As to the initial form of the wave which the action of the left ventricle produces in the arterial system, we get a hint from the experiments of Grashey, v. Kries, and Marey, where the sudden compression of a bag furnished the initial shock.[64] These changes of pressure can be represented by a curve like that in Fig. 1. So long as the contraction of the left ventricle lasts and the valves are open, the action of the heart produces a certain pressure in the aorta, but the influence of the intraventricular pressure is zero when the valves are closed. The second phase of the curve Fig. 1, where the pressure is zero, certainly gives the influence of the intraventricular pressure during the diastole, because there is no communication between the ventricle and the arterial system when the valves are closed. The question is whether the rest of the curve can represent the changes of the intraventricular pressure when the valves are open. [Illustration: Fig. 1. Changes of pressure produced in a bag by sudden compression.] [Illustration: Fig 2. Decreasing amount of liquid in a tube when the outflow is uniform.] The first curves of intraventricular pressure were traced by Chauveau and Marey. These experiments were made on a horse, and they have been repeated since it was discovered that they can be performed also on smaller animals. Besides Chauveau and Marey may be mentioned the names of Fick, Huerthle, v. Frey, Rolleston, Bayliss and Starling. The curves obtained by various observers belong to two types; one shows the so called "plateau," the other does not. Recent experiments have proved that this difference of results is due to a difference in methods. This also is suggested by the fact that different curves have been obtained from animals of the same species. Two methods have been applied lately for testing these curves of intraventricular pressure. The first was devised by Bayliss and Starling. It consisted chiefly in the photographic registration of the movement of the liquid in a manometer tube. The photographic registration is frictionless, and the mass of the moving liquid was so small that vibrations by inertia were fairly excluded for pressures which are not greater than the intraventricular pressure.[65] The second method was used by Porter. The idea of this method was to trace only a part of the curve, not the whole. The writing lever, thus, has in the beginning of the tracing no inertia at all, and the tracing may be overdrawn but is certainly correct in form up to the next point of inflexion of the curve.[66] These tests and the repeated experiments of Chauveau leave no doubt as to the existence of the plateau.