Extended Data Fig. 3: Leaf traits that are potential drivers of herbivory in mature leaves, and herbivory measurements on young leaves.

Showing both the leaf traits that varied between fixers and non-fixers in mature leaves, and metrics of herbivory and leaf retention on young leaves. For mature leaves, the difference in a, leaf area, b, leaf nitrogen concentration, c, leaf cellulose concentration, d, leaf carbon concentration, e, leaf lignin concentration and, f, leaf potassium concentration. All differences in leaf variables for mature leaves are significant as determined by two-sided Wilcoxon rank test on n = 184 fixer and n = 166 non-fixer species. N = 43 (a), 37 (b), 38 (c), 37 (d), 38 (e) and 37 (f) biologically independent samples. For young leaves, g, the incidence of herbivory; h, the proportion of leaf area lost to herbivory per day for damaged leaves (Proportion_damaged) on each seedling; i, the proportion of leaf area lost to herbivory per day on all leaves (Proportion_all) of each seedling; and, j, the proportion of sampled leaves that still remained after three months (i.e. leaves that have not been dropped by the plant). Nitrogen fixers are represented in orange and non-fixers in grey. For the measures of herbivory on young leaves there were no difference between fixers and non-fixers, as determined by two-sided non-parametric Wilcoxon rank test (n = 226 (119 fixers, 107 non-fixers)). Points represent seedlings with the lines representing means (± standard error) across seedlings; bars represent mean (± standard error).