Figure 21 shows the distribution of pollen sites known to the author that describe Mesolithic–age plant communities in the region: in the absence at a number of sites of detailed 14C–derived chronologies this estimate is based on the rise of Alnus (alder) pollen but the absence of an Ulmus (elm) decline. They are abundant: this is probably the densest array of such sites in Scotland. Those by Durno with Mercer on Jura and are of historical interest only (Edwards 2000) but many analyses are cutting–edge. Most of these are uniformly of high quality with regard to pollen taxonomy and temporal resolution, though in many there was no recording of microscopic charcoal. Most are 'sensitive' sites, describing plant communities and their changes at spatial scales small enough to identify and be supportive of archaeological evidence for similarly small–scale human activities.
The majority have not been analysed to explore anthropogenic activities, however. Some sites extend analyses of Devensian Lateglacial sediment into the early Holocene (Map 1: above). Others were constructed to understand the dynamics of deglaciation of ice sheets and valley glaciers of Loch Lomond Stadial age. Those along Loch Awe at Ford, Inverliever and Barachander (Tipping 1987, 1988, 1989a) and in Cowal at Na Lona Min (Tipping 1986) do not extend to sediments younger than c. 9000 cal BC. However, those on Rannoch Moor at Clashgour, Corrour, Kingshouse and Rannoch Station (Walker and Lowe 1977, 1979, 1981, 1985, 1987), on Loch Linnhe at Lairigmor (Walker and Lowe 1981) and on Mull at Coire Clachach, Fhuaran, Gribun and Torness (Lowe and Walker 1986) describe plant communities for the bulk of the Mesolithic period. None of these sites have records of microscopic charcoal.
Figure 21: Map of Mesolithic–age pollen records © Richard Tipping: Bealach Froige, Islay (Edwards 2000); Clashgour (Bridge et al. 1990); Coire Clachach, Mull (Walker and Lowe 1985); Coire Seilich (Bridge et al. 1990); Corrour (Walker and Lowe 1979); Coulererach, Islay (Bunting et al. 2000); Cultoon, Islay (Edwards 2000); Dubh Lochan (Stewart et al. 1984); Fhuaran, Mull (Walker and Lowe 1985); Gribun, Mull (Walker and Lowe 19872000); Iona (Scaife and Dimbleby 1990); Loch a' Bhogaidh, Islay (Edwards and Berridge 1994); Loch a' Chrannag, Mull (Sugden 1999); Fyfe et al. 2013); Loch an–t Sagairt, Coll (Wicks and Mithen 2013); Loch an t' Suidhe, Mull (Lowe and Walker 1986); Edwards et al. 2007); Loch Cholla, Colonsay (Andrews et al. 1987); Loch Cill an Aonghais (Peglar in Birks 1980); Loch Gorm, Islay (Bunting et al. 2000); Loch Lomond (Dickson et al. 1976); Lochan Mor, Iona (Bohncke in Barber 1981); Lochan Taynish (Rymer 1974); Mishnish, Mull (Lowe and Walker 1986); Oban (Davies 1997; Macklin et al. 2000); Rannoch Station (Walker and Lowe 1979); Rhoin Farm, Kintyre (Edwards and McIntosh 1988); Sorn Valley, Islay (Edwards 2000); Torbhlaren (Tipping et al. 2011); Torness, Mull (Walker and Lowe 1985).
Holocene vegetation histories at Lochan Taynish (Rymer 1974) and Cill an' Aonghais, both in Kintyre (Peglar 1980), in Loch Lomond (Dickson, Jardine and Price 1976) and on the east bank of Loch Lomond at Dubh Lochan (Stewart, Walker and Dickson 1984) are constructed by applying a uniform sampling interval, not specifically looking for or exploring in detail human activities but noting them when recorded. Their locations may in turn not be optimal for the archaeological record. Peglar's record has not been published: it is available only in a field guide which has almost no interpretation. The Loch Lomond record lacks detail.
A cluster of pollen sites on Islay were analysed to understand Mesolithic vegetation history and human disturbance to plant communities (Mithen 2000) including Coulerach (Bunting et al. 2000), Bealach Froige, Cultoon, Gruinart, Loch Gorm and Sorn Valley (Edwards 2000) and Loch a'Bhogaidh (Edwards and Berridge 1994; Edwards and Mithen 1995). Analyses at Loch Cholla, Colonsay were intended to understand the Oronsay shell middens excavated by Mellars (Andrews et al. 1987b). On Coll, the site of Loch an–t Sagairt is as yet unpublished (Wicks and Mithen 2013). Edwards, Langdon and Sugden (2007) tried to disentangle anthropogenic from climatic effects on the vegetation at Loch an't Suidhe on Mull around c. 6200 cal BC. The site of Rhoin Farm, west of Campbeltown, was analysed by Edwards and Mackintosh (1988; Edwards 1990) to explore pre–elm decline agricultural activity.
