Image of Tarbat Sculpture.Bulletin 6, 2000


The programme of analysis is aimed at establishing the data, structures and assemblages of the stratigraphic sequence. A major part of the work is being carried out in partnership with the National Museum of Scotland, co-ordinated by a research advisory group which meets twice a year.

The analytical programme includes the following areas of research:

[All enquiries through Cecily Spall please]

Interim accounts are presented here on artefact analysis and radiocarbon dating

Artefact Analysis

Cecily Spall

The Tarbat assemblage was last reported after the 1998 excavation season (Bulletin No 4). Since that time two full seasons of excavation have unearthed a growing assemblage of early medieval craft-working debris. Amongst the industries represented at Tarbat are metal-working both ferrous and non-ferrous (including gold and silver), leather-working, wood-working and possible glass-working.

Many artefact types found in abundance at Tarbat may well have been used in several different crafts and as such will be presented as stand alone assemblages, for example, the range of whetstones, burnishers and stone rubbers appear to cross many craft divisions. Other artefacts are specifically related to certain crafts and these assemblages will be presented by industry


The metal-working assemblage is dominated by ferrous slag presumably from smithing and is complemented by the presence of spherical and flake hammerscale in many of the deposits from workshop areas. Crucibles are the greatest indicator for non-ferrous metal-working within the assemblage. Many have been recovered whole and are in the typical unlidded bag-shaped crucible form (cf Dunadd Type D, Campbell and Lane 200, 143) although a fragmentary triangular crucible has been found. Both types have the external red glassy adhesions of copper oxides indicative of bronze-working. Fragments of many other crucibles have been found and these would appear to take the form of shallow dish-shaped crucibles, there is also a much larger example present of which only a large handle fragment remains.

An assemblage of clay moulds and metal-working vessels complement the assemblage of crucibles. A corner fragment of a shallow rectangular vessel has been identified as a heating tray used for the separation of precious metals (Andrew Heald National Museum of Scotland pers.comm.). Many fragmentary two-piece casting moulds have been found and a range of products is represented in the assemblage. Ingot moulds, plain pin and ring moulds, moulds for decorative items such as discs or studs and one possible cruciform object have been recovered. These are largely of medium fired clay although two ingot moulds are also made of stone.

metal-working objects
Fig 9: Moulds and crucibles, examples from the metal-working assemblage

One large stone mould recoverd in two halves has an inset in the form of a "hand mirror ". This has a parallel with a stone mould from Garranes that is smaller but has a comparable " hand mirror " - shaped inset with a cross ornament (O'Riordain 1941, p108 Fig 10 No 445). A parallel mould from the Irish monastic site at Nendrum, Co.Down has a " hand-mirror " shaped inset with no cross ornament which would be closer to the Tarbat example. The author, H C Lawlor, suggests that sheet metal may have been hammered into the mould to form a creuset lamp (Lawlor 1925,135). It is possible that the Tarbat mould is associated with hand-beaten metal rather than the casting of molten metal given the potential quantities involved.

the hand mirror mould
Fig 10: The hand mirror mould

One significant group of metal-working debris included within the site-wide assemblage has been recovered from primary dumping against a terrace wall. These dumps were associated with a rectangular, slab lined hearth situated on the terrace. It is possible that this assemblage was deposited directly over the terrace wall from the metal-working area above.

A copper alloy stud with a corroded iron pin at the back and possible silver or tin plating on the front has been recovered although it is not clear whether this object was a product of the Tarbat workshops. The stud has six double spirals connected with " C " curves and its circumference consists of a ring of small pellets contained within a plain moulded border. It has close parallels with other discs and studs found at other sites, for example a bronze mount from Clatchard Craig also with six " peltas " connected with a C curve spiral dated to the late 6th to mid 7th century (Close-brooks 1986, 168-9, Illus 28-29 No 122). Similar discs decorated with trumpet spirals and interlace have also been recovered, for example, two decorated discs from Dunnadd dated to the 7th century (Campbell and Lane 2000, Illus.4.59 and Illus.7.10). Parallels can be found in Merovingen metalwork, for example an interlace die for stamping jewellery from a metal-worker's grave at Liebnau (Cosack 1995,46) and a 7th to 8th century brooch from a female grave in Neuses an der Regnitz in Bavaria (Haberstroh 1997,42). This brooch is made of tinned bronze and also has a pelleted ring around its circumference containing interlace decoration. The brooch and its decoration have been interpreted as having Christian symbolic content and may have been worn by a member of the Christian mission in the eastern Frankish kingdom. This leaves little doubt that the Tarbat stud is part of a high status, highly decorative object perhaps a brooch or " button " similar to that recovered from Garranes (O'Riordain 1941, Fig3 No231).


