Font Size

100% 150% 200%

Background Colour

Default Contrast
Close Reset

Keiss, Kirk Tofts, 'road' Broch

Broch (Iron Age)(Possible)

Site Name Keiss, Kirk Tofts, 'road' Broch

Classification Broch (Iron Age)(Possible)

Alternative Name(s) Wic 108; Keiss West

Canmore ID 9333

Site Number ND36SW 1

NGR ND 3488 6151

Datum OSGB36 - NGR


Ordnance Survey licence number AC0000807262. All rights reserved.
Canmore Disclaimer. © Copyright and database right 2024.

Toggle Aerial | View on large map

Digital Images

Administrative Areas

  • Council Highland
  • Parish Wick
  • Former Region Highland
  • Former District Caithness
  • Former County Caithness

Archaeology Notes

ND36SW 1 3488 6151.

(ND 3488 6151) Kirk Tofts (NAT) Broch (NR)

OS 1:10,000 map, (1976)

The Road Broch is one of the best examples of a 1st phase broch (1st centuries BC and AD), although it was re-used during the Broch II phase (2nd, 3rd centuries AD) and again during the post-broch era. It was built on a mesolithic kitchen midden.

The dimensions of the original broch appear to have been 34ft internal diameter, with walls only about 12ft thick.

The second phase of occupation involved external strengthening of the walls, the creation of a secondary entrance, and the blocking of the original entrance, as at Keiss. During the third phase of occupation the secondary entrance was blocked and the original entrance re-opened again as at Keiss. The large circular court at the entrance, with the accompanying outbuildings, the internal slab compartments, and the massive wall that surrounds the complex, also belong to this period. Finds from the broch include vessels and implements of bone and stone, querns (both saddle and rotary), ingot moulds and sherds, including a fragment of 2nd century Samian ware.

Anderson notes the foundations of an oblong, rectangular building which he suggests could be the church (ND36SW 2) which is said to have stood in the vicinity; but similar buildings are associated with Keiss broch (ND36SE 2) and Whitegate broch (ND36SE 3). The foundations lie between the circumavallation, the graveyard and the road, and since they overlie the wall of an outbuilding by some 4ft, appear to be of much later date.

S Laing 1866; Name Book 1873; J Anderson 1901; RCAHMS 1911, visited 1910; A D Lacaille 1954; A Young 1964.

'Kirk Tofts' Broch, now overgrown, is generally as described above. There is no trace of the alleged church noted by Anderson and only the SE arc of the outer rampart is clearly discernible.

Resurveyed at 1:2500.

Visited by OS (N K B) 6 September 1965.

Samian sherd now in the National Museum of Antiquities of Scotland (NMAS).

A S Robertson 1970.

The NW entrance has a feature very rare in brochs, a stair running up from the W side of the passage, which recalls the design of the dun at Forse. It is probable that this broch shows how an imported fort plan was modified by the pre-existing local traditions seen at Forse.

E W MacKie 1975.

A hipped bone pin and a slotted and pointed iron object from the site are datable to the period about 600 - 934 AD by comparison with finds from Lagore crannog, Co. Meath.

R B K Stevenson 1955; L R Laing 1975.

The second of the two finds (above) is presumably the 'harpoon of iron 1ft 8 1/2ins in length, the barb 1 5/8 by 1 3/4ins with a slot below it, from the surface rubbish over the mound on Keiss Broch'.

Proc Soc Antiq Scot 1908.

No change to the previous field report.

Visited by OS (J M) 30 April 1982.

ND 3 6 (area) As part of a wider study of Iron Age Caithness, the broch settlements at Everley (ND 3699 6828), Keiss Harbour (ND 3531 6108), Keiss Road (ND 3488 6151), Whitegate (ND 3541 6120), Skirza (ND 3940 6844), and Hillhead (ND 3762 5140) were recorded by total station survey in June 2000. Aspects of the artefactual assemblage uncovered during 19th-century excavations by Laing and Tress Barry were also studied.

Sponsor: National Museums of Scotland (NMS).

