Interactions between plants and their root‐associated fungi (RAF) may influence the relative abundance of tree species and determine forest community diversity. Such plant–soil feedbacks in turn depend on the degree to which spatial distance and phylogenetic relatedness of host trees structure pathogen and mutualist communities, but research detailing these aspects of RAF communities is lacking. Here, we characterize plant–RAF associations across a diverse plant community, focusing on the degree to which RAF communities are structured by spatial distance, host phylogenetic relatedness, and host abundance. We compare results for different functional groups, including both putative mutualists and pathogens, an aspect poorly examined hitherto. We collected roots at regular intervals along ten 50 m by 2 m transects, then used DNA barcoding to identify host plants, and characterize the associated fungal community. Variance partitioning was used to measure the relative contributions of host phylogenetic relatedness and spatial distance to explaining RAF community composition. A weighted linear regression was used to measure the correlation between host abundance and RAF diversity. Phylogenetic distance among hosts was a better predictor of RAF community composition than spatial distance, but this relationship was stronger for putative pathogens than for mutualists, suggesting that pathogens show stronger host preference than mutualists. Across all functional groups, RAF showed similar levels of spatial structure. Additionally, RAF communities of locally abundant plants were less diverse than RAF communities of rare plants. Synthesis. We found that RAF communities are structured by the phylogenetic relatedness of hosts and, to a lesser extent, by spatial distance, with pathogens showing stronger host preference than mutualists. Abundant hosts had less diverse RAF communities than rare hosts, which is notable because abundant plants tend to experience weaker negative plant–soil feedback. Going forward, mechanisms underlying the host abundance‐RAF diversity relationship warrant further investigation. Additionally, the survey approach presented here could be paired with experiments linking RAF community composition to plant recruitment.

Interactions between plants and root‐associated fungi can affect the assembly, diversity, and relative abundances of tropical plant species. Host–symbiont compatibility and some degree of host specificity are prerequisites for these processes to occur, and these prerequisites may vary with host abundance. However, direct assessments of whether specificity of root‐associated fungi varies with host abundance are lacking. Here, in a diverse tropical forest in Los Tuxtlas, Mexico, we couple DNA metabarcoding with a sampling design that controls for host phylogeny, host age, and habitat variation, to characterize fungal communities associated with the roots of three confamilial pairs of host species that exhibit contrasting (high and low) relative abundances. We uncovered a functionally and phylogenetically diverse fungal community composed of 1,038 OTUs (operational taxonomic units with 97% genetic similarity), only 14 of which exhibited host specificity. Host species was a significant predictor of fungal community composition only for the subset of OTUs composed of putatively pathogenic fungi. We found no significant difference in the number of specialists associating with common versus rare trees, but we found that host abundance was negatively correlated with the diversity of root fungal communities. This latter result was significant for symbiotrophs (mostly arbuscular mycorrhizal fungi) and, to a lesser extent, for pathotrophs (mostly plant pathogens). Thus, root fungal communities differ between common and rare trees, which may impact the strength of conspecific negative density dependence. Further studies from other tropical sites and host lineages are warranted, given the role of root‐associated fungi in biodiversity maintenance.

No abstract