Skye, Flodigarry Hotel

NG47 1 DUN FLODIGARRY (‘Flodi-garry Hotel’)

NG/4639 7196

This probable unfinished ground-galleried broch in Snizort, Skye, was identified as a broch only in 1977 when the owner of the Flodigarry Hotel drew the attention of the Scottish Development Department (now Historic Scotland) to the site and suggested that it be excavated [4]. The work took place over three seasons from 1979-81 [2].

1. Situation

The site stands in the disused walled vegetable garden of the 19th century house which is now the hotel, on a low rock outcrop (4.5m high at the most) below high crags; the surrounding ground is fairly level [2, pl. l]. Excavation revealed that the wall of the prehistoric structure did not form a complete circuit and only stood to a few courses, being nowhere higher than 1.0m. The steepest slope of the knoll is on the east and the wall is missing in this arc. It thus seemed possible that the structure had been extensively robbed.

2. The surviving structure

The entrance passage faces north-west and stands only two courses high at the most [2, pl 9]; the door-frame – consisting of built, inset checks – is just over 1m from the outside. There is a ground level, intra-mural gallery of uneven width running along most of the surviving wall; on either side of the entrance it ends in a wider, wedge-shaped cell, the one on the left (south) being linked with the passage by a narrow door. The end wall of this cell, forming the south side of the passage, appears to form a straight joint with the outer half of the broch wall, and this suggested to the excavator that the quality of the masonry was not as high as usual. Exactly the same arrangement is to be seen at Dun Ardtreck (NG33 2). The cell on the north (right) has curved walls and is separated from the gallery beyond by what is termed a “blocking stone” [2, 31]; it has its own door to the interior.

There are three doors from the central court to the intra-mural spaces – at 5, 7 and 8 o'clock; the length of gallery between the left guard cell and the door at 8 o'clock is very rough. This crowding of all the doorways into a short arc of the wall is also unusual. Along the east edge of the rock knoll the only traces of the wall were some small boulders filling in natural clefts in the rock.

3. The stratification uncovered

The wall of the building was founded directly on bedrock, any turf and rubbish evidently having been cleared away from the knoll in advance of construction work. This underlying rock was uneven and a clay layer had been put down on it in places to provide a more level surface.

In the south guard cell a pebble floor had been laid on this clay, and extended through part of the entrance passage, though not far into the central court. It underlay the inner wallface in the south-west quadrant, and scatters of charcoal were found on top of the clay here, and extending into the interior. There was none of the other usual evidence of primary broch activity – such as hearths, post-holes and drains – and neither was there any sign that the rough interior surface had been flattened off. The entrance passage was paved although it lacked a lintelled drain underneath.

A rubble bank was found immediately outside, and these two areas provided the most interesting information recovered. The primary paving of the entrance emerged on top of the rubble bank which itself contained the upper stone of a rotary quern [2, fig. 12, no. 367]; both bank and quern must therefore belong to the construction phase of the broch. The rubble of this 'bank' was loose and rested directly on bedrock.

A section was cut at the foot of the rock knoll on the east side to discover whether the absence of a wall on this side could be explained by its having fallen down the slope. However there was no sign of this. This discovery, the general absence of stone debris, and the fact that the rubble bank was a primary feature and was not composed of material fallen from the wall, all led the excavator to conclude that the broch had never been higher than it was when he uncovered it; it was evidently an unfinished building.

Traces of secondary activity were found in the entrance passage in the form of two thin layers of soil, containing midden material, and a rough stone paving, all laid on the original paving. A midden layer with animal bones and shells lay on top of the secondary paving. Nothing comparable was found in the central court.

Dating: charcoal from primary deposits gave a C-14 date of 45 +/- 65 bc (GU 1662), and the mean point calibrates to about AD 55 in real years.

4. Discussion

There is evidently no way of proving that Dun Flodigarry was a hollow-walled broch since the wall is so poorly preserved that not even a trace of an intra-mural stair was found. The excavator did not consider the problem of the stair in detail, but it deserves some thought.

The intra-mural galleries in the preserved part of the wall do not seem suitable at any point to contain a stairway. At Dun Ardtreck, where the absence of the base of a stair posed a similar problem (NG33 2), there was at least a wide section of the basal gallery, with suitably smoothly built sides, which could have served as the approach to the vanished steps. However at Dun Flodigarry the only wide sections are the terminal expansions at either side of the entrance both of which are too short to have housed a stairway. The rest of the preserved galleries are narrow and with roughly built walls. The stairway may of course have been planned for the missing arc of the wall on the east side.

The hypothesis that Dun Flodigarry was an unfinished building – or, more accurately, one which had only just been started – explains these curious features and allows us to assume on the basis of the size and ground plan that it was intended to be a ground-galleried broch of standard Hebridean type. The excavator considered the possibility that it might be a D-shaped semibroch but decided against it, mainly on the grounds that such structures are invariably built against high cliffs. In addition it may be noted here that the absence of any plausible position for an intra-mural stair in the surviving wall is a decisive argument against the site being a semibroch; some trace of the stair – even if only an empty but suitably designed stretch of gallery – would have to have been preserved in such a structure.

On the other hand there is evidence which might argue against the hypothesis of an unfinished structure – particularly the long history of recent and modern settlement close by and the numerous adjacent stone structures which could have been built with material robbed from the broch. It seems unlikely however, judging from numerous examples of brochs elsewhere which are known to have been robbed, that the wall would have been so completely removed that even the foundation stones on the east arc were taken. The absence of any trace of internal furnishings, and the fact that the intra-mural gallery extended further than the inner wallface at both ends, seems to confirm that the site has not been significantly disturbed in modern times but had simply been abandoned unfinished in the Iron Age.

5. The finds

All the pottery and finds were of Iron Age type so it seems that the site was finally abandoned not long a fter building activity had halted.

The pottery comprised mainly plain body sherds but the few decorated fragments [2, fig. 6] are well matched, for example, by the 'native' incised wares of Vaul ware found at Dun Mor Vaul, Tiree (NM04 3). The two base sherds with internal decoration are also Vaul ware types. There are few signs of the Everted Rim pottery found at that site and, in its later levels, at Dun Ardtreck on Skye (NG33 2), though no. 484 [2, fig. 9] looks like one. There are also some bead- or roll-rims [2, fig. 9, nos. 243, 286, 238 and 293].

Roman material: a fragment of Roman Samian ware among the material stored in Portree museum which was not mentioned in the report [3]; the dating implications of this neatly coincide with those of the radiocarbon date.

Stone objects include many pebble hammerstones (it is suggested that they may have been used for hammering small filler stones into gaps in the wall), some faceted grinding hammerstones (the material which was ground is as yet unknown) [2, fig. 11], 2 unstratified flint scrapers [2, fig. 7], and 1 complete and 2 incomplete upper stones of rotary querns, all without handle holes [2, fig. 12]. The complete quern is a flat discoid stone, and the fragments belong to stones with a tapering edge.

6. Dimensions

The overall diameter of the broch ranges from 17.82 - 18.35 m, an average of 18.1 m; the internal diameter is estimated at about 10.0 m, and the hollow wall is therefore about 7.8 - 8.4m thick, giving a wall proportion of about 44.8%.

Sources: 1. NMRS site no. NG 47 SE 6: 2. Martlew 1985: 3. Information from Mr Roger Miket: 4. Swanson (ms) 1985, 813-15.

E W MacKie 2007