Mid-Roman occupation was chiefly identified in the eastern half of the site; the features assigned to this phase were generally smaller, sparser and less reliably dated, although this difference may be due, at least in part, to the higher level of disturbance from later activity on the western half of the site. Systems of ditches, probably both for land demarcation and for water management, were the most marked feature of this phase, although post-built structures were tentatively identified in the area formerly occupied by the Early Roman phase buildings. A small, isolated, stone building with a sunken floor, close to the course of the Potwell Dyke, was assigned to this period, and speculatively interpreted as a possible watermill associated with the villa’s farmland.

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; this phase includes a small stone structure near the Potwell Dyke, speculatively interpreted as the villa estate’s watermill.

The systems of serially recut ditches, probably both for land demarcation and for water management, were the most marked feature of this phase, although post-built structures were tentatively identified in the area formerly occupied by the Early Roman phase buildings; the number of recuts and replacements of these ditches suggests that the eastern half of the site remained vulnerable to flooding.

One main structure, a small, isolated stone building with a sunken paved floor, close to the course of the Potwell Dyke, was assigned to this period, and speculatively interpreted as a possible watermill associated with the villa’s farmland, with a mill-leat taken off the Potwell Dyke, now only partiallly traceable through a heavily disturbed area that was flooded during much of the course of the excavation, running to a stone-lined pit or trench that may have contained the undershot water wheel typical of Roman mills. This interpretation is supported to some extent by the finds of several fragments of millstones re-used as post-pads in later features, as the majority of Romano-British watermill sites were characterised by the discovery of millstone fragments. Architectural and technological advances such as water-powered mills are known in the Roman Empire from the 1st century AD, although they only became widespread on Romano-British villa estates in the 3rd and 4th centuries, the structure appears to lie towards the beginning of this date range. The dimensions of the wheels of Roman watermills can rarely be ascertained, but a 1st-century AD example at Venafro near Pompeii, preserved as an impression by the eruption of Vesuvius, could be was excavated at the beginning of the 20th century at Haltwhistle Burn in Northumberland: the building, dated to the 3rd century AD, was identified by the presence of fragments of millstones approximately 0.78m in diameter. No remains of the wheel itself survived, but the mill-race was lined with oak timbers in the immediate vicinity of the building, suggesting that a wheel estimated at 3.0m to 3.6m in diameter had turned in a close-fitting trough some 0.36m wide. The apparently stone-lined feature which was 3 to 4m long and about 0.40m wide at the base, could, therefore, comfortably have accommodated such a wheel. Assuming that the dimensions of the structure have been correctly ascertained, this interpretation appears to require the wheel to be sited at a distance of some 1.5m from the wall of the putative mill building, rather than directly adjacent to or within it, but reconstructions of Roman mills seem to incorporate beams that could bridge this distance: a 5th-century overshot water wheel excavated in Athens turned on a beam 3.5m long, which would be approximately correct to reach from a wheel turning in the stone-lined feature into the interior of the structure. The mill building that formed the first phase of the Redlands Farm Roman villa in Northamptonshire consisted of two rooms, one of which was sunk some 0.5m below the other, presenting a possible parallel to the apparent lowered floor of the structure.

The view was also put forward during post-excavation work that the structure may have been an outside lavatory associated with the villa, flushed by an early course of the Potwell Dyke or by a channel taken off it. Although the presence of a channel around the interior of the building at the base of its walls lends some plausibility to this theory, it seems to be a prohibitively long way from the villa for such a use; moreover, while several fragments of millstones have been found on the site, no worked stones of the keyhole type normally associated with Roman lavatory seats were identified (although wooden seats, such as that recently discovered at Vindolanda Fort on Hadrian’s Wall, would have left no trace). A third interpretation is suggested by the two fragments of possible hearth lining found during the excavation and the potential industrial residues from environmental sampling during the evaluation: the building may have been a smithy or other industrial building, set apart from the rest of the armstead for safety and in proximity to the Potwell Dyke for a constant water supply; the stone-lined feature might, in this interpretation, represent a water tank for activities such as the cooling of worked metal.

One pit, due to its large size and irregular sides suggest that it might have been a clay-winning pit, possibly to make daub for the construction of buildings on or near the site, or as raw material for bricks and tiles, since much of the CBM retrieved may have been locally sourced.

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