1. Induced plant responses can affect herbivores either directly, by reducing herbivore development, or indirectly, by affecting the performance of natural enemies. Both the direct and indirect impacts of induction on herbivore and parasitoid success were evaluated in a common experimental system, using clonal poplar trees Populus nigra (Salicales: Salicaceae), the gypsy moth Lymantria dispar (L.) (Lepidoptera: Lymantriidae), and the gregarious parasitoid Glyptapanteles flavicoxis (Marsh) (Hymenoptera: Braconidae).
2. Female parasitoids were attracted to leaf odours from both damaged and undamaged trees, however herbivore-damaged leaves were three times more attractive to wasps than undamaged leaves. Parasitoids were also attracted to herbivore larvae reared on foliage and to larval frass, but they were not attracted to larvae reared on artificial diet.
3. Prior gypsy moth feeding elicited a systemic plant response that retarded the growth rate, feeding, and survival of gypsy moth larvae, however induction also reduced the developmental success of the parasitoid.
4. The mean number of parasitoid progeny emerging from hosts fed foliage from induced trees was 40% less than from uninduced trees. In addition, the proportion of parasitised larvae that survived long enough to issue any parasitoids was lower on foliage from induced trees.
5. A conceptual and analytical model is provided to describe the net impacts of induced plant responses on parasitoids, and implications for tritrophic interactions and biological control of insect pests are discussed.