1. Alternate husbandry systems containing 1-, 2- and 3-year grazed leys were compared with continuous arable cropping at Jealott's Hill during the period 1945–56. Kale and wheat were used as successive test crops to measure the fertility accruing from the following crop sequences: (a) 3-year ley; (b) wheat, 2-year ley; (c) wheat, barley, 1-year ley; (d) wheat, barley, barley. A study was also made of the effect on crop yields of compost, dung and N fertilizer dressings applied to kale.2. The yields of kale and wheat were improved considerably by 1-, 2- and 3-year leys. Increments were: 5·2 tons fresh kale per acre and 8·7 cwt. wheat grain per acre from 1-year ley, 9·0 tons kale and 13·5 cwt. wheat per acre from 2-year ley and 9·2 tons kale and 15·0 cwt. wheat per acre from 3-year ley. Two- and 3-year leys were markedly superior to 1-year ley in their effects on kale and wheat. The slight benefit apparent in wheat yields from 3-year ley compared with 2-year ley was not significant and so the 2- and 3-year leys may be considered equal in their effect on the succeeding arable crops. In the third arable crop, wheat, the 2-year ley was still effective to the extent of 3·9 cwt. grain per acre but the residual effect of the 1-year ley had largely disappeared. Yields under the all-arable system were considered to have been reduced by the poor physical and nutrient status of the soil, by increased incidence of ‘take-all’ in wheat and by more weed growth.3. The response of kale to dung averaged only 1·1 tons per acre over all four rotations but this was largely obtained in the ley systems and little after all-arable cropping. Compost was generally ineffective on kale with some signs of detrimental effect, especially in the arable system. Wheat following kale benefited equally from dung and compost applied in the previous year; the average response to organics was 1·6 cwt. grain per acre and this was unaffected by rotation. The residual value of dung on subsequent crops amounted to an average of 2·0 cwt. grain per acre per annum but compost had no significant effect.4. The response to an additional 0·4 cwt. fertilizer N per acre to kale was greatest after continuous arable cropping and virtually nil after 3-year ley. Wheat grain yields were improved by 1·0 cwt. per acre on average by the addition of the extra 0·4 cwt. fertilizer N per acre to kale in the previous year.5. It was concluded that when mixed farming is practised the aim should be to have the land under grass for at least as long as it is under arable, e.g. 3 years of ley with 3 years of arable.