In June 2000 and July 2001 a geophysical survey was undertaken at Vagnari in the Regione of Puglia, southern Italy. The survey was carried out by the British School at Rome on behalf of Professor Alastair Small of Edinburgh University, aimed at locating and mapping any potential archaeological features below the present ground surface. Magnetometer survey was successful in locating and mapping the archaeological remains of the Roman settlement at Vagnari. A large building was identified comprising a number of rooms around a courtyard, and a series of structures associated with industrial activity were located along the northern edge of the gully, including a kiln and enclosure. A potential route between the settlement and the supposed line of the Via Appia to the north was also located.

During excavation of the necropolis six tombs were uncovered in trench 39. Of the five that were excavated, four were grouped in the southern part of the trench. As no evidence of other burials appeared in the eastern edge of the trench, it seems that the number of tombs decreased in this direction indicating that the limit of the necropolis must have been close by. All the tombs were “a cappuccina”, their upper parts emerging at a depth of 0.50 m. The bodies lay directly on a tile surface, in one case the head was supported on several tile fragments. With the exception of two tombs, one an east-west alignment and the other an infant burial further to the north, the burials appeared to be arranged on a north-north-east/south-south-west alignment. In the infant burial numerous nails were found along the sides of the tomb, indicating that the child was buried in a wooden coffin. Amongst the grave goods were a coin of the Emperor Trajan, found near the deceased’s chest, an intact beaker by the feet and a terracotta lamp. The lamp dates to the 2nd century A.D., is stamped with the makers name on its underside and the relief decoration on the disc shows a bearded male figure wearing a long tunic. The excavations undertaken in the late antique building revealed two masonry bases, 5 m apart, situated outside the central part of the structure on the south side. These relate to the supports for the portico that ran along the building’s south façade.

Twelve burials were excavated at Vagnari in 2008, in two trenches (nos. 39 and 49) completing the excavation of these 8 x 8 m trenches (work in Trench 39 had begun in 2006, and in Trench 49 in 2007). Of these, nine were “a cappuccina” in which the skeleton was covered with tiles set obliquely in the form a roof; and two were covered with tiles set horizontally, and equipped with a pipe for libations. In eleven cases the body was laid in the normal supine position. In the remaining burial the individual was laid semi-flexed in a simple pit. Ten of the individuals were adults, the others were a child and an infant. Some of the skeletons were poorly preserved, especially in cases where the burial structure had been damaged, but others were in better condition and will provide good material for the ongoing osteological analysis. All of the burials contained some grave goods, and six were particularly well equipped with utensils of bronze and iron. Until the restoration and study of the artefacts has been completed only a provisional date for the burials can be given. They seem likely, however, to be broadly contemporary with those already excavated in this area, i.e. predominantly of the 2nd and early 3rd centuries AD. A small excavation was also carried out near the SW corner of a large orthogonally planned complex of buildings in the north part of the site, with the aim of checking some details of the plan derived from a resistivity survey carried out in 2006 by John Hunt. It confirmed the presence of two of the walls of the 2nd century AD, and also uncovered evidence of reuse of this part of the site in the Late Antique period in the form of a midden deposit which contained numerous animal bones and fragments of pottery. Overlying this there were traces of the foundations of a building of the 5th or 6th century AD. The final report on that phase of the excavations is in preparation.

