3.3 Settlement, Landuse and Resources

Settlement forms

The settlement evidence is apparently the least responsive regional aspect of the record until the Late Bronze Age, when in the South‐east at least a distinctive regional pattern of enclosed settlements seems to make its appearance. Across much of the South‐east there is little to be seen of any settlement before this period, the notable exception being the unenclosed platform settlements of upper Tweeddale and Clydesdale. The appearance of these clusters of platforms, often on quite steep slopes, is not immediately akin to the hutcircle groups of the Highlands, though in essence they are no more than groups of roundhouses, sometimes with traces of fields and stone clearance in attendance. The hut‐circle is otherwise ubiquitous across mainland Scotland, with no distinguishing features, though in Galloway they tend to occur in ones and twos, sometimes with baffle walls around their entrances, while in Perthshire and Sutherland they often occur in larger groups of a dozen or more. The greatest contrast is to be found in the Northern Isles, where the buildings are often more oval in overall shape, while the interior is less regular, with a series of alcoves set back into the thickness of the wall.

Prehistoric houses exist in both Highland and Lowland contexts  set either in  solitary locations or clustered in the remains of a settlement. In general terms the houses survive in more immediately intelligible form in the Highland zone and, if these are typical of what existed elsewhere,  the standard form was circular or oval, with an internal diameter between 5m and 12m and with a single entrance aligned to the South or East. In upstanding sites, the walls survive as a penannular embankment of stone and earth, often with evidence of numerous phases or stages of construction. Either as part of the original design or as a by-product of use, the interior contains a penannular hollowed area running concentric to the wall. In the lowland context, many examples have this gully more deeply cut forming perhaps a basement or cellar. In both zones many sites have the penannular gully infilled and surfaced with a slab pavement within the duration of occupation, perhaps indicating a cessation of one kind of use.

Beyond this simple pattern, there seems to be no consistent orthodoxy of construction (some have posthole rings, some have occasional post-holes, some have none) or evidence of use (some have stone hearths, some have hearth pits, some have burnt ground surfaces, some have no evidence of a hearth). Many houses in the Highlands have the entrance marked by large, often quartz, boulders but again this is does not apply to the majority of cases.

This interpretation of settlement form and development may apply to parts of northern England and mainland Scotland, but there is variation elsewhere, particularly in the Northern and Western Isles – and these island group differ from each other. Early Bronze Age timber built houses occur in Shetland, the Western isles, and possibly Orkney, and , as referred to above these are more oval than round in shape. Early Bronze Age stone built houses occur also in Shetland and Orkney, where their form can be paired houses of either oval form or, in Orkney at Crossiecrown, of a form similar to Skara Brae house plans. Although the Crossiecrown houses are built to traditional late Neolithic design similar to Skara Brae, they are an isolated pair similar to other Bronze Age houses  rather than being part of a village settlement like Skara Brae. In the middle and late Bronze Age, terraced stone built houses occur at Cladh Hallan, and at Jarlshof. In Shetland and in Orkney paired stone built oval and round houses are found, such as those in Shetland at Sumburgh (see the ScARF Case Study: Sumburgh), and in Orkney at SkaillDeerness, Links of Noltland, and numerous unexcavated examples are apparent.

Rarely do the undamaged archaeological remains of Bronze Age settlement show any evidence of defence by suitable enclosure against wild beasts or aggressive neighbours. Yet it seems reasonable to suspect that both were sometimes real threats, especially in times or seasons of shortage. Large communal defences are not yet recognised as present in the Scottish Bronze Age, but some late Bronze Age instances, such as Eildon Hill, may be defensive. If the general absence of defences is real then should such a simplistic reconstruction of the primary drivers of Bronze Age life be challenged? Were the perceived threats or risks resolved or averted by procedures that lie beyond current powers of reconstruction.

See also Section 3.3.1 Burnt Mounds

Knowledge Gaps

Sites with concentrations of structures in close proximity still present huge difficulties in chronology. It remains practically impossible to identify contemporaneity between two adjacent buildings or temporal difference between two stratigraphically sequenced structures.

