A Place for the Heathlands

The agricultural revolution of the late eighteenth to early nineteenth centuries, a product of the European Enlightenment, led to heathlands across Europe being framed as unproductive "waste" on the grounds that they were unsuited to the immediate practice of intensive agriculture. The resulting decline in their traditional management means that the continued survival of European heathlands is now under threat. Heathlands are currently listed as priority habitats for conservation under Annex I of the European Union's Habitats Directive.^{1 ( 1 )}
Council Directive 92/43/EEC; 21 May 1992. The conceptualization of heathlands as marginal and economically unproductive environments was perpetuated in the scientific and archaeological narratives of the twentieth century; in recent decades, however, the imposition of Enlightenment ideals on pre-Enlightenment societies is being challenged. In this paper, we explore how Enlightenment thinking may have influenced twentieth-century interpretations of human–heathland relationships in the prehistoric landscape of Orkney. Our particular focus is on the Neolithic–Bronze Age transition in the Orkney Islands.

Heathlands occur on acidic, nutrient-poor soils in regions of cool-temperate oceanic climate where available water supply exceeds losses via evaporation and transpiration.^{2 ( 2 )}
Gimingham 1972. Some tree taxa, including Betula spp. (birch), Pinus sylvestris (Scots pine) and Quercus spp. (oak), are capable of establishing on heathland, with the result that heaths are eventually replaced by woodland. Exceptions occur when tree growth is suppressed by disturbance such as fire or grazing, or where trees are close to their modern range limits, for example in coastal locations and mountains, where heath likely is the climate-controlled potential natural vegetation. In such extreme habitats, thin and nutrient-poor soils developed following deglaciation, providing suitable substrate for grasslands and open woodlands to develop in the early Holocene. High-intensity weathering (caused by, for example, frequent strong winds, oceanic salt spray, heavy rainfall, and/or low temperatures) degraded soils rapidly by inhibiting the breakdown of organic material by soil fauna, in turn facilitating the rapid leaching of nutrients and waterlogging, leading to gleying.^{3 ( 3 )}
e.g. Bunting 1996. The resistant and/or nutrient-poor subsoil materials in these areas led to limited availability of soil nutrients. Under these conditions, substantial areas of heathland replaced woodland and grassland through natural succession within as little as a thousand years from the onset of interglacial conditions, for example in northern Scotland^{4 ( 4 )}
Peglar 1979; Birks 1996. and the Western Isles of Scotland^{5 ( 5 )}
Birks and Madsen 1979. from 9000–8600 BP, on Orkney from 7000 BP,^{6 ( 6 )}
Bunting 1996. and on Faroe from 6000 BP.^{7 ( 7 )}
Jóhansen 1982.

In less extreme environments in which woodland was able to persist into the mid-Holocene, the widespread expansion of heathland following the onset of the Neolithic around 5000 BP is largely thought to have occurred either in response to human activity (which first removed tree cover, then provided the disturbance necessary to prevent full regeneration, e.g. in Shetland,^{8 ( 8 )}
Jóhansen 1985. western Jutland,^{9 ( 9 )}
Odgaard 1994. and southwestern Norway^{10 ( 10 )}
Prøsch-Danielsen and Simonsen 2000. ), or in response to a climatic downturn shifting the competitive balance towards heathland over trees.^{11 ( 11 )}
Birks 1996; Bunting 1996.

Before the agricultural revolution, the communities of early modern Europe living in landscapes supporting significant amounts of heathland had developed distinctive farming systems in which crops were grown intensively in areas close to settlements, and extensive livestock rearing took place on surrounding heathland. Nutrients were continually transferred from grazed heathland to arable fields by applying animal dung and/or heathland turves.^{12 ( 12 )}
Webb 1998. Heathlands were often managed as "commons," in which rights to grazing and other ecosystem services were available to most or all members of the local community. Commons have been a documented feature of European landscapes since medieval times and were probably established much earlier,^{13 ( 13 )}
e.g. Brück 2000; Johnston 2005; Fyfe et al. 2008. perhaps having cosmological as well as economic importance to prehistoric societies.^{14 ( 14 )}
Haughton and Løvschal 2023. Variations on this type of farming system were practised throughout the European heathlands for centuries.^{15 ( 15 )}
And perhaps millennia: e.g. Prøsch-Danielsen and Simonsen 2000.

New technologies developed during the agricultural revolution enabled, and were enabled by, changes in attitude by landowners and the wealthier levels of society. Despite their high value to local communities, the commons (and particularly heathlands) were now reframed as "wasteland" in need of improvement. Good land management came to be defined through high productivity and the generation of economic surplus, mainly through intensive arable farming and selectively bred domestic livestock raised on improved pastures. Other types of habitats were considered more or less marginal depending on what economic returns could be extracted. The labelling of heathlands as marginal during the agricultural revolution can therefore be viewed as an example of shifting baseline syndrome, with heathland moving from a central component of a sustainably managed landscape supporting a local community to a marginal component in need of improvement to deliver specific productivity outcomes.

Heathlands continue even today to be described as "marginal" in academic literature.^{16 ( 16 )}
Diemont et al. 2013; Plunkett and Swindles 2022. It has been noted that many archaeological interpretations of heathland use are often rooted in Enlightenment ideals of maximizing production, with prehistoric heath development often seen as an unintended consequence of the abandonment of cultivated land.^{17 ( 17 )}
Ombashi and Løvschal 2023. This has often given rise to a narrative of heathlands as "marginal" areas which were exploited only when this was economically necessary, for example during times of population growth.^{18 ( 18 )}
e.g. Bradley 1980; Burgess 1984. Increases in heathland extent are often interpreted as a negative environmental pressure on human communities due to their association with nutrient-poor soils. Although these interpretations have been challenged,^{19 ( 19 )}
e.g. Brück 2000; Johnston 2005. it is notable that these views appear more strongly in regions where the technology and philosophy of the agricultural revolution were particularly firmly imposed, often by an elite group of "outsiders" with the social and economic power to reshape landholding and land-use, for example in England and Scotland.

