An Integrated Survey of Roman Urbanization at Potentia, Central Italy

Transcription

1 395 An Integrated Survey of Roman Urbanization at Potentia, Central Italy Frank Vermeulen Geert Verhoeven Ghent University Ghent, Belgium Our research contributes to the study of Roman urbanization in the Italian peninsula, both in the central Adriatic area and beyond. It focuses on the integrated use of archaeological field methods and non-destructive techniques. The study of the urban layout of the city of Potentia is an example of the use of low altitude aerial detection combined with regular grid-walking and geophysical survey to intensively investigate abandoned classical town sites. Information can be combined in a new approach to the study of urbanization processes in the heartland of the Roman world, and this integrated methodology can be applied to any large archaeological site with regular patterning such as orthogonal street layouts and planned civic centers. Introduction We present here some striking results of integrated archaeological field survey work carried out since 2000 in the central part of Adriatic Italy. Our research is part of the long term Potenza Valley Survey Project (PVS) of the Department of Archaeology at Ghent University, Belgium. The main aim of the PVS Project is to study urban and rural occupation patterns in the valley of the Potenza river, from prehistoric times into the Middle Ages, with a special focus on the period of first urbanization and the so-called Romanization of the area (ca. 300 b.c.-a.d. 500). Apart from objectives connected to wider themes such as Italian settlement history and Roman colonialism, this predominantly geoarchaeological research also pursues method- ological objectives. These include the development of interdisciplinary geoarchaeological survey methods and the refinement of integrated historical- archaeological GIS work. The Potenza region can thus be regarded as a test case for the development of methods for landscape research within a well-defined archaeological and chronological framework. The intensive surveys of the Potenza valley include systematic aerial photography of the large valley area (ca. 400 sq km) between the Apennine hills and the Adriatic coastline, combined with regular archaeological fieldwalking and detailed geoarchaeological field studies in three sample zones spread between the upper valley and the coast (fig. i). Apart from these rural surveys, which were the subject of most of the first phase of the project (Vermeulen and Boullart 2001; Vermeulen, Monsieur, and Boullart 2002; Vermeulen et al. 2003, 2005), intensive intrasite fieldwork is now also being carried out on the main proto-historic centers in the valley (Piceni culture, 9th-early 3rd centuries B.C.), as well as on the four Roman towns located in the Potenza corridor. From west to east, the latter are Septempeda, Trea, Ricina and Potentia. As argued elsewhere (Vermeulen and Verhoeven 2004), at least half of the Roman town sites in the central Adriatic region of Marche were partly or completely abandoned in the early Middle Ages and were never built upon afterwards. It is therefore not surprising that all of the former urban sites of Roman date in the Potenza valley have reasonable potential for surface survey. Because of the limited excavation prior to our work in the area, much can be learned from an integrated survey approach. Our surveys are designed to improve general understanding of the topography of proto-urban and urban settlement in this part of Italy and to deepen our knowledge of the processes of early (mostly Roman) urbanization in Marche. They integrate the results of non- destructive survey methods, such as active aerial photography and air photo interpretation, geophysical sur- veys, geomorphological surveys, detailed topographic measurements, and classic fieldwalking. The results of small-scale previous or ongoing excavation work on these