The archaeological remains of post-holes and earthen walls do not usually retain enough evidence of the duration and intensity of use of any structure. It might be more accurate to say that they do not contain the kind of evidence that is currently often recognized and nano-archaeology may well have a great deal to offer here. Occasionally excavation chances upon remains that are revealing, for example a building destroyed by fire, but the prospection techniques required to increase the chances of detecting such sites are currently lacking. It is also true that archaeologists are seldom equipped well enough to handle the wealth of bio-chemical information that can exist in well preserved contexts – opportunites need to be maximized for such analyses (as in Cladh Hallan).

How buildings were utilized and inhabited is under-researched and needs to be considered on an intra- and inter- regional basis.

A synthesis of settlement evidence for the whole of Scotland is required.

Settlements need to be considered more in terms of their relationship to landscape (see the ScARF Case Study: The Lairg Project).

The relationship between a building and a dwelling is yet to be understood. Would people dwell in buildings as families? Do variations in building design reflect social differentiation or do variations in building shape, footprint or internal layout reflect any meaningful variations in human behavior? Current research is woefully ignorant of social relations within any Bronze Age context; for example, one may suspect that slaves existed but how power was managed is not yet understood.

If the absence of defences is real then should such a simplistic reconstruction of the primary drivers of Bronze Age life be challenged? Are the perceived threats or risks beyond modern day comprehension?

Models of Bronze Age Land-use

There is little surviving tradition of demarcating territories with built boundaries in the Scottish Bronze Age landscape, and until the creation of the improved landscape some two to three hundred years ago the bounds of landholdings marched largely ‘where wind and water shear’. Distinctive natural features might also be pressed into service, but only occasionally were these supplemented with artificial markers. Thus, in contrast to the landscape of south-west England, where extensive Middle Bronze Age boundary-systems suggest the existence of fixed territories at several scales, for most of Scotland it can only be an assumption that such territories may once have existed. As a result, interpretations have tended to focus on the exploitation of the resources in the landscape, firstly through the proxy records provided by the excavation of settlements, and secondly through the locations of those settlements within the landscape and the traces of farming activity around them. More recently these have been integrated into a third category of evidence drawn from palaeoenvironmental analyses and reconstructions, operating at both local and regional scales.

Traditional approaches to the settlements of the Bronze Age landscape have assumed that the groups of hut-circles and platforms found in the uplands represent the remains of farms that occupied their sites continuously for hundreds of years and variously cultivated and grazed the ground round about. Implicit in this interpretation is that their presence in numbers in the uplands is testimony to the extent of settlement in the lowlands by the Middle Bronze Age, and that they therefore represent the expansion of a relatively dense and static pattern of sedentary settlement on the better lowland soils.

In this model it is assumed that each settlement had a specific area of relatively fertile land for cultivation, pasture for grazing, and woodlands for fuel, structural timbers, pannage and other forage; to these basic requirements for settlement and agriculture can be added water – seldom far away in the Scottish landscape. From what is known about manuring practices at this time, it is clear that cultivation and pasture went hand in hand, the one in part dependent on the other for the transfer of nutrients into the arable soils. In some cases this was by deliberate inclusion of composted household waste, animal manure and turf, in others probably by management of the grazings. Difficult to demonstrate either archaeologically or from palaeoenvironmental sources, this traditional practice in historical times, operated by pasturing the beasts on the hill by day and folding them onto the next years arable by night; known as ‘tathing’, this could have operated from after harvest to cultivation in the spring, and possibly explains the interwoven fence-lines and ploughmarks discovered beneath two stone circles on Machrie Moor. In the documented farming systems of post-medieval Scotland each holding required a ratio in excess of 10:1 between pasture and arable to work successfully, though as the population increased in places like the island of Tiree it fell to 5:1 shortly before clearance and massive depopulation and reorganisation took place (Dodgshon 1998).

