A field experiment was carried out in an arable organic cropping system and included a sequence with sole cropped fababean (Vicia faba L.), lupin (Lupinus angustifolius L.), pea (Pisum sativum L.), oat (Avena sativa L.) and pea–oat intercropping with or without an undersown perennial ryegrass (Lolium perenne L.) – white clover (Trifolium repens L.) catch crop followed by a first crop of spring wheat (Triticum aestivum L.) and second crop of winter triticale (Triticale hexaploide L.). The rotation sequence was repeated twice. Natural 15N abundance techniques were used to determine grain legume N2 fixation and 15N labeling technique to determine the fate of pea and oat residue N recovery in the subsequent crop. The subsequent spring wheat and winter triticale crop yields were not significantly affected by the previous main crop, but a significant effect of catch crop undersowing was observed. A higher soil mineral N content in the soil profile without undersown grass-clover increased the spring wheat yield. This effect was circumvented in the subsequent winter triticale, where yields in the treatments with catch crops undersown were significantly greater. The grass-clover catch crop after grain legumes had a higher grass proportion before incorporation as compared to grass-clover after oat, which had the greatest clover proportion. The dynamic response of interspecific interactions in the catch crop to the soil mineral N levels is moderating the preceding effect of main crops in the subsequent cereal – and sometimes to a higher degree than the main crop effect. For research involving rotation principles it is recommended to evaluate cumulative effects over several years and not only single seasons.