Self-help or self-sabotage? A common invader’s soil legacy does not impede, and may facilitate, potential competitors

Abstract Plant-soil feedbacks (PSFs) are increasingly implicated as major drivers of invasion success, yet models for investigating patterns in their impacts remain lacking. Here, we used Bohemian knotweed (Reynoutria × bohemica) to present a novel approach for testing whether accounting for diverse evolutionary drivers can clarify how an invader’s soil legacy mediates its success. For most tested species, R. × bohemica’s soil legacy did not influence seedling performance irrespective of its evolutionary history or the evolutionary novelty of species that share its soil. However, we found the relative importance of PSFs vs. other mechanisms of invader dominance depends on the life stage of the plants growing in soil conditioned by the invader. While the effects of R. × bohemica’s soil legacy on seedling performance were context dependent, emergence was, on average, higher in R. × bohemica-conditioned soils than in unconditioned soils. Thus, contrary to theory, R. × bohemica’s soil legacy not only fails to impede potential competitors, it may facilitate them. Our results are particularly powerful: we detected novel insights in the face of biological complexity, even for R. × bohemica, which likely does not depend on PSFs for its competitive dominance in its introduced range. Designing rigorous investigations that explore how plant species’ evolutionary histories mediate an invader’s soil legacy across diverse abiotic environments, as shown here, is key for producing insights in other systems. Methods We used a series of experimental common gardens of the globally invasive R. × bohemica to explore how plant species’ evolutionary histories mediate an invader’s soil legacy across diverse abiotic environments. To this end, we assessed whether seedling emergence and performance of 15 plant species varied when grown in field soils conditioned vs. unconditioned by R. × bohemica. We also used an established common garden to assess seedling emergence and performance of these same 15 species in soils conditioned by 10 R. × bohemica populations representative of a wide latitudinal range across the eastern United States. Results For most tested species, the soil legacy of R. × bohemica did not influence seedling performance irrespective of species identity, species type, or the latitudinal origin of R. × bohemica. In terms of emergence, plants—on average—fared better in R. × bohemica-conditioned soils than in soils unconditioned by R. × bohemica. However, the impacts of R. × bohemica-conditioned soils on seedling performance were highly context dependent. Usage notes The data we provide in excel are the raw data, and all column headings are explained in a separate excel file called metadata.