_The settlement_ The excavation of an imposing structure identified by a resistivity survey in the north part of the site was begun in 2008 and continued in 2009 (Fig. 1). _Period 3A (end of the 1st century AD)_ The first phase of construction is represented by a wall (F242) made of blocks of stone bonded in a sandy mortar. In construction technique it resembles other walls, already excavated, which can be dated to the beginning of Period 3. The fact that it terminates without walls attached to it at right angles shows that the building to which it belonged was open to the south-west. _Period 3B(i) (2nd century AD)_ Two walls, F241 and F256, which join at a right angle belong to the principal building being investigated. They were made of small stones bonded in a hard white mortar, and measure ca. 0.7 m in thickness - the thickest walls so far excavated at Vagnari. F241 corresponds to the south-east wall of the rectangular building identified by the resistivity survey, from which it can be inferred that the eastern corner of the building was situated not far outside the limit of the trench. The external walking surface was formed by the natural clay. Inside, however, the floor layer must have been much lower, because a trial trench 1.0 m wide was excavated in the fill of the room to a depth of 0.7 m without arriving at the bottom. The fill contained much demolition rubble mingled with domestic and industrial rubbish. _Period 3B(ii) (early 3rd century AD)_ The wall F236, which seems from the resistivity survey to have formed the south-west boundary of the site, belongs to this phase. It was constructed of rough stones loosely bonded in clay - a technique of construction already noted at Vagnari in some walls of Period 3B (late 2nd or early 3rd century AD). The wall was cut by the construction of the structure F238 in Period 5. _Period 4 (2nd half 4th century - end 5th century AD)_ After the construction of wall F236 a layer (308) of greyish brown silty clay was deposited, containing destruction material missed with more domestic refuse including animal bones, charcoal and pottery. The layer merges with another (F239) consisting of pot sherds, small stones, and numerous tile fragments which formed a walking surface laid down over the remains of the destroyed wall F241. A limited excavation inside this layer yielded some fragments of Late Roman Painted ware datable after the middle of the 4th century AD. A large piece of iron slag found inside the floor F239 suggests that there was a smithy in the vicinity. _Period 5 (?6th century AD?)_ The last phase of construction is indicated by three tegula fragments (F238) bonded in mortar with their flanges facing upwards, and aligned north/south.They seem to represent the lowest layer of a wall of the beginning of the medieval period which has been almost completely destroyed by ploughing. _The study of the material_ The material recovered in the excavation will be studied in detail in the spring of 2010. (Alastair Small) _The cemetery_ Excavation of the Vagnari cemetery continued in 2009 with the excavation of a 6m x 8m trench (no. 59) immediately South of trench 39 (excavated in the 2008 field season). Nine tombs were exposed, but only 6 were fully excavated due to time constraints. Another concentration of tiles was uncovered in the eastern part of the trench, but requires further excavation to establish that this is a tomb. The orientation of the tombs was variable, with the majority in an East-West direction. All of the exposed tombs were ‘a cappuccina’ (with the deceased covered with large tiles set obliquely in the ground to form an angled roof), except for one pit burial (F253) that had no apparent burial structure. The burial is atypical for this cemetery, as the individual was buried in a semi-flexed position with grave goods at the knees and feet. The majority of the depositions were in a normal supine position with the arms at the sides of the body or across the pelvis. Five of the burials contained adults and the sixth contained the poorly preserved remains of an infant (F251). Four of the six tombs contained grave goods, typically located at the feet of the deceased, with the exception of a number of metal objects and items of personal adornment. As in previous years, most of the ceramic vessels found in the tombs were broken and/or fragmented. One tomb of a young adult female (F245) contained a large number of objects including 3 metal rings found on the left hand, a ‘Butrint’ casserole dish (from Albania), another casserole dish in African Terra Sigillata, an oil lamp, a bone needle, and a plain ware vase. The infant burial (F251) contained a glass bead, but no other grave goods. Unlike previous years, no bent iron nails were recovered in any of the excavated burials. Until restoration and study of the artifacts can been completed, only a provisional date can be provided. These burials are likely contemporaneous with tombs excavated in previous seasons, that is, predominantly of the 2nd and 3rd centuries AD. Ongoing analysis of the skeletal remains and grave goods is currently underway. (Tracy Prowse)

