Download file in PDF format: TRAC 2005: Romanization in Southern Epirus: A Ceramic Perspective (pp. 12–24)
Romanization in southern Epirus: A ceramic perspective
Melissa Moore Morison
Introduction In light of recent critiques of Romanization theory and the concomitant development of new interpretive models of Roman and native provincial interaction, it is useful to reconsider the ways in which archaeological data, particularly data recovered through field survey, articulate with new theoretical constructs. Whether the complex blend of Roman and non-Roman provincial interaction is most profitably conceptualized as Romanization, globalization, creolization, discrepant experience, or some combination of these models, surface survey remains a powerful research tool characterized by an ever-increasing methodological sophistication. However, the interpretive models applied to ceramic survey data have not necessarily developed apace with the field methodologies used to recover this material. This gap is unfortunate, because effective interpretation of ceramic artefacts is clearly essential to the understanding of processes of cultural interaction in the Mediterranean at the regional level of analysis that surveys provide. Pottery litters the Mediterranean landscape, blankets both site and off-site contexts and is the most common form of data recovered by most survey projects. Moreover, when the goal of survey is exploration of rural Roman landscapes (whether political, economic, or cognitive) and comparison of urban and rural responses to Rome, interpretation of ceramic evidence is a critical component of the analytical process – especially in regions where, as in many areas of the eastern Mediterranean, the architectural remains of small farmsteads and even small agricultural settlements are relatively poorly preserved.
The issue is especially relevant in areas of the eastern empire, such as Greece, that enjoyed lively and long-standing traditions of trade in ceramics from both east and west – alongside strong local traditions of ceramic production and exportation – well before incorporation into the Roman imperial system. What relationship can then be posited between changes in ceramic assemblages and ‘Romanization’ (broadly defined)? While it may be possible to attribute quite dramatic shifts in ceramic consumption in the West to relatively clear-cut patterns of production, trade, or negotiation of elite status, this approach seems to lack explanatory rigour when applied to the analysis of cultural interaction in the Greek East. How can ceramic data, especially the notoriously problematic material recovered through survey, be used to reconstruct the social, economic, and political imperatives that underlie patterns of consumption within the complex cultural (and ceramic) entrepôt that was Greece?
A Model for Ceramic Consumption This paper is concerned with preliminary explication of a theoretical model that addresses this issue and with presentation of a case study in which the model was applied to a corpus of pottery recovered through surface survey. This approach is predicated upon synthesis of two distinct anthropological constructs. It is argued that the analysis of ceramics as manifestations both of cuisine and technological agency (Dobres 2000) provides a profitable mechanism with which to address variability in regional processes of response to Rome.
Greco-Roman pottery is frequently presented as signifying seemingly discrete activities such as advertisement of elite affiliation, transport of foodstuffs, cooking, or dining behaviours, rather than as part of an integrated cultural system or as indicative of significant patterns of change in such a system. A useful alternative argument, however, is that ceramic artefacts represent and mediate the suite of socially embedded assumptions, decisions, and behaviours that comprise the larger system defined by the term ‘cuisine’. According to this argument, ceramic artefacts (with the obvious exception of lamps, unguentaria, and the like) may be interpreted as physical manifestations of the choices made at every stage of the cuisine system, which itself functions as a connective node for other cultural constructs such as economy, social organization, and identity formation (Bats 1988; Moore 2000).
The cuisine system has been defined in anthropological terms as the complete set of activities, including manipulation of material culture, through which a self-defined cultural group satisfies its (self-determined) subsistence requirements. These activities comprise several cuisine ‘phases’ that include food production, distribution, preparation, consumption, and disposal. Each phase is composed of sub-phases and ‘aspects’ that articulate explicit links with other cultural systems (Goody 1982: 43-48). Table 1 illustrates the relationships between cuisine phases, sub-phases, and other aspects of the system.
