The area designated MON, for Monastery, was excavated over the course of five years (Fig. 1). The first trenches in 2006 were located in front of the church (BI) and inside the church, in the presbytery where previous exploration had revealed marble pavements (CI) and the northwest chapel (CII). In 2007 (Fig. 2), 2008 and 2010 the trenches were enlarged, first to the west to reveal the bell tower and other parts of the cemetery, then 11 x 10 m to the north (BII), then 10 x 15.5 m to the east to excavate the area to the north of the church. Finally in 2010 an extra .50 m was added to the north to reveal the southern wall of the second-century building which emerged near the northern section. The final size of the trench was 435 m2. The trench was aligned to the church, pegged out with a Total Station and marked with 5 m grid points for surveying. The 0.0 of the entire area is the threshold of the actual portal of the church, 206.704 asl.

Paved courtyard, arcade and exterior wall (the second century)

The earliest phase of occupation that we have examined in this area is the large courtyard paved with smoothed limestone pavers, the east–west wall in opus mixtum faced concrete and the preparation for a portico running parallel to it (Fig. 3). The pavement [4289]=[4239]=[4240]=[4492]=[4537]=[4538]=[2063]=[2209]=[2324]=[2326]=[2931]=[3539]=3805=[3853]=[3869]=[3870]=[3871]=[3872] is visible over an area measuring 12.5 x 17.5 m and is consistent in its horizontal paving with white (and the occasional black) pavers measuring some 30 x 20 cm, carefully fit together to create a smooth surface. The paving stones are set into a bed of dark yellow, very clean clay (4447), part of the original construction of the entire villa and very similar to the packing clay found elsewhere on site.¹ No materials were found in it.

At the northern limit of the trench is found a wall, [4574]=[4321]=[4460] visible for some 8.40 m, passing beyond the limits of excavation to the north (Fig. 4). The alignment of this wall is about 20° off of the alignment of the church and medieval structures, and thus the edges of the trench. The southern face of the wall is coated with 6 cm-thick mortar, preserving the impressions of revetment panels. The remains of a small bronze tang were preserved in the eastern part of the wall [4460], suggesting an original white marble revetment, one element of which remains against the base of the wall where it had slipped down from further up. The wall has a door 1.45 m wide in the western part and the doorframe is quoined in alternating brick and limestone blocks, pegged into panels of opus reticulatum; this is characteristic of the opus mixtum technique at Villa Magna in the second century. The threshold of the door [4576], at -0.60 m, is smooth concrete over a brick course, sloping up to the north by some 12 cm, which suggests that the threshold held a marble step. The yellow clay and blocks of concrete which cover it (4573) may be the packing for a later riser or paving stone. At some point a step may have been added into [4576], which was later removed 4580 and filled with silt (4579). The area to the north of the external wall of the building, that is the area inside the building itself, is visible to us only in a small section measuring 1.65 by 2.25 in the northern extension of the trench near the edge of the much later cloister. Inside this building, a smooth plaster pavement 4539 bonded to the base of the wall [4460]. This was later covered by a clay preparation (4510) for a cocciopesto pavement 4505, which slopes noticeably towards the southeast, though this may not have been for drainage but the result of settling. These floors are visible through a later cut, 4511.

The façade of the building is separated from the basoli paving by a surface 5.10 m wide, limited on the south side by a line of three pier bases, also surrounded by the dark yellow packing clay.² No foundations have been found between them and the central preserved section has a clear face on its west side, indicating that the arcade above it would not have been supported by a continuous foundation. The line of pier foundations runs west to east; the western one [4424] is 1.10 m long by .60 m at its widest; 2.05 m to its east is an upright brick plinth [4422], also 1.10 m long and enlarged by the addition of .60 m long section [4601], and 5.90 to the east is the brick and mortar base of a plinth [4526], which has a central brick/tile element 1.05 m long, probably used to cap the foundation before a plinth was added on top of it. This has no visible lateral faces, and has mortary extensions to the west and the east. Another foundation may have existed between these last two. These were probably the base of an arcade, which may have replaced an earlier structure, perhaps a portico. Their rectangular form and later extensions resemble the bases of such arcades throughout the Roman world; the Hadrianic-era arcades on the cardo maximus of Ostia, to give but one example, were brick piers in front of a paved walkway.³

The pavement of the area between the façade and the pier bases is not preserved, but a beaten earth and clay surface (4540) with small sporadic areas of reddening from exposure to heat or flame remains in the eastern section; at a later time a layer of hard yellow clay (4528)=(4521) was used as a working surface, slightly darker in colour than the earlier surface but also with small patches of badly preserved mortar with charcoal inclusions. The level of this lower, earlier surface is a few centimetres lower than the pier bases preserved to the south of it (which would have been even higher with the addition of the pavement and plinth above it), suggesting that the bases as they are presently preserved may be rebuilt from an earlier plan. The white paving stones around [4526] are slightly disturbed, perhaps from such a rebuilding process (Figs. 5, 6 and 7) The mortar and tile layers, (4543) and (4554), were added to the eastern part of the portico over the earlier surfaces as preparation for a marble or stone pavement. In the western section of the portico the earliest surface preserved is (4556), which was occupied at a date later than its actual creation, discussed below. The preparations and marble surface and the original portico were probably stripped and removed at some point in antiquity, when they were replaced with arcading, later buttressed by supporting walls which would have narrowed the arched or colonnaded openings. Later events disturbed the surfaces, and levels across the remains of the portico do not correspond closely.

Though no structures from this phase are visible in their entirety here, those that are visible are clearly part of the original layout of the second-century villa. The northernmost visible wall appears to be the southern façade wall of the large rectangular structure revealed by the magnetometry. The northern half of this consisted of a large portico—duplex or even triplex—while the southern half remains obscured by the buildings of the later monastery and castrum, and remains unexcavated. We interpret the structure as the imperial residence. It was probably originally faced with a portico, even if the one we see in the trench is not original to the second century, and faced a large open court, which seems to have been a nodal point in the topography of the villa, the main entrance from the east and the separation between public and private areas.⁴

Chronology

The clay packing which underlies all of the construction at this site parallels very closely the clay packing around the original construction of the winery in the early second century. The brick-faced masonry of the Roman building façade also follows the techniques used in the winery and their chronology.⁵

Storage in the portico (sixth century)

In the sixth century the function of the portico underwent a radical transformation. In the western end, after the removal of the floor surfaces of the portico and of the revetment from the external wall, a wall [4420]=[4246] was constructed, perpendicular to the façade, crossing the width of the portico. The wall is constructed in opus listatum, with a course of brick alternating with two courses of limestone blocks and very hard mortar, closely paralleling the opus listatum phase of the church and the narthex which must be near contemporaries. In order to build it the clay preparation of the portico (4556) was cut, 4586/(4585), and this surface was used for a time. A posthole was cut into it, 4584/(4583), and part of the surface was burned. A later beaten earth surface to the east, 4557, has fragments of pottery and charcoal.

