If we suppose as much, and no more, cold water introduced into the cylinder as would reconvert the steam contained in it into water, then we should have in the bottom of the cylinder a quantity of warm water with a vacuum above it: but the entire mass of metal composing the cylinder itself, which was previously at the temperature of boiling water, would still be at the same temperature. The warm water, resting in contact with this metal in the bottom of the cylinder, would be immediately heated by it, and would rise in its temperature, while the metal of the cylinder itself would be somewhat lowered in temperature by the heat which it would thus impart to the warm water contained in it. Under these circumstances, as we shall presently explain, steam would be produced from the water, which would fill the cylinder; and although such steam would not have a mechanical pressure equal in amount to the atmosphere, and therefore would not altogether prevent the piston from descending if it had no load to move, yet it would deprive the engine of so great a portion of its legitimate power as to render it altogether inefficient. But this defect would be removed by throwing into the cylinder a sufficient quantity of cold water, not only to destroy the steam contained in it, but also to cool the entire mass of metal composing the cylinder itself, until it would be reduced to such a temperature that the vapour proceeding from the water contained in it would have so small a pressure that it would not seriously or injuriously obstruct the descent of the piston. The piston being made to descend with such force as to render the machine practically efficient, it would then be necessary again to make it ascend; and to accomplish this, Watt found that the boiler should supply a quantity of steam many times greater than was necessary to fill the cylinder. Mature reflection on the circumstances which have been just explained, enabled him to discover how this undue quantity of steam was rendered necessary.