I n 1832, Rouanetl used a model of excised valves to demonstrate that the second heart sound may have its origin in valvular closure. Since that time a wide variety of other mechanisms have been proposed to account for the genesis of the heart sounds.2-17 Although within recent years the valvular origin of the first and second sounds has become generally accepted, the actual mechanisms which produce the sounds remain a matter of discussion. According to Wiskind and Talbot,‘* proposed theories still “do not account for the build-up of pressure and momentum which must precede sound production.” We have carried out studies on the production of sound in the left outflow tract and semilunar valves that may account for these requisites of sound production. The semilunar valves and a short section of the aorta were excised from the fresh hearts of steers and dogs. A Lucite cannula was tied into the thoracic aorta so that air could be blown retrograde into the aorta and cause the valves to snap shut. The action of the valves and the resulting sounds were recorded and photographed at 64 frames per second. Consecutive frames of a closing valve are shown in Fig. 1, and recorded sounds are shown in Fig. 2. Fig. 2 demonstrates in sequence the two types of vibration generated : (1) continuous vibrations of the valve and walls as fluid passes through the leaflets (as in murmurs), and (2) intense vibrations of short duration associated with valvular closure. When the valves were amputated or were held open by sutures, only the first type of vibration was observed. AIthough air rather than liquid was used, observation of the preparation showed that during the process of closure the valve leaflets ballooned out from the line of attachment to the wall, rolled together with increasing momentum, and then closed; the corpora arantii were the last points of apposition. Valvular closure vibrations did not appear until the closure was complete. The rolling together of the valves is demonstrated in Fig. 1. Since the cusps billow together with increasing momentun~, the corpora arantii come into apposition with greater force than do other portions of the coapting leaflets. It is the sudden snapping together of these corpora that provides sufficient energy to vibrate the tensed leaflets, much like a sounding board. Our experiments showed that malfunction of one leaflet reduced the vibration of cIosure during the apposition of the two unaltered leaflets. This may be related to the fact that the remaining leaflets are not able to attain as great a closing velocity as the cusps in the “unaltered” valve. The theory that the vibration is generated by the development of tension in