Complex Atlantic Roundhouses / Brochs / Galleried Duns

The terminological debates on broch and Complex Atlantic Roundhouses (cARs) have been simmering for decades (Armit 1992, 2003; MacKie 2002; 2007; Romankiewicz 2009; 2011; 2015; 2016; Cavers et al 2015; Barber 2017). Romankiewicz (2011, 29) defined them as ‘circular, sub-circular or oval stone-walled roundhouses that have evidence for architectural complexity in the form of intramural space and well-coursed masonry’. The key problem with poorly preserved structures is determining whether they had intramural galleries. In some cases these start from upper stories, so without good preservation the defining evidence will be lacking. The distinction between substantial sARs and cARs can therefore seem slight. Some of the sARs were substantial structures. In the past some have been defined as brochs, but without the diagnostic intramural features, they are not considered cARs.

The terminological confusion increases when galleried duns are considered: the presence of intramural features places them in the cAR group, but because generally, but not exclusively, they are of a smaller diameter in size, they are not considered brochs (Graham 1947; Cavers 2010). These galleried duns are also distinct from enclosure duns.

Map 7.3 Distribution of Brochs and Complex Atlantic Roundhouses in the Highlands (updated interactive map coming soon!)

The distribution of cARs and brochs known in the Highland region can be seen on Map 7.3; and details are available in Datasheet 7.3.

SiteAreaDatingsLab ReferenceCommentsSource
Crosskirk brochCMost last three centuries BC, but possibly earlier; with reuse 2nd century ADOld dates. Wealth of artefacts, including Roman finds. Poorly built walls. Probably no towerMHG13496; Fairhurst 1984
Elsay broch (Staxigoe)CAD 143–384SUERC-18284Excavations by Tress Barry. Dating is from a long-handled antler combMHG2079; Sheridan et al 2017  
Everley brochC1st–2nd century ADExcavations first by Tress Barry and then limited excavation by NMS in the 2000s. Dating is under secondary paving. Structure may be sARMHG1792
Thrumster Mains brochCNumber of dates, showing several phases 4th to 3rd century BC– early years ADCleared out before 1910. Recent excavations, including of the walls, showing complexity of multi-period useMHG2043; Barber and Humphreys 2012
LangwellS536–46 BC
452 BC–AD 1;
481 BC–AD 21;
297 BC–AD 249
GX 3274
The last two dates are from the same timber, but from two different labs (type of wood not specified). Galleried dun, situated in hillfort. Large roundhouse 15 by 15.5m internal, with thick walls; timber structure near opening, and timber postring. Intramural guard cell. Initial construction interpreted as mid 3rd century BC; two phases of occupation before fire, followed by 3rd occupation phase. No pottery, few finds. Remains of burnt roofing timbers.MHG7371; Nisbet 1994-1995
Scotsburn HouseER731–399 BC 537–387 BC 355–58 BCPoz-68733
Stone built roundhouse, with probable intra-mural void, with outer enclosures. Activity also between roundhouse and enclosure wall. Multi-phaseMHG8627; Hatherley 2015a
Comar Wood DunINine dates, one group 357–184 BC, the other first centuries ADClear phases of occupation and re-occupation within the central building and the enclosure outwork, and two burning events. Small stone tools assemblage, small amount metalworking debris. Interior diameter 18m by 15m, with timber structure near opening.MHG55867; Peteranna and Birch 2017a
Applecross brochWR4th to 3rd centuries BC to early centuries ADMulti-phased occupation, with at least two roundhouses before complex roundhouse which may not have had tower.MHG7680; Peteranna 2012a
Dun an Ruigh Ruadh (Rhiroy)WRLast centuries BCOld dates. Excavation in 1968 by Euan MacKie. Domestic artefacts. Postholes for supports in the centre, suggesting a raised wooden structure.MHG7808; MacKie 1980; Hudson and Humble 2015
Clachtoll brochNWSMainly 50 BC – early years ADGood preservation at modern excavation by AOC Archaeology and Historic Assynt.  Artefacts mapped to show spatial distribution.MHG13002; G. Cavers pers comm Case study; Clachtoll Broch
Dun FlodigarrySkye178 BC–AD 132GU-1662Much disturbed. Finds included pottery, Roman Samian ware, slag, animal bone, querns, stone tools.MHG5219; Armit 1996, 241
Dun ArdtreckSkye356 BC–AD 238GX-1120Old date. Excavated in 1960s. Multi-period. Large number of finds, including Roman.MHG5019; MacKie 2000
Table 7.7  Complex Atlantic Roundhouses in the Highlands with dating evidence
All dates cal at 95.4% probability. For full details of dates, see Datasheet 2.1
Note: antiquarian excavations at a number of sites in Caithness, Wester Ross and Skye have diagnostic artefacts for Iron Age, but rarely allow for detailed phasing.

