trans-10,11-Dihydroxy-5,6,6a,7,8,12b-hexahydrobenzo[a]phenan thridine (4a, dihydrexidine) has been found to be a highly potent and selective agonist of the dopamine D1 receptor in rat brain. Dihydrexidine had an EC50 of approximately 70 nM in activating dopamine-sensitive rat striatal adenylate cyclase and a maximal stimulation equal to or slightly greater than that produced by dopamine. Dihydrexidine had an IC50 of 12 nM in competing for [3H]SCH23390 (1a) binding sites in rat striatal homogenate, and of 120 nM versus [3H]spiperone. These data demonstrate that dihydroxidine has about ten-fold selectivity for D1/D2 receptors. More importantly, however, is the fact that dihydrexidine is a full agonist. Previously available agents, such as SKF38393 (1b), while being somewhat more selective for the D1 receptor, are only partial agonists. The isomeric cis-dihydroxybenzo[a]-phenanthridine neither stimulated cAMP synthesis nor inhibited the cAMP synthesis induced by dopamine. The cis isomer also lacked appreciable affinity for [3H]-1a binding sites. N-Methylation of the title compound decreased affinity for D1 sites about 7-8-fold and markedly decreased ability to stimulate adenylate cyclase. Addition of an N-n-propyl group reduced affinity for D1 sites by about 50-fold and essentially abolished the ability to stimulate adenylate cyclase. However, this latter derivative had twice the affinity of the D2-selective agonist quinpirole for the D2 receptor. The results are discussed in the context of a conceptual model for the agonist state of the D1 receptor.