The archaeological record of Britain and Ireland seems to show the abandonment of, or substantial withdrawal of settlement from, upland and heath-dominated regions at various times throughout both the prehistoric and historical periods. These episodes are often explained in terms of decreased productivity and heathland expansion in response to climatic deterioration;^{20 ( 20 )}
e.g. Parry 1978; Burgess 1985; Barber 1998; Verrill and Tipping 2010. they are cited as examples of environmental marginality, whereby a critical environmental resource is absent or in short supply, or when a critical environmental variable crosses a threshold, triggering land abandonment by communities that are already operating at the limits of that particular environment.^{21 ( 21 )}
Coles and Mills 1998. The spread of heathland is accordingly seen as reducing the resources available, leading to upland population decline through either emigration or increased mortality.

Marginality can of course also be defined in terms of economic and social or political factors. Economic marginality, for instance, has been defined in terms of the viability of a landscape for a particular form of subsistence activity, and it has been argued that marginality may be created or ameliorated by changes in the subsistence base, technology, and/or by the organization of the economy itself.^{22 ( 22 )}
Coles and Mills 1998. Social and political marginality results from the political and cultural isolation of communities living on the edge of larger social groupings. This isolation may occur due to geographic remoteness, religious, ethnic or linguistic differences, or the status of the group concerned within wider society, and may result in them having either restricted access to resources, or access only to land considered to be marginal by other communities.^{23 ( 23 )}
Coles and Mills 1998. Perceptions of land value and marginality are therefore not solely defined by environmental parameters; definitions are not static, and can vary due to cultural differences, with perceptions of the marginality and relevance of a region or environment by elite groups often dependent on its geographical distance from centres of power, for example.

The designation of particular landscapes or environments as "marginal" is therefore highly subjective and dependent upon the perspective of the designator. The process of designation is often carried out at some distance, whether through the instrumentality of elite or outsider status on the part of the designator or because the designator is looking back through time at the archaeological record. However, the degree to which a landscape or environment is considered "marginal" can, and should, also be framed in terms of the lived experience of the inhabitants of those landscapes and environments.

The "Bronze Age climate decline" and the associated abandonment of upland and heathland settlement is a classic case study of marginality found in many textbooks.^{24 ( 24 )}
e.g. Bell and Walker 1992; Champion 1999; Roberts 2014. Archaeological structures from the earlier Bronze Age such as extensive field boundaries, found in areas of Britain in the present day dominated by heathland, are widely interpreted as a response to growing populations and increasing pressure on resources, which necessitated expansion into more "marginal" parts of the landscape.^{25 ( 25 )}
e.g. Bradley 1980; Burgess 1985; Fleming 1985; Yates 2007. However recent work in some regions has challenged the interpretation of early Bronze Age occupation of the uplands as "new" settlement, for example through evidence for Neolithic upland mixed agriculture at Lairg in northern Scotland.^{26 ( 26 )}
McDonald et al. 2023. Bronze Age structures may simply be more archaeologically visible markers of ongoing activity, rather than indicating new settlement,^{27 ( 27 )}
Brück 2000; Johnston 2005. although their increased visibility may still be due to increased demands on that landscape or competition over its resources.

During the later Bronze Age in Britain, a shift to smaller, dispersed, less substantial settlement structures and reduced monumentality in heathland terrain is also generally attributed to a withdrawal of populations from these "marginal" areas. A later Bronze Age climate-driven spread of heathland and decrease in human occupation has been postulated for several upland regions of Britain, including Bodmin Moor^{28 ( 28 )}
Brown 1977. and Dartmoor^{29 ( 29 )}
Caseldine 1999; Fleming 2008. in the southwest, and parts of northern Scotland.^{30 ( 30 )}
Burgess 1985. Across northwest Europe, a widespread climatic downturn has been identified in palaeoecological records at around 850 cal BC,^{31 ( 31 )}
e.g. van Geel et al. 1996; Mauquoy et al. 2004; Blundell and Barber 2005; Swindles et al. 2007. and this is generally cited both as the primary cause for the spread of heathland and blanket peat seen at that time in climatically marginal regions, and as a possible reason for the changes in settlement. Tephra deposits from the Icelandic Hekla 3 eruption in 1159 BC have been cited as an additional driver for the changes, accelerating the process of soil acidification and the subsequent spread of heath and blanket peat.^{32 ( 32 )}
Grattan and Gilbertson 2000.

However, a growing body of palaeoclimate data for northwest Europe shows spatial and temporal variation in both climate shifts and heathland formation, even in light of a global climatic change considered strong enough to mark a subdivision of the Holocene around 4.2kaBP.^{33 ( 33 )}
Walker et al. 2012. Although the Atlantic, oceanic, and hyperoceanic parts of northwest Europe generally show less marked climatic variability due to oceanic buffering, the idea of a late Bronze Age climatic downturn forcing people to abandon "marginal" landscapes has persisted, especially in publications aimed at a general audience.^{34 ( 34 )}
e.g. Ritchie 1995. More recent palaeoenvironmental research on Bodmin Moor^{35 ( 35 )}
Gearey et al. 2000. and Dartmoor^{36 ( 36 )}
Amesbury et al. 2008. has shown that, while there is some evidence for climatic deterioration coinciding with apparent shifts in settlement pattern in these areas, this is an overly simplistic explanation for the changes observed. On Bodmin Moor, pastoral activity continued following the disappearance of Bronze Age settlement archaeology,^{37 ( 37 )}
Gearey et al. 2000. implying that any climatic deterioration was not so severe as to cause a total abandonment of the area. On Dartmoor there also seems to have been continuity of land-use, with a period of intensive pastoralism in the middle Bronze Age followed by lower intensity land-use in the late Bronze Age and Iron Age.^{38 ( 38 )}
Fyfe et al. 2008. A review of the Scottish evidence shows no clear signs of the abandonment of upland landscapes as a consequence of climatic change during the late Bronze Age.^{39 ( 39 )}
Tipping 2002. In all these regions it is now suggested that a complex combination of both environmental and socioeconomic factors was responsible for the change in emphasis of activity from upland to lowland regions seen during the later Bronze Age.^{40 ( 40 )}
Gearey et al. 2000; Tipping 2002; Amesbury et al. 2008. The shift to more dispersed settlement patterns and reduced monumentality has previously been interpreted as showing impoverishment, but this interpretation is itself likely influenced by modern ideas of progress and status.

