Question: What is the relative significance of seed source, growth form and seeding density in contributing to restoration of calcareous grasslands?Location: Experimental field site on calcareous gravel in the lowlands north of Munich (S Germany).Methods: The research question was investigated using a dominant grass (Festuca rubra) and three subordinate forbs (Buphthalmum salicifolium, Dianthus carthusianorum, Linum perenne). A randomized split-block design consisting of six blocks was established with two growth forms of F. rubra (rhizomatous, tussock) from two sources (cultivars, 'regional' plants from the region of the field experiment) seeded at two densities (800, 1600 seedÁm À2 ) and inter-seeded with forbs from cultivar and regional sources (100 seedÁm À2 ). Each combination of plant source, growth form and seeding density of the dominant species was replicated six times. Individual numbers of F. rubra (Year 1), the subordinate species (Year 1 and Year 2) and the spontaneously invading alien plant Erigeron annuus (Year 2) were counted, and total cover of vascular plants in each subplot and average number of flowers per area were recorded in Year 2. Results: Regional seed generally had higher establishment than the cultivar counterpart in both the dominant and subordinate species. Growth form and seeding density of F. rubra differed little in their effects on establishment of the forbs, but fewer E. annuus invaded subplots with regional F. rubra, and subplots with the tussock growth form. Only in D. carthusianorum did regional rather than cultivar subordinates produce more flowers.Conclusions: Seed source and growth form are the most significant factors for designing seed mixtures for restoration of calcareous grasslands.