The application of Bernoulli's law allows the non-invasive estimation of pressure drops across stenotic valves from Doppler velocity measurements. However, several assumptions are to be made, which influence the accuracy of the pressure drop estimation. Energy losses, non-uniform velocity profiles, pressure recovery, unsteady flow and omission of the upstream velocity affect accuracy and are critically reviewed. In vitro experiments, published in the literature, show good correlations between estimated and actual pressure drop. In only one of these studies does the data allow a comparison between theory and practice and to study the relationship between the pressure drop at the inlet of the obstruction and the Doppler velocity measurement. It appears that this relationship is not completely described by the simplified Bernoulli equation. In vivo verifications of Bernoulli's law show favourable correlations. However, the expected differences between peak pressure drops measured by cardiac catheterization and the pressure drops estimated by Doppler echocardiography may be as high as 25 mmHg.