Westray, Quoygrew

HY45SW 4 4433 5065.

HY 443 506 In June and July a programme of renewed reconnaissance and archaeological sampling began at the site of Quoygrew, known as Nether Trenabie in 19th-century sources. A 2 x 1m test-pit had been excavated in coastal middens by Sarah Colley (pers comm) in 1978. Although further eroded, these deposits now stretch 40m along the wave-cut bank and augering revealed that they continue inland for c 15m. The central 20m of the exposed profile (1?1.3m in thickness) was drawn. Three 50cm sample columns were excavated for fine recovery of fish bone, botanical macrofossils and terrestrial molluscs. Soil micromorphology samples were taken. A 1 x 2m test-pit was also excavated to recover a sample of mammal bone and datable artefactual material. A further 2 x 5m test trench examined a small drystone ?cellar? dug into the midden. This structure, which had a flagged floor and stone-lined drain, may represent a naust or a semi-subterranean byre.

The midden can be dated to the Viking Age by antler comb fragments and steatite vessel sherds. Other finds include bone pins, a highly corroded possible coin, a possible bone gaming piece, and perforated discs made from the unfused vertebral epiphyses of a small whale. Sherds of coarse pottery from the upper strata may imply that deposition continued into the Late Norse (medieval) period. The fill of the structure dug into the midden included coarse pottery and low-grade coal which may suggest a post-medieval date.

To landward, geophysical survey by Lorna Sharpe of Glasgow University revealed a possible rectangular structure, c 18 x 7m. Auger survey and soil test-pits delineated additional middens of unknown date and a relict anthropogenic topsoil over 90cm deep in places. The latter covers an area c 80 x 200m in maximum dimension and can probably be interpreted as an infield (tunmal).

J Barrett, I Simpson and T Davis 1997

House and outbuildings.

A coastal exposure containing extensive archaeological deposits coincides with a slight rise in the ground level in the hinterland area. The exposure extends for about 30m and stands up to 1.75m high. Structural deposits, including walling and drains, are visible among substantial midden deposits. In the hinterland, there are topographical indications of further burial remains over an area of some 25m parallel to the coast and up to 10m inland. There are kelp pits of 18th-20th century date cut into the turf above these deposits.

Orkney Coastal Zone Assessment 1998.

HY 443 506 Following previous survey and sampling (DES 1997, 61) ten weeks of excavation, opening c 150m2, were conducted in July and August, 1999 and 2000. This work has identified a well-preserved late Viking Age and medieval rural settlement (NMRS HY45SW 4). It covers an area of over a hectare and includes a coastal fish midden, a 6.5m by c 18m house with undisturbed ash floors, a 'farm mound' of kitchen midden, and a plaggen infield. Radiocarbon and artefactual evidence suggest that these deposits are broadly contemporary. All but the infield (which may only be associated with the latest phase) appear to date to between the 10th and 13th centuries.

The eroding fish midden at the shoreline, composed of peat ash, shell and fish bone, was mapped and sampled in 1997, but has since been dated to the 10th-12th centuries. A cellar or naust dug into it is now known to pre-date the 13th century.

Approximately 14m inland from the cellar, and in line with it, is a rectangular drystone building (Area F) aligned roughly NE-SW. The excavated central section has an external width of 6.5m, an internal width of 4.8m and a length greater than 11.8m. The exterior walls, c 0.8m thick, survive to a maximum height of 0.6m. They are double-faced with a rubble core. There is not yet clear evidence for an outside door. A secondary cross-wall divides the excavated area into two rooms connected by a doorway (near which was found a pivot stone). A primary cross-wall which runs into the W-facing section could represent the E end of the structure. It is more likely to be another internal division, however, given that geophysical survey suggests that the building continues 6m or more in this direction. Its W end also lies beyond the excavation and may be associated with the cellar exposed in the wave-cut bank.

