In Depth
A Consideration of the Upper Scalloway Broch Excavation
by Stephen Jennings
Niall Sharples’ excavation of Upper Scalloway, Shetland through the winter and early spring of 1989 to 1990 proved significant to our understanding of broch economy and social structure. Within a local context, it was the first broch to be extensively excavated in Shetland since Clickhimin between 1953 and 1957 and just prior to the full excavation at Old Scatness beginning in 1995. It remains one of the few brochs in Scotland thoroughly excavated to floor level. With just a few courses of the original wall still intact, the dig at Upper Scalloway was undertaken as a rescue when several skeletons were discovered during levelling for a housing development. Though this placed serious restrictions on the extent and time devoted to unpicking the complex architecture and chronology of the site, using sound methods and a controlled multi-disciplinary approach, the excavation was completed in a matter of weeks and the interpretation deemed largely secure. However, a careful interrogation of information released prior to the final excavation report, the report itself and more recent discoveries when taken together leave some lingering questions.
​
Built sometime between the 1st centuries BC and AD, the broch occupies a prominent ridge at the south end of the fertile Tingwall Valley. The constricted interior diameter of the slightly oblong structure, 7.8 to 8.4m, with a wall thickness of 5.5 to 6.0m defines it as among the stoutest of known brochs. A thick red ash layer sealing the early deposits indicates it may have been consumed by a large conflagration sometime between the mid-5th and early 6th century bringing an end to its primary occupation. Whilst there is evidence of prior site activity in the form of a Bronze Age cremation urn (Figure 1) and enigmatic parallel grooves found beneath the broch construction layer, the preponderance of activity is middle and late Iron Age. Reoccupation of the broch superstructure and surrounding area encompasses several subsequent phases of building activity (Figure 2) some of which is then overlain by 24 skeletons from the 14th and 15th century presumably belonging to a chapel cemetery somewhere in the vicinity. In all, 24 radiocarbon dates were taken from what was determined secure contexts throughout the site and a variety of artefacts, many exceptional and unique, were found (Sharples 1998).
One of the most important discoveries was roughly 10,000 charred grains of barley from which a radiocarbon date of 415-525 AD helped define the end of primary occupation (Sharples 1998: 126). The fire that consumed it is thought to have been fueled by the internal wooden structure and roof. It brought the temperature to an extreme of at least 650 degrees Celsius, enough to recrystallise the marble rock fragments in the underlying floor layer at a depth of up to 8cm (Ibid: 30). The very presence of barley in such quantity is an indication it was being stored in the broch. Sharples interprets this as evidence of settlement hierarchy and the central store of grain as agricultural control of the Tingwall Valley (1998: 209). By referencing Noel Fojut’s work (1982) on the location of Shetland brochs within the landscape, Sharples uses this to build an interpretive narrative of an isolated elite in domination of the surrounding community (1998: 208).
Figure 2. Northern broch wall with general view of external houses 1-4 (Photo Canmore 1990: AOC 1127 Film 61).
Figure 1. Bronze Age cremation urn (Photo Canmore 1990: AOC 1127 Film 47).
As one of the defining characteristics of Upper Scalloway Broch, it is appropriate to examine this interpretation more closely. It can be argued the mere presence of grain indicates nothing more than the storage of a food resource. Whether gathered as tribute, end of year harvest or trade is unknown. In contradiction to a dominating elite, alternative theories (Fojut 1982; Rennell 2015; Romankiewicz 2016) have found a convincing level of interconnectedness between brochs. It would likely have been necessary in a largely subsistence economy that may have included specialist sites in crop production and cattle management. This is something Sharples comes close to acknowledging when discussing resource management, production and the need to compare with other sites (1998: 118) but does not explore more fully, instead deferring to the preferred interpretation of competitive domination. Taken to a theoretical end point, it is therefore possible to imagine the broch attacked and burned by a rival chieftain or entities farther afield. From written sources it is known as early as the 580s the Dal Riata and the Picts were campaigning in Orkney (Foster 1996: 102) and may have extended
themselves farther north in a situation of competing control for the Northern Isles. These are events that could have also happened prior to the various historical chronicles and within the Upper Scalloway Broch destruction timeframe. Though this is not to deny the plausibility of competitive domination, it is simply to illustrate that the implications of the interpretive construct is not explored in the report to any great extent including a refinement of how it would work on a widening regional basis. The mere evidence of burnt barley alone cannot address this either.
