Male Mongolian gerbils were trained to escape electric shocks in a T-shaped maze contingent upon the presence or absence of certain drug effects (state-dependency; StD). The drug discriminative cues used were those of either delta9-tetrahydrocannabinol (THC) or pentobarbital (P-barb.) vis-à-vis the respective vehicles. Several doses of THC (0.5-16.0 mg/kg) were used and compared with P-barb. (20.0 mg/kg), a dose at which the most rapid drug discrimination occurs in the rat. When drug discrimination was established dose-time- and transfer characteristics for the training drugs were studied. Possible potentiation and antagonism was also examined in the pentobarbital trained gerbils. It was found that none of the THC doses were discriminated as rapidly as that of P-barb. Decreasing the amounts of training drug administered or increasing the injection-test intervals resulted in a decline of the number of drug associated choices. There was a maximum of 40% drug choices between THC and P-barb at the transfer tests. Mixtures of the two compounds increased the number of drug choices in an additive or even more than additive manner. Amphetamine (4.0 mg/kg) did not interact with the P-barb. induced choice responding. The analeptic drug, bemegride was found effective in antagonizing the P-barb. cued choice behavior.