A frog s eye view of the Roman market: the Batavian case

A frog s eye view of the Roman market: the Batavian case By Joris Aarts 1. Introduction Since the seventies of the previous century, many things have been said about the character of the Roman economy and the extent of its monetary economy. This discussion was led by ancient historians like Finley, Crawford, Hopkins, Howgego, and Duncan-Jones, often using monetary circulation as a measure of economic integration. 1 Both literary and archaeological sources were used to create large-scale models of the Roman economy, concentrating on questions of monetization of the economy: how fast, how wide, how deep. As archaeological evidence for monetary circulation mostly coin hoards and museum collections were used: large-scale inventories of stray finds were not yet available, or deemed too time-consuming to analyse. In the last ten years, however, much more work has been done on a regional level, leading to regional inventories and analyses of both hoards and stray finds. 2 At first, scholarly discussion about the character of the ancient economy was dominated by the question whether or not the ancient economy was fundamentally different from modern western economies and if, by extension, modern economic analysis and terminology could, and should be applied to ancient economies (the formalist-substantivist debate). This ultimately resulted in a tie, and in the last two decennia the general agreement seems to be that both substantivist and formalist models and methods can be used, depending on the topic of research. It must be said, that recently many ancient historians are inclined to embrace modernist analyses 3, particularly when dealing with macro-economic analyses. Also, much attention is paid to the extent and nature of credit systems in the (Roman) economy, and how this interacted with the circulation of physical money (coins) and bullion 4. On the other hand, more cultural approaches to coinage and money appear, although, until now, they tend to stay within the realm of the Greek world. 5 One returning point in the discussions about the Roman economy is the regionality of economic systems. Although there is some evidence for empire-wide economic integration in the long-distance trade in luxury items and foodstuffs and chains of credit relations the circulation of coin suggest that the province or even smaller regional units pressed their mark on coin circulation. Analyses of tax paying-systems show a great regional variation in time and place. The strange thing is that Hopkin s model of taxes and trade, which propagates the empire-wide rise of a monetary economy through the interplay between long-distance trade and money taxes, still is accepted by most scholars as explanation for the monetization of the Roman economy. Perhaps it is time to reconcile macro-economic views of the structure of the Roman economy with the interpretation of historical and archaeological data on a regional 1 Finley 1992 (reprint 1973); Crawford 1970; Hopkins 1980 (and 2002); Howgego 1992; Howgego 1994; Duncan-Jones 2002 (1990). 2 See for instance Aarts 2000; Katsari 2008 and many more. 3 Temin 2001; Temin 2004; Harris 2006 ; 4 Harris 2008; Lo Cascio 2003; Verboven 2009. 5 I mean the work of Von Reden, Schapps, Kurke. For an elaborate discussion, see Aarts 2005. 1

and local level, and to test the models created by the ancient historians mentioned above. What I propose to do in this article is to approach the Roman economy from the bottom up, not from the top down, to see if we can connect the macro-economic models with archaeological and historical data on the level of a civitas. I will try to do this thematically, taking the civitas batavorum as a case study. 6 The case of the Batavians 2.1 A very brief history of the Batavians The Batavian civitas can be located in the eastern half of the Dutch river area (figure 5), although its southern border may have extended to part of the sandy plateaus of the southern half of the Netherlands. Nijmegen (Ulpia Noviomagus) was the capital of this civitas, which lay on the border of the Roman empire during the first three centuries AD. The Batavians themselves were probably a construct of the Roman authorities, a group consisting of transrhenian migrants and the local population of the Dutch eastern river area, which may have been a remnant of the tribe of the Eburones. The civitas and the Batavians can be followed historically into the first half of the 3 rd century AD, after which they were likely replaced by Frankish migrants from across the Rhine border. We may assume that this also meant the end of the Batavian civitas. We know particularly much about their early history through the narrative of Tacitus, in which they play an important role in a native (Batavian) revolt during the years AD 69-70. Before this uprising, the Batavians were, according to Tacitus, a gens foederata: a people with a special treaty with the Romans, the only known but main- point of which is the exemption from paying regular taxes to the Roman government. In exchange, the Batavians were enlisted in the Roman army on a grand scale. 7 Vossen has calculated a minimum average of 1.3 young men per household (!) who were serving in the Roman auxiliaries. 8 Also, in the pre-flavian period these troops seem to have been led by officers who were recruited from the same tribe. This probably all changed after the Batavian revolt, although the treaty was continued, probably along other lines (which we do not know from historical sources). The Batavian civitas was a densely-settled area, consisting of a string of auxiliary forts along the Rhine, a city (Nijmegen), some badly-known central places (vici), a few large rural sanctuaries and a large number of simple and small rural settlements. Some of these were the object of large-scale excavations in the previous decades, providing us with a few highquality archaeological data sets. 9 Using these data, contextual analysis of material culture provides us with a detailed picture of the social, political and economic integration of the Batavians in the Roman empire. 6 The Batavians were chosen as example because they are relatively well-known, both historically and archaeologically (see Roymans 2004, Roymans/Derks 2002, Nicolay 2007, Groot 2008; Heeren 2009, Vos 2009) 7 One might argue against Tacitus that the Roman authorities in all probability considered this recruitment as a form of tributa (see for instance Aarts 2000, 16-17). 8 Vossen 2003, 421. See also Vos 2009, 216-225. 9 Tiel-Passewaaij (Heeren 2009 and Aarts/Heeren 2010); Geldermalsen Hondsgemet (Van Renswoude/Van Kerckhove 2009); Wijk-bij-Duurstede-De Horden (Vos 2009). 2

