In 1993, experiments on the restoration of calcareous grasslands on ex-arable fields were started in order to provide new habitats for species of a small nature reserve with ancient grasslands north of Munich (Germany). The effects of diaspore transfer by the application of seed-containing hay on vegetation establishment were studied on restoration fields with and without topsoil removal for 5 years. The aim of the study was to assess plant diversity for the evaluation of restoration success by different methods including determination of species with viable seeds in the hay by germination tests, phenological investigations on hay-transfer source sites at the time of harvest, and vegetation analyses on the restoration sites. Total seed content of the hay and the number and composition of plant species with viable seeds were affected by the time of harvesting and differed between a site which had been used as arable field until 1959 and ancient grassland sites. Nevertheless, the number of established hay-transfer species showed only few differences between restoration fields. The proportion of species transferred to restoration fields in relation to the number of species with viable seeds in the hay was between 69 and 89%. Five years after the hay transfer, the proportion of the established species was still between 58 and 76%. Up to now, topsoil removal had no significant effect on the number of established hay-transfer species. After triple hay application the absolute number of transferred grassland species was higher than on sites with single hay application, but restoration efficiency was lower because many of the species with viable seeds in the hay did not establish. In general, our results showed that the transfer of autochthonous hay is a successful method to overcome dispersal limitation in restoration projects.