Garbh Eilean, Shiant Islands

NG49NW 14 c. 41 98

During the first two weeks of June 2000 an archaeological landscape survey and the partial excavation of a blackhouse complex was undertaken on the Shiant Islands with a team of MA students and professionals from the Czech Republic.

Survey

The three main islands, House Island (Eilean an Tighe), Rough Island (Garbh Eilean) and Mary Island (Eilean Mhuire), were subjected to a systematic ground survey, and as many as possible were recorded by measured drawings, often stone by stone, the rest being recorded by measured field sketches. Most of the sites seem to belong the modern or Early Modern period. However, not unexpectedly, a number of these sites appear to be located on or utilise earlier sites and their stonework. In total 112 sites were located and recorded, 31 on House Island, 46 on Rough Island, and 35 on Mary Island.

Rough Island. The variety of sites appears to be richer in kind and period on this island. There are no large blackhouses, but there are a number of substantially built shielings which could have been used for more than just seasonal occupation. The kelp industry is in evidence with kelp-burners' huts and ovens, confirmed by a contemporary although romanticised print that shows the ovens in use in the distance. For the medieval period there appears to have been a small settlement on the N coast, and the Norse period may be represented by the best boat-shaped stone setting found so far. Several sites are potentially of the prehistoric period, such as a mound and stonework underlying a shieling or a massive stone-built platform and associated structures on the S coast, and several of the small cairns that dot the uplands. Of great interest are two rockshelters, but they have most likely often been used at any number of widely separated periods of time.

Sponsor: A Nicholson

P Foster 2000.

NG 41 98 The Shiant Isles Project continued in 2005 with excavations focused on Eilean an Tighe, general site survey and excavation area survey.

NG 4199 9726 site HI 15a. Blackhouse with outhouses and enclosures (DES 2004, 137). The primary occupation layers were excavated, and over 160 pottery sherds were recovered dating to the Early Iron Age. A thin charcoal-flecked clay deposit overlay the bedrock and yellow post-glacial clay. Although no pottery was recovered from this basal deposit, some 40 flakes of struck mudstone were identified, including a discoidal scraper. Pottery from other adjacent basal deposits is currently dated to the Late Bronze Age.

The sondage excavated in 2003 (DES 2004, 137) to the W of the blackhouse was extended to include all of the enclosure. Several wall footings extended out into the enclosure from under the blackhouse. The enclosure appears to have functioned as a cottage garden.

A stone mound is one of the latest features on the site. Superficially it appeared as a spread mass of jumbled stone which, when reduced to its bottom layer, revealed ordered lines of stone blocks. This is most likely the remnants of a mason's stock pile of the 18th century.

Prior to excavation it had been believed that two enclosures, divided by a wall, lay to the S and W of the blackhouse. Now it was clear that the division was formed from the remnants of the pre-blackhouse walls, combined with the abandoned remains of the stone stock pile.

An earlier building, possibly a 17th to early 18th century blackhouse, was revealed in the curved face of part of a wall, with the wall footing running S across the site. The E wall was either under, used or destroyed by the later blackhouse. Internally were patches of hard clay, some of which appeared to have trampled upper surfaces, and areas of burnt ashy soils, which most likely belong to the lower, late medieval deposits. However, a circular setting of stones just inside the W wall may be contemporary. At the S end a small sub-rectangular pit contained an articulated sheep burial truncated by the garden soils.

A building, of which only a short length of walling remained, pre-dated all other buildings. For the most part, only the bottom course of the wall survived although, at its W end, more of the outer face remained. The massive width of the wall - 1.2m widening to almost 2m - is reminiscent of Late Iron Age monumental buildings, although there is no Iron Age material in this area.

At some time the SW corner of the late blackhouse was disturbed. Excavation of this corner of the building revealed evidence for a peat-fuelled pottery clamp kiln in the form of intensively burnt peat ash. A small quantity of very underfired, oxidised pottery sherds and the kiln are datable to the mid-19th century.