2.2 Money supply The coin finds of the Batavian area were part of a recent large-scale inventory of Roman and Late-Iron-Age coins in Germania inferior and Gallia Belgica. 10 Figure 1 shows the chronological distribution of these finds in three areas of study: the Dutch river area, encompassing the civitates batavorum and cananefatium; the area between the rivers Meuse, Demer and Scheldt (part of the civitas tungrorum), and the area of Luxemburg and Trier (part of the civitas treverorum). A grand total of 62,518 coins was recorded for all three areas. 11 Because we are dealing with large numbers of coin finds, I propose that they are representative for the (physical) money supply to these areas during the Roman period. Figure 1. Coin supply to the Dutch river area, the Meuse-Demer-Scheldt area and the Luxemburg-Trier region, in percentages of the total. The main difference between the three areas is clear. In the Dutch river area, we see high levels of coin supply in the 1 st century AD (peaks in the Augustan and Flavian periods); from the beginning of the second century, a downward trend is visible, from which this area will never recover. In contrast, the curve of Luxemburg and Trier stays stable but low until the late 3 rd century AD, after which coin supply soars. Even when one takes into account the changes in the character of late-roman money, the proportional differences between these areas remain the same. The MDS area lies not only in geographical sense between the two others, but also in terms of coin supply. 10 Aarts 2000. 11 Only a part of these coins could be dated well enough to fit one of the issue periods. The missing coins and late-iron-age coins were omitted from the graph. Also, the coins of the fort of Nijmegen-Kops Plateau were excluded, because they would cause a grotesque distortion of the data of the Dutch river area. 3

The reason for the high supply levels in the Dutch river area in the 1 st century lies in all probability in the fact that it was a military zone. The presence of large numbers of troops in the 1 st century, and the building (and maintenance) of forts and other infrastructure cost a lot of money. After the army was greatly reduced in numbers at the start of the 2 nd century, coin supply decreases and never reaches 1 st -century levels any more. This implies that the army played a key role in the monetary economy of this area, in the sense that it represented its main monetized market. In the area of Luxemburg and Trier we see a different picture. Here, admittedly after a modest peak in the Augustan period, levels of coin supply stay low, until the end of the 3 rd century. Also in this case, there seems to be an evident explanation, because Trier became an imperial residence in the 4 th century AD and host to one of the mints. This does not mean, however, that the actual coin mass which reached this area was insignificant: as can be seen, the Luxemburg-Trier area shows the largest numbers of coin finds of all three areas (fig. 1), although the actual coin numbers in figure 2 show that this is an effect of the late Roman period only. Figure 2. Coin supply to the three regions until AD 260: on the y-axis, raw coin numbers The main incentive for coin supply to all regions which emerges here are governmental costs, and in particular military expenditure. And indeed, most authors seem to agree that agree that the Roman government was not interested in maintaining a steady flow of fresh coin for the sake of regional markets. 12 But in all of the three areas, the coin supply to these areas cannot be explained by governmental spending alone. The presence of the Roman army in the Treverian area is too low to account for the high numbers of coin finds there, and figure 3 12 However, contra this see Heinrichs 2000. 4

