Let A B (_fig._ 7.) be a cylinder open at the top, and let a piston P be fitted into it, so as to move in it air tight. At the bottom of the cylinder suppose an opening provided, which can be closed at pleasure, by a stop-cock, or otherwise, so that the communication may be opened and closed at will between the interior of the cylinder and the external air. This stop-cock being opened, let the piston be drawn upwards till it reach the top of the cylinder. Let the stop-cock at the bottom be then removed, and imagine that some means can be supplied by which the air within the cylinder can be suddenly annihilated. The piston, now at the top, will have above it the pressure of the atmosphere; and having no air below, it will be resisted in its descent by no force save that arising from its friction with the cylinder. If, then, the force of the air above the piston be greater than the resistance arising from this friction, the piston will descend with the excess of this force, and will continue so to descend until it reach the bottom of the cylinder. Having attained that position, let us [Pg038] suppose the stop-cock in the bottom opened, so as to allow the external air to pass freely below the piston. The piston may now be drawn to the top of the cylinder again, offering no resistance save that of its weight, and its friction with the cylinder. Having reached the top of the cylinder once more, let the stop-cock be closed, and the air included within the cylinder once more annihilated. A second descent of the piston will take place, with the same force as before, and in like manner the process may be continued indefinitely. Now, if it should appear that means could be provided suddenly and repeatedly to annihilate the air within the cylinder, and that the pressure of the atmosphere above the piston should exert a force compared with which the weight of the piston and its friction are trifling, it is evident that a moving power would be obtained which would be capable, by proper mechanism, of being applied to any useful purpose, but which would more especially be applicable to the working of pumps, the motion of which corresponds with that which has been just ascribed to the piston in the cylinder. Such were the first ideas of Papin. But in order to enable those who are not conversant with physical science fully to appreciate their importance, it will be necessary here to explain some of the mechanical properties of atmospheric air.