Landed Traders, Trading Agriculturalists? Land in the Economy of the Italian Diaspora in the Greek East

Abstract

This paper revises current understandings of the rôle of land in the economy of the Italian diaspora in the Greek East in the second and first centuries b.c., arguing that these Italians owned more land than has previously been assumed and that many of these Italian landowners practised a highly commercialized form of agriculture that focused on high-end products. This strategy shaped what empire meant both locally and in Italy and Rome, where the products they marketed fed into the ongoing consumer revolutions of the time. After discussing the evidence for the extent of Italian landholdings and examining their exploitation in three case studies, we conclude by reflecting on the long-term history of such landholdings in the provinces and the implications for our understanding of Roman imperialism more generally.

I INTRODUCTION

Scholars have been discussing the economic profile of the many Italians who went to live in the Roman provinces during the second and first centuries b.c. for about a hundred years. The details of these discussions vary but they show a marked tendency to downplay the rôle of landholding, preferring to cast these Italians as bankers and traders, at times connected with Roman military activity. While some fail to mention land altogether or emphasize the limited extent of Italian holdings, others see land as unconnected with and incidental to the Italians' otherwise commercial interests, acquired either for social prestige or as a result of debtors defaulting. These ideas about the economic profile of the Italian diaspora are crucial for how historians imagine that this diaspora shaped Roman imperialism and its local impact. ¹

As with many aspects of our understanding of Rome's imperial diaspora, this downplaying of land in the diaspora's economy has no linear historiography. Instead, three separate factors have contributed to it, which can all now be called into question. First, while the Latin negotiatores, a word often found in relation to members of the diaspora, has commonly been thought to connote bankers and traders, there is now an ever-growing set of arguments that in the late Republic the word had a very broad scope, including landowners. ² Second, in the Greek East the paradigmatic status of Delos has contributed to seeing diasporic Italians as bankers and traders. ³ Such a status, however, cannot be assumed, it must be proven; anything else would simply mean being seduced by the wealth of evidence from the site. What is more, Italians on Delos did demonstrably own land there. ⁴ Lastly, Moses Finley's insistence on the incompatibility of landowning on the one hand and commerce and profit-seeking on the other, has also played a part in sidelining the rôle of land in accounts of the economy of the Italian diaspora, a group of people thought to be mainly concerned with lucrum; ⁵ hence the suggestions that members of the diaspora purchased land to gain social prestige or that social élites would have turned over their land to tenants, receiving rents from them. ⁶ While many aspects of Finley's argument have come under attack, the static nature of the model he formulated has turned out to be its weakest aspect. ⁷ Today the greatly variable nature of the exploitation of natural resources in Greco-Roman antiquity is well-recognized. ⁸ Together these three arguments have clearly undone the assumptions on which the prevalent interpretation of the rôle of land in the economic profile of the Italians was based. Indeed, recently scholars have begun to imagine a possible ‘interdependence of their “landed” and “commercial” interests’. ⁹ The place of land in the economy of the diaspora thus now appears as a research problem waiting to be investigated, and this is the question that we propose to tackle in this article. ¹⁰

Focusing on the Greek East, where evidence is most plentiful, we argue that more Italians owned agricultural and natural resources in the provinces than has previously been recognized, that these Italians were particularly interested in producing high-end goods, and that they were involved in commercializing their products, exporting them to, among other places, markets in Rome and Italy. These Italians thus played a crucial rôle in defining what empire meant, not only in the communities in which they owned their agricultural and natural resources but also in the metropolitan centre, where their products contributed to a set of consumer revolutions. ¹¹ As such, this paper is written with the conviction that just as economic history cannot be understood in isolation from cultural, social and political history, it should not limit its investigation to the questions of modern economics either.

Our argumentation combines breadth and depth. At its heart lie an index of all known instances of Italians owning agricultural and natural resources in the Greek East during the second and first centuries b.c. and three case studies of Italian properties and their exploitation in Epirus, Cos and Chios, and Melos, which explore Italian strategies and behaviour in different ecological contexts. The case studies combine literary, epigraphical and archaeological evidence. While literary and epigraphical documents, which have informed interpretations of landownership among Italians in the provinces so far, show that Italians owned some natural resources in the provinces, they remain silent on what Italians did with them. This is where archaeology, in various guises, enters our argumentation, revealing itself yet again as an invaluable source of insight for thinking about economic life in the ancient world. ¹²

II THE EXTENT OF ITALIAN LANDOWNERSHIP IN THE GREEK EAST: SPATIAL, TEMPORAL AND SOCIETAL CONSIDERATIONS

The incontrovertible evidence for Italian landownership in the Greek East down to the end of the first century b.c. looks as follows. ¹³ We know twenty-two individual landowners in total, all of whom are attested in the first century b.c. (Table 1). Half of them are otherwise unknown, and the other half includes a freedman (Gaius Curtius Mithres), three equestrians (Titus Pomponius Atticus, Lucius Cossinius and Marcus Feridius), an exiled senator (Gaius Antonius), the son of an exiled senator (Appuleius Decianus), a praetor (Lucius Flavius), and two members of the imperial family (Agrippa and Livia). ¹⁴

Table 1 Index of individual Italian landowners in the Greek East in the Late Republic (alphabetical order)

We also have attestations of anonymous groups of Italian landowners in four different places and three more examples where groups of Italians in Greek cities are called ἐνκεκτημένοι or ἐνγαιοῦντες (Table 2). Unlike grants of ἔγκτησις, the right to own land in Greek cities, to individuals, which might just be part of a standard package of privileges that the city in question gave out, the choice to describe a local group of Italians as having this right indicates that at least some of the men so designated also had an interest in making use of it. ¹⁵

Table 2 Index of anonymous groups of Italian landowners in the Greek East in the Late Republic (chronological order)

Based on these two sets of evidence, we know of Italian landownership in twenty different places in the first century b.c.: in Epirus (Buthrotum) and on the Ionian islands (Kephallenia); in Macedon (Beroia) and the Peloponnese (Elis, Megalopolis and Messene); on both sides of the Hellespont (Parium and the Thracian Chersonese); on islands in the Aegean (Chios, Cos and Delos); in Aeolis (Cyme and Temnos), Lydia (Apollonis and Thyateira), Ionia (Colophon) and Caria (Alabanda); and in Pontus, Asia and Cilicia (precise locations unknown) (Fig. 1). ¹⁶ There is no reason to think that this evidence is representative of Italian landownership in the Greek East as regards time, place or social make-up. Instead, we want to see it as the arbitrary tip of an iceberg, the precise dimensions of which under the water are to a certain extent always going to remain unknown. Exploring what we can and cannot say about these dimensions is the subject of this section.

FIG. 1. Propertied Italians and grants of enktesis to Italians in the Greek East (third to first centuries b.c.).