2 396 An Integrated Survey of Roman Urbanization at Potentia, Italy /Vermeulen and Verhoeven Figure 1. The location of Potentia in the Marche region of Italy. cities, and artifact studies of survey materials and excavation finds, are integrated into this picture as they become available. The Roman City of Potentia Like almost half of all Roman towns in the Marche region, Potentia became from late antiquity onwards "unsuccessful" in some way and was completely abandoned in medieval times (Alfieri 1981, 1985). Almost no modern habitation is present and as agriculture now dominates the former town area, the ancient topography of the city can still be studied with our survey methodology. We concentrated on the all-important intramural part of this Roman city, where the conditions for such an integrated analysis were best. Historical Context Roman military dominance over the central Adriatic re- gion of Marche was established by 268 B.C. The majority of the population living here, mostly belonging to the Italic Piceni peoples, was soon incorporated into the Roman state. The installation of a series of mostly maritime Latin and Roman colonies during the second part of the 3rd and the 2nd centuries B.C. was a major impulse for the Romanization of a region that had no real urbanized society before the Romans (e.g., Alfieri 1977; Moscatelli 1987; Paci 1991; Delplace 1993). In the Potenza valley, one of the most important natural corridors linking the hilly Adri- atic area with the broad Tiber valley and the Tyrrhenian heartland, the Roman impact became very visible from 184 B.C. onwards. With the foundation of a coastal colony for Roman citizens, Potentia, in that same year (Livy, History of Rome: 39.44) the lower Potenza valley, and with it the whole area of central Marche, entered its first phase of real urbanization. Although for a time this city and its territory probably only functioned as a "Roman island5' in a still very conservative Italic territory, it grew into a fully devel- oped urban center, having a major impact on Piceni society as a whole. According to the written sources, which are particularly relevant for our topographic work, the colony received a circuit wall with three arched gates, a street network with sewers, an aqueduct, a temple to Jupiter, and a portico with shops to close the forum in 174 B.C. (Livy: 41.27). Although the "military or civil" character of this first Roman colony, and the number of its initial colonists (300 to 2000 persons) is still a matter of scholarly debate (Paci 2001: 22-23), the urban pattern of the coastal settlement was fully developed only 10 years after its foundation. Official sources for the later history of the town are, however, minimal. Around the middle of the 1st century before our era, probably in 56 B.C. (Cicero, de Haruspicum Responso: 62), a major earthquake destroyed part of the town. Epigraphic evidence, however, testifies to the flourishing development of the town from the Augustan Age onwards into the late 2nd century a.d. (Paci 1995, 2002). The lack of in- scriptions from the 3rd century a.d. onwards could point to a downturn in the city's prosperity. By the end of the 4th/early 5th century A.D., Potentia became a bishop's seat. There is no reference to Potentia in the work of Procopius, who, in his History of the Wars (6, 7) discusses the devastation of the Gothic-Byzantine war (a.d ) in this region of Picenum. Under Lombard dominance (after a.d. 580) the coastal cities suffered as the result of the collapse of the maritime economy, and although Potentia is mentioned at the beginning of the 7th century a.d. in the Cosmographia of the Ravennatis Anonymi (31, 34; 36, 8 [Schnetz and Zumschlinge 1940: 68, 84]) and in the Geographia of Guido (21, 35; 68, 54 [Schnetz and Zumschlinge 1940: 117, 129]), the city clearly did not survive into the Middle Ages (Alfieri 1977, 1981). Archaeological Investigations Before 2000 The Italian archaeologist Nereo Alfieri started to work intensively in Marche shortly after World War II and was the first to locate the urban site of Potentia. He established that the town was erected a few hundred meters south of the Potenza river (ancient Flosis), immediately south of the present-day coastal town of Porto Recanati (Alfieri 1947,

3 Journal of Field Archaeology /Vol. 31, ). Since this discovery, several archaeological studies have been undertaken at the site. Rescue excavations in the 1960s and 1970s revealed parts of the northern cemetery and elements of a housing sector in the ne corner of the town (Mercando, Sorda, and Capitanio 1974; Mercando 1979). A study by Moscatelli (1987) of Royal Air Force photos from World War II and shortly after revealed the town's regular street grid, while Paci (1995) produced a bibliographical synthesis and analysis of all known monuments and inscriptions. Since the mid-1980s, small-scale but systematic excavation campaigns under the direction of Edvige Percossi Serenelli have been organized on a regular basis in the monumental center of the city. These stratigraphic excavations by the Soprintendenza per I Beni Archeologici delle Marche have revealed that the site may have been already occupied in the Early Bronze Age, but nothing so far contradicts the idea that the Roman colony was the original settlement. Excavations in an area just south of the geographical center of the Roman city revealed a late Republican temple (2nd century B.C.), surrounded by a portico and other buildings of Republican and Imperial age (Percossi Serenelli 2001). According to the date of the portico surrounding the original temple and of a possible market building (macellum) to the north and a luxurious building to the east of the sanctuary, the city center was further developed during the reign of Augustus. A flourishing period under Trajan and the Antonines was followed by a period of political instability in the 3rd century. During the economic revival of the late 3rd and 4th century a.d., the central area was reorganized and most finds (e.g., the coins) suggest some prosperity until the beginning of the 5th century a.d., followed by a clear decline. The latest archaeological finds in this area are dated to the 7th century A.D., but it remains unclear what exactly the character of this late period in Potentia was. Of some relevance for our topographic work conducted in and around the city area was a 290 m-long archaeological "clean up" operation of a modern ditch that crosses the city area in its southern part (Budini and Rossini 2001). This field exercise during 2000 revealed the locations of several walls and floors of town buildings, several elements of ancient roads, as well as the circuit wall, the western section of which had been cut by the ditch. Topographic Setting At present the archaeological site of Potentia lies some 100 m from the Adriatic coastline, on a beach ridge oriented n-s (Vermeulen et al. 2003: 82). The terrain is generally flat, lying on average 3 masl, but it is possible that the Romans slightly altered the local topography when build- ing the city. The geology of the area consists of a sandy beach ridge with local gravel beds underlying a thin layer of alluvial clay. This beach ridge was cut on the south by the course of the river in Roman times. The present-day course of the Potenza river, more than 1 km north of the Roman paleo-channel, is the result of late and post-medieval human interference (Alfieri 1947). Detailed mapping of the ancient coastal plain by means of augering and electrical resistivity measurements under the direction of Morgan De Dapper of Ghent University should further elucidate the landscape situation in the immediate sur- roundings of Potentia during its long history (Vermeulen et al. 2003; Vermeulen, Verhoeven, and Semey 2005). It should also clarify the precise roles of river floods, alluvial activity, and sea level change in the process of the gradual abandonment of the town during late antiquity. The site is currently crossed from its nw edge to the se limit by the wide coastal road and adjoining railroad (figs. 2, 3). A few modern houses and gardens cover small parts of the former city area, but almost three-quarters of the original urban site is in use as agricultural land today. The central part, west of the main coastal road, is grassy, as this segment near the excavation area of the temple is now an archaeological zone protected by law. Most other fields on or near the urban site are arable land. Survey Methododology Aerial Photography From 2000 onwards, the PVS team has developed a program of intensive overflights, with yearly campaigns of systematic aerial photography between the months of April and October (Vermeulen 2004; Vermeulen, Verhoeven, and Semey 2005). Photographs, including film and digital images, were taken from a small aircraft piloted by mem- bers of the local aeroclub. The area of the town of Potentia has been continuously monitored, with regular overflights during different seasons and in different weather conditions. Frequent overflights during the dry spring of 2003 produced images that enabled location of many aspects of the urban topography including two Roman cemeteries, four roads connected to the town and several extramural settlement areas. Some flights were useful for monitoring the geomorphology of the changing river course and the precise location of the beach ridges and lagoons. The regular flights over the original intramural area of Potentia resulted in some excellent images of crop and soil marks. The latter are mostly the result of plo wed-up larger stone structures, such as the city streets and remains of the circuit wall, but also occupation layers, zones with locally more organic substances in the upper layers, and traces caused by differential drying of the soil in some parts of the town. The

