Woodland cover in Scotland reached its peak extent around 5800 cal BP (Tipping 1994; Edwards et al 2019), so the landscape of the early Neolithic was dominated by trees, with the likely exception of exposed summits and ridges and active floodplains. In Perth and Kinross, this consisted of mixed deciduous woodland, often dominated by oak with hazel and elm, but with variations in species abundance reflecting the mosaic of exposure, altitude and soil type. Hence pine is suggested to have colonised some areas, such as the upland Edradour Burn at Carn Dubh (Tipping in Rideout et al 1996), but not lowland catchments like North Mains, Rae or Stormont Lochs. There species like oak and elm were stronger competitors (Caseldine 1980; Hulme and Shirriffs 1986, Edwards and Whittington 1998a). The dominance of oak and elm also varied between low-lying, mid-slope and higher ground in eastern Perthshire (Caseldine 1980). The palisaded enclosure at Forteviot indicates that the woods included mature trees, yielding large oak timbers (Brophy and Noble 2020).
Two key features of the palaeoecological record around the start of the Neolithic are a decline in the abundance of elm and the appearance of cereal pollen. Accumulated evidence has overturned the idea that the elm decline is synonymous with the start of the Neolithic. Instead, multiple elm declines are recorded over a protracted period. For instance, in Perth and Kinross the dates for this transition vary widely, from around 6500 cal BP at North Mains (Hulme and Shirriffs 1986) to 5300 cal BP at Creag na Caillich (Tipping et al 1993). Often the elm decline coincided with reductions in the abundance of other tree species, such as at Edradour Burn (Tipping in Rideout et al 1996), Heatheryhaugh and Loch Maraich (Caseldine 1980), and Rae Loch (Edwards and Whittington 1998a). Some elm declines were accompanied by the appearance of disturbance indicators (pastoral or charcoal) as at Rae Loch (Edwards and Whittington 1998a), but this is not universal. At Edradour Burn, sustained, low intensity grazing may have prevented regeneration, generating a decline in elm from around 5800–5600 cal BP, since more intensive pastoral activity does not occur until around 5600–5300 cal BP, after the main reduction in elm (Tipping in Rideout et al 1996). Overall, the evidence suggests that the attrition of elm was a successive, not synchronous, process driven by multiple factors (Parker et al 2002).
The first farming communities were established in a generally well-wooded landscape, even at higher altitudes (eg Caseldine 1980; Tipping 1995). However, the age, distribution and extent of settlement is difficult to detect in the pollen record because of the abundance of tree ‘pollen rain’ and poor dispersal of cereal pollen, possibly from initially small fields. For instance, taking two lowland sites, cereal pollen was present during the Neolithic from North Mains (Hulme and Shirriffs 1986) but was not detected at Methven bog (Milburn 1997). Spatially patchy evidence for arable activity in the early Neolithic may reflect a combination of real variations in the spatial distribution and extent of cultivation and the proximity of activity to sampling sites for pollen that is very locally dispersed. In contrast, the absence of cereal pollen and potential arable ‘weeds’ throughout the Neolithic from upland sites like Creag na Caillich and Carn Dubh, where the records are primarily local in extent, may indicate the real absence of higher altitude cropping (Tipping et al 1993; Tipping 1995). These areas did not remain unexplored, however, with evidence for low intensity grazing from around 5800 cal BP at Carn Dubh. In contrast, the first phase of quarrying for axe manufacture around 4700 cal BP at Creag na Caillich, by Ben Lawers, appears to have taken place without grazing. This indicates a complex mosaic of landscape access and settlement (Caseldine 1980).
Soil pollen samples need to be interpreted with care due to suboptimal preservation conditions and limited dating controls. When these blurred snapshots are combined with more secure ‘off-site’ pollen sequences, they can provide tantalising indications of finer-scale dynamics in the landscapes around Neolithic and Bronze Age monuments. For instance, soil pollen evidence for a mosaic of grazed grass-heath and secondary cover from the land surface below the Cleaven Dyke earthen bank contrasts with the dominance of mixed deciduous woodland in the catchments of Rae and Stormont Lochs during the first half of the 6th millennium BP and approximately 4km away (see The Cleaven Dyke Case Study). This suggests that open ground was relatively limited in extent or only just becoming visible in the wider pollen record at the time of construction, raising questions over visibility and motivation (Edwards and Whittington 1998a and 1998b; Noble 2017). Limited local ground preparation is also indicated by soil micromorphological analysis of land surfaces beneath the monument, since the samples do not indicate extensive ground clearance (eg woodland clearance, burning) prior to construction (Simpson and Davidson 1998). Although cereal pollen only appears at Rae Loch around 5040 cal BP, the presence of pollen from taxa often thought of as ‘weeds’ of arable and pastoral ground from around 6010 cal BP and soil erosion from around 5360–5090 cal BP could suggest earlier adoption of arable within a largely forested setting within which the monument was constructed (Edwards and Whittington 1998a). Similar insights into spatio-temporal dynamics come from North Mains, where there is evidence for prolonged cultivation during the Neolithic, pre-dating barrow construction (Hulme and Shirriffs 1986). Local cereal cultivation is also indicated post-henge construction at Moncrieffe House. There was a possible period of abandonment between the timber henge and stone circle phases, which allowed secondary woodland growth, before mixed but pastoral dominated agriculture was re-established, possibly in the Bronze Age (Caseldine 1983). Data from multiple sites of different sizes within a region can be used to explore issues of landscape heterogeneity and land use visibility, and access to pollen data from multiple sites allows alternative scenarios to be explored.
