Community Composition
Overall, gopher tortoise exclusion had a weak - ~5% of the variation - albeit significant effect on plant community composition (PERMANOVA, f = 14.45; p = 0.001) (Figure 3, S3, Table 1, S4, S5). Plot spatial structure explained the highest amount of variation (~48%) in plant community composition (PERMANOVA, f = 16.11; p = 0.001), and the interaction between tortoise exclusion and plot location explained an additional 22% of the variation in plant community composition (PERMANOVA,f = 7.47; p = 0.001). The significant interaction term indicates the effect of gopher tortoise herbivory differed across the spatial layout of our plots (Figure 3). Sampling period explained ~3% of the variation in plant composition (PERMANOVAf = 10.26; p = 0.001). The change in plant composition across the growing season did not interact with tortoise exclusion (PERMANOVA, p = 0.47); however, we did find that change in composition across the growing season weakly interacted with plot location (PERMANOVA f = 1.42; p = 0.01).
Our similarity percentage (SIMPER) analysis and NMDS revealed thatSorghastrum secundum (p = 0.04), Geobalanus oblongifolius (p = 0.02), Pteridium pseudocaudatum(p = 0.01), and Toxicodendron radicans (p = 0.01) were significantly more abundant within the exclosure plots relative to the access plots and contributed the most to changes in compositional differences between plot types (Figure S4).