vary its resistance by the pressure caused by the vibrations of the diaphragm. To effect this a small circular piece, technically termed "button," of the semi-conductor was placed between two platinum disks in a small cup. Electric connection between the disks and the button was secured by inserting a small piece of rubber tubing. The first button was made of solid plumbago, and the results were quite excellent; but still the instrument was inferior to the Bell telephone. Experiments were then made upon many materials in order to obtain a button whose resistance, though small, could be greatly varied; and, when the list of substances, natural and artificial, had been wellnigh exhausted, without very satisfactory result, a fortunate accident led to the solution of the difficulty. A small quantity of lampblack had been taken from the chimney of a smoking petroleum-lamp and preserved as a curiosity on account of its intensely black color. This substance was now tried as, it would seem, a dernier ressort. The results were excellent beyond all hope, the articulation very distinct, and the volume several times as great as could be obtained with a magneto-telephone. It was found that the resistance could be varied by pressure alone from three hundred ohms to the fractional part of a single ohm. Fig. 3 shows an instrument used for the experimental determination
Fig. 3.
of the change of resistance due to pressure only. C is a piece of carbon placed between two metallic plates which are connected with the battery, B, in whose circuit is also the galvanometer G. As the current passes it must go through the carbon, the pressure on which can be varied by changing the weights placed upon it. The deflections of the galvanometer-needle indicated that the resistance of the carbon varied inversely as the pressure to which it is subjected. The best arrangement proved to be to make the resistance of the circuit 6⁄10 of an ohm, while the normal resistance of the carbon itself was three ohms.
Good results were obtained with other materials besides carbon; the following is a list of the six most useful substances for this purpose in the order of their value: 1. Lampblack; 2. Hyperoxide of lead; 3. Iodide of copper; 4. Graphite; 5. Gas-carbon; 6. Platinum-black.
In the manufacture of the carbon button great care has to be taken that the deposit of lampblack be obtained at the lowest possible temperature, and untouched by the flame; otherwise it is utterly useless for the purpose. Thus commercial lampblack offers very great resist-