A Heald and A Jackson 2000

ND 349 615; ND 370 631 As part of a wider study of Iron Age Caithness, the brochs and settlement complexes at Keiss Road, Wick (ND36SW 1), and Nybster, Wick (ND36SE 4) were surveyed in September 2004. The sites were previously investigated in the late 19th century by Sir Francis Tress Barry.

Sponsors: Highland Council, Caithness and Sutherland Enterprise, Caithness Archaeological Trust, AOC Archaeology Group, University of Nottingham, NMS.

J Barber, A Heald and J Henderson 2004


Project (1980 - 1982)

Aerial Photography (1991)

Archaeological Evaluation (July 2006)

Two evaluation trenches were excavated at the Keiss Road broch. One trench was excavated within the central area of the broch, and a second one outside and to the NE of the broch. Both trenches measured 2 x 1m. The trench within the broch revealed no surviving in situ deposits. The trench outwith the broch revealed a topsoil deposit of at least 2m depth.

Archive to be deposited within NMRS.

Sponsor: AOC Archaeology, Caithness Enterprise Trust, Caithness and Sutherland Enterprise, National Museums of Scotland and University of Nottingham

A Duffy 2006

Publication Account (2007)

ND36 7 KEISS WEST ('Road Broch', 'Kirk Tafts' or 'Kirk Tofts', near the 'Churchyard Mound')

ND/3488 6151 (visited 13/7/63)

This probable solid-based broch in Wick, Caithness, stands on flat farmland, just west of the main road north to Wick, and was partially excavated by Laing in 1864 [2], after Joseph Anderson had found some “rude bone pins and pottery” [0, 131]. Laing described the site as the "Churchyard mound" but seems mostly to have been exploring a shell midden lying between it and the road but clearly running under the broch. Sir F Tress Barry dug out the broch in 1893-95 [3] and revealed a solid-walled ring of masonry very similar to Keiss South (above) in having two entrances. Keiss West is also the clearest example in Caithness of a broch-like structure with what appears to be a second and primary intra-mural stair rising from the primary entrance. The broch sits at the centre of a complex of 'outbuildings' within an approximately circular walled enclosure about 43.9m (144ft) across.

The existence of the early photographs provides a good illustration of how the verbal descriptions of what was found can sometimes be reinterpreted on the basis of modern knowledge (below).

1. The early excavations

Samuel Laing explored an extensive shell midden at Keiss in 1864. It consisted of

"… a great mass of shells, at least five feet deep, and covering an area of several hundred square yards, (which) rests on the natural soil and is itself covered by the foundation of a massive building . . ."

Numerous stone, quartz and bone implements were found among the shells and have been considered to be of early post-Mesolithic age and "not unlike the late Larnian" (Mesolithic culture) in north-east Ireland [8, 266]. Very similar material has been found in middens at Freswick Bay, where sherds of Beaker pottery were also found [8, figs. 117 and 118]. The deposits are presumably at least two, and perhaps up to three, thousand years older than the broch.

2. Tress Barry's work in 1893-95

Laing did not explore the massive building overlying his midden and Sir F Tress Barry undertook this work in 1893-95.

2.1 The structure of the broch

Making a preliminary judgement from the close similarity between this site and its neighbour Keiss South (ND36 6), the primary entrance seems to have been that on the north. There is a stair ascending into the wall from an opening in its left wall and, when first exposed, eleven steps of it were visible [4, pl. liii]. The passage leading to it was 1.07m (3ft 6in) wide. Opposite the foot of this stair a long cell or gallery runs back into the wall for 9.15m (30ft) and the rounded inner end was corbelled; the lintels had all disappeared and it was 1.37m (4ft 6in) wide at ground level and the walls still stood 1.8m (6ft) high. This long cell was described by the Commission as being "now filled" and this probably means it had become so through dilapidation since it was excavated, rather than that it was found so at the time of its exposure by Tress Barry; Anderson makes no mention of the point.

