Let V V' (fig. 3) be two steam-vessels, communicating by the tubes T T' 

 (marked by the same letters in fig. 2) with the greater boiler D. 



Let S be a pipe, called the suction-pipe, descending into the well or reser- 

 voir from which the water is to be raised, and communicating with each of 



Fig. 3. 



3 



<B' 



the steam-vessels through tubes D D', by valves A A', which open upward. 

 Let F be a pipe continued from the level of the engine to whatever higher 

 level it is intended to elevate the water. The steam-vessels V V communi- 

 cate with the force-pipe F, by valves B B', which open upward, through the 

 tubes E E'. Over the steam-vessels and on the force-pipe is placed a small 

 cistern, C, already mentioned, which is kept filled with cold water from the 

 force-pipe, and from the bottom of which proceeds a pipe terminated with a 

 cock G. This is called the condensing-pipe, and can be brought alternately 

 over each steam-vessel. From this cistern another pipe communicates with 

 the feeding-boiler (fig. 2) by the cock M. 



The communication of the pipes T T' with the boiler can be opened and 

 closed alternately, by the regulator R (fig. 2), already described. 



Now, suppose the steam-vessels and tubes to be all iilled with common 

 atmospheric air, and that the regulator be placed so that the communication 

 between the tube T and the boiler be opened, the communication between the 

 other tube T' and the boiler being closed, steam will flow into V through T. 

 At first, while the vessel V is cold, the steam will be condensed, and will fall 

 in drops of water on the bottom and sides of the vessel. The continued sup- 

 ply of steam from the boiler will at length impart such a degree of heat to the 

 vessel V, that it will cease to condense it. Mixed with the heated air con- 

 tained in the vessel V, it will have an elastic force greater than the atmo- 

 spheric pressure, and will therefore force open the valve B, through which a 

 mixture of air and steam will be driven until all the air in the vessel V will 

 have passed out, and it will contain nothing but the pure vapor of water. 



When this has taken place, suppose the regulator be moved so as to close 

 the communication between the tube T and the boiler, and to stop the further 

 supply of steam to the vessel V ; and at the same time let the condensing-pipe 

 G be brought over the vessel V, and the cock opened so as to let a stream of 

 cold water flow upon it. This will cool the vessel V, and the stream with 

 which it is filled will be condensed and fall in a few drops of water, leaving 



