Urbanization impacts plant-herbivore interactions, which are crucial for ecosystem functions such as carbon sequestration and nutrient cycling. While some studies have reported reductions in insect herbivory in urban areas (relative to rural or natural forests), this trend is not consistent and the underlying causes for such variation remain unclear. We conducted a continental-scale study on insect herbivory along urbanization gradients for three European tree species: Quercus robur, Tilia cordata, and Fraxinus excelsior, and further investigated their biotic and abiotic correlates to get at mechanisms. To this end, we quantified insect leaf herbivory and foliar secondary metabolites (phenolics, terpenoids, alkaloids) for 176 trees across eight European cities. Additionally, we collected data on microclimate (air temperature) and soil characteristics (pH, carbon, nutrients) to test for abiotic correlates of urbanization effects directly or indirectly (through changes in plant secondary chemistry) linked to herbivory. Our results showed that urbanization was negatively associated with herbivory for Q. robur and F. excelsior, but not for T. cordata. In addition, urbanization was positively associated with secondary metabolite concentrations, but only for Q. robur. Urbanization was positively associated with air temperature for Q. robur and F. excelsior, and negatively with soil nutrients (magnesium) in the case of F. excelsior, but these abiotic variables were not associated with herbivory. Contrary to expectations, we found no evidence for indirect effects of abiotic factors via plant defences on herbivory for either Q. robur or F. excelsior. Additional biotic or abiotic drivers must therefore be accounted for to explain observed urbanization gradients in herbivory and their interspecific variation.