i8o

Of the Air-Tump.

barrel cannot be pushed back through the valveh, it will ascend through a hole in the piston,and escape through a valve at d ; and be hin-dered by that valve from returning into the bar-rel, when the piston is again raised. At thenext raising of the piston, a vacuum is againmade in the fame manner as before, between ^and e -» upon which, more of the air that wasleft in the receiver M, gets out thence by itsspring, and runs into the barrel B K, throughthe valve B. The fame thing is to be under-stood with regard to the other barrel A /; andas the handle F is turned backwards and for-wards, it alternately raises and depresies the pis-tons in their barrels; always raising one whilstit depresses the other. And, as there is a va-cuum made in each barrel when its piston i sraised, the particles of air in the receiver Mpush out another by their spring or elasticity,through the hole i, and pipe G G into the bar-rels-, until at last the air in the receiver comes i°'be so much dilated, and its spring so far weak'ened, that it can no longer get through th?valves; and then no more can be taken out*Hence, there is no such thing as making a per-sect vacuum in the receiver; for the quantity ofair taken out at any one stroke, will always bethe density thereof in the receiver : and there-fore it is impossible to take it all out, because,supposing the receiver and barrels of equal ca-pacity, there will be always as much left astaken out at the last turn of the handle.

There is a cock k below the pump-plate,■which being turned, lets the air into the reeeivctagain; and then the receiver becomes loose, # nCmay be taken off the plate. The barrels ^

sixed to the frame E e e by two screw-nuts ff*

which