Bush Farm Roman fort, Y Felinheli. NPRN 801081

Discovery

Cropmarks of a plough-levelled Roman turf and timber fort were first identified on drought vertical aerial imagery on BingMaps (probable capture date: summer 2020) by Toby Driver of the Royal Commission on 29th September 2024. Additional information from Welsh Government LiDAR confirm plough-denuded earthworks matching the cropmark information. A field inspection was conducted with the permission of the landowner by Toby Driver on 20th February 2025.

The location and siting of this early Roman fort suggests it may have been used as a springboard for the Roman invasions of Ynys Mon/Anglesey, either under Paulinus in AD 60 or under Agricola in AD 77-79. 

Setting

The Roman fort is sited at the south-west end of a prominent ridge of high ground which rises above Y Felinheli with commanding views over the Menai Strait. The fort is sited at 80m O.D. on a north-west facing hillslope which 'tilts' the site to overlook the estuary. A ridge of outcropping rock otuside the fort to the east, at SH 528 671, would have provided a naturally prominent 'look out' for military purposes.

Site description

Cropmarks show the double or triple ditches and rounded 'playing-card' corners of a likely Roman fort sited at 80m O.D. on the north-west facing coastal hillslope above Y Felinheli, overlooking the Menai Strait. The fort is defined on the south-west side by clear double-ditches of Roman military character, sited 5m apart, with a spread earthwork bank across them visible on LiDAR; this stands to 0.2m on the ground. The rounded southern corner of the fort is visible as cropmarks, and the line of the defences here is partly fossilised by a modern hedgebank which stands 1.25m high. The rounded southern corner appears to be preserved as a 0.3m high earthwork under bushes with several large boulders placed on it.

The rounded east corner of the fort is also visible as cropmarks, along with the returning ditches of the north-eastern defences which comprise triple ditches spanning a width of 12m.  LiDAR shows the survival of parallel earthwork ramparts here; these survive as shallow earthworks 0.1m high. The north-western side of the fort has been destroyed by the construction of the A487 Y Felinheli Bypass (opened 1994; see below). If the fort was originally square, it would measure 156m across externally, 138m internally, with an overall internal area of approximately 1.7ha and external area of approx. 3.7ha.

The LiDAR data shows the potential continuation of the outer ramparts of the north-east defences as a longer bank running some 140m downslope to the south-east, part of a possible annexe to the fort, although this evidence is currently more questionable.

Evaluation excavations 1991

The plough-levelled north defences of the Roman fort were evaluated during archaeological works ahead of the 1991 construction of the Port Dinorwic (now Y Felinheli) bypass, by the Gwynedd Archaeological Trust and Bradford Geophysical Surveys. However they were not recognised at the time, and were recorded as ditches relating to a recent hedgebank. The text describes the triple ditches in Field 9 thus:

‘The geophysical survey identified archaeological features running parallel to the features on the ground, which have been partly surveyed using a total station. The three linear parallel features shown in the geophysics plot on the north side of the field have been identified on the 3rd edition (1916) os 25 11 map as a field boundary. The boundary does not appear on the 2nd edition (1900) 25 11 map and was probably put up in WW1’ (Gwynedd Archaeological Trust 1991, 8).

The triple ditches were excavated in Trench F:

‘Trench F was machine dug to a depth of 0. 5m, with two ditches excavated to a depth of 2m… The most prominent features in the trench were two steep "V" shaped ditches. Ditch 1 was 1. 7m deep and 2. 4m wide at the top. There was a small slot in the base of the ditch filled with redeposited natural, above which was a fill of grey turves. Ditch 2 was 0.8m deep and 2.8m wide at the top. It was 2.2m north of and running parallel to ditch 1. The fill contained traces of turves but far less than ditch 1, the rest being dark grey-brown stone-free silt. A third possible ditch lay 3m north of ditch 2. The three ditches · were picked up by the geophysical survey. As mentioned above, there is a field boundary shown on the 3rd edition 25" OS map (1916) that corresponds exactly with the excavated ditches.· The boundary would have been of turf, earth and possibly wood and ditched on either side. The ditch on the outer side was deeper and steeper and when the boundary was removed it was simply dumped into the ditch that had been dug to create it. There is no trace of the ditch on the surface and it is not respected by the visible remains of the field system on the surface. This suggests that the cultivation marks post date WWl, and pre-date 1955, when the farm was bought by the present owners. No further excavation is required, although a complete survey of the features within the line of the bypass is needed. Aerial photographs in low sunlight would be ideal’. (Gwynedd Archaeological Trust 1991, 14-16, and Figs. 4 & 5).

With the benefit of hindsight, it appears that the triple ditches section in Trench F were indeed the substantial remains of the triple ditches of a Roman fort, the inner two being V-shaped, between 2.4m-2.8m wide and between 0.8m-1.7m deep. The most substantial inner ditch was steeply-sided and V-shaped with an ‘ankle-breaker’ slot at the base; it was full of redeposited turves. All these findings are consistent with a Roman defensive ditch. Unfortunately the much-discussed ‘field boundary’ shown on 3rd edition OS mapping is a single bank and ran at a different alignment to the fort ditches.

Field visit, Toby Driver RCAHMW, 19 Feb 2025

Reference:

Gwynedd Archaeological Trust 1991. Port Dinorwic Bypass Archaeological Assessment. Unpublished report.

Toby Driver, Royal Commission, 30th September 2024; updated February 2025