The diaspore bank (seeds of higher plants and spores of ferns and bryophytes) was assessed between 3 and 5 yr after experiments to control Pteridium aquilinum (bracken) and restore appropriate vegetation were initiated at two contrasting locations in the UK. We tested the response of the diaspore bank using univariate and multivariate analysis of variance. The two approaches were complementary and together improved the interpretation of these results. There were considerable differences in the diaspore banks of the two sites and among the experimental locations within sites. Within each experiment there were differences in species composition, with species that were (1) common to both diaspore bank and vegetation, (2) restricted to the diaspore bank and (3) restricted to the vegetation. There is a possibility of increasing the biodiversity of the developing vegetation if some of the species present in the diaspore bank can be germinated. This was especially true for ferns where four species were found in the spore bank which were not present in the vegetation. There were few significant effects of management treatment on the diaspore bank as the experiments had been in progress for only 3 to 5 yr, but a few species had different densities in the different treatments (Betula pubescens, Juncus effusus and some bryophytes). The greatest correlation between vegetation and diaspore bank was found at the top hierarchical level (entire dataset) and this progressively reduced with scale. We interpret this as a landscape/species pool effect: as the scale of the study reduces the correlation between diaspore bank and vegetation also reduces, at least over the time scale of our study. The relevance of these results for restoration ecology is discussed briefly.

Nomenclature: Stace (1997) for higher plants and ferns; Smith (1978) for bryophytes.

Abbreviations: RDA = Redundancy analysis.