Ceramic artefacts operate as representations of the technological sub-phases and aspects of cuisine phases as illustrated in Table 1. Standard functional categories already widely employed in analysis of Roman ceramics can be readily linked with appropriate sub-phases and technological aspects indicated in the Table. Transport amphorae represent the technology of distribution, for example, while mortaria represent preliminary preparation of food, ollae represent subsequent application of heat, and so on.
While the nature of individual archaeological assemblages, especially survey assemblages, may limit the extent to which the characterization of normative ‘suites’ of ceramic consumption can be achieved, the breakdown of ceramic assemblages into component cuisine phases facilitates inter-regional comparison as well as consideration of intra-regional diachronic variability in ceramic production and consumption. Regional responses to globalization processes within the Roman empire, for example, may be described in terms of interconnected, negotiated, cascading shifts in access to food resources, modifications to food distribution systems, change in cooking practices, and the like, as suggested by variable production, consumption, and discard of all classes of pottery. Moreover, changes in consumption of functional ceramic groups need not occur at the same rate or for the same reasons. Linkage of functional groups with technological sub-phases highlights and clarifies this variability of response.
In this way the anthropology of cuisine offers a robust interpretive mechanism with which to define and track specific patterns of change in each of these interlocking aspects of provincial life, explicitly connecting variability in both urban and rural production and selection of ceramics with processes of economic and social change. The cuisine model provides a well-defined framework within which to characterize change within and across ceramic assemblages and a powerful tool that links change in ceramic assemblages to change in related cultural systems. Ceramic data from both excavation and survey can then be used to consider some of the cultural consequences of Roman imperialism in both urban and rural contexts, and on both regional and site-specific levels (Moore 2000).
Table 1: Cuisine Phases and Sub-phases as defined by Goody (1982) Cuisine Phase Sub-phases Aspects Food Production Choice of plant and animal resources; organization of labour; use of productive resources (land, water); potential for surplus/famine Technology of production Food Distribution Allocation of food within the unit of production; gift exchange; market exchange; obligatory transfer; deliberate destruction Technology of storage; technology of transport; periodicity of distribution Food Preparation Preliminary preparation; application of heat or other agent of transformation; preparation for service Definition of preparation group; locus of preparation; technology of preparation Food Consumption Assembling of participants; distribution of prepared food; eating and drinking Periodicity of consumption; structure of the meal; ways of eating (‘table manners’); technology of eating; definition of dining group; differentiation of cuisine within and between groups Disposal Initial disposal; recycling; long-term disposal Technology of disposal
When combined with analysis of local/regional ceramic production and distribution, the cuisine model facilitates identification and comparison of cultural choices represented by the shifting composition of cuisine/functional groups within ceramic assemblages. Diachronic variation in selection of local and imported tableware may result as much from differences in land use and potters’ access to clay resources as from a desire to emulate ‘elite’ practice, for example; and cookware vessels may exhibit change in production technology for similar reasons. Such variation need not occur at the same rate even within a single region.
Complete analyses of ceramic consumption patterns are predicated as much on understanding of local production and distribution as on the movement of terra sigillata and transport amphorae. Discussions limited to imported wares alone fail to address a significant component of socio-economic response to Roman rule; see Dobres (2000) for important discussion of the relationships between technology, imperialism, and consumption. Analysis of ceramic technology with reference to the ceramic chaine opèratoire is now standard practice and provides important information about local economic resources and social organization. Moreover, unless care is taken to ensure collection, study, and understanding of local ceramic products, over-representation of highly visible and readily recognizable imports may significantly skew the results of field survey. See, most recently, Sanders (2004) for discussion of this issue with respect to survey in the eastern Mediterranean.
