Fig. 1. Experimental design.
In the conditioning stage, monocultures of 14 common grassland species, each replicated four times, were grown for three growing seasons in large mesocosms filled with soil collected from a single grassland. In the feedback stage, eight newly germinated seedlings of each plant species were transplanted as single individuals back into (i) four replicate conspecific soils (i.e., previously occupied by the same species), (ii) four replicate conspecific soils sterilized with gamma irradiation, and (iii) eight randomly selected heterospecific soils (i.e., previously occupied by other species). Plants were harvested after 60 days of growth. Biotic feedback was calculated as the difference between mean loge-transformed dry mass of two plants grown in each live conspecific replicate soil and two plants grown in the corresponding sterilized conspecific replicate soil (resulting in a total of four feedback values per plant species). Specific soil feedback was calculated as all pairwise differences between loge-transformed dry mass of plants in each conspecific live soil and in each soil conditioned by other species (64 feedback values per plant species). Am, Achillea millefolium; Ao, Anthoxanthum odoratum; Bm, Briza media; Cc, Cynosurus cristatus; Cn, Centaurea nigra; Dg, Dactylis glomerata; Fr, Festuca rubra; Gp, Geranium pratense; Gr, Geum rivale; Hr, Hypochaeris radicata; Lh, Leontodon hispidus; Lv, Leucanthemum vulgare; Pl, Plantago lanceolata; Ra, Rumex acetosa.