1. Plant neighbors can strongly influence the interactions between herbivores and focal plants, for instance by providing food of different quality (consumptive effects) or by changing the behavior and metabolism of the herbivore and the focal plant without being consumed (non-consumptive effects). Determining the species-specific contributions of consumptive and non-consumptive effects is important to understand the ecophysiological mechanisms which underlie neighborhood effects. 2. We quantified the impact of nine different grassland plant species on the interaction between Taraxacum officinale and the root-feeding insect Melolontha melolontha. We investigated the contribution of consumptive and non-consumptive effects to the observed patterns, and evaluated the impact of neighboring plants on the growth and physiology of T. officinale upon M. melolontha attack. 3. Melolontha melolontha growth was strongly affected by the presence of different neighboring species. The three grass species increased larval growth when growing with T. officinale, with Poa pratensis having a synergistic effect in the mixture compared to both monocultures. The forb Centaurea stoebe reduced larval growth when growing with T. officinale or alone. The other five species had no measurable impact on larval performance. Based on these results, P. pratensis and C. stoebe were selected for further experiments. 4. Diet-mixing experiments showed that P. pratensis increased M. melolontha growth when offered together with T. officinale, while C. stoebe suppressed it. When feeding was restricted to artificial diet, larval growth was not changed by the presence of P. pratensis or C. stoebe. However, when feeding was restricted to T. officinale, larval growth was increased by both heterospecific neighbors. Biomass and primary metabolism of T. officinale under attack by M. melolontha was also altered by the presence of C. stoebe and P. pratensis. Together, these results show that consumptive and non-consumptive effects can explain the positive effect of P. pratensis. By contrast, the negative effect of C. stoebe is likely driven exclusively by intoxication. 5. Synthesis. The performed experiments suggest that different combinations of consumptive and non-consumptive effects are likely to contribute to the diversity of neighborhood effects in nature. Furthermore, they show that neighborhood effects are important factors in below ground plant-insect interactions.