The Castles

NT81NW 11 82741 19199 to 82921 19090

(NT 82880 19180) Fort (NR)

OS 6" map, (1962).

Fort and Settlement, The Castles: An Iron Age fort, in part of which a settlement was later constructed, is situated on the NW end of a spur which overlooks the confluence of the Calroust Burn with the Bowmont Water.

The defences represent two distinct phases, but it is impossible to say with certainty which set of remains is the older as the two are nowhere in contact. For the purposes of description, and on the strength of reasonable probability, the inner enclosure is treated as being the earlier, but the possibility that the reverse is true must be borne in mind. What are called, with this proviso, the "first-phase" defences occupy the summit of the hill; they are formed of a single rampart and are D-shaped on plan - the chord running along the crest of the rocky ridge along the top of the spur and the arc embracing the upper part of the grassy slope to the E. The rampart (I on RCAHMS 1956 plan - see fig.439) encloses an area 250' NNW-SSE by 175'. A few facing stones still remain; they are of considerable size and suggest that the rampart was 10' thick at the base and was faced both within and without. On the N the steep hill-face was scarped away evenly below the outer face of the wall, to provide rubble for the core. There is an entrance-gap in the NW apex of the enclosure, and another gap, which may or may not be original, runs obliquely through the lowest part of the rampart on the E side. The ground within the rampart is scooped and quarried, especially on the N, but no signs of internal structures appear on the surface.

In the assumed second phase the fortified area was doubled by the building of two ramparts (IIA and IIB), of similar construction to the first but with external quarry ditches; they follow the first-phase rampart closely except on the SE, where they run on and take in an area as large again as the first-phase enclosure. The resulting multivallate fort is again D-shaped, with the chord now lying below the rock-face on the W side of the summit- ridge. Ramparts IIA and IIB are about 25' apart, and on the N, E, and S material from the external quarry-ditch of the outer rampart (IIB) has been thrown up to form a bank, IIC. While this bank is slight on the N and E, and fragmentary or non-existent on the W, it is of considerable size on the S where access is easy by way of the neck. Ramparts IIA and IIB are both better preserved than rampart I; on either face they show footings of boulders laid in courses, with a rubble core, and are respectively 11' and 7' in thickness. There are four entrance-gaps in these ramparts, all apparently original..

A minor phase of development is next to be seen in an earthen rampart (III), with a quarry-ditch inside its SW end. This runs in an arc across almost level ground from a point on the lip of Castles Craig to join the defences a few yards NE of the NW entrance, abutting on rampart IIB after overrunning bank IIC and the external quarry ditch of rampart IIB. There is a gap in it, in line with the NW entrance gap in the other defences. Its purpose was evidently to restrict the approach to the fort from the NW slope of the hill.

The very fragmentary remains of what was probably a small walled settlement, oval on plan, lie in the enclosed area to the S of rampart I. The wall, now much decayed, was about 4' in thickness; its N side abuts on the S side of rampart I, and the spread of its ruined SW part overlies the spread of the ruin of the rampart IIA. There is no gap in it to correspond with the SE entrance gap in the ramparts of the second phase, but what was probably an entrance lies at the N end of the W side. A few fragmentary foundations lie in the S half of the settlement and others are scattered about the E and S parts of the earlier enclosures.

The only finds recorded from the site are a stone axe (Rev Canon Walker 1899) and a number of "stone weapons" reported to have been collected by a shepherd, but none of these can now be traced.

RCAHMS 1956, visited 1948.

As described above.

Visited by OS (WDJ) 29 August 1960.

No change to previous field report.

Surveyed at 1:2500.

Visited by OS (EGC) 26 June 1968.

No change.

Visited by OS (TRG) 30 August 1976.

The extensive and complex remains of a sequence of fortification and settlement occupies the NW end of a long spur forming the watershed between, and overlooking the confluence of the Calroust Burn and the Bowmont Water. The spur upon which The Castles is set rises to a height of 349m OD and commands extensive views across both valleys. In both location and altitude, The Castles is similar to the site on Park Law some 2km to the E whilst some 1.8km to the W are the less tangible remains of the palisaded settlements on Craik Moor. The Castles commands the whole ridge upon which it is set and effectively controls the trackway complex that runs along the E side of the ridge eventually joining 'The Street' which crosses The Border at Windy Gyle.

Defences

The defences comprise, at least, three ramparts around the summit of the spur with a further rampart forming a secondary enclosure to the W. Relatively late activity on the site has effected considerable damage, obscuring critical points in any sequence and hindering analysis. Observations from survey do, however, permit an outline sequence to be proposed.

