It was reserved, however, for Edison (of New Jersey, U.S.A., Electrical Adviser to the Western Union Telegraph Company) to show how advantage might be taken of this discovery to make a diaphragm speak, not directly through the action of the movements of a diaphragm affected by spoken words or other sounds, and therefore either simultaneously with these or in such quick succession after them as corresponds with the transmission of their effects along some line of electrical or other communication, but by the mechanical reproduction of similar movements at any subsequent time (within certain limits at present, but probably hereafter with practically unlimited extension as to time).

The following is slightly modified from Edison’s own description of the phonograph:—

Pleasant Ways in Science — Part 187

The instrument is composed of three parts mainly; namely, a receiving, a recording, and a transmitting apparatus. The receiving apparatus consists of a curved tube, one end of which is fitted with a mouthpiece. The other end is about two inches in diameter, and is closed with a disc or diaphragm of exceedingly thin metal, capable of being thrust slightly outwards or vibrated upon gentle pressure being applied to it from within the tube. To the centre of this diaphragm (which is vertical) is fixed a small blunt steel pin, which shares the vibratory motion of the diaphragm. This arrangement is set on a table, and can be adjusted suitably with respect to the second part of the instrument—the recorder. This is a brass cylinder, about four inches in length and four in diameter, cut with a continuous V-groove from one end to the other, so that in effect it represents a large screw. There are forty of these grooves in the entire length of the cylinder. The cylinder turns steadily, when the instrument is in operation, upon a vertical axis, its face being presented to the steel point of the receiving apparatus. The shaft on which it turns is provided with a screw-thread and works in a screwed bearing, so that as the shaft is turned (by a handle) it not only turns the cylinder, but steadily carries it upwards. The rate of this vertical motion is such that the cylinder behaves precisely as if its groove worked in a screw-bearing. Thus, if the pointer be set opposite the middle of the uppermost part of the continuous groove at the beginning of this turning motion, it will traverse the groove continuously to its lowest part, which it will reach after forty turnings of the handle. (More correctly, perhaps, we might say that the groove continuously traverses past the pointer.) Now, suppose that a piece of some such substance as tinfoil is wrapped round the cylinder. Then the pointer, when at rest, just touches the tinfoil. But when the diaphragm is vibrating under the action of aerial waves resulting from various sounds, the pointer vibrates in such a way as to indent the tinfoil—not only to a greater or less depth according to the play of the pointer to and fro in a direction square to the face of the diaphragm, but also over a range all round its mean position, corresponding to the play of the end of the pointer around _its_ mean position. The groove allows the pressure of the pointer against the tinfoil free action. If the cylinder had no groove the dead resistance of the tinfoil, thus backed up by an unyielding surface, would stop the play of the pointer. Under the actual conditions, the tinfoil is only kept taut enough to receive the impressions, while yielding sufficiently to let the play of the pointer continue unrestrained. If now a person speaks into the receiving tube, and the handle of the cylinder be turned, the vibrations of the pointer are impressed upon the portion of the tinfoil lying over the hollow groove, and are retained by it. They will be more or less deeply marked according to the quality of the sounds emitted, and according also, of course, to the strength with which the speaker utters the sounds, and to the nature of the modulations and inflexions of his voice. The result is a message verbally imprinted upon a strip of metal. It differs from the result in the case of Barlow’s logograph, in being virtually a record in three dimensions instead of one only. The varying depth of the impressions corresponds to the varying height of the curve in Barlow’s diagrams; but there the resemblance ceases; for that was the single feature which Barlow’s logographs could present.

Pleasant Ways in Science — Part 187