The perpendicular wall [4420] divided the portico into sections, that to the west became a storage area, fitted with at least nine dolia set into the clay of the portico substructures and cutting through the beaten earth occupation surface (Figs. 8, 9). In two parallel rows, large round trenches some 1.0 m deep from their shoulder to their base (-1.85 m) were cut into 4557. (The northern row from west to east: 4549, 4532, 4599; the southern row: only 4602, the easternmost pit, preserves its original cut. A further eastern pit may have existed, but it was not excavated.) There is also a much smaller circular hole 4581, measuring 80 cm in diameter, which was later filled with silt (4582); it may have originally held a ninth small dolium.

Between 4549 and 4532, a small wall [4436] was constructed as a shoulder. The base of one of the cuts, 4599, preserves the cement seating of the dolium (4600). The cement is marked with the impression of a flat bottom measuring .39 m in diameter, and .69 m in diameter further up, where it broke off at a height of .27 m from the base (Figs. 10, 11). The preserved edges of the cuts measure about 1.50 m at their greatest diameters and are slightly irregular in their shape and depth, suggesting that the dolia that were placed in them may also have been irregular in size and shape. Around the dolia was laid a floor of pale grey mortar 4446, preserved only in patches.

This appears to have been a storage or fermentation facility for large quantities of wine or must. The location of the dolia in the portico in front of what has been interpreted as the imperial residence, and adjacent to the newly rebuilt church, suggests that this was more important than a simple warehouse. The highly visible location suggests that wine storage was central to the estate, deriving from the collection of taxes or rent in kind. We may imagine that the old courtyard had become the centre of the estate, surrounded by religious, administrative and impressive storage buildings.

Destruction (Fig. 12)

The dolia were all removed, some apparently breaking in extraction as fragments of dolia were present in the backfills of some of the pits. (The cuts to remove the dolia, the northern row from west to east: 4534, 4533, 4587, 4518; the southern row from west to east: 4593, 4595, 4589, 4597. The small shoulder wall [4436] was cut down 4435. A thick layer of earth, (4446), was mostly cut away by the robbing of the dolia: this may have accumulated during the use of the storage area. The fills of these cuts, in the same order: (4515), (4508), (4588), (4519); (4594), (4596), (4514), (4598).) The backfills also contained high frequencies of roof tiles, fragments of mortared bricks and stone and fragments of pilasters. These are presumably from the portico which was razed after the removal of the dolia. A small wall of stacked rounded lime stones [4523] was built into a trench 4522 cut into the destroyed portico, though subsequently cut away so much that it is difficult to know its purpose; indeed it may be a backfilled cylindrical silo rather than a small wall.

The dolia were removed and backfilled, then covered by collapsed roof tiles (4418) which fell onto a deposit (4425) of tiles and brick laid horizontally on the west side of [4420]. The extension of this deposit of collapsed tiles (4504) is visible on the east of the later monastery wall [4005]. The tiles and brick are set into a bed of clay (4520) against the wall and have a surface turned red by fire. An animal bone recorded in the clay is vitrified.

Chronology

The pottery from these phases points to the creation of the storage area in the sixth century and their destruction and backfilling in the late sixth or seventh century.⁶

Reoccupation: sunken-floored buildings (the later sixth century)

A diagonal wall runs northwest–southeast from the northern narthex wall delimiting the area in front of the church (Fig. 13). The wall presumably abutted the facade of the imperial building, though later building activity removed its northwestern limits. The wall, in fact, has two sections: the southern one [4493] consisting in large part of reused white paving stones, was constructed on top of the paved surface of the courtyard between the northwest corner of the church and the easternmost pier of the portico; the northern most section [4461] had a tenacious, pozzolanic mortar and was constructed with limestone paving stones, roof tiles and cocciopesto chunks in the aggregate, suggesting that destroyed building materials from the portico were used in its construction. Their point of juncture is not visible because of later building activity. The building of this wall meant the cutting of a shallow trench, 4547, just as wide as the wall, through the collapsed roof tiles and the earlier pavements of the portico. The southern part of the wall sits on top of the white limestone pavement, over a thin grey sandy soil (4524) some 2–10 cm thick that washed over the pavement, and is probably a colluvial deposit.

To the west of the diagonal wall, presumably inside the area it delimited, a brown, beaten earth surface 4503 accumulated over the collapsed roof tiles and destroyed portico (Fig. 14). Outside the wall, to the east, a deposit of black-brown material (4555) began to accumulate. A thick layer of rubble in a black soil matrix (4494) and a subsequent hard silty stratum (4487) also accumulated outside the wall, perhaps earlier destruction debris and refuse from the occupation inside. While the rubble accumulated outside of the walls, within the enclosure another surface of black soil (4470) was laid down on a rubble preparation. Numerous pits—4477/(4476), 4495/(4496), 4497/(4498), 4544/(4545)—appear to have been cut into its surface. Their fills yielded little material besides rubble including a white marble cornice from the interior of a structure in (4496). This surface and its backfilled pits were covered by another beaten earth surface 4395=4355 (Fig. 15).

The enclosed area was obviously dominated by the church, but there were also less monumental structures. About 18 m to the northwest of the church façade, within the area of the old portico, was a subrectangular pit cut into the backfilled dolium pits (Fig. 16). The oblong cut 4499 follows the east–west alignment of the portico and had deep vertical sides about .60 m deep from shoulder to base (-1.45 m). It passes beyond the western limits of excavation, but what is visible in the trench measures 2.25 m wide by 3.8 m long. It was surrounded by a series of postholes: 4465/(4466), 4463/(4464), 4445/(4444), 4441/(4442), 4337/(4338). The larger postholes, 4441 and 4337, had brick fragments lining the holes as braces for posts. Another posthole 4448/(4449) was located some 2.0 m to the north, perhaps part of a related structure passing beyond the limit of excavation. Inside the cut, the bottom was lined with a yellowy clay (4529) and the edges were packed with a mortared rubble (4513) forming what we presume to be supports for a floor (Fig. 17); planks could have lain across these edge supports, allowing for circulation of air underneath. The cut was filled, and the side supports were covered, by a thick black deposit with high frequency of small mortar inclusions (4440), which is presumably all that remains of the packed earth forming the side walls and roof of the structure. In this deposit a tremissis of Justinian I (527–65) was found (C207), along with pottery of the sixth century.

Another oblong pit, 4483, with vertical sides of about the same depth as 4499, was cut through the paved courtyard in front of the church; this one was oriented north–south (Figs. 18, 19). It passes beyond the eastern limits of excavation so it is impossible to know its full extent. The width must have been slightly greater than 2.50 m and the visible extent of the pit is 2.70 m. There are two deep postholes in the western side of the cut, 4560/(4561) and 4558/(4559), one of which is lined with bricks to brace a post; these must have supported the roof of the structure. This structure lacked the clay packing of the walls, and seems to have been built differently; the large number of paving stones found within its fill suggests that these were used for the foundations of the walls. This technique of reusing paving stones in the walls relates it to the boundary wall to the east.

These two sunken-floored buildings may have been small houses, though there are no traces of hearths, silos or middens associated with them (Fig. 20). It is possible that they were storage buildings or work houses, though the absence of floor surfaces and objects associated with the use of the huts makes any kind of hypothesis about their use difficult. There are traces of another structure built on top of the paved courtyard, in the form of a series of holes cut into the pavement. A number of paving stones were removed in a straight east–west line to the east of one of the huts. The fills of two of these holes were excavated—(4551) and (4553)—revealing round holes lined with mortar 4552, 4550. Another two holes may be filled and covered by the silty patches which are very closely aligned to the east of these two, (4562) and (4563). These may have supported a row of posts for a structure which used the Roman pavement as its floor, but if that was the case we have only one surviving side of it.

