22 IMPROVED BAROMETERS. 



mercury at M and N ; namely, as the area of the horizonta 

 feclion of the tube D C is to the difference of the areas of the 

 horizontal lections of the tube A B, and of the rod. That is, 

 if D, d, and r be the diameter of the tubes A B, DC, and rod, 

 asd*:D a — r 2 , which is a given ratio, and confequently 

 each quantify has to the fum of both, which is the varia- 

 tion in the length MN a given ratio, e. gr. d 7, (or B 1 -/ 1 ) : 



The variations either at M or N would therefore, if we 

 could meafure them with fufficient accuracy, exprefs truly the 

 variation in the weight of the atmofphere. 

 and lb are thofe g u t if inllead of attempting this, we move the rod by means 

 faHofrherod- of the hair until the mercury return to N, by the length of 

 if it be in every ro d moved either into the mercury, or out of it, we have an- 

 cafe adjufted to Q ^ er me afure, and one as long as we pleafe, of the variations 

 And this mea- in the weight of the atmofphere. 



fure may be en- p or ^ f U pp fi n g the mercury to have fallen at N, anyfpace 

 Proof. Ni, and rifen at M, a fpace My, for the adjuilment we 



mull draw down the rod until it has difplaced as much mer- 

 cury as will fill the fpace N x, and alfo a fpace y x in the other 

 tube of precifely the fame bulk, through which it mull rife at 

 the fame time. The length of rod drawn down for each of 

 thefe purpofes will have a given ratio to the fpace N x, and 

 confequently their fum added to My will have a given ratio 

 to it : and thence, as N x is a meafure of the variation of the 

 atmofphere, this portion of rod will alfo be a meafure : and 

 manifellly it may be increafed at pleafure by lelTening the 

 thicknefs of the rod. 



This appears as follows. The proportion of the rife in one 

 tube to the fimultaneous fall in the other upon any change of 

 the weight of the atmofphere, was above determined to be as 



•p-^ — — : —7%, Let q be the quantity of mercury transferred from 

 one tube to the other, the heights My and N x will be = 

 ttj 1 and ~TT and the fum of thefe is the variation which 



would take place in De Luc's barometer. 

 q 

 The variation My = D *~ r * is feen on tne rod witn <>ut 



q 

 moving it ; upon moving it there is added to this, firft, -jg" 



to 



