But here a difficulty presented itself, against which it was necessary to provide. The cold water admitted through the jet to condense the steam, mixed with the condensed steam itself, would gradually collect in the vessel C, and at length choke it. To prevent this, Watt proposed to put the vessel C in communication with a pump F, which might be wrought by the engine itself, and by which the water, which would collect in the bottom of the vessel C, would be constantly drawn off. This pump would be evidently rendered the more necessary, since more or less atmospheric air, always combined with water in its common state, would enter the vessel C by the condensing jet. This air would be disengaged in the vessel C by the heat of the steam condensed therein; and it would rise through the tube S, and vitiate the vacuum in the cylinder;—an effect which would be rendered the more injurious, [Pg123] inasmuch as, unlike steam, this elastic fluid would be incapable of being condensed by cold. The pump F, therefore, by which Watt proposed to draw off the water from the vessel C, might also be made to draw off the air, or the principal part of it. The vessel C was subsequently called a _condenser_; and, from the circumstances just adverted to, the pump F has been called the _air-pump_. These—namely, the cylinder, the condenser, and the air-pump—were the three principal parts in the invention, as it first presented itself to the mind of Watt—and even before it was reduced to a model, or submitted to experiment. But, in addition to these, other two improvements offered themselves in the very first stage of its progress. In the atmospheric engine, the piston was maintained steam-tight in the cylinder by supplying a stream of cold water above it, by which the small interstices between the piston and cylinder would be stopped. It is evident that the effect of this water as the piston descended would be to cool the cylinder, besides which any portion of it which might pass between the piston and cylinder and which would pass below the piston, would boil the moment it would fall into the cylinder, which itself would be maintained at the boiling temperature. This water, therefore, would produce steam, the pressure of which would resist the descent of the piston. Watt perceived, that even though this inconvenience were removed by the use of oil or tallow upon the piston, still, that as the piston would descend in the cylinder, the cold atmosphere would follow it; and would, to a certain extent, lower the temperature of the cylinder. On the next ascent of the piston, this temperature would have to be again raised to 212° by the steam coming from the boiler, and would entail upon the machine a proportionate waste of power.