Differential herbivory in contrasting environments is commonly explained by differences in plant traits. When several plant traits are considered, separate correlation analyses between herbivory and candidate traits are typically conducted. This makes it difficult to discern which trait best explain the herbivory patterns, or to avoid spurious inferences due to correlated characters. Aristotelia chilensis saplings sustain greater herbivory in shaded environments than in open habitats. We measured alkaloids, phenolics, trichomes, leaf thickness and water content in the same plants sampled for herbivory. We conducted a multiple regression analysis to estimate the relationship between herbivory and each plant trait accounting for the effect of correlated traits, thus identifying which trait(s) better explain(s) the differential herbivory on A. chilensis. We also estimated insect abundance in both light environments. Palatability bioassays tested whether leaf consumption by the main herbivore on A. chilensis was consistent with field herbivory patterns. Overall insect abundance was similar in open and shaded environments. While saplings in open environments had thicker leaves, lower leaf water content, and higher concentration of alkaloids and phenolics, no difference in trichome density was detected. The multiple regression analysis showed that leaf thickness was the only trait significantly associated with herbivory. Thicker leaves received less damage by herbivores. Sawfly larvae consumed more leaf tissue when fed on shade leaves. This result is consistent with field herbivory and, together with results of insect abundance, renders unlikely that the differential herbivory in A. chilensis was due to greater herbivory pressure in open habitats.