Case Study: Ceramic Evidence from southern Epirus The first explicit application of a cuisine model to the study of Roman ceramics was Bats’ analysis of excavated Hellenistic and Roman assemblages from Olbia/Hyère (Bats 1988). Subsequently, Berlin used coarsewares from Hellenistic and Roman levels at Tel Anafa to reconstruct the cooking practices and, by extension, the ethnicity of the site’s inhabitants (Berlin 1997). Neither Berlin nor Bats considered all aspects of the cuisine system and neither applied the full cuisine model to broader analyses of regional patterns beyond the level of a single site. However, both studies provide ample illustration of the explanatory potential of a cuisine-based approach.
It is useful at this point to consider a case study in which this approach was applied at a regional level. The study in question is based on analysis of Hellenistic, Roman, and Late Roman pottery recovered in the course of intensive surface survey in southern Epirus, Greece, by the Nikopolis Survey Project. The primary goal of this joint Greek-American project was the reconstruction of the southern Epirote landscape from Palaeolithic through early modern times. Of particular interest was the nature of land use in the Roman and Late Roman periods in general and, more specifically, the effect of Octavian’s creation of Nikopolis (shortly after 31 B.C.) on the economy and social organization of the rural hinterland.
Fieldwork and study seasons were conducted in Epirus from 1991 through 1996 under the joint direction of James Wiseman (Department of Archaeology, Boston University), Angelika Douzougli and Constantinos Zachos (12th Ephoreia of Prehistoric and Classical Antiquities), and Frankiska Kephallonitou (8th Ephoreia of Byzantine Antiquities). The first volume of the survey publication series, including discussion of survey and sampling strategy, regional geomorphology and remote sensing studies, is now available as are several brief overviews of the survey results (e.g., Wiseman 2001; Wiseman and Zachos 2003). The second volume in the series will include the complete results of the ceramic study discussed briefly here (Moore Morison forthcoming; see also Moore 2000 for preliminary results).
Epirus was, throughout most of its history, comprised of several distinct tribal areas, each of which enjoyed different forms of political organization and diverse relationships with southern Greece and Italy. The Nikopolis Project survey zone encompasses more than eight hundred square kilometres and is roughly consistent with the ancient boundaries of Cassopaia, one of the tribal sub-regions of pre-Roman Epirus. The survey zone also corresponds to the original territory of Nikopolis. This area extended from Nikopolis and the straits of Actium north along the Ionian seacoast to the mouth of the Acheron River, and east as far as the Louros River gorge. Moving east from Actium the survey zone included the northern coast of the Ambracian Gulf just into the modern nomos of Arta (ancient Ambracia), so that the Louros river delta was included but not the modern city of Arta itself (Fig. 1). Ambracia was not regarded in antiquity as a part of Cassopaia, nor was it part of the territory assigned to Nikopolis at its foundation.
Prior to the Nikopolitan synoecism Cassopaia was inhabited by a mixture of Epirotes and colonists from the Peloponnesian city of Elis. From the sixth century B.C. onward this diverse group was part of a complex web of shifting alliances with Macedonia and the inland Epirote tribes. In contrast, much of the northern coast of Epirus (e.g. Chaonia) was colonized by other Greek states such as Corinth and, in the first century B.C., by wealthy Roman equestrians known in the literary tradition as the Epirotici homines. Though the various Epirote tribes were loosely allied in earlier periods, they appear to have had different internal economic and political structures and different political relationships with Rome in the second and first centuries B.C. Thus, although Epirus was united during the reign of Trajan into a single province with an administrative seat at Nikopolis (a province subsequently divided, perhaps under Diocletian, into two separate entities known as Epirus Vetus and Epirus Nova) the temptation to view Epirus as a monolithic cultural entity with a simple or unified relationship with, and response to, Rome should be avoided.
Figure 1: Nikopolis Project Survey Zone.
The Nikopolis Project was designed, then, to address questions specific to Cassopaia rather than to Epirus as a whole. Following intensive geomorphologic study, a survey strategy was developed based upon sampling of significant environmental zones within the region, including coastal plains, inland valleys, mountain foothills and upland valleys. Fieldwork included both off-site and on-site survey with intensive coverage of areas ranging from small activity scatters to extensive settlements, including four small town sites identified as Archaic-era colonies of Elis (Tartaron 2003).