4 398 An Integrated Survey of Roman Urbanization at Potentia, Italy /Vermeulen and Verhoeven Figure 2. New simplified plan of Roman Potentia and its immediate surroundings, based on aerial photography, systematic field survey, geophysical survey, and earlier excavations.

5 Journal of Field Archaeology /Vol. 31, Figure 3. Oblique aerial view of the urban site of Potentia from the west (Spring 2003). Crop marks and the excavated temple zone (center right) can be seen. detected crop marks are connected with the colony's street grid and circuit wall, and with many building structures. The aerial photos were stored and further manipulated (image enhancement, rectification, mapping) in a GIS environment. For this mapping and interpretation work the team also made use of existing aerial photographs, including those made in 1947 by RAF flights over the area and those made by the Istituto Geografico Militare in 1972 and The integrated GIS mapping of all the aerial data was facilitated by the use of Airphoto 3.x software, developed to rectify archaeological images made with handheld non-calibrated cameras (Scollar 2002: 167). Geophysical Surrey Although the aerial survey revealed much new and valuable information, locating and mapping the remains of subsurface archaeological features in the central (protected) part of the colony, near the recent sanctuary excava- tions, was difficult. To counter this, a large-scale geophysical survey was initiated in 2004 (Vermeulen, Hay, and Verhoeven in press). An area in the south of the town was also surveyed to compare the results with data obtained from aerial photography and grid-walking. The magnetometer survey aimed to elucidate details of the layout of the Roman colony, the location of the city walls, and the position of the western gate, particularly in an area where vegetation prevented good data acquisition by means of aerial photos or fieldwalking. Also, the potential for resistivity survey was tested on a targeted area of the site. For the geophysical survey, grids of 30 x 30 m were set out to ensure that the survey traverses crossed the line of potential archaeological features at an angle of approximately 30. The magnetometer survey was undertaken using a Geoscan Research FM36 Fluxgate Gradiometer. Readings were taken at 0.5 m intervals along 1 m travers- es. An automatic encoder trigger was used to record the readings, allowing the survey to be conducted more rapidly as the area was relatively free of obstructions. The depth to which the Geoscan Research FM36 Fluxgate Gradiometer can take readings varies with the geology but is usually up to 1 m. The magnetometer survey covered in total an area of approximately 7.2 ha (fig. 4). In addition, a small area covered by magnetometry (near the western gate) was chosen to test the response to

6 400 An Integrated Survey of Roman Urbanization at Potentia, Italy /Vermeulen and Verhoeven Figure 4. Results of the magnetometer survey, showing subsurface streets and walls. resistivity survey techniques. This was carried out using a Geoscan Research RM15 Multiplexer set up with a double twin probe array in order to speed up the data collection process and therefore cover more ground. Readings were recorded at 1 m intervals along 1 m traverses. Topographic Survey In 2005, a series of topographic measurements of the urban site were made. At a low level of precision (1 m contours) this strategy allowed a secure topographic model to be produced, in accordance with the data available from the topographic map of the area at the scale of 1/10,000. This helped us to better understand the overall topography of the site and enabled us to better rectify the oblique aerial photographs and to position the results of the geophysical surveys in a GIS model. Fieldwalking and Artifact Studies In 2002 and 2003 the PVS team undertook two short campaigns of intensive and systematic fieldwalking in and around Potentia. That part of the wide urban area inside as well as outside the proposed circuit wall, which is currently used as arable land, was subdivided in regular units and large samples of datable ceramics, building materials, and other artifacts were collected. Generally 40 x 40 m blocks were walked, but during the second campaign in 2003 on a freshly ploughed plot in the western part of town, we used 19 x 19 m blocks. Every square of the 40 x 40 m grid was walked by 3 or 4 persons within a 20-minute time limit. The smaller 19 x 19 m grid was walked in somewhat shorter time spans and with only two persons per block to allow reasonable comparison with the other blocks. The collected sample is small but representative of the surface finds. The results are comparable over the whole (intramural) area of the city as conditions of visibility were almost the same in all fields and all walkers had more or less the same survey experience. Once collected, the ceramics were sorted by class (fine