Destruction and deposits

The buildings were destroyed; their pits were filled with rubble from their elevations, dark soil rich in organic materials and fragments of pottery from the sixth and seventh centuries. The area of the Roman paving was definitively covered over by a thin deposit of sandy clay and rubble (4170), probably representing the abandonment of the area. Subsequently, this deposit was covered by a dark thick layer of black earth (4330) (Fig. 21). The latter covered nearly the entire area of BII west of the line of the church façade to a depth of ca. 40 cm. This layer was consistently black in colour, with some charcoal inclusions and several fragments of metal and pottery, including fragments of globular amphorae dating to the early ninth century.⁷ There was little evidence of kitchen rubbish, or of burned surfaces, suggesting that this was earth brought to the site and deliberately deposited, perhaps for cultivation. Two round cuts 4339/(4340) and 4337/(4338) might be interpreted as holes for plantings. In the northwest of the trench, 4330 is covered by a brown-red layer of rubble and building materials (4326) in the far west edge of the trench. Also in the west, a short low southwest–northeast wall [4430], of very casual construction, was built for an unknown purpose. Near where the later wall [4005] was to be constructed, (4330) covers a thick grey deposit, mixed with residual Roman pottery and rubble, (4450)=(4516).

Contemporary Graves

Visible in the northern baulk of the trench are a series of tombs.

T372
An adult 4311*/(4310)/4312 lies on tile slabs which formed the bottom of the tomb, [4577].⁸

T392, T373 and T375
Beneath that tomb, another adult individual 4575*/(4592)/4591 lay on tile slabs [4590], T392.⁹ Bones 4323* were found further to the west of these (4324)/4325, identified as T373. T392 covers T375, a 30- to 40-year-old female 4345*/(4346)/4347 covers a deposit of fragmentary human bones (4313) on top of tiles (4314).

T384
Over the backfilled southern hut, an earthen grave 4439 contained a 35- to 45-year-old male in west–east orientation 4437*/(4438). This grave—its original limits and those of the skeleton within it much diminished by subsequent cut 4126—was made of three stones of irregular shape and size (among these a fragment of a column of cipollino marble), which were loosely bonded to the foundation of north–south wall [4237].

Chronology

The sunken-floored buildings were destroyed and backfilled with ceramic material and coins dating from the sixth and seventh centuries.¹⁰

Large and thick covering deposits formed in the early middle ages. Most of these have residual pottery of the late Roman II phases. The pottery of (4330) is markedly different than anything that came earlier in the stratigraphic sequence, because it includes Forum Ware datable to the first half of the ninth century and globular amphorae of the eighth to early ninth.¹¹ This would suggest that the large black deposit which covered the entire western part of the trench was created in the ninth century.

Early monastery: plan of complex with monastery

A major wall [4005] running north–south was constructed parallel to the façade of the church and then turned east at a right angle [2033] to meet the cemetery wall of constructed tombs in trench BI (Figs. 22, 23). The masonry wall had a foundation trench 4104/(4105), which to the west was numbered 4506=4527=4451/(4507)=(4452). Spreads of white mortar to the east, 4113 and 4106, with a burnt patch of charcoal (4103), appear related to its construction, as may a circular intrusion 4109/(4110) about 80 cm deep. The wall was built with limestone and tartara rubble, along with reused Roman marble fragments, laid uncoursed and bonded with a hard blue mortar and measuring .50 cm thick. The full height is not preserved and the length passes beyond the limits of excavation to the north. On the western face of [4005] there are two buttresses, [4016] and [4014], which indicate that this was an enclosure wall, bounding a space to the east. [4125] probably delimited buildings to the north, within the monastery proper. The space formed by [4005], [4125] and the cemetery to the south was used periodically for dumping and burials, seemingly an extension of the cemetery in front of the church. In this area a number of silty deposits formed: a brown-black deposit (4257), covering a yellow clay deposit (4247) and a purple-brown deposit (4248), possibly related to the building of the wall.

15.65 m to the east of [4005] runs a parallel wall of a similar construction technique, defining the other side of the enclosed area (Fig. 24). This wall, [4189], is built of limestone rubble of large and medium dimensions bonded with a thick hard mortar, though the materials are laid in more regular courses than in [4005]. It is about 50 cm thick, suggesting that it formed a counterpart to the western wall. It bonds with [4221]—an eastern spur that extends beyond the limits of excavation. Given the construction and composition of [4221], which is very like that of [4189], it is likely that these were part of a single building campaign or phase. [4189] is abutted on its western face by [4220], a spur added later to the complex. In the northern part of bounded area is [4188], a wall fragment with the same alignment as [4189] that extends into the northern limit of excavation. The relationship between these two walls is obscured by (4222), the rubble collapse of [4188]. The position and orientation of [4188] suggests that it was built on top of [4189], making use of the earlier wall as a foundation. Two robber trenches—4243/(4244) to the east of [4188] and 4241/(4242) to the west—are all that remain of a northern cross wall perpendicular to [4188], which extended to the east and the west.

At this time, the opus mixtum external wall of the Roman building, [4574]=[4321]=[4576]=[4460], was razed 4309=4525=4578, to a height of .90 m (about +0.02 m), and the perpendicular wall [4420], too, was cut down 4419. Whatever might have been preserved of the surfaces inside the Roman building have been compromised by later building activity. To the north of the wall [4574]=[4321]=[4576]=[4460] are two surfaces which postdate the razing of the wall and the construction of [4005]. The earliest is a light brown compact earth surface, 4373, discussed above. Over this was a thin, friable, blue-grey pavement, 4304. The pavement 4304 was cut by 4302, which also cut the wall [4005], removing stones from the wall’s elevation. It was backfilled by loose earth and rubble (4303). These surfaces must be related to the occupation of the monastic area, where the razed wall would have appeared as a small step. To the south of it a beaten earth surface 4395 covered the paving stones. Into this square cut 4398 was made to insert a hearth consisting of a layer of tiles (4399) covered by successive layers of clay and tiles, interspersed with burned patches (Figs. 25, 26) .The construction technique of the hearth is very similar to that seen in area FI¹²: layers of tiles alternate with layers of a dense, gummy yellow clay. Another square intrusion 4396/(4397) shows no sign of burning. 4395 was subsequently covered by a second beaten earth surface, 4114, which preserved the hearth for continued use.

In the eastern area of the enclosure, the latest floor is a mortared surface, 4410=4411. This stratum was a clayey compacted loam in with a fairly regular surface, sloping slightly towards the northeast. It preserved a hard mortared skin in several places and fragments of human bone were very apparent in the surface, suggesting that tombs were disturbed in its construction. The surface may the extension and equivalent of 4317 or the earlier 4373 to the west, but these are separated by the later cloister buildings (Fig. 27).

The enclosed precinct was then used for burials. A series of tombs were cut into 4395 and the clay packing of the Roman structures, (4447).

