t 3^1 ] 



thefe, which difFer only in the manner of admifllon : to illu{^ 

 trate this, and to enable the experimenter as well as the artift, 

 to determine whether his pump is imperfe£l from its ftrudure 

 or from accident ; and how far it ought to be fo from either 

 caufe ; it is neceffary, though I have (relucSlantly) been {o 

 difFufe in this account already, to ftate the following parti- 

 culars: 



If air be conftantly and uniformly admitted into the rec"' 

 from this or the machine's not being air-tight, the degree of 

 rarefadion produced by its final a61:ion, will depend on the 

 proportion between the capacities of the rec/ and cyY ; for 

 there will always be a refiduum as air is always intruding, and 

 in addition to former refiduums, will be the air admitted in 

 each fucceffive ftroke : in this cafe, the lefs the rec' is, the 

 more perfed will be the vacuum formed *. Suppofe the rec"' 

 and cyr to be equal, and the vacuum made in the latter per fe 

 perfed ; then half the admitted air will be at length pumped 

 out at every ftroke let the quantity admitted be what it will :. 



for 



• Hence it is not the faired trial of a pump, to ufe a very fmall rec', or to 

 exhauft the gage-tube inftead of a rec' ; and it would be the more fallacious, if the 

 refiduum of air under the pifton, might be variable in different ftrokes, from the de- 

 feftive and uncertain performance of valves or of any thing of valve-like ftrudure; 

 for as one ftroke with a large cyl' will nearly exhauft a tube ; fo a lucky ftroke or two. 

 among many, in which the valves perform their office, will fhew a rarefaftion, which. 

 might not be produced, where many were requifite to exhauft a larger rec'. 