In 2010 excavation was resumed in the large partly subterranean structure identified in the north part of the site by resistivity survey and already partly explored in 2008 and 2009. The purpose of the excavation in 2010 was to determine the full extent and depth of the building, and to clarify its chronology and its relationship to other structures in the area. In conjunction with the excavation an additional geophysical survey was undertaken by Nicola Masini and Raffaele Persico of CNR-IBAM at Potenza to augment the results of the previous surveying techniques. Four trenches were opened (Trenches 41-44) at targeted locations. In all four, large portions of the perimeter walls were revealed which made it possible to locate the four corners of the structure. In Trench 42 (which was effectively an extension of Trench 38 excavated in 2008 in the south quadrant of the building) the floor was encountered at a depth of ca. 1.55 m from the surviving top of the wall, and about 1.00 m below the ground level of the time. It consisted of natural yellow clay. The flanking walls were coated internally with a layer of hard, probably impermeable, plaster. When the building was abandoned, the walls were destroyed above ground level, and the subterranean part was filled with layers of destruction rubble mixed with organic material, including animal bones and carbon. Among the rubble infill were two small fragments of funerary inscriptions which must have been brought in as filling material from elsewhere on the site. _The purpose of the structure_ The external measurements of the building are as follows: length, NW side: 17.25; length, SE side: 15.65; width, SW side: 11.0; width, NE. side: 10.85. The walls range between 50 and 70 cm thick. The size of the building, the strength of its walls, its semi-subterranean construction, and its impermeable floor confirm the hypothesis, advanced in 2009, that the structure was a cistern. Its primary function is likely to have been to store water for use in a bath building situated lower down the slope. No remains of it have yet been excavated, but a scatter of hypocaust tiles collected by Carola Small’s survey team in an area ca. 20 m to the west suggests that a bath suite was located in this area. The cistern must have been served by an aqueduct which is likely to have brought water from a source in the vicinity of the Masseria Vagnari, about 400 m to the east. _The date of the building_ The building was probably constructed in the last part of the 2nd century AD (Period 3B). Its destruction can be dated by the latest material in the layers of fill which includes several fragments of African red slip of Hayes’ Form 50, with a maximum date range of ca. AD 230-400+. There are also fragments of burnished cooking-pot wares of the 4th/5th century. This material is still being studied, but a provisional analysis suggests a date near the end of the 4th century AD for the destruction.. _A later building_ Not long after the destruction of the cistern, a new structure was built over the remains of it, made of roughly cut or uncut stones bonded in clay. It appears to have consisted of a single row of rooms opening onto a yard, and was possibly used as a stall for livestock. It can be dated to the early 5th century (Period 4) and appears to have remained in use for only a short time.

Excavation of the Vagnari cemetery resumed in July 2011 under the direction of Tracy Prowse (McMaster University), with a team of archaeology students from Canada and Italy. Excavation of the cemetery commenced in 2002, with the majority of burials dating between the 1st to 3rd centuries AD, plus a small number from the 4th century AD. The goal of the 2011 field season was to complete excavation of Trench 59, initially opened in 2009, and to expose a new trench (69) immediately to the West of an area excavated in the 2007- 2008 field seasons, bringing the total area excavated to-date to approximately 400m2. A total of 19 tombs were uncovered during the 2011 field season; 6 in Trench 59 and 13 in Trench 69. Twelve tombs were fully excavated (6 in each trench), with 7 reburied for excavation in subsequent seasons. All of the burials were aligned in an East-West or an SE-NW direction. The majority of the tombs (n=10) were ‘a cappuccina’, with the deceased buried in a shallow pit and covered by a series of large tegulae in an inverted ‘V’ shape. Of the two remaining burials, one was a libation burial in which a child was interred with grave goods, covered with horizontally-laid tegulae, and an imbrex was inserted vertically in the soil above the burial for libations. The final burial excavated in 2011 was a poorly preserved skeleton found with no evidence of a tomb structure, a few ceramic fragments in association with the skeletal remains, and was likely a simple pit burial. All of the burials contained grave goods, typically deposited near the lower legs and feet of the deceased, with the exception of a small number of metallic objects and objects of personal adornment (e.g., bone pins, beads). Similar to previous years, grave goods consisted of plain ware pottery, lamps, iron nails (often bent), and a small number of bronze coins (n=3). When iron nails are found in the graves, they are typically bent and placed inside a ceramic vessel. Detailed analysis of the grave goods is in progress and will help to determine the dates of the tombs excavated in 2011. To-date, 88 burials have been located within the excavated area. The ongoing bioarchaeological analysis of the human remains within the burials provides an opportunity to investigate evidence of health, diet, social structure, and migration in a rural Roman population. Excavation will continue in the cemetery with the goals of understanding the extent of the cemetery, its chronological distribution, patterns in burial practices, and the quality of life in a rural Roman population.

