1825.] Mr, Hemes on an improved Air-pump, MS 



unexcluded, its immediate access to the receiver must retard 

 the exhaustion, and ultimately set a limit to the power of the 

 pump. 



To remedy this inconvenience, and render the machine still 

 more perfect, I have added another barrel, figure 2, (having 

 the same letters of reference in the description) which is joined 

 together by a connecting tube M. On these pipes there are 

 three stop-cocks placed as shown in the drawing. The pipes 

 DD are connected together and enter the receiver as one." L 

 and L are also connected for condensation. 



When this double pump is worked the stop-cocks N and O 

 are opened, and P is shut. The pumps may be wrought in in 

 the usual way by a rack and pinion, the one piston being made 

 to ascend while the other is descending. In this condition the 

 pumps will exhaust each individually with a double stroke, so 

 that two continued united streams of air will be thrown out. 



When the exhaustion has proceeded in this way as far as 

 may be judged necessary, say until -Jr^ths of the included air is 

 thrown out, and a still more perfect vacuum be required, open 

 the stop-cock P and shut the cocks N and O, and continue 

 working the pumps. 



The communication between the external air and figure I> 

 being cut off by the cock O, and between the receiver and 

 fig. 2, by the cock N; and a communication being opened 

 between these pipes, it will be easily seen by a slight exami- 

 nation of the plates, that the effect of fig. 2 must be to exhaust 

 the air out of that part of the pipes D and L, situated below 

 the stop-cocks N and O. Consequently the pressure of the 

 external air will be entirely removed from the valves I and I', 

 and allow the air in the barrel of fig. 1 to escape freely ; for as 

 the one piston is ascending and forcing out air, the other is 

 descending and forming a vacuum ready to receive it. 



By this contrivance the receiver may be almost completely 

 emptied of its contents. It is not, however, supposed that for 

 common purposes, in an accurately constructed pump, this 

 connected barrel will be necessary, and on this supposition 

 (where a condensing apparatus is not required), I would recom- 

 mend that the piston rod be formed of an open tube (see fig. 2), 

 having a conical valve at the bottom, so that when the piston is 

 forced downwards, the included air will escape through the 

 valve, which, if wrought with oil, will suflEiciently prevent its 

 escape ; and, of course, the pipe L. with the valve r will not b© 

 required. 



