Aims We investigated the legacy effects of a previous ley’s legume proportion on the performance of a following grass crop in a rotation. Methods In April 2015, a pure Lolium multiflorum L. crop was sown after the removal of legume containing swards (0-100% legumes), and was harvested four times over the following one-year period (3 times in 2015 and once the following April 2016). Labelled 15N fertilizer (50 kg N ha-1) was applied during the 2nd and 3rd re-growth periods to determine N fluxes. Results Across the one-year period, a significant legume-legacy induced increase in biomass yield of L. multiflorum was observed over the entire range of previous legume proportions when compared against the non-legume ley, the effect being 2.15 and 1.73 t ha-1 (P ≤ 0.001 each) in swards with 50% and 100% previous legume proportion, respectively, or up to +31%. The legume-legacy effect on biomass yield was most pronounced at the 1st harvest (June) and persisted into the 2nd harvest in August (P ≤ 0.05 both, over the entire range of previous legume proportion), though was no longer evident at the 3rd harvest (September). Importantly, the legume-legacy effect returned in the 4th harvest in April (P ≤ 0.05). Examining the source of N contributing to N yield confirmed that more N was derived from the soil at harvest 1 and 2 for previous legume containing leys (P ≤ 0.001) compared to those which contained no legumes, with a significant increase still seen for legume mixtures at harvest 3 (P ≤ 0.01). Conclusions The results demonstrate a sustained soil-transferred performance-enhancing legacy effect on a following crop in a rotation, with previous legume proportions of 50% having a comparable effect compared with that of a previous legume monoculture.