The present study was designed to further investigate the pharmacological profile of the discriminative stimulus effects of gamma-hydroxybutyric acid (GHB). Drugs acting at the gamma-aminobutyric acid (GABA)B receptor (baclofen and CGP 35348), GABA(A)/benzodiazepine receptor complex (diazepam), N-methyl-D-aspartate (NMDA) receptor complex (dizocilpine), and cannabinoid receptor (WIN 55,212-2) were tested for substitution or blockade of the GHB interoceptive cue in rats trained to discriminate either 300 or 700 mg/kg of GHB i.g. from water in a T-maze, food-reinforced drug discrimination paradigm. Baclofen completely substituted for both training doses of GHB; however, its potency in substituting for GHB increased as the training dose of GHB was increased. CGP 35348 partially and completely blocked the cue elicited by 300 and 700 mg/kg of GHB, respectively. In contrast, diazepam partially substituted for 300 mg/kg of GHB, while failing to produce a GHB-appropriate response in the rat group trained to the higher GHB dose. Neither dizocilpine nor WIN 55,212-2 substituted for GHB. Collectively, these data suggest that: a) GHB produces a compound stimulus; and b) GABA(B)- and GABA(A)-mediated cues are prominent components of the mixed stimulus of GHB. However, the quality (i.e., the proportion of the component cues) of the stimulus varies as the training dose of GHB is increased; indeed, the contribution of the GABA(A)- and GABA(B)-mediated cues were smaller and greater, respectively, at 700 and 300 mg/kg of GHB training doses.