South Uist, Bornish

NF73SW 8 729 302

NF 7294 3026 and NF 7299 3031 Field survey along the South Uist machair has demonstrated the survival of at least 80 settlement mounds in the southern half of the island. Forty-four of these have been found in an area of 5 square kilometres around the broch of Dun Vulan. They are grouped into clusters within each of the five townships in this area, Stonybridge, Ormiclate, Upper Bornish, Lower Bornish and Kildonan. On the machair of Upper Bornish there is a cluster of three mounds. Bornish 1 has not yet been excavated but it has produced pottery of Middle Iron Age (c100BC- AD500) and pre-Viking Late Iron Age date (cAD500-800). Bornish 2 and Bornish 3 have yielded grass-impressed sherds of the Viking period. The spatial patterning of the mound clusters and the apparent chronological sequence of the three Bornish mounds suggest that the origin of the township territories and organization may lie in prehistory, possibly at the beginning of the Middle Iron Age. These excavations are part of a long-term programme to test and explore this model.

Excavations on Mound 2 in 1994 demonstrated the existence of a large stone building of the 11th century, aligned E-W and containing a number of bone and iron artefacts. An earlier building appeared to have rounded rather than rectangular walls but it was not excavated. In June 1995 part of the trench was re-opened and a third stone building was located immediately above the house with rounded walls. Associated with the construction of this new building were three fragments of a decorated whalebone plaque.

Mound 3 was also excavated in June 1993 to reveal the foundations of another two Viking Age buildings, one above the other. They cannot be dated precisely nor can their full dimensions be determined without further excavation.

This excavation took place within the South Uist Machair Management Project, assessing the survival of archaeological sites, the threats to those sites and the potential management options to secure their continued preservation. The 1994 excavations demonstrated that house walls were being damaged by ploughing. One of the main lessons learned in 1995 was that the Viking Age buildings survive at a shallow depth of 20-50cm within the mounds and are very easily damaged by the burrows of rabbits, which may go down at least a metre. Since this area is heavily colonized by rabbits these sites remain under threat.

Sponsor: Historic Scotland.

M Parker Pearson, N Sharples and J Webster 1995.

NF 729 302 Three substantial mounds dominate the machair plain of the township of Bornish. Three seasons of excavations and field survey have revealed a chronological sequence dating from the Middle Iron Age to the Norse period.

Geophysical survey of mounds 2 and 3 has identified over 18 rectangular buildings which form a substantial settlement covering an area of c 0.8ha. The buildings are concentrated into four settlement foci with the two mounds simply the most prominent component of the settlement. The 1996 excavation of mound 2 revealed a complex sequence of rectangular structures interspersed with the deposition of thick midden layers, some of which contain enormous quantities of marine shell. The principal area of excavation focused on an earlier building at the E end of the original 1994 trench. This was revealed to be another rectangular building of Norse date and indicates the complexity and depth of occupation on the site.

Field and geophysical survey evidence had indicated that mound 1 was separated from the other mounds and could date to the Late Iron Age. This provisional dating was confirmed by this year's excavation. The earliest feature revealed in the SW end of the trench was a very badly damaged circular building which was the focus for some form of industrial activity in the 5th?6th centuries AD. After this industrial area was abandoned, settlement spread to the N and the well-preserved remains of a substantial multi-roomed building were partially exposed. This had been covered by heavy clay midden layers of Norse date.

The material recovered from mound 1 includes Late Iron Age and Norse pottery but a few isolated sherds of decorated pottery indicate Middle Iron Age occupation somewhere on the mound. Artefacts of note include a bronze crook-headed pin, a small broken bone plaque with an ogham inscription, a pair of bone weaving tablets, and a complete pin from mound 1. Several decorated fragments of combs and pins were recovered from mound 2. Large quantities of well-preserved animal bones were collected and flotation samples should contain a rich collection of environmental remains which will give a very important insight into the economies of both the Pictish and Viking periods.

Sponsors: Historic Scotland, University of Wales, Cardiff.

N Sharples and M Hamilton 1996

NF 729 302 Three substantial mounds dominate the machair plain of the township of Bornish. Four seasons of excavation and field survey have revealed a chronological sequence dating from the Middle Iron Age to the Norse period, and geophysical survey has revealed the plan of an extensive Norse settlement (Sharples and Hamilton 1996). The 1996 fieldwork involved the excavation of trenches on all three mounds.

The excavation of mound 1 concentrated on an area of Late Iron Age occupation. The floor of a badly preserved structure discovered last year was exposed. Despite the presence of stone piers this building appears to be sub-rectangular in plan and not a wheelhouse. The principal feature visible in the interior was a rectangular stone-kerbed hearth with one end embellished by an arc of cattle metapodials. Overlying this building was a series of charcoal-rich layers containing large quantities of burnt whalebone. Other finds include large fresh pot sherds, large quantities of hammerstones, and bone artefacts including bone combs and weaving tablets. A date in the 5th-6th century AD is indicated by the ceramics.