T377
Near the northern narthex wall, a grave, 4402, was cut with a west–east orientation and lined with a reused cipollino slab [4485] and a carved limestone slab, a piece of liturgical furniture of the early middle ages [4484]. In this was buried a 30- to 40-year-old male 4400*, which was disturbed by a later reduction of two individuals of interminable sex, one aged 15 to 20 years, 4401/4393*, then a 20- to 30-year-old male 4372/4370*/(4371) and a reduction of bones 4344/(4365), then an adult of indeterminable age and sex, 4366*. At a later date, perhaps at the same time as the other built graves in the later cloister, discussed below, this tomb was reconstructed (T374), the grave was recut 4344 and lined with an upright paver [4381]/(4385), slabs [4382], [4383] and [4386], and then a 30- to 40-year-old male 4342*/(4343) was buried and covered with stone and brick slabs (4341). Carbon dating of this individual puts his death at 1020–1160 cal. AD (95% probability; OxA-24973). Probably contemporary with the original marble-lined grave, nine other graves were cut in the enclosure.

T386
A 13- to 16-year-old of indeterminable sex 4473*/(4474)/4475.

T388
A 14- to 16-year-old of indeterminable sex 4488*/(4489)/4490.

T380
A 12- to 18-year-old person, perhaps female, 4387*/(4388)/4389 whose tomb was defined by a marble block (4394).

T381
A 20- to 30-year-old male 4390*/(4391)/4392.

T382
A 20- to 35-year-old person, perhaps female, 4403*/(4404)/4405.

T385
A 6- to 8-year-old child. 4478*/(4479)/4480. The latter of these was subsequently cut by another grave. A cut 4459, was made for a 15- to 20-year-old sub-adult, 4457*/(4458) with an upright paving stone (4471) in the side of the tomb, a semi-circular cut 4431, filled by (4432).

T389 and T390
Two tombs, one including the remains of four adult males, reduced at the bottom of the tomb 4500*/(4501)/4502 and a 40- to 50-year-old male 4541*/(4542)/earlier fill (4531)/4530 (T389) and a round pit, 4511/(4512), were cut into the cocciopesto pavement north of the razed Roman wall, wall [4460]. Radiocarbon dating of the adult male 4541* dates his death at 890–1020 cal. AD (95% probability; OxA-26153). The eastern tomb (for 4500*) had a cover made of reused roof tiles and a marble slab (4509), while the western tomb, 4530, lacked a cover and comprised a reduction of at least three individuals using part of a Roman column shaft [4546] as the shoulder wall of the tomb. The round pit served an unknown function; the rectangular cut 4416 was filled with rubble (4417). These pits and graves were covered by a beaten earth surface 4373 which contained residual sixth-century pottery.

Graves outside the monastery precinct

T383
An adult of indeterminable sex 4429*/(4428) was buried in an indeterminable orientation in a constructed tomb, a cut 4443 lined with marble slabs [4238]=[4427], [4426] and [4237] mortared together and resting on the paving stones, though only the hand is preserved after subsequent building activities.¹³

T357
An adult of indeterminable sex, 4087*, buried in a west–east orientation in a built tomb (4088)/4089 with limestone slab wall [4091], probably relates to this phase.

T371
An adult of indeterminable sex 4307* was buried in west–east orientation with legs flexed in the door of opus mixtum exterior wall of Roman building (4306)/4308 and the grave was covered by fragments of tile (4571) and earth (4572).

Into the black loamy layer, (4330), a number of tombs were cut.

T378 and T379
In the northern part of the trench, two tombs were identified: a 20- to 30-year-old male was buried in west–east orientation 4375*/(4376)/4374 and a 12- to 16-year-old subadult parallel to him in the same orientation 4378*/(4379)/4377.

In the north of the trench, a shallow oblong cut 4333 with a flat bottom was filled with rubble from construction materials and fragments of stone, (4334), perhaps related to construction activity. These were covered by a deposit of limestone flakes and loamy earth (4256), also presumed to relate to construction activity, perhaps of the buttresses or even [4005]. These were covered over by a deposit of charcoal and ashes with fragments of pottery, perhaps a deposit of rubbish intended to raise the level of the area (4290), covered by a compacted occupation surface, 4265 (Fig. 28).

T370
Into the occupation surface, an earthen grave was cut for a 15- to 20-year-old male 4274*/(4268)/4266, laid in an east–west orientation. The grave was then cut by a later posthole 4276, which was filled by a slab of cocciopesto [4267] and then later earth (4273). This posthole may have been in some way related to scaffolding or other structure used in the construction of [4005]. A shallow subrectangular hole 4298 was filled with rubble (4299). These were eventually covered by rubbly deposits (4292) and (4167), discussed below.

 

Chronology

The wall postdates the earlier ninth-century phase. Its rubble masonry, lacking the formally dressed stones of the later eleventh century, or the coursed masonry of the ninth century, may date to the tenth or early eleventh century. There is very little archaeological material which would date the contexts further; there are few pottery-rich deposits in this phase. Such a date coincides neatly with the documents describing the foundation of the monastery: a parchment in a twelfth-century hand records the foundation of the monastery of S. Maria and S. Pietro in Villamagna in the year 976.¹⁴ The manuscript is a twelfth-century copy of what we presume to be a tenth-century original; there is an original tenth-century document recording a property donation of 979 to the monastery which thus exists by that date.¹⁵ There seems little reason to doubt that these major walls defining the area adjacent to the north side of the church are, in fact, the first steps towards the building of the monastery, an institution created in the end of the tenth century. The burial of 4400*, on the inside of the enclosure, uses a piece of broken liturgical furniture [4484] typical of the period from the late eighth century and the early ninth century. However, at least some of the burials to the west of the enclosure may pre-date the creation of the wall and the monastic enclosure, as there is no direct stratigraphic connection, apart from the fact that they are both later than the dark-earth deposit (4330).

The lime kiln

To the west of the cloister, outside the building, an enormous layer of building rubble, marble and stone chips and pottery (4167) that covered the entire western part of the trench to a depth of some 40 cm was deposited, perhaps to raise the level of the area, and later covered by a pavement surface 4203 and its preparation of silt with small rocks (4230) (Fig. 29). The materials in the construction rubble layer are predominantly residual pottery of the ninth century, though there are a few later pieces. The earlier tomb 4235 and skeleton 4234* were disturbed by a cut 4236 made for ossuary pit (4231). This deposit of bones is probably related to 4168/(4169), an ossuary pit in the southwest area of the trench. Both intrusions were cut down to the pavement of white stones, 4240 and 4239, respectively, cutting through (4167).

This area was then used for the construction of a large lime kiln (Fig. 30). A deep hole, 4126, approximately 8.75 (east–west) by 4.8 m (north–south) with a depth of 1.17 m, sliced through the earlier deposits, tombs and even the Roman limestone paving. The cut was shaped like a frying pan: round with a corridor extending to the west that was the praefurnium for firing the kiln. The cut was lined with large tartara blocks, 30 x 45 cm, (4036) packed into the clay layer which underlies the Roman buildings on the entire site, (4447). The upper surface of this clay was severely charred during the lime-making process, (4059). The cut was actually rectangular at the northeast corner and dark soil (4124) was used to pack behind the stones of the kiln itself. At the mouth of the kiln this packing was coursed and particularly well-built [4123], sitting on top of the northern wall of a tomb [4238]. In a similar fashion [4235] was assimilated into the structure of the southern wall of the praefurnium, also numbered [4123] (Fig. 31). It must have originally been covered with a stone vault, but this was removed after the final firing.

