Figure 3.
Figure 1. Life cycle diagram for tree populations. The baseline and defaunation life cycles are identical except for the italicized terms. Under defaunation, the probability of dispersal (d ) changes for trees dispersed by large vertebrates but remains the same for other tree species. Seed survival (ss,u andss,d ) after defaunation is affected by changes in seed predation mortality. Seedling survival (ss,uand ss,d ) after defaunation is affected by changes in trampling-induced mortality. Seedling survival in both baseline and defaunation scenarios is density dependent (see equation 1) based on the number of conspecific adults.
Figure 2. Relative importances of all variables (the percent of the variance in the data explained by each) included in our additive model in predicting the response of tree populations to defaunation. Dark gray bars show defaunation parameters, light gray bars show species traits.
Figure 3. Variables related to defaunation responses across all tree species and defaunation scenarios. ‘Defaunation response’ is the final number of adults in the defaunation scenario divided by the final number of adults in the baseline scenario. In panels A and C, black circles represent the median defaunation response for each trait or parameter value; vertical black lines show the 95% confidence intervals. The trendlines show model fit based on all datapoints (not just median values) and include 99% confidence intervals that are too small to be seen. In panel B, the solid horizontial line shows the median, the box encompasses the 25th to 75th percentiles, and the vertical bars indicate the 95% confidence interval. There is clearly an inverse relationship between seed predation and defaunation response (panel A). Angiosperms and small-seeded species are more strongly influenced by defaunation—benefiting when seed predation is reduced and suffering when seed predation is increased; gymnosperms and large-seed species are less strongly affected (panels B and C).