On top of the Late Iron Age deposits were features associated with some form of Norse industrial activity. Two hearths and a large pit contained very large quantities of a slag-like material which must have been formed at very high temperatures. To the NE the Late Iron Age deposits were cut by a rectangular longhouse.

The mound 2 excavations focused on a large structure discovered in 1994. This was a well-preserved subterranean building over 10m long and 4m wide. Over 1m of infilling occupation layers were removed down to the primary floor level in the western half of the building. These deposits were rich in artefactual and environmental material and one of the latest layers included a composite bone comb dating to the 13th or 14th century AD.

Two trenches were excavated on mound 3. The principal trench involved the complete excavation of a rectangular building, 7.2 x 4m, with a single entrance in the northern half facing E. Three distinct floor levels were identified in the interior, and immediately in front of the entrance was a sub-rectangular hearth defined by a stone kerb. A separate ash dump possibly indicating a hearth was found in the southern half of the building. A copper-alloy buckle dating to the 13th or 14th century AD was found above the final occupation layer.

The second trench was located to explore an anomaly identified by the geophysical survey. This turned out as expected to be a building, measuring 3.4 x 4.2m, with entrances through the E and W sides. There are indications that the structure explored was the northern half of a more substantial building. It was infilled with blown sand which meant that it was well-preserved, but it did not produce any material which could be used to date the abandonment.

This year's excavation has again recovered large quantities of ceramics, animal bone and small finds. Some of the datable small finds have been mentioned above but other finds of note include three small iron spearheads, a coin of Harold Hardrada or his successors, and several steatite objects including vessel fragments. Intensive environmental sampling continued with half-metre sampling of all the house floors excavated.

Sponsors: Historic Scotland, University of Wales, Cardiff.

N Sharples 1997

NF 729 302 Three substantial mounds dominate the machair plain of the township of Bornish (Sharples 1997). Five seasons of excavation and field survey have revealed a chronological sequence dating from the Middle Iron Age to the Norse period, and geophysical survey has revealed the plan of an extensive Norse settlement.

The 1999 excavations, which lasted for eight weeks, were designed to complete the excavation of the Late Iron Age house previously exposed on mound 1 and to examine the western limit of settlement on this mound; to define the extent and plan of the large Norse house previously exposed on mound 2; to examine a new area on mound 2A where a structure was being badly damaged by ploughing; and to complete the excavation of the well-preserved ancillary structure on mound 3.

The excavation of the Late Iron Age structure on mound 1 revealed a building with two distinct phases of occupation. These were separated by a thick charcoal-rich soil, containing many carbonised planks, which must represent the burning down of the original house. The house appears to have begun life as a circular wheelhouse but after destruction this was converted into a rectangular building. However, the building was not well preserved and the W end in particular was difficult to define due to recent erosion. Each house had a rectangular hearth defined on three sides by upright stone slabs. The earliest hearth had a slab floor whereas the latest was filled with ash. Surrounding the three kerbed sides of the latest hearth were lines of cattle metapodials which had been rammed into the floor with their proximal ends showing. These were carefully arranged in patterns of metacarpals and metatarsals. Large quantities of hammerstones, bones, ceramics and pottery were found distributed throughout the destruction layer, and important artefactual material includes a parallelopiped bone dice, a whalebone axe and crucible fragments.

The excavation of the Norse house on mound 2 revealed a building estimated to be 18m long and 5.8m wide with walls up to 1.3m high. This was not fully excavated - the E end of the building lay beyond the western edge of the trench. The house was substantially modified some time after the initial occupation. A separate building was constructed in the E end of the house and the W end of the original building was reoccupied. Three small sample squares were dug into the secondary occupation layer in the western half of the building. This layer proved to be extremely productive, containing large quantities of artefactual material including complete bone pins, broken combs, an elaborately decorated antler tine, a coin, many iron objects, including two bucket handles, and complete pots broken in situ. The quality of the objects in these deposits is striking and suggests some form of ritual behaviour. Above this secondary occupation the house was infilled with metres of relatively uninteresting sand layers containing a relatively low level of material remains. Occasional stone lines and arcs indicated some form of deliberate human activity but not significant occupation.

The excavation of mound 2A exposed a poorly preserved Norse structure 3.6m wide and 6m long. It had entrances to the E and S, though the latter may be into a subsidiary chamber. The floor only survived in the northern half of the structure where it was associated with a considerable amount of waste material. This included large quantities of animal and fish bones, evidence for bone comb manufacturing and a bag of iron boat nails. A large whalebone vertebra was set on the floor against the northern wall and may have served as a seat and/or chopping board. Underlying this house was an extensive ash layer containing large quantities of slag from iron smithing operations.

