An experiment at Rothamsted during 1958–67 measured effects on yield, on K uptake and on soil K of applying all combinations of 38, 75 and 113 kg N and 0, 31 and 62 kg K/ha per cut to grass leys, which were cut and removed. Soil K was depleted most where most N and least K were given. Annual applications of 0, 33 and 66 kg P/ha were also tested; soil P was not depleted. The grass was then ploughed.
In 1968, residual effects were measured by spring wheat. In 1969 and in 1970 104 kg/ha of fresh K was applied on half of each plot; potatoes (1969) and spring wheat (1970) valued residual and fresh effects of K.
In 1971 potatoes tested 0, 104 and 208 kg/ha of fresh K, cumulatively with the three amounts given to the grass and also extra K (104 kg/ha) on half-plots, cumulatively with that given in 1969 and 1970. In 1972 winter wheat, and in 1974 and 1975 spring barley, measured residues of all treatments previously applied (the site was fallowed in 1973).
Finally, in 1976, potatoes tested 0, 156 and 312 kg/ha of fresh K on whole plots, cumulatively with the previous dressings of K, and also 156 kg/ha of extra K on half-plots, again cumulatively. All these test crops were given basal N.
Yields and K contents of wheat at ear emergence and yields of wheat grain were largest after grass given 38 kg N and 62 kg K/ha per cut, because here soil K depletion was least. Wheat grain yields benefited consistently from fresh K. K content of the wheat at ear emergence was a good indicator of the need for K, but K content of grain was not, because it was unaltered by K fertilizer. Barley was a poor test crop for K, because yields of grain were little affected by previous treatments.
Percentage K in potato leaves (in July in 1969 and 1971, in August in 1976) and yield of tubers were well correlated. Largest yields in 1969, 1971 and 1976 came where the leaves contained 3·43, 3·76 and 2·82% K, respectively, i.e. from soil containing most exchangeable K, plus most fresh K. There was no indication that maximum yields had been obtained, so the largest amounts (kg/ha) of fresh K tested (104 in 1969, 312 in 1971 and 468 in 1976) were insufficient to counteract depletion of soil K by the grass. Because the grass did not deplete soil P, the test crops benefited only little from either residual or fresh P.