Previous studies indicate that the discriminative stimulus effects of cocaine are mediated predominantly by indirect activation of dopamine (DA) D2 postsynaptic receptors, although DA D1 receptors may also be involved. In the present study, full or partial D2 agonists and D2 antagonists were tested for their ability to substitute for, potentiate or antagonize the stimulus effects of cocaine in rats (n = 15) trained to discriminate cocaine (10 mg/kg) from saline in a two-lever, water-reinforced task. The full D2 agonists bromocriptine (1.25-20 mg/kg) and quinpirole (0.013-0.2 mg/kg) engendered substantial cocaine-lever responding (> 80% drug-lever responding), whereas the partial D2 agonists preclamol (2.5-10 mg/kg) and terguride (0.313-1.25 mg/kg) produced less than 50% cocaine-lever responding. Co-administration of a threshold dose of cocaine (1.25 mg/kg) with low doses of bromocriptine (1.25-5 mg/kg) or quinpirole (0.025-0.1 mg/kg) induced higher percentages of cocaine-lever responding as compared with occasions when these D2 agonists were given alone. However, co-administration of this dose of cocaine with preclamol (2.5-10 mg/kg) or terguride (0.313-1.25 mg/kg) did not alter the percentage of cocaine-lever responding observed when these partial D2 agonists were administered alone. Pretreatment with the D2 antagonists bromuride (0.25-1 mg/kg) and haloperidol (0.125-0.5 mg/kg) significantly reduced the percentage of cocaine-lever responding. Preclamol (0.625-10 mg/kg) and terguride (0.019-5 mg/kg), but not bromocriptine (2.5-20 mg/kg) or quinpirole (0.01-0.08 mg/kg), significantly reduced the percentage of cocaine-lever responding. These results suggest that full D2 agonists substitute completely for cocaine, whereas partial D2 agonists do not produce cocaine-like responding.(ABSTRACT TRUNCATED AT 250 WORDS)