Five soils were treated with K at rates of 269, 134, and 0 kg/ha and were diluted 6:4, 3:7, and 1:9 with sand. Potassium was extracted from samples of the soils and soil‐sand mixtures by various methods. The treated soils were then sequentially cropped in a growth chamber to German millet ( Setaria italica stramineofructa ) (one cutting) and alfalfa ( Medicago sativa ) (four cuttings). After cropping, K was again extracted from samples of the soils and mixtures. The soils differed greatly in amounts of K released to the plants and to the various extracting solutions. Amounts of K removed by NH 4 OAc, HNO 3 , HCl, and NaBPh 4 from soils prior to cropping generally correlated better with amounts of K uptake than did amounts of K removed by other extractants. When all soils, K additions, and soil dilutions were combined, energy of exchange (Δ F ) values were poorly correlated with K uptake because of greater than proportional release of K in soil‐sand mixtures. Both the plants and chemical extractants removed more K from soil‐sand mixtures than that calculated on the basis of proportions of the soil in the mixtures. Cropping considerably reduced the amounts of exchangeable and boiling HNO 3 releasable K and caused Δ F values to become more negative.