The E room is 4.8 x 5.3m internally, with a stone-built bench 0.5m wide and 0.3m high along the S wall. Its E end is constructed of flat slabs laid horizontally, but towards the W the bench is constructed of orthostats backfilled with rubble and topped with flat slabs. The central feature of the room is a square hearth, 1m square, comprising a single fire-cracked flagstone surrounded on all but the W side by orthostats which jut above the central slab by up to 0.1m. This hearth was covered by a localised spread of pure peat ash and was embedded in an earth floor composed of hard-packed and finely laminated ash. The floor both pre-dates and post-dates the hearth. It supported the orthostats, but was thickest just W of the open side of the hearth and may have been augmented by repeated sweeping of its contents. An earlier hearth, not yet completely excavated, lies under this thickened area. The ash floor provided a smooth, but not flat, surface of uneven thickness over a roughly laid flagstone sub-floor. A lintelled drain lies under the flagging. It runs slightly diagonally, from NE to SW, along the length of the room.

At least two phases of stepping stones lead from the internal doorway into the E room. They were probably replaced as earlier ones became lost in the build-up of the floor deposit. These flags end at two parallel orthostats which act as a low threshold over which one must step to enter the centre of the room. They may once have secured a higher threshold board. A series of small orthostats in two parallel rows also mark off a rectangular area, 1 x c 2m, along the internal dividing wall at the W end of the room. They are embedded in the latest phase of the earth floor. Their function is unclear, but could have been to secure a wooden feature of vertical or horizontal planking. Prior to the construction of this feature, a circular, U-shaped pit measuring c 0.85m in diameter and 0.4m deep had occupied this area. It was filled with rubble and sealed prior to the last phase of the building's use. Several large stones around the edges of the room may have supported internal furniture or posts. It is equally possible, however, that they represent elements of the destruction rubble which became pressed into the relatively soft earth floor.

Large numbers of sherds of medieval coarse wares were found lying on and in the floor layers. Several clusters are likely to represent vessels crushed by an overlying stratum of destruction debris. An unfinished steatite line sinker was also found in the ash floor (sherds of circular steatite vessels were also discovered in sheet midden which built up around the outside of the building). A fragmented schist baking plate was found in a section through the core of the bench. These finds are consistent with a c 12th-century date, but radiocarbon assays and further analysis of the artefacts could refine this estimate. The floor also contained a rich ecofactual assemblage which was extensively sampled.

The E room is generally well-preserved, having been partly sealed by the overlying destruction layer, but a later ditch has removed one lintel of the bench and created a negative feature originally interpreted as a barrel impression. Conversely, the W room was heavily robbed in antiquity. Only the lowest courses of the exterior walls and remnants of an uneven flagstone sub-floor survive. It is likely that the bench along the S wall continued into this end of the house prior to construction of the secondary cross-wall, but it does not appear that any other interior fittings remain in situ. The sub-floor drain has not yet been located in this room.

Approximately 20m E of the house is a low mound reaching a height of c 2m. A modern croft and garden, abandoned in the 1930s, sit on top of the mound. A 6 x 7.6m area (Area G) was excavated in the garden. The maximum depth to subsoil was found to be 1.6m. The top c 0.7m was composed of homogenised garden soil, under which lay undisturbed middens with exceptionally good bone preservation. The midden deposits were finely stratified, but could be divided into two distinct blocks based on the ratio of fish bone and shell to mammal bone. The upper c 0.5m was dominated by marine resources, the lower 0.3m by mammal bone, including several cattle skulls. A single radiocarbon assay of AD 1004-1262 (AA-39135, 905+60 bp cal 2s) was obtained for a horse bone from the interface between the two. Steatite vessel sherds, bone or antler pins and a fragmentary bone or antler comb from the midden are all consistent with accumulation in the late Viking Age and early Middle Ages. A flagstone path runs up the side of the mound in the upper stratigraphic block, but no other architectural features have yet been found. The middens were extensively sampled for ecofactual and sediment analyses.

An 0.8 x 8m trench was also excavated as a southward extension of Area G in order to establish the relationship between the edge of the farm mound and an adjacent relict infield located in 1997. The field soil was found to overlie the Norse middens, suggesting that it is of medieval or post-medieval origin. Micromorphology samples were taken of the interface between these features. Seven additional test pits were also dug to clarify the distribution of this infield and to collect samples for OSL dating from locations where it is associated with deposits of wind-blown sand. A topographic survey of the site was also completed and an auger survey of its hinterland identified additional (presently undated) relict fields and settlement sites.