Around the crest of the wall of the kiln, two mortar pavements were created for successive use with the lime kiln, 4203=4156 and later 4017, perhaps to be associated with two periods of firing of the kiln. The earlier surface was pierced by a number of holes, 4139/(4138), 4204/(4205), 4206/(4207), 4209/(4208), 4211/(4210), 4213/(4212), post holes which may have been used for a wooden structure(s) associated with the firing of the kiln. These were covered by (4151)=(4192), a sandy grey pebbly preparation for the plaster pavement 4017. A much deeper hole, 4163/(4171)/(4164), located close to the praefurnium was cut into the later surface 4017 and rimmed with mortar. It may have worked with the praefurnium, though in what capacity we do not know. Another rectangular hole 4115 was cut into the mouth of the kiln through the pavement 4017 and filled later with light brown rubble (4116). A long and shallow rectangular cut, 5.60 m long by .70 cm wide, 4041, runs roughly east–west, cutting through the pavement 4017. It is very shallow and may preserve the traces of a low wall, removed and now backfilled with loose silt (4140).

The technique of construction and the scale of the kiln follow those known from ethnographic research in the Mediterranean and ancient textual descriptions.¹⁶ In the praefurnium near the mouth of the kiln this packing was coursed and particularly well built, in order to feed the kiln over the several nights and days that it took to fire the limestone blocks with which it was filled.¹⁷

While the kiln was being fired, three deposits of burnt earth, clay and partially incinerated fragments of marble formed. An acidy, limey deposit (4035) filled the lowest surfaces of the lime kiln and this was later covered by (4037). In the praefurnium, a sloping deposit of rubble was covered with lime (4229). A small low wall encrusted with lime partially blocked off the mouth of the praefurnium perhaps as part of an early firing (4219). A deposit of fired chunks of red cement, fragments of worked coloured marble, plaster, fragments of bark, some fragments of ceramic and glass partially fills the mouth of the praefurnium after the first firing (4128). This was partially removed by a square cut 4137, backfilled with ashes (4218) and a small piece of early medieval sculpture (AE1037). Another pack of rubble (4227) was put in, perhaps to brace the mouth of the praefurnium in a second firing. The shape of the kiln and the praefurnium mouth, with evidence of multiple firings, is similar to a late medieval kiln near L’Aquila, excavated in 2002.¹⁸ 

Against the southeastern wall of the kiln [4036] and parallel to the medieval wall [4005], a low wall some 4.5 m long and .5 m wide, [4004], was built.¹⁹ This wall, the north–south wall [4005] and the medieval buttress [4014] defined a rectangular vat, cut down, 4260, and lined with layers of quicklime (4057) (Fig. 32). This must have been used for slaking the lime from the kiln. An internal wall [4054] covers one layer of the quicklime indicating that at least two batches of lime were slaked in this vat; another layer of lime over these is (4025). As it went out of use, it was filled with rubble and mortar (4015) and later some patches of quicklime (4023) over other deposits of building materials and lime (4024). 4028 is a surface, located to the east of [4005], where lime was spread and worked and it was cut by the shallow semicircular hole, 4026/(4027), and the lateral cut 4029/(4030), for no clear purpose.

The end of use of the lime kiln was marked by a number of dirt and rubble deposits dumped into the kiln and the praefurnium, (4037) then (4035). A deposit of tacky, orangey-brown earth flecked with carbon and fine fragments of terra cotta tile (4149) served as a further levelling layer. Several skeletons were buried in (4149), after the filling of the praefurnium.

T368
The earliest of these, 4199*/(4200)/4201, was a north–south oriented skeleton of a 30- to 40-year-old female whose head was cut by 4215—a cut filled by (4214) and which may represent a re-opening of the lime kiln.

T363, T364, T365, T366 and T367
A prone skeleton of an 18- to 23-year-old male with a south–north orientation, 4146*/(4147)/4148, T363 covered 4199*.²⁰ Also buried in (4149) were 4172*/(4173)/4174, the upper half of an adult (T364), cut at the waist by the grave T365 for 4181*/(4182)/4183, a north–south oriented upper half of a 25- to 35-year-old skeleton cut at the waist by 4217, later filled by (4216), the upper portions of a 25- to 35-year-old adult 4196*/(4197)/4198 (T367) and, along the right side, a grave cut 4195 for a north–south oriented 6- to 14-year-old child 4193*/4194 (T366).

A thin layer of densely packed pebbles and earth 4154/4155, including fragments of liturgical marble (AE1036) lay on a preparation layer of tawny brown soil and pebbles (4187). This covered these graves, but was cut by a group of burials.

T362
4142*/(4143)/4144, the upper half of a juvenile skeleton, buried in a north–south orientation.

T360
4118*/(4119)/4117, an adult male, whose skeleton was cut. He was buried in a northwest-southeast orientation.

T358
To the south of that 4097*/(4096)/4095, an adult of indeterminable sex, was buried in a northwest-southeast orientation.

T361, T359
A cut for T361, 4121*/(4122)/4120, an adult of indeterminable sex was buried on a northwest-southeast orientation. It was cut by the subsequent T359, 4101*/(4100)/4099, a juvenile, located to the south of them. Some of these graves also cut 4098, a thin surface of stone and bone fragments.

T369
Nearly in alignment with the northern edge of the church façade, an oblong cut 4258 was lined with slabs of reused cocciopesto, tiles and marble [4269], [4270] and [4271], into which a 15- to 20-year-old female 4285*/(4286) was buried in a west–east orientation, covered by 4277 and then by a 35- to 45-year-old male 4272*/(4278), with a final tomb fill (4259).

An oblong ossuary, 4168, was also cut into (4330) disturbing the limestone pavement which emerged at the bottom of the cut; it was filled later by (4169). Another, sub-triangular ossuary was cut into a loose deposit (4467) at the western edge of the trench, 4468, filled and covered by (4170). Radiocarbon dating of this skeleton indicates a death in 1020–1160 cal. (95% probability (OxA-24971).

Later monastery, cloister and cistern

Between the parallel north–south walls identified above as the first monastery walls, [4005] and [4189], a new structure was built in a strikingly well executed ashlar masonry, centred between the two north–south walls and the northern wall of the church (Fig. 33). It was called Room 3 in the working documentation but it was quickly evident that this was too large to have been an interior room and it follows the conventional form of a monastic cloister. The new walls form an internal rectangular structure measuring 10.27 m wide, surrounded by a corridor 2 m wide on the west side, 2.8 m on the south and 2.5 m on the east. Its northern extent is not known as it runs beyond the northern limits of excavation. However, the east–west centre line appears to run through a cylindrical wellhead, [4133] and [4296]. As this lies 6.14 m from the southern wall, we would conclude that the full length of the cloister was 12.28 m, making it slightly rectangular, rather than perfectly square.

