Hemolytic and subhemolytic blood damage by mechanical heart valve prostheses have been observed in both clinical and in vitro investigations. A direct comparison between these studies is not possible. Nevertheless the transfer of some in vitro results to the behaviour of the valve in situ may be performed considering the similarity principle. This requires the use of dimensionless similarity numbers such as the plasma's hemoglobin concentration (PHb) or others, instead of dimensioned parameters. To evaluate the in vitro hemolysis of valve prosthesis a test chamber filled with human banked blood was used. An artificial ventricle ensuring an oscillatory flow through the valve was also used. The rise of PHb was evaluated in terms of a similarity number, called the lysis number. This number describes the probability of destroying a single red blood cell participating once in the hemolytic process under consideration. The lysis number, a Björk-Shiley valve (TAD 29), was found to be in the order of 2 x 10(-4). From this, the survival time of erythrocytes in patients with an artificial heart valve was estimated. It was found to be in the order of 20 d of T50 Cr in agreement with clinical results.