1 The sources of propagules for regeneration in an acidic grassland were identified from analysis of differences in colonization between plots subject to surface (0–5 cm) soil disturbance and plots where surface soil had been replaced by ‘seed-free’ soil from deeper soil horizons (30–35 cm), and between plots with and without the removal of rabbit pellets.
2 After 1 year, 10 species had a significantly higher cover on plots where the seed bank had been left intact. These included Agrostis capillaris (the dominant species prior to disturbance), Myosotis arvensis and Veronica arvensis.
3 Five species, including Sagina apetala, Senecio jacobaea and Veronica arvensis, showed significantly higher cover on plots where rabbit pellets were left in situ.
4 From calculations it appeared that rabbit-dispersed seeds accounted for 15% of the developing higher plant cover, other means of dispersal from outside the plot accounted for 40%, and regeneration from the seed bank accounted for 45%.
5 Similar calculations suggested that three higher plant species, Geranium molle, Myosotis arvensis and Senecio jacobaea, appeared to depend most on non-rabbit dispersed seed for colonization of bare ground.
6 High concentrations of Urtica dioica in pellets contrasted with its poor establishment in the experiment. However, the other common species in the pellets, Sagina apetala, Senecio jacobaea and Veronica arvensis, all established in greater numbers on the plots where the pellets were not removed.
7 Seed bank content correlated well with the pattern of regeneration for Agrostis capillaris, Holcus lanatus, Myosotis arvensis and Veronica arvensis. However, removal of the seed bank did not have a significant effect on the regeneration of either of the most common species in the seed bank, Rumex acetosella and Sagina apetala.
8 No species appeared to be reliant on only one mechanism for regeneration from seed in disturbed areas in this community.