Plant-soil feedbacks (PSFs), arising from interactions between plants and soil biota (mutualistic mycorrhizal fungi and soil-borne pathogens), significantly influence plant community dynamics. Positive PSFs result from the accumulation of mutualistic fungi, while negative PSFs stem from soil-borne pathogens that hinder plant growth. The role of PSFs in shaping community structure is complex and debated. Some studies suggest dominant species exhibit net positive or weaker negative PSFs than rare species, while others indicate negative PSFs for most plants, maintaining diversity through negative frequency dependency. The enemy release hypothesis (ERH) proposes that invasive plants succeed due to the absence of species-specific pathogens in their introduced range. However, evidence supporting this hypothesis is mixed. Most PSF research relies on monoculture experiments, neglecting potential alterations of PSFs by interspecific competition in diverse communities. This study aimed to determine if monoculture PSFs predict PSFs in communities and explore the role of PSFs in invasive plant success.

The literature on plant-soil feedback and its impact on plant community structure and invasive species success is extensive but presents conflicting conclusions. Some studies report dominant species having positive or weaker negative PSFs compared to rare species. Others, however, found negative PSFs to be prevalent, promoting coexistence through negative frequency dependency. The enemy release hypothesis (ERH) suggests that invasive species thrive due to a lack of species-specific pathogens in their new environment, yet empirical evidence is inconsistent. Most studies utilized monoculture experiments, failing to account for the influence of interspecific competition on PSFs in communities. A gap exists in community-level PSF research, necessitating experiments investigating the interaction between competition and PSFs to accurately understand their role in community dynamics.

This study employed a combined monoculture and community experiment using 10 native and 3 invasive grassland species. The experiment was divided into two phases: a conditioning phase, where plants were grown in soil to establish soil microbial communities, and a response phase, where plants were grown in the conditioned soil under different treatments. Three soil treatments were used: sterilized soil (control), untreated soil (PSFtot), and soil with added pathogen/saprobe filtrate (PSFpath). PSFs were calculated as the difference in biomass between treated and sterilized soil, standardized by the average biomass in sterilized soil. Monocultures and communities (with and without invasive species) were used to compare PSFs. Statistical analyses, including standardized major axis regression, mixed linear models, and ANOVA, were conducted to assess the relationships between PSFs, species biomass, functional group, and invasive species presence. Above-ground and below-ground biomass were measured.

The study revealed several key findings. First, PSFs from monocultures were poor predictors of PSFs in communities (R² = 0.29 & 0.12 for PSFtot and PSFpath, respectively). This indicated that monoculture experiments do not accurately reflect community-level PSFs. Second, dominant species in the community experienced significantly more negative PSFs (both PSFtot and PSFpath) than non-dominant species (P < 0.001). This dominance-driven negative PSF is attributed to pathogen fractioning, where pathogens specific to dominant species are more abundant in the soil. This mechanism is absent in monocultures. Third, grasses showed significantly more negative PSFs than legumes or forbs in communities (P < 0.001), but not in monocultures. This supports earlier findings linking high root-to-shoot ratios in grasses to increased pathogen vulnerability. Fourth, the enemy release hypothesis (ERH) was not generally supported. Only one of the three invasive species showed significantly higher biomass in unsterilized soil compared to sterilized soil. Fifth, PSFs had a significant net negative effect on community biomass production, and the addition of invasive species could also reduce native biomass, although this was only significant for *Avena sterilis*. Finally, the difference in PSF between community and monoculture was inversely correlated with the species biomass in the sterile soil in communities (R² = 0.87 and 0.43 for PSFtot and PSFpath, respectively).

The findings challenge the reliance on monoculture experiments to understand community-level PSFs. The results highlight the importance of community-level experiments to capture the interplay between competition and PSFs. Negative frequency-dependent effects of PSFs, particularly impacting dominant species, appear to be a key mechanism promoting plant coexistence and maintaining biodiversity. The lack of consistent support for the ERH suggests that invasion success depends on factors beyond pathogen release. The results suggest a more complex interaction between PSF, biodiversity, and invasion dynamics. Increased biodiversity may mitigate the negative PSFs experienced by native species, potentially reducing the advantage gained by invaders through enemy release. Future studies should investigate the role of PSF in communities with varying biodiversity and functional composition.

This study demonstrates that plant-soil feedbacks (PSFs) in communities differ significantly from those in monocultures, with dominant species disproportionately affected by negative feedbacks. This negative frequency dependence promotes species coexistence. The enemy release hypothesis is not consistently supported in diverse communities. Future research should focus on community-level experiments with varying biodiversity and functional groups to refine our understanding of PSFs and their role in shaping plant communities.

The study's use of a 20 µm pore-size filter for pathogen/saprobe filtrate preparation might have introduced limitations. While this method is standard, it retains some organisms and excludes others, possibly influencing PSF measurements. Some nematode eggs could have passed through the filter, and differences in recolonization rates between soil organisms could have introduced bias. The slightly lower total native biomass in unsterilized soil versus sterilized soil with added filtrate might reflect differences in pathogen recolonization or sensitivity to processing.