A total of approximately 250,000 representative ceramic artefacts, including objects not immediately recognized by the survey teams as diagnostic, were collected and analysed according to the cuisine model outlined above. In all, approximately ten thousand of these artefacts are considered in the discussion below. In 1997 and 1998 study was also made of unpublished ceramic material from well-stratified excavated sites within the survey region such as the Classical/Hellenistic town of Cassope, an early imperial farmstead/villa site at Strongyli, and a small area within the recent excavations at Nikopolis. The opportunity to study excavated ceramic comparanda made it possible to develop a good understanding of local ware traditions, an understanding that was crucial to the analysis of regional importation and production processes.
A petrographic study was conducted to characterize technological choices made in the production of the local wares in the Hellenistic, Roman, and Late Roman periods. In addition to characteristic imitations of imported terra sigillata, seven regional ware-groups were used for food preparation and service. Four of these regional wares (called A, B, C, and D) exhibit differences in firing colour but are essentially identical in mineralogy. The differences in firing colour exhibit no significant diachronic, spatial, or functional pattern. These wares were used throughout the Roman period for production of simple pitchers, jars, and cups. The remaining three Roman-era regional wares were used for food preparation vessels such as casseroles, stewpots, and mortaria. These wares are distinct in colour, texture, and mineralogy (Moore 2000: 106–126). It is also clear that each of these three wares had three distinct periods of use, as discussed below.
Table 2 presents the relative percentages of imported and regional artefacts for each of the periods under discussion here. For purposes of this paper, Roman 1 corresponds to the period from the Roman sack of Epirus in 167 B.C. to the battle of Actium in 31 B.C. The Roman 2 period extends from 31 B.C. through the early fourth century A.D., from Actium through the first wave of Slavic incursions into southern Epirus. The Roman 3 period includes the late fourth through early sixth centuries and is characterized by both economic and political prosperity in the immediate vicinity of Nikopolis despite repeated waves of Slavic and Visigoth invasion.
Table 2: Relative percentages of regional and imported wares by period and survey sub-regions Roman 1 Roman 2 Roman 3 % Regional Wares 91 73 24 % Regional, Acheron Valley % Regional, Ayios Thomas % Regional, Cheimadhio % Regional, Elean colonies % Regional, Grammeno % Regional, Michalitsi % Regional, Strongyli % Regional, Vathy % Regional, other 0 15 15 35 0 15 20 0 0 20 19 3 6 18 3 5 17 9 4 40 1 1 7 0 2 30 15 % Imported Wares 9 27 76 % Imported, Acheron Valley % Imported, Ayios Thomas % Imported, Cheimadhio % Imported, Elean Colonies % Imported, Grammeno % Imported, Michalitsi % Imported, Strongyli % Imported, Vathy % Imported, other 0 33 33 0 0 0 0 0 34 0 27 2 0 10 0 5 49 7 6 38 2 0 17 0 0 33 4
Table 2 also indicates the relative percentages of regional and imported wares in the various survey sub-regions defined below. It should be noted that these sub-regions vary considerably in size, environmental features and artefact density. Space does not permit detailed analysis here of the implications of these variables with respect to artefact interpretation, though Tartaron (2003) does provide discussion of these issues as they affected survey and sampling strategies.
Of concern here are broad patterns of ceramic consumption within the survey region. As suggested in Table 2 the consumption of regional and imported wares varied considerably both through time and across space. Consideration of variable patterning in specific functional cuisine groups, as outlined below, provides insight into economic and social mechanisms that may explain these variations.
Ceramic Consumption Patterns: Spatial Distribution of Functional Groups This section presents an overview of regional ceramic consumption patterns in the three periods of Roman/Cassopaian interaction noted above. For purposes of this brief paper, discussion of the role of ceramics in the food production and disposal phases has been omitted. The Latin literary tradition does suggest the use of certain ceramic vessels in these phases but in the context of the surface survey it was not possible to make definitive identifications of such vessels.