The widespread palynological evidence that the Bronze Age landscape was extensively grazed, and the general absence of systems of stockproof fields with permanent boundaries, suggest that once the crops began to sprout the beasts were driven to pastures at some remove from the parent settlement, though we should probably anticipate that they were always tended to prevent them straying. This raises the possibility that there were also temporary settlements occupied only in the summer months. In some cases this may provide a role for the enigmatic burnt mounds that are found scattered through the landscape.

In the uplands, the mix of the farming systems may well have leant more heavily towards beasts than arable crops, but though it is conceivable that in some places landuse was driven by wholly pastoral communities, nowhere has this been demonstrated. Indeed quite the reverse; evidence of arable farming seems to go hand in hand with hut-circle settlements, with the caveat that the crops that were being grown around them cannot necessarily be identified. If any farms specialised in cereal production for export, these will have been mixed farms with large herds to maintain the fertility of their soils. In part, the idea that grain was exported to some settlements rests on the absence of chaff from the carbonised plant remains recovered from settlements, but this is a general pattern across north Britain and is more likely to relate to the nature of the evidence and the conditions required for its preservation (e.g. the burnt hut-circle on Tormore, Arran; Barber 1997, 6-25).

More recent syntheses, however, have challenged this model (Halliday 1999; 2007; Barber and Crone 2001), emphasising that the duration of occupation of individual buildings is at best uncertain, perhaps representing periods as short as ten years, and that a characteristic of some of the multiperiod buildings that appear to represent longterm occupation is that their use is episodic and punctuated by periods of abandonment of equally unknown duration. While it is still possible to argue that a group of hut-circles represents longterm settlement on a hillside, this needs to be demonstrated rather than assumed. Equally possible is that the group of hut-circles as a whole represents episodic occupation of that hillside in a much more complex and dynamic pattern of settlement and land-use, in which clusters of apparently separate hut-circle groups represent a series of locations successively occupied and cultivated by a single settlement.

If these new approaches have any value, they disrupt the idea that each hut-circle group formed a relatively static agricultural unit, standing with its arable fields within a mosaic of grazings and woodlands in its immediate vicinity. All the various elements of the farming system – the settlement, the arable, the pasture and the woodland – are still present, but rather than being immediately around the occupied hut-circles, they may be spread over a much wider area containing half a dozen or more other hut-circle groups, to say nothing of the numerous discrete scatters of small cairns where there is no visible evidence of settlement. Presumably the arable ground would move round with the settlement, if only so that the crops could be protected from wild herbivores, but the arable plots left fallow around each of the unoccupied hut-circle groups represent the prime grazings. To some extent the management of these grazings would see some nutrient transfer onto this ground year on year; to make the comparison with the locations of the huts of later shielings, these are often distinguished by more verdant vegetation than their surroundings. The active folding of beasts onto next years arable would only be seen around the occupied settlement, after they had been brought back from the more distant pastures.

The implications of which of these two models is closest to an archaeological truth for the typical Middle Bronze Age settlement in mainland Scotland are far-reaching. With the abandonment that seems to be visited on so many upland hut-circle groups towards the end of the 2nd millennium BC, it raises the question of what changed. In the traditional model, the population collapsed catastrophically, though the post-medieval comparison drawn above for the ratios required between arable and pasture hints at one route to abandonment without recourse to natural events. In the second model, however, the impact of the whole system of settlement and land-use on the landscape is archaeologically exaggerated, with each community moving around multiple settlement foci, while continuing to graze and forage throughout the rest of the landscape. Within the resolutions of the palaeoenvironmental record, such impacts are as yet indistinguishable from those of settlements occupying single sites, but implicitly in this model neither the archaeological nor the palaeoenvironmental records need be the result of the large upland population that the traditional model demands to populate the extant settlements; after all, the Middle Bronze Age seems to be the period of most extensive settlement in either prehistory or history. But if this was an in extenso system of settlement and land-use in the first place, it could be as easily reorganised and supplanted by more intensive practice operating from a single site. Thus, the apparent abandonment at the end of the 2nd millennium BC is possibly about economic change rather than population collapse, but until Late Bronze Age settlements are discovered in the uplands the hypothesis remains untested. If there is any continuity in the settlement record into the Late Bronze Age in the uplands, sufficient work has taken place to indicate that it is unlikely to be found on the visible hut-circle groups, which do not seem to be re-occupied until later in the 1st millennium BC. Routinely the most favourable locations, and thus the most likely sites for more intensive farming, are occupied by modern farms and their improved ground, where nothing is now visible above the ground. Notably, the palynological record at Lairg (McCullagh and Tipping 1998) shows grazing continuing after the abandonment of the visible hut-circles; even if the people left, beasts apparently remained.

