Our study investigates the negative impact of nitrogen (N) deposition on species richness in acidic grasslands, based on a temporal comparison of vegetation data spanning a period of almost 70 years. We compiled a large data base of plots assigned to the Violion caninae grassland type, composed of managed, but unfertilized semi-natural grasslands on nutrient-poor, acidic soils. In total 1114 plots, mainly from Great Britain, the Netherlands and Germany, were compiled, dating back to 1939. Environmental site information included geographical and soil (mean Ellenberg values) variables as well as estimates of cumulative N and sulphur (S) deposition since 1939. Statistical analyses were carried out separately for the data subsets from the three regions. In all regions, the vegetation differentiation was mainly related to soil acidity and nutrient availability, as well as to the year of sampling and the cumulative amounts of N and S deposition. Plot-species richness of vascular plants and bryophytes (analysed for Great Britain only) decreased with time and analyses suggest these are affected by various factors, notably soil pH, but also latitude and cumulative N deposition. The latter explained more of the variation in species number than the year of sampling and cumulative S deposition, which supports the interpretation that the decline in species richness is mainly caused by increasing N availability and less by altered management and soil acidification. For Great Britain and Germany, cumulative N deposition showed a strong negative relationship with several biodiversity measures, especially the proportion of dicots, whereas it was positively related to the proportion of grass species. In general, our results give temporal evidence for the negative effect of N deposition on species richness in semi-natural vegetation.

Summary 1In a metapopulation context, the distribution of a species in a patchy landscape is interpreted as the equilibrium outcome of extinction and colonization processes. Populations are thus more likely to occur in larger, better connected habitat fragments. 2 To test whether metapopulation models explain distribution patterns of plant species with different life histories we investigated the incidence of herbaceous species in deciduous forests in four provinces of south Sweden. 3 These data were correlated with habitat quality (soil pH, nitrogen mineralization and organic matter) and habitat configuration (patch area and distance to the nearest deciduous forest patch). We also examined whether habitat configuration affected the distribution of species with different life history attributes. 4 All ground layer plant species at 81 sites were recorded, and distribution patterns of 57 species were tested against seed mass, seed number, presence of a seed bank, plant height, life span, mode of pollination, dispersal mode and habitat preference. 5 Habitat quality, especially pH, was more important for the incidence of species than habitat configuration with patch area and isolation significantly affecting only 11 and four species, respectively. Species favoured by larger area were also disadvantaged by greater isolation. 6 The importance of habitat configuration to a species varied with life history. Species that were more negatively affected by patch isolation tended to be habitat specialists and clonal perennials and to produce fewer seeds. Animal-dispersed species were more negatively affected by small stand size. 7 Habitat configuration may be less important for vascular plant distributions than habitat quality or the effects of land use history.

45Evidence from an international survey in the Atlantic biogeographic region of Europe 46indicates that chronic nitrogen deposition is reducing plant species richness in acid 47 grasslands. Across the deposition gradient in this region (2 to 44 kg N ha -1 yr -1 ) species 48 richness showed a curvilinear response, with greatest reductions in species richness when 49 deposition increased from low levels. This has important implications for conservation 50 policies, suggesting that to protect the most sensitive grasslands resources should be 51 focussed where deposition is currently low. Soil pH is also an important driver of species 52 richness indicating that the acidifying effect of nitrogen deposition may be contributing to 53 species richness reductions. The results of this survey suggest that the impacts of nitrogen 54 deposition can be observed over a large geographical range.

The concept of species pool is reviewed. It is suggested to maintain the terms regional pool and local pool but replace actual pool by community pool. The regional and local pool are considered as selections from the regional and local flora based on ecological similarity. It is also suggested to include in the community pool a selection of species present only as diaspores in the diaspore bank (including diaspores from the seed rain), the selection being based on the same ecological criteria.Four approaches to determine the species pool are discussed: ecological, functional and phytosociological similarity, and an experimental approach. The phytosociological approach appears to be promising. The species pool is elaborated as a fuzzy set in the sense that each species of a community or a local or regional flora is a member of any community, local or regional species pool with different degrees of membership. This membership is defined as a probability of a species to become part of the community pool of a target community which is a function of the ecological (environmental/functional/phytosociological) similarity of the species with the target community; the shortness of the distance of its nearest populations, the frequency/abundance, the dispersal capacity, the actual presence of dispersal mechanisms, the germinability of newly arrived diaspores, and the longevity of seeds (viability) in the diaspore bank.The information on species pools is needed for designing experiments where the number of species in a community is to be manipulated, for instance in restoration management.