Figure 2 illustrates the survey areas mentioned in the discussion below. From (roughly) north to south, these include small farmsteads in the Acheron River valley; the Elean colonies at Kastri, Kastro Rizovouni, Palaiorophoros, and Kastro Rogon; a large farmstead in the upland hills of Cheimadhio; large farmsteads/villas at Grammeno and Strongyli; a large harbour installation at Ormos Vathy; and several sites along the centuriated extension of Nikopolis along the Ayios Thomas peninsula (Ayios Minas, Metamorphosis, etc.). Reference will also be made to an area of small farmsteads in the Michalitsi hills, which extend across an area from Nikopolis to just north of Kanali.
The Roman 1 period exhibits an intriguing mix of both continuities and discontinuities in the spatial patterning of the ceramic data when compared with that of earlier periods. For example, locally produced cooking-vessels, including lopades, mortaria, and basins of standard southern Greek types, have a Roman 1 distribution broadly comparable to that of Classicaland Hellenistic-era cooking-vessels. In all three periods, the material clusters at small settlements in the Michalitsi hills and at the Elean colonies of Rizovouni, Palaiorophoros, and Kastri. Additionally in the Roman 1, examples are found at the large farmsteads at Strongyli and Cheimadhio.
Campanian Dressel 1 amphorae are the only transport vessels represented in the Roman 1, and are only found at Cheimadhio and very small scatters on the Ionian coast. The survey recovered none of the Attic finewares typical at large excavated sites in the region such as Cassope and Ambracia. The presence of late imitation Thin-Walled wares at Grammeno and in small scatters on the Ayios Thomas peninsula, however, attests to settlement in these areas of the survey region as well.
Striking contrasts are evident in the spatial patterning of ceramic functional groups. The distribution of Italian-style tableware and imported Italian wine jars is limited to large farmsteads and areas near harbour installations. Most small inland settlements lack these imports but do exhibit evidence of continuity of occupation in the form of cooking vessels of standard Greek type (e.g., chytra and lopas). Local production of such cooking vessels was apparently not significantly disrupted between 167 and 31 B.C. Rather, the ceramic evidence suggests a rebound from the sack of 167 B.C. as opposed to total devastation and abandonment of the inland settlements.
Figure 2: Settlement within the survey region.
An emerging disparity between the lifestyle of the coast and larger farmsteads, as opposed to small inland farms and villages, is also apparent. As noted above, the Italian imports in harbourages and large villa-style holdings just inland from these harbours comprise a quite different ceramic assemblage than that recovered at smaller inland sites. It is tempting to suggest that this may pattern may attest to an influx of wealthy Roman landholders, as in the area farther north near Buthrotum in this period, but definitive archaeological evidence for this argument is lacking (pace Karatzeni 2001: 171).
After the establishment of Nikopolis through synoecism shortly after 31 B.C., a decline in settlement of the rural hinterland would not be unexpected. In fact, this prediction is not borne out by the survey data. By the late first century A.D. a renewal of settlement throughout the region is apparent. The spatial patterning of the three ceramic functional classes also exhibits an interesting degree of variation in the Roman 2 period.
Not surprisingly, Italian, Dalmatian, and North African transport amphorae were recovered in significant concentrations near the Nikopolis harbour installations and the associated settlement along the Ayios Thomas peninsula, south of Nikopolis itself. Smaller quantities of these vessels were dispersed at the large farmsteads nearby at Grammeno and Strongyli. While the commodities conveyed in these vessels could certainly have been transported inland in other containers, the transport amphorae recovered in this period are not found at the smaller inland sites also inhabited in this period.