A linear pile of stone set within the enclosure soils, indicating a late period in the blackhouse sequence, may have been associated with the kiln, although it could also have been a crude revetment for the disturbed blackhouse wall.

In an area outside the blackhouse complex to the N of the barn, excavations in 2003 had revealed a layer of rubble in the trench. Amongst the larger stones were over 80 struck mudstone flakes and a few hammerstones, indicating stoneworking on a significant scale. Initial assessment of the associated pottery suggests a date in the Middle Iron Age.

Below the stoneworking deposits were several unstructured hearths, surrounded by mixed, trampled deposits of orange peat ash and clay. A few set stones may represent the remains of a hearth box. Over 300 pottery sherds, datable to the Late Bronze Age and Early Iron Age, were recovered, as was a small assemblage of struck mudstone flakes. There is little to suggest that this area originally lay within a building.

To the E, stones set to form small stalls were cut through the Iron Age deposits: earlier than the blackhouse and its barn, they may represent elusive late medieval structures. NG 4200 9723 hi 16. Barn and blackhouse. In the initial island site survey this site was considered to be a small blackhouse with an attached turf-walled outhouse on its S side.

Initial excavations have shown the supposed blackhouse is more likely to have been a detached barn, belonging to the blackhouse above, and the outhouse to be a possible early blackhouse of the 17th/18th century.

NG 4186 9767 Bothy HI 4. Excavation of the beach pebble path, which runs outside along the E wall between the two doorways, was completed. A line of walling revealed under the pebbles may belong to the chapel that stood on the side of the cemetery mound until at least 1815.

NG 4186 9762 Bothy HI 7. Excavations showed that although there is no evidence for an additional attached storage room, the internal features mirror the other lobstermen's huts built in the 19th century. Above the stone floor, a thin compacted blackened clay layer indicated a period of abandonment after which, possibly in the early part of the 20th century, fresh cobbles were laid. The fireplace was raised with further stone blocks and a fresh layer of pebbles laid

in the entrance. The islands' first coin hoard of six pennies, dating from 1900-18, was found in this phase.

Five test pits, 0.6m square, were cut in the lower settlement area. Two lay to the S of the cemetery enclosure, one sited on a lazy bed and one in a furrow; a third was sited within the cemetery enclosure, to the S of the cemetery mound itself; and two were located to the N of the cemetery.

The pits to the S of the cemetery enclosure produced over 50 sherds of 15th- to 18th-century pottery, which most likely date the use of the lazy bed system. Few sherds were recovered from inside the enclosure, but the two northerly pits produced sherds of a fabric not encountered on the islands so far. Their location, under a mass of scree and associated with a possible stone structure and burnt peat ash soil, suggests an earlier occupation in this area, either of Norse or a pre-Iron Age date.

A rapid survey of the W coastline of Eilean an Tighe and the annat bay area of Garbh Eilean found considerable coastal damage from the January storms, which put both the lower settlement of Eilean an Tighe and the annat site under threat.

The evidence for a significant stoneworking industry on the islands is now overwhelming. Several hundred struck flakes and several large blocks showing flaking scars have been recovered from the excavations, although only a few tools have been found. These include a discoidal scraper, a borer and a large cleaver. Some of the material has been found associated with Late Bronze Age pottery, but large quantities also come from Iron Age and 15th century contexts.

Over the last five years, several hundred pieces of flint have also been found, almost all of which have been un-retouched flakes, chips or small ?cores. Most of these possibly derive from beach pebbles, however, a large nodule was found in a remnant boulder clay deposit at the shore line. There are two gun flints, a borer, and a scraper. See also www.shiantisles.net.

Archive deposited in NMRS.

Sponsors: Hunter Archaeological Trust, Dorothy Marshall Bequest, CBA Challenge Fund, Society of Antiquaries of Scotland.

P Foster and J Hooper 2005

Archaeological notes continue under Architecture notes.

Continuation of archaeological notes -

NG 41 98 The Shaint Isles Project continued in May to July 2006, with excavations focused on finishing the LBA/EIA blackhouse site HI 15 and the field barn and blackhouse HI 16 on Eilean an Tighe, a general site GPS survey of all three islands and a programme of test pits across the agricultural landscape.

