The potential environmental risks associated with classical biological control arising from nontarget effects are discussed. A New Zealand research program has been designed to evaluate the effects of introducing entomophagous biological control agents on nontarget species and eventually to improve quarantine protocols for regulatory authorities to implement before approving new introductions. Using case studies, methods have been developed to assess and monitor the impact on nontarget organisms of Microctonus aethiopoides Loan and M. hyperodae Loan (Hymenoptera: Braconidae: Euphorinae), parasitoids already released in New Zealand to control the adult stages of the introduced forage pests Sitona discoideus Gyllenhal and Listronotus bonariensis (Kuschel) (Coleoptera: Curculionidae), respectively. Field surveys have shown that many species of taxonomically related native weevils occur in the same habitats and in mixed populations with the intended pest hosts. Laboratory host range tests have shown that M. aethiopoides achieved higher levels of parasitism in native species than M. hyperodae, indicating that it is a relatively more polyphagous parasitoid species. Monitoring of field populations of nontarget weevil species has supported this finding. Studies have shown that one of the unintended hosts of M. aethiopoides is Rhinocyllus conicus (Froehlich), a weed biological control agent introduced to control nodding thistle, Carduus nutans L. Some environmental, economic, and regulatory implications of nontarget effects on native and beneficial species in New Zealand are discussed.