suggests that coin supply started not much later than in the Dutch river area. It is clear that there must have been other mechanisms of coin supply in play. In the case of Trier, we may assume that the intensive building program in the 1 st century AD has had a monetizing effect on the economy of the city. At the same time, the development of villa s producing for the urban market may have profited from this monetized market. One of the disadvantages of using coin finds to analyze the money supply is that they are heavily biased towards bronze coin. 13 Silver and gold in particular were less regularly lost, and more often recovered. Also, the chance of gold coin finds being reported is many times less than that of bronze. Still, comparing the chronological distribution of gold, silver and bronze coins shows some interesting points (figure 4) 14. The Augustan peak we saw above in the Dutch river area is absent in the silver and gold. Instead, only from the Flavian period onwards silver and gold appear in large numbers. Furthermore, the Antonine peak in silver coins, that at the beginning of the 3 rd century and the two peaks in the supply of gold coin in the second half of the 4 th century spring into view. Figure 3: percentages of gold, silver and bronze coins in the DRA and MDS area (data from NUMIS) The Flavian increase in gold and silver may be linked with a change in the structure of the tax system, accompanied by an increase in (monetized?) exchange in regional pottery and metal ware. I will return to this later. 13 This becomes eminently clear when the bronze: silver/gold ratio is compared with the finds of Pompei (Duncan-Jones 2003, the critique of Andreau notwithstandingly (Andreau 2008)) or the coin finds of the battlefield of Kalkriese (Berger 1996). 14 The data of figure 3 are not exactly the same as in figures 1 and 2; here, the coin finds of the Dutch river area and the Meuse-Demer-Scheldt area are grouped together, hence the different coin totals. The data are taken from the database of the Geld-en Bankmuseum in Utrecht (NUMIS), including the coin finds after 1997 which are not part of the earlier mentioned survey. 5

Coin circulation Once Roman coin reached the Dutch river area, what became of it? Did it linger in the Roman army camps and the city of Nijmegen, never to reach the widely scattered rural villages of the hinterland? The geographical distribution of coin finds in figure 4 shows clearly that this is not the case. The coin finds are not only concentrated in the forts along the Rhine and the cities of Nijmegen and Forum Hadriani (capital of the civitas cananefatium), but are strewn across the countryside as well, particularly in the civitas batavorum (eastern half). Figure 4: coins finds in the Dutch river area before AD 70 Figure 5: coins from military and non-military sites in the Dutch river area Moreover,there seems to exist virtually no time lag between the coin supply to the forts and coins reaching the rural settlements. This becomes evident when one compares the chronological distribution of coins from forts and from rural settlements: they are virtually 6

