The Knowe of Swandro: excavating eroding archaeology on Rousay

In the concluding part of our Orkney trilogy (see also CA 394 and 395), Carly Hilts reports on her visit to the Knowe of Swandro on Rousay, where Julie Bond and Caz Mamwell took her through the latest findings from this long-lived but rapidly eroding site.


Across the waters of Eynhallow Sound, the island of Rousay was initially only an indistinct shape shrouded in mist. As the ferry drew closer, though, I could see ominously rainy clouds sweeping across its higher points, clearing just as quickly to leave the green slopes streaked with sunlight. In Orkney, I had been told, it is common to experience every kind of weather in a single day, and that had certainly been the case during my short crossing from the larger island known as Mainland – though the journey had been as beautiful as it was bracing, with gannets plunging around us like white javelins. Below deck, my car was nestled between the imposing bulk of a tractor and its bale trailer on one side, and bundles of the day’s post and milk deliveries on the other – Rousay’s c.200-strong population rely on many goods being brought in by boat, but one thing that they have no shortage of is archaeological remains. Numbering over 160, historic (and, particularly, prehistoric) sites almost outnumber humans on the four-mile-long island, and their diversity and significance have earned Rousay the nickname ‘the Egypt of the North’.

Overlooking excavations at the Knowe of Swandro on Rousay, Orkney. This photograph shows how close to the waterline the site lies; indeed, at high tide during the winter, the archaeological remains (which the team cover with protective sandbags before departing at the end of each dig season) are completely submerged. Image: University of Bradford

From the single road that encircles the island you can reach a number of enticingly named Neolithic tombs – Taversöe Tuick, Blackhammer, Knowe of Yarso – and on Rousay’s southwest side you find a small car park signposted for the largest of these: Midhowe cairn. Located steeply downslope from the car park, this 32.5m-long communal burial space is divided by upright stones into numerous stalls, in which the remains of at least 25 individuals were found during excavations in the 1930s. Today the 5,000-year-old cairn is shielded from the elements inside a structure resembling an aircraft hanger, which is furnished with suspended walkways to allow visitors an aerial view of the tomb’s interior. It also marks the start of a remarkable mile-long route along the coastal fringe, known as the Westness Walk, which guides you through tangible reminders of Rousay’s past. Immediately to the west of the cairn stand the remains of its similarly named but rather younger neighbour, Midhowe Broch, an Iron Age tower standing guard over the grey waters. Turning and walking east along the rocky, sometimes boggy, shore, I then passed the roofless ruins of long-abandoned farmsteads (the 19th-century land clearances that devastated so many communities in Highland Scotland also took a heavy toll in Orkney; Rousay’s population was once almost five times larger than it is today), the remains of St Mary’s Kirk, and a Norse hall – until I reached my intended destination: the Knowe of Swandro.

Two monuments mark the start of the Westness Walk, a mile-long coastal route lined with archaeological sites. Above is the Iron Age Midhowe Broch, while the photo below shows the Neolithic remains of Midhowe cairn, which visitors can explore on walkways suspended above the ancient stonework.

Introducing the archaeology

Preserved within an uneven beach of sea-tumbled stones, the Knowe of Swandro takes its name from its location beside (and, at particularly high tides, under) the Bay of Swandro. Its remains have been under investigation by a team of archaeologists and students from Bradford University, together with students and volunteers from as far afield as the USA (particularly from City University New York and William Paterson University), for over a decade. The project is funded by the Orkney Islands Council, the Swandro Trust, Historic Environment Scotland, and support from Bradford University, and it was born in 2009, when Dr Julie Bond was walking across this stretch of beach towards another significant archaeological site, Westness cemetery (see box below). While making her way towards the peninsula where the burial ground lay, Julie spotted a number of orthostats sticking up through the stones. These jutting uprights were clearly not natural, and Julie deduced that the sea had taken the top layer off of some kind of structure, with its surviving stonework preserved beneath the beach.

Due to the importance of the Westness finds, the area was already a Scheduled Ancient Monument, but the scheduling did not continue beyond the high-water mark. As a result, when 2011 brought a particularly low spring tide, test pits were dug below the waterline. These exposed midden deposits and what was interpreted as evidence of occupation in the Neolithic period – but it was only the beginning of what the Knowe of Swandro had to reveal.

Looking along the course of the Westness Walk – past ruined farmsteads and (below) the remains of St Mary’s Kirk – towards the Knowe of Swandro.

Today, Julie is co-director of excavations on the site, together with Dr Steve Dockrill, also of Bradford University. Last summer was the team’s first season back at the Knowe of Swandro since the onset of the Covid-19 pandemic (although the hiatus had given them the opportunity to re-examine some earlier finds from the site, including the famous ‘Pictish anvil’ – of which, more below), and it was Julie and supervisor Dr Caz Mamwell who greeted me as I arrived on site, taking me through the story so far, and explaining how the 2022 excavations were adding to this picture. To refresh the background (which you can read in more detail in CA 275): the initial excavations at Swandro had revealed mighty concentric circular walls, which at first were suggested to be the remains of an Iron Age broch. This would not have been a surprising find, as at least nine such structures still stand along the shores of Eynhallow Sound – but as investigations progressed, it soon became clear that the site was rather more complex. The stonework, while impressive, did not reflect the distinctive double-wall construction typical of a broch. Nor was there any evidence that it had once supported a second storey. Instead, it appeared to be much earlier in date – the architecture was much more akin to that of a Neolithic chambered tomb, of which 16 are known on Rousay. This interpretation still remains speculative for now, but further investigations revealed evidence of occupation spanning an impressive sweep of time, with the site’s use stretching into the Iron Age, Pictish, and Norse periods of Orkney’s history, all building on top of each other to produce an imposing mound that now needed to be teased apart.

