The role of opioid systems in the anticonflict effect of chlordiazepoxide, diazepam and pentobarbital was evaluated with a modified Vogel procedure. First, morphine, ineffective by itself, was combined with subeffective or marginally effective doses of the benzodiazepines in order to detect possible potentiation. However, the combined treatment reduced licking in the Vogel procedure as well as in a licking test where no shock was administered. Several doses of the benzodiazepines and pentobarbital were then administered in combination with several doses of the opiate antagonist naloxone. A dose-dependent inhibition of anticonflict effect was obtained. In an additional experiment, it was shown that naloxone blocked the effects of diazepam in the elevated plus-maze procedure. Motor deficiencies, as evaluated with a rotarod test, produced by the benzodiazepines and pentobarbital could not be antagonized by naloxone. It is concluded that opioids are important for the anticonflict but not for the motor effects of these drugs. An analysis of published studies concerning the interaction of opioids and benzodiazepines in several procedures supposed to reflect anxiolytic effects shows that the inhibition obtained with naloxone is reliable and not procedure specific. The mechanisms by which opiate antagonists produce this inhibition of anticonflict activity are not known. It is tentatively suggested that opioid activation associated with stress may be a necessary component of anxiolysis.