All of our evidence for Italian landownership in the provinces is ‘unintentional’. ¹⁷ There is no ancient documentation that sets out to register and chronicle it. Instead, we learn about it through occasional mentions in a variety of sources. There are references in ancient literary accounts, such as when Appian relates the conflict between Mithridates and the Chians over how the landholdings of the Italians who had fled to the city should now be taxed. ¹⁸ Greek inscriptions also sometimes mention Italians as landowners. ¹⁹ The Messenian documents concerning the levy of a one-per-cent tax on the property of all citizens and residents, including Italians, are a case in point. ²⁰ However, about half of the known instances of Italian landownership in the Greek East stem from Cicero's writings between 68 and 46 b.c., above all from the letters of recommendation that he wrote for members of the Italian diaspora introducing them to the incoming governors of the province in which their estates were located. ²¹ In light of the prominence of Cicero's writings among the evidence for Italian landownership and the relatively short period of time from which they stem, any attempt to assess the history and chronology of Italian landownership in the Greek East will have to look beyond the sources we have gathered here. As regards the social composition and origin of Italian landowners, the possible bias that results from the dominance of Cicero's writings needs to be borne in mind but these writings also present otherwise unavailable avenues for assessing the problem. Before discussing these two issues, however, we turn to the geographical extent of Italian landownership in the Greek East, the problem to which the data we gathered speak best.

Given the chronological distribution of our evidence, the map we present (Fig. 1) can be taken as a rough snapshot of the situation in the middle of the first century b.c. Though it might seem obvious, it is still worth noting: given the unintentional nature of our evidence, the absence of a dot does not mean that Italians did not own land there. In fact, there is plenty of circumstantial archaeological evidence that makes Italian landownership in various other places rather likely. We discuss one such instance in our case study of Melos, which is notably unmarked on the map. Adam Lindhagen also makes a convincing case for Italian landownership in Narona, the much neglected Roman emporion on the Illyrian coast, ²² and Sophia Zoumbaki has inferred the existence of an Italian estate near Lake Trichonion in Aetolia based on the findspot of a late second-century b.c. Latin funerary inscription of a liberta — an approach that is often used to track Roman landownership in Anatolia during the Principate. ²³ In other words, there is good reason to think that there was much more Italian landownership than we have attested. But even in light of the evidence that we have, the conclusion that Italians owned land just about everywhere — where there was a Roman province, that is — seems hard to avoid. What is more, in several places Italians came to own significant parts of various local landscapes.

In the second book of his De Re Rustica, Varro set up his discussion of cattle raising as a dialogue among the Italian ‘cattle-breeding athletes of Epirus’, as he called them. As a result, the extent of Italian holdings and the considerable number of Italians involved in the agricultural economy on the eastern shore of the Adriatic could never be denied. ²⁴ The Epirote situation, however, has been treated as an exception, often explained with reference to the mass enslavement of Epirotes after the Roman victory over Perseus. ²⁵ But there is evidence that suggests similar situations in various other places in the first century b.c. One of them is Messene. For starters, we know the names of five Italian landowners there, as well as the number and the names of their properties. ²⁶ More importantly, however, a document related to tax-collection names a category of Romans that were evaluated as part of the Messenian territorial tribes, which has been taken as an indication of their landownership there. ²⁷ The wealth of these Romans, together with that of other foreigners resident in Messene, made up more than a tenth of the total evaluation of the city's population. ²⁸ It would appear, then, that Italian landownership in the territory of Messene in the first century b.c. was not a rare phenomenon. A Coan inscription by a disparate group of Coans, metics and Romans, who all owned and farmed land in Haleis and Peles, two Coan demes, shows an analogous Italian penetration of the local rural landscape. ²⁹ In all likelihood Italian possessions on Chios were also quite substantial; Mithridates insisted on receiving tax-revenue from them. ³⁰ Lastly, there are the inscriptions from Beroia and Elis that identify groups of Italians as ἐνκεκτημένοι and ἐνγαιοῦντες respectively, suggesting that the rural population of Elis and Beroia might have been penetrated by Italian landowners in precisely the way in which we know it happened in Epirus, Messene, Cos and Chios. ³¹ All this evidence combined makes it impossible to continue casting Epirus as an exception, which gives a new dimension to the connection that scholars have made between the enslavement of large parts of the Epirote population and massed Italian landownership in the region. ³² It serves as an urgent reminder of the massive displacement of previous, local owners that these Italians and their landed estates caused. ³³

Landownership in the diaspora was not the monopoly of any one particular socio-political group. As mentioned above, we find freedmen, equestrians, senators, both exiled and not, and members of the imperial family as landowners. Moreover, many Italian landowners in the Greek East probably did not belong to any of these groups. While we might suspect that at least some of the landowners about whom we have no further information fell into this category, Lucius Genucilius Curvus, whom Cicero introduced to Quintus Minucius Thermus in 51/50 b.c., seems a quite certain case; unlike, for example, Marcus Feridius, whom Caelius introduced to Cicero in the same year, Curvus is not identified as an eques. ³⁴ The failure to mention his social position within the Roman polity would indeed be a grave mistake by Cicero in a letter designed to impress the urgency of Curvus' business on the governor of the province where his estates were located. It seems likely, then, that among the Italians whose social position we know, our evidence is skewed towards the top, simply because they and their landownership appear much more readily in our sources, especially in ancient historical narratives.

Two more aspects regarding the social composition of Italian landholders deserve highlighting. Based on a passage in Cicero's Verrines Elizabeth Rawson has posited a law that forbade senators from owning land outside of Italy, an argument that Jonathan Prag revived recently by positing a possible context for this law in the late third century b.c. together with the plebiscitum Claudianum. ³⁵ If Prag is right, the passing of such a law at the time when the earliest provinces were taking shape only helps our case here since the prohibition that it enshrined indirectly reveals that new landholdings were precisely one of the things that senators — and other Romans — were hoping for in these new provinces. At the same time, the unapologetic way in which Cicero wrote to his brother about Lucius Flavius' inheritance of an estate in Apollonis in Lydia in 57 b.c. only confirms what Cicero already implied in the Verrines in 70 b.c. — that this law was no longer consistently applied. ³⁶ The early first century b.c. thus appears a safe terminus ante quem for its desuetude. Lastly, it is worth remembering that in the first century b.c., and possibly already earlier, land located outside of Italy did not count in the census in Rome. ³⁷ As a result, the Roman socio-political status of Italian landowners in the Greek East cannot serve as an indication of the size of their estates and overall wealth. Atticus' case is illustrative here. His Epirote properties together with his urban properties in Rome made up the main part of his estate. ³⁸ In other words, possibly half of Atticus' wealth in land did not count for his census valuation in Rome.

