Is the The Wirk a Castle? Archaeological investigations in Rousay, Orkney

With HES giving the provisional sign off for the excavation at The Wirk to take place in the week commencing 21 September, and the geophys survey to take place the previous week, project leads Drs Dan Lee and Sarah Jane Gibbon outline the background to the project and what they hope to find.

Located on the south-western coast of Rousay, The Wirk is located in one of the most archaeologically rich parts of Orkney. The Orkney Research Centre for Archaeology (ORCA) is undertaking geophysical survey, archaeological excavations and 3D modelling at this enigmatic castle site (pending approval from HES).

Subscribe to our quarterly newsletter

The Wirk, meaning stronghold in Old Norse, and with the alternative name Westness Castle, is a small stone tower located close to the coast. It has variously been interpreted as a 12th century Norse Castle, a detached fortified bell-tower, a 13th century defensive tower for an incomplete church, a hall-house garderobe tower and most recently a 16th century tower and attached range.

The Wirk located on the side of  St Mary’s Church graveyard, looking south. credit: Bobby Friel @TakeTheHighView

Minor clearance and excavation in the 1920s identified similarities in construction between The Wirk and the 12th century Cubbie Roo’s Castle, on the nearby island of Wyre, considered to be one of the earliest stone keeps in Scotland. The Wirk is located in Westness which has been a large estate since at least the 12th century when it was the home of the Norse chieftain Sigurd of Westness (Orkneyinga saga). It is adjacent to Rousay parish church, likely to date from the 12th century, with standing remains of 16th century date on earlier footings.

Recently, the 12th/13th century date attribution of The Wirk has been rejected in favour of a 16th century date. This new interpretation is based on the built remains and 16th century architectural fragments which were found in the 1920s. However, architectural fragments of 12th/13th century date were also present and nearby archaeology, particularly the discovery earlier this year of a Norse hall at Skaill by the UHI Archaeology Institute would suggest this was a high-status place in the saga period. This is not to dispute that The Wirk may also have been in use in the 16th century when the estate was owned by a prominent Orkney family. One of the objectives of this project is to excavate trial trenches over Clouston’s excavation and at the eastern end of the site to identify the earliest phases of the tower and adjacent building. Upper parts of the tower were substantially rebuilt in the 19th century and so excavation will allow us to record parts of the site that have not been knowingly rebuilt.

Inside The Wirk tower, looking southeast, showing the entrance to an underground well or passage (Image: Dan Lee)

Along with the excavations, targeted geophysical survey around The Wirk aims to better understand any relationships between the tower and the buildings/features which surround it. 3D modelling of The Wirk and Cubbie Roo’s Castle will enhance our understanding of these comparable sites and allow the public to explore the remains online.

Subscribe to our quarterly newsletter

References and further reading:

Clouston, JS 1931 Early Norse Castles. The Orcadian. Kirkwall

Gibbon SJ 2017 A Survey of Norse Castles in Orkney, in P Martin (ed) Castles and Galleys: A reassessment of the historic galley-castles of the Norse-Gaelic seaways. Islands Book Trust. Laxay. 226-248.

HES 2017. Skaill farm survey, The Wirk http://canmore.org.uk/event/1040199

HES 2020. The Wirk http://canmore.org.uk/site/2282

HES 2020. Skaill farm, Rousay http://canmore.org.uk/site/351514

Tabraham, C 1997 Scotland’s Castles. BT Batsford. London.

UHI 2019: https://archaeologyorkney.com/2019/08/06/norse-hall-discovered-at-skaill-rousay-orkney/

Featured image: View of The Wirk, located on the righthand corner of St Mary’s Church yard in the foreground, looking northwest to Midhowe and the Atlantic Ocean (Image: Bobby Friel @TakeTheHighView).

Conclusions from an almost-thesis

Research on castles has outgrown the walls of architecture, sprung from the rubbly tangle of archaeology and taken flight from the pages of law texts, charters and literary exposition. In the hands of an adept writer, these are small bumps on the road to timely completion. Presently, however, I find myself in the camp of the almost-complete thesis writers (PhD fourth-yearers: here is fellowship!). It is a credit to castle studies that there is so much to think and write about and, perhaps, get lost in. This post shares some of the most interesting conclusions from my work looking at early stone castles in two polities in medieval Scotland, the Earldom of Orkney and the Lordship of Galloway. Among the themes I examined was the transition from timber to stone architecture, the relationship of castle to landscape and the political context for castle construction in these areas.

Why Orkney and Galloway? To begin with, we cannot be certain why castles appeared in non-‘feudal’ Orkney in the first place. Secondly, one of Orkney’s early castle sites, according to accepted wisdom, is the prototype for a larger group of early castle sites in the earldom – and has influenced larger debates on castles in Scotland more generally. Cubbie Roo’s Castle, this foremost plank of Orcadian castellar wisdom is, I believe, a little younger than widely believed. Nevertheless, the evidence suggests Orkney was probably home to a stone castle founded around the middle of the 12th century, decidedly early for this part of the world.

