Font Size

100% 150% 200%

Background Colour

Default Contrast
Close Reset


Rig And Furrow (Roman), Roman Fort (Roman)

Site Name Oakwood

Classification Rig And Furrow (Roman), Roman Fort (Roman)

Canmore ID 54330

Site Number NT42SW 1

NGR NT 4250 2491

Datum OSGB36 - NGR


Ordnance Survey licence number 100057073. All rights reserved.
Canmore Disclaimer. © Copyright and database right 2021.

Toggle Aerial | View on large map

Digital Images

Administrative Areas

  • Council Scottish Borders, The
  • Parish Kirkhope
  • Former Region Borders
  • Former District Ettrick And Lauderdale
  • Former County Selkirkshire

Archaeology Notes

NT42SW 1 4250 2491

For adjacent temporary camp, see NT42NW 12.

Roman Fort, Oakwood. The only Roman works which have so far been identified in the county - the fort and adjacent temporary camp at Oakwood, on the S. side of the Ettrick valley, 3 1/2 miles SW. of Selkirk were discovered in 1949 during a routine examination of the National Survey air-photographs in preparation for the present (1957) Inventory. Under the direction of two of the Commission's officers, excavations were carried out on the defences of both works in 1951-2, and the following account is a summary of the published report. (K A Steer and R W Feachem 1954)

The fort occupies the flat top of a small boulder-clay knoll lying on the S. side of the minor road from Oakwood Mill to Inner Huntly.

The N and E flanks of the knoll fall sharply to a nameless burn which threads its way, in a marshy gully, between the two works, while the W and S slopes, though more gently inclined, are sufficient to ensure immediate dominance of the situation.

To N and W the outlook is good, and it is significant that to the E the tops of the Eildon Hills, 9 miles distant, can be seen in a depression in the ridge that otherwise blocks the view from the fort on this side. Direct communication between the forts at Oakwood and Newstead would thus be possible by means of the signal-post recently identified on the summit of Eildon Hill North. (K A Steer and R W Feachem 1954; RCAHMS 1956)

Apart from a small plantation of conifers, approximately in the centre of the fort, the site is at present in rough pasture; and, although the defences have been severely mutilated by former cultivation, they are still discernible on the W side and on parts of the N and S sides. On the S. side, too, the existence of an annexe is revealed by a single ditch which runs southwards for 200ft from the SW corner of the fort and thence eastwards as far as the boundary wall between Oakwood and Hartwoodmyres.

Excavation showed that the fort was founded at the time of Agricola's advance northward, probably in AD 81, and was abandoned some twenty years later, but as at Newstead (I A Richmond 1952; RCAHMS 1956) two structural phases could be distinguished within this unbroken period of occupation. It will be convenient to describe each of these phases separately.

Phase I. In its original form the fort faced E, was square on plan, and covered 3 1/2 acres within the rampart - an area appropriate for the accommodation of an infantry cohort, 500 strong, reinforced with a mounted detachment (cohors quingenaria equitata).

The rampart, 18ft to 23ft thick, was built of turf, and in front of it on every side lay two ditches, each about 13ft in width and 3ft to 4ft in depth. The S half of the W side of the fort was further protected by an isolated length of a third ditch 12ft wide and 3ft deep.

All the four gates were of timber, and were basically of the same, highly unusual design. Each was set back about 30ft from the front of the rampart, at the narrow end of a funnel-shaped passage between the rampart terminals, and comprised twin portals placed side by side and flanked by square towers.

At the W gate (porta decumana) the towers were not closed structures but simply frameworks for fighting-platforms of the kind depicted on Trajan's column: they measured respectively 10ft 9 in (N) and 11ft 4 in (S) in width, by 10ft 6 in in length, while the portals, the southernmost of which had been blocked up before the gate was used, were each 9ft 6 in. wide. The oak stumps of six of the ten main uprights of the gate were found still in situ, together with two erection-pits for the derricks which had been employed to raise these timbers into position. A conjectural restoration of the original aspect of this gate is given in Fig. 97. (RCAHMS 1957)

Only one other gate, that on the E side (porta praetoria), was examined in detail. Its dimensions closely corrsponded to those of the W gate, but whereas the S tower was again an open framework, the N tower had an enclosed basement which presumably served as a guard-chamber. The stumps of five main uprights were recovered from this gate, and two derrick-holes were again present in positions corresponding to those at the W gate.

Stratified pottery and a somewhat worn denarius of Vespasian minted in AD 69-70 associated the erection of the fort with the Agricolan invasion of the Lowlands in AD 80, while the pottery suggested that, as at Newstead, the end of the first structural phase came shortly after AD 86.

