A field experiment testing the effects of potassium and magnesium sulphates applied from 1960 to 1967 and sodium chloride applied in 1966 and 1967 on a range of crops (Bolton & Penny, 1968) was sown with ryegrass in 1967. Only basal N and P was given to all plots until 1971 when basal K was also applied. The experiment terminated in autumn 1974.

Grass yields were increased by K residues for 3·5 years but overall yields became small due to K deficiency. After giving basal K, residual K no longer increased yields but did slightly increase % K in the grass until the end of the experiment. Some residual K from the 1960–7 dressings was found in the sandy loam soil after the final harvest.

Magnesium fertilizer residues slightly increased grass yields in the 1967–70 period but not later. However, % Mg in the grass dry matter was increased by the residues for the whole 7·5 year period. Exchangeable Mg was also increased in soil sampled after the final harvest, especially in the 23–46 cm subsoil.

Sodium fertilizer residues increased grass yields until the basal K was applied, especially on plots without added K. Significant increases in % Na in the grass and Na uptakes were detectable in all years, i.e. up to 7·5 years after the last Na application. This finding is contrary to most accepted views on the longevity of Na fertilizer effects and could be important in areas where sodium deficiencies in cattle and sheep occur.

Increases in % Cl in the grass from the sodium chloride were only detectable in the first two cuts in the year following the final application.