Soil seed bank composition in relation to the above-ground vegetation in fertilized and unfertilized hay meadows on a Somerset peat moor.
In a study on a peat moor in Somerset (SW England, 51°12′N, 2°49′W) in 1991, assessments were made of the soil seed bank in March, and of the aboveground vegetation in the following May. Meadows that had received repeated applications of inorganic NPK fertilizers during the preceding 5 years were compared with those that were unfertilized. The meadows had been cut for hay followed by aftermath grazing each year. A substantial number of seeds of species that normally germinate soon after shedding (i.e. type I species) had survived in the soil until March. This was probably due to a rapid change from very dry to waterlogged soil in the autumn, with both conditions inhibiting germination. Type I species accounted for a disproportionately large fraction of total ground cover relative to the number of species involved, compared with more persistent species. Based upon variance between individual 8-m2 plots, the type I group showed the closest correspondence between vegetation cover and seed bank abundance, while the type IV (long-term persistent seed bank) group showed the least. However, this relationship was close within both the Juncaceae and the Cyperaceae species groups, most of which are type IV species. This was attributed to marked between-plot variation in the specific hydrological conditions to which these species are preferentially adapted. Fertilizer application increased the vegetation abundance of a few species and reduced the abundance of a larger number. These effects were mirrored by differences in seed bank composition. The effects of fertilizer application were greatest within the type I group. This result was more pronounced when species were further subdivided into those that had reacted positively or negatively to applied N in previous experiments in the same meadows. Fertilizers caused a fourfold shift in the balance of species in the seed bank towards those that were previously identified as responding positively to N in terms of vegetation cover, with a smaller concomitant shift from type IV species to type III. Where the aim is to restore high species diversity to previously fertilized meadows, these results show the importance both of reducing soil fertility and of encouraging seed production in those species that have declined.