A small but significant assemblage of leather-working artefacts have been recovered from the excavations all of which appear to relate to a single workshop area. An iron "lunellum" or crescentic leather-working knife was recovered as well as three pumice rubbers. The pumice rubbers, unlike a parallel pumice rubber from the leather-workers toolkit from Evie in Orkney, are finished, shaped objects all with circular perforations for suspension from a belt or neck. A possible leather-working needle of bone has also been recovered. The workshop area associated with this assemblage included dumps of burnt seaweed which would have been highly alkaline. Alkalis are used in the tanning process thus supporting the interpretation of the artefactual assemblage.

leather-working tools
Fig 11: Leather-working tools

There is as yet no indication of the quality of the leather being produced but the relative quality of the tools in the assemblage found so far would suggest a specialism, perhaps vellum or parchment production.This is particularly so given that pumice is used in the final stages of vellum or parchment production. Within this workshop floor small clusters of white quartzite elongated pebbles were also recovered. They were present to the exclusion of other gravel or pebble inclusions and were generally uniform in size tending to be no longer than two or three centimetres. The production of parchment involves stretching skins on a wooden rectangular frame. Small smooth pebbles are used to form knobs in the skin which are then secured with cord and lashed to the wooden frame. The skin is then stretched from the knobs and tightened on the frame with pegs, to stretch the skin taught. This avoids piercing holes in the skin which would stretch and tear as the skin dried out (de Hamel 1992, 11). It is possible that these small stones are being specifically collected for such a purpose.


Evidence for wood-working is less tangible than that of leather-working. Iron objects recovered from an occupation layer within a workshop area exhibit mineralised wood shavings within their corrosion layers, face down as the object was deposited. This clearly indicates that they were deposited onto a surface rich in wood shavings, the assumption being that the wood shavings were created within that workshop and were not imported. One iron object recovered displays mineralised wood shavings and has been tentatively identified as a wood-working tool. The object appears to be a long-handled chisel although this object is awaiting x-radiography.


The evidence for glass working is still ambivalent. The sherd of glass vessel reported (Bulletin No 4 p20) may have been from a vessel imported to the settlement that was then broken or it could have been imported as cullet and missed the recycling procedure although this interpretation has lost favour in recent years. During the 2000 season a small cylindrical blue glass bead with lost zig-zag inlay was recovered. Again it is unclear as to whether this was a product or an import. A small spacer bead measuring no greater than 2mm diameter was also recovered. A possible lump of raw blue glass and a fragment of blue glass with fired clay adhesions are perhaps better indicators for site production. However, they were both recovered from a secondary context backfilling the inner enclosure ditch and as such are separated from a workshop context. The deposit that yielded these two finds also contained two crucial objects.

glass mould
Fig 12: Possible glass mould

One circular clay mould with a floreate design has parallels with symmetrical compass designed moulds found at Iona (Reece 1981, Fig III.I). It is possible that the Tarbat example was a one piece mould used to manufacture a glass stud, although the circumference of the decorated area has a possible recieving ring for a mould back and a two piece mould might indicate metal rather than glass-working (Dr Justine Bayley, Ancient Monuments Laboratory, pers.comm.).

A blue glass mould with inlaid metal wire and polychrome glass was also recovered. The form of the metal wire remains unindentified awaiting x-radiography since the design of the wire within the glass is obscured by the decay of the glass inlay. The object is damaged but otherwise would appear to be finished and it is possible that this was a product of the Tarbat workshops. This glass stud has parallels with highly decorative liturgical objects from Early Christian Ireland and from glass working debris from Garranes and more particularly with a light blue glass stud from Dunmisk Fort, Co.Tyrone (Henderson and Ivens 1992,Fig.4;1). A fragment of a shallow, hard-fired clay vessel with a coating of opaque yellow glass has also been recovered indicating that glass-working may have been carried out at Tarbat. Similar finds from Frisian and Merovingen sites have yielded shallow trays with adhesions of lead rich opaque yellow glass associated with glass-working although the Tarbat example awaits analysis. The example from Wijnaldum was, like the Tarbat example, recovered from debris from metal-working suggesting these crafts took place in close association with one another (Saberolles in Bestman et al 1999,280).

