Journal Code Article ID Dispatch: 16.08.17 CE: O J O A 1 2 1 2 3 No. of Pages: 16 ME: 1 50 2 51 3 52 4 FLORIN CURTA Q153 5 54 6 55 7 COINS, FORTS AND COMMERCIAL EXCHANGES IN THE SIXTH- 56 8 AND EARLY SEVENTH-CENTURY BALKANS 57 9 58 10 59 11 60 12 61 13 Summary. The nature of settlements in the sixth-century Balkans is a matter of 62 14 current debate. Amphorae and hoards of iron implements and weapons have been 63 15 discussed in relation to this controversy. A key problem is that of the use of coins in 64 16 an economic environment without any large-scale agricultural production. While 65 17 hoards of coins have been analyzed in relation to the presence of the military in 66 18 the Balkans, single finds of coins remain a category of archaeological evidence 67 19 commonly neglected in discussions of the sixth-century economy. The article 68 20 offers an explanation connected with the quaestura exercitus implemented in 69 21 536, and its conclusion is that the small copper denominations discovered on 70 22 hilltop sites in the Balkans were not obtained on the market (none existed in 71 23 any of the many hilltop sites known so far), but piggybacked on transports of 72 24 annona. 73 25 74 26 75 27 76 28 77 INTRODUCTION 29 78 30 The nature or character of the settlement pattern in the sixth-century Balkans has been a 79 31 subject of debate among historians and archaeologists. The chief reason is that nearly every aspect 80 32 of the history of the last century of Roman rule in the Balkans has been revised in the past twenty 81 33 years. The problem of ‘ruralization’–the transformation of urban centres into something less than 82 34 cities, but more than just villages – is certainly the largest area of revision. Only slightly less 83 35 extensive revision has occurred in the scholarship of the non-urban settlements of the sixth century, 84 36 where archaeologists have sketched a picture of a great number of fortified sites. 85 37 Recently, the debate has focused on these sites: were they civilian (fortified villages) or 86 38 military (fortresses inhabited by soldiery)? According to Archibald Dunn, fortified hilltop sites in 87 39 northern Greece offered shelter to the urban and rural populations fleeing the lowlands under the 88 40 continuous threat of barbarian raids (Dunn 1997, 144; 2004, 551–2). Mihailo Milinković believes 89 41 that the very large number of fortified sites cannot be interpreted either as refuges or as purely 90 42 military sites. There is in fact evidence of women and children inside the forts, and the empire, 91 43 according to him, could not possibly have administered all those sites. They were therefore fortified 92 44 villages, and their rural character is betrayed by finds of agricultural implements (Milinković 2012; 93 45 2014, 251). Andrew Poulter, meanwhile, denies the existence of any identity or even similarity 94 46 between the hilltop sites in the northern Balkans, which he regards as temporary refuges, and those 95 47 96 OXFORD JOURNAL OF ARCHAEOLOGY(9999) 439–454 2017 48 © 2017 John Wiley & Sons Ltd 439 97 49 98 COINS, FORTS AND COMMERCIAL EXCHANGES 1 ‘regularly built fortifications on the frontier, which more obviously performed a military role’ 50 2 (Poulter 2004, 247; 2007, 380) Like Milinković, Chavdar Kirilov points to the archaeological 51 3 evidence of agricultural occupations as an argument in favour of the idea that hilltop sites were 52 4 fortified villages, not military forts (Kirilov 2007, 337–8). 53 5 54 6 55 7 56 FORTS AND AGRICULTURE 8 57 9 However, agricultural implements found on hilltop sites in the central and northern 58 10 Balkans have often been misdated, as many of these sites were reoccupied in the ninth and tenth 59 11 centuries. In several cases, the tools in question are to be associated with the early medieval, and 60 12 not with the late antique, phase of occupation. This is also true for implements found in hoard 61 13 assemblages. Out of the three hoards mentioned by Kirilov, Shumen is most likely to be of early 62 14 medieval, and not late antique, date, as indicated by the coulters and the spade frame found in that 63 15 assemblage (Curta 2013, 831 and 835; 827 fig. 6). On the other hand, the mattocks and pick-axes, 64 16 as well as the sickles and bill-knives found in abundance in late antique hoards, fit very well within 65 17 the picture of small-scale cultivation of crops either within or just outside the fort walls. Large 66 18 ‘open spaces’ existed, for example, on the northern side of the early Byzantine fort built in the 67 19 south-eastern corner of the ancient city of Nicopolis ad Istrum (Nikiup); there is no sign of 68 20 large-scale grain cultivation, and the open spaces may have been used for garden cultivation of 69 21 millet and legumes (Poulter 1995, 181). Analysis of palaeobotanical