The potential effect of two sowing dates of ryegrass, Lolium perenne L. var. Elka as an interplant in winter leek, Allium porrum L., on weed control, soil nitrogen allocation and crop yield was studied at two sites in Switzerland. In order to reduce potential competition with the ryegrass, row application of the herbicide methazol (as additional treatment) and increased fertiliser doses in spring (as split-plot treatment) were used, and compared with a weed-free control. Ryegrass was sown 4 weeks (early) and 6 weeks (late) after planting of the leek, allowing for one and two mechanical weed control treatments, respectively, prior to sowing. Under the experimental conditions, intersowing ryegrass 6 weeks after planting resulted in crop yield similar to that in the control plots. Crop quality variates were increased and the overall production system was environmentally more advantageous. The combination of two mechanical weed control treatments followed by ryegrass intersowing suppressed weeds sufficiently without herbicide applications. Approximately 20 kg N/ha could be retained from being washed out over the winter, and c. 50 kg N/ha was stored in the interplants up to harvest. The interplants were incorporated into the soil after the harvest to serve as N-source for the subsequent crop. Surplus fertiliser doses in spring increased N-allocation to leek in the control plots. However, when ryegrass was present, it mainly derived a benefit, resulting in only slightly increased N-allocation to leek, or even reduced allocation when overall N-levels in the soil were low. Depending on soil conditions at harvest, full mechanical harvest could possibly be impeded when ryegrass is present. Early ryegrass sowing resulted in severe yield reductions. Row application of the herbicide methazol reduced crop yield in both ryegrass treatments, most possibly due to a direct negative impact of the high herbicide doses on leek growth. Copyright (C) 1996 Elsevier Science Ltd