NG 4199 9726 Site HI 15A: Blackhouse with outhouses and enclosures

Area E/F The cottage garden enclosure

In 2005 the excavation of this area had reached a point where most of the early 19th-century garden soils (F207) had been removed and the surviving remains of the ?17th- to 18th-century blackhouse F221 had been revealed and recorded. The site was now all at the same level, but not at the same contextual level, since it was clear that the S side of the site had begun to slope down.

The slope becomes pronounced to the W and a possible mid-18th-century limpet shell midden F231 had been deposited at this point. The midden was fully excavated to its base, which proved to be on bedrock. Bedrock was now exposed over a large extent of the W side of the site, sloping unevenly down to the E away from the garden enclosure wall. The wall was now sitting high along the steep edge of the bedrock hump that also marked the limit of the occupation on the W side. A further line of lower rounded humps of bedrock rose up approximately am from the W wall of the late blackhouse HI 15A. This configuration of the bedrock formed a hollow, open-ended to the S into Area E, which contained the bulk of the garden soils. The depth of soils in this hollow suggests that considerable amounts of soil have been introduced, possibly the result of gathering 'scalped' turf from the surrounding landscape.

The ?17th- to 18th-century blackhouse F221

The excavations in 2006 revealed further features of the early blackhouse F221. A drain F83, earlier recorded mistakenly as a wall foundation, ran from the centre of the blackhouse in a curving line to the N where it disappears under the SE corner of the enclosure garden wall. From the E, from under the W wall of blackhouse HI 15A, a wall foundation F79, also previously recorded, ran to meet the end of the drain in the centre of the blackhouse. This arrangement is reflected in the more substantial and later house of HI 15 and indicates a separation of the floor space into the domestic, southern portion and the northern animal byre section with its drain.

As usual it proved impossible to identify a floor surface associated with the house. The base of the garden soils changed to a stiffer mix of burnt soils, clays and general occupation debris F235-236, which produced a number of pottery sherds datable to the early to mid-18th century, but these deposits appeared to continue below the level of the house walls and the floor hearth. However, there can be little doubt that the surface of this deposit was used as a floor level.

During the life of the house more material was trampled into

the rooms, for example a burnt deposit F234 in the area where the house wall had been reconstructed from the N face of the earlier wall F205. However, the general lack of ash on the house floor hearth suggests that the floor surface was kept stable and relatively clean. A thin brown sterile deposit (F235), found in some areas and noted as occasionally overlying some of the ashy deposite of F236, may have been either a soil contemporary with the house or the sign of a short period of non-occupation.

An iron cold chisel (SF647 ) was found under the house wall, resting on the surface of the red ash. This is the third iron cold chisel found on the site and there have been abundant metalwork finds.

Sections in earlier wall F205 showed tipped ashy soils, thought to relate to the dismantling of the northern face and its rebuilding as the inner wall of blackhouse F221. Ceramics from these deposits appears to be of early 18th-century date, but their often close similarity to the S-profile jars of the medieval period leaves room for doubt.

Pre-blackhouse deposits and stone building (wall F205)

The next series of deposits under the blackhouse F221 are either a brown, generally sterile, soil (F235) or a widespread and compacted, occasionally thinly laminated soil (F236) composed mainly of peat ash impregnated clayey soils. This originated and was trampled out from a number of structured and unstructured hearths, and it appeared to have been the surface upon which some at least of the 18th-century activity took place. The ashy soils butted up to and were level with the base of the northern wall stones of the massive wall F205 and in some places swept up the sides of the base wall stones. Where the wall has been robbed away the wall line can still occasionally be traced by intermittent occurrences of the reddened ashy soil, which was not beneath the wall itself. Where the wall had been dismantled and reconstructed as the SW of the 18th-century blackhouse a similar line of burnt soil ash showed the former line of the wall. This strongly indicates that wall F205 was contemporary with the hearths and their ash deposits.