Based on the evidence that forms the starting point for this section, Italian landownership in the Greek East would appear to be a predominantly first-century b.c. phenomenon and in most cases would also post-date Mithridates' revolt. As argued above, the dominant position of Cicero's writings as a source casts doubt upon this picture. More importantly, there is evidence that makes Italian landownership in second-century mainland Greece and in pre-Mithridates Asia Minor rather likely. Italians were present in many cities in Asia Minor before 88 b.c.; Mithridates famously had them killed. ³⁹ They also owned land in at least one of these cities, in Chios. ⁴⁰ It seems unlikely that the island was an exception; Mithridates, at least, anticipated taking over all of their possessions. ⁴¹ Also, a dispute between Colophon and its resident Italians in the 110s b.c. appears to have concerned the landholdings that these men had acquired in the city's territory. ⁴² As regards mainland and insular Greece, Italians were present in cities in the third and second centuries b.c., and some of them were granted the right of ἔγκτησις (Fig. 1). ⁴³ Sophia Zoumbaki has convincingly argued that at least three awards of enktesis to Italians in central Greece — by the Acharnanian League, Amphissa and Delphi — were not simply part of a standard set of honours that foreigners could receive in these polities; instead, they are exceptional enough in their context or formulation to suggest that the awardees were actively interested in this particular right and in making use of it. ⁴⁴ The late second-century b.c. Latin funerary inscription for a liberta from Lake Trichonion should also be added here, in addition to the four Italians who rented properties on Mykonos and Rheneia in the 150s b.c. In this context it is also worth mentioning Pandusinus' involvement in the agricultural economy in Thisbe in 170 b.c., which most likely accompanied Roman military campaigns in the region. ⁴⁵ However, it remains hard to say anything more specific than that we need to look to the late third and second century b.c. — to the period that saw increasing Roman military and administrative involvement in the region — for the beginnings of Italian landownership in the Greek East.

While Italian landownership increased over time, it was also disrupted, intensified and manipulated in the context of the ‘big events’ of the period. As cities in Asia Minor joined Mithridates' cause, they expropriated the landholdings of the Italian populations living in their midst. After Mithridates' defeat Italians returned. The Chians, for example, made sure to get the Senate's guarantee for the conditions under which Italians were to live in Chios henceforth. ⁴⁶ The Sullan settlement also meant a windfall of land for Italian creditors, as cash-strapped cities offered them their public property, including land, as collateral, and more generally, we should imagine that after Mithridates' defeat barely any city dared resist the requests and desires of the diaspora. ⁴⁷ Later in the first century b.c. Pharnaces' invasion of Pontus and Bithynia caused a similar dynamic concerning Italian landownership in these parts of the empire. ⁴⁸ However, the colonial foundations of Caesar and Augustus in the second half of the first century spelt the most fundamental and lasting alterations in the existing patterns of Italian landownership in the Greek East. ⁴⁹ Pompey had already settled some of his veterans in Bithynia, Cilicia, Crete and Macedon, but the settlements of Caesar and Augustus were on an unprecedented scale. ⁵⁰ These two men increased the number of Italian landowners in the provinces, but their foundations also threatened many existing ones. Atticus' frantic correspondence with Cicero about Caesar's testament, Mark Antony's execution thereof, and the plan to place a colony at Buthrotum, where Atticus' own estates were located, provides a good illustration of the disruption in the patterns of Italian landownership in the Greek East that these colonies could mean. ⁵¹

The Caesarian and Augustan colonies have rarely been treated in connection with the Italian diaspora, let alone with the diaspora's landholdings, a scholarly division that we have followed as well in this section so far. ⁵² The overwhelming geographical extent of Italian landownership as well as the density of the Italian penetration of several rural landscapes of the Greek East that we have uncovered in this section makes this relationship an interesting research problem. Again, Adam Lindhagen's research on Narona, in particular on the amphora production on Vis, a small island off the Croatian coast where the city was located, provides some intriguing insights. In the late first century b.c. Narona became a colony. While this refoundation of the town did not interrupt the amphora production in Vis, it did change the names of the people whose stamps could be found on the vessels: members of senatorial families seemingly took over the organization of the production from freedmen and other Italian families. ⁵³ In other words, as the socio-political profile of the community changed, its economic mainstay, the production and export of wine, did not. Strabo also tells us that the Roman colonists at Patras made sure to own the rights to fish a lake near Aitolian Calydon that was renowned for its oysters, which were also consumed in Rome. ⁵⁴ In so doing, these colonists followed precisely the same economic strategies that many Italian landowners in the Greek East had done before them. This, at least, is what the following three case studies suggest.

III THE EXPLOITATION OF ITALIAN ESTATES IN THE GREEK EAST: THREE CASE STUDIES

In this section we present three case studies that examine the exploitation of Italian estates in Epirus, Cos and Chios, and Melos, respectively. We chose these regions with a view to examining Italian behaviour in different ecological contexts and in relation to different types of resources and products. This variation helps us to discern a pattern and strengthens our argument that such a pattern existed. At the same time, the radically different types of evidence that form the basis for each study bring to the fore more clearly aspects of the exploitation of Italian estates of which other case studies might only reveal glimpses. As such, the different cases also build on each other argumentatively. Overall, each case study constitutes a chapter in the history and geography of the region in question, an examination of how the Italians shaped the environment that they encountered and what this meant for local social, political and economic life. Put in the language of the Corrupting Sea, each case study explores how these regions became ‘the outposts of the demographic and economic dynamics of a wider world’. ⁵⁵ In our case this world and its demographic and economic dynamics were those that Roman imperialism created.

Epirus

As Nikola Čašule has argued, the Adriatic was an interconnected entity. The networks that shaped these connections — networks that the Romans most likely joined when they founded colonies on Italy's Adriatic coast in the third century b.c. — were constantly bringing Italians to Illyrian and Epirote cities on the other side of the sea. These were the men whose alleged mistreatment in the third and second centuries b.c. repeatedly involved Rome in military conflicts with the dynasts that ruled over the eastern shores of the Adriatic. ⁵⁶ As a result, Samnite and Messapian names at third-century b.c. Dyrrachium should not come as a surprise, and neither should the funerary inscription of an Οὐεργιλία Λευκίου — Vergilia, daughter of Lucius — in second-century b.c. Buthrotum or the dedications to Pan and one of his nymphs by a certain Κασιανός, probably a Greek transcription of the Latin name Cassianus. ⁵⁷ Just like the Italians that manumitted their slaves under the auspices of Asclepius in second-century Buthrotum, Vergilia and Cassianus participated in local social and commemorative practices that were common in the city. ⁵⁸ Beginning in the second half of the second century b.c., however, archaeology hints at the possibility of a very different type of Italian presence in the Epirote landscape.

