The theory of the electric transmission of sounds in electro-magnetic telephones is somewhat complex. It has been seen that they can be obtained from diaphragms of non-magnetic substance, and even from simple mechanical vibrations produced by shocks. Are we to ascribe them in the first case to the inductive reaction of the magnet on the vibrating plate, and in the second case to the movements of magnetic particles before the spirals of the helix? The matter is still very obscure; yet it is conceivable that the modifications of the inducing action of the magnet on the vibrating diaphragm may involve variations in the magnetic intensity, just as we can admit an effect of the same kind due to the approach and withdrawal of the magnetic particles of the spirals of the helix; M. Trève, however, believes that there is in the latter case a special action, which he has already had occasion to study under other circumstances, and he sees in the current thus caused the effect of the transformation of the mechanical labour produced amidst the magnetic molecules. The question is complicated by the fact that these effects are often produced by purely mechanical transmissions. There is another point to consider, on which Colonel Navez has made some interesting remarks; that is, whether the effects in the receiver are stronger with permanent than with temporary magnets. In the first model of the telephone, exhibited by Mr. Bell at Philadelphia, the receiver was, as I have said, made of a tubular electro-magnet, furnished with a vibrating disk at its cylindrical pole; but this arrangement was abandoned by Mr. Bell, with the object, as he states in his paper, of rendering his instrument both a receiver and a sender.[13] Yet Colonel Navez maintains that the magnet plays an important part, and is even indispensable under the present conditions of its form. ‘It is possible,’ he says, ‘under certain circumstances, and by making the instrument in a special way, to make a Bell receiver speak without a permanent magnet, yet with an instrument of the usual construction the sound ceases when the magnet is withdrawn and replaced by a cylinder of soft iron. In order to restore the voice of the telephone, it is enough to approach the pole of a permanent magnet to the cylinder of soft iron. It follows from these experiments that a Bell telephone cannot act properly unless the disk is subjected to an initial magnetic tension obtained by means of a permanent magnet. It is easy to deduce this assertion from a consideration of the theory.’ The assertion may be true in the case of Bell telephones, which are worked by extremely weak currents, but when these currents are relatively strong, all electro-magnets will reproduce speech perfectly, and we have seen that M. Ader made a telephone with the ordinary electro-magnet which acted perfectly.