. 71 at Nijmegen (author’s data).

A.D. 71 Hofheim Roma 21% Victoria 14% Securitas 11% Other 54% n = 28 Figure 4: Ratios of the different reverse-types on bronze coins of the year A.D. 71 at Hofheim (Gorecki 1994).

A.D. 71 Richborough Eagle 28% Providentia 14% Securitas 13% Victoria 12% Other 33% n = 90 Figure 5: Ratios of the different reverse-types on bronze coins of the year A.D. 71 at Richborough (Reece 1981).

Is this merely coincidence, or are other factors at work? To answer this question we should look at the political situation in those three areas in the year 71. At Nijmegen the Batavian revolt of 69/70 (Bijvanck 1943: 221–266), as well as the civil war of 68-69 had led to great upheaval. The Batavians had been defeated, but to the cost of great losses on both sites. To control the Batavians, the Tenth Legion Gemina was sent to Nijmegen, which had not seen such a large military presence in decades. At the same time the garrison forces at the forts to the west of Nijmegen were replaced by new ones, not consisting of local recruits. The Batavian revolt had not reached as far as Hofheim, but the troubles of the civil war had not gone unnoticed. The two legions at the fortress at Mainz had been the first to support one of the usurpers in the year of the four emperors, Vitellius (Wellesley 2000: 15–16 and 34–35). In Britain the year 71 saw the arrival of the new governor Cerialis, who was to embark on large campaigns against the Brigantes (Webster 1970: 196-197; Birley 1973: 187-90), together with the newly appointed commanding officer of the Twentieth Legion Valeria Victrix, Agricola (Gaheis 1917: 128–129). Is there an attempt to communicate some basic messages to those regions by means of pictures on the coins?: Victory, Security and (above all) Stability to the Lower Rhine, Stability, Security and Roma(n rule restored?) to the Upper Rhine and Victory and the eagle (the emblem of the legions) to Britain?

Not at random The best way to verify the tantalizing statement made above is to work it the other way round. Is there a political event known for a specific region in a specific year, does the message on the coin finds match this event and, most importantly, is it different in another region? A likely candidate for this test is the revolt of the governor of Upper Germany, Saturninus in 89, who declared himself emperor, supported by the troops of Upper Germany (Strobel 1986). However, the legions of Lower Germany remained loyal to the emperor Domitian and crushed the revolt. As a sign of his gratitude, the legions of Lower Germany were henceforward entitled to call themselves ‘pia fidelis Domitiana’. As was shown in Fig. 2 there is a peak in coin finds from the year 90–91 in the entire Rhine-area. As can be read from Fig. 6 at Nijmegen nearly fifty percent of the coins issued in these years show the reverse type Virtus (bravery, courage, martiality). Is this a message of the government to their loyal troops, who defended the rightful emperor?

A.D. 90-91 Nijmegen Virtus 47% Fortuna 20% Moneta 24% Other 9% n = 183 Figure 6: Ratios of the different reverse-types on bronze coins of the year A.D. 90-91 at Nijmegen (author’s data).

It should be noted however, that the pictorial repertoire of Domitian’s bronze coinage in these years was far from diverse. Only the reverse types of Fortuna (Fortune), Moneta (the personification of money), Virtus, Salus and Iupiter Custos (Iupiter the protector) were issued in noteworthy quantities. On the other hand, the reverse type distribution at Hofheim, supporting the wrong side in 89, is quite different (Fig. 7). Though Virtus is present, other types dominate. Unfortunately the number of coins for this year from Hofheim is rather limited, but sites with larger numbers are rare. As a final check therefore, we grouped the coin finds from a number of sites in the area of Hofheim, viz. Frankfurt-Heddernheim (Gorecki 1994) and Mainz (Franke 1960) together. As shown in Fig. 8 the three main reverse types of Domitian’s bronze coinage appear in comparable numbers, quite different from the overwhelming dominance of Virtus at Nijmegen.

A.D. 90-91 Hofheim Virtus 19% Fortuna 31% Moneta 37% Other 13% n = 16 Figure 7: Ratios of the different reverse-types on bronze coins of the year A.D. 90-91 at Hofheim (Gorecki 1994)

A.D. 90-91 Hofheim area Virtus 31% Fortuna 26% Moneta 26% Other 17% n = 58 Figure 8: Ratios of the different reverse-types on bronze coins of the year A.D. 90-91 in the Hofheim-area (Franke 1960).

Conclusions Although the sites with coin assemblage large enough for a meaningful analysis are rather limited in number, several general remarks can be made. Based on the evidence presented above, it looks like the supply of bronze coinage in the Flavian period was regionally differentiated. Those regions were not so much defined by their administrative division but by their military presence. Since paying the army was the most important motivation for the Roman state to issue coins in the first place, and, certainly in the first century in the frontier zone, the only way coins were supplied to a region was by means of military pay, this is perhaps not so surprising. What is remarkable however, is that the supply of bronze coins to the various units was not a homogeneous bulk. Rather two factors seem to be at work: in the first place economic factors in a given region determined the kind of coins supplied to that region, though probably only if the government was aware of it and, more importantly, considered it a problem too. Secondly the differentiation in dominant reverse types seems to have been caused, at least to a certain degree, by the government’s intent to distribute some basic messages, related to the political situation in a given region.

This final aspect brings us to the sensitive theme of propaganda by means of the pictorial repertoire on coins. Though this discussion has been, and still is, long and vehement (for a summary of the discussion Wolters 1999: 255–265), most scholars have failed to look at the distribution of coin types in the Roman empire. On the basis of this paper I would argue, that especially the reverse types on bronze coins certainly served a propagandistic purpose. Though the first interest in minting coins was of course the payment of troops, an interesting sideaspect of those coins will have been their propagandistic value. This propaganda was perhaps, as demonstrated above, not very elaborate, but the repetition of the same basic message over and over and over again on coins used on a daily basis (which bronze coins were to a much larger extent than silver, let alone gold coins) might have been effective. The results of this paper are intended to start a discussion on this subject. To reconstruct the Roman monetary policy regarding the supply of coins to the troops a lot more research has to be done. What is needed is the publication of coin lists of the major Roman military forts and fortresses. Only in this way will it be possible to provide answers to the question of regionality and coin supply, not just regarding the distribution of reverse types, but also regarding the differing peaks in coin supply per year.

Department for Provincial-Roman Archaeology, Radboud University Nijmegen, The Netherlands