VAGNARI - _Vicus_ The University of Sheffield began a new fieldwork project in the _vicus_ at Vagnari in July 2012 under the direction of Maureen Carroll. The excavation concentrated on a stone-built structure revealed by resistivity survey on the northern edge of the _vicus_. This large building appeared on the resistivity plan to have a long range of cell-like rooms of uniform size on its northern side, however the excavation revealed that, in reality, this wing consisted of rooms and corridors of varying sizes. Some of the ‘walls’ in the older survey proved to be well-preserved drains that clearly carried considerable quantities of waste and/or water from an unknown structure up-slope whose character and function remain to be clarified. The pottery and coins suggest that the structure was built in the 1st or 2nd century AD and remained in use until it was dismantled or abandoned in the latter part of the fourth century. The excavation shed important new light on the economy of the estate. Substantial evidence was retrieved for metal working with lead, iron and bronze and for glass production. The lead objects, especially, consisted of roughly torn and cut pieces taken from other things such as pipes, vessels and tools, suggesting that they were being re-cycled, processed and re-worked. An additional find of some significance was a small hearth containing a large quantity of charred plant material. The composition of this material suggests that macaroni wheat was grown here on the estate as a cereal crop. Fieldwork in 2013 will focus on the exploration of the new evidence for industrial activity and domestic habitation, including the excavation of a possible kiln. The connection between industrial production and domestic habitation is of significant interest and importance in understanding the socio-economic complexities of living and working on this imperial estate. VAGNARI – Cemetery The cemetery is located on the South side of the site, across a small ravine from the settlement area where the University of Sheffield team was excavating. Excavations of the cemetery continued in 2012 under the direction of Tracy Prowse (McMaster University). Two trenches previously exposed in 2011 (59 and 69) were reopened to excavate 6 remaining burials, and a third new trench (79) was opened immediately to the West of Trench 69. The number and density of burials was lower in Trench 79, possibly indicating the western limit of the cemetery. A total of 10 tombs were excavated in the three trenches this year, bringing the total number of burials excavated to-date in the Vagnari cemetery to 98. Seven of the tombs excavated this year were ‘a cappuccina’, consisting of the deceased buried in a shallow grave and covered by a series of tegulae in an inverted ‘V’ shape. Two tile burials contained the remains of infants, and the final burial was a simple pit without any burial covering. The grave goods recovered from the burials are similar to those found in previous years, consisting mainly of ceramic vessels (sometimes broken), iron nails, small glass vessels, and objects of personal adornment. Some of the ceramic vessels appear to have been intentionally broken prior to deposition, while others were found intact. Only one of the burials excavated in 2012 contained a bronze coin. This burial, F294, is noteworthy for the relatively large number of grave goods found with the skeleton, including: the previously mentioned bronze coin, a stamped oil lamp (“AVFRRON”), two glass vessels, two ceramic vessels, and hobnails deposited around the feet indicating that the deceased was buried with footwear. The lamp dates to the 2nd century AD, which is consistent with the dates for the majority of tombs excavated in the cemetery. A smaller number of burials in the cemetery date to the 3rd and 4th centuries AD. Ongoing bioarchaeological analysis of the skeletal material recovered from the cemetery is investigating diet, health, and disease in this rural Roman skeletal sample.

