We used linear mixed effects models (LMMs) to test how parasite specialization in experiments was related to 1) host community variation (relative crop area harvested, Q1), 2) parasite occurrence across hosts (from herbarium records, Q2), and 3) predicted specialization from abiotic environment (host-specific ENM contrasts, Q3). For each of the three predictors of host specialization, we modeled mean relative emergence of an S. hermonthica population i, on a focal host as follows:
Emergenceij = b0 + b1Xi + S0j + I0i + R0i + 𝜖ij