Sponsors: Historic Scotland, Social Sciences & Humanities Research Council of Canada, Orkney Islands Council, Hunter Archaeological Trust, Society for Medieval Archaeology, University of York, University of Glasgow.

J Barrett, H James, T O'Connor and S Dobson 2000

HY 4433 5065 Continued excavation aimed to expose the eastern end of the medieval to post-medieval building partly exposed in 1999 (DES 2000, 69-70), and to further sample Viking Age and medieval middens of the adjacent farm mound. The structure proved to have a primary entrance in the centre of the E gable. This doorway later provided access to a secondary extension in line with, but slightly narrower than the original building. The external dimensions of the addition were 7.2 x 5.8m, its internal measurements were 5.6 x 4.3m. The extension originally lacked an external door, but one was inserted near the SW corner in a late phase. The floor deposits consisted of superimposed layers of (now largely robbed) flagstone, redeposited midden and ash. This area must originally have functioned as a dwelling space, but seems to have served as a workshop or agricultural building in later phases. A Dutch duit/doit of Batenburg, struck between 1616-22 was found in the floor and two billon pennies of James IV, second issue, c 1500-10, were recovered outside its walls. Imported pottery associated with the extension includes sherds of Scottish Red Ware of probable 13th to 15th-century date.

The upper stratum of the farm mound middens yielded a single sherd from a Borders straight-sided cooking pot of White Gritty Ware, suggesting a 12th-century date. These deposits were otherwise aceramic, but produced large numbers of steatite vessel sherds. They were excavated to subsoil and intensively sampled for ecofactual analysis.

Report to be lodged with the NMRS.

Sponsors: Historic Scotland, University of York, Orkney Archaeological Trust, Orkney Islands Council.

J Barrett 2001

HY 443 506 Excavation continued in July and August 2002 with the aim of fully exposing and excavating a sequence of medieval to post-medieval buildings (DES 2001, 73). Area F was extended to 10 x 25m in order to locate the W gable of the primary building (structure 1) and to relate it stratigraphically to eroding deposits at the wave-cut bank. The E half of structure 1 was very well preserved in all phases. Its W end was disturbed by a shelving negative feature, but this robber trench itself marked the position of the W gable and a few stones remained in situ. It was thus possible to ascertain that the original building was 10 x 4.8m internally, with entrances in the centre of both the E (previously excavated) and W ends.

Approximately 1m W of the western gable of structure 1 another building (structure 3) was discovered in direct line with it. This building overlay the Viking Age middens eroding at the shore. Presently, however, only the tops of the walls are exposed and the stratigraphic relationship between it and structure 1 remains to be clarified.

Excavation of the interior of structure 1 exposed and recorded four successive phases of occupation, each represented by one or more hearths and other internal fittings (including flagstone paths, pits, orthostatic features and a wall bench). Three phases were above a sub-floor drain which was previously assumed to represent the original construction of the building. In the event, this feature proved to be cut into earlier floor levels which continue for another c 30cm in depth based on the sections of two internal pits.

To the E of structure 1, excavation of a secondary extension (structure 2) first discovered in 2001 was completed. This room seems to have lacked an exterior door in its earliest phases. Its early floors consisted of a series of superimposed layers of redeposited ash and flagstone. The main internal features were a central post-hole and a series of rectangular stone settings along the N wall (set into subsoil and even bedrock) which must have supported a substantial stone or wooden superstructure. The room lacked a hearth. The later phases of structure 2 have been discussed elsewhere (DES 2001, 73).

The buildings produced numerous finds of pottery, soapstone and whalebone. Notable examples include 25 sherds of Scottish redware probably dating to between the 13th and 15th centuries, and a square-sided steatite vessel of Shetlandic type. Three hearths in structure 1 were sampled for archaeomagnetic dating and the house floors and external deposits were both extensively sampled for ecofact and sediment analyses.

