1. Renal medullary blood flow is relatively insensitive to angiotensin II (Ang II)-induced vasoconstriction, due partly to AT(1)-mediated release of nitric oxide and/or prostaglandins. AT(2)-receptor activation appears to blunt AT(1)-mediated vasodilatation within the medullary circulation. This could affect long-term efficacy of antihypertensive pharmacotherapies targeting the renin/angiotensin system, particularly in Ang II-dependent forms of hypertension. 2. We tested the effects of AT(1)- and AT(2)-receptor blockade on basal cortical and medullary laser Doppler flux (CLDF and MLDF), and on responses to renal arterial infusion of Ang II, in rats with 2 kidney, 1 clip (2K1C) hypertension and sham-operated controls. Studies were carried out in thiobutabarbital (175 mg kg(-1), i.p.) anaesthetised rats, 4 weeks after clipping, or sham surgery (n=6 in each of eight groups). 3. Candesartan (10 microg kg(-1) h(-1), intravenous (i.v.)) reduced mean arterial pressure ( approximately 17%) and increased CLDF ( approximately 24%), similarly in both sham and 2K1C rats, but did not significantly affect MLDF. PD123319 (1 mg kg(-1) h(-1), i.v.) increased basal MLDF (19%) in 2K1C but not sham rats, without significantly affecting other variables. 4. In sham rats, renal arterial infusion of Ang II (1-100 ng kg(-1) min(-1)) dose dependently decreased CLDF (up to 44%), but did not significantly affect MLDF. These effects were markedly blunted in 2K1C rats. After PD123319, Ang II dose dependently increased MLDF (up to 38%) in sham but not 2K1C rats. Candesartan abolished all effects of Ang II, including those seen after PD123319. 5. Our data indicate that AT(1) receptors mediate medullary vasodilatation, which is opposed by AT(2)-receptor activation. In 2K1C hypertension, AT(2)-receptor activation tonically constricts the medullary circulation.