Phase II. In this phase the layout of the fort remained the same, but the defences were reconditioned - the weathered front of the W rampart being refaced with clay and the roads resurfaced. The W gate was now strengthened by the construction of a clay baffle-wall across the S. half of the entrance-passage, and a clay buttress was added to the face of the rampart on the W side of the E gate.

The precise duration of the occupation following this overhaul is uncertain, but as the pottery from the site include nothing that need be later than the end of the 1st century AD, it can be assumed that the fort was evacuated about AD 100, when, probably owing to preparations for Trajan's Dacian Wars, the Roman troops in Scotland, with the possible exception of the garrisons in lower Annandale, were withdrawn to the Tyne-Solway isthmus. The appearance of burning on the timbers from the E and W gates is also consonant with the evidence from Newstead, where the second phase of occupation ended in disaster accompanied by a general conflagration, but unlike Newstead the fort at Oakwood was seemingly not rebuilt in the Antonine period.

THE ANNEXE. The annexe, which covers an area of a little over one acre in extent, was defended by a rampart, 19ft thick, composed of ditch upcast revetted with cheeks of puddled clay, and by a ditch 16 ft. wide and 3ft deep. An entrance, 40ft wide, can be seen on the S side, close to the SW corner, but it remains unexcavated.

An examination of the junction between the ditches of the fort and the annexe suggested that the annexe was not laid out at the same time as the fort but was added at a later date, although it does not necessarily follow that it belongs to the second structural phase observed in the fort defences.

RCAHMS 1957, visited 1952

Full reports and discussions on excavations and discoveries, which may be summarised in the following;

K A Steer and R W Feachem 1953; W Roy 1793; I A Richmond 1932; 1939; 1948; 1952; J K St Joseph 1947; J P Gibson and F G Simpson 1909; F G Simpson, I A Richmond and J K St Joseph 1935; I A Richmond and K S Hodgson 1936

NT 4250 2491. The fort is generally as described by the RCAHMS. There is now little evidence of the rampart which defended the fort, it having been ploughed down by former cultivation. As for the ditches, in no place does the depth exceed 0.3m. The only entrances discernable are in the N and W sides of the fort. The rampart which defended the annexe is evident as a slight mound, 0.1m high, on the W. and part of the S sides only. The ditch around the annexe reaches a maximum depth of 0.5m near the junction with the fort at the NW corner: there is a mound, 0.4m high, on the counter-scarp of the ditch on the S side. Old plough-lines cross the fort in a NE/SW direction while within the annexe they run NW/SE.

Surveyed at 1/10560.

Visited by OS (EGC) 27 June 1962

As previously described.

Surveyed at 1/2500.

Visited by OS (WDJ) 27 April 1965

Discrepancies between the two AO OS surveys are due to the tolerances relative to the two scales.

Visited by OS (EGC) 29 April 1965.


Publication Account (1985)

Though little more than a flat-topped, gentle grassy knoll, the site is sufficiently elevated to dominate the surrounding countryside. The outlook is particularly good to the west and north; to the east the tops of the Eildon Hills some 14-15km distant are visible in the depression in the ridge. Given the signal-post on top of Eildon Hill North (no. 84), communication was possible direct from Oakwood to the Roman nervecentre of New stead.

The fort was established cAD 81 during Agricola's northward advance. During this occupation there were two phases of building, the second mainly reconditioning the defences of AD 81-86 which consisted of a thick turf-built rampart surrounded by two ditches. These are clearly visible on the west side and on parts of the north and south sides. Adjoining the south side, evidenced by a single ditch, was an annexe running over 60m south from the south-west corner and thence east to the modem drystone dyke dividing Oakwood and Hartwoodmyres farms.

Though part of the main street running north-south is covered by a small plantation, all four gateways can be identified, set back over 9m from the front of the rampart at the narrow end of a tunnel-shaped passage. All four gates were of timber and comprised twin portals placed side-by-side, flanked by square towers-an unusual design. The towers at the west gate were not enclosed; they were simply frameworks for fighting platforms similar to those shown on Trajan's column in Rome.

A short distance away, across the road, traces survive of a temporary camp (NT 425255), probably built to house mainly labourers working on the fort.

It seems likely that Oakwood was abandoned cAD 100 when most Roman troops in Scotland were withdrawn to the Tyne-Solway line. Given burning on the timber stumps of the east and west gates, identified during excavation, the occupation may have ended with a disastrous (or deliberate?) conflagration, as at Newstead. Unlike Newstead, however, Oakwood was never rebuilt.

Information from 'Exploring Scotland's Heritage: Lothian and Borders', (1985).


MyCanmore Image Contributions

Contribute an Image

MyCanmore Text Contributions