The year following the Upper Scalloway Broch excavation, Niall Sharples and Mike Parker Pearson began excavating Dun Vulan in South Uist. Here the political hierarchy of exploitation model was again explored in two journal articles published before either excavation report (Parker Pearson, Sharples and Mulville 1996; Sharples and Parker Pearson 1997) using evidence from Upper Scalloway to supplement the theory. A vigorous discourse ensued over the 1996 article (Armit 1997; Gilmour and Cook 1997; Parker Pearson, Sharples and Mulville 1996; Parker Pearson, Sharples and Mulville 1999) encompassing several aspects of method and interpretation. One strand rested on the demarcation of broch location on the liminal edge of settlement as a display of power and fertility and culmination of elite status expressed in monumentality (Parker Pearson, Sharples and Mulville 1996). Ian Armit questioned this with the smaller Hebridean islands and the differentiated settlement patterns throughout the Atlantic zone (1997: 267). Much later Rebecca Rennell showed this basis of location can fall apart when looking at other areas of the Outer Hebrides where brochs are found in all four landscape zones she identified (2015: 27). It is something also not particularly true of Upper Scalloway which is firmly ensconced in the middle of a highly fertile landscape, though it is 170m from the shoreline. Moreover, Sharples and Parker Pearson believe the differential midden material dumped to the south of Upper Scalloway Broch and Dun Vulan as an important fertilizer represented the intentional construction of an “olfactory boundary” (Sharples and Parker Pearson 1997: 261) and symbol of fertility separating the broch inhabitants from the rest of the community (Parker Pearson, Sharples and Mulville 1996) which implies further demarcation of elite status via control of these middens. Perhaps true, but it is used to support the interpretation without consideration of more prosaic explanations followed by evidentiary refutation. In Upper Scalloway, for instance, the proximity to cultivated fields or protection from dominant northerly winter winds would likely account for the midden location.
​
However, the issues with the Upper Scalloway evidence used in conjunction with Dun Vulan to underpin a competitive domination interpretation is a bit more problematic, especially with the 1997 article. There is a selective reinterpretation and presentation of material frequently in contradiction with the published excavation report. For example, a rampart at Upper Scalloway appears as certainty (Sharples and Parker Pearson 1997: 257) where there were only unexcavated hints of one by virtue of a ditch in Upper Scalloway (Sharples 1998: 34). The “extensive excavation” to the south of the broch (Sharples and Parker Pearson 1997: 257) was a handful of small trenches beyond that of the immediate southern broch wall (Sharples 1998: 73). A larger issue is the presentation of material inside the broch as representative of elite status. The simplified graph of artefact categories by location in Upper Scalloway (Sharples and Parker Pearson 1997: 260) is misleading. Used to enhance the sense of prestige goods from an elite found within the broch versus more common goods outside, it glosses over the fact very few were found in layers prior to the destruction (Sharples 1998: 210). Of the roughly 600 artefacts from the primary broch occupation (Ibid: 31), the vast majority of which are potsherds believed to have come from the upper levels during the fire, all are associated with standard domestic function. Most prestige goods and metal artefacts found inside the broch are within the context of later buildings described as poorly built (Ibid: 210). Therefore, whilst the site may exhibit an elite status based on architectural monumentality, it does not appear the primary occupation artefacts would agree. Instead, it could more closely resemble a community store and workplace which would also explain the barley.
​
With the vast preponderance of high status finds coming from the period after a wheelhouse and later structures are inserted into the abandoned tower, the final period up to the 8th century the most wealthy (Sharples 1998: 210), it is a matter in need of more reflection. Early in the excavation Sharples disagrees with regional archaeologist Val Turner over the interpretation of a substantial wall to the east of the broch. She believes it may be evidence of a nearby wheelhouse, he instead interprets it as a boundary wall (Ibid: 34). This differs from the overall impression of the site retaining evidence of a northern boundary only and the wall is hardly again addressed. It also remains uncertain why the wheelhouse inside the broch was deemed poorly constructed and probably abandoned before being finished based on a single course of stones (Ibid: 43). It does not seem enough evidence to make that determination, yet the minimal remains may have been difficult to fully understand when truncated by a later structure. Nonetheless, there are indications Upper Scalloway Broch could have shared a much closer settlement affinity with Old Scatness and Jarlshof than we currently understand based on the excavation report. More recent geophysics and trial excavation conducted on the open field in front of the 1991 housing development show the village may have extended downslope to the east, as much as 75m from the broch.
