CARMARTHENSHIRE

(1) Ffarmers, Bryn Dolau (SN 65006 44310): an archaeological evaluation trench identified a section of the surviving upper road surface of a Roman road as well as associated ditches flanking each side of the road. The road clearly extended beyond the extent of the excavation trench. Unfortunately no dating evidence was recovered but the site lies directly on the line of the Sarn Helen Roman road leading north from Pumpsaint.Footnote ¹

CEREDIGION

(1) Abermagwr (SN 6688 7416)Footnote ² : two trenches were opened to the west of the villa building. One trench, placed diagonally to the west of the house to sample the courtyard in that area, produced very few finds beyond abraded sherds of pottery, brick and tile fragments and pieces of broken stone roof-tiles. No deliberately surfaced courtyard horizon was reached to match the cobbled yard close to the villa; however, lower layers of hard-packed clay and gravel with an undulating surface flecked with charcoal, burnt stone and sparse artefacts suggested the likely character of the Roman farmyard here. No building footings were encountered, beyond four putative stake-holes encountered in the lower subsoil. The other trench across the western perimeter of the bivallate villa enclosure produced a section of both the inner (east) and outer (west) ditches. These too were largely devoid of finds. The inner (east) ditch, however, yielded a complete Roman roof-slate laid flat in the upper fills, perhaps used as a ‘step’ across the infilled ditch during the robbing of the villa in the fourth century. The slate is the most perfect specimen from the entire site. A crude ‘maul’ in the form of a heavy spherical river cobble with pronounced areas of pecking was recovered from the lowest fill of the ditch, perhaps indicative of the more mundane tasks of a working farm. The lowest fill of the outer (west) ditch yielded a lump of charcoal comprising both oak stem wood and hazel charcoal. The latter was successfully radiocarbon dated, providing a date of 145–382 cal a.d. at 95.4 per cent probability, 211–382 cal a.d. at 92.4 per cent probability (SUERC-65621). This is a later date range than that previously obtained from the inner ditch, but one tallying with the acceptable date range as currently understood for the villa.Footnote ³

ISLE OF ANGLESEY

(1) Cemaes, Rhyd y Groes wind farm (SH 39281 92917): trial-trench evaluation confirmed the presence of archaeological features shown on aerial photographs and the results of a previous geophysical survey. No artefacts were discovered; however, it is possible to broadly date some of the features recorded on the basis of their form or on their likely relationship to a known complex of Iron Age and Roman enclosures at the eastern end of the site. The environmental samples broadly corroborate these conclusions.Footnote ⁴

(2) Cemlyn, Roman fortlet (SH 3251 92451): a fluxgate gradiometer survey revealed a Roman fortlet, 45 m by 50 m in size, with strongly rounded corners and a slightly in-turning entrance in the centre of the north side ( fig. 2). Traces of several rectangular buildings, probably barracks, were detected in the interior. The fortlet is surrounded by a slightly meandering sub-circular ditched enclosure, 74 m in diameter. Parallels for this can be drawn from Old Burrow and Martinhoe fortlets on the north Devon coast.Footnote ⁵ Reports of late first- to early second-century finds from the site support the identification of this as the first military site from the immediate post-conquest period to be identified on Anglesey.Footnote ⁶

FIG. 1.

Key to numbered sections of Part I (Sites Explored).

FIG. 2.

Fluxgate gradiometer survey of Cemlyn Roman fortlet. (© Gwynedd Archaeological Trust)

MONMOUTHSHIRE

(1) Caerwent: (a) Coach and Horses (ST 47074 90491): an archaeological evaluation revealed a Roman foundation wall with possible facing stone and plaster only a few hundred millimetres beneath the ground surface.Footnote ⁷

(b) Myrtle Cottage (ST 46841 90611): an archaeological evaluation comprising three trenches identified significant and generally well-preserved archaeological deposits. The majority of the features were of Roman date and included deeply stratified and, in places, complex soil deposits, as well as a possible stone-built feature, a possible drain, a number of substantial cut features, numerous pits and a cobbled surface. Roman artefacts recovered included relatively unabraded pottery, which spanned the first to fourth centuries, roofing-tile, a flat tile and at least one piece of box-flue, fragments of painted wall-plaster, fragments of glass, metal objects, including iron nails, and animal bones. The Roman features and finds are entirely typical of those to be expected within a Roman town such as Venta Silurum.Footnote ⁸

NEWPORT

(1) Caerleon: (a) Bulmore House (ST 35996 91491): an excavation to the top of the archaeological resource revealed important Roman remains dated from the late second century to the early third century. The interpretation of the features was limited by the extent of the excavation, but the work confirmed that the Roman occupation extended to the south of the area studied in the 1980s.Footnote ⁹ The features recorded comprise two drains and two large pits that were only partially excavated. A small assemblage of 222 sherds of pottery was recovered and consisted mainly of black burnished ware (BB1), South Wales grey ware, oxidised fabrics, Dressel 20 amphorae, mortaria and a few sherds of mainly South Gaulish samian. A silver denarius of Faustina II struck in a.d. 161 and a small copper-alloy hollow-cast female head, probably part of a goddess statuette, were also recovered. The Roman cut features could be interpreted as pits probably associated with domestic settlement or as industrial features associated with a workshop or industry.Footnote ¹⁰

(b) 33 Goldcroft Common (ST 33805 90853): an archaeological watching-brief observed stratigraphy consisting of topsoil overlying reddish brown silty clay and bedrock. A possible Roman layer of foundation river-bed cobbles was revealed and a small amount of abraded Roman red ware pottery was recovered from the heavily disturbed topsoil. The observed features were suggestive of the possible remains of Roman barrack block foundations.Footnote ¹¹

POWYS

(1) Walton, Hindwell Ash marching camp (SO 2548 6098): a magnetometer survey was conducted in two adjacent fields to the north of the lane leading to Four Stones (centred at SO 2553 6112 and SO 2575 6109 respectively) to investigate the potential continuation of the western ditch and prospect for the northern side of the putative marching camp.Footnote ¹² The results from this survey, together with a small-scale excavation which examined a section of one of the ditches, cast doubt on the interpretation of these features as part of a marching camp. Although the putative western ditch of the camp is now known to extend to the north of the land to Four Stones, the small-scale excavations demonstrated the slight nature of the feature which, at only 0.6 m wide and 0.28 m deep, is perhaps too shallow be associated with a Roman military structure.Footnote ¹³

VALE OF GLAMORGAN

(1) Wick (SS 91 71): a hoard of 91 silver denarii, deposited a.d. 164 or later, was found by use of metal-detectors.Footnote ¹⁴ The date range of the coins was principally from Nero (a.d. 54–68) to Marcus Aurelius/Lucius Verus (a.d. 161–80); the latest coin was struck in a.d. 163–4. The hoard also contained three coins issued by Mark Antony in 31 b.c. The hoard was contained in a pottery jar of local manufacture.Footnote ¹⁵