During the Roman period, the site was inhabited, and appears to have formed part of the known villa’s ‘farmyard’, occupied by ancillary or agricultural buildings such as sheds and barns. An early phase of post and beam buildings, with a possible northern boundary wall, was identified, overlain by a second phase in which a sequence of ditches were excavated and the original buildings demolished and replaced (see MNT28606 for second phase).

The Roman finds assemblage broadly indicates that the earliest occupation on this site was pre-Hadrianic (in or before the first quarter of the 1st century AD). Although much of the early Roman dating evidence is sparsely distributed and erratically stratified, activity from this period does seem to be concentrated within a roughly north-to-south-aligned band across the approximate centre of the site. The principal groupings of this early phase of features all suggest the presence of a range of small buildings, altered and extended several times before being demolished; the presence of post-holes and beam-slots, together with wider linear features that may have been foundation trenches and large amounts of roof tile and stone rubble, indicates buildings partly of stone and partly of timber, with tile roofs. The sets of intercutting post-holes probably indicate that timber posts rotted and needed to be replaced frequently in the wet ground.

Other complexes of post-holes and beam-slots have been identified directly adjacent to the ‘buildings’ allocated to this phase, but these feature groups were even more poorly dated; these possible buildings could only be identified as Roman, although their position and alignment may suggest that they formed part of the earliest settlement on the site. The two ditch or foundation/robber trench fragments, with the stone wall remnant, may have demarcated the north side of the property on which the buildings stood; if the wide, a shallow linear feature can genuinely be attributed to this period, it may also represent a remnant of an earlier boundary, on broadly the same alignment as its successors, obliterated by later activity across the rest of the site.

Structural Details:

At the north-western corner of the area, a closely positioned double row of pits, post-holes and other features ran north-north-east to south-south-west, parallel to the north end of two ditches at a distance of approximately 6m to the east. These features appear to comprise a single structure, although some of the pits intercut; the majority of the pit alignment is recorded as a group, while the stratigraphically later features, which cut features in the group while clearly forming part of the structure, are also grouped together. The features varied so widely in size and shape that they had little in common apart from their alignment, with the two early features apparently representing a stake-hole, 0.16m in diameter, at the end of a 0.20m wide beam slot truncated by a pit, while the largest feature, a sub-rectangular pit, was slightly over two metres square and 0.60m deep. Only two of the pits could be reliably dated: twelve sherds of mid- to late 1st to early 2nd century pottery, with a quantity of animal bone including the articulated leg of a horse, were retrieved from a fill in the most northerly pit in the western pit alignment, while six sherds of early Roman pottery were retrieved from the upper fill of a pit, directly to the south of most northerly pit: the feature complex has been assigned to this Early Roman phase on the strength of the dating evidence from these two pits. A further 19 sherds of pottery that could only be identified as Roman, with a very large assemblage of heavily abraded Roman ceramic building material (CBM), were also retrieved from the sub-rectangular pit; small amounts of Roman CBM were also retrieved from two other pits, at the south end of the eastern pit alignment, and a single sherd of Roman pottery was retrieved from another pit in the same alignment. Most of the CBM retrieved was roof tile of the interlocking tegula and imbrex types, with a much smaller amount of brick. Unlike other post-holes and post-pits on the site, there was little evidence that the CBM had been deliberately inserted as a pad beneath or packing around a post; however, one pit in the middle of the western row of the complex, contained part of a worked stone, possibly a fragment of a millstone, which seemed very likely to have been placed to support a post, while the rubble inclusions in the pit in which 6 sherds were recovered from the upper fill, occupied the centre of the fill, potentially suggesting that they were deposited following the removal of one. The millstone fragment could not be retrieved for assessment, but photographs of it in situ were submitted to a specialist who considered it likely to be Roman. Environmental sampling was inconclusive, with a sample from one post-hole producing three charred barley grains, a few charred seeds of sedge, grass and cabbage/mustard species and a small amount of charcoal. The dating evidence for this feature group being among the earliest on the site was borne out by its stratigraphic relationships. To the west, a ditch fragment ran close to the group on a converging alignment, cutting the edge of one pit in the earlier alignment; this ditch has been ascribed to the mid-Roman phase. The two most southerly pits in the group were both cut by, and therefore stratigraphically earlier than, the possible robbed-out wall, which formed part of a substantial late Roman phase ditch sequence. The dating of this pit complex is called into question by the presence of an apparently re-used millstone fragment within one of the pits; however, grain was presumably being processed on the villa’s land, and so worn- out millstones would have been periodically discarded, throughout the period of Roman occupation here, and this fragment is not necessarily associated with the putative late Roman watermill structure.