Although there were several unstructured hearths in the area it was dominated by a hearth (F237) of usual dimensions, around 0.70m diameter, made of blocky stones set in a rough circle, some of which have burnt to the point of disintegration. Surrounding this hearth was a circle of intensely baked clay, approximately 2.60m in diameter and thus disproportionate in size to the core structure of the hearth. There was no regular structure to the edge of this baked clay disk and it was penetrated by numerous rat burrows, leaving it uneven. The clay was extremely hard, indicating that high temperatures had been reached, and layering within the matrix indicates prolonged use. The disk also encompasses the small unstructured hearth (F247), and allowing for the missing burrowed edges of the disk, two further hearths (F248 and possibly 245) would also have been included. An alignment of three small shallow post pits F253/4 and 275 was covered by the baked clay, with a further posthole F253 beyond the disk to the NW.

On the S side of wall F205 in Area E the surface of a less burnt clayey soil was revealed in which several stone features were set running directly to the S from the base of the wall face. These may correspond with the hard baked surface C121 and deposit C120, which were revealed below the garden soils within the winnowing barn HI 15C Both the surface and deposit also ran out of the excavated area in the barn into Area E. It is fairly certain that these features and deposits related to the large wall F205. The finds of steatite from deposit C121 may therefore be of importance in dating wall F205. Amanda Forster has reported that the steatite from HI 15 can be dated to the late Norse period around the late 13th to early 14th centuries and that one of the bowl fragments is of a Shetland type.

Throughout the excavations in Area F and E a small number of recognisable medieval jug/flagon rim forms and sherds in a distinctive fabric type associated with many of these rims have been found. The dating of most of the pottery is, until a more expert analysis can be made, hampered by the abundance of globular S-shaped jars. This form, with a great profusion of minor differences, appears to be common from at least the 13th century on into the mid-18th century. One of the difficulties is being able to recognise residual medieval items in the later assemblages.

On the northern side of wall F205 the baked clay and ash deposits were removed and the lesser unstructured hearths were sectioned. One hearth (F249) was found to be a pit filled with pure red/orange peat ash. At its base was a large fragment of a globular jar.

It was now possible to see that to the N of the wall a number of features were all evidence of activity associated with wall F205: the unstructured hearths F245/6/7/8 and 251; a structured hearth F237; pit F249; soils F235/6; patches of possibly dumped ash; a very dark brown soil (F238) below the ashy layers; and probably postholes F253/4/5 and 275. At some time before the construction of blackhouse F221 in the 17th or 18th century the structured hearth F237 became the centre of a massive bonfire F237b. Wall F205 did not go deeper than these deposits and features and it is therefore highly unlikely that it formed part of a monumental construction of the Late Iron Age. The deposits associated with the wall are now seen to be very thin, not the kind of deep accumulation usually associated with Late Iron Age monumental buildings. However, the difference between the base of wall F205, the burnt deposits and the surface of the small hearths is so thin and the interface with the Late Iron Age deposits so immediate that it is possible that the contexts could be spread over different periods.

The structured hearth F237 is a formal stone setting, suggesting that it may have been an internal feature of a room on the N side of wall F205. However, there is at present no evidence of a turn in the wall to form such a room.

The lower Iron Age levels

Excavations within the floor space of the blackhouse complex HI 15 had revealed that Iron Age features and deposits continued up to and under the W walls of the standing structures. The excavation of Area F was expected to reveal more of these deposits and features as they re-emerged from under the blackhouse walls. This expectation was not fully realised. Only thick clayey soil deposits containing Iron Age pottery and some stone rubble scatter were revealed.

Area G. The stalls/sheep crush

At the end of the 2005 season the open excavation in Area G was concluded without resolving the date and function of either the wall B90, which ran under the N barn B into Area G, or of the stall-like structure built against its E side in Area G. An attempt by Linda to address this problem determined that the stalls may have been a sheep crush and that probably belonged to the mid-18th-century blackhouse.

