The Late Devensian ice sheet maximum (Dimlington Stadial) covers the period 26–13 kyr BP (24000 – 11000 cal BC) and probably reached its maximum around 20 kyr BP (18000 cal BC), although this would have varied spatially. By 13000 14C years BP (11000 cal BC) deglaciation had probably affected virtually everywhere in Scotland. The disappearance of the main ice sheets from towards the end of the Devensian ice age left northern Britain devoid of a vegetational cover and with many areas having only glacially-derived material overlying the bedrock. From this time, a vegetational recolonization occurred. Although the ensuing interstadial period was relatively warmer, it was also one of overall declining temperatures and temperature oscillations. At the end of the Lateglacial Interstadial there was a severe cooling which lasted c. 1000 radiocarbon years from c. 11 kyr BP (9000 cal BC). During this time local ice sheets were re-established and some valley glaciers reappeared (the Loch Lomond Stadial period).
The accumulation of palynological information for the Late Devensian period has been extensive. There are virtually no regions of Scotland that have not been explored. Even such areas as the Outer Hebrides and the Shetland Islands, where it was once thought that records for this period were lacking. In some areas the sedimentary records are interrupted due to the resurgence of ice during the Loch Lomond Stadial period, revealing a retrogression in the vegetational recolonization process.
The vegetational history of the Late glacial Interstadial period may be considered from several standpoints, although it is not possible to consider all of these in depth here. The number of sites with records from this time allows a general picture of the recolonisation progress to be established. Given the latitudinal extent of Scotland, it might be asked whether, within the overall development, a north-south contrast in vegetation at any one time came into existence or was the recovery in temperature swift enough to nullify this effect? It is also feasible that an east-west contrast came into existence due to the dominance of oceanicity in western areas and greater continentality in the east. Depending upon the concentration of sample sites and the level of pollen counting resolution, it is also possible to show variation in vegetational cover on a local scale arising from differences in topography and especially of aspect. A frequent spectre in such deliberations, however, is the inadequacy of the dating evidence. The bulk of this dating for Late glacial sites involves radiocarbon and much of this was undertaken during the 1970s, a time when dates were relatively expensive to obtain and they consisted of bulk 14C dates covering considerable thicknesses of deposit rather than small samples (especially of plant macrofossils). The advent and wider availability of AMS dating has helped to change this situation, although a reduction in research related to Late glacial palynology means that good dating frameworks are lacking. Published exceptions include Loch an T-Suidhe in Mull (Lowe and Walker 1986) and West Lomond in Fife (Edwards and Whittington 1997c).
The earliest sediments of Devensian Lateglacial limnic deposits are either devoid of or possess very few pollen grains. At this time vegetational colonization was being undertaken by liverworts and mosses and the filaments of Drepanocladus are frequently encountered. Throughout the interstadial period the landscape was open and the flora displays many arctic or alpine associations typical, in part, of bare or unstable soil; at the same time, the sediments of the period are not infrequently relatively organic, reflecting areas of some soil stability. Among the commonest taxa were Poaceae (grasses), Cyperaceae (sedges), Betula nana (dwarf birch) and Salix herbacea (dwarf willow). Although it is sometimes difficult to be certain of separating the pollen of these two latter taxa from their arboreal relatives, the presence of macrofossils, particularly leaves, bears witness to the species involved. Further common components of the vegetation included Empetrum nigrum (crowberry), Juniperus communis (juniper), Artemisia (mugworts), Asteraceae (daisy familily), Filipendula (meadowsweet), Rumex (sorrels), Ranunculaceae (buttercup family) and Caryophyllaceae (pink family). Aquatic vegetation also developed – shallow water bodies would have allowed rapid warming – with seed dispersal undoubtedly assisted by wildfowl and migratory birds. The pollen of Myriophyllum alterniflorum (alternate water milfoil) does in some instances, reach values of 40% total pollen, while Nymphaea (white water lily) and Potamogetonaceae (pondweeds) were also found.
From the above it should not be concluded that the landscapes of the Late Devensian Interstadial in Scotland were not only very open with unstable soils, but also a dreary waste of grasses and sedges with some dwarf shrubs and heathland. This would fail to recognize that certain dominant taxa (especially Cyperaceae, Poaceae and Betula nana) are prolific pollen producers. Present throughout most of this period were many herbaceous taxa, many insect-pollinated, which do not need to produce abundant quantities of pollen. To the plants already mentioned may therefore be added, for example, Brassicaceae (cabbage family), Lactucaeae (dandelion group), Chenopodiaceae (goosefoot family), Sedum (stonecrops), Thalictrum (meadow rues) and even Koenigia islandica (Icelandic purslane) which today is only found at restricted locations in Skye and Mull.
Attention has already been drawn to the fact that the period of the Lateglacial interstadial was affected by considerable climatic oscillations. These had been shown palynologically and sedimentologically in many pollen sequences from Continental Europe (e.g. Fletcher et al. 2009; Ilyashuket al. 2009), but Scottish ones, in general, seemed to be somewhat insensitive to such episodes. It was the availability of the record obtained from cores taken from the Greenland ice sheet which first provided proof of these oscillations, although correlating them precisely with events in the Continental pollen records has been hampered by problems of dating (Lowe et al. 2008; Walker et al. 2009). Whatever the dating precision, eastern Scotland (and perhaps sites elsewhere; Tipping 1991a, b; Edwards et al. 2000a) seems to provide sites which reveal sequences of many of the traditional climatic interludes, e.g. Stormont Loch (Caseldine 1980), West Lomond (Edwards et al. 1997c), Lundin Tower (Whittington et al. 1996), Pickletillem (Whittington et al. 1991) and Wester Cartmore (Edwards and Whittington 2010). These sites have often benefited from higher pollen counts and sampling resolution.
The Greenland ice cores show that around 11k radiocarbon years ago there was a renewed decline in temperatures. There is no doubt that this event was also experienced in Scotland. The very large number of pollen records that have data from this period all show a change from the pollen assemblage of the interstadial period. This heralded the start of the Loch Lomond Stadial period. During this time tundra conditions were experienced, but the constituent taxa of the vegetation involved apparently varied from area to area. Almost everywhere Poaceae and Cyperaceae pollen became the dominant herbaceous taxa along, variously, with Artemisia and Rumex while Salix herbacea and Betula nana frequently increase. Spore-bearing plants such as Huperzia selago (fir clubmoss) and Selaginella selaginoides (lesser clubmoss) appear to have been prevalent. Virtually all such polleniferous deposits are found within a minerogenic sediment matrix deriving from eroded soliflual soils.
Expansions in microscopic charcoal have been found in pollen samples of Loch Lomond Stadial age over many parts of Scotland, as well, to a lesser extent, in sites of Lateglacial Interstadial age (Edwards et al. 2000a). Although human activity might be implicated in this phenomenon, there is probably more support for a natural cause associated with climatic aridity. It would be useful if charcoal analyses were to be carried out in close proximity to any archaeological sites proven to be of Palaeolithic age.