The relative merits of adding a "pulsatile" component to flow during cardiopulmonary bypass (CPB) has long generated controversy, the resolution of which has been hampered by lack of quantification of the "pulsatility" delivered by different devices. The present experimental series had two goals: to quantify the "pulsatility" of blood flow during CPB in terms of pulse rate and pulsatility index (PI) and to examine which aspects of a "pulsed flow" provide clinical benefits. A flow waveform can be expressed in terms of its baseline rate and its PI, the sum of the square of its harmonics components divided by the square of the mean flow. We used PI to quantify the pulsatility of blood flow in the descending thoracic aorta and used changes in the serum lactate level as an indication of end organ flow. In one experimental series seven adult mongrel dogs were placed on roller pump CPB at a constant flow of 100 ml/kg/min. After a 20-minute stabilization period a roller pump wave and three different pulse shapes (generated by a computer-controlled hydraulic pump) were evaluated for 15 minutes each. The pulse wave shapes were graded, with C being the sharpest and A the least sharp. In a second series six other dogs were placed on CPB and were subjected to roller pump perfusion and three pulse waves of identical shape but at different rates. The results indicated that a combination of a minimum PI of 1.88 and a minimum rate of 80 bpm were necessary to significantly reduce lactate production as compared with roller pump perfusion. Thus the same mean flow can have very different physiologic effects depending on how it is delivered. This quantification method permits direct comparison of different "pulsatile waveforms" and provides a means for identification of optimal pulsatile flow.