Excavations begun in 2012 in the vicus at Vagnari continued in the summer of 2013. Work concentrated on a stone-built structure (North Building) on the northern edge of the settlement, a tile-roofed building almost 30 m long consisting of a series of rooms and corridors with plastered and painted walls of daub and with floors of beaten earth or mortar. The pottery and coins suggest that this building was in use from the late 1st to the mid-4th c. A.D., although residual pottery and coins of the 3rd to 1st c. B.C. suggest that the imperial vicus was not the first settlement on the site. The excavation shed important new light on the economy of the estate. Substantial evidence has been retrieved for metal working with lead, iron, and bronze and for tile production. Additional significant finds are charred plant material from several contexts inside and outside the building that points to the cultivation of free-threshing bread wheat and durham wheat as cereal crops on the estate. The North Building may have had multiple functions and served as dwellings for slave and/or free labour, as well as for the storage of goods and for activities associated with manufacturing and processing. Fieldwork in 2015 will focus on the remains of industrial activity and domestic habitation in this structure and an adjacent southern building at the summit of the hill. The connection between industrial production and domestic habitation is of significant interest and importance in understanding the socio-economic complexities of living and working on an imperial estate. Excavations in the Vagnari cemetery have been underway since 2002 and continued in the summer of 2013. A 5m (East-West) x 16m (North-South) trench (89) was opened to the East of trenches excavated during previous field seasons. Another trench (59), originally excavated in 2009, was reopened to complete the excavation of three tombs. A total of 10 tombs were excavated in 2013, consisting of 9 inhumations and one cremation burial. Three additional tombs were identified along the eastern baulk of Trench 89, but were not excavated due to time constraints. All but one of the graves contained modest grave goods, similar in quality and quantity to items recovered in previous years, with the notable exception of a large number of grave goods found in the cremation burial. Cremation burials are relatively infrequent in this cemetery, with only 3 uncovered to-date out of a total of 108 excavated burials. All 3 cremations date to the same time period as inhumation burials in the cemetery (2nd century AD), but the greater number and grander quality of grave goods found in these cremations suggests that wealthier individuals at Vagnari may have opted for the practice of cremation. Preliminary osteological analysis indicates that 5 adults, 3 children, and 2 infants (i.e., less than 1 year) were buried in this part of the cemetery. Ongoing bioarchaeological research at this site is investigating geographic origins, diet, and health of the people living in this rural Roman settlement.

Work concentrated on a hitherto unexplored area adjacent to the substantial stone-built structure (North Building) excavated in the previous seasons on the northern edge of the settlement. The excavation revealed exciting new evidence for viniculture, consisting of a paved area in which three large pitch-lined ceramic vats (dolia defossa) were sunk into the floor. It is likely that wine was stored and fermented in these vessels, and that this room was a cella vinaria. Only a small fraction of this room could be explored, but it will have been much more extensive, almost certainly with rows of fixed dolia. Associated facilities, such as a pressing room, a treading basin, or a must container, were not found in the trenches opened in 2015, but the area will be expanded in 2016 to locate them and the complete range of dolia. In another part of the building, outside the cella vinaria, finds such as pieces of marble wall or floor cladding, panes of window glass, a wide range of pottery, and bone implements, shed light on the appointment of a part of the building and associated domestic activity. The project is shedding light on the diversity of the economy of the estate and the role of the vicus and its inhabitants in organising and managing work and income for the emperor. VAGNARI – Cemetery After a study season in 2014, excavations in the Vagnari cemetery resumed in the summer of 2015 with a team of students and supervisors from Canada and the United States. Two trenches were opened to the South (Trench 99) and East (Trench 109) of previous excavations. Trench 99 revealed a surprisingly small number of burials (n=3) given the large dimensions of the area excavated (16m E-W by 8m N-S), which suggests that we may have found the SE edge of the cemetery. The smaller Trench 109 (5m E-W by 9m N-S) to the NE of Trench 109 contained a total of 9 burials, 5 of which were excavated this year. All of the burials were cappuccina tombs, with the exception of one pit burial of an adult male. Most burials contained one interment with a modest number of grave goods located around the feet, as in previous years, with the exception of the pit burial that contained no grave goods. Two burials (F308 and F314) contained the remains of multiple individuals. F308 was associated with the remains of 3 individuals. The poorly preserved skeletal remains of a subadult (3-5 years) were found outside the grave at the SW end. Inside the burial, two adult individuals were discovered with the earlier one showing evidence of disturbance and redeposition with the later burial stratified over top. There were some disturbed tiles located to the North of F308, which may be part of the earlier tomb. It is not clear if the disturbance was unintentional (i.e., the earlier grave was not visible and was accidentally disturbed) or deliberate, possibly indicating a relationship between the two individuals and the intentional burial of the two in the same grave.

