Pigeons were trained to discriminate IM injections of 1.0 mg/kg d-amphetamine from water, 5.6 mg/kg pentobarbital from water, or 1.0 mg/kg d-amphetamine from 5.6 mg/kg pentobarbital by requiring them to peck different response keys depending on which substance was administered prior to the session. Excellent stimulus control was achieved under all conditions with close to 100% of the responses occurring on the injection-correlated key. In tests with doses different from those used in training, the percentage of responses on the drug key was directly related to drug dose. When d-amphetamine was given to birds trained to discriminate pentobarbital from water or when pentobarbital was given to birds trained to discriminate d-amphetamine from water, responding occurred predominately on the water-correlated key. d-Amphetamine produced a dose-related antagonism of the effects of pentobarbital for birds trained to discriminate pentobarbital from water or from d-amphetamine. Rates of responding on the drug key were generally highes after administration of the drug doses used in discrimination training; but response rates were not systematically related to the percentage of responses occurring on the drug key. All birds were subsequently trained to discriminate a combination of 1.0 mg/kg d-amphetamine and 5.6 mg/kg pentobarbital from either 1.0 mg/kg d-amphetamine or 5.6 mg/kg pentobarbital alone, demonstrating that the discriminative stimulus properties of amphetamine-pentobarbital combinations are different from either drug alone. Several of the drug effects reported were related to the drug discrimination that had been established.