From the researches of a number of experimenters with the phonograph, it appears that:

(1) The embossing point should be abandoned for a process of engraving - that is to say, instead of pressing the registering surface with a style, as in the original phonograph, an outline should be produced parallel to the surface of inscription.

(2) The best substance answering all the requirements of the question raised by engraving by means of the style, is beeswax hardened by paraffin or some other similar substance.

(3) To speak in a high tone is unnecessary, the ordinary tone giving the best results, although the intensity of the sound produced does not excel that of the telephone.

In 1886, Dr. Chichester, Dr. Bell, and C. S. Taint er took out a patent for engraving the phonogram and for the use of wax, and, as a result of their labours, produced, in the spring of 1887, the graphophone, the first really practical apparatus of the phonographic type, which was exhibited at Washington, This instrument reproduced the voice with the intensity of a good telephone; but the deformations due to the engraving were great enough to render the voice unrecognisable, unless the fancy was exercised, and one had a previous acquaintance with the speaker. The impression was obtained upon a cardboard cylinder covered with wax. As soon as the graphophone became known in America, Edison, encouraged by the results obtained with this instrument, recommenced his experiments with the phonograph, and after having tried the embossing point anew abandoned it for the engraving, thus confirming the accuracy of the conclusion of Bell and Tainter. The graphophone and the new phonograph of Edison seemed then to be practically the same instrument, only differing in form and the nature of the motor used.

The difficulties which the embossing point presented would disappear, as Prof. Thomson observed, if a surface could be found whose resistance to the embossing point was exactly proportionate to the depth of this embossing. Every trial confirmed these conclusions, for the more one attempts to speak loud the more indistinct becomes the articulation. It is necessary, then, to trace the vibrations as in the old phonauto-graph of Scott, parallel, and not perpendicular to the sheet, and to reduce to a minimum the resistance offered by the substance engraved. Berliner observed that lampblack was impracticable and insufficient; but he saw that the greyish deposit of lampblack, produced by a kerosine flame, gave a finer line than the thick dark deposit produced at the top of the same flame, and this observation led him to previously grease the registering surface by applying a layer of printing ink or oil paint by the aid of a roller or a brush. The smoke then gives a deposit of a fatty appearance, and of such a consistence that if it is touched with a style a free line is produced without any blister, even when examined with a microscope.

The phonautogram obtained is then engraved mechanically, chemically, or photo-chemicnlly. After numerous attempts, Berliner has arrived at a process by which a reprodaction can be 15-20 minutes if necessary), and which can afterwards be multiplied indefinitely by the electrotyping process, which Berliner culls the art of engraving the human voice with aquafortis, and which we call more simply phonogravure. The registering apparatus (Fig. 129) consists of a case K, which serves to support the whole. The centre of the diaphragm A carries a small brass cylinder B riveted to it, and divided in front. E is a piece of rubber tube fitted into the division in K, and supporting the end of the style C, which is formed of a thin sheet of metal vibrating on two steel pivots F. D is a piece of writing-paper strengthened by a spring plate, the end of which constitutes the tracing point. There is a piece of rubber tube placed round the style to control its vibrations. Hits disc of felt placed between the case of the diaphragm and the diaphragm itself, to control the superfluous vibrations and prevent resonance.

The whole is placed upon a light support (Fig. 127).

Phonogram and graphophone.

Phonogram and graphophone.

The apparatus reproducing the words of the speaker (Fig. 128) is constructed on the same plan, but is of smaller dimensions and more rigid in the vibrating parts. The end of the style consists of an iridium point, to avoid the wear caused by continual robbing with the metal.

Phonogram and griphoplione.

Phonogram and griphoplione.

Harder metals give a more intense sound than substances less resisting or elastic, as rubber or plaster of Paris. 'The hard metals, such as copper, nickel, or brass, speak louder than zinc; but there is a grinding, unless the writing surface is very smooth, and polished with the greatest care. This grinding is scarcely perceptible on polished glass, and Berliner hopes, by moulding the regisrelief which shall give intense sounds with a mini-