Furosemide (frusemide)-induced renal medullary hypoperfusion provides a model for studies of the dependence of local circulation on tissue tonicity. We examined the role of medullary prostaglandins (PG) and adenosine (Ado) as possible mediators of the response to furosemide. Furosemide was infused i.v. at 0.25 mg kg(-1) h(-1) in anaesthetized rats, untreated or treated with intramedullary indomethacin (Indo) or Ado. An integrated set-up was used to measure renal medullary laser-Doppler flux (MBF) and medullary ionic tonicity (electrical admittance, Y), and to infuse Indo and Ado directly into the medulla. The cortical flux was measured on kidney surface. The excretion of water, sodium and total solute was also determined. Intramedullary Indo (1 mg kg(-1) h(-1)) decreased MBF 18 +/- 5% and increased tissue Y 14 +/- 3% (both significant); the treatment abolished the post-furosemide decrease in MBF (-22% in untreated group) and enhanced slightly the increase in renal excretion. Intramedullary Ado (5 mg kg(-1) h(-1)) did not change baseline MBF or Y; the post-furosemide decreases in MBF (-22%) and Y, and the increase in renal excretion were preserved. We conclude that a decrease in intramedullary PG activity secondary to decreased medullary hypertonicity mediates the fall in medullary perfusion in response to furosemide; the hypoperfusion may help restore the initial tonicity. Together with the earlier evidence on the dependence of post-furosemide medullary hypoperfusion on angiotensin II, the study exposes its interaction with PG in the control of medullary circulation. Adenosine is not involved in medullary vascular responses to decreased tissue hypertonicity.