Parasitic wasps are excellent learners. When searching for herbivorous host-insects, they make use of so-called herbivore-induced plant volatiles (HIPVs) that are emitted by the plant in response to insect feeding damage. Upon a first host encounter they learn to associate these HIPVs with the presence of the host, and subsequently increase their response- en preference level towards the learned HIPVs. During associative learning, the host insect thus serves both as the reward and as the inducer of HIPVs. Field studies have shown strong effects of associative learning on parasitoid foraging success. This holds great promise for the optimization of biological pest control. In this project we aim to boost parasitoid efficacy for the control of citrus mealybugs by developing a method for training parasitoids with mealybug-induced crop volatiles upon their release into the field/ greenhouse. To achieve this, we will first test the innate response towards crop plant HIPVs and the learning rate, and select the best performing parasitoid species. To simplify the training procedure, we will identify alternative rewards to real mealybug-hosts, as well as artificial inducers that can mimic mealybug-induced plant volatiles. Subsequently, we will further improve the learning rate of the selected parasitoid species through an artificial selection procedure. Finally, we will test the efficacy of the developed methodology and the selected parasitoids in both small- and large-scale greenhouse experiments.