By Kristian Pedersen and Alexander Brown
Investigating and understanding the interactions between hunter-gatherers and the environment has been identified as a key research theme for Mesolithic studies across Britain at both a national and regional level (eg. ScARF; Blinkhorn and Milner 2013).
Mesolithic activity has been recorded across the region in the form of flint assemblages (ScARF Mesolithic), but it is more challenging to identify the impact of these communities on their environment. Episodes of woodland disturbance are difficult to detect in the palynological record, and where visible, are equally difficult to interpret as categorical evidence of anthropogenic disturbance or manipulation of the vegetation.
Evidence for synchronous openings in the woodland have been tentatively identified in the Cheviots at Yetholm Loch and Sourhope dating to the late Mesolithic, with higher frequencies of microscopic charcoal recorded at Yetholm Loch and Ravelrig during the early Mesolithic (Tipping 2010; GUARD Archaeology Ltd 2014). Additional more poorly defined woodland disturbance episodes have been identified at The Dod (Innes & Shennan 1991).
Tipping (2010) suggests woodland disturbances by hunter-gatherers likely operated at the decadal scale. Many older pollen studies from the region lack the necessary fine-scale sample and chronological resolution required to resolve the details of vegetation dynamics. The low resolution of many palynological studies makes it extremely difficult to compare pollen data against other increasingly high-resolution climate proxies, in turn presenting a challenge to distinguish between disturbances resulting from natural or anthropogenic agencies.
Identification of anthropogenic impacts on woodland are increasingly complicated by the identification of episodic and sudden climatic fluctuations during the Holocene (Bond et al 1997; Anderson et al 1998; Croombe 2018). For example, climate-driven periodic short-term (c 50–100 years) synchronous declines in oak have been identified in the dendrochronological record from north-western Europe (Leuschner et al 2002; Spurk et al 2002) and should not be dismissed as a potential explanation for some of the oak declines apparent in pollen sequences from the region (see Tipping 2010).
Additional signs of woodland disturbance during the Mesolithic are apparent in the evidence for fires preserved in the charcoal records from peat and lake deposits, and increasingly quantified alongside palynological analysis (eg Tipping 2010; GUARD Archaeology Ltd 2014). However, many older pollen studies across the region lack associated quantification of microscopic charcoal. Moreover, there has been a tendency to interpret charcoal as evidence for anthropogenic-induced disturbance, heavily influenced by the evidence for upland burning from England and Wales (synthesised by Simmons 1996) and the close association between burning and occupation on lowland Mesolithic sites such as Star Carr (Mellars and Dark 1998).
There is little doubt that hunter-gatherers were capable of manipulating their environment, including using fire. It was used as a tool to encourage increased grazing by ungulates or to promote the growth and predictability of seasonally available plant resource. However, charcoal in pollen sequences must be interpreted with caution, considering the range of conditions under which natural fires might have occurred. Not all fires are necessarily anthropogenic, nor where they are associated with archaeology do they have to reflect human manipulation of the surrounding environment.
The role of multiple causal processes in woodland disturbances is perhaps best exemplified in the study of the elm decline, widely documented from pollen records across North West Europe (Parker et al 2002). The elm decline is a broadly synchronous event across Britain, occurring between c 6350 and 5280 BP (Parker et al 2002), straddling the late Mesolithic and early Neolithic. The decline of elm has been variously ascribed to disease, a shift to a continental climate, human interference, competition, soil deterioration or a combination of one or more of these processes.
Elm declines are recorded at all sites in and around the region. At a few sites two elm declines are recorded (cf Hirons and Edwards 1986; Tipping 1995; 2010). Radiocarbon dates for elm declines in southern Scotland suggest the first decline occurred between c 6100 and 5450 BP (Tipping 1997; Cayless & Tipping 2002), although a later date of c. 5200 BP at Ravelrig Bog (GUARD Archaeology Ltd 2014) still sits within the date range suggested by Parker et al (2002). However, a much later second elm decline is dated at Yetholm to around 4590 BP (Tipping 2010).
Disease, climate and human activity are generally considered to be amongst the most likely causes of the elm decline (Parker et al 2002), although there is no clear evidence from the region in support of either a single or multiple causal hypothesis. The disease hypothesis relies heavily on the identification of declines in elm pollen closely associated with fossil remains of the elm bark beetle (Scolytus scolytus), thought to function as a vector for the fungus also responsible for Dutch elm disease (Girling & Greig 1985; Girling 1988). Support for the disease hypothesis therefore requires complementary coleopteran analysis. Likewise, support for a climate hypothesis requires independent palaeoclimate data. Moreover, although human activity cannot be excluded as a causal factor, not all sequences with elm declines have related human activity (eg. Tipping 2010; GUARD Archaeology Ltd 2014).
Although well-documented from pollen analyses, a fuller understanding of the comparative role of different causal factors in the elm decline can only be appreciated through an integrated approach to analysis that utilises the full range of applicable techniques (including but not restricted to pollen, plant macrofossils, coleoptera and palaeoclimate data).