Orkney, a group of islands situated off the northern coast of Scotland (Figure 1), provides an excellent opportunity to study the possible influence of environmental conditions on prehistoric human activity within a geographically defined area with a rich archaeological record. Today most of the islands support relatively limited heathland, but some areas are rich in heath - for example the West Mainland hills, and the islands of Hoy, Rousay, Eday, and Stronsay. The modern Orcadian landscape is one with clear divides between green "lowlands" of improved grassland and crop fields and brown, heath-dominated coastal fringes and "uplands" (Figure 2), known locally as "hill" and used mainly for sheep grazing, nature conservation, or tourism and outdoor recreation, with settlement and intensive agriculture clustered in the lowlands.^{41 ( 41 )}
Lee 2007.

A broad understanding of Orkney's environmental history based on an earlier generation of palaeoenvironmental studies^{42 ( 42 )}
Moar 1969; Davidson et al. 1976; Keatinge and Dickson 1979; Davidson and Jones 1985. was incorporated into twentieth-century syntheses of Orcadian archaeology.^{43 ( 43 )}
e.g. Renfrew 1985; Ritchie 1995. The general narrative was that birch-hazel "scrub" woodland covered the islands until c. 3500 cal BC, when it was cleared for agriculture by Neolithic settlers. The resulting mixed agricultural landscape was then altered by spreading heathland at around 1800 cal BC in response to a shift to more oceanic climatic conditions combined with increased grazing pressure.^{44 ( 44 )}
Keatinge and Dickson 1979. However, as more palaeoenvironmental research has accumulated, a more nuanced understanding of prehistoric land-cover and land-use in the islands has emerged, with both woodland history and heathland development being more complex than originally presumed.^{45 ( 45 )}
Bunting 1994; 1996; de la Vega-Leinert et al. 2007; Farrell et al. 2014; Farrell 2015; Bunting et al. 2022. Even though the main phase of development and expansion, whether caused by human activity, natural processes, or a combination, seems to have been during the Bronze and Iron Ages,^{46 ( 46 )}
Keatinge and Dickson 1979; Farrell 2015; Timpany et al. 2022. it is notable that at some exposed locations (e.g. Loch of Knitchen, Rousay), heathland became established during the Mesolithic, long before any clear signal of human activity is recorded.^{47 ( 47 )}
Bunting 1996.

Archaeological research in Orkney has tended to focus on the spectacular and well-preserved Neolithic and Iron Age remains. Bronze Age (defined in Orkney as the period 2200-800 cal BC) features are much less visible, with the best recorded type of domestic site being "burnt mounds" - large heaps of burnt stone and dark soil containing charcoal and ash, often associated with a stone slab-lined trough and situated close to a source of freshwater. Suggested functions include cooking, bathing, tanning of hides, fabric processing, and brewing.^{48 ( 48 )}
Brown et al. 2016. In Orkney few of these mounds have been scientifically dated, but those that have span the early Neolithic through to the Iron Age, with most dating to the Bronze Age.^{49 ( 49 )}
Anthony 2003. The apparent lack of Bronze Age settlement combined with the "insular" material culture from the period^{50 ( 50 )}
Øvrevik 1985: 137. has led to suggestions that the Neolithic-Bronze Age transition represents an economic or cultural "decline" in Orkney, described as a "prehistoric recession".^{51 ( 51 )}
Ritchie 1995: 86. This has been attributed to environmental factors such as climatic deterioration, soil exhaustion, deposition of tephra from the Hekla 3 eruption, and/or the spread of heathland and blanket peat.^{52 ( 52 )}
e.g. Øvrevik 1985; Ritchie 1995. Blanket peat development and heathland spread at three sites in West Mainland is dated to the mid to late Bronze Age,^{53 ( 53 )}
Keatinge and Dickson 1979. providing some support for the narrative of heathland spread as a contributing factor to an environmentally induced cultural "decline."

The story of Bronze Age decline - a story that frames heathland development as an intrinsic part of the processes leading to the increased marginalization or growing insularity of populations - is persistent in discussions of Orkney's prehistory. The striking contrast between the monumental Neolithic archaeology and the smaller-scale, dispersed Bronze Age sites fits the explanation well. The islands today have an unusually productive agricultural lowland for their location in northern Scotland, thanks to an underlying geology of mainly sandstone, which produces relatively fertile soils. Heathland is largely confined to more exposed areas only used for recreation, conservation, and extensive sheep grazing, reinforcing the narrative. There is also clear evidence for the removal of substantial areas of heathland by agricultural improvement from the mid-eighteenth century onwards, particularly on islands owned by a single landowner,^{54 ( 54 )}
e.g. Shapinsay, North Ronaldsay, and Burray; Mamwell 2017. a removal made possible and profitable by the relatively fertile soils. At some locations, small patches of heath persist within the agricultural zone – notably inside the Ring of Brodgar monument within the Heart of Neolithic Orkney World Heritage site (Figure 3) – demonstrating the role of human activity in the presence or absence of heath. The effective replacement of heathland via historical agricultural "improvements" has been used to support the interpretation of heathland as a sign of the reduced productivity of the Bronze Age landscape and its unfavourable conditions for human habitation.

