Table 1.
Summary of initial sample sizes and final sample sizes of plants used in analyses after accounting for Epichloë coenophiala infection and plant mortality in the field.
Fig 1.
Conceptual diagram of key experimental steps, data processing, and data analysis.
Fig 2.
Graphical table of the four parasite species and pictures of their associated disease symptoms on tall fescue leaves.
Necrotrophic parasites feed on dead host cells, while biotrophs feed on living cells. Hemibiotrophs employ both biotroph and necrotroph feeding strategies throughout their lifecycle. Pictures were taken by B. N. Joyner.
Fig 3.
Disease progression measured as the area under the disease progress stairs (AUDPS) of A) C. cereale, B) R. solani and C) P. coronata varied in response to individual hosts’ initial inoculation with symbionts. No symbiont (mock inoculation) Col (C. cereale inoculation), Rhiz (R. solani inoculation), Col + Rhiz (simultaneous C. cereale and R. solani inoculation), and Epi (E. coenophiala positive plants). Letters denote grouping based on Tukey-HSD post hoc comparisons conducted on disease progression of each parasite species. Filled circles are estimated means and their 95% confidence intervals. Smaller unfilled circles display the raw data that are jittered to show the distribution of the data.
Fig 4.
The cumulative temporal change of parasite communities (i.e., the total distance moved by a parasite community).
Filled circles are the treatment means and their 95% confidence intervals. Letters denote grouping based on Tukey-HSD post hoc comparisons. Smaller unfilled circles display the raw data that are jittered to show the distribution of the data.
Fig 5.
Divergence of parasite community structure over time.
Each panel shows parasite community patterns for an inoculation treatment group. The arrows show the direction of parasite community assembly over repeated field surveys, and the green arrow indicates the final survey and end of the trajectory. The location of each parasite species name (the first four letters of the genus) indicates the species projection on the PCoA axes.
Fig 6.
Early in the experiment, variation in parasite community structure among hosts increased over time, consistent with ecological drift.
The increase in variation of community structure was no longer apparent after day 21 of the field survey, consistent with the prediction that drift would be more important earlier in the experiment. Plotted is the fitted line (solid) and 95% confidence interval (dashed lines) from the piecewise regression mixed model, and each point is the mean variation in community structure across all hosts at a given survey event.