550 cnv.Ai? Atn imPiEAfK Tvup. 



c 18 a conical brass valve, well ground into the top of the 

 barrel, on one side of the collar of leathers, and kept from 

 being thrown out of its place, when open, by a pin and a 

 small semicircle of brass, d is another brass valve, only 

 this must be made much thinner than the other, that it may 

 open with more ease by the condensation of the air below it. 

 This valve has a long wire, //, fixed to its under side ; 

 which wire, when the valve is shut, must project about | of 

 an inch below the piston, e e, e e^ are two cross pieces, hav- 

 ing a hole drilled through the middle of each, through which 

 the wire^// passes. (One of these is represented separate at 

 fig. 2, PI. VIII.) When the valve rf, is open, it is prevented 

 from flying out of its place by a small button, at the end of 

 the ^'ueff, g gig gi are the passages by which the barrel 

 communicates with the receiver. After the valves are fi- 

 nished, and fixed in their places, the top of the piston, and 

 the under-side of the top of the barrel, are to be ground 

 accurately to each other; so that when the piston is drawn 

 to the top of the barrel, there may not be the least vacancy 

 between them. The action of this pump will be as follows : 

 Arjionofthis When the piston is moving upwards, the valve d will 

 i^ump, ^i^y^ jjy ^^^ ^^^ weight, and by the condensation of the air 



above it, which having no other exit, will necessarily open 

 the valve c in the top of the barrel ; and when all the air is 

 expelled, this last valve will also shut by its own weight; 

 and as the top of the barrel and the piston are ground to fit 

 each other, it is evident, that the whole of the air, above the 

 latter, must be expelled. When the piston descends, the 

 valve c (tig. 2.) will still continue shut, by its own weight, 

 and now also by the pressure of the atmosphere, and the 

 valve d will open, by the condensation of the air below it, 

 and in the receiver; air from the latter, therefore, will rush 

 through the passages g g^ g g, into the upper part of the 

 barrel, and will be expelled again when the piston is again 

 drawn upwards; and this expulsion of a fresh barrel full of 

 air, iBvery time the piston ascends, will continue, till the air 

 in the receiver is so far rarified, as not to be able to lift the 

 mcialljc valve d. It is now that the use of the wire/y will 

 be manifest, for by projecting below the piston, it is evident, 

 that it must reach the bottom of the barrel, sooner than the 



piston 



