Supplementary Methods 1: Evaluate herbivore biomass in the field survey
We used herbivore biomass to represent herbivore pressure (Allen et al., 2021) as higher herbivore abundance may not always translate to greater herbivore damage if feeding by particular herbivore species is relatively minor (e.g., aphids and beetles) or extensive (e.g., larvae). To determine whether the herbivore biomass correlated with leaf damage, we checked all leaves and recorded the number of each herbivore species on each non-native plant. We classified observed herbivores into five orders (nine categories), including Lepidoptera (larvae), Orthopter, Thysanoptera, Hemiptera (leafhoppers, stinkbugs, whiteflies, and aphids), and Coleoptera (beetles and herbivorous ladybirds) since some larvae were difficult to identify to species. A total of 20,723 insects on non-native plants were counted (Native plants were not measured as these measurements were primarily to correlate damage to insect biomass). Then, we evaluated herbivore biomass on each plant by multiplying abundance and mean herbivore biomass of each category and then summing herbivore biomass of all categories. To obtain the mean biomass of each herbivore category, we carried out a field survey at eight sites in 2019. We collected all herbivores on non-native species, stored them in plastic tubes with leaves of host plants and determined the fresh weight of individual insects to the nearest 0.01g in the lab. The mean biomass of each herbivore category was as follows: Lepidoptera, 0.25 g/ individual; Orthoptera, 0.45 g/ individual; Thysanoptera, 0.01g/30 individuals; leafhopper, 0.03g/ individual; stinkbug, 0.19 g/ individual; whitefly, 0.01 g/30 individuals; aphid, 0.01 g/5 individuals; leaf beetle, 0.05 g/ individual; ladybird, 0.04 g/ individual.