7 Journal of Field Archaeology /Vol. 31, wares, coarsewares, and amphorae) and type (e.g., Black glazed, African Red Slipped, etc.), and the building materials were sorted by category (e.g., mosaic tesserae, bricks, etc.). After a first interpretation of the finds (Vermeulen et al. 2005; Monsieur, Verreyke, and Boullart 2003) their relative densities across the site were mapped within the GIS. This particular field approach was necessary not only for the control and confirmation of remotely- sensed images (such as our aerial photographs and the results of the magnetometer survey), but also for a more detailed under- standing of the chronology, functional zoning, and spatial development of the urban center. This intensive field survey of Potentia was carried out in close collaboration with the geomorphology team in order to take into account biases introduced by physical processes at the site, such as erosion and riverside sedimentation. Artifact density maps of the presumed habitation center revealed differences in occupation density in several sectors of town. Although their chronological and functional significance needs to be further analyzed, they generally confirm the chronological span of the town's life as deduced from the recent excavations in the monumental center (Percossi Serenelli 2001): from the early 2nd century B.C. to the 6th or early 7th century of our era. Re-examination and Integration ofexeavated Data The comparison of the excavated data with the survey data is essential not only for artifact identification and dat- ing purposes, but also for the interpretation of the survey results. Thanks to collaboration with the Italian excavators we can use the published material from the old excavations in the ne quarter of the town, and integrate still unpublished structural data and finds from the ongoing excavations in and near the Republican sanctuary situated at the heart of the city. When the ne quarter of the city was studied by Liliana Mercando, the site was interpreted as an elaborate farm building (villa rustica). Later the buildings were assigned to two intramural town quarters (insulae) of Roman Potentia (Percossi Serenelli 2001), an identification which helped us to propose a new reconstruction of the city plan. The Survey Results and their Integration Although Potentia was first identified in the 1940s, and some small-scale excavations have revealed some of its topography and stratigraphy, the field surveys started in 2000 made a detailed plan of the urban pattern possible. Our results identify the location of the town defenses (including the gates), the street grid, the forum, several other monumental complexes, many elements of city housing, three extramural funerary areas, and a large segment of the suburban and rural settlement system and roads, connect- ing the city to its territory. Grid-walking and subsequent artifact studies and their integration with partly re-studied excavation data and with the very valuable historical information about the early years of Potentia, have allowed us to make major progress in understanding the topography, urbanization, and settlement history of this Adriatic town, building upon the excellent recent synthesis by Percossi Serenelli (2001), who focused on the excavation data. Location and Geomorphological Context The remote sensing information, in conjunction with data from geomorphological and geophysical fieldwork, demonstrate that the ancient linear beach ridge, lying parallel with the coastline and just north of the Roman mouth of the Potenza river, was narrow and not completely straight. In the area of the geophysical survey the patterns become faint towards the western edge of the town site. There is no doubt that the ancient topography is the cause of this: the beach ridge must slope relatively sharply down to the west, a fact confirmed by augering. The depth of mostly clayey material now covering the Roman town masks the archaeological remains so that their detection using magnetometry is very limited. This phenomenon could account for the fact that the Roman streets and other structures immediately near the town wall were not identified from the air. The deeply buried remains cease to affect the growth pattern of the grass seen on the surface, which is the fundamental way of distinguishing features in aerial photographs. The same difficulty with depth of remains applies to the still lower southern and sw parts of town. Here the arable fields are much eroded due to post-roman flooding of the Potenza river in this area, an observation well attested by resistivity and augering (Vermeulen et al. 2003: 14-16). Dominating the magnetometer results in this area is a curvilinear positive feature in a ne-sw orientation that cuts across the se corner of the survey area. To the south of this line it is evident that no trace of the town layout remains: the clearly visible city streets (see below) stop abruptly at precisely the point where they meet this feature. This feature is visible on our aerial photographs and on earlier vertical photos, and is further indicated by the sharp reduction in surface artifacts towards the south. The total station survey will, when finished, allow very detailed mapping of the Roman aspect of the full beach ridge. It is obvious that its quite limited surface, surrounded by an originally very wet landscape, must have put major constraints on the Roman town builders. City Wall As identified by Moscatelli (1987) on the basis of older