Irrespective of the model of land-use that is favoured, the question arises as to how far the evidence from the uplands is representative either of the whole Bronze Age, or of the entire area of Scotland. To some extent the lowland settlements that have been excavated reveal similar chronologies, with this same focus of occupation in the mid to late 2nd millennium BC, though at Kintore, Aberdeenshire, after a statistically significant hiatus in the sequence of radiocarbon dates, the site was re-occupied early in the 1st millennium BC in the Late Bronze Age and Early Iron Age (Cook and Dunbar 2008). In this same part of North-east Scotland, the recumbent stone circles that have been excavated also seem to show a sudden burst of activity at the end of the 2nd millennium BC, with the insertion of cremation deposits after an interval of 1000 years or more since they were constructed (see Welfare 2011, 164-5, 167, 260-3). In the case of the undated circle at Loanhead of Daviot it appears that the cairn was reconstructed as a ring-cairn at this time, as if it was believed that this was its original form. South of the Dee at Cairnwell, a small ring-cairn was newly built on the site of a small timber enclosure and again associated with cremation burials (Rees 1997). The creation of monuments that replicate those built a millennium before is extraordinary, and suggests powerful forces at work. That it seems to coincide with the abandonment of so many hut-circle groups from Arran to Sutherland and Caithness is surely no coincidence and hints at a widespread dislocation of society. The new chronologies of settlements enclosed by palisades, banks and ditches that have been established on the Lothian Plain, which now can be traced back into the Late Bronze Age (Haselgrove 2009, 193-6), reveal the equivalent change in the South-east of Scotland, heralding the new regional settlement patterns that emerge in the Iron Age, with the striking contrasts between the drystone architectures of the north and west, and the timber and earthwork architectures of the east and south. Indeed, the end of the 2nd millennium BC is probably the key period of change that shapes the regional character of the Iron Age settlement record.

Loanhead of Daviot, Aberdeenshire ©RCAHMS

While mainland Scotland presents a general impression of synchronous periods of Middle Bronze Age settlement expansion and then collapse, particularly in the uplands, it would be a mistake to assume that this was the case everywhere. In Shetland, for example, the general tenor of the settlement record seems to run some 500 years in advance. The moorland settlements there, none of them at any great altitude but nonetheless subsumed into peatlands, seem to represent Early Bronze Age settlement, and at Scord of Brouster, for example, are falling into disuse in the middle of the 2nd millennium BC (Whittle et al. 1986), though this is not necessarily the case everywhere in Shetland. With a relatively low-lying landscape, the differences between upland and lowland, coastal and inland, moorland and pasture are probably finely drawn in complex patterns that do not lend themselves to simple generalisation. The extensive systems of stone dykes round the moorland settlements also hint at different land-use models, as do the areas of deepened soils, some of which seem to be of this early date. These relationships are only poorly understood.

In the Western Isles too, the records of buildings associated with Beaker pottery and agricultural soils in the machair on the coast of South Uist (Parker Pearson et al.  2004), for example, may represent different models of settlement and land-use, adapted to the local environment and its capacity for agriculture. The inland areas of many of the Western Isles appear unremittingly bleak in agricultural terms, the moorland peats often lying directly on to little more than bare rock. Thus physical evidence of Early Bronze Age settlement and fields is otherwise largely missing from the mainland, though at Cnoc Stanger, in Caithness, in a coastal setting there also seem to be deep cultivated machair soils of comparable antiquity (Mercer 1996).

