Variation in herbivore pressure has often been predicted from patterns in plant traits considered as antiherbivore defences. Here, we tested whether spatial variation in field insect herbivory is associated with the variation in plant quality by conducting a meta-analysis of 223 correlation coefficients between herbivory levels and the expression of selected plant traits. We found no overall correlation between herbivory and either concentrations of plant secondary metabolites or values of physical leaf traits. This result was due to both the large number of low correlations and the opposing directions of high correlations in individual studies. Field herbivory demonstrated a significant association only with nitrogen: herbivore pressure increased with an increase in nitrogen concentration in plant tissues. Thus, our meta-analysis does not support either theoretical prediction, i.e., that plants possess high antiherbivore defences in localities with high herbivore pressure or that herbivory is low in localities where plant defences are high. We conclude that information about putative plant defences is insufficient to predict plant losses to insects in field conditions and that the only bottom-up factor shaping spatial variation in insect herbivory is plant nutritive value. Our findings stress the need to improve a theory linking plant putative defences and herbivory.