The failure of my experiment at the Academy was simply due to the fact that I spoke at too great a distance from the vibrating disk, and that my lips did not touch the sides of the mouthpiece. Some days later, at the request of several of my colleagues, I made repeated trials of the instrument, and I soon succeeded in making it speak as well as the supposed ventriloquist; but I learned at the same time that practice is necessary to ensure success. Some words are reproduced more readily than others; those which include many vowels and many _r_’s come out better than those which abound in consonants, and especially in _s_’s. It is therefore not surprising that, even in the case of an experienced manipulator like Mr. Edison’s agent, some of the sentences uttered by him are more audible than others. The simultaneous repetition of several sentences in different languages by registering one over the other is one of the most surprising effects of the phonograph. As many as three different sentences have been obtained; but in order to distinguish them through the confused sounds which result from placing one over the other, it is necessary that different persons, giving special attention to a particular sentence, should thus separate them and understand their sense. Vocal airs may, in the same way, be registered over the word tracings, and in this case it is more easy to distinguish them. There are several models of phonographs. The one represented in fig. 66 has been chiefly used for public experiments, but there is a small model, generally sold to the public, in which the cylinder is much longer, and serves at once for register and fly-wheel. This instrument gives good results, but can only be used for short sentences. In this model, as indeed in the other, the words are more easily registered by fastening a small tube in the form of a prolonged speaking-trumpet to the mouthpiece; the vibrations of the air are thus concentrated on the vibrating disk, and act with greater energy. The tenuity of the vibrating disk adds to the efficiency of the instrument, and the tracing point may be fitted directly to this disk. [Illustration: FIG. 69.] I need not describe particularly the phonograph which acts by clockwork. The instrument resembles the one represented fig. 66, except that it is mounted on a rather high table, to give room for the descent of the weight which moves the clockwork; the mechanism is applied directly to the axis of the cylinder, supplying the place of the winch, and is regulated by a small fly-wheel. The wheel used in an English system has been adopted, but we prefer that of M. Villarceau, which has small wings.