In recent years archaeologists working on the Thesprotia survey in the Kokytos valley to the north-east of the Ambracian gulf have discovered a villa complex on one of the spurs of the Paramythia mountain range that border the valley to the west. ⁵⁹ Located at Agios Donatos, this complex covers an area of about 90 by 40 m and stretches over three terraces. On the site the excavators found a tile with a Latin stamp — COS — and one of the rooms had well-preserved wall-paintings in the Second Pompeian style. Most importantly, however, the earliest phase of the complex, which coins and pottery date to the second century b.c., was already built in opus incertum. This site now constitutes the earliest known villa complex in Epirus, and the evidence found there is among the best that archaeology can provide for identifying an Italian owner. ⁶⁰ This owner must have had a very different relationship to local economic, social and political life from that of the Italians attested in the epigraphic material from Buthrotum. Among other things, his diet demonstrably differed from that of many members of the local population: animal bone assemblages from the site contain the bones of wild animals such as deer, luxury food items that other contemporary bone assemblages from the region lacked. ⁶¹ Most tellingly though, the entire villa complex was built on the site of a Hellenistic fortress. For now it remains impossible to know the relationship between Aemilius Paulus' slave-hunting campaigns in 167 b.c., the destruction of the Hellenistic fortress, and the Italian man who acquired the estates that we should assume accompanied the villa complex. Crucially though, the sequence of buildings at Agios Donatos illustrates that the Italians who began acquiring land in Epirus in the second century b.c. did not simply move into empty lots in the landscape that the slave-hunting of Aemilius left; their presence also altered social, economic and political life in the region. ⁶²

Italian landownership in Epirus in the first century b.c. has never been in doubt. In the second half of that century Varro could make Italian landowners in Epirus five of the six interlocutors in the second book of his De Re Rustica: T. Pomponius Atticus, L. Cossinius, Qu. Lucienus, Murrius and Vaccius. ⁶³ And these were clearly not imaginary characters. Atticus was of course Cicero's friend, and L. Cossinius probably was the man of the same name who appeared in their correspondence. ⁶⁴ According to members of the Thesprotia survey, the COS on the tile they found abbreviated his name, making him the owner of the villa complex at Agios Donatos. ⁶⁵ Atticus, we know, purchased his estate in Epirus, his emptio Epirotica, in 69 or 68 b.c., and he had a domus in Buthrotum and properties along the Kalamas river further south. ⁶⁶ The villa complex at Agios Donatos, which is 18 km from the sea, now conclusively shows that the Italian presence in Epirus was not limited to the coastal plains. One of Cicero's letters of recommendation to L. Culleolus, governor of Macedon, further confirms this picture: he asked Culleolus to concern himself with the dispute that a certain L. Lucceius had with the town of Byllis, located on a hill overlooking the Aous river as it entered the coastal plain around Apollonia. ⁶⁷ The town was the centre of the koinon of the Bylliones, which controlled large upland territories in southern Illyria. ⁶⁸ While we do not know whether the dispute involved land, Italian interest in land that lay not only beyond coastal, but also beyond river plains seems highly likely in light of the type of agriculture in which Italians in Epirus were notoriously involved: transhumant pastoralism. ⁶⁹

Already in the Classical and Hellenistic periods Epirus and its hills were famous for pasture land and the cattle raised there. Pindar knew the Epirote mountains as ‘pasture loved by cows’ and Aristotle attributed the Panhellenic fame of Epirote cattle, which was due to their extraordinary size and the large amount of milk that they gave, to the fact that the region had an appropriate grazing ground for every season. ⁷⁰ The fact that up until the early nineteenth century the coastal plains in Epirus, in particular around the mouths of rivers, were substantially smaller than today will have made the highlands all the more important in the economy of the region. ⁷¹ No archaeological evidence exists that would allow us to quantify the rôle of pastoralism in the Epirote economy, but anecdotes about the Epirote kings speak to its importance. Pyrrhus and his family, Aristotle and Pliny tell us, took pride in having their own flocks of cows and sheep to whose well-being they tended and whose extraordinary qualities — qualities that they in fact bred for — might also have been meant to reflect back onto the king and his family. ⁷² In this titbit of royal ideology the king and his family emerge as the most successful pastoralists in Epirus — a reflection of the specific economy of the region in which these Hellenistic kings originated. Thus when Italians in Epirus became involved in cattle raising, they tapped into the distinct ecological potential of the region. The second book of Varro's De Re Rustica, in which Italian landowners in Epirus expound upon the principles of cattle raising, provides a good indicator for how these men might have approached their holdings. Two separate but related aspects deserve highlighting: the interest of Varro's characters in the entire production process, including the buying and selling of the animals, and the importance of the animals' quality, which pervades the entire book.

Varro's characters did not let out their estates to tenants but retained control of the exploitation of their agricultural resources through the unfree labour of a magister pecoris and other herders under his command. ⁷³ This exploitation also involved selling the cattle. The advice of Varro's characters for how to deal with each type of cattle always included the legal formula according to which it should be sold. ⁷⁴ What is more, where and why their animals were sold was a source of pride for these men. One of them bragged that he had sold some of his asses in Arcadia, a place that was renowned for its mules. ⁷⁵ More generally, Varro probably chose the Italians in Epirus — the ‘cattle-breeding athletes of Epirus’ as he calls them — as the characters in his book on pastoralism because of their animals’ renown. Together these interests and attitudes point to a highly commercialized form of pastoralism focused on high-end animals.

And indeed, different animals from Epirus were luxury goods in the late Republic, albeit on different levels. Most prominent are Epirote horses, which Romans in the late Republic and early Principate liked to use for racing. ⁷⁶ Just like asses, they could fetch high prices and most likely were the object of focused breeding efforts. ⁷⁷ Much more affordable — a cheaper luxury, one might say — were Epirote cows, known for their size, which according to Varro were used in just about any context in Italy except for sacrifice. ⁷⁸ This widespread use and consumption of Epirote animals in Italy should be taken seriously. In the case of Epirote racehorses, for example, it seems possible that Italians in Epirus bred their horses purposefully with consumers in Rome and Italy in mind, among whom horse racing was becoming extremely popular. ⁷⁹ As regards cattle, one might wonder whether the Italians involved in cattle raising in Epirus reorganized the existing networks for ferrying these animals across the Adriatic to Italy. ⁸⁰ Overall, animals from Epirus appear to have fed directly into the consumer revolution that was taking place in Italy in the second and first centuries b.c. ⁸¹