Four pollen sites are included in Macklin et al.'s (2000) paper on the Oban region. These are very problematic. The full data–set is found in Fay Davies's (1997) PhD thesis. The pollen sum (the basis for all percentage calculations) is unconventional in that it includes spores and pollen of aquatic plants. This renders the pollen diagrams hard to compare with others and difficult to interpret. Interpretation requires the full data–set but Macklin et al. (2000) present summaries only, of all trees and shrubs, Gramineae, cereal–type pollen and Plantago major, together with concentrations of microscopic charcoal (their Figure 3). Grouping all trees and shrubs together is too simplistic an approach to describing woodland loss. Choosing P. major to describe pastoral activities is odd: P. lanceolata is the accepted indicator. These choices make these pollen records far less than the corner–stone of future work they should be. No details of how the chronologies for the pollen diagrams were constructed is given by Macklin et al. (2000): construction appears not to have been straightforward (Davies 1997). The discussion by Macklin et al. (2000) is synthetic from the start, lacks detail as a result and is too concise.
The difficulties of defining an unambiguous hunter–gatherer signal in pollen records are many, not least because new data on the intensity of abrupt climate change introduce mechanisms by which vegetation can be naturally disturbed. Tipping (1996) argued, for instance, that relative drought increased the likelihood of wildfires in northern Scotland, contested by Bishop, Church and Rowley–Conwy (2015) though true at the global scale (Carcaillet et al. 2002). Random events like lightning–strike fires and tree–felling in storms (eg Tipping (2008) at Clashgour on Rannoch Moor) are feasible sources of woodland change. Edwards et al. (2007) elegantly evaluated human–climate agencies at Loch an't Suidhe on Mull across the major early Holocene abrupt climate change event at c. 6200 cal BC. Losses of deciduous woodland around 6300 cal BC are accompanied by increases in microscopic charcoal, which were argued to have been anthropogenic in origin because climatic deterioration was recorded from chironomid analyses after, not during woodland loss. Errors in 14C dating may explain why the climatic deterioration appears not to be that at c. 6200 cal BC (though see Edwards et al. 2007 for other interpretations) but do not detract from their more general conclusions.
On Islay, Mesolithic anthropogenic activity has been suggested from closely–spaced pollen analyses seeming to describe short–lived vegetation disturbance, accompanied by burning (Edwards 2000). Events are diachronous across the island, which Edwards (2000) uses to dismiss climatic causation. Disturbance was particularly widespread in the 6th millennium BC. The astonishing find of burnt hazel nuts – the product of perhaps some 15000 trees – at Staosnaig on Colonsay (Mithen and Finlayson 1991) excited much interest, with the suggestion that the pollen record might be able to detect this event (Edwards 2000) but the key factor for the human community on Colonsay would have been the sustainability of the resource through management. We cannot currently demonstrate woodland management from pollen data with conviction (Waller, Grant and Bunting 2012), but tree–ring data from hazel stems can suggest coppicing cycles. Such evidence needs to be sought. Andrews (in McCullagh 1989) also identified possible anthropogenic woodland disturbance on Islay, around the Sorn Valley, around 5900 cal BC. In the Oban area significant woodland loss, associated with burning, is also recorded for short time intervals throughout the Mesolithic.
Anthropogenic signals in the Mesolithic were not recorded at any of their fourteen sites by Lowe and Walker (1986: Walker and Lowe 1977, 1979, 1981, 1985, 1987). This may in part be because microscopic charcoal was not recorded by them, or because the time interval between pollen samples was too coarse, but this is an intriguing if negative result. Andrews et al. (1987b) also did not report Mesolithic–age anthropogenic impacts on Colonsay though the sampling interval is comparatively coarse and microscopic charcoal not recorded. Tree pollen rarely exceeds 20% total pollen and spores in the diagram, though, and the landscape must have been quite open, maintained by a high degree of exposure to winds. Walker and Lowe (1985) also infer a high degree of exposure to wind from the low representation of trees at sites on Mull, and around Iona Loch the dominant vegetation in the Mesolithic is a grass–heath with few trees (Scaife and Dimbleby 1990). It is difficult to see in these open settings why people would have needed to disturb woodland: models for anthropogenic interference derived from the closed deciduous woodlands of eastern England (Simmons 1996) need not have applied to the Scottish west coast. Other landscapes appear more wooded, but nevertheless the precise 'ecology' of these disturbances in western Scotland, and if purposeful what they achieved, needs to be explored much more fully. This is now a pressing need because if anthropogenic signals in Mesolithic–age pollen records along the Scottish west coast can be demonstrated, the technique is one of the few that could test Wicks and Mithen's (2013) suggestion of population collapse in and after the 6200 cal BC abrupt climate change.
The possibility that Mesolithic people might have precociously planted cereal crops was for a time an intriguing idea (Edwards 1989; Edwards and Hirons 1984). Macklin et al. (2000) report cereal–type pollen as early as c. 9250 cal BC at Gallanach Beg but interpret them as deriving from wild grasses. The site at Rhoin Farm was analysed (Edwards and Mackintosh 1988) to test this by scanning for cereal type pollen for much longer than normal in pollen counts. Cereal type pollen in pre–elm decline contexts was found, but was not shown unambiguously to be from cereal grasses: almost all the pollen grains are of Hordeum type (barley type), which includes many wild grasses known to produce cereal–size pollen it habitats such as the sand dunes that surround the site today and probably also in prehistory.