[ "t ] 



the two furfaceS, fp is the apparent variation in the tube to the 

 fum of the variations in the tube and box, which gives the 

 true variation. But as applying this corredion to all the 

 feveral variations in a feries of obfervations would be trouble- 

 fome and tedious, I think it would be much betcer to form, 

 at once a fcale which fhould need no correcflion ; and this 

 may' be done by reducing the common fcale of inches, that is, 

 by making a fcale whofe divifions fhall be lefs than the cor- 

 refpondent divifions of the common fcale, in the fame proportion 

 that the apparent variation in the tube of your barometer is 

 lefs than the true one ; and this proportion is always conftant 

 in the fame barometer ; for it is that proportion which the 

 furface of the mercury in the box bears to the fum of its 

 furfaces in box and tube. If this contra<5led fcale be annexed 

 to the tube of the portable barometer, it is evident that, when 

 the mercury has varied through any of the contradled divifions 

 of this fcale, it will have varied, at the fame time, through 

 the correfponding divifions of the common fcale annexed to 

 the tube of a fixed barometer. Therefore this con cradled' 

 fcale will always point out the variations and altitudes of the 

 mercury truly, or fuch as the common inch-fcale fliews them 

 to be at the time in a fixed barometer whofe box is of the 

 largeft dimenfions. To illuflirate this by an example : fuppofe 

 that in a portable barometer the furface of the mercury in 

 the box is to that in tiie tube as 49 to i, then it will be to 

 the fum of the furfaces as 49 to 50 ; and when the mercury in 

 the tube falls through A-I of an inch, it will rife in the box y'- ; 

 io that its true fall, at that time, will be one inch. If then 

 Vol. V. Q^ to 



