The shunt consists of a wire leading from the copper to the carbon rod with a condenser and an inductance coil inserted in it. The latter coil also forms one part of that coil by which the oscillations in the local circuit are transferred to the antenna. The electrical explanation of what happens when the current is turned on to an arrangement like this is rather too complex to set out here. It depends upon a curious behaviour of the arc. It is really a conductor, yet it does not behave as ordinary conductors do, and the result is that the continuous current flowing through the arc is accompanied by an oscillating current in the shunt circuit. And the important feature of the arrangement is that these oscillations are continuous, in one long train, not in a succession of trains. The advantage of this has already been referred to. One other feature of the apparatus just described should be mentioned, since it will seem curious to the general reader. For it to work properly it is necessary that the arc should be enclosed in a chamber filled with hydrogen or a hydro-carbon gas. Coal-gas is generally used. Hertz' original discovery was that small sparks could be seen to pass between the ends of a curved wire when the electric waves fell upon it. Such "spark detectors," as they are called, are useful in the laboratory, but not for practical telegraphy. [Illustration: FIG. 12.--Diagram (simplified) showing how Poulsen generates oscillations. Current from a dynamo flows through the arc, whereupon currents oscillate through the condenser and coil (as described in the text).] Several people seem to have noticed in years gone by that a mass of loose metal filings, normally a very bad conductor of electricity, became a much better conductor when an electrical discharge of some sort occurred near by. The demand for a wireless receiver had not then arisen, however, and so the discoveries were not followed up. Consequently it remained to be rediscovered by Branly, of Paris, in 1890. He placed some metal filings in a glass tube, the ends of which he closed with metal plugs. Lying loosely together the filings would not conduct the current of a small battery from one plug to the other, but when a spark occurred not far away they suddenly became conductive and allowed it to pass. Several years after this Sir Oliver Lodge took up the idea as a receiver for wireless messages, and believing that its action was due to the waves causing the filings to cling together, he christened it "Coherer." Marconi succeeded in making a very delicate form of this, although working on strictly the same lines.