There were rather less methodical forces at work, too, to dismantle the archaeology, however – since their project began the excavation team have been engaged in a race against time and tide to document as much as they can before the emerging remains are lost to the sea. In the winter, the waters of Swandro Bay rise beyond the excavated area, meaning that at the end of each digging season the team have to cover the remains first with sandbags (filled with sieved spoil), and then rocks – not for them the tarpaulins pinned by tyres that I had seen at the Ness of Brodgar the previous day (see CA 395). Despite these precautions, though, the sea is still taking its toll on the underlying archaeology, stealing from beneath as well as from above; the beach is made up of loose stones with gaps between them, and when water rushes through these under pressure and sucks back out, it takes archaeology with it and jams the resulting gaps with beach stones and sand. Larger, heavier stones are left behind, but are destabilised, making their excavation a painstaking process. Adding to these challenges, the site also ‘enjoys’ a particularly high water table, and the team also faces water running off the nearby hills during particularly unfriendly weather. In spite of these challenges, though, the excavations have proven very productive – and at the time of my visit, further illuminating clues had already emerged.

The Knowe of Swandro’s archaeology is complex: archaeologists have to make sense of ancient stonework that has been heavily disturbed by the sea, and which lies among tumbled beach pebbles.

The Westness cemetery

Lying just beyond the site of the Knowe of Swandro excavations (and marking the culmination of the Westness Walk), Westness’ archaeological significance was revealed in rather unprepossessing circumstances, when a farmer digging a hole to bury a dead cow in 1963 uncovered what would prove to be the wealthiest Viking Age female burial yet found in Scotland. There, a young (under 30) woman had been laid to rest along with the remains of a newborn baby; it is thought she had died in childbirth. She was accompanied by a range of grave goods including an ornate silver brooch pin decorated with amber and gold (known as the Westness brooch, it is held by National Museums Scotland; see She was not alone on the peninsula, however.

Investigations by the University of Bergen in Norway established that the woman’s grave was part of a much larger cemetery of around 50 Norse and Pictish burials, including two boat graves containing the remains of adult men – one of whom had an arrowhead still embedded in his skeleton (see Subsequent isotope analysis by Professor Janet Montgomery at Durham University has revealed that both men came from the very far north of Europe, with one possibly having his origins north of the Arctic Circle, while the woman with the brooch came from either the Western Isles or Ireland.

Structures on the site were very well built, leading to initial interpretations that this could have been the remains of an Iron Age broch – nine more examples are already known to have stood guard over the surrounding waters. More recently, it has been proposed that the site’s origins could have involved the construction of a Neolithic tomb, although for now this remains speculative. We do know, though, that a substantial Iron Age roundhouse was built on the site.

The presence of both local and Scandinavian individuals within the same cemetery might indicate that the Norse newcomers had integrated with the community – indeed, their burials were not the only indications of such activity in the area; in the 1950s and 1960s, Raleigh Radford excavated the remains of two longhouses a stone’s throw away from the peninsula, and right on top of Swandro, probably representing the last phase of its settlement.

Uncovering the Iron Age

As work progressed, the site’s chronology was coming into sharper focus. At the very base of the mound the team have found a series of concentric stone walls with the gaps between them filled with rubble. As mentioned above, questions of a funerary aspect to the site’s earliest use remain unresolved, but its later story is relatively easier to piece together. Delving into more recent prehistory, tiny traces of Bronze Age activity have emerged in recent seasons, including little bits of steatite soapstone imported from Shetland, and coarse stone tools that are known from burials of this period elsewhere in Orkney. It is also fairly common, Caz said, to find that Orcadian Neolithic tombs had been subsequently remodelled by Bronze Age communities – something that can be seen at Tresness stalled cairn on Sanday (excavated in 2021 by Professor Vicki Cummings and Dr Hugo Whymark-Anderson), for example, and which might prove to be relevant if the Knowe of Swandro remains also turn out to have been originally intended for the dead rather than for the living.

Overlooking the interior of the Iron Age roundhouse, which was subdivided a number of times during its lifespan.

Most of Swandro’s surviving stonework, however, dates to the Iron Age, when a large roundhouse was built on top of the earlier archaeology. This was an impressive construction with an entrance passage and an imposing outer wall; dates of c.800BC have been recovered from its lower levels, and c.550 BC from further up. Inside the roundhouse the team has identified plentiful evidence of the everyday, including a series of hearths, quernstones, stone spindlewhorls, and a little pot with its Orkney flagstone lid still in place. This was no ordinary residence, however; ground-penetrating radar survey by Professor Chris Gaffney (also at Bradford University) has revealed the presence of a substantial ditch encompassing the occupied area, hinting at its importance. This might have been due to the activities taking place within its walls, which appear to include high-status metalworking.

