Descriptive analyses
We began with descriptive analyses and illustration of the virus co-occurrence structure. We assessed the nestedness of the data by organising the virus community matrix based on overlap in virus presences among plants and decreasing fill (Almeida-Neto et al.2008). We calculated C-scores (Stone & Roberts 1990) for the virus community at the level of individual host plants and populations (Figure 2). We calculated the numbers of pairwise virus co-occurrence combinations in host plants, as well as the pairs of viruses that never co-occur. We used the full dataset of 25 viruses (See Results) for these descriptive analyses.
We quantified the relationship between the cumulative virus richness and co-occurrence patterns with respect to increasing area sampled by calculating the mean species–area–curve and mean coexistence–curve (Hart et al. 2017). Both curves were constructed by randomly selecting one sampled host plant, and then increasing the spatial scale and sample by including the next closest plant to the species richness or co-occurrence calculation. We calculated the mean curves by repeating this 100 times, selecting a different initial plant for each round. We also calculated the maximum number of co-occurring virus pairs from the number of taxa observed thus far. For more details, see Supplementary Material and Methods and Supplementary Results Figure S2-3.