National Taiwan University, Taiwan (Republic of China)
In plant-soil feedback, plants and soil microbes interact to shape community assembly above and belowground. Plants can ‘condition’ the soil by recruiting species-specific pathogens or mutualists that affect the next generation of plant coexistence. Classic feedback experiments grow plant A to ‘condition’ the soil, then grow a ‘response’ plant B in soil A to study the feedback of plant A on plant B. This design fails to consider (1) the effect of litter decomposition and (2) that plants do not immediately colonize and grow after a plant has died. We hypothesize that (1) litter, and (2) the time gap between soil conditioning and plant response, may shift the microbial community, changing the strength and direction of plant-soil feedback. We conditioned field soil with California native annual plants of three functional groups: legume, grass, and forb. We then used the conditioned soil as microbial inoculant on three ‘response’ treatments where each of the plant species were grown: immediately after conditioning, six months delayed after conditioning without litter, six months delayed with litter. In a fully factorial greenhouse experiment, we collected dry plant biomass to quantify stabilization and fitness differences between species pairs, and soils to sequence for bacterial and fungal communities.
Our results show that in the immediate response treatment, legumes competitively excluded grasses and forbs, while grasses and forbs can coexist. This was associated with the recruitment of beneficial nitrogen-fixing bacteria in legume-conditioned soil that allowed conspecifics to perform better, destabilizing coexistence. In contrast, pathogenic microbes found in grass and ford-conditioned soils generated negative frequency-dependent dynamics allowing plants to grow more in heterospecific soil. In the treatment where response plants were grown six months delayed without litter, all pairwise species interactions destabilized relative to the immediate response, and competitively excluded heterospecifics. The microbial community also shifted. In delayed response with litter, forbs interaction with grasses and legumes destabilized and competitively excluded them relative to the immediate response, but legumes and grasses could coexist due to priority effects generated by pathogenic microbes. Our results quantify microbially-mediated fitness differences to show that incorporating (1) litter decomposition and (2) time delay in plant-soil feedback can change plant coexistence outcomes. These variables could explain why plant-soil feedbacks are often found to be stronger in greenhouse experiments than in the field where litter and seasons, as we have modeled in our experiment, change the magnitude of feedback effects.