On mound 3 the second trench excavated in 1997 was reopened and extended S. This exposed a small kiln attached to the building, excavated in 1997, by a long passage. This kiln had been built through an earlier kiln which had a short passage and circular bowl. The two phases of kiln coincide with two phases of floor inside the building. Analysis of the upper floor, which was excavated in 1997, has revealed the presence of very large quantities of carbonised oats and barley. This structure appears to have been cut into a mound of windblown sand and represents a relatively short phase of activity near the end of the occupation of the settlement.

This year we successfully completed the excavation of the structures previously exposed on mounds 1 and 3 and excavated a threatened structure on mound 2A. However, the large Norse building on mound 2 proved to be much larger than expected and a complete understanding of this very important structure will require further excavation. The excavation recovered very large quantities of animal bones and a wide range of artefactual material. Extensive flotation has recovered a very large assemblage of carbonised plant remains and fish bones. The analysis of this database will provide an invaluable source of information on the economy in the period from the 8th through to the 14th century AD.

Sponsors: Historic Scotland, University of Cardiff.

N Sharples 1999

NF 729 302 Three substantial mounds dominate the machair plain of the township of Bornish. Six seasons of excavation and field survey have revealed a chronological sequence dating from the Middle Iron Age to the Norse period, and geophysical survey has revealed the plan of an extensive Norse settlement (NMRS NF73SW 8; DES 1999, 90-91). The 2000 fieldwork involved the excavation of areas on mounds 2 and 2A.

The mound 2 excavations focused on a large bow-shaped hall, oriented E-W and dating to the late 11th century AD. The 1999 trench was reopened and extended to the E and W to define the full extent of the structure, which can now be defined as 19.3m long and up to 5.8m wide. A considerable portion of the floor was excavated and this revealed a large central hearth area surviving as a mound of peat ash. Thin occupation layers surrounded this hearth and at the W end of the house these deposits were preceded by 15 pits. The final floor layer contained a large number of artefacts, and this year the exceptional quality of this material was emphasised by the discovery of an elaborately decorated bone cylinder depicting an animal in characteristic Ringerike style.

Extending the trench to the E revealed a later rectangular building oriented N-S. This building was 12 x 5.4m with walls standing over 0.7m high. A 2m wide slot was excavated through the floor of this structure and revealed a complex sequence of deposits preceded by a series of shallow charcoal-filled pits. After the abandonment of the house a structure was constructed in the S end. The W extension revealed that the deposit initially thought to be the primary floor of the bow-shaped hall belonged to an earlier timber structure.

Excavations on mound 2A were largely concerned with defining the complete sequence of deposits in this area. Trenches were dug to the N and E and these revealed a consistent sequence that began with sterile wind-blown sand. The initial deposit was a ploughsoil dating to the Viking/Norse period. This was overlain with a thick accumulation of grey sand, which in the NW corner of the trench was associated with structures producing large quantities of slag. These were in turn sealed by midden deposits, particularly rich in fish bone, which seemed to be associated with a sequence of domestic structures in the centre of the mound. The final deposits were associated with large quantities of antler waste from comb making.

The assemblage recovered from the site is now very substantial and artefactual material from mounds 2 and 2A has been particularly prolific. This season alone we recovered over 482 pieces of worked antler, 160 pieces of worked bone, 13 whalebone objects, 22 lead objects, 48 copper-alloy objects, 423 iron objects, 54 stone tools (24 of steatite), 5 glass beads and over 189 flints. Most of the worked antler represents debris from comb making but there are still over 100 objects made from antler and bone.

Sponsors: Historic Scotland, University of Cardiff.

N Sharples 2000

NF 729 302 The basic topography of the site has been outlined previously (DES 2000, 96-7). The excavations of mounds 2 and 2A were continued in June and July 2003. On mound 2 worked focused on the early and late houses in a sequence of three large houses.

Only the E end of the early house, dating to the 10th century AD, was available for examination. It was a large timber structure with a well-preserved floor layer that sealed a series of large pits. Artefacts associated with the occupation include a small lead cross, a copper-alloy strap end, large fragments of steatite vessel and large fragments of pottery. The house was partially subterranean and had been dug into an earlier Late Iron Age settlement. The limited excavation of the Late Iron Age occupation deposits recovered pottery and a bone pin and comb which would suggest a 7th to 8th-century date for this settlement.

The late house, probably constructed by the 13th century AD, was completely excavated. It proved to be a markedly rectangular building, 13 x 6m, distinctively different from the previous bow-walled houses. The house was oriented N-S, was defined by an internal revetment wall, and had an entrance in the E, close to the N end. The internal deposits were complex and the house had clearly been occupied for a long time. Several hearth areas could be defined by large circular spreads of ash, and these were located on the central axis eventually ending up opposite the entrance. The floor was not particularly rich in finds, but a coin of Henry III was found in situ.