VAGNARI – Vicus Excavations conducted since 2012 in the vicus at Vagnari continued in the summer of 2016. Work concentrated on the eastern side of a substantial stone-built structure partially excavated in 2015 in which a cella vinaria was preserved. This wine storage area did not continue further east, as we expected, but it is now clear that the room housing the large pitch-lined ceramic wine vats (dolia defossa) had a north-south orientation, and it is to the north and south of the area excavated in 2015 that we expect to clarify the extent of this facility. In 2016, the physical evidence retrieved allowed us to fill important gaps in the occupation history at Vagnari, both for the period prior to the establishment of the imperial estate and for the earliest period of Roman settlement. The most exciting discovery was a building that predates any structure found thus far in the vicus. The pottery, loom weights, ceramic oil lamps, and iron implements that were retrieved from a circular storage pit in this building can be dated to the Hellenistic period, to the 2nd century B.C. at the latest. This building continued in use in the early first century A.D. when it was adapted and enlarged, precisely at the time in which the site became an imperial property. This must have been a relatively high-status structure, with floor or dado coverings of white and grey marble slabs. Also several large panes of window glass retrieved here suggest a well-appointed structure. Some of the walls of this early Roman building were preserved, although in places they had been robbed out. On top of one of them, a coin of Vespasian was found, indicating that the building was dismantled after A.D. 70. The building was replaced with another structure on the same alignment, of which well- preserved walls and floors, including a cobblestone floor, remained. This building was occupied well into the third and possibly even the fourth century A.D. The tile roof covering two of its rooms had collapsed onto the beaten earth floors, without being retrieved in antiquity, and a coin of the early fourth century lying immediately on top of a robbed-out wall here is a useful piece of evidence to date the abandonment of the building. Although only a part of the Hellenistic and the early Roman buildings could be excavated due to time constraints, the structural and in situ artefactual evidence demonstrates beyond doubt that there was a predecessor settlement here, perhaps a villa, which was taken over and adapted when the imperial estate was created in the early first century A.D. The region clearly had not been depopulated or empty after the Roman conquest in the third century B.C., although it remains to be clarified who, in fact, built and inhabited the Hellenistic settlement. VAGNARI – Cemetery Excavations in the cemetery at Vagnari continued in the summer of 2016. We reopened Trench 109 from the 2015 excavations in the eastern part of the cemetery to expose 4 burials that were not fully excavated during that field season. We also extended this trench to the North by an additional 6m, creating a trench that was 11m N-S by 6m E-W in size. We uncovered 4 new burials in this northern extension, three of which were excavated this year. All but one of the tomb structures in Trench 109 were cappuccina burials. The exception was one adult individual (F331) who was interred in a simple pit with no evidence of a burial structure. We also opened up a new trench to the North of Trench 109 and Trench 89 (from 2013), measuring 10m E-W by 6m N-S. Five burials were identified in this trench (Trench 119) and 3 of these were excavated. All 3 were the characteristic cappuccina burial structures. Nine of the 10 burials excavated in 2016 were inhumations, with the deceased typically buried in an extended supine position with grave goods deposited around the feet and legs. The grave goods found in 2016 are similar to those found in other burials from the cemetery, including objects that were likely used by the deceased during life (e.g., hobnails, iron tools, lamps, and pottery vessels). One burial in Trench 119 (F325) was a cappuccina burial containing an in situ cremation. This is only the fourth cremation burial found in the cemetery to-date, and similar to the previous 3 cremation burials this one contained a large number of high-quality grave goods. Preliminary osteological analysis indicates that 6 adults, 2 children, and 1 neonate were buried in this part of the cemetery. The cremation burial is also likely that of an adult, but age and sex could not be determined. Ongoing bioarchaeological research at this site is investigating geographic origins, diet, and health of the people living in this rural Roman settlement.

