6.4.3 Building Traditions

Despite the increase in available data since Wainwright’s limited exploration of ‘Houses and Graves’ in the seminal Problem of the Picts (Wainwright 1955), early medieval settlement studies continue to be problematic. Settlement evidence has been notoriously elusive in mainland Scotland, with the notable exception of the Pitcarmick-type houses (Carver et al 2013; Strachan et al 2019). Elsewhere in Scotland, settlement evidence broadly indicates a widespread and complete movement away from the prehistoric roundhouse form and towards less monumental multi-cellular structures (Hedges 1990; MacKie 2007). In Caithness and on the Northern Isles, where stone architecture was typical, wheelhouses and WAG-type buildings show a trend towards cellular buildings (Ralston 1997). These are often associated with earlier settlements and sometimes used the abandoned structures for building materials such as at the Iron Age sites of Gurness, Wag of Forse and Buckquoy. More locally in Fife, where timber architecture predominates, single cell structures which employ similar construction techniques to those of the Bronze and Iron Ages are apparent. Importantly the round building forms were superseded by rectangular structures, such as at Easter Kinnear and Hawkhill (Driscoll 1997, 91).

Early medeival turf structure in Glen Shee ©️ Perth and Kinross Heritage Trust

Pitcarmick-type houses

These are large, elongated buildings with rounded ends, sometimes with bowed walls, often narrowing to one end, which is often sunken. With turf-built walls over stone footings, they range from 10–30m in length and are frequently constructed on an incline with the narrow end down slope (RCAHMS 1990). They are architecturally similar to the byre blackhouses once common in the Highlands and Western Isles until the 19th and early 20th centuries (Carver et al 2013; Strachan et al 2013). They take their name from the type-site, excavated by Barrett and Downes (1993; 1994), as discussed above. While Continental parallels have been recognised, they most likely evolved from the local roundhouse tradition as part of a northern European trend (Strachan et al 2019). Recent analysis and reconstruction of turf buildings in the Netherlands suggested that the rounded corners were used to prevent long walls from tearing apart at the corners (Nicolay and Postma 2018). The byre-houses were constructed with turf or turf-and-stone layered walls with timber roof-supports. They were constructed on inclines with the domestic ‘house’ including a hearth upslope, while animals were stalled on either side of the lower end, with a drain removing waste downslope out of the building.

Coloured drawing of a long structure with curved ends and a wattle roof. The drawing show the inside of the building on one side, which has wooden pillars and is sloped towards  the end.
Reconstructed drawing of a Pitcarmick house ©️ Alan Braby, courtesy of Perth and Kinross Heritage Trust
Excavation image of a stone hearth which is made of large, flat stones as a rectangular frame and smaller stones in the centre.
Excavated hearth from Pitcarmick structure (Carver et al 2013)

Buildings 1 and 2 (MPK4456) at Lair, also in Glen Shee, were initially described as ‘Pitcarmick related’ and may represent the development of the Pitcarmick style over time (RCAHMS 1990). Building 1, built almost entirely with turf, timber and thatch, was very long (25m internally) with rounded ends and straight sides. The downward slope towards the north-west end suggesting the need for drainage supports the interpretation of this structure as a byre-house. Building 2 was constructed using considerably more stone along with turf. It was much shorter than building 1 (13m long) and had the characteristic bowed sides of a Pitcarmick building. The possible function of this building ranges from an outhouse to the domestic dwelling associated with non-domestic building 1. Building 3 was large (20 x 5.5m) and had at least two phases. Evidence within this building supports an interpretation of the building being used for storage, working and as a byre-house over the course of its lifespan. Buildings 3, 4 and 5 are referred to as the Pitcarmick-type proper and seem to represent a network of structures which developed as additional storage and workspace was needed (MPK4384; Strachan et al 2019). The excavation at Lair offered a new perspective on Pitcarmick-type sites, suggesting specialist use of the annexes associated with the Pitcarmick buildings. This defining feature of the site type was previously unrecognised (Strachan et al 2019, 150).

Excavation at Lair, Glen Shee in 2013 ©️ Perth and Kinross Heritage Trust

The transition of form from roundhouses to elongated buildings remains unclear. Whether it was the consequence of external influences, or an emergence from the existing roundhouse tradition, remains an open question. The rounded ends of Pitcarmick-type buildings may relate to turf construction and/or the maintenance of turf walls against erosion by cattle and the weather. It is likely that some roundhouses remained in use for some time after the introduction of the Pitcarmick-types, but as yet no clear early medieval examples have been discovered.

