A combination of on- and off-site palaeo-environmental and archaeological investigations of the upper Allen valley of Dorset conducted in 1998–2000 has begun to reveal a different model of landscape development than those previously put forward. A combination of off-site geoarchaeological and aerial photographic survey and palynological analyses of two relict palaeochannel systems, and sample investigations of four Bronze Age round barrows and a Neolithic enclosure, have been combined with inter-regional summaries of the archaeological and molluscan records to re-examine the prehistoric landscape dynamics in the study area. Preliminary results suggest that woodland development in the earlier Holocene appears to have been more patchy than the presumed model of full climax deciduous woodland. With open areas still present in the Mesolithic, the area witnessed its first exploitation of the chalk downs, thus slowing and altering soil development of the downlands. Consequently, many areas perhaps never developed thick, well structured, clay-enriched soils (or argillic brown earths), but rather thin brown earths. By the later Neolithic these under-developed soils had become thin rendzinas, largely as a consequence of human exploitation. The presence of thinner and less well-developed soils over large areas of downland removes the necessity for envisaging extensive soil erosion and thick aggraded deposits in the valley bottom in later prehistory. The investigations have suggested that, if there were major changes in vegetation and soil complexes, these had already occurred by the Neolithic rather than in the Bronze Age as suggested by previous researchers, and the area has remained relatively stable since.
