1. Twelve experiments were carried out in Bedfordshire and Hertfordshire in 1955–57 to compare combine-drilling and broadcasting muriate of potash for barley. Ordinary mineral soils were used and only one contained more than 25% of calcium carbonate.

2. On average of all the experiments in each year 0·25 cwt. K2O/acre drilled with the seed gave higher yields than 0·5 cwt. K2O/acre broadcast and worked into the seedbed. The average extra yields of barley from combine-drilling, as compared with broadcasting, were 1·2 cwt./acre when 0·25 cwt. K2O/acre was applied and 2·0 cwt./acre of grain when twice as much K2O was given. In twenty-one of the total of twenty-four comparisons drilling gave higher yields than broadcasting and in onethird of these comparisons the effects were significant. Combine-drilling gave higher average yields of straw than broadcasting.

3. The average level of response was not great, the low and high rates of combine-drilled muriate of potash increasing yields by 13 and 16%, respectively. Only a small proportion of the applied potassium was taken up by the crops; apparent recoveries by grain plus straw averaged 11% of the light-drilled dressing and 8% of the heavy dressing.

4. The level of response to potassium varied from year to year, but was much greater on average in the 1955 experiments than in 1956 and 1957. There was no close relationship between the degree of response and values for dilute hydrochloric-acidsoluble potassium or exchangeable potassium in the soils used.