With the exceptions of the Northern and Western isles, however, the Bronze Age settlement and land-use archaeologies seem remarkably homogenous. The caveat that might possibly be offered concerns south-eastern Scotland, where the clusters of house platforms on relatively steep hillsides in the headwaters of the Clyde, the  Tweed and other rivers, strike a contrast with the scatter of ring-banks found in the large cairnfields of northern Lanarkshire. The significance of this is uncertain, for elsewhere in the South-east the succession of later land-use seems to have largely erased any visible evidence of Bronze Age settlements, upstanding examples of which only re-appear in north Northumberland. The recently published palynological work in the Bowmont Valley in the northern Cheviots, however, clearly demonstrates Bronze Age farmers were moving into the uplands (Tipping 2010). At Swindon Hill, for example, at an altitude of about 365m OD, Tipping recovered near continuous records of Barley cultivation from c. 2850 to 1150 cal BC, albeit with a relatively low temporal resolution, and at Sourhope at 260m OD from 1800 to 1350 cal BC, while he more tentatively suggested cultivation was also taking place at the higher Cocklawhead plateau, above 500m OD, from 2100 to 1250 cal BC ( 2010, 174). The archaeological evidence for the settlements and fields from which the pollen is derived is as yet more tenuous. But while some of the pollen diagrams clearly show Early Bronze Age farming impacting on the uplands, it is notable that the arable farming at all these locations had ceased by the end of the 2nd millennium BC (Tipping 2010, 177), though grazing generally continues into the Late Bronze Age (Tipping 2010, 178), reflecting the broad trends noted elsewhere in Scotland.

Example Pollen diagram from Carn Dubh, near Pitlochry (Tipping 1995) ©Richard Tipping

Settlement technology

It has to be assumed that settlements were located with some regard to drainage, shelter and water. However, as many societies studied in recent centuries demonstrate, such considerations often lie well hidden within often magical or religious locational choice processes.

The two major imperatives of avoidance of in-breeding and survival of winter impose requirements of scale and complexity on Bronze Age buildings. In principle, the fully developed Bronze Age tool kit would permit sophisticated joint and trenail carpentry but for most buildings the limitations of binding and simple lap-joints impose limited ranges of ground-plan and roof forms. More sophisticated technologies were rapidly developed for soil management and perhaps the defining signature of the Bronze Age was the retardation of the declining trajectory of soils through relatively sophisticated land management (ploughing, manuring, land protection, etc).

Within the life experience of people living in Bronze Age Scotland, the life-span of all buildings was governed by the short-life span of the wood, thatch, turf and stone building components. It is unlikely that any building constructed mainly of wood could last, as a coherent structure, for more than 5 to 10 years without repair, and replacement of some components. Although Pope (2003) suggests a more generous estimate of 30-50 years for the majority of structures, this study should not be used to generalize across Scotland as there is great variation across time, and regional differences as outlined above. Indeed in the Northern Isles the large paired or double houses could have been occupied (with modification and repairs) for several hundred years.  

Knowledge Gaps

There is a gap between the well-exercised ability to micro-reconstruct the location for a Bronze Age settlement (e.g. using geomorphological and mapping techniques) and the little that is often done generally to model the topography within settlements.

It is necessary to undertake research to challenge the current architectural models for Bronze Age buildings, to test their applicability for various regions of Scotland.

See also the ScARF Case Studies: The Lairg Project and Teasing out regional settlement traditions


When we talk about duration of Bronze Age settlement (or indeed of prehistoric settlement in general), are we all thinking along similar lines when we use the term 'generational'? Does that represent c.25 years? And if so, is that more applicable as a generation in terms of modern life expectancies?