Site: HI 16 Field Barn (A) and possible Blackhouse (B)

NG 4200 9723

Description of the site

Site HI 16 is shown on the 1st Edition Ordnance Survey map of 1854 as a ruin. This agrees with the documentary evidence that this was the period before the Campbell family began their term of occupation in the 1860s, a time when no shepherding families were willing to live on the islands. In 1827 Lord Teignmouth informs us that the shepherd's family were 'just packing to leave their cottage with attatched shed' (the blackhouse with barns HI 15), and no inducement could persuade them to stay on the islands. In 1859 TS Muir described the islands as still uninhabited.

The outline of the building on the map appears to be exaggerated in size, since it is shown as larger than the blackhouse HI 15 just to the N. This remained unaltered in the 2nd Edition in 1899, for which it is unlikely that the islands were resurveyed, but on current maps it is shown in a much reduced form, which is closer to the ground plan of the standing building.

In our initial rapid survey of 2000 we made a measured field sketch of HI 16 which showed that the building was slightly more complex than shown on the OS maps. barn A was found to have a possible outhouse B attached to its southern end. The barn walls (externally surveyed as 8 x 5.20m) were built of stone and earth, standing on average 0.80m high, in common Hebridean style, but the outhouse was visible only as a low mounded turf outline barely more than 0.20m high, adding an extra 5m to the total length of the building. Both had E-facing entrances.

One important feature of the building was recorded at the time but did not appear unusual. The southern outhouse (B), while generally conforming to the line of the E house wall, was wider than the house, and its W wall was outside the line of that of the house by up to 2m. However, it appears to fade away as it reaches the line of the southern house wall, allowing for the possibility that it turned inwards to the E and butted up against the house wall.

Now that the barn has been partially excavated it is clear that the so-called outhouse is the southern end of an earlier blackhouse. The unusual thickness of the E wall at the northern end beyond the entrance was found to be a free-standing stone setting of massive boulders (12) infilled with earth rubble and a limpet midden. The very large boulders outside the NE corner of the barn A may have been the remnants of the original walls of an early blackhouse (B) now known to underly A. The low mound of earth forming the southern wall could be the remains of the blackhouse wall core after its stone facings were robbed to build the barn. If this is correct, then a building of substantial proportions, not very different from those of blackhouse HI 15, may be envisaged, which would correspond with the dimensions suggested on the Ordnance Survey maps.

The excavations by Linda Foster

Introduction

The excavation of this building began in 2005, but only proceeded as far as removing the bulk of the collapsed walling material and the accumulated outcast soils (1/5). Constant rat activity in the upper already disturbed and loose material combined with deeper burrowing into archaeological deposits below produced a quantity of cultural material, including recent pottery, glass and the fly sheet of a tent.

NG 4200 9723 HI 16 barn and blackhouse Phase 1

The primary phase, which interfaces directly with the post-glacial subsoil (23) and exposed glacially smoothed bedrock, involved the construction and occupation of an 18th-century (a 17th-century foundation date is also possible) blackhouse (B) (c 10 x 6m) with an unstructured hearth (20) and drains (14 and 16). The wall foundations exposed at the southern end appeared as low earthen mounds with a little stone content, presumably the remnant foundations of a basic turf superstructure. The apparent E-facing entrance may or may not be original. An alternative, much more substantially constructed building, is proposed in the discussion below.

The present excavations only affected the part of the building inside the barn (A). This meant that only a part of the interior floor space was examined and that the shell of the building was beyond the excavated area, either below or incorporated into the standing walls of the barn A or are even further outside the barn wall lines (?12). The northern part of the house floor area is also problematic since it is mainly composed of exposed bedrock over which any surviving floor deposits have been eroded away or removed for the insertion of the early 19th-century drain (4) of the later barn (A).