Excavation photo of a trench with an abundance of stones and boulders appearing to take the shape of at least two elongated buildings.
Buildings 1 and 2 during excavation at Pitcarmick (Carver et al 2013)


Crannogs are artificial island dwellings ‘enclosed’ by the surrounding water. The majority have Iron Age origins; however, there are Scottish examples spanning the Neolithic to the medieval periods (Crone 2012). They are found in the lochs of Perth and Kinross and those in Loch Tay have formed part of a number of studies (Dixon 1981; 2004; Dixon and Shelley 2006). One sampling programme, which identified mainly Iron Age activity, did include evidence of early medieval reuse during the 6th to 9th centuries AD at Craggan  and Dall Bay North  (Dixon 2007). The Living on Water project targeted seven crannogs in Loch Tay to explore the chronological relationships of Iron Age occupation using radiocarbon wiggle-matching. Reuse in the 7th century AD was indicated at Dall Bay North, and 5th to 7th century AD reuse was identified at Eilean Breaban (MPK180), Fearnan Hotel (MPK495) and Milton Morenish (MPK170) (Hamilton pers comm). A field survey in 2004 at Loch Drumellie crannog returned an early medieval date from an oak timber (Dixon and Shelley 2006). While there has been no comprehensive excavation of an early medieval crannog in Perth and Kinross, excavated sites of the period elsewhere in Scotland have evidence of high-status occupation. The use of crannogs as high-status centres may have been widespread (RCAHMS 1994). For example, E-ware pottery and crucibles for fine metalworking were recovered at Loch Glashan, Argyll, which may link it to other high-status sites such as the forts of Dunadd and Dundurn (Crone and Campbell 2005). To the east, Castle and Prison Islands, Loch Kinord, Aberdeenshire, have demonstrated phases of use sometime in the 6th–12th centuries AD (Stratigos and Noble 2014: 2018). The potential of further late first millennium AD crannogs across all of northern Scotland including in Perth and Kinross has also recently been highlighted (Henderson et al. 2021; Stratigos and Noble 2021).

Image of a large circular building on stilts in the middle of a loch. The structure has a pointed roof and a long bridge on stilts from land.
Reconstructed crannog at The Scottish Crannog Centre on Loch Tay ©️ Gunther Tschuch
Croft-na-Caber crannog under reconstruction ©️ HES

Early medieval finds from crannogs in the area are scarce as a result of the lack of deliberate work investigating crannogs dating to the period. Yet, the potential for illuminating the early medieval period with material culture from crannogs has to be very high given their usual waterlogged or submerged conditions. Preserved wooden artefacts and cloth at the Iron Age Oakbank crannog (MPK484) demonstrate the preservation possible in waterlogged deposits. It also shows the potential for further understanding of early medieval activities not usually detected on terrestrial sites. Although, this presumption should be tempered as not all excavated crannogs produce excellently preserved and abundant artefact assemblages (eg Cults Loch crannogs, Dumfries and Galloway; Cavers and Crone 2018).

While the recorded find circumstances are vague, it is worth mentioning the late 7th- or early 8th-century hoard from Tummel Bridge (MPK540), at the western end of Loch Tummel. The hoard comprises three silver penannular brooches, and fragments of a bronze hanging bowl and a cup, probably of Roman origin. The fragmented bronze items contrast with the complete, but possibly unfinished, silver brooches, and it has been interpreted as a Pictish metalworker’s hoard (Anon 1888; Cessford 1999). The material would not be unexpected at a high-status fort but non-ferrous metalworking also took place at crannogs, recalling Hebridean examples, such as Eileann Olabhat, North Uist (Armit et al 2008, 82–92). In a similar vein, the 19th-century stray discovery of two Pictish, silver penannular brooches of 8th–9th-century date from close to Clunie Castle, a crannog site in Loch Clunie (MPK5520; Youngs 1989, 114–5) may support elite residence and manufacture at that site. A metal-detector find of a fragment of a third brooch in 1990, allocated to the National Museums Scotland to accompany the two complete brooches, may substantiate that supposition. Viking raiding in the area may have prompted their deposition or loss, and the crannog, which was later occupied by the bishop of Dunkeld’s castle, may have been at this time a Pictish lordly residence (RCAHMS 1994, 91).

The early medieval reuse of crannogs on Loch Tay also offers the opportunity to develop tree-ring chronologies for the period as oak timbers, so far only radiocarbon dated, survive. Furthermore, one oak from the multi-period submerged woodland at Craggantoul in Loch Tay returned a radiocarbon date of AD 410–570 (MPK 17641; Dixon 2007). This potentially suggests a cohort of oak remains which could augment tree-ring data from the early medieval phases on some of the Loch Tay crannogs. However, a tentative indication from the Living on Water project suggests the use of oak in early medieval contexts on crannogs decreased in comparison to the Iron Age. This may limit the targeted acquisition of sample material with the ability to develop 1st millennium AD dendrochronologies in Perth and Kinross.