Figure 5.

Aboveground plant biomass 18 months after planting for the five original coffee cultivars in the experiment (A). Biomass was positively affected by the rainout treatment across the cultivars, though to a lesser degree than fruit weight (Fig. 4A). Coefficient plots for control intercepts (B) and rainout effects (C) show estimates as offsets from the overall mean (for more detail, see  Supporting Information Table S3). Where the 95 % confidence interval does not overlap 0, we can infer that the estimate differed significantly from the mean.