Phase 1; Initial enclosure

The original enclosure appears to have been D-shaped, measuring 80m from NW to SE by a maximum of 60m transversely, occupying the N or highest part of the summit area although it is conceivable that the whole of the area, measuring 135m by 60m, within the later multivallation, was enclosed at this time. The size and shape of the enclosure, so formed, was largely determined by the topography of the hilltop. Stone for the rampart was gained from quarrying in the immediate vicinity both external to, and from within the interior of the enclosure. There is no indication of an external ditch at this stage. A simple entrance, 4m wide, is situated in the NW corner of the enclosure. From here, the rampart extends E in good condition for 40m, past evidence of internal quarrying, only to disappear in a maze of later quarry scoops, post-dating any initial occupation, for about 20m. At this juncture a much slighter and somewhat mutilated bank, crosses the summit to the S end of a steep, sometimes rocky, W-facing scarp. Occupying the crest of this scarp are residual traces of a rampart which strengthen towards the N end as it approaches the gateway described above.

Phase 2; Possible expansion

The faced-stone wall that flanks the E side of the summit and possibly cuts across the S end of the primary enclosure, is very different in constructional character from its predecessor and may represent a strengthening of the defences on the E and S side as well as the possible enlargement of the enclosure to encompass the hilltop. Beyond the quarrying at the SE corner of the initial enclosure, there is a further 20m stretch of rampart, similar in character to that to the N of the disturbed area, adjoining a further possible gateway, again measuring 4m in width. From this gateway the rampart continues in reasonably intact condition around the S end of the enclosure, as far as the SW corner where it has been the object of intensive stone robbing leaving only a minimal kerb on its inner face and beyond which there are only the faintest traces.

Phase 3; Multivallation

Whilst the RCAHMS (1956) accepted, for the purposes of description and on the strength of reasonable probability, that the rampart sequence displayed at The Castles is one of expansion from the core, it is, nevertheless, stressed that the defensive sequence could, just as logically be reversed into a model of contraction.

The construction of a bi-vallate defence, with intermittent traces of a counterscarp bank over the gentlest approaches on the N through to SW sides would appear to follow an expansionist model. On the W side, the bi-vallate rampart diverges from the initial enclosure line to pursue a course below a steep natural scarp while the counterscarp is replaced by a another natural scarp as far as a gateway. The gateway, which measures 4m in width, is of complex form with both ramparts linked by returning banks flanking the passageway; suggesting that the two banks are contemporary. North of the gate, the bi-vallate ramparts continue to the NW gateway and some 15m before this gateway the counterscarp bank is re-established on the more vulnerable approaches. Like the initial enclosure rampart, the inner rampart of the bi-vallation has been heavily robbed, particularly in proximity to the later settlement on its SE flank.

Phase 4; the Annex enclosure

A large area defined by the Castles Craig on the W and the fort itself on the E was enclosed by the construction of a further stone and earth bank, with an internal shallow quarry ditch across the N end of this area. The bank commences above the steep ground at Castle Craigs and continues across gently-rising ground for 200m to cross the counterscarp bank of the outermost (phase 3) rampart of the fort 10m to the NW of the NE gate. The bank is pierced by a single 5m wide gate in line with the aforementioned entrance. Whilst the RCAHMS (1956) show and describe this bank as crossing the outermost (or counterscarp) bank (their IIc) and external ditch of rampart IIb to abut rampart IIb, this survey could not establish such a clear-cut relationship.

The occupation sequence

Occupation within the fort occurs within the interior of the primary summit enclosure. There is no evidence for occupation spreading onto the denuded ramparts, unlike at Park Law (NT81NW 157), until the construction of the late settlement at the S end. Several types of building are represented. There are at least ten platforms within the NE part of the enclosure and further S, between the primary enclosure and the late settlement, a single ring-groove house was recognised during the 1985 survey. Nearby there are four other 'stances' of similar dimensions though all lacking visible traces of 'grooves'.

Platforms

The platforms are all confined to the area defined by the initial enclosure. They are predominantly NE-facing, levelled into the summit ridge along 'the grain' of the hillside and, in some cases, occupying broader quarry scoops behind the initial rampart on the E side. They average 8m in length from NW to SE by 5m.

Possible ring-groove house

Within the N part of the initial expansion there is a possible ring-groove house measuring 6.4m in diameter within a faint groove. There is a possible entrance on the E. In the open space to the E of this house, there are up to four more stances, with an average diameter of 6m but none show any evidence of for ring-groove bedding trenches.

Settlement

To the S of these stances, there are the fragmentary remains of a later settlement whose internal plan is difficult to unravel. The settlement occupies the S end of the fort, in an area roughly corresponding to that taken in by the expansion of the primary enclosure From the NE angle of this settlement, where there is a possible entrance measuring 4m in width, the enclosing wall, averaging 1.5m in thickness, follows and overlies the innermost rampart (dating to the secondary expansion to enclose the whole hilltop) of the earlier fort to just S of the SW angle of the primary hilltop enclosure. Here the settlement utilised the S bank of the primary enclosure until the angle midway along its course where it turns NE. From here a short stretch of bank turns SE to return to the gateway described above. Internal banks are restricted to the S part of the enclosure. They form a series of four subrectangular 'cells' ranged around the E, S and W sides of a larger central 'enclosure'.

Information from RJ Mercer (University of Edinburgh) 1 April 1985

RCAHMS MS 2598. No. 34/449