Publication Account
Date 2007
Event ID 587621
Category Descriptive Accounts
Type Publication Account
Permalink http://canmore.org.uk/event/587621
NT76 1 EDINSHALL
NT/7724 6031
This broch-like stronghold, or large dun, inside an enclosure is associated with an older bivallate hillfort and with a later open settlement centred on a roundhouse (visited in 1989). Edinshall is included here because the round, drystone building has always been termed a ‘broch’ despite the several fundamental differences from true broch architecture which its design exhibits (see ‘Discussion’). The site was first explored in about 1870 by Turnbull [2] and has recently been re-examined, and its significance re-assessed, by Dunwell [6]. The following description is based on this new work.
Description
Sequence of occupations: the latest fieldwork at the site [6] appears to confirm the traditional sequence outlined at the beginning of this entry. Presumably the enclosure began to be constructed while the ‘broch’ was in use, since it provides a large area around the latter, and was gradually expanded. There is no obvious difference in age between its various parts, and no relevant evidence was revealed by the recent excavations.
The ‘broch’ lies within a sub-rectangular enclosure measuring about 58m north/south by 54m east/west. It is slightly oval in shape and the external diameter varies from 27.5 to 28.2m, the wall being from 4.8 to 6.4m thick. The central court looks distinctly elliptical on Dunwell’s plan. The outer face is vertical, in marked contrast to the sloping, stepped profile of tower brochs. The building stone appears to be local and the technique of construction – in which large blocks are underpinned by smaller slabs – is familiar from sites in Atlantic Scotland.
Edinshall is another example of a broch-like building which is known to have been better preserved in the recent past [6, 315]. At present the outer face stands to between 1.0 and 1.8m, though this is partly due to modern reconstruction. At the end of the 18th century, before the documented stone-robbing occurred, the wall is said to have been between 2.0 - 2.5m high (Turnbull 1857, 9). This should mean that, if there were once any raised galleries in the wall, that in Level 2 should have been partly preserved just over two centuries ago. On the other hand if the wall was never much higher than that it could have been solid.
The paved entrance passage – just under 5m long – faces just south of east and the outer part is 1.3m wide; at a distance of about 2.5m from the outside is the usual door-frame with rebated checks but no sign of any bar-hole or socket. The width then increases to 1.6m. There are two guard cells with rounded ends and with opposed doorways; the battering of the walls of the north (right) cell indicates that the roof was probably once corbelled, which would take the original wall height up to at least 2.5m and probably more. The doorway to this cell was once wider but has been narrowed by a block of unbonded stonework. The southern cell has been divided into two parts by another such block; its doorway has been walled off, probably during modern restoration. It once had a high sill [6, 316].
There are three intra-mural cells with doorways at 9.30, 11.30 and 3 o’clock respectively, all of which are approximately dumb-bell shaped with their doorways at the centre. The historical evidence for their once having been partly corbelled is somewhat contradictory [6, 316] though it seems very likely that they were once roofed with domes. The sizes of the cells, and various other dimensions, are given by Dunwell [6, 313, Table 1].
The cell at 9.30 contains the remains of an intra-mural stair at its right end; the left end is a distinct oval cell. Dunwell observes that the steps of the stair have little sign of wear. A secondary block of masonry projects from the outside wall, just to the right of the doorway, as if to narrow the approach to the stair. The original clearance by Turnbull failed to find any evidence of paving.
The cell at 11.30 has been divided into three interconnecting parts by two added blocks of masonry, one on either side of the doorway. Some paving was found by Turnbull, as well as a hearth in the middle of the left half.
The cell at 3 o’clock shows evidence for one-time corbelling at its ends and it too has been subdivided into two chambers by a secondary block of masonry opposite the doorway.
Shape of central court: in 1989 the author carried out an angle-and-distance survey of the inner wallface. It is clear that the inner face of the wall was not laid out using a peg-and-string compass; the shape is rather irregular and nothing like circular or elliptical. the maximum and minimum internal diameters are 17.2m and 18.3m according to Dunwell).
