Early vs. Late Succession
The four Frankia soil clades discerned in our phylogenetic analysis were non-randomly distributed across early and late-succession soils derived from A. tenuifolia rhizospheres (Table 1, pseudo-R2=0.09, likelihood ratio (LR) P < 0.0001). Between early and late succession, proportional representation of AT clade sequences decreased significantly (Fisher’s exact P < 0.0001), representation of soil clade B increased significantly (Fisher’s exact P < 0.0001), and ‘typicalAlnus ’ representation trended toward a significant increase (Fisher’s exact P =0.07).
Distribution of sub-clades was also non-random between early and late succession (Table 2) (pseudo-R2=0.17, LR P < 0.0001) soils. The largest difference occurred for the early-succession nodule specialist RF7 (Fisher’s exact P < 0.0001), which was not found in any late succession soils but made up >25% of the clones derived from early-succession soils (Table 2). Significant differences also occurred for soil sub-clades B1 (P =0.01), B3 (P < 0.0001), and B4 (P = 0.007), which yielded 6.1%, 16.7%, and 7.0% of clone sequences from late succession alder soils, but were not found in early succession soils. Sequences related to the ‘typicalAlnus ’ group that is dominant in nodules of A. viridis in our study sites were found in seven (6.1%) clones from late succession alder soils, but only one (0.9%) early succession clone. This difference trended toward significance (P = 0.07). Late succession soils, both within and outside of A. tenuifolia rhizospheres, appear to support higher OTU richness than early succession soil (Figure 2A).