The intensity of intraspecific plant-soil feedbacks in alien Impatiens species depends on the environment

Abstract

Plant-soil feedback, i.e. the relationship in which a plant affects the composition of the soil and such modified soil affects plant growth, is becoming an important concept for explaining plant invasiveness. Impatiens parviflora is one of the most widespread invasive plant species in Central Europe, but it is unknown whether this species exhibits any form of plant–soil feedback, previously seen in other invasive species. In this study, we examined intraspecific plant-soil feedback of I. parviflora and compared it to feedbacks of other three alien, but non-invasive, Impatiens species growing in Central Europe. Moreover, we studied the effect of environmental conditions on this feedback.The four species were studied in a two-phase feedback experiment. In the first phase, soil was conditioned by a species or left unconditioned (control). During the second phase, plants were grown in soil conditioned by the same species and in control, not cultivated soil, under four different environmental conditions (two levels of watering × two levels of shading).All the studied species showed positive plant-soil feedback in some types of environment indicating they can potentially become invasive. The feedbacks for total biomass and for root-shoot ratio were significantly affected by environmental conditions. Individual species and studied performance measures responded differently to individual treatments. In most cases, the feedback was changing from positive in optimal treatment to neutral and in I. balsamina to negative under some suboptimal treatments. No effect of the environmental conditions on the feedback for germination was observed.These results indicate that environmental conditions play an important role in plant-soil feedbacks. We show that some feedbacks detected in highly controlled experiments using just one type of environmental conditions might not exist in the field, and similarly some existing feedbacks might remain undetected. We therefore recommend using multiple environmental conditions or conditions closely resembling conditions in natural sites to increase the realism of the results.