Diverse plant communities are consistently more productive and support other desirable ecosystem functions. While productivity gains from diversity are often attributed to increased resource partitioning, direct evidence of this mechanism in plant biodiversity experiments has been limited. This could be due to inadequate measures of resource partitioning or the importance of other, unmeasured mechanisms. Pathogen dilution is a prominent alternative explanation, suggesting that productivity benefits arise when compatible pathogen hosts are buffered by unrelated neighbors, thereby diluting negative impacts like disease incidence and inhibited growth. Past research on pathogen dilution has mainly focused on wildlife and human diseases, with meta-analyses indicating that biodiversity decreases disease prevalence. However, the generality of pathogen dilution across systems remains controversial, and studies of pathogen dilution in plant communities have been less common, with sparse direct evidence of its contribution to plant productivity and ecosystem functioning. Testing the importance of pathogen dilution in plant communities is challenging due to the high taxonomic diversity of pathogens and difficulties in measuring their abundance. Conclusions about pathogen importance may depend on which pathogens are selected, and their effects may depend on interactions with omitted pathogens. Studies often focus on patterns of pathogen symptoms, while hypotheses about pathogen mediation of biodiversity-productivity relationships depend on pathogen dilution alleviating negative impacts on plant growth. An alternative approach focuses on the net impacts on plant fitness from host-specific changes in microbial composition—the plant-soil feedback (PSF) framework. This framework recognizes that pathogen growth rates are host-specific, leading to differential pathogen accumulation on particular hosts. While the effects of multiple pathogens are hard to evaluate, the net effect of pathogen dynamics on plant-plant interactions can be characterized using reciprocal inoculation experiments. This analysis reveals that pathogens can contribute to plant species coexistence when intraspecific effects are more strongly negative than interspecific effects. This occurs when a plant species' fitness declines due to pathogen accumulation on its own species (net pairwise negative feedback), a common phenomenon in native plant communities. This stabilizing influence of pathogen-generated negative feedback might also contribute to statistical complementarity between different plant species, resulting in overyielding. Accumulation of host-specific pathogens could reduce productivity in monocultures, while these negative impacts decrease in diverse communities due to reduced densities of compatible hosts. Manipulations of microbiome composition in mesocosms and in the field provide direct evidence of pathogen mediation of overyielding, and negative plant-soil feedback models have predicted patterns of overyielding. However, no study has demonstrated the causal connections between host-specific pathogen accumulation, the strength of negative feedback, and the magnitude of complementarity effects observed in the field. Understanding this causal mechanism is crucial for predicting the context dependence of biodiversity-productivity relationships. For instance, pathogen dilution-mediated complementarity should be more common in phylogenetically diverse plant communities, as phylogenetically similar species are more likely to share pathogens. Pathogen dilution should increase with plant species richness, as should pathogen dilution-driven complementarity. This study aimed to investigate the role of plant-pathogen feedbacks in driving plant diversity-productivity relationships through an integrated approach of field manipulations, greenhouse assays, and feedback modeling.

