So we believe that the little molecules lie just anyhow, wherefore they neutralize each other and the mass of iron is powerless. When another magnet comes near, however, they are more or less pulled into the right position and the iron becomes magnetized. When the magnet is removed the magnetism which it produced is largely lost, and if last of all we give the iron a smart blow with a hammer even the residual magnetism vanishes too. Now, oscillations taking place in the neighbourhood of a piece of iron possessing residual magnetism have much the same effect as the blow of a hammer. Probably because of its rapidity an oscillating current shakes the molecules up and strews them about at random, entirely destroying any orderly arrangement of them. And Marconi used that fact in detecting oscillations. Two little coils of wire are wound together, one inside the other. Through the centre of the innermost there runs an endless band of soft iron wire. Stretched on two rollers this band travels steadily along, the motive power being clockwork, so that it is always entering the coil at one end and leaving it at the other. As it travels it passes close to two powerful steel magnets, so that as it enters the coil it is always slightly magnetized. The oscillations are passed through one of the two concentric coils, and their action is to remove suddenly the residual magnetism in that part of the moving wire which is at the moment passing through. That sudden demagnetization then affects the second of the concentric coils, inducing currents in it, not of an oscillating nature but of an ordinary intermittent kind which can make themselves audible in a telephone which is connected with the coil. This arrangement, then, causes the oscillations, which will not operate a telephone, to produce other currents of a different nature which will. The reason why oscillations have no effect in a telephone is no doubt because they change so rapidly, at rates, as has been mentioned already, of the order of a million per second. The telephone diaphragm, light and delicate though it is, is far too gross and heavy to respond to such rapidly changing impulses as that. In the magnetic detector the difficulty is overcome by making them change the magnetic condition of some iron wire which change in turn produces currents capable of operating a telephone. The Crystal Detector achieves the same result in another way. There are certain substances, of which carborundum is a notable example, which conduct electricity more readily in one direction than the other. Most of these substances are crystalline in their nature, and hence the detector in which they are used gets its name. Carborundum, by the way, is a sort of artificial diamond produced in the electric furnace and largely used as a grinding material in place of emery.