The present study used a drug discrimination paradigm to characterize the contribution of separate receptor systems to the stimulus effects of different training doses of ethanol. In a two-lever drug discrimination paradigm two groups of adult male Long-Evans rats (n = 8 per group) were trained to discriminate either 1.0 g/kg ethanol from water or 2.0 g/kg ethanol from water, administered intragastrically (i.g.), 30 min prior to the start of daily sessions in which responding was maintained under a fixed ratio 20 schedule of food presentation. Following training, cumulative dosing substitution tests were conducted with the GABAA positive modulator pentobarbital (1-17 mg/kg, i.p.), the uncompetitive NMDA antagonist dizocilpine (0.01-0.3 mg/kg, i.p.) and the 5-HT1B/2C agonist m-trifluoromethylphenylpiperazine (TFMPP 0.17-1.7 mg/kg, i.p.). Next, the rats initially trained at 1.0 g/kg ethanol were retrained to discriminate 2.0 g/kg ethanol from water, and the rats initially trained at 1.0 g/kg were retrained to discriminate 2.0 g/kg ethanol from water. Both groups were then re-tested with the same ligands. Regardless of training history, animals currently discriminating 1.0 g/kg were more sensitive to the ethanol-like effects of TFMPP and pentobarbital compared to rats discriminating 2.0 g/kg ethanol. However, no difference in sensitivity to the ethanol-like effects of dizocilpine based on ethanol training dose was detected. These results support the view that ethanol is a heterogeneous discriminative stimulus comprised of GABAA, NMDA and 5-HT1B/2C receptor-mediated activity. Furthermore, changes in sensitivity to GABAA and 5-HT ligands as a function of training dose could be indicative of overshadowing by other components of ethanol's heterogeneous cue. Finally, it appears that the current profile of ethanol's heterogeneous stimulus effects, rather than an interaction with ethanol training history, determines the substitution pattern of specific receptor ligands.