The extent to which the agricultural strategies of Italian landowners in Epirus differed from those of their Epirote counterparts remains impossible to know but in the second and first centuries b.c. the Epirote socio-economic landscape underwent significant changes that might have been the result of the Italian involvement there. While at least two enclosures that were used as part of transhumant regimes in northern Epirus show continued use across Hellenistic and Roman times, settlement patterns within the region do not reveal such continuity. ⁸² Instead, over the course of the first century b.c. settlements located in the uplands regions were increasingly abandoned. ⁸³ Many factors can explain this combination of change and continuity, but the dynamics of transhumant pastoralism suggest a very specific one. As has been shown in the case of the pastoral regime in Thessaly and between the Zagros mountains and the Mesopotamian floodplains, people involved in transhumant pastoralism can be based in the mountains or in the plains. ⁸⁴ In other words, herds might migrate from their base in low-lying valleys into summer pastures in the mountains or the other way around. Changes in the balance between these two regimes were highly contested, and in the Mesopotamian context at least were also often brought about through military conflict. Heracleides Lembos, writing in the second century b.c., preserves a glimpse of such upland pastoralists in Epirus. ⁸⁵ The abandonment of upland sites in Epirus thus very likely reflects a change in the fraught balance between highlands and lowlands, with pastoralists based on the coastal plains extending their grazing rights at the expense of those based in the mountains. The chronological correlation of this change with Italian landowners in Epirus and the heavily commercial type of pastoralism they practised is tempting and raises the distinct possibility that the Italians played a part in displacing the people in these highland settlements. ⁸⁶

In sum, starting in the second century b.c. Italians were involved in the agricultural economy of Epirus, acquiring large estates and showing a distinct interest in the pastoralist potential of the area. They were particularly keen on high-quality animals, on their breeding and subsequent marketing, to Italy in particular. As such, their agricultural strategies were highly commercial in nature and most likely had deep-running consequences for the forms of social organization prevalent in the region. But Epirus, one might argue, was exceptional — not only because of Aemilius Paulus' slave-hunting campaigns, but also due to its various connections with the Italian peninsula that long predated Rome's rise to prominence. In the next case study we focus on the islands of Cos and Chios, and the mainland opposite them to argue that what for Italians in Epirus was cattle, for those in the western Aegean was wine. Empire, it would appear, brought certain parts of the world closer together.

Cos, Chios and Beyond

By the first century b.c. Italians owned land in both Chios and Cos. As regards Chios, Appian tells us that the city and Mithridates came into conflict in the aftermath of the murder and exile of the Italians on the island because the citizens of Chios cultivated the properties of the Italians who had left, but refused to pay taxes on these properties to Mithridates. ⁸⁷ This episode not only testifies to the existence of Italian landholding on the island before 88 b.c., but Mithridates' eagerness to collect taxes on these landholdings also suggests that the revenue he derived from them was substantial, which could be taken as an indication of their extent. A few years after Mithridates' defeat Italians appear to have been back on Chios in sufficient numbers to prompt the definition of their legal position in a senatus consultum in 80 b.c. ⁸⁸ These Italians also left their footprint in the island's epigraphic material from the first century b.c. ⁸⁹ On Cos, the first attestation for landownership stems from the Augustan period when a group of people calling themselves ‘those residing in the deme of the Halentians and those having the right to own land and those farming in the demes of Haleis and Peles, of the citizens, the Romans, and the metics' honour a public physician. ⁹⁰ While this is the first attestation for Italians’ involvement in agricultural production on Cos, they were clearly already present before that date. ⁹¹ Stamps on Coan and Chian amphorae make it likely that at least some of them were interested in the viticultural fame and potential of these islands.

It has been known for a long time that Coan amphora handles, easily recognizable by their double-barrelled handles, also bore Latin stamps. ⁹² Susan Sherwin-White first mentioned this intriguing fact in print and attributed her information to the late Virginia Grace, who had spent her entire life working on these stamps without ever publishing the long-awaited corpus. Our argument here relies for the most part on the stamps we were able to consult in Grace's personal papers that are currently housed in the archives of the American School for Classical Studies in Athens. We begin by outlining our findings there. In 1958 Grace compiled a list of names written in Greek on Coan amphora handles, which she sent to Peter Fraser for the Lexicon of Greek Personal Names. In 1984 she drew up an additional list to be added to the previous one. These two lists contain five Italian names written in Greek. ⁹³ More significantly, Grace also worked up a card catalogue for the names found on Coan amphorae, often also including a picture of the stamp itself and annotations about where she found and saw them. The catalogue again only concerns the names written in Greek, but at the back of it, behind the cards with stamps that she had not read or identified yet, are stored six index cards and many more photos of Coan amphora handles bearing stamps of Latin names. ⁹⁴ They make up a total of twenty-eight distinct handles showing the stamps of twenty-one distinct individuals. ⁹⁵ As indicated by the annotations on the back of the pictures, twenty-one of the twenty-eight handles come from the Benaki Collection in Alexandria, which Grace helped sort and catalogue in 1955. ⁹⁶ Another two she saw in Athens, and three more in Antioch, Cos, and in the collections of the British Museum respectively. The origin or location of the remaining two is unknown. This relatively small number of stamps says nothing about the volume of production in which these Italians were involved since Coan amphorae appear to have been stamped very rarely. ⁹⁷ By contrast, the names of the individuals on these stamps reveal a lot about the place and organization of land in the economy of Italians in the diaspora.

Both common interpretations of stamps on Italian amphorae and recent models proposed for Greek amphorae suggest that the men whose names these stamps record were involved in agricultural production on the island. ⁹⁸ Most likely they had these amphorae produced for the storage and transport of the wine from their estates. Based on this assumption, it seems worth noting that at least four of the individuals named on Coan handles — Publius Arpinius, Postumius Curtius, Lucius Eumachius and Publius Sulla — have also been recorded on Italian amphorae. ⁹⁹ Intriguingly Murrius, the character in Varro's dialogue discussing asses and horses, states that he came from Reate, the best region for these animals in Italy. ¹⁰⁰ One might wonder whether he had also already bred animals there. At least some of the Italians in the diaspora, then, extended the type of agriculture they practised in Italy to the provinces.

Furthermore, some Italians were demonstrably involved in viticultural production in more than one city. Postumus Curtius, whose name can be found on Coan amphorae handles, provides a good example. Probably he became Rabirius Postumus by adoption, Cicero's client ‘cuius res in pluribus provinciis versata est’. ¹⁰¹ His freedman Gaius Curtius Mithres had a house in Ephesus while also being involved in a dispute over land in Colophon. At the same time, he was also honoured by the city of Naxos, yet another place famous for its wine. ¹⁰² Curtius Mithres, then, was a man with properties in different cities and we might wonder whether he also looked after the interests of his patron there. Notably, both Varro and Cicero repeatedly see Atticus and the other Italians on the eastern coast of the Adriatic as being simply in Epirus rather than in distinct cities there. ¹⁰³ They most likely saw an analogous region in Western Asia Minor, in which Italians were involved in agricultural production across civic boundaries, exploiting, just as in Epirus, a particular ecological potential of this part of the world: viticulture. ¹⁰⁴

