THE hypothesis that a functional excess of dopamine underlies many of the abnormalities seen in schizophrenia is receiving much attention. The supporting evidence is principally pharmacological. There is a correlation between antipsychotic drug action and the ability of a neuroleptic drug to block dopaminergic transmission1–3. The exact site of antipsychotic drug activity is not known and evidence for the presynaptic inhibition of dopamine release4, as well as the blockade of postsynaptic receptors5,6 has been presented to explain neuroleptic drug action. Studies of binding properties of the dopamine receptor, however, provide a coherent explanation of neuroleptic drug action. Butyrophenones, phenothiazines, thioxanines and diphenyl-butylpiperadines inhibit 3H-haloperidol binding to a central nervous system (CNS) membrane fraction in exactly the order of their clinical potency7–10: the inhibition occurs at drug concentrations similar to those expected in the CNS of patients taking these drugs. These results with 3H-haloperidol, a dopamine antagonist, are used to argue that neuroleptics bind to postsynaptic dopamine receptors, and are effective by decreasing the activity of pathways using dopamine as their neurotransmitter. On the other hand, dopamine-sensitive adenylate cyclase is associated with the postsynaptic receptors5,11, yet the inhibition of this enzyme by neuroleptic butyrophenones does not correlate with in vivo or clinical potency. The difference between butyrophenone inhibition of 3H-haloperidol binding and inhibition of adenylate cyclase has been attributed to “variable degrees of coupling of dopamine receptor sites with the adenylate cyclase”9. Alternatively, there could be several dopamine sensitive adenylate cyclases with differing binding affinities and localisations : this is suggested by the presence of this enzyme in several tissue culture lines of glia12. Lesion studies show that the enzyme is not localised presynaptically in the caudate nucleus13 or in the substantia nigra14. The only defined dopaminergic receptors in the substantia nigra are localised on dopaminergic neurones15, and the elimination of these cells as the site of the enzyme was established by unilateral injections of 6-hydroxydopamine which removed these cells14. It therefore seems necessary to define more exactly the site of the dopamine binding receptor and the dopamine-stimulated adenylate cyclase. We report here that a large proportion of CNS dopamine haloperidol binding sites seem to be present on glial membranes and may be associated with an adenylate cyclase localised on these membranes.