Soil abiotic and biotic interactions govern important ecosystem processes. However, the mechanisms behind these interactions are complex, and the links between specific environmental factors, microbial community structures, and functions are not well understood. Here, we applied DNA shotgun metagenomic techniques to investigate the effect of inorganic fertilizers N, P, K, and NPK on the bacterial community composition and potential functions in grassland soils in a 54-year experiment. Differences in total and available nutrients were found in the treatment soils; interestingly, Al, As, Mg, and Mn contents were variable in N, P, K, and NPK treatments. Bacterial community compositions shifted and Actinobacteria were overrepresented under the four fertilization treatments compared to the control. Redundancy analysis of the soil parameters and the bacterial community profiles showed that Mg, total N, Cd, and Al were linked to community variation. Using correlation analysis, Acidobacteria, Bacteroidetes, and Verrucomicrobia were linked similarly to soil parameters, and Actinobacteria and Proteobacteria were linked separately to different suites of parameters. Surprisingly, we found no fertilizers effect on microbial functional profiles which supports functional redundancy as a mechanism for stabilization of functions during changes in microbial composition. We suggest that functional profiles are more resistant to environmental changes than community compositions in the grassland ecosystem.

Abstract. Nutrient enrichment of habitats (eutrophication) is considered to be one of the main causes of plant diversity decline worldwide. Several experiments have shown a rapid loss of species in the first years after fertilization started. However, little is known about changes in species richness in the long term. Here, we use a 50-year-old field experiment with a range of fertilization treatments in grasslands that were mown twice each year in the center of The Netherlands. We show that species richness in all plots initially declined but started to recover after ;25 years of continued fertilization. This was also true for the heavily fertilized treatment (NPK). In NPK-fertilized plots, the decline was strongest and associated with a strong divergence of plant trait composition from the control, reflecting a shift to a plant community adapted to nutrient-rich conditions. During the subsequent period of increase in species richness, the trait composition remained stable. These results show that plant species richness can, at least partially, recover after an initial diversity decline caused by fertilization.

No abstract

No abstract

Restoration of formerly nutrient‐poor and species‐rich grasslands generally leads to an increase in species diversity. However, species without a persistent seed bank and with poor dispersal ability often do not re‐establish spontaneously. Here, reintroduction is an option. If existing populations are comparable in their genetic composition, any population will do. This is not the case if populations have local adaptations. Unfortunately, whether populations are adapted locally is not easily determined, in contrast to assessing differentiation using neutral genetic markers. We used AFLP to study genetic diversity of Cirsium dissectum and Succisa pratensis within and among several Junco‐Molinion plant communities in the Netherlands (up to 200 km apart) that were potential source populations, and followed the reintroduction using seeds from these populations. Also, vegetative growth phase characteristics of three populations of C. dissectum were analyzed under controlled conditions. Most of the genetic variation in these cross‐fertilizing species was found within populations. Small but significant genetic differences in band frequencies were found among populations (Fst 0.100 ‐ 0.135). The first generation of reintroduced plants contained less polymorphic bands than the source populations. The genetic differences caused by reintroduction using a limited number of seeds (founder effects) were significant in all except one case (Fst 0.012 ‐ 0.101 between source and corresponding reintroduced population), but the magnitude was smaller than the source population differentiation. In assignment tests, reintroduced populations resembled their source population more than any other population, but all populations contained sizeable proportions of plants that were assigned to most similar plants from other populations, indicating that the populations are only marginally distinct. Calculations show that reintroduction from more than one source population introduces significantly more polymorphic bands into the new population, capitalizing on the existence of band frequency differences among populations.