A small group of features near the south side of the site have been tentatively interpreted as the remains of a structure, dated largely by their stratigraphic and spatial relationships to the Early Roman phase and the beginning of Mid-Roman phase, and tenuously connected by similarities in ceramic building material. The shallow, sub-circular pit, approximately 0.92m in diameter, contained frequent medium to large limestone fragments, cobbles and large but very abraded Roman CBM fragments in its fill, suggesting that demolition material might have been used as packing in a large post-hole; it produced three sherds of mid-1st to 2nd-century pottery. Three tegula roof tile fragments retrieved from the pit fill appear to have come from the same roof. About 2m to the east of the shallow pit lay a truncated fragment of an elongated or linear feature 0.47m wide and 0.08m deep, from which a further five sherds of the same date were retrieved; the construction of the school had destroyed the east end of the feature and any potential stratigraphic relationships to the east, so it could not be ascertained whether or not this feature had been cut by the large ditch sequence adjacent to it. A truncated linear feature some 5m to the south of both features, excavated with difficulty in flooded conditions, is tentatively assigned to this complex, since its stratigraphic relationships suggest an early date and it appears to run parallel to the elongated feature. It survived to a length of 3.0m, and ran approximately west-north-west to east-south-east on an alignment close but not exactly parallel to that of the north side of the Roman building. It was 0.75m wide and 0.25m deep, and Roman CBM fragments, stone rubble and mortar flecks in the upper fill of a section through the feature suggested an interpretation as a robbed-out wall, although no such inclusions were seen in a second section. Although the single sherd of pottery retrieved from the second section through this feature could only be identified as Roman, pieces of tegula from both the elongated feature and the truncated linear feature were sufficiently similar to those from the pit to suggest that they had all come from the same roof. The feature was cut into a clay layer, a deposit which survived only within the footprint of the school building and which may have been an occupation layer, but which produced no finds other than a fragment of an iron nail, and so seems more likely to have been a flood deposit. Another pit, a shallow, flat-based pit cut by the eastern edge of a ditch, produced no dating evidence, but as the ditch is considered to represent one of the earliest Mid-Roman phase features, the pit can tentatively be assigned to the Early Roman phase. The features making up a structure may represent the vestigial remains of a building composed of posts and ground-beams. The west end of a gully was cut by a perpendicular gully, which is very provisionally assigned to the Early Roman phase on the grounds of its alignment with the cluster of this phase features to the north, and may represent a later alteration to the structure.

A group of more precariously dated early Roman features in the centre of the site may also have represented portions of a building that was serially altered and extended; it is possible that the structure forms a part of this feature complex, with the connections obliterated by later redevelopment. The stratigraphically earliest features were a cluster of six post-holes or small pits, intercutting in two groups of three, and truncated by later features and by the school construction. Three pits all had flat bases and steep sides, suggesting a common function, possibly as post-pits; one pit was stratigraphically the earliest of these, and another pit, which had no other stratigraphic relationships and might in fact have been associated with the overlying the mid-Roman phase structure was the latest. Directly to the north of these was the flat-based, a stone-filled pit, which may also have been a post- pit, although no post-pipe was identified within a stony fill that might have suggested that the stones had been inserted as packing to stabilise a post, rather than dumped in a pit that had gone out of use; it was cut by the two smaller, shallower pits. The sub-rectangular pit can be speculatively dated by a sherd of 1st to 2nd century pottery, although this may have been redeposited from the fill of one of the shallower pits, while the other shallow pit had been almost obliterated by later features. These pit groups were cut by a complex of narrow linear features that may have represented beam-slots. One of the earliest of these was some 8m long and continued the line of a perpendicular gully north-north-eastwards: if it was, as it appeared, a part of the same feature disturbed by later groundworks, it suggests that at least one of the triple pit-groups was of the same date as the structure, as the earliest linear feature cut a possible post-pit. Another linear feature ran parallel to the earliest linear feature, but could not be traced as far to the south; it was only 6m long and had no possible southern portion corresponding to the perpendicular gully. It was wider than earliest linear feature at 0.55m, suggesting the base of a foundation trench rather than a beam slot, and could be dated only by its stratigraphic relationships. The north end of the earliest linear feature was in turn cut by a narrow linear feature on a perpendicular west-north-west to east- south-east alignment: one section produced a single sherd of Roman pottery, with small fragments and flakes of Roman CBM. This feature also cut through the pit cluster. The latest linear feature in the sequence extended some 14m to north-north-east with a small right-angled return to west-north-west, which cut through the lnarrow linear feature and into the north end of one of the linear features, suggesting a combined recutting and extension of the feature complex; the U-shaped profile of this feature suggested a beam slot more strongly than the shallower, more rounded profiles of the other linear features in the complex did.

