Cryptonomicon

by Neal Stephenson

As of 1861, some 17,500 kilometers of submarine cable had been laid in various places around the world, of which only about 5,000 kilometers worked. The remaining 12,500 kilometers represented a loss to their investors, and most of these lost investments were long cables such as the ones between Britain and the United States and Britain and India (3,500 and 5,600 kilometers, respectively). Understanding why long cables failed was not a trivial problem; it defeated eminent scientists like Rankine and Siemens and was solved, in the end, only by William Thomson. In prospect, it probably looked like it was going to be easy. Insulated telegraph wires strung from pole to pole worked just as one might expect, and so, assuming that watertight insulation could be found, similar wires laid under the ocean should work just as well. The insulation was soon found in the form of gutta-percha. Very long gutta-percha-insulated wires were built. They worked fine when laid out on the factory floor and tested. But when immersed in water they worked poorly, if at all. The problem was that water, unlike air, is an electrical conductor, which is to say that charged particles are free to move around in it. When a pulse of electrons moves down an immersed cable, it repels electrons in the surrounding seawater, creating a positively charged pulse in the water outside. These two charged regions interact with each other in such a way as to smear out the original pulse moving down the wire.