Numerous studies have explored the relationship between biodiversity and ecosystem functioning, often focusing on the positive correlation between plant diversity and productivity. Early experiments, such as those in European grasslands, demonstrated increased productivity with higher plant diversity. These findings spurred extensive research into the underlying mechanisms driving this relationship. One prominent hypothesis centered on resource partitioning, suggesting that diverse plant communities utilize resources more efficiently due to niche differentiation. However, studies examining resource partitioning in biodiversity experiments have yielded mixed results, with limited empirical evidence directly supporting this mechanism as the sole driver of diversity-productivity relationships. This has led researchers to consider alternative mechanisms, including the role of soilborne pathogens and the concept of pathogen dilution. The dilution effect hypothesis posits that the presence of multiple plant species can reduce the impact of specialist pathogens by diluting the density of susceptible hosts. However, the generalizability of this effect across various ecosystems has been a subject of ongoing debate, with studies providing conflicting evidence. Meta-analyses have suggested that biodiversity can indeed decrease disease prevalence in various systems. However, the extent to which pathogen dilution contributes to plant productivity and overall ecosystem functioning in plant communities has remained relatively unexplored. The plant-soil feedback (PSF) framework provides a complementary approach to studying the role of pathogens in shaping plant community dynamics. This framework focuses on the net effects of host-specific changes in microbial composition on plant fitness. It acknowledges that pathogens differentially accumulate on various hosts, and the net effect of these dynamics can be analyzed through reciprocal inoculation experiments. Research using this framework has shown that pathogens can contribute to plant species coexistence when negative intraspecific effects are stronger than interspecific effects. Several studies have suggested a link between plant-soil feedbacks, pathogen dynamics, and overyielding, where the combined biomass in mixed-species communities exceeds the sum of monoculture biomasses. However, a direct causal link between host-specific pathogen accumulation, the strength of negative plant-soil feedbacks, and observed complementarity effects in the field has been lacking.

This study employed a multi-faceted approach combining field experiments, greenhouse assays, and theoretical modeling to investigate the role of pathogen dilution in generating diversity-productivity relationships in plant communities. **Field Experiment:** The field experiment was conducted at the University of Kansas Field Station, using 18 native prairie plant species from three families (Poaceae, Fabaceae, and Asteraceae). A total of 240 plots (1.5m x 1.5m) were established, manipulating species richness (1, 2, 3, and 6 species), plant community composition (phylogenetically under- or over-dispersed), and precipitation (50% or 150% ambient rainfall). Under-dispersed treatments contained plants from a single family, while over-dispersed treatments contained species from multiple families. Soil and root samples were collected four months after planting to analyze bacterial, fungal, arbuscular mycorrhizal fungi (AMF), and oomycete communities via high-throughput sequencing. **Greenhouse Assays:** To assess plant-soil feedback (PSF) effects and pathogen dilution, greenhouse assays were conducted using the soil samples collected from the field experiment as inocula. All 18 plant species were grown in pots inoculated with soil from conspecific monocultures or heterospecific monocultures, resulting in 81 full-factorial pairwise feedback tests. Plant shoot and root biomass were measured after two months. Pairwise PSF effects were calculated using the log response ratio of plant performance in conspecific vs. heterospecific soils. **Feedback Modeling:** A general feedback model was parameterized using the empirically derived plant-microbiome interaction strengths. The model was used to explore the potential of PSFs and pathogen dilution to promote long-term community coexistence and overyielding. The model describes plant abundance as proportions, with plant-soil microbiome effects driving frequency-dependent feedbacks. Changes in plant frequencies are modeled using differential equations incorporating plant fitness and interaction terms representing the effects of one species on another's fitness. The model was parameterized using empirically quantified dissimilarities in fungal pathogen, root fungal pathogen, and oomycete compositions of plant hosts. The model was used to assess the feasibility, community-level feedback, and local stability of various multispecies communities assembled from the 18 plant species. Complementarity effects were calculated based on the difference between equilibrium fitness and monoculture fitness. **Statistical Analysis:** Pairwise Bray-Curtis dissimilarities were calculated to analyze microbial composition differences between plant monocultures. Plant biomass in greenhouse assays was analyzed using linear models, with plant species, inocula species, and their interactions as fixed effects. Pairwise PSF effects were compared within and between families. Regressions were performed to assess the relationships between microbial community dissimilarities and pairwise PSFs. Model selection using AICc was performed to identify the most important microbial predictors of PSF. A best linear model was derived to characterize pairwise PSF effects. Complementarity (CE), selection effects (SE), and relative yield total (RYT) were calculated for the field data using aboveground biomass from the second growing season. ANOVA and t-tests were used to analyze the effects of various factors on biomass, CE, and RYT. Linear regression was used to analyze the relationships between independent variables (predicted PSF effect and predicted pathogen dilution) and dependent variables (CE and RYT).

