The design of the cable was a subject of experiment by Professor Morse and others. It was known that the conductor should be of copper, possessing a high conductivity for the electric current, and that its insulating jacket of gutta-percha should offer a great resistance to the leakage of the current. Moreover, experience had shown that the protecting sheath or armour of the core should be light and flexible as well as strong, in order to resist external violence and allow it to be lifted for repair. There was another consideration, however, which at this time was rather a puzzle. As early as 1823 Mr. (afterwards Sir) Francis Ronalds had observed that electric signals were retarded in passing through an insulated wire or core laid under ground, and the same effect was noticeable on cores immersed in water, and particularly on the lengthy cable between England and the Hague. Faraday showed that it was caused by induction between the electricity in the wire and the earth or water surrounding it. A core, in fact, is an attenuated Leyden jar; the wire of the core, its insulating jacket, and the soil or water around it stand respectively for the inner tinfoil, the glass, and the outer tinfoil of the jar. When the wire is charged from a battery, the electricity induces an opposite charge in the water as it travels along, and as the two charges attract each other, the exciting charge is restrained. The speed of a signal through the conductor of a submarine cable is thus diminished by a drag of its own making. The nature of the phenomenon was clear, but the laws which governed it were still a mystery. It became a serious question whether, on a long cable such as that required for the Atlantic, the signals might not be so sluggish that the work would hardly pay. Faraday had said to Mr. Field that a signal would take 'about a second,' and the American was satisfied; but Professor Thomson enunciated the law of retardation, and cleared up the whole matter. He showed that the velocity of a signal through a given core was inversely proportional to the square of the length of the core. That is to say, in any particular cable the speed of a signal is diminished to one-fourth if the length is doubled, to one-ninth if it is trebled, to one-sixteenth if it is quadrupled, and so on. It was now possible to calculate the time taken by a signal in traversing the proposed Atlantic line to a minute fraction of a second, and to design the proper core for a cable of any given length.