After much experimenting Hertz, of Carlsruhe, discovered the fact that when a discharge was taking place in an oscillatory circuit tiny sparks passed between the ends of a curved wire held some distance away. His apparatus is illustrated in Figs. 6 and 7. The former, which is termed nowadays a "Hertz Oscillator," is simply two metal discs almost connected by a thick wire. The wire is broken, however, at the centre, and the two halves terminate in two metal balls. Each ball is connected to one terminal of an induction coil. Now the current comes from an induction coil in a series of spurts. It is not an alternating current exactly (since every alternate current is so feeble as to be negligible), but is practically an intermittent current always in the same direction. Thus we may call one the positive end of the coil and the other the negative. A short current comes along with every backward movement of the little vibrating arm which forms a part of the apparatus. This breaking of the "primary" circuit may take place perhaps fifty times per second, so that the intermittent "secondary" currents will succeed each other at intervals of a fiftieth of a second, or even less. The brain reels at the attempt to think of a fiftieth of a second, but it is really quite a long interval as these things go, and during that interval quite a lot happens. For the current first of all charges the two plates as a condenser. [Illustration: FIG. 6.--The apparatus by which Hertz made his discoveries, hence called the Hertz Oscillator. _a a_ are metal plates; _d_ is the spark-gap between the two metal balls; _b_ is the battery, and _c_ the induction coil.] When they are as full as they will hold the current overflows, as it were, across the gap between the two balls. Now an air-gap--a gap that is filled with air, between two conductors--is a very strong insulator. But when current has once broken through it it becomes a fairly good conductor. Hence as soon as the first spark has passed between the two knobs the plates become connected almost as if a wire were passed from one to the other. And there we have quite a good oscillatory circuit. There is capacity at each end and a fairly long length of wire to provide the inductance. Consequently that breakdown of the insulation of the air in the spark-gap is followed by electrical oscillations which take place with inconceivable rapidity. Yet because of the resistance of the spark-gap, which is considerable even after it has been broken through, the oscillations do not continue for long. They have died away long before the lapse of a fiftieth of a second, when the next impulse comes along from the coil. In the meantime the air-gap regains its insulating properties, and so, on the arrival of the next impulse, the whole thing occurs once more.