Vedius Pollio, probably the most well-known member of the Italian diaspora, illustrates both these aspects very well: he most likely owned properties in both Chios and Cos — his name, at least, can be found on both Coan and Chian amphorae — and his father, Horace tells us, worked 1,000 iugera of Falernian land. ¹⁰⁵ While Pollio's rise to political prominence under Augustus might have been exceptional — he acted as a representative of the first emperor in Asia — his economic strategies in the provinces, though in part certainly located at the very high end of the luxuries available from there, would thus appear to have been less so. ¹⁰⁶ As a result, one might speculate that his fiscal privileges recorded in the Lex Portorii Asiae, while probably being a result of his political position, also reveal economic behaviour that was representative of his fellow viticulturalists in Western Asia Minor. ¹⁰⁷ From this document it would appear that Pollio enjoyed a complete tax immunity on exports from and imports to the province of Asia, a situation that was slightly modified in 17 b.c., when only shipments of his that did not exceed the value of 10,000 denarii were to enjoy such immunity. ¹⁰⁸ In order for this limit to be a meaningful restriction on Vedius Pollio's privileges, he must also have been involved in transporting goods on a large scale, possibly including the products of his own estates. ¹⁰⁹ So what in the case of the Italians raising cattle in Epirus we could already see glimpses of, also seems probable in the case of Vedius Pollio and his fellow landholders on Cos and Chios; they not only owned agricultural resources in the provinces, but were also involved in moving and commercializing their products.

The onomastic material from Coan stamps also reveals aspects of the social composition of the Italians involved in the production of wine on the island. With Publius Sulla, whose stamp also shows a caduceus and who was possibly the nephew of Lucius Cornelius Sulla, they included at least one member of a senatorial family, who himself was at some point a member of that body. ¹¹⁰ Analogously to Atticus in Epirus, we also find at least two equites: Rabirius Postumus and Vedius Pollio. ¹¹¹ Plotius Tucca, whose stamp was recently identified on a Coan amphora at Carthage and who most likely is the man known as the executor of Virgil's literary estate, should probably also be counted among them. ¹¹² Coan stamps also reveal members of different Italian gentes: the Albii and the Antonii. Based on the information provided on the stamps, it is impossible to identify these men any further, but they left their mark on the social make-up on the island of Cos; men and women carrying these names can be found in the epigraphic material on Cos from the first century a.d. ¹¹³ It is impossible to say whether these were their freedmen or direct descendants, but we should allow for the possibility that at least some of the Italians involved in viticultural production on the island also settled there. ¹¹⁴

While the trade in Coan wine has been chronically under-estimated, recent research locates Coan amphorae on sites all over the Mediterranean and the Black Sea, including Italy. ¹¹⁵ At second-century b.c. sites in the Levant such as Marissa, Coan amphorae are second in quantity only to those of Rhodes and far ahead of amphorae from other Greek cities. ¹¹⁶ A preliminary and certainly still insecure estimation of volume development on sites in the Levant, Cyprus and Egypt has revealed a sharp increase starting c. 150 b.c. ¹¹⁷ By the late second century b.c. slaves working in vineyards and the production and sale of particular wines on Cos were regular and reliable tax bases. ¹¹⁸ Thus when Italians arrived on Cos, the population of the island was already heavily involved in the production and export of wine on a large scale. As the findspots of the Coan handles bearing Latin names show, the wine that these Italians produced was distributed along the same routes as other Coan amphorae in the eastern Mediterranean. We can then only speculate how these Italian viticulturists shaped or changed the production and movement of wine on Cos, and note that their presence appears to have coincided with the peak of Coan production and, interestingly enough, with the decline in the volume of the Rhodian export of wine, which, in late second-century Labraunda at least, was demonstrably replaced by Coan wine. ¹¹⁹

The history of the perception of Coan wine in Italy points to an additional way in which these men might have transformed the production on the island. In the middle of the second century b.c. Cato knew Coan wine as a good variant of Greek wine — wine mixed with sea water, that is — but for him this was simply the type of wine that one would make to give to the familia working one's estate. ¹²⁰ This type of wine was also appreciated for medicinal reasons. ¹²¹ Imitating it thus simply was a matter of good and responsible household management. However, by the late first century b.c. this type of wine was no longer simply known as Coan wine, it was a particular type of it: white Coan wine. ¹²² At the same time, people in late Republican Italy also knew different types of Coan wine. Festus, summarizing the work of first-century b.c. grammarian Verrius Flaccus, explains the meaning of vinum hippocoum: it was a wine from a particularly excellent field on Cos called Hippo. ¹²³ The existence of such local knowledge of the viticultural geography of Cos reveals that in Italy Coan wine had indeed become a high-end product, known for its high quality and drunk in great quantities. ¹²⁴ This transformation of the status of Coan wine in Italy is remarkable, and just as in the case of cattle and horses from Epirus, we can speculate whether the involvement of Italians in the production of wine on this island had anything to do with it. By contrast, in the next case study — the Italian exploitation and processing of alum on Melos — the way in which Italians with properties in the provinces shaped local production as well as consumption in Rome and Italy is no longer a matter of speculation. The Italians there arguably put Melian alum on the map; they created it as a product in the western Mediterranean.

Melos

Like many cities in the Greek East, Melos witnessed the increasing arrival of Italians over the course of the second century b.c. ¹²⁵ Grave inscriptions from this period bear Roman praenomina and gentilicia (Μάαρκος, Τιβέριος, Κοίλιος, Νευμήνιος, Πλωτία), and by the late first century b.c. Italians, possibly of freedman origin, were setting up altars and stoas on the island. ¹²⁶ In the earliest Latin inscription on the island, which dates to the first century b.c., Gaius Caelius Eros, libertus of Gaius, is identified as a mercator, a merchant. ¹²⁷ Tantalizing though it may be, this inscription is the only glimpse of the Italians' involvement in the economic life of Melos that the epigraphic evidence provides. Without the unusually intensive and thorough archaeological exploration that the island's landscape has witnessed over the past forty years, the striking transformations in the island's economy that the Italians wrought would have escaped historians completely. ¹²⁸

Melos is rich in mineral resources that are exploited to this day. ¹²⁹ In ancient times the island was known for its pumice and melinum, Melian earth, but above all for its alum and sulphur deposits. ¹³⁰ The astringent qualities of alum and sulphur made them essential ingredients in the production of various goods, including metals, leather, textiles, cosmetics, and pharmaceutical remedies. As Diodorus wrote in relation to similar deposits on Lipari, these minerals were ‘of great usefulness’. ¹³¹ So useful, necessary and rare were these minerals that the inhabitants of Lipari could apparently place an exorbitant tax on their export.