Dining vessels exhibit a similar distribution pattern in the Roman 2 period. A limited number of Italian sigillata platters, and a larger quantity of regional imitations of Italian platters, characterize the first century A.D. Both imported sigillatas and regional imitations appear in excavated contexts at Strongyli and Nikopolis; the distribution of comparable vessels recovered through survey is limited to clusters near the harbours and large farmsteads in association with transport amphorae. Very small quantities of Eastern sigillata and African Red Slip appear between the second and early fourth centuries and exhibit the same distribution pattern. In no case have types been found that are without frequent parallel in other regions of Greece.
More surprising is a dramatic shift in the production and distribution of cooking vessels in the Roman 2 period. A few examples of Pompeian Red Ware were recovered at Strongyli and Grammeno, but the vast majority of cooking vessels datable to the three centuries following the Nikopolitan synoecism belong to a round-bottomed olla type common in southern Greece in this period – although the rim shape is somewhat unusual and finds closer parallels to shapes from sites in northern Epirus and Dalmatia. The local Cassopaian examples exhibit a notable degree of standardization of rim diameter and are produced in a clay body notably different from that of the earlier Classical- and Hellenistic-era cooking vessels. The latter essentially disappears at this point from both survey and excavated contexts.
Petrographic study of cooking vessels and regional clay resources suggests that the foundation of Nikopolis may have been accompanied by significant change in the nature of local coarseware production. Local potters began to use a new type of clay with a considerably more restricted regional distribution by the end of the first century A.D. Figures 3 and 4 are photomicrographs illustrating the differences in mineralogy and textural composition between the clay bodies of a Hellenistic lopas (Fig. 3) and a Roman 2 olla (Fig. 4). The Hellenistic vessel’s clay body is characterized by small, somewhat rounded grains of plagioclase feldspar, quartz, and (rarely) chert, with some mica present as well. Figure 5 illustrates a type of clay found in abundance in the Michalitsi hills and many other parts of the survey region. This clay sample was fired in an oxidizing atmosphere to about 750° C and exhibits textural and mineralogical features comparable to those of the Hellenistic cooking vessel. In contrast, the Roman 2 cooking vessel is characterized by the presence of quite large, angular grains of chert, quartz, and feldspar, with very little mica. Figure 6 illustrates a clay type typical of the Ayios Thomas peninsula. This sample was fired in an oxidizing atmosphere to about 900° C and exhibits features comparable to those of the clay body of the Roman 2 cooking vessel.
It should be noted that the differences between the clay bodies are primarily textural rather than mineralogical and that the sample size of this pilot study was small (50 sherds and 28 clay samples). Combined with the changes in standardization, shape, and spatial distribution of the cooking vessels, however, these textural differences do suggest an important shift in production practice in this period, perhaps the result of reorganization of labour, land use or access to raw materials. Nevertheless, a reliance on local industry for production of these wares is still evident and there is little evidence outside Nikopolis itself for adoption of standard Italian cookware forms such as flat-bottomed baking pans.
Figure 3: Photomicrograph of Hellenistic cookware in thin-section, crossed polars
Figure 4: Photomicrograph of Roman 2 cookware in thin-section, crossed polars
Figure 5: Photograph of local clay sample in thin-section, crossed polars
Figure 6: Photograph of local clay sample in thin-section, crossed polars
The distribution of cooking vessels contrasts strongly with that of amphorae and tableware in the Roman 2 period. Transport amphorae, for example, were seen to cluster at the Nikopolis harbourage and the coastal extension of the city along the Ayios Thomas peninsula; no doubt these were meant to supply Nikopolis and the larger farmsteads. The new cookware, however, is present in significant quantities at the former Elean colonies, Michalitsi, the Acheron valley, and the large farmsteads and coastal settlements as well as the Nikopolis harbourage. This broader distribution suggests that the foundation of Nikopolis had little long-term effect on settlement of the countryside. Intensification of settlement occurred along the Nikopolis harbour installations, but the hinterland, at least in terms of settlement pattern, seems to have been largely unaffected. Many rural sites, indeed, exhibit what may be a continuity of settlement from the Hellenistic through the Late Roman periods. It should be noted that only a close attention to definition of typologies for local coarsewares has made this understanding of the settlement pattern possible.