identical (figure 5). 15 A second indication for the quick spread of Roman money over the countryside is the frequent occurrence of coins with pre-flavian countermarks on rural sites. 16 The next question is of course by which mechanisms these coins flowed so easily from the prime target of Roman money supply (the forts) to these villages. Does this say anything about the speed of monetization of the Batavian economy? Earlier, I have made the point that a large part of the Batavian male population served in the Roman army in the early Roman period. The effect of this was that not only they were paid in Roman coin, but they also quickly learned how to use it in a Roman way. Part of their wages were probably spent in the forts themselves and the adjacent camp villages, but since they were able to visit their home villages on a regular basis during leave, it is certainly not impossible that they brought the money with them, along with their acquired knowledge of monetary ways. Althought this did not mean that exchange within these villages became monetized, the people who lived there possessed money and the knowledge how to use it. The monetary markets in this area were located in, and just outside of the forts and in the city, and this is were monetary exchange took place. In the pre-flavian period, there is little evidence for a lively economic interaction between the rural settlements and the urban and military markets: percentages of local pottery wares remain high in the rural world. There is no surplus production of grain and other crops that could be sold on the market. We have evidence for surplus production of cattle and horses, but this seems to have started around the middle of the 1st century AD. 17 The possibilities for the conversion of agricultural surplus into money seem limited, in any case for the pre-flavian period. Also, the tax system did not invite people to use money on a substantial scale. Batavian taxes were paid in manpower, not in money, nor in agricultural produce. This all changed in the Flavian period. Although the ethnic recruitment of Batavians did not stop, it is not likely that it remained the basis of the taxation in the Batavian civitas. The Batavian revolt came as an extremely unpleasant surprise for Roman authorities, and they would have looked very critically to the recruitment of Batavian soldiers and where they were based. 18 The re-evaluation of the terms of the treaty after the Batavian revolt and the creation of a formal civitas organization under Domitian imply that the rules of taxation may well have changed. Archaeological indicators for this change are the increased production of meat, hides and horses for the military and possibly urban markets. Also, the inner structure of rural settlements changed: in the Flavian period, field systems appear which re-structured both the internal space of the settlements themselves and the surrounding fields. Vos has made a plausible case for the field systems of the settlements at Wijk-bij-Duurstede De Geer and De Horden having been based on the Roman actus. 19 In the settlements themselves, larger structures appeared for the storage of agricultural produce. All these changes imply a different approach to the ownership of land and its yield. Perhaps this can be linked to a shift in the taxation system to payments in agricultural produce and/or money; if this is the case, the area would have been subjected to a new census. 15 The only difference seems to occur in the Claudian period: the military sites show an increase here, but this is probably an effect of the fact that most of the forts in the western zone of the Dutch river area were established somewhat later than those in the east. 16 See Aarts 2003. 17 Vos 2009, 253. 18 See also Van Driel-Murray 1994, 93 for the probability of a restriction on the possession of weapons by citizens after the Batavian revolt; this is corroborated by the change in the finds of militaria in rural contexts (see Nicolai 2007; Vos 2009, 197. 19 Vos 2009, 109-216. 7

At the same time, percentages of regional pottery and metal wares increased in the material culture of rural settlements. 20 Coins reached another high in the Flavian period, and this time silver and gold coin joined the bronze (fig. 3). The latter could have been an effect of the alteration in taxation, but may also have been a consequence of the increased building in the area. The city of Nijmegen grew steadily, although it stayed essentially a backward town compared to others. Nevertheless, also in the countryside there is evidence for monumental architecture in the temples of Kessel-Lith and Elst. Additionally, also the military infrastructure continued to grow during the Flavian period. The civilian architecture would have been paid for by elite members of the civitas batavorum. In order to do this, they must have lent sums of money (from patroni outside of the civitas), because this would not have been available (yet) through the workings of a monetary economy on the civitas level. This argument for the role of credit in the monetary economy is not backed up by archaeological evidence. There is no complementary flow between supply of credit and physical money. That is, not in the sense that credit took over when money supply was low: if anything, the coin evidence seems to imply that an increase in credit meant an increase in the physical money supply as well. Figure 6. Model of the coin circulation in the Batavian civitas before and after AD 69 (after AD 69 box above left added ). The drastic reduction of Roman troops in the Dutch river area at the beginning of the 2 nd century took its toll. After this, the general trend in the coin supply is a downward one (fig. 1). In any case, a substantial less amount of money flowed to the area in the form of army wages. 20 Van Kerckhove 2009, 189. 8

This must also have had repercussions for the exchange between the rural settlements and the army markets, whether or not through urban channels or the tax system (direct army supply contracts?). Both the official supply of money and the income from surplus production dwindled through the shrinkage of the military market. At the same time, debts had to be paid back. This must have put some economic pressure on the urban elite. Perhaps in recognizance of this, Trajan extended municipal rights to the city of Nijmegen (Ulpia Noviomagus). However, for the Batavian population there was no real alternative market to turn to. No economic interactive exchange system between the urban settlements and the countryside developed, like in the case of the civitas treverorum. In this light, the Antonine peaks in silver coins and in both gold and silver during the Severian period seem unexpected. The Antonine increase in the supply of silver may have had some relation with the construction of a city wall in Nijmegen, but it is harder to explain the increase in gold and silver coin under the Severians. However, the same Severian increase in silver coin is visible in the coins finds of Luxemburg, and perhaps we must look for an explanation in the changing money system and higher price levels. The absolute bottom: the coin finds of two rural settlements Descending form a regional level of analysis to a local one, what can the coin finds of two average rural settlements in the Batavian area contribute to a discussion of the Roman economy? In the manner of research, the settlements of Tiel-Passewaaij and Geldermalsen- Hondsgemet are far from average. Both were excavated almost completely (Tiel-Passewaaij) or for the largest part (Geldermalsen-Hondsgemet); also, they were excavated by the same team of archeologists which used a system of intensive metal detection. In this way, large numbers of coins were retrieved from both sites: 347 coins at Tiel and 320 coins at Geldermalsen. Considering that these were simple rural villages consisting of only a few farmsteads per phase, the number of coins is impressive. Tiel-Passewaaij Geldermalsen- Hondsgemet coins coins Hoards 105 30 Settlement finds 221 290 Cemetery finds 21 - Total 347 320 Table 1. The coin finds of Tiel-Passewaaij and Geldermalsen-Hondsgemet This is even more so when one extrapolates these numbers to the other rural villages plotted on the map in figure 7; we have no reason not to, since the rural villages of Tiel and Geldermalsen are representative for the majority of rural settlements in the Batavian civitas. During large-scale excavations at Tiel-Passewaaij a rural settlement and a large cemetery (400 graves) came to light, lying in the heart of the Batavian civitas. Its occupation starts in the late Iron Age and ends in the first half of the 4 th century AD. The village itself consisted of five to seven farms which were occupied at the same time. In the last few decades of the 3 rd century AD, the population of the settlement changed dramatically, which is attested by changes in the material culture. It is thought that new groups of Franks settled here, who had little or nothing in common with their Batavian predecessors. We do not know 9

