Abstract Question: How do various plant functional groups and types are related to regeneration of limestone grasslands and to temporal changes in cover of trees and shrubs? Which mechanisms are responsible for changes in differently treated grasslands? Location: Southern Poland, 50° 01′N, 19° 50′E. Methods: A 12-year restoration experiment after the clearing of a 35-yr-old secondary pine wood developed on limestone grassland, and a parallel observation of succession in neighbouring abandoned grassland were carried out. Changes in cover of species with different functional traits and established strategies in relation to time and cover of shrub layer were analysed using multiple linear regression. Results: Over 12 years the composition and cover of species changed in the studied grasslands, but significant differences still occurred between the old grassland and the grassland restored in former wood gaps and that developed in former closed wood. Despite the very close proximity of the old grassland, the composition of restored grasslands was still much determined by initial conditions. In these grasslands, the cover of Festuco-Brometea species was significantly lower than in the old grassland. In all sites, the cover of species with large leaves and with leaves distributed regularly along the stem, and the cover of CS strategists increased. With increasing shrub cover in restored grasslands, the cover of annuals and biennials, tall species, and R and CS strategists decreased. Conclusions: Limited availability of seeds seems the principal reason for the weak regeneration of xerothermic limestone grasslands. Trees and shrubs inhibit the developed of both early successional species and perennials occurring in established communities. They significantly hinder regeneration of grasslands and should be cut every five or six years. Nomenclature: Tutin et al. (1964–1980).