A I R 



pvincipal improvement wliich it received about tliis- pcviod. 

 To whom this waj owing, it is not cafy to decide : fomo 

 Blcrihe it to Div Hoolce, other; to I'apiii, and others again 

 to Haukfbee. An eni^jine of this kii.d, with a double 

 tulie, iodefcribtd by Mr. Boyle, in the fecoiid continuation 

 of riivfieo-mechanical Experiment.-;, (works, vol.iv. ;). 510.) ; 

 but tiie manner of working it, by mtauo of a pulley and 

 v\' iro;i llirrups or triddlcs, upon which the operator Hood, 

 mull have been extrenuly inconvenient. However, by the 

 ufe of a feeond barrel and pifton, contrived to rife and fall 

 alternately with the other, and by t!ie ir-troduclion of valves, 

 which in this tiiird air-pump of Mr. Boyle fupplicd the 

 place of the plug and Hop-cock wiiieh he had before ufed, 

 ao well as by the fubfcquent iu'.provement.s of Haukfbee, 

 the preffure of the atniofphere on the defcending pillon al- 

 ways nearly baknieed tliat of the afceuding one ; fo that the 

 wiiich which worked them up and down waseafdy moved by 

 a gentle force with one hand ; and the exhaultion was alfo 

 made in much lefs time. See HaukP.iet's Phyfico-Mecha- 

 nical Experiments, p. 1, &c. Mr. Vreani, a pneumatic 

 operator, employed by Defaguliers, made an improvement in 

 Haukdiec's air-pimip, by reducing the alteniate motion of 

 the hand and winch to a circular one. In his method the 

 winch is turned ipiite round, and yet the piilons are alter- 

 nately raifed and deprelled ; by which the trouble of ibift- 

 inp- the hand backwards and forvvards, as well as the lofs of 

 time, and the Ihaking of the pump, are prevented. See 

 Defagulitrs's Courie of Exp. Philof. vol. ii. p. 378. For 

 a brief account of the progrelTive improvements of the 

 air pump, fee Cotes's Hydroftatical and Pueumatical Lec- 

 tures, left. xii. p. 156, &c. 



The llruCtnre of the air-pump, thus improved, is repre- 

 fented in Plate v. Pneumatics, Jig. 33. It confifts of two 

 brafs barrels or cylinders, a a, a a, which communicate with 

 each other by the ciilern, d J, and with the receiver, 0000, 

 which is ground level at the bottom, and fet over a hole in 

 the plate, by means of the bent pipe, /j h. In thefe barrels 

 the piltons, which are fallened fo tight that no air can get 

 between them and tiie barrels, are worked by a toothed 

 wheel, turned by the handle, bb\ and thus the racks, c c, 

 with their piftons, are worked alternately up and down. 

 The gage tube, //, is immerfcd in a bafon of quickfdver, »;, 

 at the bottom, and coinmunicates with the receiver at the 

 top ; from which it may be occafionally difengaged by turn- 

 ing a cock ; and /; is another cock, by turning of which the air 

 is again let into the exhaiiiled receiver, paffing into it with 

 a hiffing noife. Tiie attion of the toothed wheel and pif- 

 tons is reprefented in Jig. 3^. 



As the handle is turned backwards, it raifes the pifton d e, 

 in the barrel B K, by means of the wheel E, and rack 

 D d : and as no air can get between the pifton and barrel, 

 all the air above d ia lifted up towards B, and a vacuum is 

 made in the baiTcl froin e \.q h ; upon which part of the air 

 in the receiver by its fpring rufhes through the hole in the 

 brafs plate of the pump along the pipe G G, communicat- 

 ing with both barrels by the hollow trunk I H K, and pudi- 

 ing up the valve b, enters into the vacant part b e, of the 

 barrel B K. Then, as the handle F, is turned forward, the 

 pifton d e, will be deprcffed in the barrel ; and the air 

 which had got into the barrel, finding no way of efcape 

 through the clofed valve b, will afcend through a hole in 

 the pifton, and make its way into the external air through a 

 valve at d ; and it will be prevented by that valve from re- 

 turning into the barrel, when the pifton is again raifed. 

