The different parts of the research focus on the species-rich grasslands restoration. The first part concerns the definition of the optimal pedologic requirement of these covers. Namely, the major negative influence of high quantities of exchangeable phosphorus on plant diversity have been highlighted. This influence could be explained by the influence of these quantities on those of available nitrogen. In this case this last element should remain the major element controlling plant diversity. It has nevertheless been shown, by a chemical analysis of these elements in the standing biomass, that phosphorus is related to plant diversity independently to nitrogen. The restoration of such covers often requires a new sowing. The second part of this research concern studies of the germination (tests, germination related to nutrients quantities, light quantity or quality) or the dormancy breaking. At a more practical level, experimentations about the best methods to be applied for such restoration (the sowing date, the optimal seed quantities of the grass/ dicots proportion,...) are also carried on. After implantation, the vegetation has to evolve correctly in a way to give a species-rich cover. Short-term studies focus, in the laboratory, on the inter or intra-specific competition (at root or shoot level) between grasses. Long-term study of this evolution has begun in 1993 on plots established on different soils (rich or poor in nutrient) after new sowing of with standing vegetation. Different management practices are tested on these plots (1 or 2 cuts per year, in June, July or August). The measures concern the botanical evolution and the yields and qualities of the forages. This constitutes the third part of ourresearch : the opportunities to adapt such cover and such forages into a conventional farming system.