Fields in Bronze Age Scotland
Since the beginning of the 1970s and Richard Feachem’s (1973) summation of the work on highland zone field-systems by the OS Archaeology Division, it has been clear that there are extensive traces of prehistoric agriculture across the uplands of Scotland, represented by scatters of small cairns, stony banks and lynchets. Following excavations in the late 1970s on Arran (Barber 1997) and more recent work at Lairg, Sutherland (McCullagh and Tipping 1998) it is also irrefutable that the majority of these remains are probably Bronze Age in date. However, by the yardstick of lynchetted field-systems and other enclosures taking in huge blocks of countryside in southern Britain (Fleming 1988; 2008; Yates 2007), the Scottish field-system has always appeared comparatively small-scale and incoherent, so much so that individual fields with four sides are unusual, and recognisable systems of fields so rare that they can only be presented as exceptions rather than any rule. By way of example, Drumturn Burn in Perth and Kinross (RCAHMS 1990, 47-9), with its photogenic trackway wending down between fields to a cluster of hut-circles often appears as a text book illustration of a Scottish field-system, and yet it is the only one of its kind amongst the dozens of hut-circle groups of north-east Perthshire. Another half dozen have a recurrent arrangement of banks in the immediate vicinity, but none of these define any fields as such. Small wonder that Stuart Piggott’s Celtic Cowboys ( 1958), written at a time when Celtic field-systems were thought to date from the Iron Age, proved an enduring explanation of subsistence strategies in northern Britain.
This same contrast between northern and southern Britain has to some extent been recast by more recent work, most notably by David Yates ( 2007), albeit that the data are fuller and the analysis draws on more sophisticated social and economic arguments. Yates’s model couples the coincidence of large tracts of enclosure in southern Britain with concentrations of metalwork deposition to argue the case for the intensification of stock production in a competitive prestige goods economy. On the basis of the random sampling of commercial excavation he has established a reasonably robust distribution for the extensive enclosure of parts of the southern English landscape in the Middle Bronze Age. This pattern, which roughly echoes the division between lowland and highland Britain coined by Sir Cyril Fox in The Personality of Britain ( 1932), leaves little doubt that settlement and agriculture in northern Britain developed on a different trajectory without resorting to large scale enclosure, though whether this can be equated with non-intensive farming systems is more contentious, particularly when developed into core periphery models that emphasise the relative importance of one area above another in the manner that Kristiansen has argued in Scandinavia ( 1987). The work on Arran (Barber 1997) and at Lairg (McCullagh and Tipping 1998) is now amongst a host of excavations revealing the massive expansion and intensification of settlement in the Middle Bronze Age throughout the Scottish landscape (e.g. Kintore; Cook and Dunbar 2008). As in southern England, this expansion of settlement does not extend unbroken into the Iron Age, and in the uplands at least is suddenly fractured at the end of the 2nd millennium BC. While Iron Age settlements in southern England reveal evidence of much larger quantities of cereals than their Bronze Age counterparts, it is far from clear where these were being grown. In southern England, it seems, cereal production in the Iron Age did not require extensive enclosure as such and earlier Bronze Age field-systems were only brought back into use in the late Iron Age (Yates 2007, 110), the same period when settlement reappears on many Bronze Age settlements across Highland Scotland (e.g. Lairg).
Despite the contrast between north and south, Scotland has a rich and varied record in a wide range of environmental settings. In many of these there is also the huge potential to exploit palaeoenvironmental data as a proxy record of local vegetation cover from which to reconstruct the history of land-use and contemporary farming systems. The majority of the visible remains lie in areas of rough pasture in the margins of the modern landscape, often under a blanket of peat. Inevitably, their equivalents are largely missing from the areas that now comprise Scotland’s best agricultural land. Though skewed in this way, the excavation of settlements in advance of development in the eastern lowlands is providing records of the agricultural economy in the form of carbonised plant remains. At Kintore, Aberdeenshire, for example, small quantities of cereals were recovered, but little evidence of chaff, echoing the findings at Lairg, while stratified beneath the Early Bronze Age barrow at North Mains, Perthshire, a ridged field surface was uncovered (Barclay 1989). The potential of large standing Bronze Age monuments to preserve fragments of earlier field surfaces across the lowlands is an important resource.
