Font Size

100% 150% 200%

Background Colour

Default Contrast
Close Reset


Temporary Camp (Roman)

Site Name Raedykes

Classification Temporary Camp (Roman)

Alternative Name(s) Garrison Hill

Canmore ID 37153

Site Number NO89SW 2

NGR NO 841 902

Datum OSGB36 - NGR


Ordnance Survey licence number 100057073. All rights reserved.
© Copyright and database right 2016.

Digital Images

Administrative Areas

  • Council Aberdeenshire
  • Parish Fetteresso
  • Former Region Grampian
  • Former District Kincardine And Deeside
  • Former County Kincardineshire

Archaeology Notes

NO89SW 2 841 902

See also NO88NW 23.

(Centred NO 841 902) Camp of Raedykes (R)

OS 6" map, (1927)

The plan of Raedykes marching camp is most irregular, to suit the terrain, but conforms to the orthodox rules in having opposed gates protected by traverses, though nothing convincingly Roman has been found. Much of most interesting feature is the strengthening of the defences in the direction of expected attack, namely, the two northern corners and most of the east side. Here the ditch, excavated by MacDonald, was V-shaped, about 15 feet wide and 7 feet deep; and the rampart was higher and broader than elsewhere. On the south and west Garrison Hill slopes rather steeply, and the defences were accordingly not so strongly made. This weakness may have been compensated by a formidable earth rampart running for 580 feet parallel to the south side at a distance of 800 feet south of it. It has a deep ditch on the southeast side now used as a drainage-channel. Crawford (1949) has no doubt that it is Roman although there is no evidence of its age. Crawford suggests that Raedykes camp was Agricolan in date and the most convincing of the many alleged sites for the Battle of Mons Graupius in AD 84. But Dr St Joseph suggests that it could be Antonine or Severan and places it in his 120-acre category, despite MacDonald's statement that its area is only 93 1/2 acres. Roman coin hoards have been found at Hill of Megray (NO88NE 3) and Cowie Common (NO89SE 2). In each case the latest coins were Severan.

O G S Crawford 1949; G Macdonald 1916; J K S St Joseph 1958.

Raedykes marching camp is generally as described above. From ground observations and Macdonald's measurements, the angle in the SE side occurs at the entrance "C" and not to the W of it, as shown on his plan (Fig.6) (Macdonald 1916). Similarly the SE angle has been repositioned. Resurveyed at 1/2500.

The outer earthwork is obscured by dense afforestation.

Visited by OS (NKB) 28 July 1965.

The outer work, referred to by authority 5, extends 190.0m from NO 8423 8955 to NO 8440 8961 and is situated in dense afforestation astride a ridge, some 230.0m S of the marching camp and parallel to the SSE rampart of the latter. It terminates at the NE end on marshy ground and at the SW end on the steepening slopes of the ridge. It comprises a strong earth-and-stone rampart with a ditch facing SE along which is a modern ditch. Two breaks occur in the rampart, of which the more westerly is either caused by, or accentuated by a drainage ditch. The position of the work relative to the marching camp, effectively blocking any approach along the ridge from a low hill to the SE, confirms that it is an outwork to the camp.

Revised at 1/2500.

Visited by OS (NKB) 29 June 1966.


J K S St Joseph 1969.

This temporary camp is situated on Garrison Hill; it is irregular on plan and encloses about 38 ha. Each of its six gateways was protected by a titulum and it probably dates to the first century AD.

G Macdonald 1916; 1939; O G S Crawford 1949; J K S St Joseph 1978; RCAHMS 1984, visited March 1984.


Field Visit (2 May 1957)

As Macdonald reports (1916, 318 f), there are two sizes of rampart and ditch used here—the smaller rampart being for all practical purposes a stone wall. Macdonald accounts for the existence of the two sizes by differences in terrain, but his argument breaks down on the north front where the ground facing the lesser rampart us just as much in favour of the attacker as elsewhere on this side. It seems probable the true explanation is that the smaller size of rampart represents unfinished work.

There are numerous shallow pits on Garrison Hill which, on the analogy of similar pits at Cawthorne [Cawthorn Camps, Yorkshire], may indicate the sites of tents, but otherwise the interior is featureless.

Visited by RCAHMS (KS) 2 May 1957.

Publication Account (17 December 2011)

The camp of Raedykes was first recorded by Maitland (1757: 202) and planned by George Brown and Barclay of Urie in 1778 (Barclay 1792; Roy 1793: Pl. L). It was the most northerly known for some time, and became a favoured site for the battle of Mons Graupius, with Kempstone Hill, 4km to the east, one potential location of the battle (Barclay 1792). Melville, however, did not favour the site for the battle (Gough’s edition of Camden’s Britannia 1790: iii, 416; Macdonald 1916: 319). It is worth noting that John Stuart thought that the camp was the Caledonian camp before the battle, being ‘an awkward imitation’ of the Roman one (1831: 300). Douglas notes that the tradition of the country was that a Scottish army lay in the camp, to oppose a Danish one encamped on the Links of Arduthie (1792: 261). Macdonald 1916 provides a useful summary of the historical sources for the site.

The camp is irregular in shape, and encloses Garrison Hill, with height-differences of over 40m between the highest and lowest points enclosed. It commands good views of the surrounding countryside, particularly to the sea at Stonehaven some 5km to the south-east, and is sited on the north side of the Cowton Burn, a tributary of the Cowie Water.

The camp measures around 663m from NNW to SSE by a maximum of 590m transversely and encloses 38.9ha (96 acres), if the projected line of the north-west corner is correct (having been taken from earlier sources).

The camp remains in a remarkable state of preservation, with the rampart and ditch clearly visible for much of the perimeter, owing to the site lying mostly under rough and improved pasture. At various points, the rampart measures up to 5.1m in width and 0.8m in height, with a 0.9m berm before the ditch, which measures up to 4.3m in width and 1.3m in depth. Tumbled stone in the ditch on the ENE side has led to suggestions that it may have been rock-cut in places, but this has not been confirmed. Entrance gaps with tituli (surviving as low swellings) are visible in the centre of the NNW side where the rampart angle changes – two on the ENE side and another in the SSE side; a further entrance gap is visible at a change of alignment in the southern part of the west side.

Excavations by Macdonald in 1914 identified that the ditch was V-shaped, about 4.5m in width and 2.1m in depth, faced with puddled clay up to 5cm in thickness, although was smaller elsewhere on the perimeter (Macdonald 1916: 332–4). He also suggested that where the ditch was smaller, the rampart was also smaller. Two of the tituli were recorded, the ditch of one measuring 18.6m in length by 3.4m in width and 1.5m in depth, the ditch of the other only 11.6m in length by 3.8m in width and 1.4m in depth (1916: 337). Historical finds from Raedykes include two Roman hastae (Douglas 1782: 261), a small hoop or ring of iron (Stuart 1831: 301) and a complete wheel (NSA 1845: xi Kincardine, 249–50).

A linear earthwork located some 230m south of the camp has been presumed to also be Roman in date and is associated with the camp as an additional defensive earthwork (Crawford 1949: 108–10; RCAHMS: NO 88 NW 23). However, this earthwork is of a rather different character and much larger in scale. It is also questionable whether the camp would have needed an additional outwork at this point. A medieval or post-medieval date for this feature seems more likely.

R H Jones.


MyCanmore Image Contributions

Contribute an Image

MyCanmore Text Contributions