1. It has been suggested that an ideal antiparkinsonian treatment requires stimulation of both D1 and D2 dopamine receptors. Bromocriptine and lisuride are regarded as pure D2 receptor agonists, whereas pergolide and apomorphine are thought to stimulate both D1 and D2 receptors. 2. The aim of this study was to compare the affinities of bromocriptine, lisuride, pergolide, and apomorphine for the D1, D2, and D3 receptors in postmortem human striatum. The dissociation constants (Ki values) of the dopamine agonists were determined from competition binding experiments with selective radioligands. 3. The Ki values of the orally administered agonists--bromocriptine, pergolide, and lisuride--for the D2 receptors were proportional to their optimal doses against parkinsonism. Ki(D1)/Ki(D2) ratios were 23 for lisuride, 67 for pergolide, 60 for bromocriptine, and 2.6 for apomorphine. Ki(D3)/Ki(D2) ratios were 0.4 for lisuride, 1 for pergolide, 5.4 for bromocriptine, and 21 for apomorphine. 4. The present results support the hypothesis that the antiparkinsonian effect of dopamine agonists is mediated primarily by D2 receptors. Apomorphine is a mixed D1/D2 agonist, but pergolide has no more D1 agonist properties than bromocriptine and lisuride. The role of the D3 receptors is unknown, but their activation might either be associated with the generation of psychiatric side-effects or dyskinesias, or alternatively add to antiparkinsonian activity.