Published online by Cambridge University Press: 01 June 2009
Recently, forest seed banks were proven to not only reflect former (decades-old) but also ancient (centuries-old) land use. Yet, as land-use intensity determines the magnitude of seed-bank changes in recent forests, this study aims to identify whether an ancient land-use gradient would also be reflected in the seed bank. On a forested 1600-year-old archaeological site, five different land-use intensities were mapped and sampled. Apart from seed density, species richness and composition, functional seed-bank types, defined by nine seed-bank-related plant traits, were related to the land-use intensity gradient. The land-use gradient from gardens to undisturbed sites was still clearly reflected in the soil seed bank. Six emergent functional seed-bank types, characterized by specific plant traits, changed significantly in abundance, parallel to the land-use gradient. In particular, dispersal agent (and related traits) proved an important explanatory trait of present (functional) seed-bank patterns. Poor dispersers (large and heavy seeds) were not found in the intensively used areas, contrary to animal-dispersed species. Wind-dispersers may have been inhibited in the extension of their distribution by recruitment bottlenecks (low seed production) and/or competitive exclusion. Additionally, the agricultural land-use probably introduced ruderal species into the seed bank of the most intensively used areas, yielding a simultaneous increase in vegetation–seed-bank dissimilarity with land-use intensity, eliminating present vegetation as a driver behind the differences over the seed-bank gradient. We conclude by arguing how coppice-with-standards management possibly maintained the seed-bank gradient.