The surviving remnants of the floor of the blackhouse are composed of a base accumulation of mixed brown and grey clayey soils (17 and 19) blackened at the surface, which have been compacted and consolidated to form a very hard surface. This deposit was associated with an unstructured hearth base (20) and goes under the southern walls, as did the overlying soil 0.25m deposits, which represented continous floor accumulations (11, 15 and burnt peat ash 18). The hearth was set slightly off-centre towards the W in the southern half of the blackhouse, but in the excavation of the barn floor area it appeared in the SW corner of the barn where a much larger part of it disappeared under the standing W wall of the barn. It was left intact although the clay floor base deposit was taken down to the subsoil/bedrock. The hearth, an area in the SW corner of the barn of bright orange/red baked and burnt peat ash, was a roughly rectangular shape a 1 x 0.40m in extent. Unfortunately the shape was broken up and distorted by old rat runs, but despite this it could be seen that the exposed baked area was relatively large. Although it was not excavated it could be seen by comparison with the excavated base floor accumulation that it was not a thick, piled deposit. No structural elements were found and without dismantling the barn walls it is not possible to say more. However, it is possible that this was only a thin peripheral spread from a much more structured hearth hidden beneath the walls.

The drains (14 and 16) are considered contemporary with the blackhouse and with each other, although their junction was almost totally destroyed by the later insertion of the stone levelling blocks (10). Small blocks and slabs of stone that lined and capped the 0.10m deep drain channels cut into the underlying glacial clays.

Phase 2

The settlement was vacated in the late 18th century and the community later replaced by a single family living at the main blackhouse site HI 15. Presumably the HI 16 blackhouse was abandoned at this time. The landlord may have ordered the roof to be destroyed and scattered since there was no evidence for it. Any stone walling would probably have been used in the construction of the Phase 3 barn (A).

Phase 3

A windowless field barn (A) (c 7.5 x 5.5m) was constructed over the northern part of the blackhouse site in the 19th century, with a wide stone-faced earth-filled cavity wall standing c1m high and an E-facing doorway. A stone-capped and -lined drain (4) was inserted into the floor and an apparent hollow in the central NW floor area was levelled up with a setting of stone blocks (10). This barn was probably built at the time when the farmstead was being expanded with the construction of the N barn (B) and the S winnowing barn (C) at the main blackhouse HI 15, and possibly the conversion of the site at HI 17 to a kiln house. This could also have been the point at which the island was first occupied by a shepherding family and the stock changed from mainly black cattle to sheep or, taking into account the map evidence, this may more probably have happened in c 1865 when the Campbell family took up residence. There was no permanent hearth in the building at any time after Phase 1. The entrance is approached over some rough stone cobbling (22). The space between the outer northern part of the E wall and the boulder walling (12) may have been used as a storage space. In the entrance threshold exposed bedrock appears to have served in place of paving or cobbles.

The Campbell family occupied the main blackhouse HI 15A until their new cottage HI 6 was constructed in around 1870 in the lower settlement area. They may have built or repaired the roof of barn HI 16A with a small section of a boat hull, since on its later collapse timbers and iron boat rivets in (8) were found scattered, especially in the central floor space. Whether they used the barn for livestock is difficult to determine.

Phase 4

After the Campbell family left the blackhouse complex HI 15A they appear to have continued to use the barn, perhaps intermittently for shieling activities. This is witnessed by dumps of peat ash mixed with animal bone (6) and some small limpet midden deposits (7). A limpet midden (21) was possibly dumped in the cupboard space (12) outside the E entrance.

Rats continued burrowing in the walls and the resulting ejection of wall core earth probably accounts for the brown soils (9) that appear as an accumulation at this level. Their burrows are most visible along the inner edges of the wall at their junction with the floors. These show up as strips of brown soils that do not relate to the surrounding colourful ashy soils and midden deposits.

At some time in the later 19th or early 20th century the site was left derelict, although the Campbells continued to live in their cottage in the lower settlement area until 1911. The building decayed and the roof (8) fell in. Rats continued to eject wall core soil (5) into the interior and walling began to collapse. Although shown on plan the exposed bedrock at the N end probably had at least a thin film of soil and would undoubtedly have supported some grass and nettles.

