. chapter one . Coins and Hoards Kevin Butcher INTRODUCTION respectively. The latter two issues were probably produced at Antioch rather than Zeugma itself.5 Although intermit- The assemblage of 790 coins presented here is largely the tent, the size of some of the bronze issues may have been product, directly or indirectly, of a major trauma in the his- large, but Zeugma was not a particularly important civic tory of settlement at Zeugma: the mid-third-century sack mint and neighboring cities such as Hierapolis and Samo- by the Sasanian forces of Shapur I.1 Buildings were burned, sata issued significant quantities of coin more frequently. abandoned, and collapsed, and coins were deposited, either In fact, just before the Sasanian sack it seems that the coin- singly or in groups, in the process. The event and its after- age of neighboring Edessa predominated at Zeugma over math led to the deposition of a range of objects that might the city’s own issues. not otherwise have found their way into the archaeologi- cal record. This includes a number of hoards of silver and bronze coins. AppROACHING THE ZEUGMA FINDS The material from Zeugma invites comparison with the site of Dura-Europos further downstream on the Euphra- The numismatic material can be divided into two categories: tes, which was also sacked by the Sasanians in the same single finds and hoards. It is generally accepted that these period, and where excavations produced a very large sam- are different sorts of evidence. If someone were to analyze ple of coins, published by Alfred Bellinger.2 Indeed, there coin use at a site without distinguishing single finds from are notable similarities: a large number of coins of Edessa; a hoard coins, numismatists might rightly admonish him or significant number of coins from Pontus and the Pelopon- her for doing so. In general, hoards are seen as deliberately nese; and a higher proportion of silver coins than might chosen assemblages of more valuable coins, whereas single otherwise be expected on a Near Eastern site. But there are site finds are thought to represent a random selection of also notable differences. low-value pieces lost in the course of everyday use, with a The 288 single finds and four hoards discussed here bias in favor of those that were not worth spending much were recovered in the PHI excavations undertaken by time trying to recover. While assumptions about loss might Oxford Archaeology in the summer of 2000.3 It should be questioned, this opposition of high value to low value be understood that the term assemblage here refers to the certainly holds good for much of the numismatic evidence coins from these excavations, and that other excavations from antiquity (in those cases where a range of denomina- at Zeugma have produced coins and hoards that are not tions and metals existed), to the extent that little can be discussed here. Publication of the latter may modify some deduced about the circulation and use of high-value coins of the views expressed in this chapter. Most of what follows from site finds, and hoards provide only limited evidence concerns the coins of the first half of the third century, as for base metal. However, the real distinction seems to lie the material of other periods is too scarce to form any over- between those coins that were hoarded but rarely occur as all picture. single finds, and those coins that occur both in hoards and as single finds.6 To that extent, at least, the need to distin- guish hoards and single finds is justified. The Coinage of Zeugma Yet it is much harder to move from these general obser- Unlike Dura-Europos, Zeugma issued its own coins in the vations about the nature of the material to any conclusions Early and Middle Imperial periods (62 of which were found about its meaning. Different specialists use different meth- at Dura4). Most of these bore a design showing a temple ods and assumptions, and it is necessary to outline those on a hill (almost certainly Belkis Tepe), accompanied by methods and assumptions that I agree with, and those that the simple inscription ΖΕΥΓΜΑΤΕΩΝ (“of the citizens I disagree with, before proceeding with an analysis of the of Zeugma”). Production was intermittent: There were material. three issues of bronze under Antoninus Pius (A.D. 138–161), which probably included a small issue of “autonomous” Single finds coins (those lacking an imperial portrait or titles); another The simplest approach to single finds is to regard them as under Marcus Aurelius and Lucius Verus (A.D. 161–169); the product of casual loss, meaning that the deposition of an issue of silver tetradrachms in the sole reign of Cara- each coin was accidental, and occurred during the course calla (A.D. 212–217); and two further issues of bronze of its useful life as money. With this approach there is a under Elagabalus (A.D. 218–222) and Philip (A.D. 244–249), tendency to posit a direct correlation between the number . 1 . butcher . 2 of coins lost and the number in use. The number of coins the eastern Roman Empire issued their own coinages down recovered from a site can be divided up into issue periods, to the middle of the third century, and we know very little or imagined periods of use, to produce a picture of coin about the circulation and use of these coins. A number of circulation over time. Rises and falls in the number of coins these “foreign” civic mints in the assemblage from Zeug- recorded for each successive period are thought to indicate ma are represented by a single coin, which is hardly proof changes in the quantity of coins circulating. The number that their coins circulated there, still less of economic links of coins per period can then be divided by the number of between Zeugma and the cities concerned. years in a given period to produce an index of coin loss per As I have suggested elsewhere,10 cities may have regu- year, which is supposed to inform us about economic activ- lated the coinage circulating in their territories rather than ity at a given site. The number of losses is therefore taken to accepting each and every coin that was brought to the mar- correlate with the number of transactions conducted using ketplace, and the value (if any) that they assigned to the coins (where a transaction is assumed to be the arena in coins was not necessarily the value that they had elsewhere. which a coin is most likely to be accidentally dropped). A For example, the coinage of Samosata, a city upriver from rise in the number of coins being lost equals a higher num- Zeugma that issued large quantities of coins more regularly ber of presumed transactions taking place, which equates than Zeugma, is entirely absent from the assemblage.11 It with a thriving economy. would seem rash to conclude from this that there were few Such methods have misled historians, and sometimes or no economic links between Zeugma and its neighbor, spurious conclusions about ancient monetary economies and yet something was preventing coins of Samosata, circu- have resulted.7 There are numerous problems with the lating in quantity at Zeugma. It cannot have been distance, approach. First of all, we know comparatively little about for Edessa is further away than Samosata but Edessene how long various sorts of coin remained in circulation. The coins are common (and travel between Edessa and Zeugma periods of “use” are largely either speculative ones based on had to be undertaken by land, whereas travelers between observed changes to the designs of coins as produced, or Samosata and Zeugma had the option of the Euphrates). derived from changes in the patterns of hoarding, conve- An alternative that students of Roman coinage commonly niently ignoring the admonition not to confuse hoards with invoke whenever they encounter apparent boundaries to single finds. Where work has been done to devise periods circulation is that the coins are absent because they were of coin use by looking at single finds in their archaeologi- overvalued — in this case, that would mean that coins of cal contexts, the results suggest that it would be prudent Samosata were given a higher face value in proportion not to confuse the two categories of evidence in this way,8 to their bullion value than coins of Zeugma or those of although they may form a useful complement to each other any other cities where coins of Samosata were absent. The when studying circulation in general. As yet we do not have assumption is that metal content and weight were some- enough information about periods of coin use at Zeugma, how critical in determining value, though this seems to fly and I will not be attempting to create an index of coin loss in the face of another widely held assumption that bronze per year here. coins had a “token” face value that was divorced from their In any case, the evidence from Zeugma itself would bullion value. According to this view, the absence or rarity throw such a method into doubt. Here there would be a of Samosatene coin can be explained by rational economic very low figure for coins per year in Seleucid times, con- preference: People outside the territory of Samosata chose tinuing through the Commagenian phase and the first cen- not to use these coins, preferring other coins with a metal tury, with a slight rise in the second century. The majority content closer to their face value.