The consequences of fertiliser addition to semi-natural grasslands are well understood, but much less is known about the consequences of cessation of nitrogen fertiliser regimes, including rates of recovery. This study aimed to investigate whether the effects of nitrogen (N) additions to a mesotrophic grassland were still apparent 15 years after the cessation of N inputs. A long-term experiment at Tadham Moor, UK, received N additions at rates of 0, 25, 50, 100 and 200 kg N ha(-1) yr(-1) between 1986 and 1994. Fifteen years after the cessation of N additions soil chemistry, plant tissue chemistry, plant biomass and Ellenberg N values were assessed. KCl-extractable ammonium-N, total soil N, total organic carbon and microbial biomass N differed between the controls and the higher historic levels of N addition. Plant tissue chemistry showed no significant effects of previous N addition. Above-ground biomass was higher where N had been added, although this response was only weakly significant. The species composition of the vegetation showed effects of the N addition with mean Ellenberg N values significantly higher than the control in most treatments. The effects of long-term N addition can be seen for many years.