GLOUCESTERSHIRE

(1) Eastington, Land at M5, Junction 13 (SO 7778 0645): excavation adjacent to the site of Whitminster Roman villa identified a small number of pits and short ditches containing transitional late Iron Age to first-century a.d. pottery; a buried soil also probably dates to this period. The commonest remains were a typical suite of rural Roman features comprising ditched enclosures and a trackway, pits and post-holes, inhumation graves, and the possible stone foundations for a timber building. Part of the ditched precinct around Whitminster villa was also found.Footnote ¹²⁶

(2) Gloucester, Gloucester City Campus, St Aldates Street (SO 8331 1860): evaluation trenching revealed remains dating to the early Roman occupation of Gloucester, including a possible buried soil and alluvial deposits dating to the late first to second centuries. These were post-dated by a second-century wall footing. A cobbled surface dating within the second to fourth centuries corresponds with the location of an east–west street identified during excavations in the 1950s.Footnote ¹²⁷

SOUTH GLOUCESTERSHIRE

(1) Wickwar, Churchend Extension, Wickwar Quarry (ST 7142 9108): excavation recorded elements of the Roman road between Berkeley and Bitton (Margary route 541a), the course of which was identified by a prior geophysical survey and archaeological evaluation. Remnants of a cobbled surface and flanking ditches were identified as well as Roman or post-Roman limestone quarrying.Footnote ¹²⁸

SOMERSET

(1) Bridgwater, Bridgwater Tidal Barrier Scheme, Phase 2 (ST 2950 3990): a trial-trench evaluation comprised 241 trenches and 54 geotechnical test pits. A large number of archaeological features were present in the western part of the site (known as Chilton Trinity Borrow Pit), suggesting agricultural and settlement activity of Iron Age, Roman and post-medieval date. A handful of ditches in the central and southern part of this area contained later Iron Age pottery. A dense concentration of Roman and post-medieval ditches was recorded in the northern part of Chilton Trinity Borrow Pit. A fired clay bobbin was recorded in one of these trenches, and the same context contained Roman pottery dating to a.d. 240–410 and a late Iron Age/early Roman brooch. Disarticulated human bone was also recorded in two trenches in this western part of the site, indicating the potential for inhumation burials in this area. A handful of features were recorded within the western haul road area, including over 400 sherds of Roman pottery recorded in a buried soil. This suggests there may have been a second Roman settlement located in this area.Footnote ¹²⁹

DEVON

(1) Crediton Hamlets/Newton St Cyres, Roman road from Gunstone Cross to Half Moon Village (SX 778 996 to SX 888 975): interpretation of LiDAR and Google Earth (GE) imagery has identified a previously undetected section of the known Roman road (Margary 492a) north-west of Exeter. Previous work in 2019, building upon the observations of the late Hugh Toller, had allowed the road (believed to originate from a starting point at the mid-first-century a.d. legionary fortress at Exeter) to be mapped westwards from Gunstone Cross (SX 778 996) to the North Tawton military camps complex (SX 660 998) and beyond the River Taw to Trehill Farm (SX 636 972).Footnote ¹³⁰ From Trehill Farm the road appeared to be heading on a south-west alignment towards a point somewhere just to the south of the Okehampton Roman fort.Footnote ¹³¹ New work by Kaye has added a c. 9 km, mostly eastwards, extension of the road which follows a general alignment WNW-ESE from Gunstone Cross through Uton (SX 227 985) as far as Choldon Brake at SX 857 974 (fig. 30, panel A); from there the road turns eastwards to West Holm before returning to the WNW-ESE alignment as far as Half Moon Village (SX 888 975); this takes it to within c. 5 km of the Exeter fortress.Footnote ¹³² In addition to the LiDAR observations the road is visible in a field west of the River Culvery (SX 832 984) where GE imagery for 2018 shows a crop marked agger and ditches (fig. 30, panel B). The Roman road crosses the Culvery and continues upslope in a WNW–ESE direction. In fields centred at SX 836 983, the LiDAR shows an elongated depression. This is matched by GE 2018 imagery of crop marks of an agger, bounding ditches and quarry pits possibly used to supply material for the Roman road (fig. 30, panel C). Eastwards from Broomfield (SX 284 981), the Roman road probably underlies an extant minor road for c. 550 m. At the eastern end the existing road curves southwards, while the Roman road follows a largely terraced, WNW–ESE trajectory across fields south of Rudge Farm to Choldon Brake (SX 857 974). The GE imagery for 2020 provides evidence of the Roman road in a field centred at SX 852 975) with aligned and extensively elongated crop marks together with possible quarry pits (fig. 30, panel D). Eastwards from Choldon Brake, the Roman road changes trajectory, becoming easterly as it progresses to West Holm (SX 876 975) and the crossing of the Shuttern Brook. From the crossing of the Shuttern Brook the interpreted Roman road traverses upslope before crossing an elevated area of higher (c. 70 m AOD), level ground in the vicinity of Ridge Way House and Hayne Farm (SX 884 973). Here the remains of the road are shown by elongated and partially aligned depressions shown in the LiDAR data before the road arrives at Half Moon Village (SX 888 975) where it meets the modern A377.Footnote ¹³³

