Fig. S1 Coassociation networks of common bacterial and Streptomyces OTUs, for each plant richness treatment (n = 12 samples per plant richness treatment). Nodes correspond to OTUs, and connecting edges indicate correlations in abundance across samples. Nodes are coloured in each panel according to their membership in network modules defined for the monoculture samples. For OTU identifiers associated with each module, see Table S1.

Fig. S2 Patterns of correlation between OTU relative abundance and Streptomyces inhibitory activity, between network modules and between plant richness treatments for given sets of OTUs. Shown are the mean (±SE) Pearson correlation coefficients for OTUs within each monoculture network module, and for the same OTUs across samples from the other plant richness treatments. * indicates a significant difference compared with the corresponding sample from monoculture (anova with Dunnett's test, P < 0.05).

Table S1 List of OTUs included in network analyses, their consensus taxonomic identification and placement in network modules for the whole data set, for the A. gerardii-specific network and for the monoculture-specific network.

Table S2 Chi-square test for the distribution of sequences among phyla, comparing module composition to the composition of the input data set as a whole.

Table S3 Characteristics of coassociation networks for bacterial OTUs, across all samples, or by plant host or plant richness treatment.

Table S4 Bacterial and Streptomyces diversity and richness were sometimes significantly correlated with soil edaphic factors, but patterns of relationship differed between data sets (bacterial vs. Streptomyces).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.