1. If an armature of soft iron is applied to the poles of an electro-magnet, with its two branches firmly fixed on a board, and if pieces of paper are inserted between this armature and the magnetic poles, so as to obviate the effects of condensed magnetism; if, finally, this electro-magnet is connected with a speaking microphone, of the form given in fig. 39, it is possible to hear the words spoken in the microphone on the board which supports the electro-magnet. 2. If two electro-magnets are placed in communication with a microphone, with their poles of contrary signs opposite to each other, and if their poles are separated by pieces of paper, speech will be distinctly reproduced, without employing armature or diaphragm. These experiments are, however, delicate, and demand a practised ear. 3. If, instead of causing the current produced by a microphone to pass through the helix of a receiving telephone, it is sent directly into the bar magnet of this telephone in the direction of its axis--that is, from one pole to another--the words pronounced in the microphone may be distinctly heard. This experiment by M. Paul Roy indicates, if it is exact, that the electric pulsations which traverse a magnet longitudinally will modify its magnetic intensity. The experiment, however, demands verification. Another point was obscure. It was important to know whether the diaphragm of a telephone really vibrates, or at least if its vibrations could involve its displacement, such as occurs in an electric vibrator, or in wind instruments which vibrate with a current of air. M. Antoine Bréguet has made some interesting experiments on the subject, which show that such a movement cannot take place, since speech was reproduced with great distinctness from telephones with vibrating plates of various degrees of thickness, and he carried the experiment so far as to employ plates fifteen centimètres in thickness.[10] When pieces of wood, caoutchouc, and other substances were laid upon these thick plates, the results were the same. In this case it cannot be supposed that the plates were moved to and fro. I have moreover ascertained, by placing a layer of water or of mercury on these plates, and even on thin diaphragms, that no sensible movement took place, at least when the induced currents produced by the action of speaking were used as the electric source. No ripples could be seen on the surface of the liquid, even when luminous reflectors were employed to detect them. And indeed it can hardly be admitted that a current not more intense than that of a Daniell element, which has traversed 10,000,000 kilomètres of telegraphic wire--a current which can only show deviation on a Thomson galvanometer--should be powerful enough to make an iron plate as tightly stretched as that of a telephone vibrate by attraction, even if we grant that the current was produced by laying a finger on the diaphragm.