The doorway from the interior to this stair was in fact a passage right through the wall and, as at Keiss South (above), a massive stone slab was wedged in position against the door-checks; a large, long stone leaned against its inner face as if to prop it in position. Presumably the original wooden door was removed at some stage. This door-frame was about 90-120cm (3-4ft) in from the outside, judging from the plan, and there was a bar-hole in both sides of the passage wall behind the slab. The outer end of this entrance had been blocked by a secondary facing of masonry against the outer face of the broch wall, again as at Keiss South (ND36 6), and this can also be seen on the plan. The Royal Commission's account makes no mention of the door-frame or blocking slab, thus giving the impression that this broch only had one entrance [4].

At about 9.30 o'clock from this, the assumed main entrance, there was a second doorway (on the east, pointing to the bottom right corner on the plan) which is 75cm (2ft 6in) wide at the outer end and which also has a pair of slabs set into the wall as a door-frame 2.13m (7ft) in from the outside; within these the passage is 96cm (3ft 2in) wide. There is no north point on any of the reproductions of the plan of this site. Swanson’s plan supplies the information At a distance of 76cm (2ft 6in) behind these checks is a doorway in the right wall (looking in), apparently leading to some kind of guard cell or mural chamber which is now ruined and inaccessible. No lintels were found in position over its doorway, and no details of the chamber were recorded on the plan.

At 12 o'clock from the assumed main entrance, on the south, is a doorway from the interior which leads to a second mural stair rising to the right and having a long stair-foot guard cell on its left; the latter was 3.66m (12ft) long by 1.52m (5ft) wide, and twelve steps were found when it was first exposed.

At 4.30 o'clock is a shallow mural cell (in the space between the long guard chamber of the north-west entrance and the inside wallface) with a curious doorway – an opening cut in the middle of a large stone slab set upright in the wall [3, fig. 16]; the cell to which this leads measures 1.37m (4ft 6in) long by 1.14m (3ft 9in) wide by 1.38m (4ft 7in) high and is corbelled.

The central court

In the middle of the central court was a partly rock-cut well or underground chamber 1.68m (5ft 6in) deep and 1.55m (5ft 1in) long by 91cm (3ft) wide: it was roofed with stone flags. None of the available photographs of the interior show this feature clearly. Two tanks formed of slabs on edge were set into the floor of the central court and rows of large flags on edge, set at right angles to each other, divided the central court into four approximately equal areas [3, fig. 15: 4, pl. 54]. One of these rows was a double one which extended the line of the north-east doorway inwards to the centre. A stone vessel having a cavity 30cm (12in) square and 23cm (9in) deep stood in the floor [3, fig. 19].

The outer face and the outbuildings

The northern entrance emerges into a sort of circular courtyard with a diameter of 10.07m (33ft) – about the same diameter as the broch interior. The inner face of this court (which seems to have been the only part exposed) runs up against or under the masonry of the outer end of the broch entrance is a manner which is not altogether clear despite the survival of some good photographs of these features taken soon after they were excavated [3, fig. 17]

The plan clearly implies that a thick mass of masonry has been added to the outer face of the broch here, on both sides of the entrance, and Anderson's account confirms this [3, 137]. The wallface on the right, projecting at right-angles to the broch wall, must be the end of the thick added “casing”, just to the north of the passage. He also says that this added wall is "partly founded on an accumulation of debris 2.5ft in height." and this could be the earthy deposit at its foot.

A photograph shows the outer end of the eastern broch passage on the left (with a high sill) with what appears to be the original foundations of the broch wall in the foreground. This feature is not included in Anderson's cut-down reproduction of the photograph [3, fig. 17]. The curved face on top of and set back from this is presumably the end of the inner face of the large, added circular building, partly overlapping the reduced broch foundations. If this interpretation is correct the photograph rather neatly demonstrates how the eastern entrance has to be a secondary one; this passage, with its paved floor, can be clearly seen to be framed by the wall of the circular enclosure and to be emerging on top of the original broch foundations (which may run continuously in front of it). There is a short flight of small steps leading up from the lower level formed by the broch foundations up to the passage floor. The thick “casing” wall on the right evidently abuts against the wall of the circular enclosure, as is implied by the plan.