Figure 1. View of site of Cubbie Roo’s Castle, Wyre, Orkney, in the distance atop the highest hill on the small island. The farm on the right may represent the 12th-century home farm. The ruined building on the left in the distance is the chapel site. ©William Wyeth.

If the castle site (but not the surviving architecture) is genuine, then the attendant landscape must be examined too, for Cubbie Roo’s and other Orcadian castles. Following a conventional model of castle as the centre of lordship, one would expect to find churches and important farm sites close to suspected or documented castle sites. This was simply not the case in Orkney. Cubbie Roo’s and the other castle sites I examined were, on present (chiefly place-name) evidence, sited on marginal lands of indifferent quality. Some preserved evidence of associated chapel sites, which fits with the conventional model of early castellar lordship.

The disconnect between terrestrial wealth and castle location opened three possibilities about the origin and siting of castles: firstly, were these defensive sites? If defence included a consistent grasp of the surrounding area, including major seaways through the Earldom, then the answer is no. Secondly, were these sites connected to maritime wealth? It is entirely plausible: recent research has demonstrated the wealth of the earldom derived massive fisheries exploitation. Thirdly, were these sites connected to ‘new’ arrivals to the Earldom? The political upheavals in the Kingdom of Norway (to which Orkney belonged) saw magnates appear in the Earldom with no obvious familial connection. The lack of obvious relationship with good farming land may be taken to suggest the castles were built on land acquired more for the purposes of castle building (and architectural showmanship) and maritime exploitation than the inherent wealth of the soil. The new men’s power derived from proximity to the Earl, not a direct ancestral claim to a portion of Orkney’s economic output. Orkney’s earliest castles, whatever their form, were the product and reflection of a shift in how comital wealth and power operated.

Figure 2. View from the top of Cubbie Roo’s Castle, overlooking the possible home farm. The island in the distance on the right (Egilsay) is home to the Church of St Magnus (visible by its distinctive round tower mid-way along the island’s profile); the cult of St Magnus was instrumental in the formation of distinctly Orcadian political identity in the medieval period. ©William Wyeth.

Galloway’s castles present a different challenge; though as a group they are more numerous than their Orcadian counterparts, the evidence for them is drastically more erratic. The exquisite (and displaced) Loch Doon Castle is more clearly understood than the hummocky mound of Castledykes outside Kirkcudbright, for example. On a cursory examination of both sites’ landscapes shows that Castledykes is much easier to understand. Kirkcudbright was the centre of the powerful Lordship of Galloway.  Its hinterlands feature no fewer than three monastic foundations by the Lords, and the probable extent of demesne estate concentration in the area around Kirkcudbright is one of the highest in the Lordship.

What of Loch Doon Castle? Any modern visitor will appreciate the eponymous loch is remote and difficult to access from the big towns of Scotland’s south-west. However, it was originally built on an island in the Loch and later dismantled and reconstructed on the shore. It initially sat on a substantial route-way between the Glenken in Galloway – another demesne area of the Lords – and the Scottish coastal royal burgh and castle of Ayr. Another route south of Loch Doon offered access to the southern area of Carrick and, via the River Cree, the Wigtownshire portion of the Lordship of Galloway, with its probable administrative centre at Cruggleton Castle. Though its landscape is presently dominated by fishing holiday cottages and forestry, place-name evidence suggests in the medieval period the area was exploited for its pastoral suitability, and analogous documentary evidence from monastic sources hint at mineral wealth too. Doubtless Loch Doon Castle also formed an upland centre in Carrick in counterpart to the lowland, maritime-oriented centre at Turnberry Castle (also surveyed during this project).

Figure 3. Loch Doon Castle, view from wall-top towards loch. The castle was moved stone-by-stone before the construction of damming schemes in the mid-20th century. It was originally located on the partially submerged island in the loch visible here. ©William Wyeth.

Lastly, an overview for all of Scotland. A small, prefacing section of my thesis examined the monuments record for all possible castle sites in Scotland, including the sites above and less conventional secular power centres in the 12th-14th centuries – palaces, enclosures, crannogs (artificial islands in lochs), duns and brochs. A growing body of evidence in Scotland suggests many sites typically understood as prehistoric (chiefly the last three aforementioned categories of sites) were re-occupied during the broad medieval period. One conclusion from this country-level study suggested that there were more medieval power centres (e.g. bearing evidence for occupation) per square kilometre in the western counties of Scotland than the east. This likely reflects the underlying patterns of early medieval lordship in Scotland over which the culture of castles was overlain. This may act as another, belated, nail in the coffin of the military-architecture thesis of castle studies in Scotland.

The framework of understanding castles through landscape as much as architecture and archaeology is one I hope to apply more widely to Scotland and, as I engage more fully with English castle studies in my new job, in the wider medieval world.