The accretion, survival and destruction of sediments is no less an issue in the Highlands, where the ongoing succession of settlement and land-use prior to the growth of blanket peat differentially preserves earlier soils and surfaces. Hut-circles are often found to have been built on previously cultivated soils, and may equally be ploughed over after their abandonment, as can be found at both Lairg (McCullagh and Tipping 1998) and on Arran (Barber 1997). Indeed, at Tormore on Arran ploughing also encroached upon one of the hut-circles between two major phases of occupation (Barber 1997,8-11), while at Cnoc Stanger, in a machair environment, a cultivated soil was found separating two phases of a hut-circle that used elements of the same stonework at the entrance (Mercer 1996). Occasionally, a series of banks give the impression of greater order, such as the example noted already at Drumturn Burn, or Tulloch Wood in Moray (Carter 1993), but in general there is no tidy Bronze Age landscape preserved in these remains as such, so much as fragmentary archaeological deposits providing insights into complex histories of settlement and land-use over long periods of time. The field remains simply do not conform to concepts of field-systems driven by southern English models. Essentially, the field-system of Bronze Age Scotland is an untidy, cumulative and haphazard piece of ground, shaped on the one hand by the presence of earlier remains, and on the other by the intensity and extent of the cultivation practices that have taken place successively within its compass. Cumulatively, through time and the ongoing process of successive cultivations, it consumes its own history, and while relatively deep sediments may accumulate as a record of this history, trapped against an undisturbed baulk or an earlier boundary, dating the beginning of the process is as fraught with difficulty as dating its end. Field soils are by their very nature stirred and mixed contexts, and in Scotland rarely yield any cultural material that may assist in the establishment of a coarse chronology. As often as not the cessation of cultivation is conferred by a basal peat date and the assumption that the onset of peat growth followed the final season of cultivation in relatively short order (see Carter in McCullagh and Tipping 1998, 157).
The apparent absence of large scale landscape enclosure is not universal. On Shetland around settlements such as Scord of Brouster (Whittle et al. 1986), which although often referred to as Neolithic settlements are probably mainly Early Bronze Age and later, there are extensive enclosure walls which can be followed for hundreds of metres, dipping in and out of the peat blanket. As has happened in mainland Scotland, detailed investigation will probably reveal a more complex history, but at face value the Scord of Brouster field-system, with its enclosing walls, low lynchets and clearance heaps, seems to form part of a much bigger enclosed landscape (see also Whittington 1978). In Argyll, on the Atlantic seaboard, the dykes beneath the Black Moss of Achnacree, at Black Crofts, North Connell, hint at a series of large enclosures of perhaps 10ha extending back from the shore, but this is very unusual and has no immediate parallels (Carter and Dalland 2005).
Evidence of Bronze Age arable farming, though rarely of the form of the plots themselves, is also routinely recovered from machair deposits. These blown shell sands are found mainly around bays in the Northern and Western Isles, and though prone to catastrophic deflation, equally accrete in ways that with the addition of midden material create long and deep sequences of successive field soils with finely preserved evidence of the tools that were used to till them. Such soils underlay the hut-circles at Cnoc Stanger, Caithness, and have been found extensively in Orkney, such as at Tofts Ness, where there is evidence of the addition of manures including heathland turf (Dockrill et al. 2007, 386). Many of the deepened soils found in the Northern Isles around Iron Age settlements at places like old Scatness are also likely to have their origins in the Bronze Age. Examples of hut-circles and field enclosures are also beginning to be discovered in the moorland in the interior of some of the large Orkney islands, whereas in the Western Isles there is little sign of settlements penetrating inland. Again the evidence of ploughing and manuring is extensive, for example on South Uist at Cladh Hallan or at Rosinish, Benbecula (Shepherd and Tuckwell 1977). This sort of evidence, however, is found everywhere that there are coastal sands, for example at the mouth of the River Ythan at Rattray, Aberdeenshire, where in addition to ploughmarks and cultivated soils, part of a burnt and carbonised hurdle fence was found (Murray et al. 1992). Machair seems to have attracted farmers throughout the Bronze Age.