This part of the original broch has evidently been extensively altered. If one supposes that the broch was very substantially demolished in Phase 3 (below), particularly in the eastern arc where a new doorway was to be constructed, the two photographs (Illus. 7.147 and 7.148) are probably showing clear evidence of this event. In the foreground is the original line of the broch wall – now reduced to one visible course although there are doubtless several more still buried. This extensive demolition was undertaken to allow the new entrance to be built easily.

Above this is the end of the inner wallface of the external circular building where it approaches the new entrance, evidently formed at a considerable distance above the original ground surface outside (presumably the equivalent amount of debris had accumulated in the interior). A short flight of steps leads up to the sill of the new passage and, on the right, the end of thick, added “casing” wall can be seen in the upper photograph, abutting against the outer circle (also as shown on the plan). The masonry of the latter shows very high quality masonry, including one very neat infill of small slabs) (Illus. 7.148).

A massive outer wall enclosed the whole site, an area about 43.9m (144ft) in diameter; at the time of excavation the wall stood to a maximum height of 2.2m (7ft 3in). Many stone buildings were between this and the broch, and an even later rectangular building was found on top of these.

2. Discussion

It is likely that the same error that was evidently made in the early interpretations of Keiss South (ND36 6) – of assuming that the most massive and best defended entrance passage was secondary because there was a stair rising from its left wall – was also made at this site. In this case too the entrance passage diagnosed as primary – that on the north-east – is more likely to be a secondary one (though here the existence of a doorway to a cell in its right wall makes the original interpretation easier to understand). Even so a number of facts suggest that it was the former entrance which was the primary broch doorway; they include the door-checks, bar-holes and long guard cell of the north-west entrance, together with the fact that a stone slab was chocked in position in place of the wooden door, and the outer end of the passage blocked with a secondary facing of masonry which continued along the outer face of the broch wall.

It seems possible that this entrance too went out of use when most of the upper wall of the tower-like building was largely pulled down (when the building ceased to be primarily a tall defended roundhouse) and when some of the masonry removed from the upper walls was used to block the outer end. The other doorway could well have been driven through the wall at this time, perhaps either through or next to a pre-existing mural cell. As we have seen the photographs of the outer end of this appear to support this view.

It is true that, if one considers Keiss West by itself, this interpretation is perhaps only moderately more plausible than the original one – particularly because one could argue that the north-west doorway, if secondary, was driven through the wall at the most convenient point where it was already thinnest (the base of the mural stair). This argument certainly applies to Yarrows (ND34 17). Nevertheless, in the author's opinion, the very clear evidence from Keiss South renders it much more plausible.

This being the case, the highly unusual mural stair rising from the left wall of the main entrance, opposite the long guard cell, is of great interest. This time there can be no doubt that it is a proper mural stair and not a second guard cell and that, if the structure was a true broch, it must somehow have run up among the intra-mural galleries. Though it is behind the main door it seems a very curious and not altogether satisfactory arrangement; it seems to offer an opportunity for attackers breaking in to force their way into the broch via the intra-mural galleries and to take the defenders in the rear. Some further comments on this feature are made below.

3. Probable site sequence

As has been indicated above, less information is available about the history of Keiss West than was the case with Keiss South, because the whole of its interior was cleared out by Tress Barry who, as usual, left no records of any sequence of later buildings he may have encountered, as Laing did with the other site. However this broch was evidently much less well preserved than the other; the photographs imply that not more than about 1.5m (5ft) of wallface remained. This is less than half of that found by Laing at Keiss South and there is unlikely to have been the same wealth of later structures at this site.

Pre-broch era – the shell midden

Laing recovered a quantity of material from this midden including some potsherds, chipped flints, stone implements including hammer-stones, some smaller pebbles interpreted as sling stones, and various pointed bone implements. The age of this midden is uncertain but, if the potsherds were truly found in it (of which one cannot now be certain), it need not antedate the broch by any great time. If it is analogous to the Freswick midden a date in the third millennium BC seems plausible.

