5^4 Defmpihn of a new Air-pump, 



is almoft cylindrical, tapering but a little from the bafe upward, and being laid on tbe plate, 

 ■with a fmall drop of oil interpofed, the hollow cylindrical cap E, very little wider internallT 

 than the valve, is put over it, being faftened down on the proje£ting circular part of the plate 

 a a. To let die air pafs through this cap, three little holes, e, e, are made in its fides, the 

 centres of which are juft as high as the lower edge of the valve D, when it is raifed to touch 

 tlie top of tlie cap ; the holes being higher than the plate under the valve, that the oil may 

 not run out by them. When this valve is found to admit air into the barrel, it is occa- 

 fioned either by the oil's being blown away, or fomc particle of dufl, &c. getting between it 

 and the plate (which would produce the fame efFeft in any valve), and is remedied by taking 

 it off, wiping it clean, and applying another drop of oil : the greater the quantity of air is, 

 which paffes by the valve, the more is the oil diffipated ; and, confequcntly, this chiefly hap- 

 pens in the beginning of the exhauftion, when much air is drawn from the receiver, and 

 thrown out of the barrel at each ftroke ; (b that whenever great rarefa£Hon is requifite, I 

 often found it convenient, towards the end of the operation, to apply a frefli drop of oil to the 

 valve, keeping the pifton in the mean time elevated to the top of the barrel, whilft the valve is 

 taken off, until It be replaced; which is done in an inftant : thus it will perform perfectly 

 well if quite clean and free from duft ; but as the fmalleft particle of dirt, or mucus (which 

 is apt to be produced from the oil's corroding the brafs, and becoming clammy), will render 

 die valve not air-tight, I am, therefore, perfuaded, that no valve will fo perfectly or eer« 

 tainly exclude air as a ftop-cock ; and this I take to be the chief reafon why Mr. Smeaton's 

 pump, as improved by Meflrs Haas's and Hurter's contrivance, for raifmg the valve at the 

 bottom of the barrel, does not, in pradtice, anfwer as well as any other : for if fome air did 

 not return into the receiver by that valve in pumps on his conftru£tion, it would be hard to 

 tell why it fhould not perform as well as a pump of any other conftruQicn. 



To the bottom of the barrel is faftened by four fcrews, paffed through its prominent rim, 

 the circular plate R, forming In pari the fliell of the cock S, and which has the high ridge 

 T on the back of it ; all thefe being only one piece of metal, which is reprefented fepa- 

 rately in fig. 5-. and was caft in that form : the round plate is about Jth of an incH 

 thick ; but the cavity of the cock intrudes fo far on this, that when the key of the 

 cock is put in,, the thicknefs of the plate in the part direftly over the key, where the 

 dufts open into the barrel, is not more than ith of an inch: the length of the ftieii 

 of the cock is, three inches; that of the key -Jth of an inch lefs : the diameter of the 

 key, at the thicker extremity is i to^is of an inch, and at the fmaller end i ;|th of an 

 inch. It: is turned by the handle u, faftened on its axis as near to the end of the fhell 

 as may be; viz. at the diftance of l^h. of an inch. Care was taken that the metal of the 

 cock fhould be free from pores, by which, if air were admitted in the working, it would be 

 very hard to difcover it, as this might take place only in certain pofitious of the key ; and the. 

 ointment might fometimes prevent and fometimes fuffer it, fo that it might elude trials ; . 

 the cock was alfo fo truly and fmoothly ground, as to produce an intimate contaft of the' 

 key and ftiell . throughout (for air will penetrate where liquids would not); it was alfo' 

 jnade of a true conical fliape, its fides being perfectly ftraight from end to end. The cock' 



muii 



I 



m 



