Does land use change affect the interactions between two dry grassland species?

Over the last 20 years, a change in traditional land use practices has taken place in central Germany. Formerly species-rich dry grassland communities have been converted into communities with greatly reduced diversity in many places. Whereas grass species have expanded, several forbs have declined in abundance. For the present study, plant-plant interactions were assessed between the expanding grass Festuca rupicola and the forb Dianthus carthusianorum - two typical, companion grassland species - to ascertain any associated effects of land use change. A competition experiment was set up with replacement design, in which pots were placed in the open field and monitored over two consecutive years. In order to assess the performance of both species in monoculture (D9, F9) and mixture (D3F6, D6F3) the parameters biomass, length of flowering stems, number of flowering stems and number of flowers per individual, or length of panicles, were analyzed. Positive and negative interactions were evaluated using the "Relative Neighbour Effect" index (RNE). To simulate land use by grazing or mowing, individuals in half of the pots were clipped in the first year after sowing. At this time, performance of D. carthusianorum - in terms of length of flowering stems, number of flowering stems and number of flowers - was significantly affected: whereas the first parameter increased, both of the other parameters decreased with increasing presence of the grass species. However, RNE indicated facilitation with a low number of F. rupicola individuals, and competition with a high number of F. rupicola individuals. In contrast, F. rupicola did not show any significant differences in performance with increasing presence of D. carthusianorum, and the RNE switched from neutral interaction to facilitation. In the second year after sowing, the performance of D. carthusianorum in the various treatments did not differ, except for the length of the flowering stems. The simulation of land use practices did not affect the performance of D. carthusianorum, but F. rupicola showed significant differences in biomass and in length of the panicles, as was the case also between monocultures and mixtures. The performance of the grass species was clearly improved in the unclipped treatments. The RNE reflected competition against D. carthusianorum which became reduced in unclipped treatments, while F. rupicola became facilitated. There were no interactions between the different treatments (monocultures and mixtures) and the simulated land use change (clipped and unclipped). Our data suggest that although traditional land use practices, i.e. grazing and mowing, do not affect the development of D. carthusianorum, the abandonment of these practices has improved F. rupicola performance, which has led to the increasingly invasive character of this grass species over the last 20 years. (C) 2011 Published by Elsevier GmbH.