. 



126 CAUSES which affeEl the ACCURACY 



mined, by comparing the formula that has been given here 

 with obfervations. But this would lead into difquifitions far 

 exceeding the bounds of the prefent inquiry, the object of 

 which is, to afcertain the form, rather than the abfolute quan- 

 tity of thefe corrections. 



41. IT is evident, that, in the preceding inveftigation, as 

 well as in all the other methods of meafuring heights by the 

 barometer, it is fuppofed, either that the one of the barometers 

 is vertical to the other, or that a perfect aequilibrium prevails 

 through that part of the atmofphere intercepted between 

 them. The determination of the conftant quantity in the fore- 

 going integrations, by fuppofing that b j3 when x o, or that 

 the mercury in the two barometers ftands at the fame height in 

 them, when they are at the fame diftance from the furface of 

 the earth, obvioufly involves in it either the one or the other 

 of thefe conditions. But the laft of them, the asquilibrium of 

 the atmofphere, never takes place ; and, therefore, it is necefla- 

 ry, in order that barometrical meafurements be perfectly accu- 

 rate, that the one barometer be immediately above the other, 

 or, at leaft, that the horizontal diftance between them be very 

 fmall. If this be not the cafe, the unequal diftribution of the 

 heat through the different parts of the fame ftratum of air will 

 render it impomble to deduce the difference of the heights of 

 the barometers from a comparifon of the columns of mercury 

 contained in them. 



FOR inftance, let there be three barometers; the Jirji at 

 the furface of the earth, the fecond raifed up into the air per- 

 pendicularly above the^^r/?, and the third removed into a colder 

 climate, but raifed up alfo into the air, fo as to have in it a co- 

 lumn of mercury of the fame length with that in the fecond. 

 Thefe two laft, when compared together by M. DE Luc's, or 

 by the preceding rules, will appear to be at the fame height 

 above the furface, or above the firft barometer. But, if each 

 of them be compared with the Jirfl^ the fecond will appear 



more 



