Defcription of a portal U Barometer. 353 



fuffering it to efcape when it does not dilate. Ivory ftop-cocks 

 are attended with the fame advantage. E is a cylindric cup, 

 the diameter of which nuift be known; about an inch ap- 

 pears to me to be the mod convenient dmienfiou for a baro- 

 meter of this kind. 



F is a piece of ihamov leather with which the cup is co- 

 vered. It permits the air to cxercife an aftion on the mer- 

 cury, and prevents it from efcaping when the inlirument is 

 carried. 



After having boiled the mercury in the tube, the cock and 

 cup muft be cemented at the interior bending. A little mer- 

 cury is then to be added, that it may float above the ftop- 

 cock, when it is fliut after the initruinent has been inclined. 



It mav readily be conceived that the interior diameter of 

 the long tube, and that of the cup, being known, the greatcft 

 precifion mav be obtained when they are perfeihtly cylindric ; 

 for, the diameter of the two cylinders being known, the ratio 

 of their I'olidity may be known *. 



Thofe who wi(h to avoid the trouble of this fhort calcula- 

 tion may cement to the ftop-cock a very large cup ; and 

 then the error occafioned by the elevation of the level of the 

 mercury will be a very fmall fraction of a linc^ even if the 

 barometer fliould defcend feveral inches. 



In cafe it may not be thought proper to adapt to the ftop- 

 cock a large cup, which would render the barometer much 

 heavier, nothing will be neceflTary but to meafure the height 

 of the mercury above the line of the level, and to fubtra6l it 

 from that of the mercury in the long tube. Let us fuppofe, 

 for example, that the barometer marks 23 in. 6 lin. and that 

 the mercury in the tube is elevated only one line; the re- 

 mainder will be 23 in. 5 lin. for the height of the mercury. 



ty takinn; a view of the height of the barometer, without 

 making the leaft calculation, it will be eafy to devife how the 

 fcale mav be crraduatcd fo as to make this correction needlefs. 



A nonius, to indicate the /,,th or even t4i^^' of a line may 

 be adapted to this inftrument ; but the variations in the up- 

 per (Irata of the atmofphere arc too well known to admit of 

 our having recourfe to fo exa«ft a graduation in the meafuring 

 of heights. Ought we not to conlider as very correal a baro- 

 meter that indicates -Vthofa line ? The errors in the meafure- 



I c 



• Let us fuppofe that the cHaincrtr of the barometric tube is two lines, 

 and thit of the cup twclvt, the afccnfion or' the mercury in the cup will bi: 

 to the dcprcffion of that in the tube as the Iqunre of the lurfaccs ; which 

 in this Ciiie is as i. to 144, or I to 5O. When the baromeicr indicatci 

 1^ inches we muU then Inbtiaft from that h':ij;ht i-jOth, that is 10 fay, a 

 liae, aiiiil wc fliall have tLc real height, 24 inches ii lines. 



ment 