where these Frankish people buried their dead. This group continued to live there until around 350 AD, after which the settlement was abandoned. The settlement at Geldermalsen is very similar to that of Tiel, but we do not know where they buried their dead. Almost a third of the coins of Tiel came from three hoards: two (or possibly one) Severan silver hoards which was buried inside a house, and a hoard of 60+ barbarous radiates which was recovered from the cemetery, buried at a moment it was no longer in use (table 1). The only hoard of Geldermalsen is more modest in size: a hoard of 30 barbarous radiates found in the corner of a ditch surrounding the late-roman settlement. The chronological distribution of the coin finds of both settlements are displayed in figure 8, set off against the general coin supply to the Dutch river area. I will not present an exhaustive discussion of all aspects of the coin finds of these two settlements, but rather select two points which are pertinent to our theme. 21 Figure 7. Coin finds in the Dutch river area 50 BC- AD 402 The first point is that the spatial distribution of coins does not imply that they were lost during monetary transactions. They were found in or nearby the houses of the village. There is no indication whatsoever for local market places at which periodically monetary exchange took place. In fact, the only probable monetary exchange which I can think of is that with outsiders. The people of the village were closely related, and it is hard to imagine people of such communities exchanging things through the intermediation of Roman money. In consequence, the population would not have run around daily carrying money purses. This leaves us with two options for the interpretation of the stray finds: either they were used in a ritual context, as we can clearly see in the case of the cemetery finds, or they were part of (small) hoards waiting to be used again in a monetary context (outside market). This does not mean, of course, that the rural population was not monetized. As I have shown, Batavian communities learned quickly how to use money in market exchange and did so, albeit in monetary market centers. 21 For a more elaborate account, see Aarts 2007; Aarts 2009; Aarts and Heeren 2010. 10

The second point is related with the spatial distribution of coins: contextual analysis revealed that from the second century onwards, there was a large percentage of old money in circulation. This phenomenon was already known from funerary finds, but many 2 nd -century coins could be assigned to 3 rd -century structures. Thus, the seemingly continuous supply of coins emerging from the chronological distribution from the settlement finds (figure 6), is misleading. A more likely scenario is that, particularly from the second century onwards, money reached the settlement in periodical spurts. Again, this perceived slowness in the circulation of rural money (or periodicity in coin supply) does not imply that the Batavians living inside the villages were not involved in the monetary economy. It simply means that they not always had the means (surplus) or need to participate in monetary exchange. Figure 8. The coins of Tiel-Passewaaij, Geldermalsen-Hondsgemet and the Dutch River area. 11

