no pressure on the upper surface, and to this extent the behaviour of the water resembles that of the iron. Finally, suppose that the box be filled with gas or vapour, this substance will, like the solid or the liquid, press on the bottom of the box, with the entire weight of the gas, it will like the liquid exert a lateral pressure, but it will differ from both the other substances by its manifesting an upward pressure. It will generally, therefore, be necessary to provide a cover for the box, or if there be no cover to expose the upper surface of the gas to pressure in some other way, in order to retain it within the limits of the cube. No doubt under ordinary circumstances we may fill a vessel to the brim with carbonic acid gas, and the gas will remain much in the same manner as if the vessel had been filled with water. But the carbonic acid in such a case is kept down by the pressure of the superincumbent atmosphere. If that pressure were removed the gas would speedily expand and overflow, unless a cover were provided by which it was restrained, and then the gas would exert an outward pressure upon the cover, showing that it had the tendency to expand even though actual expansion was not permitted.
It will be seen that in this behaviour a gas is totally different from a solid or a liquid. No doubt evaporation is generally speaking taking place from the upper surface of a liquid, and the vapour thus produced acts as a gas, so that in this respect some slight qualification of the above statements might be necessary. But nothing can be more obvious than the fact that the upper pressure of the gas indicates some profound difference between its molecular structure and that of a solid or a liquid, We are so accustomed to this
