The results are reported from an experiment on the effects of cutting date (14 June, 21 July and I September), fertilizer application (none or 80 kg ha−1 N plus 40 kg ha−1 P and K) and grazing treatments (none, autumn or autumn plus spring) on the vegetation of an upland mesotrophic grassland in Upper Teesdale. northern England, UK.

Effects on plant species number and cover are reported for 4 years (1989–93) of treatment. Effects on ‘species -attributes’ are given for the fourth year.

The cessation of grazing combined with the use of fertilizer progressively reduced species number by about 25%. Under traditional management (no fertilizer, cutting date on 21 July, autumn and spring grazing) the species number and cover remained relatively static over the 4 years.

Comparison between treatments in the fourth year showed a reduction in species number under the fertilizer application, cutting date on 1 September and no-grazing treatments. Fertilizer use together with cutting date on 1 September particularly lowered species number and cover.

Analysis of variance was used to assess the effect of treatment on species that occurred frequently in the sward. A cutting date of 1 September favoured Agrostis capillaris. Alopecurus pratensis, Poa trivialis, Phleum pratense and Trisetum flavescens, The absence of grazing favoured Dactylis glomerata and Holcus lanatus. The use of fertilizer particularly favoured A. pratensis and H. lanatus. Ordination methods were used to assess the effect of treatment on the less frequent species. These were primarily associated with the treatment combination that matched ‘traditional’ management. Deviations from this ‘traditional’ regime acted separately, rather than in combination, and favoured different grass species. Traditional management was associated with ruderal, stress-tolerant ruderal and competitive ruderal strategists and with longer seed germination times, heavier seeds, some of which needed scarifying or chilling to break dormancy, and transient seed banks that germinated in the autumn.

The original sward was an Anthoxanthum odora-turn-Geranium sylvatirum grassland, Briza media subcommunity (MG3b). After 4 years, Festuca ovina-Agrostis capillaris-Galium saxatile grassland, Holcus lanatus-Trifolium repens subcom-munity (U4b) and Lolium perenne-Alopecurus pratensis-Festuca pratensis grassland (MG7c) were found in many of the fertilized and late-cutting treatments.