Phase 1: the broch

Structurally this building is very similar to its neighbour Keiss South. The overall diameter is about 17.7m (58ft) but this is inferred from the only two measurements which are actually given – the internal diameter of 10.41m (34ft) and the primary wall thickness of some 3.66m (12ft). The north-west entrance is inferred here to have been the primary one; as noted the stone door – or at least a blocking slab – was still in position, wedged upright against something by a leaning slab; a thin secondary facing of masonry, laid against the outer wallface, blocked the outer end.

This passage was equipped with a bar-hole and socket, and also presumably with door-checks, though these are not actually mentioned in the early accounts [3, 135]; the plan seems to show them and the stone slab must have rested against something. Immediately behind the upright blocking slab a passage opened from the left wall (looking in) in which was the second intra-mural stairway. Opposite this a long guard cell opened from the right side of the passage, running back for 9.15m (30ft).

On the opposite side of the central court, at 12 o'clock, is the doorway to the second mural stair – apparently also a primary feature of the building – of which 12 steps remained at the time of the excavation, rising to the right as usual. A stair-foot guard cell opened to the left of this doorway and was 3.66m (12ft) long by 1.52m (5ft) wide. In the floor of the central court there appear to have been sunk a well and some water tanks. The broch may have had an outer defensive wall in the shape of the perimeter wall shown on the plan as surrounding both the broch and the outbuildings.

Phase 2: primary occupation

Presumably the broch was occupied as a defended building of some kind for a considerable period but no information about this phase can be recovered now. The observation that the outer “circle” was partly founded on 76cm of debris suggests that this phase was a long one.

Phase 3: demolition and conversion to a low roundhouse

The blocking of the northern entrance suggests that – as was evidently the case with Keiss South – most of the upper part of the high wall of this broch was pulled down after the building ceased to be needed as a defended roundhouse. According to Anderson [3, 131] a thick secondary 'casing wall' was added against the outside of the broch, varying in thickness from 61cm (2ft) to 1.09m (3ft 9in). On the plan this is shown most clearly on the arc from north-east to south-east, around the eastern entrance; as we have seen one photograph appears to show the end of this added masonry very clearly. There also seems to be a thinner ‘casing’ on the north-west arc, blocking the outer end of the primary entrance.

The eastern entrance was evidently pushed through the wall – or, more probably, built on top of its reduced foundations – at the time of demolition and, as described, the photographs clearly show that its paved floor emerged well above the broch foundations. The door-checks mentioned by Anderson are not apparent on the plan but they evidently consisted of massive opposing slabs set into the walls about half way down. The doorway to the filled mural cell in the right wall has already been mentioned (traced to a length of 8ft); this suggests that the new passage was laid out at a point where there was already a mural cell. The large external circular enclosure was laid out at the same time, and the thick “casing wall” – presumably built from debris from the demolished broch wall – was added against it.

None of the demolition debris seems to have been used to build a secondary facing wall around the central court. Anderson states that no such "scarcement" was found, using the word in the same (now old-fashioned) sense as with Keiss South.

Yet there are one or two suggestions that there might have been such an added facing. In one of the original photographs a distinct kink is visible in one wall of the passage of the main entrance near its inner end, as though the end of a secondary wall did not quite fit flush with the corner of the primary masonry [3, 135, fig. 16]. Also the curious mural cell close to the left of this entrance might have been built in Phase 4; a single stone slab with a doorway cut into it forms its front wall and may imply that a secondary facing had previously been built. Perhaps it is less likely – though not impossible – that such a slab was let into the primary wall after part of the core was torn out to make the cell. However the circularity of the central court (below) suggests that its masonry is original.

Presumably the various arrangements of upright slabs in the floor of the court were put in at this stage; they divide it into four quadrants which are clearly aligned on the secondary or north-east entrance. Also the many outbuildings which were partly revealed by the excavations might be said to fit best into a period after the site had ceased to be primarily defensive, and when plenty of building material was available from the demolished high wall of the broch. The inner part of the primary (north-west) entrance, with the 9.15m (30ft) long guard cell, may have remained in use in Phase 4 since its inner end seems to have remained unblocked.