Three broad hypotheses can be proposed to explain the changes observed in the archaeological record across the Neolithic to Bronze Age transition in Orkney. These are population collapse, environmental change leading to a shift in land-cover (e.g. heathland expansion), and cultural change. We briefly summarize each hypothesis below and consider how it might be reflected in the palaeoecological record, before bringing together all the available palaeoecological data to evaluate the evidence for each.

This hypothesis argues that the end of the Neolithic in Orkney was a period of upheaval leading to rapid loss of population, either through famine and death or emigration, in response to environmental stressors (see below). A reduction in the amount of land suitable for agriculture and poorer agricultural yields^{55 ( 55 )}
e.g. Øvrevik 1985; Ritchie 1995. would mean that the large population inferred from the abundant and substantial late Neolithic archaeology could no longer be supported.

In response to a population collapse, the palaeoecological record would be expected to show both the existence of environmental stressors and a decrease in the level of human activity manifesting as a reduction in cereal-type pollen and pastoral indicator taxa,^{56 ( 56 )}
Behre 1981. and potentially an increase in "natural" regenerated vegetation. There might also be evidence for changing land-use patterns, with a move away from parts of the landscape where agriculture was less productive, since the needs of a smaller population would not require exploitation of these areas.

This hypothesis argues that some combination of a shift to wetter and/or cooler climatic conditions, human-induced soil exhaustion due to intensive late Neolithic agriculture, and/or soil deterioration due to the acidifying effects of volcanic ash deposited by Hekla 3 occurred. All of these phenomena would have contributed to the spread of blanket peat and heathland across the landscape. At some locations in Orkney, soil changes in response to high levels of exposure are argued to be responsible for the replacement of woodland by heath before the arrival of Neolithic people;^{57 ( 57 )}
Bunting 1996. agricultural pressures could have accelerated this process in other parts of the islands. These changes would have reduced the resources available to support the human population, meaning that the above-subsistence levels of activity (including monumental construction) seen during the later Neolithic could no longer be sustained.

If the proposed climatic changes had occurred, the palaeoecological record should show evidence of increasing bog surface wetness, along with more heathland and grazing and less arable cultivation in the wider landscape. If soil deterioration were part of the problem, a shift from arable cultivation to pastoralism in more productive areas might accompany the heathland expansion in marginal areas; if Icelandic volcanic ejecta had affected soil quality, tephra would be recorded in sediment sequences and a widespread, simultaneous "crisis" should be observed in the palaeoenvironmental record. In all these situations, there might also be evidence for increased pressure on more fertile land, or a change in emphasis from arable cultivation to pastoralism. There might also be evidence from the archaeological record for the adoption of new technologies and subsistence methods in response to increasing environmental stress.

Across Scotland and the rest of Europe, Bronze Age cultural practices seem to give greater significance to individuals than to membership in larger communities.^{58 ( 58 )}
e.g. Ashmore 1996; Harding 2000. This might reflect increased social fragmentation, which in turn might drive the division of the landscape into territories that took place throughout Britain during the Bronze Age.^{59 ( 59 )}
e.g. Fleming 2008. Personal identities in this period might have been defined less by lineage and more by territory, with control over land being as important as control over people.^{60 ( 60 )}
Parker Pearson 1999. Agriculture, animals, and land were central to Bronze Age cosmology;^{61 ( 61 )}
Yates 2007; Downes and Thomas 2013. it has been suggested that the Bronze Age world revolved around an "agricultural metaphor".^{62 ( 62 )}
Williams 2003: 244.

If the changes seen in Orkney are solely the result of cultural processes, then a lack of evidence for a change in farming type preserved within the palaeoecological record might be expected, although given the centrality of agriculture to Bronze Age ways of life, there might be some indications of intensification. It might also be possible to observe changing patterns of land-use driven by a more dispersed population.

In this section we discuss palaeoecological records from across Orkney, with a particular focus on three sites investigated by the first author of this text, Michelle Farrell.^{63 ( 63 )}
Farrell 2009. The three sites were chosen because they differ in the degree of perceived environmental "marginality" in the present day, and the discussion aims to provide a more detailed and up-to-date synthesis of Bronze Age environments and land-use in the islands than has previously been available.^{64 ( 64 )}
e.g. Davidson and Jones 1985. The locations of the three small wetland basins studied by Farrell^{65 ( 65 )}
Farrell 2009. are shown on Figure 1: Whaness Burn, Hoy; Hobbister, Orphir, Mainland^{66 ( 66 )}
Farrell 2015. and Blows Moss, South Ronaldsay.^{67 ( 67 )}
Farrell 2009; Farrell et al. 2014. Of these three sites, Whaness Burn is the most environmentally "marginal" today, being situated in an upland, peat-covered, heathland-dominated landscape, and Blows Moss is the least "marginal," being low-lying and surrounded by relatively fertile soils formed over boulder clay which support improved pasture and arable cultivation. Hobbister lies between these two sites, in a location where the land is mostly below 60 m above sea level (the proposed upper limit of prehistoric settlement and farming in Orkney^{68 ( 68 )}
Lamb 1989. ), with a varied surficial geology of peat and boulder clay supporting a mix of heathland, improved grassland, and arable cultivation. Figures 4-6 illustrate the present-day setting of the three sites.

