Fungal colonization
First, we analyzed how level of fungal relatedness affected plant growth and fungal colonization in the whole plant experiment. We found that level of fungal relatedness affected the intraradical colonization of host roots: intraradical colonization of the focal plant root was significantly higher in treatments of non-selfing fungal combinations – meaning that decreased relatedness of the fungi was associated with higher colonization of the focal root. Intraradical colonization of the focal plant connected to a fungal network with high genetic relatedness (i.e. network containing selfing A5 only) was half of that observed when the partner plant was inoculated with the non-selfing partner strain B12, and a fifth of that when the partner plant was inoculated with Agg (One-way ANOVA: F2,23=5.732, p<0.001) (Fig. 2a). The intraradical colonization of the partner plant roots showed the opposite pattern, with higher colonization of partner roots inoculated with A5 versus with B12 or Agg (One-way ANOVA: F2,21=5.617, p=0.011) (Fig. S1 in Supporting Information). We then converted these numbers to a ratio of fungal abundance in focal roots/partner roots so we could compare colonization dynamics within replicates. A ratio of 1:1 means equal fungal growth in both roots. As expected, we found a ratio of ~1:1 in the plants inoculated with only A5. However, as relatedness decreased, this ratio increased (~10 times higher), meaning a bias in favor of colonization in the focal root when relatedness decreased (one-way ANOVA: F2,21=15.812, p<0.0001) (Fig. 2b).
In the in-vitro experiment, we quantified both extraradical and intraradical fungal abundance. We found a similar trend as in the whole plant experiment: decreased relatedness (non-selfing) was associated with increased total fungal abundance. Specifically, we found that the extraradical abundance of highly related A5-A5 networks were roughly eight times lower than when the partner plant was inoculated with the non-selfing partner strain B12 or Agg (one-way ANOVA: F2,37=9.0833, p<0.001) (Fig. 3a). However, we found intraradical fungal colonization did not differ statistically among the treatments (Focal roots: one-way ANOVA: F2,38=0.418, p= 0.662; partner roots: F2,38=0.68, p=0.513) (Fig. S2a&b). We also found that partner roots inoculated with Agg, were also colonized by fungal strain A5 (7.09·105 ±3.2·104 copy numbers, Fig. S2c). Since A5 and Agg are not known to fuse, this indicates that the fungus of the focal root crossed the central compartment and into the partner root compartment. We then tested whether level of relatedness affected investment ratio of the fungus, defined as fungal allocation to intraradical versus extraradical growth. We found that low fungal relatedness resulted in a bias towards more extraradical growth (one-way ANOVA: F2,37=12.343, p<0.001) (Fig. 3b).