Previous seasons had already uncovered the presence of hearths and a furnace within the Iron Age occupation area, and the furnace was particularly well-preserved, with its lining and traces of fired-on vitrified slag surviving around three of its sides. While I was on site, there came a shout of excitement across the main trench – one of the team had found what appeared to be the furnace’s top. This prompted great interest as it appears that there might be enough intact to get some archaeomagnetic dates from it. To facilitate this, the dig’s blog attests, the furnace has been carefully protected until next year when it can be fully excavated. This particular feature was too fragile to simply cover over, however, so the team have built a little sandbag bunker around it that was next filled with bubblewrap and topped with a convenient piece of driftwood. Only then was this construction covered by a protective membrane and stones like the rest of the site – it will be fascinating to see what secrets it reveals when it is brought to light once more and dating analysis can take place.

ABOVE & BELOW During CA’s visit to the site, the location of a known furnace was under investigation when a chunk of material thought to represent part of the top of this feature was found.

Next to the roundhouse remains, the team have also excavated a small, semi-subterranean structure that has been interpreted as a smithy. It had been built into the stone-lined ditch surrounding the site, and illuminating details of its layout have survived, vividly evoking how the building may have operated – the team have even identified bolt holes on either side of its entrance, together with the door jamb and a pivot stone, while a large upright stone had been set up in front of the doorway, standing between the entrance and the hearth in order to help keep the draught off the fire. Furnishings including little stone cupboards and two stone anvils have also been identified, and thanks to analysis of floor deposits and crucible fragments within the space we are also able to learn more about what the smith was doing there.

A specialist smith

This was clearly the domain of a very skilled individual, as the smith had been using a technique called fire-welding, which involves bending, folding, and reworking iron at a very high temperature. This produces a very distinctive spherical form of slag like ball bearings, which the team has observed within the smithy, as well as a diagnostic form of hammerscale. This was not everyday metalworking, but the art of a real specialist.

A Roman coin known as a minim.

As for the materials being used, magnetic signatures attest that the smith was working with brass and bronze. The presence of the latter metal is not surprising – there is some copper on Rousay which could have been easily accessed, though the tin required for this alloy presumably came from Cornwall, hinting at very long commercial connections. Brass, though, is more unusual in Orkney, as to make this you need zinc, which was less readily available. It could have been traded through the Roman Empire, though, and the team wonder if a stash of Roman metal might have been being recycled at the Knowe of Swandro.

Evidence of contact with the Roman world has already been found on the site, in the form of bits of glass and coins – including a small but surprisingly heavy minim which I was shown during my visit. Two small glass toggle beads represent particularly significant finds; they are rare in Scotland, with perhaps 15 known in all, and it is very unusual to find two on the same site. Many previous examples of these finds have proven to be reused Roman glass, and with blobs of molten glass within the Iron Age remains at Swandro suggesting that glass-working was taking place on site, the two examples found there might indicate that this was an important Iron Age craft-working centre.

The ‘Pictish’ – now Iron Age – anvil. Photo: Steve Dockrill

As mentioned above, the enforced pause in digging caused by the pandemic allowed closer examination of the larger of the two anvils from the site. When we reported on its discovery back in CA 343, there had been great excitement after the stone was lifted and cleaned, as the team swiftly realised that they could see carbon imprints left by the smith’s knees and hands – a powerful evocation of this long-vanished individual at work, kneeling at the bellows. At the time, it was described as a Pictish implement, but radiocarbon dating has since revealed it to be rather earlier in date, placing it in the 1st century BC-1st century AD, in keeping with the Iron Age activity seen across the rest of the site.

ABOVE & BELOW Some of the finds from the latest work on the site: a bone point and a bone comb.

Aside from the now-renamed anvil, though, there were more concrete clues to a Pictish presence at the Knowe of Swandro. This came in the form of an archaeomagnetic date from a hearth reflecting activity c.AD 500, and a semicircular wall dated to the 8th century, from which the team recovered distinctive hipped bone pins and pieces of composite comb. This period of the site’s past also contained a dramatic event, as the roundhouse’s walls appear to have suddenly collapsed inwards while Picts and their animals were still in residence – the articulated skeleton of a piglet was found trapped between two particularly large stones that had fallen to one side. The team have been working to take out the rubble from this collapse, but they suggest that there is not enough tumbled stone to reflect the scale of what probably happened. It is also notable that while the wall core survives, its facing stones are missing. This points to the material being robbed for recycling at a later date – probably during the Viking Age, as new groups began to make their mark on this regularly reused stretch of beach.

This is a site with a long story, a story of artisans and craftspeople, of intrepid journeys and skilled work – and it is a story that is still emerging from the stony beach, even as some of its details are swept out by the sea.

ALL IMAGES: Simon C B Jones, unless otherwise stated
Further information
To read more about the Knowe of Swandro, excavations on the site, and how you can support the project, see