Suppose that a very small carbon rod with pointed ends is placed at the bottom of a box, of about the size of a watch. One end of the rod rests against a morsel of carbon, which is fastened to a very thin steel diaphragm, placed before a mouthpiece which acts as a lid to the box, and is screwed above it. Next suppose that a small piece of paper, folded in two, in the shape of the letter V, is fixed above that part of the carbon in contact with the carbon of the diaphragm. This constitutes the instrument, and in order to work it, it must be held in a vertical position before the mouth, at a distance of about three centimètres, and it is necessary to speak in the ordinary tone. If the telephone is placed in direct communication with this instrument, it will send the voice to a distance. Without employing a Leclanché cell, the voice may be heard at the distance of ten yards, if one of the carbons used for the phonograph is placed before the mouthpiece of the telephone. In this system, the sensitiveness of the instrument is entirely due to the slightness of the contact between the two carbons, and the slight elasticity of the folded paper constitutes the contact. Perhaps the paper itself has some influence; at any rate the most delicate spiral spring is incapable of producing the same effect, and it is necessary to suspend the instrument vertically, in order that the weight of the moveable carbon may not affect it. It can be regulated by depressing or elevating that part of the paper which rests on the carbon rod. Although it is possible to work all telephones with this instrument, some are more effective than others. The mouthpiece must be concave, and the diaphragm must be close to its rim, and must be made of a particular kind of tin. The ordinary diaphragm does not act well, and M. Boudet de Paris has tried several, so as to obtain the maximum effect. It is certain that when the instruments are as well regulated as those which the inventor has deposited with me, their results are really surprising. It is even possible, by using several microphones at the sending station, to obtain the reproduction of duets, and even of trios, with remarkable effect. With this kind of microphone speaker M. Boudet de Paris is able to transmit speech into a snuff-box telephone, merely consisting of a flat helix of wire, placed before a slightly magnetised steel plate, and without insertion of a magnetic core. A single Leclanché cell was enough. An experiment of the same nature was tried in England, but it was found necessary to use six Leclanché cells. [Illustration: FIG. 43] [Illustration: FIG. 44.]