Imagine a short tube into which you have inserted a pellet, or wad of any kind, so that it fits tolerably, yet moves easily back and forth in the bore of the tube. If this pellet or wad is at one end of the tube you may, by inserting that end in your mouth and putting air-pressure upon it, make it slide to the other end. You do not touch it with anything; you may push it back and forth with your breath as many times as you wish, not by blowing against it, so to speak, but by producing an actual air-pressure upon it which is confined by the sides of the tube and cannot go elsewhere. The only pressure necessary is enough to move the pellet. Now, if you push this little pellet one way by the air-pressure from your mouth, and then, instead of reversing the tube in the mouth and pushing it back again in the same way, reverse the process and suck the air out from behind it, it comes back by the pressure of the outside atmosphere. This was the way the first steam engines worked. Their only purpose was to get the piston lifted, and air-pressure did all the actual work. If you turn the tube, and put an air-pressure first at one end and then at the other, and pay no attention to vacuum or atmospheric pressure, you will have the principle of the later modern, almost universal, high-pressure, double-acting steam-engine. But now you must imagine that the tube is fixed immovably, and that the air-pressure is constant in a pipe leading to the tube, and yet must be admitted first to one end of the tube and then to the other alternately, in order to push the pellet back and forth in it. It seems simple. Perhaps the young reader can find a way to do it, but it required about a hundred years for ingenious men to find out how to do precisely the same thing automatically. It involves the steam-chest and the slide-valve, and all other kinds of steam valves that have been invented, including the Corliss cut-off, and all others that are akin to it in object and action. But now imagine the tube closed at each end to begin with, and the little moving pellet, or plunger, on the inside. To get the air into both ends of the tube alternately, and to use its pressure on each side of the pellet, we will suppose that the air-pipe is forked, and that one end of each fork is inserted into the side of the tube near the end, like the figure below, and imagine also that you have put a finger over each end of the tube. [Illustration: Fig. 1]