VAGNARI – Vicus Excavations continued in the vicus at Vagnari in the summer of 2017, the sixth season of research here by the University of Sheffield. Work concentrated on the eastern side of a building partially excavated in 2015 in which a Roman winery was preserved. In 2016 in this part of the site, circular storage pits filled with pottery and other artefacts of the 2nd century B.C. were found which suggested that there had been a settlement at Vagnari predating the Roman imperial estate. To investigate this more thoroughly, work in 2017 concentrated on the investigation of an adjacent unexcavated sector of the vicus and on the re-excavation of the deepest deposits and features in three small trenches that had not been completely explored in previous seasons. The results of this work are outlined in the following. The existence of a settlement in the 2nd to the mid-1st centuries B.C. is attested by storage pits filled with refuse and by pottery found in various locations immediately above the natural subsoil. This settlement, possibly in private hands as part of the senatorial elite’s appropriation of Roman state land following the conquest of Apulia in the 3rd century B.C., predates the establishment of the vicus of the imperial estate. In the next phase of the site, new buildings with very well-built walls and drains were erected, and in them were several storage pits. The ceramics confirm that this second phase dates to the early 1st century A.D., and its material culture may indicate a new influx of settlers. Therefore, two distinct phases of settlement activity are clear, with a clear “break” in occupation and a change in ceramic assemblages between the middle of the 1st century B.C., when the pre-imperial settlement came to an end, and renewed building activity and occupation in the early 1st century A.D., when the Roman emperor acquired the territory as a revenue base and established the imperial vicus. When a winery was added to the vicus, the basins cut into the ground for the wine dolia sliced through two drains of the first half of the 1st century A.D. and made them obsolete. The winery, therefore, appears to date to the late 1st century or 2nd century A.D. It is not yet clear precisely how long it was in use, but in the fourth century some of the dolia were either completely removed or smashed, leaving only the base or the bottom third of the vessel in place. Fragments of pottery of 5th-century date in the fill of a robber trench of a wall indicate clearly that vicus structures were being quarried for their stone and building materials, almost certainly for the new, smaller village that was established at this time on the other side of the ravine, as excavations by the University of Foggia in 2001/2002 indicated. Our work in 2017, therefore, has shed light not only on the earliest phases of occupation at Vagnari, but also on the end of the settlement and the salvaging of materials from it. VAGNARI – Cemetery The 2017 field season in the Vagnari cemetery involved opening a new trench (10m E-W by 6m N-S) immediately to the West of the 2016 excavations, further exposing the northern extent of the cemetery. We also reopened a trench in the southern area of the cemetery that was initially excavated in 2003 by Hans vanderLeest (Mount Allison University, Canada). This southern trench was originally opened to investigate a strong anomaly indicated by geophysical survey, but excavation revealed the presence of additional cappuccina burials, so work in the area was halted and no burials were excavated. The 2017 trench was 4m (E-W) by 5m (N-S) and three burials were exposed, with the participation of a team of high school students from the community of Gravina in Puglia. We identified a total of 16 tombs and excavated ten of them in the 2017 field season. As in previous years, most of the tombs excavated were ‘alla cappuccina’, containing inhumations of both children and adults with the individual in a supine extendend position and grave goods deposited around the legs and feet. Two burials had flat tile covers reinforced with large quantities of rock and mortar. Both these burials also had libation tubes (constructed out of curved imbrices) inserted vertically into the burial and held in place by the surrounding rocks and mortar. Of particular interest was the presence of a flat, circular, perforated piece of lead covering the libation tube of one of the burials (F333), which is the first to be recovered from this cemetery (Fig. 2). Three of the burials excavated this season appear to have been disturbed in antiquity, including one burial (F342) that is unique for the cemetery as it is the first burial with a ‘cassa in laterizio’ structure inside the disturbed cappuccina tomb structure. Fortunately, a bronze alloy coin was recovered from an undisturbed section of the burial, so it may be possible to provide a date range for this interment. Another disturbed burial appears to have been a simple pit burial (F339) that was disturbed by the construction of a later tomb (as yet unexcavated). Captions Fig. 1 Vagnari vicus 2017. Documenting the excavated walls, drains, and pits of the early 1st century A.D. Fig 2. – Flat lead piece with perforations (SF 129-011) found on top of a libation tube in burial F333. The perforated part of the object (a sieve?) has collapsed into the libation tube.