More than a thousand prehistoric circular fortified structures built of unmortared stone are known in Scotland, mostly on the western seaboard and islands and among the Orkney and Shetland Isles. In the past they have been categorized as ‘brochs’ or ‘duns’, brochs being the larger category, but the distinction is often unhelpful and recently the more neutral term ‘Atlantic roundhouses’ has gained currency. However most of the buildings formerly classified as brochs are notably more sophisticated in their construction than the others, and in the new nomenclature these have become ‘complex’ Atlantic roundhouses or, more prosaically, ‘cARs’.

CARs as defined in this study ‘have an external drystone wall of circular or sub-circular-to-oval shape, about three to six metres in thickness. This wall is split into two separate wall leaves and the resulting void between them was lintelled over. The resulting superimposed intramural passages are commonly referred to as “galleries”. Some galleries contain a stair that supposedly led to upper floor levels.

Cells and galleries at ground-floor level were generally accessible from the internal circular central area.’

Some cARs may ‘only have been one storey high; others appear to have evidence for several upper floors’. More than 500 sites in Scotland fit these defining criteria, mostly in the core western and northern Highland areas and among the western and northern isles, though 11 ‘outliers’ have been recognized in the Lowland south. While relatively few have been scientifically dated they appear to have flourished from the late 1st millennium BC to the early centuries of the Christian era. Many though by no means all are situated close to the sea.

Variously explained in the past as bolt-holes against Roman slave-raiders, ‘Pictish’ towers, or Viking strongholds, they are now firmly regarded as indigenous vernacular structures of the Late Iron Age which functioned as focal points of some kind, usually in locations where terrestrial and/or maritime resources might be exploited. But what this tells us about those who built them, and for what purpose, is far from clear—were they parts of a defensive system; fortified farmsteads of an agrarian/fishing elite; tribal muster points; or safe gathering-places for scattered and perhaps more egalitarian societies?

What is certain, however, is that their unmortared stone architecture is massive and complex, and shows a deep understanding of structural principles and the properties of the building materials involved—principally stone and wood, supplemented by turf, thatch, and cordage. Some at least were tall—Mousa in Shetland still stands 13m high and its walls curve inwards towards the top like a modern cooling tower (Fig. 1).


Figure 1. a) The partly collapsed complex Atlantic roundhouse at Dun Telve in Glenelg, Highland Region, showing the double wall construction; b) The well preserved complex Atlantic roundhouse at Mousa, Shetland, which still stands 13 m tall; c) The complex Atlantic roundhouse at Mousa, Shetland, in its maritime setting. (All photos: Colin Martin)

Download figure to PowerPoint

While most cARs are hollow-walled structures of the ‘broch’ variety, another form of stone buildings formerly called ‘wheelhouses’ also fits this general category. These are substantial circular houses with radial masonry piers which presumably supported a roof and separated the building into peripheral bays around a central hearth. The cAR category also includes substantial timber roundhouses, although our knowledge of these is restricted to ground plans only, revealed by slots and post-holes in the subsoil.

For all their uniqueness and interest, stone-built cARs have been patchily studied and rarely excavated to modern standards. Questions of how they evolved, how they were fitted out and roofed, and what activities took place inside them, have long been matters of speculation and controversy. However an alternative approach is to analyse the existing remains on objective architectural principles—given what has survived, what solutions might have been adopted to fill in the missing elements? Would such solutions have been within the competence and resources of prehistoric builders, and would they have been viable and effective in practical terms?

This is the approach adopted by Tanja Romankiewicz, whose undergraduate training included a study of architecture and architectural conservation before she undertook the PhD research of which these volumes are the outcome. It is an impressive body of work. Volume 1 comprises the main text and Volume 2 is a descriptive catalogue of the 148 sites investigated. The latter contains a vast corpus of plans and photographs, the fruits of a systematic trawl of the literature and many site visits. This format allows the user to follow the argument by reading Volume 1 with Volume 2 to hand for ready cross-referencing and amplification of the relevant sites. It may sound clumsy, but it works.

