Reciprocal interactions between plants and soil in an upland grassland

Through the production of litter, plants with different life history strategies are predicted to both affect and be affected by the properties of soil. Competitive species are expected to increase the fertility of, and have a positive growth feedback with, soil, whereas stress-tolerant species should decrease fertility but show no growth feedback. We maintained monocultures of competitive ( Lolium perenne and Agrostis capillaris ) and stress-tolerant ( Festuca ovina and Nardus stricta ) grasses on an unproductive grassland for six years. The Nardus soil developed significantly greater inorganic nitrogen than the Agrostis and Festuca soil, and significantly greater soil moisture content than the Festuca soil. However, there were no differences in organic matter content, phosphate or bulk density between the soil types. In a greenhouse assay, each species was grown in soil cores from the different monocultures as well as natural turf. There were significant differences in growth between plant species and soil types. As expected, L. perenne produced the greatest amount of biomass. However, plants grown on Nardus soil were twice as large and had a 21% lower root allocation than plants grown on any of the other soil types. Lolium perenne, A. capillaris and F. ovina had significant negative growth feedbacks with their own soil (−0.460, −0.821 and −0.792, respectively) and N. stricta had a significant positive feedback (0.560). This study highlights the difficulties of predicting how plant traits will affect soil properties.