The walls [4053] on the west (with foundation trench cut/fill 4079/(4078)=4367/(4368)), [4129] to the east (4414/(4415)) and [4006] to the south, have solid square foundations upon which rose walls of carefully squared and dressed stones, a double row of level courses of limestone, bonded in the centre with mortar. The ashlar elevation is preserved only in parts of [4053] (Fig. 34), though dozens of blocks were found in later destruction deposits, permitting the reconstruction of the elevation (AE1001, AE1002, AE1004, AE1005, AE1006, AE1009, AE1010, AE1011, AE1012, AE1013, AE1014, AE1015, AE1016, AE1020, AE1021, AE1024, AE1029)(Figs. 35, 36).²¹ There were clearly double-arched windows, with a double row of voussoirs overtop—a simpler, local limestone version of the cloister of S. Scholastica, Subiaco, of the first half of the thirteenth century.²²

The wellhead was built into a vault [4455](Fig. 37). A course of limestone at the base of the wellhead below [4133], [4296], forms the joint between the wellhead and a cut in the vault below, 4456. It is semi-circular, with an elongated north–south axis, and is only present in the western part of the wellhead. This vault covers a stone-built cistern, whose existence in this location was known when our excavation began, because one of the vaults had collapsed and the hole was visible in the topsoil. In the 2006 season we entered the cistern, measured its size, recorded its construction techniques and the locations of the drain spouts leading into the tank from above. Where the wall surfaces were visible (the cistern was partially filled with rubble, stagnant water and other debris) they were covered by hydraulic cement. Where the walls met the vaults, finely cut ashlar blocks formed the joints. In the interest of safety and preserving the integrity of the remaining vaults, we backfilled the cistern (4162) and continued backfilling it as the fill settled. Excavating from above, we arrived at the top surfaces of the vaults and the mouths of the drainage spouts that had been visible from below.

The upper surface of the cistern vaults, 4056, is a red-brown hydraulic cement, with finely crushed terracotta and pozzolanic ash, over coarser cement with a rubble aggregate. It is quite thin (15 cm) in places, where the edge of the collapsed vault, 4161, allows us to see the section. The upper surface is dynamically moulded to facilitate drainage into the lower cistern through four squared holes (see Fig. 36). From a level horizon near the outer walls the cement surface slopes in regular but not symmetrical curves down into the spouts and the four points with drainage spouts may very well have been located where the overhangs from a roof over the corridor met (Fig. 39). In the funnels moulded into the surface, two different materials were packed: a layer comprising white limestone blocks of small dimensions mixed with a fine sand of dark black volcanic pumice, (4357), and over this a thick red sandy layer (4354) which covers each of the four drainspouts and the entire central area of the rectangular structure. At some point a post, 4362, was put into the red fill (4354), some 40 cm in diameter and 40 cm in depth, filled with loose earth (4363). These deliberate fills must have been intended to filter the water draining into the cistern, like a modern French drain, and may have been covered by soil and planting. We have not been able to identify other examples of this kind of cistern. Other monasteries had cisterns in the cloister, such as that at S. Michele alla Verruca, but none are as large or monumentally built as this one.²³

Around the wellhead runs a low, semi-circular wall [4226] abutting two parallel straight walls [4134] and [4329], which in turn abut the eastern wall of the cloister [4129]. This apsidal structure measured 3 m by 4.9 m. A number of bonded highly dressed stones [4135] defined the collapse, face down, of an arched portal leading from the apsidal structure to the eastern arm of the cloister, probably the original entrance to the well house (Fig. 39). The little building was paved with a mortared masonry surface [4328] made of bricks, limestone and very hard black mortar, which turns blue when exposed to the sun and air. The pavement slopes towards the east and there is a square hole cut through the eastern wall of the cloister, [4129], which permitted water to run out of the apsed well house down a drain under the floor of the cloister. The drain then ran down the slope of the hill, defined by [4359] and [4360]. It may have run down to fields or a garden located beyond the limit of the excavation. The exterior faces of the well house are plastered with a cement coating, (4453) and (4454), to seal the corners. To the east of the well house is the eastern arm of the cloister, reached by a step (4332) down from the well house.

The pavement of the well house was repaired at some point and the level of the pavement was raised by the deposit of a layer of loose earth and broken pottery, (4364). It was then paved with limestone pavers of medium dimensions (20-30 cm), [4297], over a surface of cement and mortared rubble preparation [4328] which bonds to the side walls of the apsidal structure. There is a construction joint running directly over the eastern edge of the cistern, which separates the paving of the area around the wellhead from a separate, later deposit of cement blocks of medium size and varied shapes (4327) filling in the area below and east of the pavement [4297] and sloping down to the drain to the east. The cement blocks were covered by a deposit of loose earth, rubble and chips of pottery (4320). The foundation trench for [4294] was covered by mortar and small fragments of building rubble, the construction surface of the cloister walls (4408) and its equivalent of (4380).

An east–west wall [4294], abutting the much earlier wall [4188], was constructed to create a return for that wall and protect the entrance to the well house. It had the foundation trench 4406/(4407) and the deposits of rubble. New surfaces were laid over the earlier pavements, a deposit of brown soil and stones, uniformly horizontal and compacted, 4384=4369. Another beaten earth surface, 4361=4356, covers these, maintaining the edges of the drain. This surface was cut near the entrance to the well house but against the outer wall of the cloister arm, [4188], for a rectangular tomb, 4350 (T376). In it a 25- to 30-year-old male, 4348*/(4349), was laid with a north–south orientation, and the grave was covered with a flat long limestone slab (4358) placed on the surface horizontally over his feet. The man had the point of a lance in his skull.²⁴

Outside of the monastery, a wall was built perpendicular to wall [4005], running west [4125]. It was robbed out 4034 at a later date, leaving a scar on the western face of [4005]. There was a parallel cut for a wall running 1.9 m to the south, though it does not continue as far as [4005]. The southern cut 4141 measured 5.6 m long and passed beyond the limits of the excavation to the west (Fig. 40). Both cuts were about 70 cm wide. The southern wall may have been the stylobate of a portico or arcade running in front of the building, though this would not have led into the monastic cloister itself, at least not at ground level. These walls may have been part of a building which was open to the laity, like a guesthouse or the abbot’s residence, though with only the traces of two walls, we really cannot even hypothesize about their function.

Chronology

The fine ashlar masonry of the cloister elevation points to date after the mid- to late eleventh century, when specialist masons began carrying out major architectural commissions in Italy.²⁵ The apse and façade of the Duomo of Anagni, dated to the late eleventh or first half of twelfth century, is carried out in this kind of ashlar masonry in travertine, with blocks similar in size to those of the cloister (Duomo: 33-40 cm long x 19-29 cm high; Villamagna cloister: 48-72 cm long, 32-37 cm in height, 15-25 cm deep).²⁶ Daniela Fiorani has examined masonry in Lazio, dating the use of travertine ashlar masonry with newly cut blocks and very fine seams, sometimes using mortar (IIa in her classification) to the eleventh or early twelfth century in ecclesiastical and domestic architecture in cities, like Anagni and Ferentino and in fortifications (IIb) to the thirteenth century. Her chronotypology of ashlar masonry in tartara, like that at Villamagna, is as follows: IIIa: blocks that measure over 25 cm in height, dating to the late eleventh or first half of the twelfth; IIIb: blocks that measure under 20 cm in height, dating to the thirteenth or fourteenth centuries, regardless of whether they appear in rural or urban, defensive or ecclesiastical structures.²⁷ 