Another shift in production method occurred between the third and fourth centuries A.D. A new technology was adopted in which cooking vessels were fired in a reducing rather than an oxidizing atmosphere. This practice resulted in a dark grey clay body, though the clays used are mineralogically comparable to those used in production of Roman 2 cookware. From the fourth through early sixth centuries A.D., moreover, the distribution of these vessels across the survey region was more restricted than the cookware distribution of earlier periods, suggesting a possible contraction of rural settlement in response to invasion. This pattern has also been noted by Karatzeni (2001: 171–172).
Transport amphorae and tableware datable to the same period (Roman 3) reflect the continued integration of the region in pan-Mediterranean trade networks despite the threat of barbarian incursion. Large quantities of African and spirally-grooved eastern Mediterranean amphorae were recovered by the survey. Predictably, these were distributed in more limited quantities near the inland villas and in far greater numbers in the coastal settlements near Nikopolis. A few examples of spirally-grooved amphorae also appear in association with cookware at small farmsteads along the Acheron valley.
Conclusion Some profitable conclusions can be drawn at this point regarding the social impact of the Roman conquest of southern Epirus and the subsequent political and economic integration of the area into the Roman imperial system. It is interesting to note that while the regions from which luxury food items and tableware were imported shifted over time, importation patterns from the late second century B.C. onward are generally in accordance with established patterns of trade throughout Greece, and do not reflect an unusual range of choices on the part of Epirote consumers. Rather, it seems reasonable to hypothesize that the geographical position of southern Epirus along east-west trade corridors guaranteed at least its urban population access to a trans-Mediterranean flow of luxury foodstuffs and dining vessels. The extent to which much of this material penetrated the hinterland beyond the large farmsteads established by the end of the first century B.C. was apparently rather limited, although settlement in these regions exhibits, at least as evidenced by local cookware products, a strong degree of continuity across several centuries.
Urban and rural response to Roman hegemony, as indicated by ceramic evidence for cooking and practice and the use of dining vessels thus appears to have been quite different. In the rural interior of the region, importation of luxury food items was minimal as was use of imported food service vessels such as Italian sigillata and African Red Slip, whereas the opposite pattern appears to apply in the more urbanized coastal area near Nikopolis and the large farmsteads located in the immediate hinterland. Regional ceramic production strategies do exhibit some significant change from the period of the Nikopolitan synoecism onwards, however.
Southern Epirus had maintained throughout its history a strong reliance on local ceramic production. Although a wide range of imported products was used, particularly in urban centres, the overall quantities of imported wares appear always to have been small in comparison with local products. On the other hand, many locally produced dining vessels were imitations of popular imported types – and imitation of at least some of these types, such as mould-made sigillatas, would have required potters either to learn new technologies or to make way for itinerant potters from other areas. Either way, the southern Epirote potters can be seen as a highly skilled and far-from-provincial group of craftsmen, particularly in light of the notable shift in production of cookwares.
Further examination of patterning in this body of data will provide more nuanced information regarding economic and social responses to the Roman presence in this region. Additional exploration of the evolution of land ownership, taxation, and other aspects of the southern Epirote economy will be necessary to fully contextualize the results of the ceramic study. Nevertheless, this general overview of a simple application of the cuisine model suggests that the study of survey pottery has significant interpretive potential.
Department of Classics, Grand Valley State University, Michigan
Acknowledgements The gracious support and assistance of the Greek Archaeological Service (12th Ephoreia of Prehistoric and Classical Antiquities and 8th Ephoreia of Byzantine Antiquities), the members of the Nikopolis Project team, the American School of Classical Studies in Athens, and the organizers of TRAC 2005 are gratefully acknowledged.