1. A ‘reference plot’ rotation experiment was begun at Rothamsted in 1956 on a field where grass had grown for many years. The soil contained very little HCl-soluble P and K but much N. The soil had been acid, but was limed to pH 6·5 immediately before the experiment started; this pH has been maintained. Very small individual plots (0·00128 acre) were used and nearly all work was done by hand. The results showed that such small plots can be used to do complicated rotation experiments satisfactorily provided that all work is done carefully and movement of soil over plot boundaries is prevented.

2. Responses to combinations of N, P and K fertilizers and farmyard manure (FYM) were measured on wheat, kale, barley, grass-clover ley, and potatoes grown in rotation. Two rates of N were tested. A strip of permanent grass had the same treatments as the arable crops.

3. Winter wheat yielded most dry matter, but wheat yields were increased least, proportionally, by added nutrients. Potatoes gave the smallest yield without manure and responded most to nutrients, yields being increased almost sixfold by NPK fertilizer plus FYM. Kale, barley and grass-clover ley were intermediate in their unmanured yields and in their responses.

4. All crops except the grass-clover ley responded to N; kale responded most and wheat least. Potatoes kale and barley responded roughly equally to P and needed this nutrient more than did wheat and the ley. Potatoes responded most to K, yields of wheat and clover ley were also small without K, but barley and kale needed K less. Interactions between nutrients were large with most crops, particularly with N on kale and K on potatoes, wheat and clover.

5. FYM greatly increased yields of all crops but responses were less when fertilizers were also given. Potatoes responded most to FYM whether or not fertilizers were also used. FYM also caused large increases in yields of grass-clover ley, kale, and permanent grass.