Excavation of the adjacent farm mound (Area G) did not continue this season. However, a radiocarbon date of AD 780-1000 (AA-50702, 2s) on a pig skull from the basal stratum confirms a Viking Age date for these deposits. Work at Quoygrew this year was associated with continued field and auger survey of its hinterland.

Archive to be deposited in the NMRS.

Sponsors: HS, University of York, Orkney Archaeology Trust, Orkney Islands Council.

J Barrett, J Gerrard and T O'Connor 2002

HY 443 506 During this season the excavation of a long-lived medieval building with four rooms (structures 1-4 in Area F) was completed (see DES 2002, 89-90). This work also clarified the stratigraphic relationship between it and eroding middens at the wave-cut bank recorded in 1997 (DES 1997, 61). Moreover, it revealed an underlying building (structure 5) on a different alignment (roughly E-W) associated with aceramic strata yielding steatite vessel sherds.

The multi-roomed medieval building is of 12th/13th- to 16th-century date. The hall room (structure 1) had nine major phases, each characterised by one or more sub-phases of hearth setting. The underlying building (structure 5) is of Viking Age or early

medieval date. Two trial trenches S of Area F (Areas J1 and J2) revealed a roughly paved yard enclosed by a boundary wall.

An additional trial trench (Area G3) was excavated adjacent to a past intervention (Area G1) on the farm mound approximately 30m inland from Area F. This revealed the intact walls of two or three buildings associated with Viking Age and early medieval midden deposits. The two indisputable structures overlay one another.

The earlier of the two was a semi-subterranean building incorporating both coursed and upright stones in its revetment walls. It included a hearth and produced a fragment of an antler comb and a steatite vessel sherd. Only some of the post-abandonment fill of the later building, which was of massive coursed stone construction, was excavated. It was aceramic, but produced two steatite vessel sherds. Another exposed wall and resistivity survey strongly suggest that a third building existed adjacent to the later of the two obvious ones. In this case, the late buildings would appear to represent contemporary and parallel structures with touching side walls of the thick and curving style associated with Viking Age architecture. Work at Quoygrew this year was also associated with continued auger survey of its hinterland.

Archive to be deposited in the NMRS.

Sponsors: HS, British Academy, Orkney Islands Council, Society for Medieval Archaeology, University of York, Orkney College, Orkney Archaeological Trust.

J Barrett and J Gerrard 2004

Late Viking Age and medieval settlement HY 443 506 July and August 2005 saw the final season of excavation in Area F near the eroding shoreline (see DES 2004, 97-8). Work focused on completing the excavation of structure 3 (part of a medieval sequence of end-on-end buildings including structures 1 and 2) and structure 5 (an underlying house newly

discovered in 2004).

Structure 3 was a room with original internal dimensions of 4.9 x 7.9m and drystone walls c 1m thick. On exposure of its primary phase, it proved to be a byre built together with structure 1, probably around AD 1200.

Under this byre was an earlier house on a different alignment (structure 5). It was internally of 3.9 x >8.5m with drystone walls, c 1m thick. The long walls were very slightly curved and the eastern gable significantly bowed. The western gable was obscured by the end of structure 3 (the walls of which were left in situ for public display), but the house could not have been more than 10m long in total. The remains of structure 5 may represent a stone platform for a turf superstructure, as there was no stone demolition material between it and the immediately overlying structure 3.

Structure 5 produced three recognisable phases of internal organisation, during which the main doorway moved from the centre of the S wall to the SE corner. The internal use of space followed the Icelandic pattern of slightly raised side aisles, demarcated by edge-set stones in this instance. The earth floors were aceramic, but yielded sherds of steatite vessels, antler combs, a piece of a weaving batten, fishing weights, a fishhook and hones of Eidsborg schist. Based on the combs, radiocarbon dates from adjacent middens and finds from the overlying stratigraphy, this house is likely to be

of 11th/12th-century date. The 12th century is most probable if the putative turf walls were not subjected to regular repair.

Archive to be deposited in NMRS.

Sponsors: HS, HLF, Orkney Islands Council, University of York.

J Barrett, J Gerrard, J Harland 2005