8 402 An Integrated Survey of Roman Urbanization at Potentia, Italy /Vermeulen and Verhoeven Figure 5. Aerial reconnaissance and old vertical aerial photos show subsurface elements of the urban topography. 1) Details of streets and the city wall visible on RAF images; 2) Rectified oblique photograph of a street pattern; 3) Rectified and filtered oblique photograph with streets and buildings. aerial photographs (fig. 5), the town defenses of Potentia can be better distinguished with our new data. Re-examination of these images in the GIS overlay with our 2003 oblique aerial photographs and with the information from the magnetometer surveys demonstrates that the situation is quite complex. Although we must be cautious when con- sidering the chronology of these defenses, as no regular stratigraphic excavations have been undertaken, we propose three major phases in the development of the Potentia circumvallation, and suggest that some of the detailed historical information about the early years of Potentia can be linked to this phasing (fig. 6). In the first phase, no doubt coinciding with the official foundation of the colony in 184 B.C. as mentioned by Livy, the chosen town area was probably surrounded by a ditch (fossa) 2 m wide, possibly flanked by an earthen bank {agger) made up of soil from the ditch. This ditch structure was only clearly seen as a dark and sharply delineated crop mark on the vertical photographs of the northern periph- ery and on some of the best oblique photographs of the eastern and western parts of the city area. It seems to surround a very regular rectangular area (525 x 300 m) with a nnw-sse orientation, possibly representing the initial settlement of the first colonists. Shortly after in 174 B.C., according to Livy, (Livy: 41.27), the site of the colony was fully urbanized. Thanks to the financial intervention of the censor Fulvius Flaccus, the new citizens were alotted sufficient funds to develop from a simple defended military-looking settlement into a real town with structures providing the base for full social and economic development. These urban structures included a temple for Jupiter, a circuit wall that contained three arched gates, a regular street network with sewers, an aqueduct, and porticos with shops to close the forum square. The archaeological data support this detailed historical information, for instance in the case of the (second phase) of the town defenses: the building of a circuit wall. The

9 Journal of Field Archaeology /Vol. 31, Figure 6. Urban structures at Potentia based on the integration of survey and excavation data. presence of a circuit wall was first attested during the cleaning in 2000 of a modern ditch, crossing the beach ridge from w to e (Budini and Rossini 2001). Spatially the wall structure was also detected during the geophysical surveys and especially on the vertical and oblique aerial pho- tographs, as a quite sharp but pale crop mark, only some 2 m wide. Its observed building technique, an ashlar structure of dry masonry 2 m wide made with large regular blocks of regional sandstone {opus quadrntum)^ fits well with the chronology of the early urbanization of the city ac-

10 404 An Integrated Survey of Roman Urbanization at Potentia, Italy /Vermeulen and Verhoeven cording to Livy (Percossi Serenelli 2001: 36). The observed town wall seems to delimit a very regular, rectangular area of some 525 x 343 m, or almost 18 ha (measured intra muros). In the ancient Roman system this represents some 15 x 10 actus. When we include the walls in the total city surface area, it covers almost 19 ha. It is certainly significant that this size is close to the 19.3 ha presumed for the central Adriatic colony of Pisaurum (Pesaro), founded in the same year as Potentia (Sommella 1988: 81). Still hypothetical is our identification of a third phase in the development of the town defenses. We suggest that the circuit wall initially surrounded an area of the same size as the space confined by the first ditch of the colony, but that this intra-mural area was enlarged some 50 m to the east in a later stage of the settlement's history. The initial eastern side of the city wall was then replaced by a street- evidenced by a larger and more pronounced trace in the aerial photographs than most other streets- and the new wall (and ditch?) was built on the edge of what is now a local coastal road bordering the current beach area. This hypothesis would certainly explain why the sector in the ne corner of the city, excavated by Mercando in the 1970s, was only inhabited from the Augustan era (late 1st century B.C.) onwards. It would also explain why in our aerial photos the e-w streets of the housing blocks in the eastern periphery of the town have a different look than the other city streets. Such a partial reorganization of the town is also suggested by the stratigraphic discovery of a destruction layer in the temple sector dated to around the mid- 1st century B.C., when Cicero mentions a major earthquake destroying Potentia (Percossi Serenelli 2001: 39). It could indicate that the town might have suffered greatly from this catastrophic event, and that major reorganization and rebuilding was necessary in the decades following the earthquake, as can also be seen in the whole temple area. Town Gates and their Suburban Links Whatever the exact phasing of the defenses may be, it is likely that the town wall had three major gates, as suggested by Livy. As can be expected in a regularly planned colonial town built from scratch, they were more or less centrally placed on their respective sides of the town, namely the north, west and south. The presence of a gate on the eastern, coastal side cannot be attested. Thanks to our surveys the northern and western gates have been well located, and await confirmation by excavation. For the northern gate there is evidence from several aerial photographs displaying a clear breach in the city wall and ditch traces, and the grid- walking of the plowed field in this area revealed a dense surface concentration (covering an area 15 x 10 m) of building materials, consisting of many tiles, some pieces of white limestone, and several large river pebbles. Some aerial photographs along with geophysical evidence hinted at the location and quite com- plex phasing of the second entrance to town on the west. Combining both magnetometry and resistivity in this area has revealed a zone 15 m long and oriented e-w, full of anomalies that seem to indicate the presence of high concentrations of building materials and a complex gate sys- tem. A third gate in the southern part of the wall can also be located, despite the fact that it is in an area now eroded by post-roman stream movements. This southern gate links up neatly with a road leaving the town area to the sw towards the main Roman bridge over the Potenza river (Vermeulen and Verhoeven 2004; Lilli 1999). This road, and also the major roads leaving the northern and western gates, were discovered and mapped thanks to the study of vertical photographs and especially with the help of our aerial photography since Because of these aerial surveys and connected ground surveys we now know that cemeteries with funeral monuments facing the roadways bordered all three roads leaving the city gates. A large cemetery north of the city was already known thanks to rescue excavations in the 1970s (Mercando, Sorda, and Capitanio 1974) and was recently further confirmed by still unpublished excavations in 2004 by the Soprintendenza per I Beni Archeologici delle Marche, during which the traces of several square limestone funerary monuments along the road, seen as crop marks during one of our flights, were excavated and dated. A comparable picture emerges along the road (oriented sw) leaving the city from its southern gate. Here crop marks of at least five such funerary monuments along the road traces were observable from above and, although no excavations were conducted, surface concentrations of marble, limestone and early Imperial pottery (1st century a.d.) connected with them support this interpretation. Finally, the situation in the areas outside the supposed west- ern gate seems almost identical. Here, the main road leadthe still stand- ing out of town in a westerly direction along ing concrete core of a Roman funerary monument (HI Toracciov), had suggested another cemetery area (Vermeulen and Verhoeven 2004). In 2005 the magnetometer survey revealed, immediately beyond the limits of the town wall, several possible traces of mctusolea flanking the sides of the road. They can be distinguished as a number of rectilinear and curvilinear positive anomalies, further confirmed by the results of the restricted grid survey of this area in The latter not only suggested the presence of the circuit wall, and its persistence from late Republic to late antiquity, but also demonstrated that the surface finds, mostly building materials, pottery, and glass, were much

