Plant-associated microbes play a key role in mediating the relationship between plant diversity and productivity. However, previous studies have generally focused on a sole microbial guild (i.e. plant-beneficial microbes or pathogens), and on either aboveground or belowground microbes. As a result, the interplay among different microbial guilds and the overall impact of above- and belowground microbes on plant diversity-productivity relationships have rarely been investigated. Here we carried out an experiment where we applied microbial inocula collected from leaves and soils in the field onto plant leaves and soil in a greenhouse experiment with a herbaceous plant community. We showed that microbial inoculation of leaves reduced plant productivity and this negative effect was weaker at higher plant diversity, which promoted positive diversity-productivity relationships through complementarity effects. In contrast, microbial inoculation of soil alone had no impact on plant diversity-productivity relationships, but it counteracted the negative effects of leaf inoculum on plant productivity and weakened the leaf microbe-induced positive diversity-productivity relationships. We found that the abundance of arbuscular mycorrhizal fungi increased when soil microbes were inoculated, and such increase was more significant at lower plant diversity, potentially explaining the effects of soil inoculation on plant productivity. These results suggest that the belowground plant beneficial microbes can counteract the effect of aboveground plant pathogens in mediating positive plant diversity-productivity relationships. Simultaneous study of plant-pathogenic and -beneficial microbes both above- and belowground is required to better understand the contributions of plant-associated microbes to biodiversity-ecosystem function relationships.

The widespread observation that rare species have stronger conspecific plant-soil feedback (PSF) than common species raises more questions than answers on how rare species can possibly win the dance with abundant species. Here, we test soil feedback effect of phylogenetically related species on seedlings of contrasting local abundance in a subtropical forest. The results showed that although rare species suffered strong negative PSF in soils of conspecifics or phylogenetically close relatives, no such feedback was found in the soils of distant relatives. In contrast, although common species had weak conspecific PSF, they suffered consistently strong heterospecific soil feedback. These mechanisms ensure that rare species would fare well in the neighborhood of phylogenetically distant heterospecifics but do poorly under their close relatives, while common species perform relatively well in their own neighborhood but poorly in others’. This phylogenetic conservatism in PSF facilitates the persistence of rare species in a community.