Evolutionary novelty shapes predator-mediated insect herbivory in exotic oaks.

Phylogenetic distance between exotic and native plant species is a well documented predictor of insect herbivory, but whether it also influences predator-mediated (top-down) control remains unresolved. In this study, we manipulated predator access on 18 exotic oak (Quercus) species grown in a common garden, spanning a gradient of phylogenetic distance to four european oaks (Q. robur, Q. petraea, Q. suber, or Q. pyrenaica). We excluded three predator guild treatments: ground-dwelling arthropod predators (hereafter ant exclusion, because the sticky-barrier method used primarily excludes ants), vertebrate predators (mainly birds), and both predator groups combined. We quantified cumulative leaf herbivory and plant chemical traits (volatile organic compounds and phenolics). Under ambient predator access, herbivory declined with increasing phylogenetic distance from native oaks (r = - 0.49, P = 0.038), consistent with previous work in this system. This relationship disappeared following ant exclusion (r = - 0.08, P = 0.76), and a phylogenetic generalized least-squares model detected a significant treatment × phylogenetic distance interaction (F = 2.96, P = 0.033) that remained robust after accounting for phylogenetic non-independence among species (phylogenetic regression slope, P = 0.001). Excluding vertebrate predators increased herbivory consistently across the phylogenetic gradient. Total VOC emissions and phenolic concentrations were unrelated to phylogenetic distance; however, multivariate VOC composition diverged with phylogenetic distance (Mantel r = 0.26, P = 0.049), unlike phenolic composition (P = 0.62). These results indicate that ant-mediated predation, not bulk chemical defences, underlies the phylogenetic herbivory pattern, highlighting predator identity as an underappreciated axis of plant-herbivore-predator interactions in novel communities.