11 Journal of Field Archaeology /Vol. 31, more abundant intra muros than outside the town area. We can now link the presence of fragments of marble and mosaic tesserae in the extramural area to this funerary area along the western road out of town. Finally, we must stress that some of the extramural surface finds and traces found during remote sensing are certainly to be linked to small areas of habitation and other activities outside the city walls. This is particularly evident in the southern sector where ongoing geoarchaeological research suggests the presence of a Roman harbor near the ancient river mouth. Street Network The gates link up well with the regular nnw-sse orthogonal street grid. Several tracts of this orthogonal street grid, mostly seen on our aerial photographs as pale and linear crop marks (fig. 5) and soil marks 5-6 m wide, were identified on the ground and complemented by the results of the geophysical surveys. Most tracts are seen in the plowed surface as clear concentrations of river pebbles mixed with some artifacts, mostly pottery and tile. Excavations near the temple showed that most streets are ca. 5 m wide and consist of a surface of river pebbles. The excavations also indicate that at least some of them had one or two sidewalks and brick sewers, the latter lying underneath the length of the roadbed (Percossi Serenelli 2001: 80). According to our aerial photography, the somewhat broader (7-8 m) street segments oriented e-w in the younger (>) peripheral eastern part of town seem to be even simpler, with a roadbed of rammed earth and a stone lining. The two dominant main n-s and e-w street axes of the planned Roman town, respectively called cardo maximus and decumanus maximus, were probably constructed with much more care. It is likely that both streets, which link up nicely with the town gates, were paved. This can be deduced from their different aspect in the crop marks seen from above, and the stronger linear positive anomalies seen in the magnetometer survey. As their traces are not continuous we presume that the pavement of these main city roads did not survive post-roman plundering of the building materials of the town. We can now map the parallel city streets with some confidence and subdivide the urban space into housing blocks (insulae) of different sizes. It is difficult to calculate the exact number of blocks in Potentia, and it is also not clear whether some streets are part of the grid or local subdivisions of insulae. We think there are a total of 51 to 57 insulae, dividing the city grid into three main zones: northern, southern and central. The northern and southern zones have, respectively, five and four rows (oriented e-w) of five insulae. The width of these city blocks, not incorporating the streets, is consis- tent and measures some 35 m or exactly 1 Roman actus. The length of the insulae are less constant, between 2.5 actus (ca. 87 m) and 1 actus. This system is less regular, but comparable to the proposed street pattern with longitudinal blocks of 70 x 35 m (oriented e-w), or 2 x 1 actus, for the Republican colony of Sena Gallica (Senigallia), north along the Adriatic coast (DalTAglio, Maria, and Mariotti 1991: 154). The central zone of Potentia, determined by the presence of a series of large public complexes, seems to have six larger insulae oriented n-s, with a length of some 3.5 actus or a little more than 120 m, and widths varying between 1 and 1.75 actus. If all or some of these insulae were subdivided by a central e-w street, then a total number of twelve for this central zone is also possible. The Civic Center Basic rules of Roman regular town planning and some 15 campaigns of excavations in the south-central area of the urban space had made clear, long before the start of our systematic surveys, that the forum of Potentia was to be situated in or near that area. The temple found during these excavations is most probably to be identified as the late Re- publican sanctuary of Jupiter, mentioned by Livy (41.27). Although good epigraphic support for this is still lacking, its prominent location and architectural and decorative style, as well as the dates from the sacred pits (favissae) found during the excavations, place this important sanctu- ary at the very beginning of the life of the colony, around the time of the financial support given by the censor Fulvius Flaccus to build a sanctuary and a forum (Percossi Serenelli 2001: 36). As the excavation campaigns were restricted to the area of the temple and parts of some surrounding buildings, and as our aerial photographic evi- dence was still limited because of the vegetation and lack of deep plowing in the central part of the site, the exact location of the forum square remained unknown. The 2004 geophysical surveys in the large zone adjoining the excavated temple area succeeded in locating the forum and the central insulae near the junction of the cardo maximus and decumanus maximus, representing the civic center of Potentia. The magnetometer results were very explicit in situating the forum square, a long rectangular open area of some 120 x 30 m (oriented n-s), immediately sw of this junction. It was bordered on both long sides by rows of shops fronted by a portico, some of which, according to the many tesserae found here during the grid survey, probably contained mosaic floors. The magnetic anomalies testify that several large public buildings surrounded the forum. To the east of the excavations we have already identified a temple and part of a food market (macellum) direct-

