In areas such as parts of Britain where many closely spaced sites have been investigated, between-site diversities prevent any simple correlation between vegetation response and climatic variables. These diversities reveal the influence of other factors in modulating this response. Analysis of profiles of allochthonous lake sediments has provided evidence for the importance of soil factors. The otherwise inexplicable spatial .variation in the response of trees, in Britain and neighbouring parts of the European mainland, to the climatic warming (inferred from faunal evidence) at the most recent glacial termination, can be explained by the postulated spatial differentiation of raw skeletal soils. Differences in particle-size, composition and drainage, consistent with the spatial differentiation of pre-arboreal vegetation, would affect the water-retaining capacity of immature late-glacial soils and hence the reproductive success of tree birches. The lag in response of these to climatic amelioration appears to have varied from 500 to 1 500 yr, and it seems likely that a similar lag may have characterised the early warming period of each interglacial. The conclusion must be that in this situation the degree of success of trees measured by pollen values is not a reliable indicator of palaeotemperatures. Secondly, it appears from evidence in northern Scotland that at an ecotone between forest types, the relation between vegetation and climate is likely to be obscured except in regions of uniform soils. In areas of differentiated bedrock, the vegetation pattern is likely to have been influenced most by the effects of soil maturation processes in developing a soil mosaic.