The main phase of peat initiation and heathland spread in Orkney seems to have been during the Bronze Age. Heathland was present at Mill Bay on Stronsay from c. 2100 cal BC in the early Bronze Age,^{69 ( 69 )}
Tisdall et al. 2013. the spread of heathland began in the West Mainland hills between c. 1750 and c. 1100 cal BC,^{70 ( 70 )}
Keatinge and Dickson 1979. and heath was present at Tuquoy on Westray by the end of the Bronze Age.^{71 ( 71 )}
Timpany et al. 2022. Some heathland was present at Whaness Burn^{72 ( 72 )}
Farrell 2009. and Hobbister^{73 ( 73 )}
Farrell 2015. from the start of the Bronze Age, although widespread heathland development at these sites did not occur until later.

Evidence from organic deposits associated with archaeological sites in the more fertile areas of West Mainland and South Ronaldsay suggests low-intensity mixed farming practices in the early to middle Bronze Age.^{74 ( 74 )}
Jones 1975; 1977; Downes 1994. Pollen data from some sites within modern hill areas such as Lesliedale Moss^{75 ( 75 )}
Jones 1979. and Glims Moss^{76 ( 76 )}
Keatinge and Dickson 1979. in West Mainland imply little Bronze Age agricultural activity, although other pollen assemblages from similar settings indicate quite intensive pastoralism in the uplands.^{77 ( 77 )}
Keatinge and Dickson 1979. Sequences from the modern agricultural lowland provide evidence for cereal cultivation at Loch of Skaill^{78 ( 78 )}
Keatinge and Dickson 1979. and of continuing pastoralism from the Neolithic until the present day around Scapa Bay.^{79 ( 79 )}
de la Vega-Leinert et al. 2007

The three sites investigated by Farrell^{80 ( 80 )}
Farrell 2009; 2015. suggest that agriculture expanded into modern hill areas during the Bronze Age. At Blows Moss the surface of the wetland itself was probably exploited for grazing from the late Neolithic until the mid-Bronze Age;^{81 ( 81 )}
Farrell 2009. pastoralism continued in the surrounding landscape until the sequence ends in the late Bronze Age at c. 1010 cal BC. There is also evidence for cereal cultivation near to this site throughout the Bronze Age, with possible intensification in the later part. A similar sequence of events is seen in the pollen record from nearby Liddle burnt mound, with pastoralism and low-intensity arable cultivation during the early Bronze Age and a shift to more intensive mixed agriculture in the middle Bronze Age.^{82 ( 82 )}
Jones 1979.

At Whaness Burn on Hoy^{83 ( 83 )}
Farrell 2009. there is evidence for grazing on the valley floor between c. 1750 and c. 1170 cal BC (early to middle Bronze Age), probably associated with the occupation of an enclosed settlement within the valley, which - although unexcavated - is believed to be Bronze Age.^{84 ( 84 )}
Lamb 1989. Heathland was managed by deliberate burning, probably to maintain the quality of the vegetation for grazing. An increase in surface wetness at the site occurs during the later Bronze Age, which may explain the decline in pastoral activity indicated by the pollen record at c. 1170 cal BC, perhaps suggesting that the settlement in the valley was abandoned then. No evidence of cereal cultivation is present at this site.

At Hobbister,^{85 ( 85 )}
Farrell 2015. grazing activity continued throughout the Bronze Age, perhaps more intensively than during the Neolithic. Sub-peat dykes identified at the site^{86 ( 86 )}
Sharman 2007. may represent Bronze Age field systems buried by later peat development. Heathland here was also managed by deliberate burning, and grazing probably took place on the site itself as well as in the surrounding landscape. Towards the end of the Bronze Age there seems to have been a slight decline in grazing activity, perhaps due to an increase in surface wetness of the sampled mire rather than a change in the wider landscape, where cereal crops were grown throughout the Bronze Age.

In summary, the palaeoecological record shows that mixed agriculture was practised in more fertile regions of Orkney throughout the Bronze Age, with some evidence for slight intensification during the later Bronze Age (e.g. Blows Moss^{87 ( 87 )}
Farrell 2009. ). More use began to be made of hill areas (e.g. Whaness Burn;^{88 ( 88 )}
Farrell 2009. West Mainland hills^{89 ( 89 )}
Keatinge and Dickson 1979. ) and the surfaces of fens and peatlands in the lowlands (Blows Moss and Hobbister^{90 ( 90 )}
Farrell 2009; 2015. ) for livestock grazing. There is an apparent decline in grazing activity in the late Bronze Age, with grazing on lowland wet mires such as Blows Moss reduced, though it is likely that grazing continued in the wider landscape. These changes might be artefacts of increasing surface wetness at the cored peat sites, which both makes local grazing less common and affects the sensitivity of the pollen record itself.^{91 ( 91 )}
Farrell 2009; 2015. This general increase in surface wetness may be indicative of regional climatic change,^{92 ( 92 )}
e.g. van Geel et al. 1996; Mauquoy et al. 2004; Blundell and Barber 2005; Swindles et al. 2007. or may simply be the natural growth of peat bodies over time.

Today's division of the Orcadian landscape into lowland and "hill" (Figure 1), with heathland overwhelmingly confined to hill and food production focused in the lowland areas, reinforces the Enlightenment perception of heathland and upland areas as less productive and therefore marginal for human occupation. The hill areas clearly do have economic value - for example from dispersed grazing, conservation value, recreation, tourism and provision of ecosystem services - but in forms that were either less significant or entirely ignored by the paradigms of the agricultural revolution. The narrative of Bronze Age decline in general texts about Orkney is one of deterioration as the landscape becomes more marginal and less suited for human settlement, resulting in a less-visible archaeological record. While the growing body of palaeoecological data does indicate an increasingly heath-covered landscape, it does not show the expected human response. In this section, we will revisit the three hypotheses outlined above.

