r 449 ] 



3tio. Let the thermometrical degrees above 32° (that is their 

 difference with 32°) = D. and let 1000 N = », then D x — », and 

 X This eives the coefficient fought. 



Example. 



Thus in the firft experiment, barometer 29,3, thermometer 53^, 

 W = 302,89 grains, and t = 322,1 grains. Now fince 322,1 grains 

 denote 1000 cubic inches, 302,89 at the fame temperature fhould 

 denote 940,36 inches = N. and 1000 — 940,36 = 59,64, the 

 number of inches gained by 21° of heat; therefore putting * for 

 the coefficient, 2 1 ar = 59)64, and x = 2,84. 



In the fecond experiment I find, by fimilar reafoning, the co- 

 efficient to be nearly 2,000, but as this relates to air much drier 

 than any hitherto found in the atmofphere, it is only ufeful fo 

 far as to fhew that the moifter air is the higher, muft be the co- 

 efficient. It varies therefore according to the ffate of the air as to 

 moifture between 2. and 3. 



In the fourth experiment, hygrometer 75°, I found it 2,77, but 



with us the hygrometer is generally above 80^. If, therefore, the 



hygrometrical mean be unknown, I take it that 2,5 is, at leaft of 



round numbers, the fafeft coefficient. 



To 



