Abstract

A combination of on- and off-site paleoenvironmental and archaeological investigations of the upper Allen valley of Dorset, conducted from 1998-2002, has begun to indicate a different model of prehistoric landscape development to those previously put forward for this part of the southern English chalk downlands. Woodland growth in the earlier Holocene appears to have been slower and patchier than the presumed model of full climax deciduous woodland rapidly attained in a warming environment. With open areas still strongly present in the Mesolithic, the area witnessed its first exploitation, thus slowing and altering soil development. Consequently, many areas perhaps never developed thick, well-structured, brown forest earths, but more probably thin brown earths. By the later Neolithic period, these soils had become thin rendzinas, largely as a consequence of human exploitation and the predominance of pastoral land use. The early presence of thinner and less well-developed soils over large areas of downland removes the necessity for envisaging extensive soil erosion and the accumulation of thick colluvial and alluvial deposits in the dry valleys and valley floor as often postulated. If there were major changes in the vegetation and soil complexes in this area of chalk downland, these had already occurred by the Neolithic rather than the Bronze Age as often suggested, and the area has remained relatively stable ever since. This has major implications for models of prehistoric land use in the southern chalkland region, such as a much greater degree of stability in prehistoric and historic times, variability within sub-regions, and differences between different parts of the chalk downlands than had previously been envisaged.