Semi-natural grasslands : the effects of cutting frequency on long-term changes of floristic Composition

The objective of the experiment was to study the effects of different frequencies of cutting on long-term changes in the floristic composition of grassland, with an emphasis on the effects of very late cutting. Cutting at 2, 4, 6, 8, 10 and 12-weekly intervals was applied from 1995 to 2006 in a randomized block experiment on semi-natural grassland fertilized with 170 kg N, 40 kg P and 150 kg K ha[^-1]. The floristic composition of treatments was determined in May 1995, 1999, 2002 and 2006. In the final year (2006) the proportion of grasses in the dry matter of harvest was highest (77%) in the treatment with the least frequent cuts, and it was the highest obtained during the whole experiment. Under very frequent cuts, the proportion of grasses ranged only from 45 to 57%. Among legumes, only Trifolium repens L. was represented in high proportions in the dry matter of the harvest (from 19 to 25%) in 2006, but only in the treatments involving high cutting frequencies. The proportion of non-legume forbs in treatments involving low cutting frequency did not exceed other treatments; moreover, their proportion in the dry matter of the harvest in the treatment with the lowest cutting frequency decreased significantly to the lowest value (19%). The increased proportion of grasses in treatments with low cutting frequency indicates that a very late cut does not necessarily mean that forbs increase in the floristic composition, but can help grasses to retain their dominant and stable proportion in the semi-natural sward. Consequently when farmers are being provided with financial incentives with the aim of protecting diversity and the natural environment, it is worth recommending also a very late cutting, which includes natural reseeding to a greater extent. This can help to find a compromise between farmers' expectations of stable yield and quality forage, and the public interest in having grasslands of high biodiversity.