Font Size

100% 150% 200%

Background Colour

Default Contrast
Close Reset


Roman Fort (Roman)

Site Name Cappuck

Classification Roman Fort (Roman)

Canmore ID 57050

Site Number NT62SE 39

NGR NT 6950 2123

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 Oxnam
  • Former Region Borders
  • Former District Roxburgh
  • Former County Roxburghshire

Archaeology Notes

NT62SE 39 6950 2123.

(NT 6950 2123) Roman Station (NR) (Remains of).

OS 6"map, Roxburghshire, (1923).

Roman Fort, Cappuck (Site). The Roman fort at Cappuck, of which nothing now remains visible on the surface of the ground, stood on the E bank of the Oxnam Water, overlooking the point where Dere Street crossed that stream, 11 miles SE of Newstead. Since there is no good outlook in any direction the fort must be regarded as a bridgehead post. The site was first excavated in 1886 by Walter Laidlaw (W Laidlaw 1894; 1900) for the Marquess of Lothian, when a stone-built granary and other buildings were exposed and a neat building-inscription of the Twentieth Legion, now in the National Museum of Antiquities, was recovered. In 1911 much more extensive excavations (G H Stevenson and S N Miller 1912), were undertaken by the late Messrs S N Miller and G H Stevenson, resulting in a first definition of the defences. These were further examined (I A Richmond 1953), in 1949, when ditches disclosed for the first time by air-photography were tested and related to previous discoveries. The result, as at Newstead (see RCAHMS 1956 No. 604), was to reveal four stages of occupation two Flavian and two Antonine.

(i) The Agricolan fort, dated to this period by Samian ware and by association with Newstead, measured 290ft from N to S by 218ft from E to W over its rampart, 23ft 6in thick, of beaten clay. It contained timber buildings, but their plan has not been recovered in any detail. The rampart, which was underpinned by cobbling at the front, was protected by a single V-shaped ditch, 12ft wide and 4ft 9in deep.

(ii) The late-Domitianic fort, as at Newstead (No.604), was strengthened by a frontal thickening of the Agricolan rampart carried upon a new strip of underpinning, 5ft 6in wide. This overlapped the inner lip of the Agricolan ditch, which was now filled up and obliterated, its place being taken by a new ditch 16ft wide, set so close to the old one as to cut its outer lip on the S side of the fort and to cut deep into it on the E side, so that the two features, though actually distinct, seem to merge with one another. The effect of this change, as at Newstead, was presumably to give greater height to the rampart. Internal buildings are not yet defined, except for two isolated square erections, reminiscent of those in the fortlet of Haltwhistle Burn,(J N Gibson and F G Simpson ) Northumberland. No evidence for the duration of this post-Agricolan occupation is forthcoming at Cappuck.

(iii) The early-Antonine fort, to which doubtless belongs the elegant building-inscription of the Twentieth Legion, took for the line of its ramparts the N, E, and S sides of the late-Domitianic fort, but rebuilt the front completely on a new underpinning of cobbles which, on the E, coincided with the older front but on the S lay a little outside it.

The W rampart was on new lines and lay beyond the earlier ditch, so that the fort now became 303ft by 260ft over its ramparts. On all four sides, therefore, the late-Domitianic ditch was filled and obliterated, and two new ditches were dug on all sides, their outer lips at a distance of 57ft 6in and 73ft from the new rampart front. The outer ditch was W-shaped and probably intended to hold a hedge-like obstacle. The internal buildings comprise the substantial granary discovered by Laidlaw in 1886, (W Laidlaw 1894), courtyard building resembling a commandant's house, a little two-roomed bath-house (like that of Castleshaw II), and at least one barrack-block. There is no headquarters building of the usual kind, as if the fort were a convoy-post rather than the quarters of a garrison of normal type. Certainly its size would not accommodate a cohort.

(iv) The later-Antonine fort is marked by a new N rampart, which cut off, as at Ardoch, one end of the early-Antonine defences so that the fort now measured 252ft by 260ft over the ramparts. The space thus disused was intersected by a new ditch, lest it should serve as the rallying-ground for an attack. On the S side the outer Antonine ditch was obliterated while the inner Antonine ditch and the filled Domitianic ditch were re-dug. On the E side of the fort, which faces uphill, two extra ditches, making four in all, were added in order to prevent the defences from being rushed. The new N rampart was a broad one, and there is evidence that the older ramparts on the other sides were enlarged at the back to match it, while a new intervallum road and at least one new barrack-block was supplied. The courtyard building was also rebuilt and to some extent re-planned. The date of these alterations is unknown, but below both the new rampart and the new intervallum road there occurred Samian vessels of the Antonine potter Cinnamus, whose peak period of output is about AD 160. The fort is thus likely to be later than the troubles of AD 155-158. The development of the fort as a whole thus closely resembles that of the fort at Newstead (No.604).

Inscribed Stone:

A portion of an inscribed stone of some interest was found at Cappuck in 1886 and is now in the National Museum of Antiquities, Edinburgh. It is the dexter side of a dedication tablet from a building erected by the Twentieth Legion, attested by the first figure of its numeral which survives on the penultimate line of the stone. The die containing the inscription is bordered by a cohort-standard, decorated with three paterae and a wreath, and erected upon a base. Beyond this comes a decorative pelta, taking the place of an ansate border; the surface of the pelta is adorned with two large leaves and its terminals end in spiral scrolls. Below the pelta a boar canters to the right, while the field above it is occupied by a bird pecking at a flower with two blossoms. The style of the stone quite certainly connects it with the first Antonine fortlet. The surviving fragment of the stone is 12in high by 8in wide.(Ephemeris Epigraphica, F Haverfield [ed.] 1913). The text can be restored (RCAHMS 1956, fig.494) as LEG. XX. V.V. FECIT, the upright stroke of the F still remaining.

Native Palisade:

The section cut across the S defences of the fort in 1949 disclosed the existence of a curvilinear palisade-trench, which had been obliterated when the fort was built. The palisade-trench was 3ft wide by about 1ft 6in deep and, like the palisades at Hownam Rings, Hayhope Knowe and elsewhere (RCAHMS 1956, cf., 19 ), presumably enclosed a pre-Roman native settlement.

RCAHMS 1956, visited 17 September 1949.

No remains. The fence shown on the RCAHMS plan is incorrect, the OS 25" being correct.

Visited by OS(EGC) 8 February 1967.

This is a Roman fortlet, not a fort

Transcript amendments to RB4 J K St Joseph 8 April 1978.


Sbc Note

Visibility: This site is visible as a cropmark.

Information from Scottish Borders Council


MyCanmore Image Contributions

Contribute an Image

MyCanmore Text Contributions