External structures. Immediately out-side the entrance is a series of added walls. The earliest is a grassed-over wall which curves round from the right side of the passage to form a porch protecting the doorway. A small rectangular building was erected on top of this, but its design does not suggest that it had anything to do with defending the entrance passage.
Finds [6, Illus. 20 and 21]
Included in this list are the probable Iron Age items from both Turnbull’s work and that of Dunwell [6, 332]. Nothing from the central court or the cells has a proper context except for the copper ingots. The numbers in brackets refer to Dunwell’s drawings.
Stone: 1 rough-out for a bead or pendant (1), 1 perforated disc of sand-stone (2); 1 ring or pendant (or hair ornament) of cannel coal (3). There are also some coarse stone tools such as a hammerstone (11), a whetstone (9), a whetstone/polisher (8) and a rubber (10). Some querns of unknown type are recorded by Turnbull but their present whereabouts is not known.
Glass: 1 armlet of opaque white glass, of Kilbride’s Type 3A, made of re-used Roman glass and dateable to the 1st and 2nd centuries AD (4).
Amber: 1 perforated bead (5).
Copper: 2 plano-convex ingots one of which is in the National Museum, the other being lost. Both were apparently found with a metal detector in 1976, buried 60cm deep under the floor of the southern cell, near the base of the intra-mural stair. The one in the National Museum was reported, unsurprisingly, as having been found well away from the site but Dunwell managed to contact the finder twenty years later and to establish the true find spot. Both ingots were found together but the second was obtained by an antiques dealer and is now missing. The dating of such ingots is discussed in detail [6, 339-40].
Bronze: 1 medieval ring-brooch (6) dated to the 13th century onwards: 1 decorative domed stud-head (7).
Pottery: two coarse sherds, probably of Iron Age type.
Discussion
Finds: apart from the two copper ingots the Iron Age finds are “common-place” and shed little light on the date or purpose of Edinshall, apart from suggesting an occupation in the Roman Iron Age and later. The suggestion is made that the ingots – derived from local copper mines – show how the inhabitants obtained their wealth and prestige. The second and third centuries were the time of the manufacture of the massive bronze armlets so characteristic of north-east Scotland and a considerable amount of copper would have been needed for them.
Geometry: The shape of the central court is irregular – in marked contrast to the surveyed circular brochs of the Atlantic Province. The standard deviations of the radii of the best-fitting circles for the central courts are very small – sometimes as little as 2% of the radius or less. It is clear that the ritual accompanying the setting out the shape of a new broch on the ground involved a carefully measured circle for the central court, possible using a standard unit of length and certainly using a peg-and-string compass. This did not happen at Edinshall and this fits well with the author’s impression that the structure is a late, outsize, crude copy of the broch tradition.
Terminology: The confusion produced by using the relatively new but vague terms ‘Atlantic roundhouse’ and ‘complex Atlantic roundhouse’ – the second of which the Edinburgh school wishes to substitute for ‘broch’ (in the sense of ‘hollow-walled tower broch’) – is well illustrated by Edinshall. This huge site is called a ‘broch’ in the title of the latest description [6] – even though it patently is not one. Yet the same excavator describes Durcha (NC50 2) – which is in the middle of the north mainland broch province (where almost no other kinds of small, round, thick-walled stone buildings are known) and is of the appropriate size and shape – as an ‘Atlantic Roundhouse’ in the title of the report on that site which immediately follows the one on Edinshall.
Sources: 1. NMRS site no. NT 76 SE 6: 2. Turnbull 1856 and 1882: 3. RCAHMS 1915, 60-4, no. 115: 4. Stevenson 1976, 49-50: 5. RCAHMS 1980b, 35, no. 190: 6. Dunwell et al 1999b.
E W MacKie 2007