Presumably also the occupation deposits accumulated in the interior at this time and many of the recorded finds must have come from this secondary stratum. However no information is available about the stratigraphy of the site.

4. Finds

From Laing's excavations [2, 20-22]

This material all came from the shell midden under the broch.

Flint and stone tools included chipped flints, hammerstones made from beach pebbles, a rude mortar and some small beach pebbles interpreted as sling-stones.

Bone tools included 3 points (called "arrowheads" [2, figs. 26-28], about 18 "skewers" or awls [2, figs. 29-31] made from both bone and horn.

There was also some "rude pottery". However in the National Museums of Scotland there are sherds among Laing's collection which do not fit this description, being a thin, fine, smooth ware with a near black surface (nos. 00-00). Presumably this material, which is clearly of Iron Age date, belongs to the broch horizon.

Food refuse included periwinkles ("4/5ths" of the total) and some animal bones (unidentified).

From Tress Barry's excavation [6]

Metal objects included 1 small bronze ring decorated with small diagonal slashes on the outer side (diameter 19mm or 0.75in).

Bone implements included 1 long-handled comb, 2 awls, 1 double-pointed tool perforated in the centre, 2 broken needles, 1 pin with ornamented head and a swelling on the shank of late Iron Age type [6, 285], 1 sheep shank (centrally perforat-ed) is probably a bobbin, and 1 toggle.

Stone tools included 11 slaty discs, diams. 57-178mm (2.25-7.0in), 1 polished schist disc, diam. 83mm (3.25in), 1 cup without a handle, 1 broken handled cup, 1 sandstone lamp, 1 frag-ment of a jet armlet, 12 upper and lower rotary quern stones, 2 large saddle querns, 2 moulds (one, for an ingot, carved in the rubbing face of a piece of rotary quern), 1 crudely hollowed vessel, 2 oval vessels, 2 pivot stones, 3 hammerstones, 8 whetstones, 2 sandstone whorls and 1 of steatite, 2 small discs. diams. 45mm (1.75in), 2 broken quartzite pebbles, grooved on both faces – presumably strike-a-lights – and 1 carved sandstone disc [3, fig. 18].

Pottery included 1 base sherd from a Roman Samian decorated bowl of Form Dr. 37 (2nd century) [10, Table II], and several native sherds including one with an indented cordon.

Food refuse included bones of ox, sheep, goat and pig?, with pieces of red deer horns, and edible mollusc shells. Small quantities of charred grain (bere?) were found and the canine tooth of a bear, Ursus arctos.

5. Dimensions

The internal diameter of the broch was given as 10.2m (34ft) [4], and the original wall was apparently c. 3.6m (12ft) thick; thus the external diameter should be c. 17.4m (58ft) and the wall proportion c. 41.4%. A new survey of the interior by the author in 1971 showed that its plan was extremely close to a true circle, with a radius of 5.18 +/- 0.04m (diam. 10.36 m, or a fraction under 34 ft). This suggests that the inner wallface is the primary one, not a secondary addition, and that the “kink” in the masonry of the passage is a door-check near its inner end.

Sources: 1. NMRS site no. ND 36 SW 1: 2. Laing 1866, 19 ff.: 3. Anderson 1901, 131-9: 4. RCAHMS 1911b, no. 517, 157-8, fig. 41, pls. 53 and 54: 5. Young 1962, 181-2: 6. Stevenson 1955, 285; 7. Proc Soc Antiq Scot 43 (1908-09), 12-13 (finds): 8. Lacaille 1954, 266: 9. Laing and Huxley 1966: 10. Robertson 1972: 11. Swanson (ms) 1984, 586-89 and plan: 12. Heald and Jackson 2001, 129-47: 13. Armit 2003, 106:

E W MacKie 2007


MyCanmore Image Contributions

Contribute an Image

MyCanmore Text Contributions