Of BAROMETRICAL MEASUREMENTS. 91 



exactly the law which connects thefe variations, and we mufl 

 have recourfe to reafoning, in order to fupply this defect. Let 

 us fuppofe that air of a given temperatiire, for inflSnce, of 32, 



by the lofs of one degree of heat, is contracted or the part m of 



its whole bulk; its bulk, therefore, when of the temperature 3 1 , 

 will be i m. By the lofs of another degree of heat, its tem- 

 perature will be reduced to 30, and its contraction will not 

 be m, as before, b\\t m(im), which, fub tracked from I m, 

 its bulk, when of the temperature 31, will give its bulk when 



of the temperature 30, = I 2m+m z (i mf. In like man- 

 ner, after the lofs of 3 of heat, the bulk of the fame given 



quantity of air is {hewn to be (i m) ; and, in general, its bulk 

 is as that power of i m, which is denoted by the difference 



between 32 and the given temperature. If, therefore, b be 



32 h 

 the heat of a given quantity of air, (i m) will be the 



fpace occupied by that air, fuppofing always that the compref- 

 fing force is given. 



6. THIS formula affigns a finite magnitude to the air as long 

 as the diminution of its heat is lefs than infinite ; for as 



i m is lefs than unity, when h becomes negative and infi- 



32 A 

 nite, ( i m) becomes then, and not till then, = o. When 



3 2-A 



b is affirmative, and greater than 32, (i m} becomes 



greater than i, and increafes continually, being infinite 

 when h is infinite. When 32 b is not very great, then 



3 2-b 



(im) =!+(,& 32) m nearly, which agrees with the hy- 

 pothefis of uniform contraction and dilatation in moderate 

 temperatures. 



THIS formula alfo reprefents, with tolerable exactnefs, the 

 experiments which General ROY made with the manometer, ex- 



M 2 cepting 



