Three separate groups of rats were trained to discriminate the putative D3 dopamine receptor agonists (±)-7-hydroxy-dipropylaminotetralin (7-OH-DPAT) (0.03 mg/kg), PD 128,907 (1.0 mg/kg) and quinpirole (0.03 mg/kg) from saline. Food was presented after each 10 (7-OH-DPAT and PD 128,907) or 20 (quinpirole) consecutive responses on one lever after administration of the training drug, and the other lever after the administration of saline. Once stable performances were obtained, the effects of various doses of several dopaminergic agonists were assessed during test sessions in which responses on either lever were reinforced. The substitution tests were conducted to determine if differences in potencies would be obtained, which would be suggestive of differences in the mechanisms underlying the discriminative effects of the training drugs. Non-selective agonists with activity at both D2 and D3 dopamine receptors (D2-like agonists) substituted for each of the three training drugs. In addition, the selective D2 dopamine receptor agonist U91356A also generalized to both 7-OH-DPAT and PD 128,907. The potencies of the D2-like agonists in substituting for each training drug were highly correlated with potencies in substituting for the others. SKF 82958 and SKF 81297, agonists with selectivity for D1 and D5 dopamine receptors (D1-like agonists), partially substituted for 7-OH-DPAT but not PD 128,907. The D1-like partial agonist SKF 38393 did not substitute for any of the training drugs for which it was tested. Cocaine produced intermediate substitution in 7-OH-DPAT- and PD 128,907-trained subjects and did not substitute at all in quinpirole-trained subjects. The dopamine D1-like antagonist SCH 39166 (0.001–0.03 mg/kg) did not alter the discriminative stimulus effects of PD 128,907, whereas the D2-like dopamine antagonist spiperone (0.001–0.1 mg/kg) produced at the highest dose an insurmountable antagonism of the discriminative effects of PD 128,907. In contrast, there was no appreciable antagonism of the effects of PD 128,907 on response rates. The data collected are consistent with a distinction between the effects of each of these training drugs and the indirectly acting agonist cocaine. Further, these data indicate that there are differences in the mechanisms underlying the discriminative effects of PD 128,907 and its effects on response rates. Moreover, these data indicate that each of the training drugs is distinct from drugs acting through D1 dopaminergic mechanisms. However, there were no data that clearly distinguished these training drugs from each other or from drugs acting through D2 dopaminergic mechanisms.