This study was designed to compare the relative merits of soft and rigid artificial ventricles. A cascade mock circulation was used to measure cardiac output under different circumstances. The data show that these soft air driven ventricles show a Starling's-like response over a wider range of filling pressures than identical, but rigid, ventricles. Compression of soft ventricles by high intrathoracic pressures was simulated in vitro. Air pressures up to + 20 mm Hg did not seriously affect soft ventricles. Cardiac tamponade was simulated by compressing the ventricle in a closed fluid compartment. Tamponade became severe when volume reduction of the ventricle rose to 60 ml. Hemolysis caused by soft and rigid ventrides was tested in a blood bag set-up and was ---I _