 At the next elevation of the pifton, a vacuum is again made 

 in the fame manner as before, between b and :- ; upon wJiich 



A I R 



more of the air that was left in the receiver will get out by 

 it:; fpring, and flow into the barrel, B K, through the 

 valve b. The other pifton and barrel aft in the fame man- 

 ner ; and as the handle, F, is turned back'.\ ards and for- 

 wards, it alternately raifes and deprefies the piftons in their 

 baiTels ; one being raifed v.hilft the other is deprefted. By 

 thus repeating tlie operation again and again, the air in the 

 receiver is at length rarefied to fuch a degree, that its dcn- 

 fity does not excek;d the thin air remaining in the barrel when 

 the pifton is raifed : which done, the effeft of the air-pump 

 is at an end ; the valve cannot now be opened, or if it 

 could, no air would pafs it ; there being a juft equilibrium 

 between the air on each fide. 



To judge of the degree of exhanftion, there is added 

 the gage-tube, //, open at both ends, and about 34 inches 

 long {Jig. 33-)> affixed to a wooden ruler, which is divid-d 

 into inches and parts of an inch, from tlis bottom where it 

 is even with the quiekfilver In the bafon, w, and continued 

 to the top, a little behnv the plate of the air-pump, to 30 

 or 3 I inehe?. Hence the air in the tu!)e raref) ing as faft 

 as that in the receiver, in proportion as the exhauftion ad- 

 vances, the mercury will be raifed by the prefl\ire of the 

 column of external air, prevailing over that of the column 

 of air included ; till the column of air, and mercury toge- 

 ther, become a balance to that of the external air. When 

 the mercui-y is thus rifen to the fame height as it ftands in 

 the barometer, which is indicated by the fcale of inches 

 added to the gage, the inftrument is a juft Torricellian 

 tube ; and the vacuum may be concluded to be as perfeft 

 as that in the upper end of the barometer. When the 

 cock, n, is turned, fo as to make a communication with the 

 external air ; this rufties in, and the mercury in the gage 

 immediately fubfides into the-bafon. See G.1GE. 



In eftimating the gradual afcent of the quiekfilver in the 

 gage, it is evident that, as we continue to pump, the mer- 

 cuiy continues to afcend ; and that it approaches always 

 more and more to the ftandard altitude, or about 29^ inches, 

 more or lefs according to the variety of feafons. 7\nd it is 

 eafy to prove, that the dcfeft of the height of the quiek- 

 filver in the gage from the ftandard altitude is always pro- 

 portionable to the quantity of air which remains in the re- 

 ceiver ; that the altitude itfelf of the quiekfilver in the gage 

 is proportionable to the quantity of air which has been 

 exhaufted from the receiver ; and that the afcent of the 

 quiekfilver, upon every turn of the pump, is proportion- 

 able to the quantity evacuated by each turn. Let it 

 be confidered, that the whole preffure of the atniof- 

 phere upon the ciftern of the gage is equal to, and may be 

 balanced by, a column of f^uickfilver of the ftandard alti- 

 tude ; confequently, when the quiekfilver in the gage has 

 not yet arrived to the ilandard altitude, the dcfeft muft be 

 fupplied by fome other equal force, and that force is the 

 elaftic power of the air remaining in the receiver ; which 

 communicating with the upper part of the gage, hinders 

 the quiekfilver from afcending, as it w'ould otherwlfe do, to 

 the ftandard altitude. The elafticity of the air in the re- 

 ceiver is then equivalent to the weight of the deficient 

 quiekfilver ; but the weight of this is proportionable to the 

 fpace it fhould polfefs, or to the defeft of the height of the 

 quiekfilver in the gage from the ftandard height ; therefore 

 the elafticity of the remaining air is alfo proportionable to 

 the fame defeft. But the denfity of any portion of air is 

 proportionable to its elafticity, and the quantity in this cafe 

 is proportionable to the denfity ; and therefore the quantity 

 of air remaining in the receiver is proportionable to the defeft 

 of the quickfdver in the gage from its ftandard altitu<h'. 



Hence 