John Barber has investigated the structural aspects of brochs with forensic detail, including with experimental archaeology (Cavers et al 2015; Barber 2017; 2018 Rhind lectures), while Tanja Romankiewicz (2011; 2016) has approached the subject from an architectural perspective, looking at design aspects, spatial layouts and structural issues. Both are useful approaches, providing greater insights from simply typological considerations. Romankiewicz (2016; Romankiewicz and Ralston 2020) has highlighted the variety of architectural detail which react to regional environmental circumstances and availability of materials. For example, brochs built in areas that had been substantially deforested, such as Caithness, have smaller internal dimensions, probably due to issues of sourcing timber for roofing and other internal work (Romankiewicz 2016, 18).

Issues regarding broch development and dating as well as the social factors relating to construction have also been the subject of much debate, which is still ongoing (see Armit 2003; MacKie 2007; Cavers et al 2015; Romankiewicz 2011; 2016; Barber 2017). These factors are important for the Highlands, though one should not presuppose one explanation fits all examples. It is therefore important to think about alternatives (Romankiewicz and Ralston 2020 72–73).

The dating of cARs is now seen to begin in the 4th or 3rd century BC with many showing occupation into the early centuries AD, though whether the occupation is continuous or episodic is often debated (Dockrill et al 2006). Dating remains key, as it is probable that every surviving broch was altered and modified either in the Iron Age or later (Barber 2017). Only at Thrumster broch in the Highlands has it been possible to show the chronological changes in detail (Barber and Humphreys 2012).

Diagram showing the modifications and dating sequence at Thrumster Broch. ©John Barber/AOC Archaeology

Heald argues strongly that it is therefore necessary to dig the walls and the interiors to illustrate phasing and re-use, especially since interior remains may relate to re-deposited fill after a structure has gone out of use. For this reason the Clachtoll broch finds are important, because they were sealed by a wall collapse (Case Study Clachtoll Broch). Romankiewicz (2016) also notes that only with large scale excavations will some of the questions relating to the social dimensions of broch building and use be tested. The cost of such investigations is, however, a major factor.

A great deal of work has been undertaken over the years on Highland brochs, on the east and west coasts. These include in Caithness 19th to early 20th century excavations in Caithness by Tress Barry, Rhind, Anderson and Laing (Heald and Barber 2015). Modern investigations of brochs by AOC Archaeology have taken place at Nybster (MHG1593), Whitegate (MHG1645), Keiss Harbour (MHG1659), Keiss Road (MHG1650), Everley (MHG1792) and Thrumster (MHG2043). On the west, MacKie (2007) excavated a number of Wester Ross and Skye brochs in the later 20th century. AOC Archaeology and Historic Assynt have recently excavated Clachtoll broch in northwest Sutherland (MHG13002). There the preservation was exceptionally good, allowing for a range of detailed analysis, including the study of the spatial distribution of artefacts (Case Study Clachtoll Broch). A number of new, detailed surveys of Highland brochs have also been undertaken in recent years (Cavers et al 2015). 

Several key points have emerged from this work that had provided evidence for the terminology debates and issues relating to the people who built and used the cARs. Most of the sites are multi-phased, and detailed dating is needed.

Some Highland cARs are built on sites that were probably earlier hillforts as accepted in the Atlas of Hillforts, for example: Ness (MHG2151) and Skirza Head (MHG655) in Caithness; Dun Chealamy (MHG19718), Kilphedir (MHG9856), and An Dun (MHG9158) in Sutherland; Dun Cruinn on Skye (MHG6488); and Dun Lagaidh, Wester Ross (MHG7812). Other evidence suggests the siting of cARs near an even earlier prehistoric past. For example at Carn Liath, Golspie, the broch was built near a Bronze Age cist burial (MHG10872). There is also great scope re-analysing finds from old antiquarian excavations; a good example of what can be gleaned by looking at old finds from Caithness brochs it the publication by Heald and Jackson (2001).


Case Study: Clachtoll Broch

Leave a Reply