12 406 An Integrated Survey of Roman Urbanization at Potentia, Italy [Vermeulen and Verhoeven ly north of it. On the southern short side geophysical re- search suggests a large building, such as a civic basilica, but the magnetic field of this area was too disturbed by the excavation fence to allow further precision. On the northern short side a whole complex of buildings can be distinguished, among which may be a temple and/or bath complex. An early Christian church and/or different civic buildings cannot be excluded, however. The presence in this northern area of many older epigraphic finds having a public character (Paci 2001) and the great number of marble fragments found during our fieldwalking could support this suggestion. The presence of good quantities of Late Roman fine wares in this general area north and nw of the forum seem to suggest to a long and complex history. Houses and City Quarters Finally, our surveys and a re-evaluation of all excavated evidence have also succeeded in producing some information about city housing and the spatial and temporal de- velopment of city quarters. Thanks to the magnetometer surveys and intensive site monitoring from the air, fragments of buildings are visible all over town, as a rule neatly lined up according to the street grid. House walls are generally less pronounced than city walls, broad streets and large public buildings, so it was necessary to constantly in- tegrate these two bodies of data. Thanks to this effort, and despite the fact that the magnetometer surveys did not produce very clear evidence for houses, a good number of partial plans of houses are now discernable. In the most prominent groupings of traces, some of which were clearly linked to the main n-s street axis, we can distinguish remains of large buildings, probably single family residences (domus type), with several rooms grouped around what seem to be central courtyards. At the plowed surface level they are characterized by a high density of finds, especially pieces of brightly decorated wall plaster and sherds of fine wares. Fragments of marble and mosaic tesserae together with masses of roof tiles and amphorae hint towards a built-up area rich in style. In some other cases where the overall plan of the buildings remains obscure, a series of rectangular patches in the crops, visible from the air, could be interpreted as individual floors. As these particular plots often coincide with concentrations of mosaic tesserae, a certain quality of architecture is indicated. In several areas simple rectangular structures, with one of the short sides facing the street, suggest the presence of shops and modest houses of the so-called taberna type. Still, as the overall form and plan of many groups of features is not clear, an interpretation as to their function (houses, shops, ware- house, etc.) cannot be made. Nevertheless, all these new data add much knowledge to the already available partial ground plans of two insulae near the ne corner of the city, where a full battery of house walls, floors, sewers, etc., was excavated in rescue operations (Mercando 1979). Thanks to these rescue excavations and to the evidence from our grid surveys, some chronological information is now also available about housing and about the general evolution of the use of urban space. Although surface finds picked up in the plowed fields- mostly building materials (tiles, build- ing stones, mosaic tesserae, pieces of wall plaster, etc.), pottery, and glass- were abundant throughout the whole intra-mural area, differences in date and density of the artifacts suggest some interesting phenomena. While a clear drop in the density of surface materials in the most southern and se blocks of our survey grid coincides well with the effects in this area of post-roman flooding and erosion by the original Potenza streambed (Vermeulen et al. 2005), a few marked density peaks can be observed elsewhere. The regular distribution over the city surface of low quantities of black glazed pottery is a good indicator of late Republican settlement activity and indicates that most intra-mural sectors were already inhabited or used in the first two centuries of the life of this town. It is likely that the eastern fringe of the urban area, possibly as a result of later expansion of the town in the direction of the coastline, was only fully exploited after the earthquake of the mid-lst century B.C., as can be deduced from the dating evidence of the excavations in the ne corner of the city (Mercando 1979). During the Principate (lst-3rd centuries a.d.) probably the full intramural area was intensely built over, and indications from our surveys suggest that some suburban extramural areas (e.g., sw of the city) were also occupied then (Vermeulen and Verhoeven 2004). In different city sectors strong concentrations of fine terra sigillata (Samian) pottery, an indicator of early Imperial date, is evidence for the uneven spread of rich housing during the booming first two centuries of our era (fig. 7). In late antiquity (4th-7th centuries A.D.), however, there was a serious contraction of the inhabited space. The artifact maps show a remarkable and consistent concentration of abundant Late Roman fine wares in the centralnorthern part of town. This seems to suggest a clear regrouping of the town population and activity in the area around the forum and immediately north of it. These data agree well with the youngest date (4th century a.d.) of the excavated housing structures in the ne corner of town and with the discovery in that peripheral zone of a series of late graves whose proposed date is 5th-6th century a.d. (Percossi Serenelli 2001: 64-65). All this demonstrates that parts of the city, and at least the outer ne edges near the town walls, were most likely not inhabited in the later life

