Theoretically, induced defenses should be prevalent within low resource environments like the forest understory where constitutive defenses would be costly. Also, the induced response should be stronger when the herbivore is a generalist rather than a specialist, which often have mechanisms to avoid or overcome plant defenses. These ideas have been previously tested for herbaceous species, and we examined these predictions in Lindera benzoin (spicebush), a common woody shrub of the eastern deciduous forest. Lindera benzoin plants in contrasting light environments served as control plants or were subjected to one of four treatments: application of jasmonic acid, clipping, herbivory by the specialist Epimecis hortaria (tulip tree beauty) and herbivory by the generalist Spodoptera exigua (beet armyworm). Following treatment, we assessed induced responses by measuring leaf chemistry (C/N ratio, protein content, and peroxidase activity), and by using insect bioassays with E. hortaria larvae. We found no difference in peroxidase activity between light environments in controls, plants treated with clipping or jasmonic acid. In plants subject to insect herbivory, peroxidase activity was greater in shade plants than in sun plants. The magnitude of this increase in the shade varied between the herbivores, with a 32 fold increase in plants exposed to the generalist S. exigua and a 9 fold increase in plants exposed to the specialist E. hortaria. Leaves from shade plants had more protein and lower C/N ratios than leaves from sun plants, regardless of induction treatment. In control plants, E. hortaria larvae consumed more leaf biomass and achieved greater final weights in the sun than in the shade, but these differences disappeared with induction treatments were applied. These results are among the first to show rapid induction in a woody plant, and different levels of induction with light environments and with specialist versus generalist herbivores.