Dispersal limitation is often involved when the species composition of a dry abandoned grassland shows a slow response to resumed regular mowing. A seed-addition experiment, using 32 species which do not belong to the local species pool, was performed on Monte San Giorgio (southern Switzerland) to test whether the low recruitment success was due to dispersal limitation or due to unfavourable microsite conditions. In October 1997, 20 species were individually sown in six 3 × 4 m blocks of a 2 × 2 factorial “partial” split-plot design with treatments of abandonment vs. mowing and undisturbed vs. root-removed soil, this last being applied in small naturally-degradable pots. Moreover, 12 species were sown only in the treatments on undisturbed soil. Seedlings of sown and spontaneously germinating seeds were observed on 16 occasions over one 12-month period.
Seeds of 31 out of the 32 species germinated. Twenty-four species showed germination rates higher than 5% and different seasonal germination patterns. Established vegetation, especially the tussocks of Molinia arundinacea, reduced the quality of microsites for germination. Whereas a few species germinated better under the litter of Molinia arundinacea, many more germinated better under the more variable microsite conditions of a mown grassland.
Only a few seedlings of 25 species out of the 31 germinated species survived until October 1998. Seedling survival was negatively affected by litter, unfavourable weather conditions (frost and dry periods followed by heavy rains) and herbivory (slugs and grasshoppers). Tussocks of Molinia arundinacea, however, tended to protect seedlings.
The poor establishment success of “new” species observed in abandoned meadows on Monte San Giorgio after resumed mowing is due to dispersal and microsite limitations.