During the Classical and Hellenistic periods the inhabitants of Melos exploited the island's mineral resources, alum in particular. Hippocratic authors writing in the fifth and fourth centuries b.c. knew of the island's alum and advised its use in the treatment of wounds and infertility, and so did Bolus of Mendes, who used alum from Melos in his experiments in second-century b.c. Alexandria. ¹³² While this exploitation left some traces in the archaeological record, the Roman period witnessed a dramatic increase in the number of sites related to the extraction and processing of mineral resources — especially of alum — on the island with close to twenty such sites identified so far. ¹³³ What is more, the connections between and interdependence of these sites makes it possible to assess the organization of the exploitation of Melian alum in the Roman period.

Sites related to the exploitation of Melian alum in the Roman period fall into three interdependent categories: mineral extraction sites, ceramic production sites, and processing and export sites. ¹³⁴ The raw materials from the first type of sites and the pottery produced in the second were then used in the sites of the third type to process the alum that would eventually be exported. ¹³⁵ At least five geographical clusters have been identified on the island, each of them including one of all three site-categories. Furthermore, each of these clusters was located in the vicinity of a villa-style building. ¹³⁶ Importantly, the connection between these sites does not rely on geographical proximity alone. In at least two cases, pottery from different sites in a cluster shares stamps that are found nowhere else on the island. ¹³⁷ The existence of these clusters strongly suggests the integration of all aspects of production — from extraction through processing and export — under the control of one individual. The inclusion in these clusters of the villa-style buildings — some of them decorated with Roman cemented pebble mosaics and polychrome marble revetment (at Soleta, Provatas and Agathia) — indicates that some of these individuals were Italians or at least claimed to have a particular relationship with that part of the world. ¹³⁸ The amphora stamps on Milo 1a amphorae, in which Melian alum was exported, lend further support to this idea. Most of them clearly abbreviated an Italian name. They include COR, CO[- -], [-]OR, PRISCIL, EROT(IS), COLONUS, and TI. CLAVDI SOSISTRATI. ¹³⁹ Significantly, these stamps were not only found on amphorae. Stamps reading COLO[- -], COR and CO[- -] were also found on lekanai at ceramic production sites together with amphorae bearing the same stamp. ¹⁴⁰

The following picture emerges. In the Roman period several members of the Italian diaspora, who — we should assume — owned the natural resources of Melos that they exploited, oversaw separate, but analogous integrated systems of alum production. The fact that at least three of the families were important members of the Melian élite throughout the first, second and third centuries a.d. supports the idea that they owned these resources. ¹⁴¹ While it is hard to date the emergence of the sites that constituted these systems with any precision, Milo 1a amphorae, the type of amphorae most commonly associated with the export of minerals from Melos in the Roman period and that were demonstrably produced at the ceramic production sites on the island, have been found in deposits from the last decades of the first century b.c. in Arles, Milan and Padua. ¹⁴² In Arles the deposit was located next to a tannery that was in operation in the middle of the first century b.c. It would appear, then, that the second half of the first century b.c. presents a likely terminus ante quem for the establishment of at least some of the systems of exploitation that the clusters of sites constituted.

These systems of exploitation that the members of the Italian diaspora introduced constituted an intensification of the production of alum on the island. The dramatic increase of Roman-period sites related to this production, as well as the fact that very little Hellenistic pottery has been found at these sites, suggests as much. ¹⁴³ The findspots of Melian amphorae outside of Melos also allow us to speculate that the increased production on the island mainly fed into the production of textiles and leather elsewhere. ¹⁴⁴ In short, the archaeology of Melos clearly shows what the evidence for Epirus and Cos could only hint at: as Italians in the diaspora acquired landed resources, they capitalized on a particular ecological potential of a region and intensified its exploitation, thus fundamentally transforming the economy of the place in question.

Milo 1a amphorae have only recently been identified and connected to the production and export of alum from Melos. ¹⁴⁵ As a result, their findspots cannot in any way be taken to be representative of the export of alum from Melos, but they are nonetheless significant for our purposes here. Starting in the first century b.c., these amphorae and the alum that they contained made their way to the western Mediterranean — to Italy, to the north in particular, and to southern Gaul, where some of the same Latin stamps were also found. ¹⁴⁶ This is particularly noteworthy because Lipari, situated to the north of Sicily, had plentiful alum deposits that were exploited at the time. In other words, Melos and Lipari did not supply the western and eastern Mediterranean respectively, as has been hypothesized. ¹⁴⁷

Two factors explain this phenomenon. First, just as we argued in the case of Epirote cattle and Coan/Chian wine, the Italians who owned alum resources on Melos and organized their exploitation were seemingly interested in the export and shipping of at least a part of their products. The location of their processing sites, which are all close to natural harbours, ¹⁴⁸ points in this direction, as do several tituli picti that have been found on the necks of their amphorae. It has recently been suggested that these tituli picti pointed out the recipient of the amphora in question. ¹⁴⁹ The Melian amphora with SEPUL written on its neck that was found in Padua, where the Sepulli were a well-known family engaged in various craft activities, is a particularly striking piece of evidence in support of this idea. ¹⁵⁰ In other words, it seems likely that the Italians on Melos acted as connectors, that their relationship with people in and from Italy, rather than the need for minerals, brought the Melian alum to the western Mediterranean.

Second, at least by the first century a.d., Melian alum was not simply any alum. Among the many sources of alum through the empire with which Pliny and Dioscorides were familiar, it was considered the best — the best for tanning, for processing wool, and for medicinal purposes. ¹⁵¹ These references stand in stark contrast to Diodorus' thinking about Melian alum, who, writing in the middle of the first century b.c., emphasized the Melian deposits' paucity. ¹⁵² In other words, in the western Mediterranean at least, the alum that the Italians on Melos produced became a high-end product. Pliny referred to alum and other Melian minerals as nobilissimum, laudatissimum and optimum ex omnibus, adjectives that nicely capture the intertwining of the product's quality and its consumer's social standing. Melian alum thus presents yet another case in which the Italians in the diaspora engaged in the production and marketing of goods that helped people in the imperial centre transform economic change into social difference. The Italians, we suggest, were involved in the making of such high-end goods — not only physically on the ground in the provinces but also as regards their reputation in Italy.

IV DISCUSSION

The economic lives of Italians in the diaspora included a great many things: they were slavers, managing the sale and movement of Roman armies' captives; ¹⁵³ they were art dealers, acting as middle-men between Greek production and Italian tastes; ¹⁵⁴ they were also actors, catering, in part at least, to the tastes of Italians abroad. ¹⁵⁵ We have argued here that landholding and involvement in local production were not exceptional among these Italians, but should be considered the norm along with the provision of credit and the movement of people and goods. Although the evidence is not always straightforward, it seems likely that the widespread nature of Italian landownership that we have diagnosed in the late Republic should also be imagined elsewhere in the Greek East and in other parts of the empire. ¹⁵⁶ As a result, Italian landholding in the provinces, which is so familiar to scholars of the imperial period, has a distinct and significant pre-history in the late Republic.

