Let us suppose the piston at the top of the cylinder, and the space in the cylinder below it, filled with steam so as to balance the pressure of the atmosphere above the piston. Under such circumstances the steam, as will presently be explained, must have the temperature of boiling water. But that the steam should have, and should maintain, this temperature, it was evidently necessary that the inner surface of the cylinder in contact with it should have the same temperature: for if it had a lower temperature, it would take heat from the steam, and reduce the temperature of the latter. Now the cylinder being a mass of metal, has a quality in virtue of which heat passes freely through its dimensions, so that its inner surface could not be maintained at a temperature more elevated than that of its dimensions extending from the inner surface to the outer surface. Therefore, to maintain the steam contained in the cylinder at the proper temperature, it was essential that the whole of the solid metal composing the cylinder should be itself at that temperature. Things being in this state, it was required that a vacuum should be produced under the piston to give effect to the atmospheric pressure above it, by relieving it from the pressure below. This, indeed, would appear to have been attained by introducing as much cold water within the cylinder as would be sufficient to reconvert the steam contained in it into water; but Watt found, in his experiments on the atmospheric model, that the piston would not descend with the proper force, unless a vastly greater quantity of water were introduced into the cylinder than the quantity which he had ascertained to be [Pg088] necessary for the reconversion of the steam into water. The cause of this he perceived and fully explained.