Given that a number of the techniques used to test drug abuse liability are not free from criticism, a series of oral free-choice experimental procedures was adopted. When simultaneously offered as alternatives to glucose using the classical polydipsic procedure, no preference for buprenorphine (0.025 mg/ml), morphine (0.5 mg/ml) or fentanyl (0.005 mg/ml) solutions was shown by premedicated rats. The same result was obtained when the two-bottle procedure was used for at least one month to offer etonitazene (10 micrograms/ml), buprenorphine (60 micrograms/ml), cocaine (300 micrograms/ml) and haloperidol (25 micrograms/ml) solutions as simultaneous alternatives to aspartame. This absence of preference was maintained even when the rats showed evident pharmacological effects and, in the case of the opiates, tolerance and withdrawal syndrome. However, when a gustatory marker (quinine) was introduced into one of the two solutions, preference was always shown for the other. Finally, in a conditioned taste aversion (CTA) test, etonitazene (5 or 40 micrograms/kg, i.p.) and haloperidol (0.5 or 2 mg/kg, i.p.) did not induce any reduction in saccharin consumption, while morphine (40 mg/kg) did. Pretreatment with naloxone (120 micrograms/kg, i.c.v.) did not antagonize morphine-induced CTA, while it did antagonize morphine-induced analgesia.