Seasonal variation visible despite undersampling, but subtler
differences were masked
With 30 mg subsampled per plant we were able to see broad seasonal
differences between May and September (Figure 6A). However, this reduced
our power to differentiate among plants that harbored more similar
communities. This became apparent when analyzing multiple replicates per
plant (Figure 6B). When we visualized the communities recovered in each
additional extraction, we observed significant variation among plants
that was not apparent when we only had community data from single 30 mg
extractions. With the additional extractions we observed strong host
filtering among individual plants (PERMANOVA, R2 =
0.91, P < 0.001). Subsequent pairwise analyses revealed that
the fungal community within individual plants varied significantly from
all other plants (Pairwise PERMANOVA, P < 0.01), and that
three of the plants appeared much more similar to each other than the
other two plants, potentially due to factors associated with their
spatial distribution within the site.