​
Such omissions and interpretive statements of site evidence are an occasional but peculiar feature throughout the report. One of the more unique finds, a leaf-shaped spearhead of probable Anglo-Saxon type, is regarded as most likely “the possession of a mercenary or exile” (Sharples 1998: 159) rather than an import. It is an astonishing statement with no elaboration or additional data to confirm such a conclusion. The banal explanation is that it is an import or, perchance, a trophy. To be sure, the sheer numbers of trophy samurai swords that found a way into the United States after World War Two could, sans our abundance of multimedia source material, someday lead to the conclusion a sizeable number of Japanese warriors were wandering the American countryside in the mid-20th century. A similar issue arises with a statement on how the broch functioned on a seasonal cycle with the stalling of animals on the ground floor in winter months (Sharples 1998: 40). The included micromorphology report provides no direct evidence other than worms present in situ and layered accumulation of damp plant material prior to the fire, the latter interpreted as fodder (Ibid). Alternatively, as entertained earlier in the excavation report, this could have been accumulated barley residue and the spread of charred grains related to this rather than mass storage (Ibid: 31).
​
Puzzling too are a couple of structures south of the broch. The first is a figure of eight house just outside the southern walls with parallels to the structure at Buckquoy uncovered by Anna Ritchie (1977). Although the unexcavated baulk obscured the juncture between the northern and southern cells, only half of each still surviving undamaged, the planned sketch (Sharples 1998: 71) does not show a definitive relationship between the two. With the only evidence of a hearth lying just east and outside of the northern cell, there could be something different going on. The house may be trefoil in nature or, with evidence in the sketch plan of a possible southern passage to the larger cell, part of an additional amorphous multicellular structure. More significant, further south the remains of a turf structure with stone foundation, outwardly bowed walls and a paved floor (Ibid: 73-4) is mostly ignored in the interpretation of the site. This shows a high level of affinity with early Norse structures (Ritchie 1993: 33) and with strong evidence of Scandinavian activity occurring on site around the 9th century, especially in mixed midden deposits heavy in fish bones (Sharples 1998: 206), this may have been a missed opportunity to further enhance a chronological understanding of the site and fill the gap between sudden abandonment in the late Iron Age and later appearance of a Medieval cemetery.
.jpg)
Figure 3. Four by ten metre trench with rotary quern (Photo S Jennings 2020).
The significance of this additional evidence and, perhaps, overlooked chronology is relevant to what came out of a small 4 x 10m trench in the Spring of 2020 (Figure 3). Digging the foundations for a bicycle shed the local homeowner began to uncover human remains in an area roughly 30m east of the broch. Historic Environment Scotland initiated an emergency excavation where a total of 26 skeletons were recovered and determined to be an extension of the graveyard uncovered in 1989. After removal, substantial stone foundations were beginning to show and over the course of five weeks excavation the trench revealed myriad multicellular structures of Pictish typology. Whilst many unique and significant items were recovered from the trench - fine metalwork, imported flint, gaming pieces, bone dice, more thick red ash layers of midden mixed with heaps of animal bone and thousands of potsherds - a more common item has garnered much attention, a rotary quern (Figure 4, 5). Although eleven were found on site along with two fragmented saddle querns (Sharples 1998: 140), very little
discussion of them made it into the excavation report. Unfortunately, the lack of information on the original rotary querns gives little basis on which to compare but the more recent rotary quern has the appearance of being ritually closed with several fragments of white quartz pressed into two of the holes. It is unknown whether any of the others from the original excavation showed a similar phenomenon.
Figure 4. Recently excavated rotary quern (Photo C Leith 2020).
Figure 5. Possible ritual closing of rotary quern with quartz fragments (Photo C Leith 2020).
Not fully interrogating the variety of ritual more closely at Upper Scalloway Broch is surprising with Sharples’ interest in exploring related ideas elsewhere (Parker Pearson, Sharples and Mulville 1996; Sharples 2006; Sharples and Parker Pearson 1997). Though he ascribes structured deposition to a nearly complete pot with several artefacts found in a hollow (Sharples 1998: 36) and three steatite objects in house 1 (Ibid: 57-8), both outside the broch, a human skull found crushed underneath the first course of walling for the last phase of secondary broch occupation was not addressed as such (Ibid: 51-2). There is also little inquiry into the Bronze Age urn cremation burial (Figure 1), surely ritual, and the possibility of more associated material in toto from this period. Although acknowledging the uniqueness of the urn (Ibid: 17), the interpretation of the ceramic specialist who ascribed pottery sherds found in a hollow to the late Bronze Age/early Iron Age is later dismissed because of a lack of radiocarbon dates (Ibid: 185). The recent find of another urn cremation burial tentatively assigned to the Bronze Age provides additional evidence of more extensive use of the site before the broch. It would have been beneficial to investigate the possibility of Bronze Age structures more thoroughly since the report acknowledges the dates from the ditch indicate it could be a pre-broch enclosure (Ibid: 86-7).