Conclusion In the above I hope to have shown you how analysis of coin finds on a regional and local level can be tied to the general discussion of the Roman economy. The comparison of trends in the coin supply to different regions supports the view that the particular historical setting of a region leaves its marks on the chronological distribution of coins, not only of bronze coin which is believed to remain where it ended up, but also silver and gold coin display chronological patterns which are characteristic of the region. In the Batavian civitas, there are clear indications that the presence of the Roman army was an important influence on the levels of coin supply, and at the same time serving as a monetary market for surplus products. This market function started to play a significant role in the regional economy from the second half of the 1 st century AD, and seems to have had a monetizing effect. However, through the extensive enlistment of Batavian soldiers in the Roman army, Roman money and ideas about its use had already spread across the countryside. As a result of the Batavian revolt, the tax system probably underwent rigorous change in the Flavian period, from a tax based on supply of manpower to the Roman army to a regular land tax. The revolt was only an immediate cause for this change, but the fundament for the new tax system was perhaps already laid down by the monetization of the Batavian mind in the pre-flavian period and by the increasing surplus production for a monetary military market. The flexibility of the Roman tax system would leave enough room for periodical local shortness of cash money, as is shown in figure 9. Figure 9. The various possibilities for the flow of money, goods and taxes in the Batavian civitas. The dependence on the military market also was the cause of a severe blow to the monetary economy at the beginning of the 2 nd century with the drastic withdrawal of Roman troops from the area. However, the monetary market was not limited to the army alone: there is evidence for monumental building both in Nijmegen and the countryside, which was paid for by the urban Batavian elite, at first using loans from extra-regional patroni. These building 12

projects must have had a further monetizing effect on the Batavian economy, although in itself it was not enough to replace the vacuum left by the army. Non-military large-scale spending may be linked to the peaks in gold and silver coins in the second half of the 1 st century and the 2nd century AD. If this assumption is correct, it implies that the extension of credit (most clearly in the Flavian period but also possible for later periods) had a multiplying effect on the supply and circulation of physical money as well. The analysis of coin finds of individual settlements show that a lot of old money was around, particularly from the second century onwards. It seems that one of the effects of the reduction of troops was a decrease in the velocity of circulation of money in the rural world. The connection of rural sites with the monetary economy was continuous, but clearly periodical: periods of money flow seem to alternate with periods of less activity. However, it would be wrong to think that the countryside was less monetized than the cities and military sites: the same people moved in different circles and whether people participated in the monetary economy was determined not so much by where they lived, but by the context in which they found themselves. Most coins found on rural sites are by consequence either used in a ritual way (in graves and other ritual depositions) or were part of small hoards, which were disturbed in later periods. Bibliography Aarts, J. G., 2000: Coins or money? Exploring the monetization and functions of Roman coinage in Belgic Gaul and Lower Germany 50 BC -AD 450. Dissertation Free University Amsterdam. Aarts, J. G., 2003: Monetization and army recruitment in the Dutch river area in the early 1st century AD, in: R. Wiegels and S. Seibel Kontinuität und Diskontinuität. Germania Inferior am Beginn und am Ende der römischen Herrschafft. Berlin: 145-161. Aarts, J., 2005: Coins, money and exchange in the Roman world. A cultural-economic perspective, in: Archaeological dialogues : Dutch perspectives on current issues in archaeology, ISSN 1380-2038 12(1), 1-27 (27). Aarts, J. G., 2009: Romeins geld in Geldermalsen-Hondsgemet, in: J. v. Renswoude and J. v. Kerckhove Opgravingen in Geldermalsen-Hondsgemet. Een inheemse nederzetting uit de Late IJZertijd en Romeinse tijd. Amsterdam: 287-299. Aarts, J. and S. Heeren, 2010: Het grafveld aan de Passewaaijse Hogeweg. Opgravingen in Tiel-Passewaaij 2. Amsterdam, Zuidnederlandse Archeologische Rapporten 41. Andreau, J., 2008: The Use and Survival of Coins and of Gold and Silver in the Vesuvian Cities, in: W. V. Harris (ed.), The monetary systems of the Greeks and Romans. Oxford, 1, 208-226. Berger, F., 1996: Kalkriese 1. Die römischen Fundmünzen. Mainz. Crawford, M., 1970: Money and exchange in the Roman world, in: The Journal of Roman Studies 60, 40-48. 13

