The inherent potential of soils to deliver K from the mineral soil material becomes increasingly important in the development of sustainable agricultural systems. However, soil mineralogy is seldom given any attention in advisory systems. Despite the long history of K application to agricultural land in many areas, there are numerous reports of agricultural systems that run with an annual negative K field balance. This implies that these production/farming systems depend on K pools intrinsic to the soil. The reports include a wide range of farming systems, including both intensive and extensive types— arable cropping, dairy, grazing, conventional and organic— indicating that this is a universally occurring phenomenon. Apart from added mineral fertilisers and manures, the only sources of K in soils are the K bearing phyllosilicates and feldspars. There is, however, very little direct information on the role of different minerals as sources of K to crops. There is also a particle size aspect, where most K typically is thought to come from the finer fractions, especially the clay fraction (< 2 μm). However there are studies that suggest that the coarser fractions also may play a role as a source of K for crops. Different kinds of tools are used world-wide in routine analysis of agricultural soil to estimate pools of K that could constitute a long-term reserve of K for plants. These pools are often referred to as non-exchangeable K, reserve K, or slowly available K. The analysis often involves an acid extraction, for example HCl and HNO3, but physical analysis like determination of clay content is used as well. One often used method is extraction by 2 hours boiling in 2 M HCl. The aims of the present study are to quantify the effects on the mineralogical and chemical composition of the soil during this extraction. It also aims at estimating the relative contribution of different particle size fractions to HCl extractable K. Soils from three different soil suborders (Oxyaquic Hapludoll, Humic Dystrocryept and Typic Haplaquept) (Soil Taxonomy) and from two different climatic regions (northern temperate and southern boreal) were colleted for the study. Bulk soil (< 2000 μm) from six soil depths (between 0 and 110 cm depth) and, in addition, five particle size fractions from one of the surface soil layers were analysed. Aliquots of the soil samples were treated with 2 M HCl, which involved boiling for 2 hours in the acid. The extracts were analysed for K by ICP, whereas the solids were recovered for post treatment analysis. The latter involved assessment of the total weight loss of solids due to the treatment, and Differential X-Ray Diffraction (DXRD) to assess the loss of individual mineral phases. Preparation of samples for mineralogical analysis involved spray drying and addition of corundum as internal standard. After manual identification of mineral phases, a full pattern fitting procedure was used to quantify the different phases and DXRD patterns were calculated. The weight loss of the treated bulk soil (< 2000 μm) samples was positively correlated with the clay content in the different soils and horizons. This indicates that most of the extracted K originates from dissolution of K bearing phyllosilicates, including both di and trioctahedral varieties. Feldspars do not appear to be affected in any obvious way. Among different particle size fractions, the contribution of K was largest from the fine fractions 2-20 and < 2 μm. However, in the sandy soil, the coarse fractions (20-60, 60-200, 200-2000 μm) together accounted for 45 % of the total HCl extractable K, indicating that minerals in the sand and coarse silt fractions may be an important K reserve in soils with low clay content.