The now roofless shell of the barn may have been used as a small enclosure for livestock and a sheep crush (3) was built in the NW corner at this time. It is also possible that at this time the entrance was deliberately blocked and some of the collapsing internal walling may have been repaired around the area of the sheep crush. Brown soil (5) from rat burrowing continued to accumulate and is found under and against the sheep crush structure. It cannot be determined whether the Campbell family, who could still have been using the main blackhouse HI 15 as a shieling at this time, built the crush or whether it was the later, non-resident, shepherds in the 20th century. Use by the Campbell family is the preferred option, supported by the recovery of a fragmented 19th-century soup plate. However, the extensive rat burrowing makes the contextual integrity of the site highly suspect.

Phase 7

At some time in the early 20th century, the ruined shell of the barn was finally abandoned and the rats continued to undermine much of the internal walling. The site became green as grass grew over the ruin and nettles rapidly colonised the northern part of the site. A recent camper buried the flysheet to their tent under this turf, but otherwise the soil cover below the turf is a uniform organic blackened peaty soil, often as a result of the more recent rat burrowing. Some of the collapsed stonework was deeply embedded in the soil, resting on only a thin soil deposit above the bottom level of the surrounding wall stones. At this level the soil became mixed with reddened and orange burnt peat ash bought up from lower levels and a small number of hand-made pottery sherds were recovered.

Today the rats are still in occupation and still burrowing. Walls will continue to collapse and soils will be thrown out from the wall cores. The nettles will rapidly return to their former dominion in parts of the site.

Discussion

Before the excavations HI 16 had been considered, from the survey evidence, to be a relatively uncomplicated site with a short life, possibly spanning only a single phase of use; a site whose exposed bedrock suggested that it had few if any deeply stratified contexts. Perhaps unsurprisingly these islands continue to confound such assumptions. Never densely populated, the Shiants nevertheless have a well used and highly complex landscape in which the archaeological sites, although not exceptionally numerous, more often than not present us with histories spanning several periods, each of which contains many phases. HI 16 is such a site and once again it may have been the presence of a convenient source of stone in the form of a primary structure that influenced the choice of site for the construction of a later building.

After our initial rapid survey we considered that this building could be a small, unaltered, early form of blackhouse belonging to a possible 17th- or 18th-century clachan that was thought to exist in this area. However, during the excavation of the large late 18th- to 19th-century blackhouse complex HI 15 to the N, it became apparent that at least some if not all of the buildings that were so prominent in the upper settlement landscape, including the stone-built HI 16, formed a single farmstead complex dating to the period after 1790. This was after the island's community had been replaced by a single shepherd family and therefore the buildings do not form part of an earlier clachan settlement.

Before the excavations at HI 16 the remnants of an early blackhouse (221) had been found under the garden soils in the enclosure (F) to the W of the farmstead blackhouse HI 15. The initial excavations of HI 16 in 2005 had already revealed that it was more likely to be a field barn than an early blackhouse, but that an earlier phase of activity was present below the stone-built building. The current excavations have shown that building A was indeed a field barn with several phases of use and disuse and that the supposed outhouse B was part of an early blackhouse that extended N below the barn.

Both the HI 15 and the HI 16 blackhouses are oriented N to S, in contrast to the later farmstead blackhouse, but whereas the HI 15 house is constructed of stone the HI 16 appears on the evidence of the exposed southern portion to be constructed of turves. However, the ground plan of HI 16 shows that the E wall at the southern end could be in line with the setting of large boulders (12) outside the line of the E wall of the barn. They may therefore represent a surviving fragment of the original stonework of the blackhouse, perhaps too large or inconvenient to move. The earthen bank walls at the southern end may have been the remnants of the earthen wall core left after the facing stones had been robbed out to be reused for the construction of the barn. It is also likely that much of the original earth wall core and perhaps some of the floor accumulations were reused in building the barn. This may account for the insubstantial nature of the remains, the lack of further banking along the W side and the residual handmade pottery recovered from the barn.

