We investigated the effects of angiotensin II (Ang II), microinjected into the supraoptic (SON) and paraventricular (PVN) nuclei of rats, on the urine outflow rate and underlying mechanisms. Ang II produced antidiuretic effects in a dose-dependent manner with ED50 values of 0.1 and 0.05 nmol in the SON and PVN, respectively. [Sar1, Ile8]Ang II at 0.1 nmol diminished the Ang II (0.5 nmol)-induced antidiureses in the SON more markedly than in the PVN. A high dose of [Sar1,Ile8]Ang II, 1 nmol, completely inhibited the effects in both the nuclei. In addition, the Ang II (1 nmol)-induced antidiuretic effects were partially inhibited by phenoxybenzamine (80 nmol) in the SON and by phenoxybenzamine, timolol (100 nmol) and propranolol (100 nmol) in the PVN. The microinjection of Ang II (1 nmol) into both the nuclei, after pretreatment with a vasopressin V1V2-antagonist, d(CH2)5-D-Tyr(Et)VAVP (i.v.) significantly increased the urine outflow rate. These findings suggest that 1) Two mechanisms account for the Ang II receptor-mediated antidiureses resulting from an increase in vasopressin release: direct stimulation on vasopressin-containing neurons and indirect stimulation on them through alpha-adrenoceptors in the SON and alpha- and beta-adrenoceptors in the PVN; 2) The Ang II-induced antidiuretic effect in the SON is slightly less potent than that in the PVN; and 3) Ang II receptors in the nuclei may possibly produce the diureses through mechanisms that are not presently understood.