Abstract The mechanisms of the heart sounds have been investigated by means of animal experiments and the recording of tracings in normal human subjects. Simultaneous tracings of pressure have been recorded in dogs from the two ventricles, from one ventricle and its respective atrium, and from the aorta and pulmonary artery. The time relationship of the motion of the various valves was revealed by these tracings plus simultaneous external phonocardiograms. Simultaneous intracardiac phonocardiograms confirmed these data. Selective phonocardiograms have been recorded through various filters in man, together with low-frequency tracings of the precordium, carotid and aortic tracings, and jugular tracings. The following conclusions have been reached (Fig. 11): 1. 1. The first sound is made of 3 phases: low-pitched beginning, due to 2. myocardial tension; higher-pitched central phase, due to valvular events; lowpitched final phase, due to vascular phenomena. The central phase contains at least 4 vibrations which correspond to the motion of the 4 valves in the following order: mitral closure, tricuspid closure, pulmonic opening, aortic opening. Whenever one or more large vibrations correspond to the phase of ejection, they are likely to be of a vascular nature. 3. 2. The second sound is made of 3 phases: low-pitched beginning, due to eddies preceding the valvular closure; higher-pitched central phase, due to closure of the semilunar valves; and low-pitched final phase, due to final vibrations plus opening of the A-V valves. The central and final phases correspond, firt, to the closure of the semilunar valves and, then, to the opening of the A-V valves, in this order: aortic closure, pulmonic closure, tricuspid opening, mitral opening. 4. 3. Rapid filling of the ventricles is usually nonsimultaneous, in this order: rapid filling of right ventricle, rapid filling of left ventricle. These data confirm some interpretations of previous workers but exclude others. Time intervals for the various phases are given for large dogs. Measurements taken in phonocardiograms of normal persons, recorded with a technique which is identical to that used in dogs, indicate that similar figures can be used in man.