III

Of the Air-Pump.

then be held down to the piate by the pressure of the externa! air, oratmosphere. For, as the handle (Fig. 2.) is turned backwards, itraises the piston de in the barre! BK, by means of the wheel F andrack Dd: and as the piston is leathered so tight, as to fit the barre!exactly, no air can get between the piston and barre!; and therefore,all the air above d in the barre! is lifted up towards B, and a vacuumis made in the barre! from e to b ; upon which, part of the air in thereceiver M (Fig 1.) by its spring, rulhes through the hole r, in thebrass piate LL, along the pipe GCG (which communicares with bothbarreis by the hollow trunk IHK (Fig. 2.) and pushing up thevalve b , enters into the vacant place be of the barre! BK. For,wherever the resistance or pressure; is taken off, the air will run to• that place, if it can find a passage.—Then, as the 'handle F be turnedfor ward, the piston de will be depressed in the barrel; and, as theair which had got into the barrel cannot be pustied back throughthe valve b , it will afcend through a hole in the piston, and efcapethrough a valve at d ; and be hindernd by that valve from returninginto the barrel, when the piston is again raised. At the next raisingof the piston, a vacuum is again made in the fame manner as before,between b and e ; upon which more of the air, which was lest in thereceiver M gets out thence by its spring, and runs into the barrel BK,through the valve B. The fame thing is to be understood with regardto the other barrel AI-, and as the handle F is turned backwards andfor ward s, it alternately raises and depresses the pistons in their barrels;always raising one whilst it depresses the other. And, as the re is avacuum made in each barrel when its piston is raised, every particle ofair in the receiver M pulhes out another, by its spring or elasticity,through the hole i, and pipe G G into the barrels ; until at last theair in the receiver comes to be so much dilated, and its spring sofar weakened, that it can no longer get through the valves; and thenno more can be taken out. Hence, there is no such thing as makinga perfect vacuum in the receiver ; for the quantity of air taken out atany one stroke, will always be as the density thereof in the receiver :and therefore it is imposiible to take it ali out, because, supposingthe receiver and barrels of equal capacity, there will be always as muchlest as was taken out at the last turn of the handle.

There is a cock k below the pump-plate,which being turned, lets theair into the receiver again ; and then the receiver becomes loose, andmay be taken off the piate. The barrels are fixed to the frame Eeebytwo screw-nuts ff which press down the top piece E upon thebarrels: and the hollow trunk H (in Fig. 2.) is covered by a box,as Gf/in Fig. 1. There