It has been noted that the apparent scarcity of Bronze Age settlement evidence in Orkney is probably the result of failure to identify it, rather than a real lack of occupation at this time.^{93 ( 93 )}
Downes 2005. While few Bronze Age houses have been excavated in Orkney, several probable sites have been identified,^{94 ( 94 )}
Downes and Thomas 2013. including "double houses" at Skaill, Deerness,^{95 ( 95 )}
Buteux 1997. Wasbister (within the World Heritage Site),^{96 ( 96 )}
Robertson 2005. the island of Auskerry, and the Links of Noltland on Westray.^{97 ( 97 )}
Moore and Wilson 2011. Evidence for settlement continuity across the Neolithic-Bronze Age transition is also seen (e.g. at Crossiecrown in West Mainland;^{98 ( 98 )}
Card et al. 2016. Links of Noltland on Westray;^{99 ( 99 )}
Moore and Wilson 2011. Tofts Ness on Sanday^{100 ( 100 )}
Dockrill et al. 2007. ). While the known Bronze Age sites are mainly in hill areas, this may not be a true picture of their original distribution; archaeological sites of all ages have been destroyed by agricultural improvements in Orkney since the mid-eighteenth century.^{101 ( 101 )}
Mamwell 2017.

If population decline had occurred, the palaeoecological record should show an overall decrease in the level of human activity, especially in the less productive parts of the landscape, since a smaller population would mostly be supported by the more productive areas. However, the record shows, if anything, an increase in anthropogenic activity, for example intensification of cereal cultivation at Blows Moss, and greater pastoral activity around sites in hill areas such as Whaness Burn and Hobbister.^{102 ( 102 )}
Farrell 2009; 2015. Population decline is therefore an unlikely explanation for the changes seen in the archaeological record.

The Bronze Age "decline" in Orkney is often attributed to environmental factors such as climatic deterioration, soil exhaustion, or tephra deposition, all of which are seen as causing the spread of blanket peat and heathland and thereby reducing the productivity of the landscape and leading to a decline in standards of living and reduction in the surpluses needed to support a centralized society constructing impressive monuments.^{103 ( 103 )}
e.g. Øvrevik 1985; Ritchie 1995. However to date no Hekla 3 tephra has been found in Orcadian sediment sequences,^{104 ( 104 )}
Bunting 1994; 1996; Farrell 2009; Timms et al. 2017; 2018. making it unlikely that eruption deposits contributed significantly to heathland formation and apparent cultural "decline".

There is evidence for intensive manuring and modification of soils from across Orkney since the Neolithic.^{105 ( 105 )}
e.g. Simpson et al. 1998; Guttmann et al. 2006. If at any point these practices had become insufficient to maintain soil fertility, deterioration of soils via leaching and podsolization might have resulted in the spread of heathland in previously cultivated areas. Blanket peat formation and spread of heathland did begin in many parts of Orkney during the Bronze Age, but in hill locations where no evidence of prehistoric cultivation is seen. Overgrazing may have been a possible cause, but stocking densities were likely relatively low.^{106 ( 106 )}
e.g. Boyle et al. 2015; Rowley-Conwy et al. 2020. Archaeological evidence from Tofts Ness on Sanday and Links of Noltland on Westray shows that intensive manuring of calcareous sandy soils allowed sustained arable cultivation into the Bronze Age on soils subject to episodes of windblown sand deposition^{107 ( 107 )}
Dockrill et al. 2007; Simpson and Wilson 2011. which would otherwise have been suited to heath formation. Heathland actually seems to expand further in the Iron Age,^{108 ( 108 )}
Jones 1979; Bunting 1996; de la Vega-Leinert et al. 2007; Farrell 2015; Timpany et al. 2022. when monumental construction returned in Orkney. Heathland expansion due to soil exhaustion therefore seems an unlikely driver of the changes seen in the Bronze Age archaeological record.

Evidence from Whaness Burn and Hobbister shows that the surfaces of basin mires do become wetter as the Bronze Age continues. This, along with heathland expansion in the wider landscape as demonstrated by numerous palaeoecological sequences, might reflect a climatic shift.^{109 ( 109 )}
Farrell 2009; 2015. However, the presence of anthropogenic indicator taxa in the pollen record remains fairly constant throughout the Bronze Age, suggesting that these changes did not affect the level of human activity in the landscape as a whole. Grazing in hill areas and on the surface of peatlands and mires in the lowland^{110 ( 110 )}
Keatinge and Dickson 1979; Farrell 2009; 2015. suggests, if anything, increased exploitation of these heathy environments. Bronze Age pastoralism in Orkney was dominated by sheep,^{111 ( 111 )}
e.g. Nicholson and Davis 2007; Downes and Thomas 2013. marking a shift away from the predominantly cattle-based economy that characterized the Orcadian Neolithic.^{112 ( 112 )}
e.g. Mainland et al. 2014; Card et al. 2018. Sheep may have been better suited than cattle to grazing the developing heathland, and a more diversified suite of animals might reflect a response to changing resources. Widespread, intensive Bronze Age sheep husbandry is also seen widely in the English lowlands,^{113 ( 113 )}
Pryor 1996; Yates 2007. challenging assumptions that sheep farming was practised only in "marginal" upland regions as arable cultivation in the lowlands intensified.

Comparison of the three Orcadian pollen records investigated by Farrell^{114 ( 114 )}
Farrell 2009. implies that there were variations in land-use depending on local environmental conditions, as also seen at Garbh Allt in northwest Scotland following a shift to wetter surface conditions c. 950 cal BC.^{115 ( 115 )}
Tipping et al. 2008. In Orkney, arable cultivation occurred during the Bronze Age at Blows Moss (lowland) and to some extent at Hobbister (intermediate), while at Whaness Burn (upland, "hill") there is no evidence for cereal cultivation. Pastoralism took place at all three sites, with indications that wet heath and mire habitats were used for grazing at Blows Moss and Hobbister, perhaps reflecting increased use of former pasture for arable or fodder crops. Alongside the agricultural differentiation around these sites, heathland at Hobbister and Whaness Burn seems to have been managed by deliberate burning throughout the Bronze Age, probably in order to maintain the quality of the vegetation for grazing.