The drains of the blackhouse (14 and 16) indicate that stock was kept in the N end of the house, as in the other blackhouse F221 to the N, but later disturbances have removed all evidence from that area. However, this would lend weight to the possibility that the original entrance to the blackhouse was in a more central position and that the entrance to the barn was probably in the same position. The outer stones (12) appear to show a return in the correct position for them to be in situ placements of the NE door jamb of the blackhouse.

Both of the early blackhouses F221 and HI 16B are oriented N to S. Although it should not be assumed that this represents a surviving tradition, it draws attention to the building to the W of HI 16, the supposed kiln house HI 17. In the survey of 2000 it was observed that the enclosure attached to the kiln house had a disproportionately wide NW corner (E). It was considered that this was an early phase of the site and that some other substantial building had stood there before the kiln house was built. A close examination of the survey plan shows that the corner in question is end of a N to S oriented building. Perhaps the 18th-century clachan is beginning to make itself clear at last.

The main aim of this excavation was to recover some 'pure' ceramic material. While the quantity was a little disappointing, the material is very promising. The decorated 'craggan' jar is an especially rewarding find and its similarity to another decorated jar from Lewis is of considerable interest. The jar, like other excavated examples including the assemblage excavated at HI 15, reinforces the impression that the large globular milking 'craggan' has a long currency, spanning some two centuries. Within this timespan the form remains essentially the same, reflecting the function for which it was designed. There is, however, some evolution, from a much more finely made, thinner-walled vessel in the 18th century to a rather crude, thick-walled, heavily finger-moulded, massive version in the 19th and early 20th centuries. There also appears in the 17th to 18th centuries to be a greater number of thin-walled globular S-shaped rimmed jars against fewer bowl forms. The fabrics for both periods are generally very fine and all the vessels were almost certainly potted by local hands using the local Shiant materials. However, more study of the material by a specialist with more extensive knowledge of the Scottish early modern western potting traditions and material assemblages from other relevant sites is required.

NG 4188 9746 Site HI 60

This is a new site located at the littoral edge on the southern side of the bay on the W coast of Eilean an Tighe. The site is at the same location as the already recorded site HI 13, a modern sheep fank, which is marked on the Ordnance Survey map. Although we had walked around and past this site many times the disturbances of the fank had masked the earlier site. As part of the construction of the fank holding pens and dip the area had been levelled down to the hard surface of the postglacial mineralised clay and stone which covered the bedrock at this point. Without knowing the original extent of the site it is impossible to estimate how much of the site has been destroyed, but it cannot be less than approximately 75%.

The site sits at the edge of the raised beach platform and lazy beds run down to it from the incline from the S. The beds stop before reaching and rising slightly up onto the narrow strip of humps and bumps which are presumably elements of the new site. This slight disorganisation at the coastal edge must have been observed many times, but it had not previously been recognised as an unknown site. Even after the January storms of 2005, which completely erased the fank and with some of the undisturbed sea front face of the site, it was still recognisable. This is all the more galling since the storm-damaged area had been photographed that year.

The site was discovered this year purely when walking over the denuded fank platform. A small quantity of burnt bird bone was noticed in the soil face which had been cut back by the storm. These were collected along with a few sherds of crudely made coarse pottery. This led to a more thorough search and 19 sherds were found.

Although there were no rim forms fragments of a base show a pot with a flat bottom and with near vertical sides. The fabric is not exactly similar as there are only a few large stone inclusions, but the general form and poor clay preparation resembles the Early Iron Age material found at HI 15. However, this also differs only slightly from some of the early medieval Pictish plain ware pottery found on Garbh Eilean at site RI 41B. At present without more material it is considered that the early medieval material tends to be a little thinner walled and slightly better manufactured, especially in respect of clay preparation and the elimination of most of the larger stony inclusions. An early Iron Age date is currently preferred.

Archive to be deposited with NMR Edinburgh, with copy to local (Stornaway Museum) SMR by 2006. Full report available on the Shiant Isles Project home web page. www.shiantisles.

Sponsors: the Hunter Archaeological Trust, CBA Callenge Fund. Society of Antiquaries of Scotland Dorothy Marshall Bequest.

Patrick Foster, 2006.