The rig uncovered at North Mains also holds the possibility that there were ridged field surfaces, although the successive restatement of the agricultural landscapes has effectively erased what evidence may once have been visible. The recorded stratigraphic relationships of cord rig in the Border counties (Halliday 1982; Topping 1989; Halliday 1993) suggest that this form of cultivation is mainly of Iron Age date and later, and it remains absent from the traces of fields that occur around several unenclosed platform settlements. In south-eastern Scotland it also seems likely that several of the field-systems normally considered Romano-British in date are probably Bronze Age, such as on Ellershie Hill in Lanarkshire, where an unenclosed house platform belonging to an adjacent settlement is apparently cut into the leading edge of a lynchet. Elsewhere, the only stratified rig that is significantly earlier cuts across a deposit probably dating from the Late Bronze Age or Early Iron Age in the machair at Kilellan Farm, Islay, Argyll (Ritchie 2005, 38fig. 46).
The final category of evidence worth considering is provided by the artefactual record, in the Northern Isles principally represented by stone ard points and other bar tools that have probably been used as hoes and mattocks (Rees 1979; Clarke 2006). Many of these are probably Bronze Age in date, including those ard points found broken off at the ends of ploughmarks at Kebister, Shetland (Owen and Lowe 1999, 263-5), though how they were mounted is open to debate. Waterlogged deposits elsewhere have also preserved a number of remarkable wooden agricultural implements. Although perhaps Iron Age rather than Bronze Age, the Milton Loch share and the Lochmaben plough beam have been long known (Fenton 1968), together with several ard tips from peat in Shetland, but the dating programme of the National Museum of Antiquities has also taken in several older discoveries with surprising results. These include an ox yoke from Loch Nell, Argyll, and a swingle-tree from Shapinsay, Orkney (Hedges et al. 1993; Cowie and Shepherd 2003, 164) both of which are firmly Bronze Age in date.
The combination of sources – palaeoenvironmental, excavation proxies, deep stratified soils and visible field remains – suggests that Bronze Age arable agriculture in Scotland was not only extensive, but in places sustained and intensively practiced. There was almost certainly some regional variation in agricultural practice, which in the Northern Isles gave rise to large patches of deepened soils; these were artificially created through intensive manuring regimes and the addition of moorland turf. The benefits of these practices and the addition of domestic rubbish were evidently understood throughout the Bronze Age, the adding of turf and peat representing an extreme form of nutrient transfer from grazing land. Whether this latter practice, which persisted throughout Scotland into the post-medieval period, was limited to the arguably hostile farming environment of the Northern and Western Isles remains to be demonstrated. The equivalent locations in the mainland have been intensively cleared, dug over and improved for at least the last two centuries. Although areas of deepened soils have been mapped in the lowlands of eastern Scotland, these are generally believed to be of medieval date. On higher ground, the emphasis of the agricultural regime may well have leant more heavily towards stock farming, and it is known from palynology that the uplands were extensively grazed. In these areas arable farming may have been more akin to gardening, though it is still a persistent feature wherever there is settlement. The general absence of fixed field boundaries is probably the most significant contrast with the evidence from southern England. Possibly the pattern of arable farming was part of a much more dynamic and flexible approach to farming in which arable may have been constantly shifting in location and extent. This seems to be manifested in later cord rig plots recorded in the southern uplands, where the visible rig seems to represent the last season of cultivation within much larger areas of turned and smoothed ground demarcated by fainter traces of cultivation scars (Halliday 1993). This too is perhaps the explanation of the multiple lines of stake fencing and the intercutting ploughmarks found beneath the stone circles on Machrie Moor (Haggarty 1991). Enclosure, such as it was, was temporary, designed for folding beasts onto the next year’s arable rather than to formalise the layout of the farming landscape.
To return briefly to Yates’ model of farming in southern England, the supposed contrast with the north may be more apparent than real. While enclosure can be used to increase stocking ratios on downland, this is only achieved by intensive management of the pasture within the fields. Indeed, the hypothesis that enclosure per se represents intensification rests on the ideas of improved farming in the historic environment. As yet, there is no other yardstick of the use of these fields by which direct comparisons can be made into the integrated palaeoenvironmental and archaeological datasets that have been recorded in Scotland. The greater contrast is probably between the implicit long-term stability implied by current interpretation of the enclosed landscapes of southern England, and the sheer complexity of land-use pattern that the Scottish datasets reveal. In themselves, these latter may well hold a lesson for researchers further south.