Experiments were conducted in 12 pentobarbital-anesthetized dogs with sinus denervation and vagotomy. The chest was opened, and the heart was replaced by a roller pump with two perfusion lines. The systemic and pulmonary circulations (SC and PC) were perfused with a constant and adjustable flow (Q). Venous outflows were directly driven by pumps without passing through a reservoir. In each closed circuit, the total blood volume remained constant because inflow and outflow were simultaneously and equally altered. In both SC and PC, arterial pressure (Pa), i.e., systemic arterial and pulmonary arterial pressures (SAP and PAP), was a positive function of Q, and venous pressure (Pv), i.e., right atrial and left atrial pressures (RAP and LAP), was a negative function of Q. The first series of experiments involved three equal step reductions in Q from baseline to zero flow. The venous-to-arterial compliance ratio (Cv/Ca) was calculated from delta Pa/delta Pv and vascular resistance (VR) from (Pa - Pv)/Q. The values of Cv/Ca in SC increased from 9.3 +/- 0.4 to 14.5 +/- 1.1 and 21.9 +/- 1.4 (P less than 0.001) in the three-step Q reduction. VR in the SC was not significantly dependent on Q. In the PC, the Cv/Ca was approximately 2.0 and VR was 0.16 mmHg.ml-1.min.kg, both values being independent of Q. Multiple-step reduction in Q for regression equations was carried out in 8 of the 12 dogs. We found that only the SAP was a linear function of Q (ml.min-1.kg-1): SAP = 18.458 + 0.953Q.(ABSTRACT TRUNCATED AT 250 WORDS)