​
The ignoring or refuting of Bronze Age material is a peculiar aspect of the excavation. In a very real way, it is the other bookend to the seeming reluctance to look meaningfully at the Viking activity on site after it was abandoned. Disappointingly, had the latter been examined more closely it could have given additional insight into the Pict/Viking horizon and may even have shown to be the catalyst for final site abandonment by native inhabitants. With the former it would have provided a much wider and interesting portrait of site activity, let alone adding to our knowledge of funerary rites. Yet it appears the focus was on the broch and Iron Age exclusively. One understandable reason for this may have been the limited time to undertake the dig. Another might simply have been the importance and prestige of an undiscovered broch with a chance at full excavation.
​
Resolving these conflicting issues is important because of how significant Upper Scalloway Broch has become to the debate and understanding of broch economy and social structure. Much of the interpretation and discourse included earlier ongoing commentary over Dun Vulan where the theory of competitive elite domination put a lot of emphasis on the charred grains in Upper Scalloway. In many cases, more prosaic interpretations and a more thorough interrogation of available evidence would have been welcome. Some of the hypotheses – the provenance of an Anglo-Saxon spear or the symbolic reasons for a large organically-rich midden south of the broch – are not readily testable. However, though there is room for disagreement on some of the interpretations, the multi-disciplinary approach, later widened and strengthened at Old Scatness, and methods employed in such a short space of time are exemplary. The attempt to treat the artefacts in a thematic way, integration of specialists and comprehensive cataloguing of all finds were quite an extraordinary effort to bring sense to an exceptionally complicated site (Smith 2002: 810). Failing to consider the wider site chronology is a most regrettable aspect. Allowing some speculation in the report to go unchallenged, perhaps more so.
​
​
References
Armit, I. (1997) ‘Architecture and household: A response to Sharples and Parker Pearson’. In Reconstructing Iron Age Societies. Ed. by Haselgrove, C. and Gwilt, A. (Eds.), Oxford: Oxbow Books.
Fojut, N. (1982) ‘Towards a Geography of Shetland Brochs’. Scottish Archaeological Journal 9 (1), 38-59.
Foster, S. (1996) Picts, Gaels and Scots. London: B T Batsford Ltd.
Gilmour, S. and Cook, M. (1998) ‘Excavations at Dun Vulan: a reinterpretation of the reappraised Iron Age’, Antiquity 72 (276), 327-337.
Pearson, M.P., Sharples, N. & Mulville, J. (1996) ‘Brochs and Iron Age society: reappraisal’. Antiquity 70 (267), 57-67.
Parker Pearson, M., Sharples, N. and Mulville, J. (1999) ‘Excavations at Dun Vulan: a correction’. Antiquity 73 (279), 149-152.
Rennell, R. (2015) ‘Re-Engaging with the Iron Age Landscapes of the Outer Hebrides’. Journal of the North Atlantic 901, 16-34.
Ritchie, A. (1977) ‘Excavation of Pictish and Viking-age farmsteads at Buckquoy, Orkney’. Proceedings of the Society of Antiquaries of Scotland 108, 174-227.
Ritchie, A. (1993) Viking Scotland. London: B T Batsford Ltd.
Romankiewicz, T. (2016) ‘Land, Stone, Trees, Identity, Ambition: The Building Blocks of Brochs’. Archaeological Journal 173 (1), 1-29.
Sharples, N. (1998) Scalloway: A Broch, Late Iron Age Settlement and Medieval Cemetery in Shetland. Oxford: Oxbow Books.
Sharples, N. (2006) ‘The first (permanent) houses: an interpretation of the monumental domestic architecture of Iron Age Orkney’. Journal of Iberian Archaeology 8, 281-305.
Sharples, N., Parker Pearson, M. (1997) ‘Why were brochs built? recent studies in the iron age of Atlantic Scotland’. In Reconstructing Iron Age Societies: New Approaches to the British Iron Age. Ed. by Gwilt, A. and Haselgrove, C., Oxford: Oxbow Books: 254-265.
List of Figures
Figure 1: Canmore (1990) ‘SC 1866269: AOC 1127 Film 47’. [online] Available from <http://canmore.org.uk/collection/1866269> [19 November 2020]
Figure 2: Canmore (1990) ‘SC 1866285: AOC 1127 Film 61’ [online] Available from <http://canmore.org.uk/collection/1866285> [19 November 2020]
Figure 3: Jennings, S. (2020) ‘Trench’
Figure 4: Leith, K. (2020) ‘Rotary Quern 1’
Figure 5: Leith, K. (2020) ‘Rotary Quern 2’
Past In Depth articles can be found in the Archive.