13 Journal of Field Arch neology /Vol. 31, Figure 7. Results of a systematic grid survey in 2002 and 2003 showing the number of sherds per ha of terra sigillata (Samian ware), adjusted according to man-minutes per sq m. Total of terra sigillata sherds collected is 491.

14 408 An Integrated Survey of Roman Urbanization at Potentia, Italy /Vermeulen and Verhoeven of Potentia, when some formerly inhabited space was used for funerary activities. Conclusions: New Approaches to Studies of Roman Urbanization It is commonly accepted that the study of towns is central to any understanding of the Roman world, since they were the nodes through which the administration, economy, and dominant culture were negotiated (Keay et al. 2004). Among archaeologists involved in intensive urban research it is agreed that there is a firm need for the collection of data about the internal organization and chronology of the full range of urban sites in different Roman provinces. As much of our knowledge about Roman urbanism in Italy and other parts of the Empire is based on an uncertain sample of well-preserved but atypical towns such as Pompeii and Ostia, there is a need for more systematically collected data, different from the large body of earlier work on Roman towns, which far too often focused upon individual buildings taken out of the context of their urban environment. Furthermore, Roman urban archaeol- ogy is constrained by medieval and modern developments in those towns that remain occupied today. We believe that carefully defined strategic use of non- destructive field techniques and systematic aerial reconnaissance in investigating such ancient city sites can be useful. We recommend the careful selection of Roman urban sites to study, focusing on the still available range of abandoned sites with well-preserved subsoil structures that allow a fully integrated geoarchaeological approach. To illustrate the value of this approach and to contribute to the broader historical-archaeological debate on Roman urbanization, we have presented the synthetic results of a project that examines urban settlement patterns in the Potenza valley. Our work complements earlier town-based research in a region of Italy (the Marche), which has a long and distinguished tradition of historical and archaeological work on the development of Roman towns (Luni 2003). That research provides important and detailed information about the development of parts of individual towns and town buildings, involving the painstaking excavation and publication of particular structures and the analysis of material evidence, such as that of pottery, inscriptions, and sculpture. Our approach, by contrast, provides broader information for a series of urban settlements located in a central area of Marche. We are interested in the patterns of urban development and their distribution, size, and form, and we believe that new information should be based on the combination of intensive and integrated surface survey and fieldwork rather than solely on excavation or detailed artifact studies. If detailed excavation evidence is available it must, however, be used fully. Integrating data correctly can advance our knowledge of the topography and internal organization of the ancient towns and contribute to a wider vision of urban development in the Roman era. It has been argued that systematic surface survey and large-scale geophysical survey have contributed to our understanding of Roman urban topography at classical town sites such as Tanagra in Boeotia (Bintliff and Evelpidou 2001), and Falerii Novi and Portus in the Tiber valley (Keay et al. 2000, 2004). Both survey techniques are easy to use and allow archaeologists to cover substantial areas in a short time span. As we have demonstrated, it is also beneficial to add detailed investigation of old aerial images and intensive monitoring of the urban site with the help of lowaltitude aerial photography during different seasons and years. We argue that a simultaneous application of these non- destructive techniques with feedback from geomorphological research creates a method ideally suited to characterize the extent, organization, and chronology of urban sites typical of central Italy and other parts of the western Mediterranean. The Potenza Valley Survey project has attempted to refine the non- destructive exploration of one Roman urban landscape. Despite the success of the magnetometer and resistivity surveys, the low altitude aerial photography, and the analyses of these data in a GIS model, it must be stressed that they revealed little with regard to the chronology or specific function of the buildings detected (Keay et al. 2000: 9; Vermeulen, Verhoeven, and Semey 2005). Hypotheses about functions and general chronologies of Roman urban patterns remained problematic. Therefore, we conducted systematic fieldwalking in grids over large sectors of town, contextualizing the imagery of the remote sensing efforts. Together with further integration of the stratigraphic data from excavations within the town, our research has proven to be of crucial importance for the understanding of the city topography and part of its evolution. We are convinced that our integrated methods can be applied beyond the geographical and chronological limits of Roman archaeology and should be particularly useful on all types of large open archaeological sites that have some form of regular or understandable patterning. Acknowledgments The authors are grateful to all those who have contributed to the fieldwork. For the geophysical surveys we thank in particular Sophie Hay of the Archaeological Prospection Services of Southampton, Eamonn Baldwin and Rose Ferraby of the British School at Rome, and Lieven Verdonck and Sarah de Seranno of Ghent Univer- sity. For artifact surveys and finds processing we are very