Fig. 3. The relationship between mean relative emergence from empirical studies and metrics of predicted specialization for focal hosts. Each point is representative of the relative emergence of a parasite population on a single host genotype and predicted specialization based on either A-C) relative crop harvested area, D-F) herbarium occurrences, or G-I) host-specific ENM contrasts and plotted against actualized specialization (mean relative emergence). Lines are shown based on coefficient estimates from linear mixed models. Beta coefficients with an asterisk indicate that the added fixed effect of a specialization parameter is significant for model predictions of mean relative emergence. The axis for host-specific ENM contrasts (G-I) is inverted such that more negative ENMs reflect higher specialization (see Materials and Methods)