During the late Bronze Age, grazing activity on the surface of the wetlands studied by Farrell^{116 ( 116 )}
Farrell 2009; 2015. seems to have ceased, probably due to increases in surface wetness which might be due to climate change, wetland development, or a combination. Arable cultivation continued around Blows Moss and Hobbister, with slight intensification at the former site, and livestock grazing continued in the landscape surrounding all three sites, suggesting that while the wetlands themselves were changing, this was not related to wider landscape changes that would substantially affect human activity.

Overall, it seems that during the Bronze Age in Orkney people adapted their farming practices in response to changing environmental conditions. Orcadian farmers have modified their practices to suit local environmental conditions ever since the introduction of agriculture during the Neolithic, a key example being the use of seaweed as sheep fodder from the moment of their introduction to the islands,^{117 ( 117 )}
Balasse et al. 2019. and the Bronze Age environment does not seem to have presented significant new challenges. The apparent increased use of hill areas might be related to an increase in the number of pollen records available for the later Holocene or to a genuine change in demand for agricultural production - whether this was becoming more extensive in response to environmental change, increasing due to population expansion, changing in location in response to the reorganization of settlement patterns, or reflecting the more central role of agriculture in Bronze Age lifeways compared with the Neolithic.^{118 ( 118 )}
e.g. Downes and Thomas 2013.

There are clear cultural shifts at the transition from the Neolithic to the Bronze Age in Orkney, with a change from communal to individual burial and the establishment of a more dispersed settlement pattern. Similar shifts occurred across Britain at this time, with communal burial monuments falling out of use and the landscape being divided up for agricultural purposes.^{119 ( 119 )}
Parker Pearson 2005. Extensive boundary markers such as the Dartmoor reaves^{120 ( 120 )}
Fleming 2008. show clear demarcation of divisions in the landscape, perhaps as a consequence of social fragmentation. In Orkney, massive linear earthworks known as "treb dykes"^{121 ( 121 )}
Lamb 1983. run for long distances across some islands and were probably constructed during the Bronze Age. Social fragmentation in Orkney is hypothesized to have begun towards the end of the Neolithic period^{122 ( 122 )}
e.g. Richards 1998; Bayliss et al. 2017. and continued during the Bronze Age. The settlement evidence and the barrow cemeteries indicate a scattered settlement pattern, and treb dykes may have marked out boundaries between settlements. An alternative explanation has been proposed,^{123 ( 123 )}
Downes and Thomas 2013. arguing that rather than representing land allotments, these monumental divisions demarcate different domains, allocating space to particular beings, whether these be human or animal, real or supernatural. It is suggested that both Bronze Age domestic and funerary architecture and landscape organization in Orkney reference a north-south "sacred" axis, bisected by east-west-oriented paths and boundaries,^{124 ( 124 )}
Downes and Thomas 2013. similar to a cosmology proposed for the Scandinavian Bronze Age.^{125 ( 125 )}
Kristiansen and Larsson 2005. Division of the land in Bronze Age Orkney may therefore have been as much about ritual significance as practical considerations.

During the Bronze Age, Orkney may have become more culturally isolated from the wider European world. In both the Neolithic and the Iron Age, the islands were situated at the centre of trade routes across the North Sea and along the Atlantic seaboard of Europe, resulting in a dense network of intercommunication. In contrast, evidence for outside contact during the Bronze Age is limited to steatite from Shetland and a few metal objects like the gold discs from Knowes of Trotty barrow cemetery that indicate contact with southern England,^{126 ( 126 )}
Coles 1969. suggesting much reduced exchange with the wider world. There have been relatively few finds of Bronze Age metalwork from Orkney, leading to suggestions of growing cultural insularity.^{127 ( 127 )}
e.g. Øvrevik 1985. It is possible that climatic deterioration made the Atlantic sea routes more difficult to navigate; phases of increased storminess during the Bronze Age are inferred from a sediment sequence from Mill Bay on Stronsay,^{128 ( 128 )}
Tisdall 2013. and similar ages for windblown sand accumulations have been reported from Sanday.^{129 ( 129 )}
Sommerville et al. 2007. This evidence provides some support for the indirect effect of environmental change on cultural activity through disruption to trade networks.

However, there is no palaeoenvironmental evidence for a decline in agricultural activity at the Neolithic-Bronze Age transition in Orkney. If anything, the evidence points to an extensification and intensification of farming, with regional specialisms developing in different parts of the islands. Cereal cultivation intensified in areas of more fertile soils such as South Ronaldsay, and a pastoral specialism, likely dominated by sheep grazing, developed in heath-dominated areas such as the island of Hoy and upland areas of Mainland.^{130 ( 130 )}
Farrell 2009. It has been argued that many of the changes seen in material culture, burial practices, and landscape organization in the Orcadian Bronze Age are directly related to sheep rearing.^{131 ( 131 )}
Downes and Thomas 2013. It is likely that transhumance was practised, with people and sheep moving short distances on a seasonal basis, usually to upland heath pastures in the summer and lower valleys in winter. This intensification of stock and extensification of pastures would have necessitated the use of land that had not been used for farming before, and which could previously have been proscribed by cultural beliefs, necessitating new forms of tenure which were not wholly about ownership.^{132 ( 132 )}
Downes and Thomas 2013. Living in and moving through the landscapes of Bronze Age Orkney referenced a horizonal and vertical "ritual grid" in which the north and the uplands were more sacred,^{133 ( 133 )}
Downes and Thomas 2013. and the changes in farming practices and patterns of land-use evidenced by the palaeoenvironmental record may be a reflection of these new forms of engagement with agriculture, animals and land, and between people and the cosmos.

