Publication Account
Date 2002
Event ID 575397
Category Descriptive Accounts
Type Publication Account
Permalink http://canmore.org.uk/event/575397
HU43 4 SCALLOWAY ('Upper Scalloway')
HU/406399
Probable solid-based broch in Dunrossness which was discovered as a result of human bones having been found during work on a housing development; it was excavated during generally bad weather in the winter of 1989-90 [1]. There had been no knowledge of a broch on the site previously, which suggests that there could be many more of these structures than is known at present.
The excavation has been thoroughly reported and, since in addition structural remains were minimal, only a summary of the discoveries is given here. Some very important dating evidence was recovered.
The site stands on a limestone ridge running N-S. The broch itself had been almost completely destroyed, a complete cross section of the wall being preserved only on the E side. The excavations revealed five phases of occupation.
Phase 1 was a cremation burial and various structural features of unknown pre-broch date. Phase 2 saw the construction and occupation of the broch and its outer defences during the middle Iron Age period. This period was ended by a destruction by fire. Phase 3 was the secondary occupation of the broch and an adjacent settlement during the late Iron Age, while Phase 4 saw the use of the site as a cemetery and Phase 5 the construction of a Medieval farm. Only Phase 2 is discussed here in more than outline; the material culture and dating of Phase 3 is commented on because of its relevance to the length of the Phase 2 occupation.
The wallfaces rarely stood to more than two courses high and most of the rubble core had gone. The interior appeared to be slightly pear-shaped towards the east and the foundations of one intra-mural cell were found on the south-east. No clear traces of the entrance passage or of an intra-mural stair could be found (although the latter doubtless started at first floor level). There was a paved hearth in the centre of the court and the foundations of a secondary wall, built against the broch inner wall face, had sealed part of the primary floor deposits; the excavator is confident that he had isolated most of this layer over the interior. There were some pits in this floor.
The broch's outer defences consisted of a ditch cut across the promontory on the northern approach. There was also a hollow between this ditch and the broch which yielded a number of finds.
A severe fire at the end of Phase 2 caused the deposition of a red ash layer which was diagnosed as the remains of a collapsed and burned roof. Two C-14 dates for this ash layer imply that there is a 98% chance that the fire took place between about AD 345-505 and AD 415-525. A large number of artefacts were found in the primary floor levels, nearly all of which were in this burnt layer; the assumption is therefore that the broch was regularly cleaned out while occupied. A very large amount of burnt barley was also found, suggesting that the building had served as a grain store.
Dating: a total of 24 radiocarbon dates was obtained for stratified organic samples from the site; none were from the pre-broch period [1, Table 24). When arranged in stratigraphical order [1, fig. 71] they give a fairly clear picture of the chronology of the site's history.
The earliest stratified dated sample for Phase 2 also gave the earliest date which falls in the 1st centuries BC/AD; this is exactly what is to be expected if one adopts the 'short' chronology for hollow-walled brochs (see Section on dating in Part 2). The rest of the dates, including the two specifically from the destruction deposits, are scattered through the first half of the first millennium AD but there is a good cluster of three which shows fairly clearly that the middle Iron Age occupation probably went on until about AD 500.
The post-broch phase 3 was divided into three stages on the basis of the excavation evidence and in broad terms these stages seem to have lasted from about 500-650, from 650 to 900 and between about 900 and 1000 respectively. Presumably the broch itself, although burned out once and presumably much reduced in height, must have been roofed and standing to at least 3m in height until about AD 1000.
Finds from the end of Phase 2: these are important because they give us a picture of the material culture of the inhabitants of a Shetland broch probably at the end of the 5th century AD. However the way the artefact appendices, and their accompanying illustrations, are organised in the report (to illustrate various economic activities) does not make it easy to get a quick and clear impression of what finds are characteristic of each Phase.
Pottery includes some Everted Rim ware of both the neck-band (fig. 80, 6 and 7) and the fluted rim varieties (fig. 80, 5). There were also some simple barrel-shaped pots, including an almost complete example with a remarkable pattern of vertical and diagonal fluting on its exterior (fig. 80, 1). This last also had an internally decorated base.
Metal objects include a piece of a bronze bowl (fig. 85) which may be Roman;
Carinated pottery: there is one strange phenomenon which, though commented on [1, 133], is not really explained in the report. From deposits confidently allocated by the excavator to late Phase 3 (end of the first millennium AD) comes some talc-tempered pottery which looks strikingly like the early Iron Age carinated ware from Clickhimin and Jarlshof which should date to the 7th/6th centuries BC [1, fig. 83). In particular the sharply carinated wall sherd no. 5 has decoration which is quite similar to that on a similar piece from Clickhimin (Hamilton 1968, fig. 00, no. 00). Presumably some mixing of deposits must have occurred during the extensive destructions and re-buildings which took place on the site. Thus there may have been an important early Iron Age occupation on the site which could not be recognised among the stratified layers.
Discussion: the
Dimensions: external diam. 19.6m, internal diam. 7.8-8.4m, wall thickness 5.5-6m; thus the wall proportion would vary from about 56-61%, probably above 60%; this at the upper range of the measurable Shetland sites [1, fig. 34].
Sources: 1. Sharples 1998.
E W MacKie 2002