Fiorani clearly could not have seen the ashlar cloister blocks excavated at Villamagna which would seem to fit her chronotypology of the late eleventh­-early twelfth century. Small colonnettes have been recovered in the excavations (AE1032, AE1033) indicating that the bifora arches were separated by these colonnettes, much like the cloister of nearby Casamari (perhaps mid-twelfth century²⁸), or the elegant Cistercian cloister of Valvisciolo (thirteenth century²⁹), which has a door profile which matches closely the profile of the cornice from the cloister at Villamagna (Fig. 41). The double-arched arcades resemble those at the cloister of S. Scholastica (Subiaco), signed by the mason Jacopus Romanus of the Cosmati family ca. 1220 (Fig. 42).³⁰ Indeed, these architectural forms were in use in this part of Italy through the twelfth and thirteenth century, and ours were recovered without capitals, which might provide clearer dating evidence. Elsewhere in Lazio, ashlar masonry in tuff, like that of the cloister, has a slightly later chronology than that suggested by Fiorani for southern Lazio; in Tarquinia, for example, the largest masonry blocks date from the first half of the thirteenth century, measuring between 32.5 and 34 cm high³¹, while at Acquapendente, the largest blocks of ashlar masonry dated from the second half of the fifteenth century.³²

The fortification of the site as a castrum and the reuse of the earlier buildings (Fig. 43)

The area inside the cloister over the cistern filled up with several separate layers of materials and a single grave, signifying a major change in function of the area.

T356
The original red deposit of the cistern filters was cut for a tomb, 4086, of a 25- to 35-year-old male oriented north–south, 4084*/(4085). At this time the elevation of the well house was apparently suffering, as a post was put in on the southern side of [4134] of which the hole 4224 remains, filled by the remnants of the post (4225). The area of the cloister then filled with layers of accumulated earth and materials. First was 4228=4185, a compact beaten earth surface comprising mortar, small inclusions and flecks of charcoal. The surface is very thin towards the corners of cloister and not present in the southeastern corner. The surface of the pavement is made up of very small granules of mortar and fragments of pottery, and the loose silty quality of the soil suggests a high content of acidic organic material, perhaps horse manure. A subsequent deposit comprised a large quantity of marble fragments of fragments of finely worked coloured marble (opus sectile) (concentrated in the eastern edge of the room) and tile and pottery, (4184) (Fig. 44). The marbles seem to have been stored in this area. These appear to be exclusively Roman—no medieval carving was found in this context, though a small ‘fleur-de-lys’ marble ornament (AE1043) was recovered from later deposits in this room. There is an abundance of large pieces of late thirteenth-century and early fourteenth-century pottery: ceramica laziale, local lead-glazed pottery and tin-glazed wares mostly of storage vessels.

In the eastern wing of the cloister, the drain running from the well house out to the east filled with silt and fragments of pottery, (4353) and (4319). A horizontal cut, 4351, removed the top slabs of the drain. The drain had already filled with gray silty deposit, the removal of the slabs resulted only in the absence of covering and the redepositing of two slabs and gray mortary fill (4352). The corridor was covered by a uniform grey surface, (4331). A shallow circular cut, 4336, may have removed building materials and was backfilled by loose silt (4318).

Chronology

The occupation layers associated with the reuse of the cloister area, such as (4184), include fragments of pottery from the thirteenth century. At that point, the porticoed cloister no longer fed an underground cistern but the central area went on to be used for other purposes, including storage of animals and building materials, until they were abandoned and then destroyed. The situation in the well house provides a concise example of the chronology of the transitions in this area: the mix of materials in the backfilled hole 4300/(4301) date from ca. 1250-1350 reflecting the date that these leveling deposits were formed, rather than the date of the excavation and backfill which nonetheless happened shortly thereafter, probably in the late fourteenth century. The silt in the drain from the well house, (4353) and (4319), included fragments of pottery from the fourteenth century and a coin (C221), dating from the mid-thirteenth century. Inside the well house, the collapsed portal entrance was covered by (4250), a rubble deposit related to the collapse of [4329] with pottery from the fourteenth century. This chronology is equally true outside the monastery. The enormous layer of building rubble, marble and stone chips and pottery (4167) includes residual material from the mid-thirteenth or fourteenth century. These phases of changed use, abandonment and destruction thus coincide with the papal document of 1297 supressing the monastery and awarding all buildings, properties and vassals to the bishopric of Anagni.³³

Destruction and accumulation of deposits

The accumulations and surfaces in the central cloister were covered by a destruction layer of mortar and rubble, (4186)=(4223), extending some 50 cm from the wall [4134]. It probably is the collapsed elevation of the southern wall of the well house and, as such, marks the beginning of the phase of destruction of the cloister buildings. The worked stones of the arched entrance to the well house, [4135], fell on to a deposit of stones and silty soil in the apsed room, (4291), and a layer of fluffy brown soil (4245), the abandonment of the area.

The walls of the well house were knocked down to a height of about 30 cm, 4335. The entire cloister area was covered by a destruction layer, (4180)=(4058), rich with bone, rubble, mortar and tile fragments, tied to the razing of the walls 4202. Inside the well house, the collapsed portal entrance was covered by (4250), a rubble deposit related to the collapse of [4329] with pottery from the fourteenth century. In the area to the east of the cloister two deposits of rubble, (4275) and (4433), are the remains of the upright walls [4188], [4189] and [4294]. This layer, in turn, was covered by a dark brown occupation layer (4145) containing large amounts of thirteenth- and fourteenth-century pottery. To the north of the well house, a deposit of limestone of medium and large dimensions (20-30 cm) forms a coherent horizontal surface, 4305, on top of (4180). In the area adjacent to the well house to the west, there was a level surface of worked limestone blocks (4179) and four faced marble tiles 4175. These lay on a layer of dark brown soil and rubble, (4176). The entire area was covered with a layer of fine, fluffy brown soil, (4136), which is probably evidence for the abandonment of the cloister after its destruction (Fig. 45). (4136) was covered by a layer of rubble (4132) which included modern stamped bricks as well as medieval worked stone and pottery. A quarter of the well house was dug out at an unknown point. A semicircular cut, 4300, in the southeastern section of the room went down to the original paved layers, removing all layers of deposit, presumably looking for stone. The hole was refilled by the materials from these layers (4301) thrown back in.

In the western wing of the cloister (4055) formed, an irregular firm deposit of compacted soil with stones and frequent tile, as well as snails, mortar and charcoal. Over it came a destruction layer (4048) which filled the area with sandy grey blue silty ash, covered by (4069), a yellowish brown layer with ashlars and rubble, (4052), a grey brown later of silt with ashlar rubble, covered by a thin white mortary patch (4090), a burned patch (4103) and a mortary pile of stone and tile rubble abutting [4053], (4092). These contain ashlar stones from the elevation of [4053], once it had been cut down, 4050, and roof tiles, along with pottery from the fourteenth century. An east–west cut 4068 filled by (4051) and (4070) removed part of the destruction layers, perhaps created in a search for stone.

