A modification of the infusion test with saralasin, an angiotensin II antagonist for the detection of renin-dependent high blood pressure was studied in renal hypertensive rats and in normotensive and hypertensive subjects. Infusion was started at a rate of 0.01 microgram/kg x min saralasin and the dose was increased ten-fold at 15 min intervals. A significant fall of diastolic blood pressure was observed at the dose of 0.1 microgram/kg x min in renal hypertensive rats, in healthy subjects treated with diuretics, and in patients with renovascular hypertension (saralasin responders). Plasma concentrations of angiotensin I, angiotensin II and of saralasin as well as plasma renin activity were measured. At the lowest infusion rate of 0.01 microgram/kg x min, saralasin plasma levels were 40-fold higher than plasma angiotensin II levels. The decrease in arterial blood pressure occurred at lower doses of saralasin than the increase of plasma renin due to inhibition of feedback on the renin secreting cells. It is concluded that if the saralasin test is performed by a stepwise increase of the infusion rate, potentially dangerous complications such as hypo- or hypertensive reactions can be avoided. The diagnostic reliability is improved by such a procedure since false positive and false negative responses may be prevented. The pressor effect of saralasin in non-renin dependent patients is an advantage since it causes a more marked difference of blood pressure change between saralasin responders and non-responders.