Abstract To test the validity of a theoretical model of aortic valve closure, based upon the observations in a two-dimensional analogue, the effect of some hemodynamic factors on aortic valve behaviour was studied in open-chest dogs. Direct cinematography was used to record aortic valve movements. The ECG, instantaneous ascending aortic blood flow as well as left ventricular and aortic pressures were registered simultaneously. The experiments revealed that: (1) at higher stroke volumes complete valve opening was reached earlier in the cardiac cycle; (2) the higher the peak aortic flow the better a circular shape of the completely opened valve was approximated; (3) the higher the systolic aortic pressure the earlier valve closure started in the cardiac cycle; (4) initial valve closure during flow deceleration was faster in the case of a larger systolic aortic pressure drop; (5) the backflow in the aorta increased at a larger end-systolic valve orifice area; (6) a rise in heart rate does not affect the mechanism of valve closure; (7) valve behaviour was not affected significantly by fluid viscosity. The closing behaviour of the valve as determined in animal experiments under different hemodynamic circumstances showed reasonable agreement with the valve closure as predicted by theory. However, due to the simplifications assumed in the model, the theoretical description must be used tentatively, especially with large pressure changes within the valve orifice during the cardiac cycle, with low peak flows or with high Strouhal numbers.