The possible beam-slot complex was cut by a number of pits and large post-holes, all but one of which have been assigned to the Early Roman phase. The best-dated of these features was a sub- circular pit which produced three sherds of pottery dating to AD 100-135, with a substantial assemblage of 88 fragments of Roman CBM. The variety of tile fabrics represented indicates that they did not all come from the same building, although some tegulae were sufficiently similar to those retrieved from a pit and two other linear features (previously mentioned) to suggest that these had all come from one roof, providing a tenuous connection to the earlier Phase 1 feature group structure. The sub-circular pit measured 0.96m x 0.88m and was 0.35m deep, with a flat base suggestive of a post-pit; it cut the terminal of possible beam-slot as well as the smaller pit, and it is possible that the early pottery was redeposited from the fill of one of the earlier features, although the CBM was probably intrinsic to the pit, as only one further fragment was retrieved from the fill of the smaller pit. The pit fill was sampled, but no environmental evidence was retrieved. The extensive CBM assemblage and the presence of wall plaster fragments in the fill suggest that this feature cannot be very early, as Roman buildings must have already been demolished in its vicinity in order to provide these inclusions. However, as with the CBM inclusions in the pits of the feature complex, this material did not appear to have been inserted into the pit in order to support or stabilise a post, but to have been dumped into the pit at the end of its period of use, and it seems plausible that both feature groups represent the remains of early structures, demolished and back-filled with their own demolition materials. Directly to the south of the first sub-circular pit, a second sub-circular pit also cut the west end of a linear feature: again, the fill of this pit contained a substantial assemblage of Roman CBM, whose arrangement within the pit suggested a rubble dump rather than the positioning of a post; the tile fabrics present were commensurate with an Early Roman phase date.

A large post-pit which cut a possible beam slot near its junction with a narrow linear feature, contained stone and CBM fragments which did appear to surround a post-pipe. Two sherds of early Roman pottery were retrieved from a post-pipe fill, while both the post-pipe fill and the surrounding post-packing produced small amounts of Roman CBM. A smaller post-hole was positioned at the north-east angle of the beam slot, separating the two branches and making it impossible to tell whether the feature, immediately to the north, was a pit or post-hole in its own right or a severed northward extension of the beam slot). While the post-hole could not be dated, its position at this junction seemed unlikely to be coincidental, and the stone and CBM rubble in its fill also suggested that it formed part of the complex. One pit positioned about half-way along the beam sot, was filled by a brownish-grey clayey silt that could not be distinguished from the fill of the beam slot; as the pit was flat-based and no deeper than the beam slot, its presence could only be detected where it cut the natural to either side of the beam slot, or cut the fill of the earlier pit, which also pre-dated the beam slot. The pit seems likely to have been a post- hole within and contemporary with the structure represented by the beam slot, with the earlier pit possibly representing an earlier phase of development.

To the east of the north end of one ditch, the wide, a shallow linear feature was cut by the construction trench of later, possibly mid-Roman building structure. This feature produced no finds, but has been assigned to this phase on the basis of its north- north-east to west-south-west alignment and its stratigraphic relationship to the structure.

Although the main Early Roman phase of activity appears to have taken place across the centre of the site, two features at the north-western edge could be attributed to this phase. The earliest identifiable feature in the area appears to be the west-north-west to east-south-east aligned linear feature, cut into undated former ground surface, which contained the remains of a stone structure. The structure was made of rough- hewn, dry-laid tufa limestone blocks, and appeared to represent the base of a wall within foundation trench: it was some 7m long and 1.0m wide, but had survived only as the basal course over much of its length. Its position and alignment suggest that it may have been a boundary wall around the villa’s ‘farmyard’. No dating evidence was retrieved from this feature, which is assigned to this phase on the basis of its stratigraphic relationships. The line of the wall remnant was continued westwards by a linear feature; this feature was 0.80m wide with a flat base, and seems likely to be a part of the aligned linear feature from which the wall had been completely robbed out: frequent fragments of stone in one fill may have been demolition rubble. This phase is otherwise represented only by a single sherd of late 1st to early 2nd-century pottery retrieved with small pieces of Roman CBM and two iron nails from a small pit, directly south of the aligned linear feature. The find may have been redeposited from a pit remnant, which cut the small pit; in either case, it is possible that one or both features were associated with the possible boundary.

R. D. Savage and J. Sleap, 2015, Proposed Residential Development, Former Minster School Site, Church Street, Southwell, Nottinghamshire: Archaeological Excavation Report (Unpublished document). SNT5953.