The influence of catch crops on the accumulation of mineral nitrogen (ammonium and nitrate N) in soil during autumn and winter was studied in 22 held experiments in South and Central Sweden during 1989-92. Five treatments were compared: spring barley sown without a catch crop and spring barley with red clover, white clover, perennial rye grass and a mixture of red clover and perennial rye grass undersown as catch crops. In all these treatments, ploughing was carried out either in early autumn (average date: 6 October), in late autumn (average date: 26 November) or in early spring (average date: 25 March). In seven of the 22 experiments the effects of catch crops on the subsequent crop, spring barley, were also studied. The catch crops reduced the grain yield of the main crop, spring barley, in the first year by 1-3%. At yellow ripeness of this main crop, 41 kg ha(-1) of mineral N remained within the 0-90 cm soil layer as an average of all five treatments. From this stage, mineral N increased in bare soil and under the clover catch crops until late November, whereas the amounts remained low or even decreased under rye grass and under the red clover-rye grass mixture. After early ploughing of soil without a catch crop and with the red clover-rye grass mixture, the accumulation of mineral N increased during autumn. After ploughing in early or late autumn, mineral N levels increased until sampling in early spring, especially following red and white clover. These accumulations of mineralized N indicate larger risks of N leaching, particularly after the clover catch crops. In treatments left unploughed until early spring, the average amounts of mineral nitrogen remained largely unaltered during the winter, indicating a smaller risk of N leaching than after ploughing in autumn. Here mineral N levels were lowest under rye grass and under the red clover-rye grass mixture, not only in late autumn but also in early spring. The clover catch crops and the red clover-rye grass mixture increased the nitrogen supply to the barley in the second year by 30-40 and ca. 15 kg N ha(-1), respectively, compared with the treatment without a catch crop. Pure rye grass had only a small or no additional N effect. The nitrogen effects were mainly due to increased net N mineralization during the growing season of the second year. After autumn ploughing, the clover and clover-grass catch crops increased grain yields of the subsequent barley crop by 10-16% and the pure rye grass catch crop by 9%. Ploughing or, on heavier soils, shallow tillage, in spring caused lower yields, espe cially after rye grass and after the red clover-rye grass mixture, than ploughing in late autumn. This was obviously partly due to reduced supply of plant-available nitrogen in the spring in treatments where the catch crops remained over the winter.