The benzofurane (+)-S 14297, the benzamide nafadotride, the aminoindane U 99194 and the arylpiperazine GR 103,691 have been proposed as "selective" antagonists at dopamine D3 vs. D2 receptors. Herein, we compared their in vitro affinities and in vivo actions to those of the aminotetralin D3 antagonists (+)-AJ 76 and (+)-UH 232. Affinities at recombinant, human (h)D3 and/or hD2 sites were determined by employing the mixed D2/D3 antagonist [125I]-iodosulpride and the preferential D3 ligands [3H]-(+)-PD 128, 907 and [3H]-(+)-S 14297. [3H]-(+)-PD 128,907, [3H]-(+)-S 14297 and [125I]-iodosulpride yielded an essentially identical pattern of displacement at D3 sites, which suggests that they recognize the same population of receptors. The rank order of potency (Ki values in nM vs. [3H]-(+)-PD 128,907) was GR 103,691 (0.4) approximately nafadotride (0.5) > haloperidol (2) approximately (+)-UH 232 (3) approximately (+)-S 14297 (5) > (+)-AJ 76 (26) > U 99194 (160). The rank order of preference (Ki ratio, D2D3) for D3 receptors (labeled by [3H]-PD 128,907) vs. D2 sites (labeled by [125I]-iodosulpride) was (+)-S 14297 (61) approximately GR 103,691 (60) > U 99194 (14) > nafadotride (9) approximately (+)-UH 232 (8) approximately (+)-AJ 76 (6) > haloperidol (0.2). (+)-S 14297 and GR 103,691 also showed greater than 100-fold selectivity at dopamine hD3 vs. hD4 and hD1 sites. However, GR 103,691 showed marked affinity for serotonin1A receptors (5.8 nM) and alpha-1 adrenoceptors (12.6 nM). In vivo, all antagonists except GR 103,691 prevented the induction of hypothermia by (+)-PD 128,907 (0.63 mg/kg s.c.) and a further preferential D3 agonist, (+)-7-OH-DPAT (0.16 mg/kg s.c.). On the other hand, haloperidol, (+)-AJ 76, (+)-UH 232 and nafadotride all induced catalepsy in rats, whereas (+)-S 14297, U 99194 and GR 103,691 were inactive. Haloperidol, (+)-AJ 76, (+)-UH 232, nafadotride and (weakly) U 99194 also enhanced prolactin secretion and striatal dopamine synthesis, whereas (+)-S 14297 and GR 103,691 were inactive. However, despite its high affinity at 5-HT1A receptors and alpha-1 adrenoceptors, both of which are present on raphe-localized serotonergic neurons, GR 103,691 (0.5 mg/kg i.v.) failed to influence their basal firing rate or the inhibition of their electrical activity by the 5-HT1A agonist (+/-)-8-OH-DPAT (0.005 mg/kg i.v.), a result that casts doubt on its activity in vivo. In conclusion, both (+)-S 14297 and GR 103,691 are markedly selective ligands that permit the characterization of actions at hD3 vs. hD2 receptors in vitro, but (+)-S 14297 appears to be of greater utility for the evaluation of their functional significance in vivo. Nevertheless, to develop a better understanding of the respective roles of dopamine D3 and D2 receptors, we need additional, chemically diverse antagonists of improved potency and selectivity.