Mesenteric blood flow was measured in anaesthetized rats with a non-cannulating electromagnetic flow probe around the superior mesenteric artery. Reductions in blood flow were produced by intravenous bolus injections of angiotensin II (1-300 ng) and noradrenaline (3-300 ng) before and after the administration of frusemide (5 mg kg-1, i.v.). Loss of volume after frusemide was prevented by either a urinary bladder-intravenous shunt or replacement of urinary output by intravenous saline. Frusemide administration caused a small increase in baseline blood pressure of 3.2 +/- 1.3 mmHg (P less than 0.05) but did not change mesenteric blood flow. This dose of frusemide inhibited the vasoconstrictor responses to both angiotensin II and noradrenaline (P less than 0.01, two way analysis of variance). Responses to angiotensin II were inhibited to a greater extent. Acute bilateral nephrectomy or treatment with indomethacin (2 mg kg-1, i.v.) completely prevented the inhibitory effect of frusemide on the responses to angiotensin II and noradrenaline. To test whether frusemide-induced increased endogenous levels of angiotensin II may be responsible for the effects of frusemide on the vasoconstrictor responses, a separate group of rats were not given frusemide but were infused with exogenous angiotensin II (12.5-25 ng kg-1 min-1). This produced a small increase in mean blood pressure (4.0 +/- 1.4 mmHg, P less than 0.05) but did not change baseline mesenteric blood flow. Unlike frusemide, the responses to bolus injections of angiotensin II and noradrenaline were not changed by the infusion of angiotensin II. 5 It is suggested that frusemide may release directly or indirectly a prostanoid from the kidney (or a substance from the kidney which leads to the formation of a prostanoid) which inhibits constrictor responses in the peripheral vasculature.