Several studies have indicated that acquiring discriminative stimulus control for a certain anxiolytic drug influences its subsequent anti-conflict properties. To further elaborate on the question whether drug discrimination procedures affect behaviour in a conflict paradigm, a classical two-lever drug discrimination procedure was combined with an operant conflict procedure within the same animals. To this extent, rats were trained to discriminate the anxiolytic chlordiazepoxide (CDP, 30 mg/kg, po) from saline (SAL), and subsequently punished responding periods were introduced within the same session. In addition to the rats that were trained to discriminate CDP from vehicle, a group of rats was trained on a random relationship between CDP and the rewarded lever. CDP and alprazolam completely substituted for CDP, whereas mianserin did not. Responding during punished components in a session was increased by CDP and alprazolam, but not by mianserin in rats that were trained to discriminate CDP from vehicle and in randomly trained rats. The data indicate that rats can be reliably trained and tested in drug discrimination and conflict procedures within a single session and that CDP's discriminative stimulus does not alter its anti-conflict effects.