Dry rust forming between the fish plates and the rails of the track, at the joints, makes a poor conductor, and hence the low current, from only one or two cells of battery used in the rail circuit for signaling is very liable to find sufficient resistance at the joints from this cause to prevent the continuous passage of the current through the rails to the relay. Mr. Robinson discovered this difficulty in his first experiments in rail signaling in 1872 and the necessity for making a reliable electrical connection from rail to rail in order to insure the reliability of his closed circuit signal system. As heretofore stated, therefore, he at that time conceived the invention of the bond wire, Fig. 11, for this purpose, the connection to be made by drilling holes in the adjacent rails, driving the ends of the wires tightly into these holes, and making the connection so close that there would be no room for moisture to penetrate or rust to form. And as an alternative form he proposed to secure the ends of the wire, or of a plate, to the adjacent rails by soldering, as shown in Fig. 12. In those early days there were serious technical objections to both of these methods. _First_: The difficulty and expense of boring holes in all the rails of the section and connecting them up, and the difficulty of getting the railroad company to consent to such an innovation to test what at that time might be regarded as an experiment, and _Second_: Soldering seemed impracticable on account of the difficulty of heating up the rail quickly enough at the required point. Mr. Robinson, therefore, postponed the application of the bond wire until he could secure better facilities for applying and using it. He, meantime, experimented along other lines, however, for the purpose of securing good electrical connection between adjacent rails without boring holes therein. One of these methods was very successful. It consisted in the use of elastic split springs having their ends resting on the flanges of the adjacent rails, and held in place by small blocks secured to the ties. The passing of a train depressing the rails slightly caused a slight frictional movement between the rails and the springs, thus preserving good electrical contact. In the West Somerville installation, near Boston, made in January, 1876, as heretofore described, Mr. Robinson used the bond wire shown in Fig. 11. In applying this, holes were bored in the rails and the wire, fitting the holes as closely as possible, were forced in. A semi-circular punch was then carefully used to set the metal up close around the wire. There has been no better bond wire devised since then except in mechanical construction. Bonds of various designs have been made heavier, and with heavier end plugs for mechanical connection to the rails.