While engaged in these experiments I was struck with the remarkable disproportion in weight between the membrane and the bones that were vibrated by it. It occurred to me that if a membrane as thin as tissue paper could control the vibration of bones that were, compared to it, of immense size and weight, why should not a larger and thicker membrane be able to vibrate a piece of iron in front of an electro-magnet, in which case the complication of steel rods shown in my first form of telephone, Fig. 3, could be done away with, and a simple piece of iron attached to a membrane be placed at either end of the telegraphic circuit. Figure 6 shows the form of apparatus that I was then employing for producing undulatory currents of electricity for the purpose of multiple telegraphy. A steel reed, A, was clamped firmly by one extremity to the uncovered leg _h_ of an electro-magnet E, and the free end of the reed projected above the covered leg. When the reed A was vibrated in any mechanical way the battery current was thrown into waves, and electrical undulations traversed the circuit B E W E', throwing into vibration the corresponding reed A' at the other end of the circuit. I immediately proceeded to put my new idea to the test of practical experiment, and for this purpose I attached the reed A (Fig. 7) loosely by one extremity to the uncovered pole _h_ of the magnet, and fastened the other extremity to the centre of a stretched membrane of goldbeaters' skin _n_. I presumed that upon speaking in the neighbourhood of the membrane _n_ it would be thrown into vibration and cause the steel reed A to move in a similar manner, occasioning undulations in the electrical current that would correspond to the changes in the density of the air during the production of the sound; and I further thought that the change of the density of the current at the receiving end would cause the magnet there to attract the reed A' in such a manner that it should copy the motion of the reed A, in which case its movements would occasion a sound from the membrane _n'_ similar in _timbre_ to that which had occasioned the original vibration. [Illustration: Fig. 7] [Illustration: Fig. 8]