The excavation of mound 2A continued from 2000. A major new discovery was the location of the principal domestic structures, located to the S of the area previously excavated. At least three buildings were found to be present in this area. To the N, the sequence exposed in 2000 was confirmed, and it was possible to expose a much larger area of the ploughsoil that underlies this mound. The importance of craft specialisation in the excavated area was emphasised by the presence of two small kilns, the discovery of more comb-making debris, and the identification of a cluster of copper-alloy objects, including a penannular brooch fragment, an ingot and a buckle, which may be a collection of scrap for remelting.

Finds recovered in 2003 include over 121 pieces of worked antler, 86 pieces of worked bone, 4 whalebone objects, 2 pieces of worked ivory, 9 lead objects, 33 copper-alloy objects and fragments, 279 iron objects (mostly nails), 37 stone tools (26 of which are steatite), and over 56 flints. Large quantities of animal bone were recovered, and as in previous years extensive flotation has recovered a large assemblage of carbonised plant remains.

The 2003 excavation has confirmed that the site has a continuous sequence of settlement that includes Middle Iron Age, Late Iron Age I and II, Viking and Norse settlement. The major period of disruption appears to be in the middle of the 1st millennium AD when the occupation moves from mound 1 to mound 2. This shift is similar to that which occurred at the Udal in North Uist and may be quite a significant historical event.

Report lodged with the NMRS.

Sponsors: HS, University of Cardiff.

N Sharples 2003

NF 729 302 The basic topography of this site has been outlined previously (see DES 2000, 96-7). Between May and July 2004 the excavations of mounds 2 and 2A were continued (DES 2003, 132-3). On mound 2, worked focused on house 2 which dates to the 11th century AD, though there was limited excavation of earlier and later deposits.

The basic dimensions of house 2 were confirmed, and this year's excavation involved the removal of the floor deposits in the eastern half of the building and the exploration of the entrance. The entrance faced S and was located close to the E wall. It was constructed on a paved area which extended under the walls and outside the excavation area. It seems that this paving was associated with the E end of house 1, which was oriented on a slightly different alignment to house 2. It represents a significant area of overlap between the two buildings. The interior of the E end of house 2 was not radically different to the W end and there was no evidence for cattle stalling in this area. The central hearth area continued up to the entrance. The quantity and quality of the finds dropped off towards the E end of the house, with the largest concentrations occurring in the centre.

These finds include some important objects. There is a group of largely complete composite combs which stand out from the broken fragments normally found. Close by, a fragment of green porphyry from Laconia was found. This is the second fragment from the site: an unrecognised piece was recovered from mound 1 in 1997. Other finds from the floor include a piece of amber, which may be from a large cross, and two cut silver coins, one a quarter and the other a half.

A trench was excavated on the edge of mound 2, and the depth of stratigraphy indicates that there is a substantial quantity of archaeology in the area surrounding the houses that have been the focus of our excavation. This work stopped after the

exposure of the remains of a substantial structure close to the base of the sequence, which may be Pictish in date.

The excavation of mound 2A continued on from 2000. A larger area of the southern half of the mound was cleared to expose another building, and confirmed that there is a sequence of overlapping houses in this area, similar to that on the other mounds. The main focus of the excavation was removing the deposits in the northern half of the mound to get down to the ard-marked natural that marks the base of the archaeology in this area. This was successfully accomplished. The distinctive nature of activity in this area was confirmed by the presence of a sequence of hearths that preceded the kilns noted in previous years. The ploughsoil was rich in artefactual material and a large assemblage of animal bone was recovered. There is no evidence for any Pictish material in this assemblage, and so it should provide an important collection of early Viking material.

The assemblage recovered from the site is now very substantial, and this season alone we recovered over 110 pieces of worked antler, 140 objects of worked bone, 10 whalebone objects, 7 pieces of lead, 33 copper alloy objects and fragments, 315 iron objects or fragments, 2 glass beads and a counter, 35 stone tools including a substantial number of whetstones, and over 170 flints.

This was the last year of excavation at Bornish. The vast quantity of material recovered over the last decade requires detailed analysis and publication and any extensive delay would compromise the analysis. The remains of house 2 and house 3 on mound 2 have been completely excavated and we have bottomed a large area of mound 2A. The stratigraphic integrity of the recovered assemblages will provide a solid basis for the analysis of cultural change during the Norse occupation of the Western Isles. However, much more could be done and it may be desirable to return to the site in the future.

Archive to be deposited in the NMRS.

Sponsors: HS, University of Cardiff.

N Sharples 2004