The productivity of permanent swards of mixed species composition was compared with that of sown Lolium perenne, cv. Melle at five fertilizer‐N rates (0, 150, 300, 450 and 900 kg N ha−1) and with L. perenne/Trifolium repens, cv. Grasslands Huia at 0 kg N ha−1. The investigation was conducted under two cutting frequencies at sixteen sites in England and Wales, representing a range of grassland environments. Annual total herbage dry matter (DM) production from both permanent and reseeded swards increased with successive increments of fertilizer‐N up to 450 kg N ha−1. Herbage DM production from reseeded swards in the first year after sowing was signficantly higher than from the permanent swards, at all fertilizer‐N treatments. In subsequent years the production advantage of the L. perenne reseeds was maintained only at the higher N rates, though sown L. perenne/T. repens was the most productive sward type at 0 N. Average differences in modified acid‐detergent fibre suggested small advantages in herbage quality to the reseeded swards. It is concluded that, while reseeded swards are more productive in the year after sowing, many permanent swards are capable of high levels of production and that reseeding to a L. perenne sward cannot always be justified, particularly for grassland receiving low or moderate inputs of fertilizer‐N.

Research on the effects of fertilizer N on grassland has mainly been conducted on sown ryegrass swards, yet about half the grassland in England and Wales is over 20 years old; it usually contains relatively small proportions of sown species and is frequently associated with low-input systems.A small-plot cutting trial was established in 1983 by GRI and ADAS at 16 sites in England and Wales on existing permanent pasture swards. Initial swards contained less than 30% perennial ryegrass (Lolium perenne) and had previously been managed with a relatively low level of inputs. The objective ofthe trial was to determine the effect of fertilizer N rate and cutting frequency on the yield, quality and botanical composition of permanent pasture compared with identically managed sown swards of perennial ryegrass on the same sites. The treatments comprise five rates of fertilizer N (0,150,300,450 and 900 kg ha"' N) cut at intervals of four weeks at all the five N rates on all sites; plus either (1) an 8-week cutting interval for all N rates (at eight ofthe 16 sites) or (2) an 8-week cutting interval for the 300 kg ha~' N rate only (at the other eight sites). A reseed of perennial ryegrass with white clover (Trifolium repens) is also included at all sites in the comparison at the zero N level.Sites were chosen to represent a range of elevation, rainfall and soil types. Initial deficiencies in soil pH, P and K were corrected at the outset and are being maintained at levels to