Stone artefacts

A growing assemblage of stone artefacts consists of utilised pebble whetstones and finer imported whetstones and burnishers. They range in size, form and material considerably. Some are in the local red sandstone and some utilised pebbles could easily have been gathered from the surroundig coastline. However some are clearly not made of native stone and most are yet to be identified. Two examples have been identified as Norwegian schist (although not more precisely) (Colleen Batey, pers.comm.). Another example has a scratched interlace graffito perhaps of the owner although it is too crude to be considered as a motif piece. Whetstones with motif pieces are known from Ireland and are often used after the motif has been incised (O'Meadrha 1979, Plate27 Nos 66074). Included within the assemblage are two smooth and finished circular stone discs sometimes referred to as potlids. The favoured interpretation of these at Tarbat are as whetstones or rubbers since one example was recovered with an assemblage of whetstones, honestones and rubbers that had been discarded in an occupation layer within a workshop area.

The definition and excavation of the early medieval workshop complex is now well underway at Tarbat. The assemblages defined here by craft association would also appear to have a spatial relationship on site. The workshop strata that await excavation are crucial in terms of further defining the craft processes at work not only from the tools being used, or broken or finished artefacts but also from craft working debris and the raw materials involved. The picture that has emerged from Tarbat over recent years is one of a highly organised and diverse production centre. The products of which may have been glimpsed in the form of the glass and metal stud and the decorative stud mould.

Radiocarbon dating

A preliminary series of radiocarbon dates have been determined as follows:

Oxford University radiocarbon accelerator unit

OxA-9662. Charcoal from the ultimate backfilling of an inner enclosure ditch in Sector 1. (Int 11/F18/C1143)----890 (68%) 985 AD Early 10th century.

OxA-10159. Wooden stake in situ in the side of the outer enclosure ditch in Sector 1. (Int 11/F158/C1490)----690 (68.2%) 780 AD Early 8th century.

OxA-9699. Skeleton from one of the earliest long cist graves in the church. (Int 20/F152/C1373)----535 (65.8%) 605 AD Late 6th century.

OxA-9664. Burnt wood from the destruction layer over workshops. (Int 26/C1030)----400 (68.2%) 540 AD Late 5th century.

Scottish Universities Research and Reactor Centre

GU-9298. Skeleton with head wound from cemetery in the church. (Int20/F93/C1222)----1189 (1 sigma) 1258 AD. Early 13th century.

GU-9296. Skeleton with head wound from cemetery in the church. (Int20/F138/C1238)----733 (1 sigma) 886 AD. 8/9th century.

GU-9297. Skeleton with head wound from cemetery in the church. (Int 20/F132/C1307)----890 (1 sigma) 981 AD. 9/10th century.

Preliminary assessment

These results seem to justify an assumption that the site as a whole will be found to begin in the later 6th century. The outer ditch was in use by the early 8th century. An inner ditch, which may have been earlier, was still visible as a shallow depression which acquired glass debris and charcoal with a C14 date in the early 10th century. The end date of the site is difficult to determine. In an object lesson in the difficulties of selecting material for radiocarbon dating, samples of large pieces of wood were collected from the charcoal layer sealing 8th century artefacts which covered much of the workshop area and had been attributed to a Viking raid. They gave a date in the early 5th century. This is explained by the samples being of heartwood (so less well burnt) which had died in the tree in the late 5th century, cut for building in the late 6th century. The timber could still have been burnt in the 9th century. New dates will be sought using sieved identified macro fauna and flora which will have died at or near the date of the fire.

NEXT: Interpretation
BACK: Excavation

Last updated 10 October, 2003.
Contact the Tarbat Discovery Programme

Image of Field Archaeology Specialists Logo Image of TDC logo.

Image of The University of York logo.
© 2003-2006 University of York

Link to W3C XHTML 1.0 Validator

Link to W3C CSS Validator

Level Double-A conformance icon, Link to W3C-WAI Web Content Accessibility Guidelines 1.0