The evidence from this earlier period also suggests that these landholders were much more socially diverse than imperial sources, which mostly concern senators, would lead us to believe. These blind spots in scholarship on both sides of the republican-imperial divide are no accident. They reflect a significant but under-appreciated moment in the history of Italian, and more generally Roman, landownership in the provinces — the moment, most likely part of Augustus' social legislation, when land in the provinces began to count for the census evaluation of Roman citizens. ¹⁵⁷ A far-reaching change in the political economy of the empire, which probably provoked another wave of Roman land-grabbing in the provinces, this reform, and the patterns of evidence that it produced, should not lead us to obfuscate the early days of Italian landholding in the provinces. The figure of the vilicus/οἰκονόμος that imperial inscriptions from Anatolia attest so frequently already featured in the province of Asia as Cicero knew it. ¹⁵⁸

In these early days, then, members of the diaspora exploited their landholdings in a variety of ways, leasing them out, for example, or growing grain. ¹⁵⁹ At the same time, our case studies reveal a distinct pattern of exploitation, a particular strategy that several Italian landowners pursued. Focused on the production of high-end goods with a view to marketing at least some of these goods in Italy and Rome, these Italians appear to have followed the same economic and socio-political imperatives that underpinned pastio villatica — the cultivation of high-end food items for the Roman dinner table in the suburbium of Rome and in the Bay of Naples — on an empire-wide scale. ¹⁶⁰ Rome and other Italian cities, the inhabitants of which were eager to join the Roman élite in consuming empire, provided stable markets and high returns for high-end goods from the provinces. At the same time, Italy and Rome were precisely the places where, just as their fellow Romans and Italians, they would hope to put their products and productivity on display — something that their contemporaries often did in the context of their villae. ¹⁶¹

This perspective reveals these Italian landowners as entrepreneurs, which they are often called, only to the extent that, just as for their modern counterparts, the rules of the game that they were playing were firmly set. The goals of this game, though, were not limited to maximizing income. As we have seen, several Italian landowners in the provinces were equestrians, and others probably had the land that would have allowed them to join that group had it not been located in the provinces. While rarely aspiring to a career in politics, these men focused on producing and marketing goods that allowed them to claim a place in the society of the imperial centre. Some negotiatores, Cicero suggests, staged their return to Italy on the model of the reditus of a Roman magistrate, with large crowds in attendance. ¹⁶²

The diagnosis of this economic and socio-political strategy makes it possible to posit further places that became ‘outposts of the demographic and economic dynamics' that Roman imperialism created in analogous ways to Epirus, Cos and Chios, and Melos. Phrygia, and in particular the Lycus valley, seems a prime candidate. While Italians, both as individuals and as collectivities, are already attested in Phrygia in the early first century b.c., there is no direct evidence for Italian landholding at the time. ¹⁶³ That being said, Stephen Mitchell has shown that members of Italian families that we know owned land in Phrygia in the first three centuries a.d. were already present in the late Republican period. ¹⁶⁴ Importantly, cities in the area of the Lycus valley — the towns of Laodicea, Hierapolis, Philadelphia, Colossai, and possibly Acmonea, that is — saw an analogous reshaping of their ecological potential to the one we found in our three case studies. While Phrygia was previously known for textile-production, a passage from Strabo suggests that the wool and fabrics of the Lycus valley only became famous as high-end products — as even finer than Milesian wool, the epitome of luxury since the Archaic period, Strabo and Pliny claim — in the late second and early first centuries b.c. ¹⁶⁵ To us, at least, the coincidence of this change in local textile production with Italian presence and possible landownership in the area suggests that the origins of the Lycus-valley region as the most important centre for textile production in the eastern Mediterranean quite possibly lay with the particular economic and socio-political strategies that members of Rome's imperial diaspora regularly pursued. ¹⁶⁶

Locally, in the places where these Italians had their landholdings, these strategies meant displacement, changing economic landscapes, and the disruption of previous ways of life — their behaviour constituted fractals of Roman imperialism, one might say — with local variations, of course. While on Cos the Italians joined in an already flourishing production of wine and, just as in Messene, owned a portion of the estates in the city, in Melos they and their agents appear to have completely taken over the ownership and exploitation of mineral resources on the island. Similarly, the dramatic disruption of previous ways of life that the Roman villa at Agios Donatos built on the site of a Hellenistic fort symbolizes does not need to be imagined on the Aegean islands.

On a different level, Italian landholdings spelt changes for the Greek cities in which they were located. In principle, land in these cities was the basis of political power, and Italians with landholdings there began to participate in civic life. The alum-producing Italians on Melos who held civic offices on the island are one example of this phenomenon. ¹⁶⁷ At the same time, several Italians owned land in more than one city — a circumstance that made them less dependent on individual ones, while also increasing their power in relation to each. The fact that Nicias Curtius, who became tyrant of Cos in the middle of the first century b.c., was a freedman of one of the Italians producing wine on the island, who also had several properties in other cities in the region, is a good illustration of just that dynamic. ¹⁶⁸

However, not only the patterns of Italian landholdings, but also Italian economic strategies might have posed new challenges for the idealized, but unobtainable autonomy of Greek cities. A second-century b.c. decree from Abdera that granted M. Vallius the right to export one hundred medimnoi of grain per year for his private use might reveal a glimpse of the potential conflict between local concerns with the food supply and export-eager Italians interested in highly specialized production. ¹⁶⁹ While there certainly existed various local ways of negotiating these conflicts, it seems likely that the Italians' particular strategies of exploitation further exacerbated the dependency of the communities in which their estates were located, providing yet again opportunities for Italians to take on rôles locally. In short, if Italians exploited their estates in part at least with a view to markets and audiences in Rome and Italy, the fact of landholding and these particular strategies of exploitation also got them involved in politics locally. Atticus seems to have practised such politics, both in Buthrotum and on Corcyra. ¹⁷⁰

Lastly, we would like to suggest that in addition to raising new questions about the history of ownership and exploitation of Italian landholdings in the provinces and how to explain it, our arguments here also have implications for how the Roman Empire fits into the history of empire and imperialism more generally. Arguably, the widespread nature of Italian landholding and involvement in primary production that we have sought to establish, make Roman imperialism in the late Republic look more similar to the imperialisms of early modern Britain or of the Dutch Republic, where diasporic landholding and plantation agriculture have long dominated the picture, than it has ever done before. Teasing out and explaining the similarities and differences between these cases — whether they concern the types of goods produced, the relationship of the imperial diaspora to imperial power and institutions, or the diaspora's relationship and attitude towards local populations — now appear as intriguing research problems. Admittedly, the comparison between modern empires and the Roman case has a long and problematic history, being often carried out by nineteenth- and early twentieth-century imperial administrators themselves. The perspective that we propose — the perspective of comparative expropriation, one might say — we hope can escape the pitfalls of this tradition and contribute to refining our understanding of the Roman case and of modalities of empire and imperialism more generally.