4154/4155 covered the later mortar pavement 4017 forming a continuous surface over most of this area. This pavement system was bounded on the north by the wall [4125] that ran east–west and bonded with the west wall of the monastery wall [4005]. The surface was also cut by two large holes that suggest occupation and domestic use—a grain storage pit in the northwest corner of the trench, 4131/(4130) and lower fill (4249), and 4160/(4159), which may have been used for kitchen refuse.

Over these surfaces, sitting directly on a thin surface with fragments of brick and some bones (4080), a metre-long wall [4018] was built, with foundation trench 4093/(4094), but no other structures are related to it. A late surface (4046), to the north of the abandoned lime kiln, was cut for a series of tombs against [4005].

T350
A 25- to 35-year-old female 4040* was buried in 4038/(4039) with north–south orientation, though the grave was significantly disturbed by later building activity.

T351
A 30- to 40-year-old male 4062* was buried in 4060/(4061) with a north–south orientation.

T353 and T354
Another 30- to 40-year-old male 4073* was buried in a north–south orientation in 4071/(4072) partially cutting through an earlier grave, T354, of another adult, probably male, in a north–south orientation 4077*/(4076)/4075 slightly to the west, cutting through 4017 and (4046).

T352
Another north–south earthen grave 4065 was cut for a 30- to 40-year-old adult of indeterminable sex 4067*/(4066).

T355
A 25- to 35-year-old adult, probably male, in a north–south orientation, 4083*/(4082)/4081 was only partially preserved.

Along the northern baulk of the trench, separated from the other contexts by the robbing out of the east–west wall 4034 are some other graves buried in a deposit of loose black soil and rubble (4568) which covers the cut-down wall [4321].

(4564) and (4570) are both loose deposits of human bone disturbed by later graves. 4569* is a reduction of two individuals located near to (4564).

T391
All of these were cut by 4567, the tomb of a 30- to 40-year-old male 4565*/(4566).

Chronology

The destruction and accumulation layers of this phase contain pottery of the fourteenth century, including pitchers of tin glaze wares.³⁴

Modern

A modern semi-circular cut 4158 on the south side of the trench sliced through a number of these layers: (4136)/(4145)/(4180) as well as removing part of wall [4006]. It was backfilled with a mix of rubble, a fragment of ancient statuary (S26) and modern material, (4157).

The eastern corridor of the cloister was filled with modern deposits, which covered the even grey surface (4331). A grey, rubbly layer (4322) contains modern roof tiles, and a later dark yellow brown layer (4288) contains many nails and fragments of pottery. A thin brown layer (4287) runs in a strip north–south, quite compacted against wall [4129].

A semi-circular cut, 4190, cut away part of the hill and removed the medieval and early modern layers of the area to the east of the cloister, Room 4 (Fig. 46). It was cut from (4177), a churned up mix of pottery, rubble and soil very near topsoil, measured 6.50 m (north–south) and 4.0 (east–west) and filled by (4191), a great deal of material from the Second World War, suggesting that this was a German foxhole dating from the occupation of the site during World War II. A cache of spent rifle bullets (2009) appeared just below topsoil in the northern part of BI.³⁵

In the western part of the trench (4020), (4021) and (4064) contain fragments of Roman statuary, residual pottery of the late fourteenth century, and represent accumulations filled with material after the abandonment of the lime kiln. These, and the thinner surface (4019), were cut by 4034/(4000) to remove the east–west wall which bonded with [4005], perhaps to find stone for another building project. Parallel to 4034 is a second robber trench 4141/(4140), which does not run the full length of the trench but ends in the centre. The fills were covered by (4049), including modern building materials, and (4063) in the far west of the trench.

Building works in the modern period have left several deposits of rubble, mortar and building materials: some filling cuts 4107/(4031)=(4108), others (4032) simply deposited on work surfaces (4033) near the north side of the church.

A long, low east–west wall [4003]=[4007]=[4008] was built to terrace the hillside to the north of the church and prevent its encroachment on the church itself (Fig. 47). A horizontal cut, 4009, removed part of the earlier medieval walls [4004] and [4005] to level the area. A gray-blue mortar and stone layer (4047) may be associated with its construction and an accumulation of rubble (4012) to the north may have been gathered to fortify the wall. On the far west side, (4074) and (4127) are gravelly rubble deposits which support the wall on its uphill side. An occupation surface (4022) south of the terrace wall was covered by (2147)=(2027)=(2148) in BI. This deposit was in turn covered by (4031), a spread of mortar and stone near the corner of the church, which might have had the aim of shoring up the northwest corner of the church, or may have been otherwise related to a construction project. To the north of the terrace wall, two long channels for plantings are visible with their black soil, (4010) and (4011). A horizontal cut 4009 removed surfaces to cut back part of the hill. Later, the terrace wall collapsed partially, leaving the slumping deposits (4001) and (4002) sliding down the hill towards the church.

Topsoil, (4000), covered all of this.

---

1^ See: Young, Stratigraphic report from Areas A and G.

2^ Margaret Andrews contributed much to this discussion of the portico.

3 Regio I — Insula V–VI, Calza (ed) 1953: 201–3.

4 See: Dufton, print volume: 64–9.

5 See: Young, Stratigraphic report from Areas A and G.

6 See: Ciausescu, print volume: 252; Rice, print volume: 119–21; Hoffman and Lepri, print volume: 252–3; Totten, print volume: 240–52. For full catalogues, see: The pottery and glass.

7 See: Rascaglia, print volume: 278–84; Medieval Pottery.

8 This individual was not excavated, but was visible in the northern baulk of the trench and so was recorded.

9 This individual was not excavated, but was visible in the northern baulk of the trench and so was recorded.

10 See: Totten, print volume: 240–52.

11 See: Rascaglia, print volume: 278–84.

12 See: Totten, Stratigraphic report for Areas A and F.

13 These contexts were numbered in 2008 [4238], [4235], [4237], 4234*, (4233), not excavated, and then renumbered in 2009 when they were excavated.

14 Flascassovitti 1994: no. 1.

15 Flascassovitti 1994: no. 2.

16 See: Cato, Agr, XLIV, 38, discussed in Adam 1994: 65–71; Delaine 2000: 230–68.

17 On firing times based on ancient sources, see: Adam 1994 and Delaine 2000.

18 Redi 2003.

19 For discussion of the cemetery wall, see: Fenwick, Stratigraphic Report for Area BI.

20 1050–1260 cal. AD (95% probability; OxA-24968).

21 See: Goodson, print volume: 284–300.

22 See: n. 30, below.

23 Francovich and Gelichi 2003: 24.

24 See: Candilio, print volume: 390–2.

25 Bianchi and Valenti 2009.

26 De Meo 1998: 75 measures the blocks of the façade of the Duomo. He dates the masonry to the second half of the eleventh or early twelfth century, following Fiorani. Cf his Tav. 22/a, 24.

27 De Meo 1998: 75 measures the blocks of the façade of the Duomo. He dates the masonry to the second half of the eleventh or early twelfth century, following Fiorani. Cf his Tav. 22/a, 24.

28 Farina and Fornari 1978: 62.

29 Cristino 2002.

30 Claussen 1987: 77-80.

31 Chiovelli 2007: 180.

32 Chiovelli 2007: 176-9.

33 Flascassovitti 2007: no. 177.

34 See: Rascaglia, print volume: 315–25.

35 See: Fenwick, ‘Stratigraphic Report for Area BI.’