Injection of carbachol into an indwelling third ventricular cannula evoked milk ejection in anesthetized lactating rats as measured by increased intraductile mammary pressure. There was no alteration in blood pressure. Similar doses were ineffective when injected systemically, either intravenously or into the carotid artery. Larger doses of carbachol given intravenously evoked milk ejection, probably via a direct effect on the mammary myoepithelium. The effect of intraventricular injection of carbachol was blocked by prior injection of atropine into the third ventricle. These results were interpreted to mean that a cholinergic synapse lies in the central pathways controlling oxytocin release. Intraventricular carbachol appeared to be less effective in evoking oxytocin release in hydrated rats. In the hydrated animals, intraventricular carbachol evoked an antidiuresis accompanied by increased concentrations of Na+ and K+, indicative of vasopressin release. Antidiuresis frequently occurred in the absence of milk ejection, which indicates that the threshold for vasopressin release to this stimulus is sometimes less than that for oxytocin release. An increase in excretion of sodium and an increase in the Na+/K+ excretion ratio also occurred after intraventricular carbachol, but were not mimicked by injection of vasopressin or oxytocin.