The first practical use of induced currents in telegraphy was when Mr. Edison, in 1885, enabled the trains on a line of the Staten Island Railroad to be kept in constant communication with a telegraphic wire, suspended in the ordinary way beside the track. The roof of a car was of insulated metal, and every tap of an operator's key within the walls electrified the roof just long enough to induce a brief pulse through the telegraphic circuit. In sending a message to the car this wire was, moment by moment, electrified, inducing a response first in the car roof, and next in the "sounder" beneath it. This remarkable apparatus, afterward used on the Lehigh Valley Railroad, was discontinued from lack of commercial support, although it would seem to be advantageous to maintain such a service on other than commercial grounds. In case of chance obstructions on the track, or other peril, to be able to communicate at any moment with a train as it speeds along might mean safety instead of disaster. The chief item in the cost of this system is the large outlay for a special telegraphic wire. The next electrician to employ induced currents in telegraphy was Mr. (now Sir) William H. Preece, the engineer then at the head of the British telegraph system. Let one example of his work be cited. In 1896 a cable was laid between Lavernock, near Cardiff, on the Bristol Channel, and Flat Holme, an island three and a third miles off. As the channel at this point is a much-frequented route and anchor ground, the cable was broken again and again. As a substitute for it Mr. Preece, in 1898, strung wires along the opposite shores, and found that an electric pulse sent through one wire instantly made itself heard in a telephone connected with the other. It would seem that in this etheric form of telegraphy the two opposite lines of wire must be each as long as the distance which separates them; therefore, to communicate across the English Channel from Dover to Calais would require a line along each coast at least twenty miles in length. Where such lines exist for ordinary telegraphy, they might easily lend themselves to the Preece system of signalling in case a submarine cable were to part.