The study revealed rapid differentiation of soil pathogens in response to plant community composition four months after establishing the field experiment. Greenhouse assays confirmed that this pathogen divergence generated negative pairwise plant-soil feedbacks (PSFs). On average, pairwise PSFs were negative, indicating that pathogens suppressed plant growth more strongly in monocultures than in mixtures. Different feedback patterns were observed between species from different families. PSFs were negative within composite and grass families and between composite and grass families, but positive within the legume family. Soil fungal and oomycete pathogen dissimilarities mirrored the patterns observed with negative feedbacks, with greatest average pathogen dissimilarities found within composites and grasses and between composites and grasses or legumes. Pairwise PSF values significantly decreased with both soil fungal and oomycete pathogen dissimilarities, suggesting that fungal pathogens played a key role in driving negative PSFs. The most abundant fungal pathogens differed for most plant species, suggesting candidate pathogens that may drive negative feedbacks. The study found a strong relationship between the strength of negative PSFs and the degree of complementarity and productivity in the field the following year. More negative PSFs corresponded to higher complementarity and productivity gains with increasing plant diversity. A parameterized general feedback model corroborated that these relationships are causal and can generate linear relationships between plant diversity and productivity over time. The results provide strong evidence that dilution of host-specific pathogens contributes to diversity-driven yield advantages. The model showed positive relationships between complementarity and plant richness, a negative relationship between complementarity and predicted PSF effects, and a positive relationship between complementarity and expected pathogen dilution. These findings highlight the importance of pathogen dilution in generating biodiversity-ecosystem functioning (BEF) relationships in plant communities.

This study provides compelling evidence that dilution of host-specific pathogens can significantly contribute to the observed productivity gains associated with increased plant diversity. The findings demonstrate a clear causal link between pathogen-driven negative feedbacks, pathogen dilution in diverse communities, and increased productivity. The results support previous studies manipulating pathogen abundance and those connecting negative feedback to positive diversity-productivity relationships. While the generality of pathogen dilution across systems needs further investigation, this study shows its independent ability to generate productivity increases with plant diversity. The context-dependent nature of pathogen dilution is likely influenced by factors such as the coevolution of plants and pathogens and the effects of environmental change on plant-microbe interactions. The observed strong complementarity in the first two years of the biodiversity manipulation, potentially influenced by the use of native prairie soil microbiome, highlights the importance of considering soil microbiome composition in BEF experiments. Global change drivers, such as warming, high precipitation, and nutrient enrichment, can strengthen BEF relationships by favoring pathogen growth. While this study focused on productivity gains, future research should explore the link between pathogen dilution and other positive ecosystem responses associated with plant diversity. The theoretical results suggest that pathogen dilution can promote community coexistence, maintaining productivity increases over extended periods.

This study offers strong evidence that pathogen dilution, driven by host-specific pathogens, significantly contributes to the positive relationship between plant diversity and productivity. The integration of field experiments, greenhouse assays, and a theoretical model provides a robust understanding of this mechanism. The findings highlight the importance of considering pathogen dynamics when investigating biodiversity-ecosystem functioning relationships and emphasize the potential for environmental changes to decouple plant-microbe interactions, thereby affecting productivity benefits from diversity. Future research should explore the generality of this mechanism across different ecosystems and investigate the interactions between pathogen dilution and other factors influencing BEF relationships.

While the study used a robust experimental design and comprehensive data analysis, some limitations should be acknowledged. The study focused on a specific prairie ecosystem with native plant species and soil, limiting the generalizability to other ecosystems with different plant communities and soil microbiomes. The specific identity of pathogens driving the observed effects was not fully determined, although the data strongly implicates soil fungal pathogens and oomycetes. The study did not explicitly investigate the relative contribution of pathogen dilution to other potential mechanisms driving diversity-productivity relationships, such as resource partitioning. Finally, long-term studies would strengthen the conclusions concerning the maintenance of overyielding over time.