This is not a book for the casual or faint-hearted reader but, for those interested in these enigmatic structures and what they imply about the societies that created them, the considerable effort required to digest it pays rich dividends. Romankiewicz leads us carefully through the fundamentals of cAR design from an architectural historian's viewpoint, defining characteristics which inform the investigation and interpretation of the surviving structural remains. How the buildings were roofed, and how timber may have been used for flooring and internal partitioning, is addressed at some length. Her interpretations are aided by analogies from post-medieval vernacular architecture in the region, where traditions of unmortared building and the use of wood, turf, thatch, rope, and stone weights for roofing survived into the comparatively recent past. Reconstructions based on this evidence show cARs in their landscape contexts, suggesting that one function at least was to impress and dominate. Intervisibility between some examples—particularly ones in maritime settings—further suggests that such buildings may have been elements in linked networks of surveillance and communication, presumably to enhance control, security, and defence. Among Romankiewicz's other conclusions is that while cARs were sophisticated and highly developed buildings within a common tradition, they do not adhere to standardized designs, and many show evidence of individuality, even experiment, in their construction.

It was once thought that buildings of such complexity could not have originated locally, and elaborate diffusionist theories placed the sources of this architectural tradition on the Continent from where, driven by population pressures, invaders moved across the Channel into SW England and thence via Ireland into Scotland bringing their building traditions with them. These simplistic invasion theories have been replaced by a recognition that from the earliest times Europe's Atlantic seaboard has been a network of maritime routeways, along which peoples and ideas have travelled and interacted. Within this zone discrete geographical groupings, receptive to external ideas but each with its own distinctive culture, evolved and flourished. Thus, although some similarities can be detected in architectural traditions from Armorica (Brittany) through Cornwall and Ireland to western and northern Scotland, none seem to have been ‘copied’ or directly influenced by the others. The concept of a maritime ‘Atlantic façade’ as a prime driver in European prehistory, first articulated as long ago as 1963, has recently been developed by scholars such as Cunliffe and Henderson (see reviews in IJNA 32.1 and 40.2). Thus we can now see the cARs of Scotland as a unique phenomenon, created indigenously by those who lived there, but they were not developed in ignorance of what was going on elsewhere.

What, the reader may ask, does this have to do with nautical archaeology? Though no boat finds associated with cARs have yet been made, the proximity of many of them to the sea strongly implies a relationship with the maritime environment and the exploitation of its resources. Though we do not know how societies associated with the Scottish Atlantic roundhouse tradition were structured, it is clear from their architectural achievements that they were capable of drawing together and organizing skilled work-forces with which to undertake co-ordinated technological enterprises of considerable sophistication, replicating what had been successful in the past but not afraid of trying out new ideas. The best of the surviving structures come close to the limits of what is achievable with unmortared stone. We can surely assume that cAR dwellers applied similar levels of skill, adaptability, and occasionally innovation to building the water-craft that the maritime bias suggested by the locations of many of the structures imply. It follows that they would have exploited to the limits the properties of wood, skin, and associated materials to fashion seagoing craft as elegant and successful on the water as the cARs were on land. Exactly what forms such vessels may have taken must await nautical archaeological evidence, which we must hope may one day be found.

But useful work can still be done to set cARs into the maritime landscapes and cultures of which they were clearly a part. This is not an aspect that Romankiewicz's research has specifically addressed, though proximity to the sea is mentioned as a factor in the siting of some of them. Taken as a whole, however, this corpus (together with the outstanding resource of Scotland's National Monuments Record, accessible at, provides a solid foundation for wider contextual studies of this distinctive architectural tradition, seen from a maritime perspective. Such work is already under way through the aerial photography programme run by the Royal Commission on the Ancient and Historical Monuments of Scotland, in which this reviewer is participating. Initial results are promising, with many of the sites now being examined in their wider maritime contexts. Many were clearly located to take advantage of natural harbours and landing-places, some of which show evidence of artificial enhancement (which may, of course, be later). Some of these putative harbours are ‘secret’, in the sense that ships berthed inside them would not have been visible from seaward, while others have entrances which could easily be negotiated by those familiar with the hidden dangers but not by strangers unaware of them.

All this makes the point that the terrestrial elements of maritime cultural landscapes, though primarily the preserve of ‘muddy boots’ archaeologists, should also be studied by those who view them from seaward perspectives. That the book under review barely touches on these aspects is not a criticism in itself—on its own terms it is a remarkable and ground-breaking achievement. But it challenges us all to see maritime archaeology—and indeed history—as integral parts of the wider disciplines, and not simply as optional extras.