The changing view of heathlands during the agricultural revolution of early modern Europe clearly demonstrates the somewhat subjective nature of marginality as a concept as well as its dependence upon particular social, economic, or environmental perspectives. The late eighteenth- and early nineteenth-century elite "improvers", often from outside the community, defined the long-established and effective farming systems in heathland landscapes as inefficient; valuable communal resources were deemed unproductive "waste." This reframing made the imposition of new approaches to farming, often associated with the destruction of heathland, both morally and economically necessary.

Many archaeologists of the twentieth-century who sought to explain the patterns of prehistoric activity were educated within the same elite traditions; the influence of this perspective can be seen widely in the literature. However, the subjectivity of defining an environment as "marginal," and the interaction between sociopolitical, economic and environmental factors in creating this perception, are now becoming more widely recognized in both the archaeological^{134 ( 134 )}
e.g. Turner and Young 2007; Svensson and Gardiner 2009; Costello 2021. and the palaeoecological^{135 ( 135 )}
e.g. Tipping 2002; Plunkett and Swindles 2022. literature, especially with growing awareness of the importance of decolonizing our views both of other places and of other times.

In Orkney, heathland did become a more substantial part of the landscape mosaic in the Bronze Age. In the archaeological narratives of the twentieth-century, this expansion of heathland was interpreted as a deterioration in the agricultural resource available to the islands' inhabitants, reflecting a post-agricultural-revolution, Enlightenment attitude to landscapes not suited to intensive, high-control exploitation. Although there are distinct differences in the archaeological record that reflect cultural changes between the Neolithic and Bronze Age populations of Orkney, there is no clear evidence to suggest that the increase in heathland was linked to any "decline" in the resource base available.

Palaeoenvironmental analysis^{136 ( 136 )}
See e.g. Farrell 2009. indicates continuity of activity across the Neolithic-Bronze Age transition, rather than abandonment of areas that might be deemed more marginal. Such analysis indicates that the response to spreading heathland was greater geographical diversity in land-use, rather than any form of population collapse or abandonment of "marginal" areas. Orcadians adapted their practices in response to available resources throughout the Bronze Age. For example, as heathland spread around Hobbister and Whaness Burn, deliberate burning is inferred, probably to manage the quality of grazing. At Tofts Ness on Sanday and Links of Noltland on Westray, intensive manuring was carried out to allow continued cultivation in increasingly marginal coastal environments.^{137 ( 137 )}
Simpson et al. 1998; Dockrill et al. 2007; Simpson and Wilson 2011.

The changes seen in the archaeological record of Orkney at the Neolithic-Bronze Age transition are thus likely to have resulted from a combination of environmental and cultural factors. Cultural factors drove decentralization and changing settlement patterns, and heathland expansion changed the variety and distribution of natural resources across the islands. The available palaeoecological evidence indicates that during the Bronze Age a pastoral specialism, likely dominated by sheep farming, developed in the more "marginal" parts of the islands, while in areas of more fertile soils arable cultivation intensified. This may have occurred partially in response to the expansion and/or fragmentation of population that started at the end of the Neolithic;^{138 ( 138 )}
e.g. Bayliss et al. 2017. but it may also have been encouraged by a Bronze Age cosmology in which agriculture played a central role and upland/hill regions may have held particular significance.^{139 ( 139 )}
Downes and Thomas 2013. An increase in storminess may have led to a degree of cultural isolation and marginality implied by the lack of metal and far-traded objects in the archaeological record, but cultural marginality is not necessarily directly associated with scarcity of subsistence resources.

There are clear changes in the way that land was exploited between the Neolithic and Bronze Age in Orkney, and there is an increase in the structured use of different parts of the landscape typical of traditional heathland management across northern Europe. The evidence presented here suggests that these differences were the result of changes in social organization and cosmological beliefs, in combination with adaptation to changing environmental conditions. The formation and expansion of heathland in Orkney did not constrain human activity, but rather enabled a range of new possibilities in both ritual and practical terms. Heathlands, central to the new customs and practices of the Bronze Age, were abundant in the "sacred" uplands;^{140 ( 140 )}
Downes and Thomas 2013: 88. they also provided a range of useful resources, of which extensive grazing land was only one. Heathlands also provide fuel, bedding, and thatching and building materials, and these would have been highly valuable in what was by then a largely treeless landscape;^{141 ( 141 )}
Bunting et al. 2022. turves from grazed heathlands can also be used to improve soil quality.^{142 ( 142 )}
Kaland 2014. Bronze Age barrows in Orkney were often constructed in areas of heathland, and turves from the heath were used for fuel in Bronze Age cremations^{143 ( 143 )}
e.g. Downes 1994; Bunting et al. 2001. and at domestic sites such as the Links of Noltland.^{144 ( 144 )}
Hastie 2011. Increased numbers of small arable "infield" plots were created, and at Tofts Ness these were fertilized by the addition of heathland turves.^{145 ( 145 )}
Dockrill and Bond 2009.

Bronze Age Orcadians were skilled in managing the natural resources available to them. In response to heathland expansion they developed a local variant of the characteristic heathland farming systems practised throughout the European heathlands prior to the agricultural revolution.^{146 ( 146 )}
Webb 1998. The new interpretation presented here challenges the imposition of modern ideas of marginality onto past societies. The reframing of heathlands as unproductive "waste" during the early modern agricultural revolution contributed strongly to later assumptions in the twentieth-century about the negative impact of increasing heathland on Orcadian populations. It also contributed to the interpretation of less-visible Bronze Age archaeological remains as representing a period of decline or collapse of society. Challenging this paradigm through ongoing exploration of the palaeoecological and archaeological records may help to rebalance current attitudes to heathland in northern Scotland, where large areas were removed by historical "improvement" and where heathlands are still viewed as economically problematic despite new recognition of their biodiversity value. Such a challenge may encourage exploration of reinstating traditional farming practices as part of sustainable management and the conservation of these ecosystems.