Derks, T. and N. Roymans, 2002: Seal-boxes and the spread of Latin literacy in the Rhine delta, in: Journal of Roman Studies-Supplementary series- 48, 87-134. Duncan-Jones, R., 2002: Structure and scale in the Roman economy. Cambridge. Duncan-Jones, R., 2003: Roman Coin Circulation and the Cities of Vesuvius, in: E. Lo Cascio Credito e moneta nel mondo romano : atti degli Incontri capresi di storia dell'economia antica (Capri, 12-14 ottobre 2000). Bari: 161-180. Finley, M.I., 1999: The Ancient Economy, Berkeley. Sather classical lectures 43. Groot, M., 2008: Animals in ritual and economy in a roman frontier community: excavations in Tiel-Passewaaij. Amsterdam, Amsterdam Archaeological Studies 12. Harris, W. V., 2006: "A revisionist view of Roman money." in: The Journal of Roman Studies 96: 1-24. Harris, W. V., 2008: The nature of Roman money, in: W. V. Harris (ed.), The monetary systems of the Greeks and Romans. Oxford, 174-207. Heeren, S., 2009: Romanisering van rurale gemeenschappen in de civitas Batavorum: de casus Tiel-Passewaaij. Amersfoort, Nederlandse Archeologische Rapporten 36 Heinrichs, J., 2000: Überlegungen zur Versorgung augusteischer Truppen mit Münzgeld, Ein neues Modell und daraus ableitbare Indizien für einen Wandel in der Konzeption des Germanenkrieg nach Drusus, in: L. Mooren Politics, administration and society in the Hellenistic and Roman world, Proceedings of the international colloquium, Bertinoro 19-24 July 1997. Leuven: 155-213. Hopkins, K., 1980: Taxes and trade in the Roman Empire (200 BC-AD 400), in: The Journal of Roman Studies 70, 101-125. Hopkins, K., 2002: Rome, Taxes, Rents and Trade, in: W. Scheidel and S. v. Reden, The ancient economy. Edinburgh: 190-230. Howgego, C., 1992: The supply and use of money in the Roman world 200 BC to AD 300, in: The Journal of Roman Studies 82, 1-31. Howgego, C., 1994: "Coin circulation and the integration of the Roman economy." in: JRA 7: 5-21. Lo Cascio, E., 2008: The Function of Gold Coinage in the Monetary Economy of the Roman Empire, in: W. V. Harris The monetary systems of the Greeks and Romans. New York, Oxford University Press, USA: 160-173. Lo Cascio, E. (ed.), 2003: Credito e moneta nel mondo romano. Atti degli incontri capresi di storia dell economia antica (Capri 2000). Bari. 14

Nicolay, J., 2007: Armed Batavians: use and significance of weaponry and horse gear from non-military contexts in the Rhine Delta (50 BC to AD 450). Amsterdam, Amsterdam Archaeological Studies 11. Scheidel, W. and S. von Reden (eds.), 2002: The ancient economy. Edinburgh. Temin, P., 2001: "A Market Economy in the Early Roman Empire." in: Journal of Roman Studies 91: 169 (113). Temin, P., 2004: "Financial intermediation in the early Roman Empire." in: The Journal of Economic History 64(03): 705-733. Van Driel-Murray, C., 1994: Wapentuig voor Hercules, in: N. Roymans and T. Derks (eds), De tempel van Empel. Amsterdam, 92-107. Van Kerckhove, J., 2009: Aardewerk, in: J. Van Renswoude and J. Van Kerckhove (eds), Opgravingen in Geldermalsen-Hondsgemet:een inheemse nederzetting uit de late IJzertijd en Romeinse tijd. Amsterdam, 115-192. Van Renswoude, J. and J. Van Kerckhove, (eds.), 2009: Opgravingen in Geldermalsen- Hondsgemet: een inheemse nederzetting uit de Late IJzertijd en Romeinse tijd, Zuidnederlandse Archeologische Rapporten, Amsterdam. Verboven, K., 2009: Currency, bullion and accounts. Monetary modes in the Roman world, in: Belgisch Tijdschrift voor Numismatiek en Zegelkunde/Revue Belge de Numismatique et de Sigillographie 155, 91. Vos, W., 2009: Bataafs platteland: Het Romeinse nederzettingslandschap in het Nederlandse Kromme-Rijngebied. Amersfoort, Nederlandse Archeologische Rapporten 35. Vossen, I., 2003: The possibilities and limitations of demographic calculations in the Batavian area, in: T. Grünewald and S. Seibel (eds), Kontinuitat und Diskontinuitat. Germania inferior am Beginn und am Ende der römischen Herrschaft. Berlin, 414-435. 15