Acid treated sediment samples collected from Pickens Hole during excavations in 1965 are predominantly silt sized but are poorly sorted. The sample from Unit 6 is significantly coarser than those from the other units. The latter are similar in grain size distribution to non-acid treated samples from the base of the Pleistocene sediments at Brean Down. The grain size data suggest that these sediments are either directly derived from loess, or more likely derived by colluviation from pre-existing soils developed on a loessic parent material. The significant clay content may be derived by weathering of unstable minerals such as feldspar and chlorite and supports the suggestion for a period of pedogenesis prior to redeposition in the cave. This loessial origin is supported by the morphology of quartz in the silt size fraction which differs significantly from that of acid insoluble residue from the Carboniferous Limestone, but is comparable to that of quartz from Mendip soils which are thought to be aeolian in origin. More generally heavy mineral data for Mendip soils have been suggested to indicate derivation from the east, although a more local origin for the high chlorite and mica may also be considered, either locally from the Mercia Mudstone or from the braid plain of the River Severn to the west. A westerly source is also indicated by heavy mineral data from coversands in the area and also more widely in the UK. Soil parent materials mapped by Findlay (1995) in the Mendip area are coarser to the west than the east. The coarser materials are confined to lowlands and slope base situations, whereas extensive and thick fine deposits have accumulated on the Mendip plateau, especially to the east. This evidence also strongly suggests a derivation of aeolian material from the west and confirms the results from heavy mineral data. This suggestion is however contrary to the generally accepted model of Catt (1978) for an easterly North Sea source.