assemblages from Iatrus 70 22 (Krivina) has revealed that the diet of the soldiers in the fort garrison consisted of oats and peas, 71 23 both of which may have been cultivated on site (Hajnalová 1982). But there is also sufficient 72 24 evidence to suggest that such small-scale cultivation on plots inside or outside the city walls was 73 25 not sufficient for the subsistence of the relatively large number of people living inside sixth-century 74 26 hilltop sites. The distribution of sixth-century amphorae (particularly LR1, LR2 and spatheia)in 75 27 the Balkans has been interpreted as evidence of a state-run distribution of food supplies to the 76 28 garrisons stationed in forts (Curta 2001, 187). Palaeobotanical assemblages from the late sixth- 77 29 and early seventh-century military site at Svetinja comprised mixtures of wheat, rye, barley and 78 30 millet – an indication of supplies of corn coming from outside the military settlement, probably 79 31 from neighbouring Viminacium, to which they may have been shipped via the annona-like 80 32 distributions signalled by finds of Late Roman amphorae (Borojević 1987). The author of the 81 33 Strategikon – a late-sixth- or early-seventh-century military treatise – recommended that when 82 34 campaigning north of the Danube River, in Sclavene territory, Roman troops should not destroy 83 35 provisions found in the surrounding countryside, but instead ship them on pack animals and boats 84 36 ‘to our own country’ (Strategikon XI 4.8). 85 37 That Roman soldiers needed to rely on food supplies captured from the enemy suggests 86 38 that there was no large-scale production of food in or around the fortified sites in the Balkans. 87 39 Similarly, the analysis of faunal remains from Iatrus shows that the soldiers in the garrison relied 88 40 heavily on hunting for meat procurement (Bartosiewicz and Choyke 1991, 196). Even if, on the 89 41 basis of the evidence of hoards of iron implements, one has to admit that some inhabitants of the 90 42 fortified sites in the sixth-century Balkan provinces of the empire turned to small-scale cultivation 91 43 of crops in order to supplement insufficient or irregular annona distributions, no evidence exists 92 44 that such activities were anything more than temporary or economically marginal. Hilltop sites in 93 45 the Balkans may not all have been military, but none of them appears to have functioned as a 94 46 fortified village. Behind or just outside the walls of the sixth-century forts, no agricultural 95 47 96 48 OXFORD JOURNAL OF ARCHAEOLOGY 97 49 440 © 2017 John Wiley & Sons Ltd 98 FLORIN CURTA 1 occupations could be practised in such a way as to satisfy the needs of the existing population. 50 2 The ‘ruralization’ of the late antique Balkans must therefore be understood as a militarization 51 3 of the countryside. 52 4 The agrarian technology revealed by the analysis of hoards is one of limited resources, 53 5 which could in no way be linked to a self-sufficient rural economy. Since the size of the fields 54 6 depends upon the implements used to till them, one might ask how it was possible to feed the 55 7 population – military or otherwise – living within the ramparts of the numerous sixth-century 56 8 forts. It has long been suggested that the archaeological evidence, particularly that of amphorae 57 9 and lead seals, points to the great significance of the quaestura exercitus, an administrative unit 58 10 created in 536 by means of connecting rich provinces overseas (islands in the Aegean sea, Caria 59 11 and Cyprus) with border provinces such as Moesia Inferior and Scythia Minor, in order to secure 60 12 both militarily and financially the efficient defence of the Danube frontier (Curta 2002 and 2016). 61 13 The main responsibility of the quaestor exercitus was the annona for the army in Moesia Inferior 62 14 and Scythia Minor: the taxes collected in Caria, Cyprus and the Aegean islands were redirected 63 15 towards the troops stationed in the northern Balkans either in cash or, more likely, in kind. The 64 16 distribution of LR 2 amphorae in the Balkans overlaps with that of lead seals with inscriptions 65 17 referring to officials of the imperial administration, many of whom worked for the quaestura 66 18 exercitus. 67 19 68 20 69 21 70 COMMERCE, ANNONA AND THE NUMISMATIC EVIDENCE 22 71 23 But how did the distribution of the annona actually work in the sixth-century Balkans? 72 24 Catherine Abadie-Reynal has contrasted the ‘free-market commerce’ indicated by finds of 73 25 Palestinian amphorae (particularly the so-called ‘Gaza amphora’, LR4) with the annona 74 26 distributions to the army signalled by LR1, LR2 and spatheion-type amphorae, thus implying that 75 27 the latter were not ‘commerce’, but exchanges of a rather different nature (Abadie-Reynal 1989; 76 28 2010).