All this you know, for you have followed through such a cycle of changes before. You know also how we can use the audion characteristic to tell us what sort of changes take place in the plate current when the grid voltage changes. The plate current increases and decreases alternately, becoming greater and less than it would be if the grid were not interfering. These variations in its intensity take place very rapidly, that is with whatever high frequency the sending station operates. What happens to the plate current on the average? The plate current, you remember, is a stream of electrons from the filament to the plate (on the inside of the tube), and from the plate back through the B-battery to the filament (on the outside of the tube). The grid alternately assists and opposes that stream. When it assists, the electrons in the plate circuit are moved at a faster rate. When the grid becomes negative and opposes the plate the stream of electrons is at a slower rate. The stream is always going in the same direction but it varies in its rate depending upon the changes in grid potential. [Illustration: Fig 55] When the grid is positive, that is for half a cycle of the alternating grid-voltage, the stream is larger than it would be if the plate current depended only on the B-battery. For the other half of a cycle it is less. The question I am raising is this: Do more electrons move around the plate circuit if there is a signal coming in than when there is no incoming signal? To answer this we must look at the audion characteristic of our particular tube and this characteristic must have been taken with the same B-battery as we use when we try to receive the signals. There are just three possible answers to this question. The first answer is: "No, there is a smaller number of electrons passing through the plate circuit each second if the grid is being affected by an incoming signal." The second is: "The signal doesn't make any difference in the total number of electrons which move each second from filament to plate." And the third answer is: "Yes, there is a greater total number each second." [Illustration: Fig 56]