FIG. 30. Roman road from Gunstone Cross to Half Moon Village. Panel A: LiDAR location map of the Roman road (thin yellow line) from Uton to Choldon Brake. The thicker yellow lines B, C and D, are the locations for the Google Earth (GE) images in the panels below. Panel B: 2018 GE image of a field west of the River Culvery; crop mark of the agger. Panel C: 2018 GE image of a field To the east of Old Ridge Road showing the agger, bounding ditches and possible quarry pits. Panel D: 2020 GE image of a field to the south of Rudge Farm showing Roman road terracing, bounding ditches and possible quarry pits. (LiDAR data (National LiDAR Programme, 1 m, 2021) produced by the Environment Agency (UK) released under the Open Government Licence. The Google Earth imagery is covered by a fair use policy that applies to non-commercial applications)

CORNWALL

(1) Roche, (a) Trerank Farm (SW 9837 5956): evaluation in advance of a road project identified a Roman camp, first recorded through geophysical survey in 2018.Footnote ¹³⁴ Subsequent geophysical survey on areas outside the road corridor by Borlase and Wright has identified a rounded return at the north-eastern corner giving a total length of 425 m for the northern ditch (fig. 31). If the southern edge is of similar length, that gives a total area of around 11.7 ha for a camp, the size of which could potentially house a legion. Excavation undertaken as part of the road project identified the section of the camp within the road corridor, with the full western side of the camp ditch visible, measuring 275 m, with a 7.5 m wide gap visible slightly south of centre (fig. 32). The corners were rounded and the northern and southern arms were visible over distances of 124 m and 165 m respectively. Evaluation and excavation have confirmed the presence of a ditch, varying in depth but generally getting shallower towards the south. Where the ditch has depth an ‘ankle-breaker’ slot in the base of it is clearly visible. Excavation and exposure has also confirmed the rounded north-western corner and the presence of a 10 m wide entrance on the western side (fig. 31). The excavation has included only a thin strip of the interior of the camp. Few features that can be ascribed to the camp have been identified but these include one keyhole oven partially sealed by bank material just inside the ditch. A Roman-period burial was found cut across the southern ditch of the camp; due to the acidic nature of the local geology, only hobnails remained. The camp and other features of the prehistoric, Roman and later periods are intended to be published in monograph form at a future date.Footnote ¹³⁵

FIG. 31. Trerank. Geophysical survey results. Inset top left, location; inset top right, the camp (top) and fortlet (bottom) with contours and the River Fal (N.B. the heavily contoured hills to the south-west are modern china clay tips; that to the east of the camp is Roche Rock). (© Crown Copyright and Cornwall Council all rights reserved)

FIG. 32. Trerank. The western arm of the camp facing east towards Trerank Farm.

(b) Trezaise (SW 9874 5890): investigations related to the road project at Roche took place on a ditched rectilinear enclosure at Trezaise, 700 m to the south-east of the Trerank camp (see above and fig. 31). This enclosure was located over four modern fields and was constructed in at least two phases. The northern and north-eastern part of it lay beyond the limit of excavation. The first phase comprised a relatively small ditch enclosing a space 25 m (NNE–SSW) by 22.5 m (WNW–ESE). It was difficult to make out the exact shape of the enclosure in plan due to recutting but the internal south-western corner was rounded and the interpretation is that of a possible Roman fortlet, albeit the finds assemblage might suggest a native feature. No entrances were identified other than a small 0.6 m gap in the circuit to the north-west. Finds recovered from this ditch were limited to two base sherds of burnished Romano-British pottery and a base sherd of coarse ware. The ring-ditch of a post-ring roundhouse appeared to respect the enclosure ditch, the eastern terminal of its southern section running up to and stopping just short of it. This feature was associated with Romano-British burnished pottery and a sherd of Dressel 1 amphora, all of potentially first-century a.d. date, while sherds from Cordoned ware and Type 13 large storage vessels are likely to be later.

The second phase of the enclosure was marked by the construction of a larger and deeper ditch, up to 1.35 m deep, enclosing an area 25 m (NNE–SSW) by 24 m (WNW–ESE). In plan the enclosure was roughly square with rounded corners (fig. 33). The base of the ditch contained a vertical slot 0.45 m wide and 0.4 m deep in most sections. There was an entrance 3.5 m wide in the southern side, off-centre to the west; this had partly been formed by deliberately backfilling the earlier ditch with stony material in this area. Finds from the ditch included sherds of Romano-British pottery, including a Trethurgy Type 1 burnished vessel of late first- to early second-century date, a fragment of a stone hone, and a fragment of a stone mould that may be part of a mould for cold-hammering pewter dishes. Across the external side of the entrance, 0.7 m forward from it, was a sub-rectangular pit, 2.5 m long, 0.85 m wide and 0.8 m deep. This contained a burial from which a fragment of tooth from the western end and fragments of probable tibia from halfway along the eastern end were the only surviving skeletal remains. A piece of copper-alloy wire was found towards the western end and hobnails were found at the eastern end. A whole burnished pot that had been smashed during the backfilling process has been identified as a Trethurgy Type 2 jar of first- to second-century a.d. date. The pit appeared of unusual dimensions and shape for a grave and it has been speculated that the pit may have been first employed as a defensive ditch placed in front of the entrance (together with an accompanying rampart) forming a simple titulum. The possible fortlet and other features of the prehistoric, Roman and later periods are intended to be published in monograph form at a future date.Footnote ¹³⁶

FIG. 33. Trezaise. Overhead view of the possible Roman fortlet, looking north-east. The adjoining post-ring roundhouse is visible at top left. The grave/titulum is visible bottom centre, in front of the second phase entrance (photo: Cornwall Archaeological Unit).