Velocity of Sound, 133 



temperature could not be thereby raised from 32*^ to 212° ; nor 

 could the ignition of tinder be caused by the heat evolved from 

 air on its being condensed. 



But though such results be out of all shape, yet the law now 

 stated is, if possible, still more at variance with itself than with 

 facts ; for we soon obtain a very different result, by taking the 

 condensation at several successive steps. Thus, by doubUng 

 the density of air at 32° Fah., or putting t = 32° and g = 2, 

 we obtain i = 90°, which raises the temperature to 122°, 

 Doubling again the density with t = 122° and § = 2, we get 

 i = 106°.875. The density has thus been only quadrupled, 

 and yet the rise of temperature, viz. 90° + 106°.875 =: 

 196°.875, exceeds the rise due to compressing all the air of the 

 atmosphere into a point, at one operation *. I may remark, 

 by the bye, that so far as is yet known, an elastic fluid cannot 

 be liquefied by condensation, if it lose no heat. 



If the air, which, by having its density quadrupled, has 

 attained the temperature of 228°.88, be now restored to its 

 original density, ought it not to be, in every other respect, re- 

 stored to its former state ? and, in particular, if it have neither 

 gained nor lost heat, ought it not to resume its original tempe- 

 rature of 32° ? So far, however, from this being the case^ we 

 shall find that by restoring the original density, the resulting 

 temperature will be lower than 32° by the enormous quantity 



of 564°.6. Thus, putting r = 228°.88 and /? = -^ , (for the 



law being general, must suit any initial temperature or den- 

 sity,) we obtain i = — 76I°.48. Hence, in place of 32°, the 

 resulting temperature is 228°.88 - 761° 48 = — 532°.6 Fah., 

 or 84°.6 below the absolute zero of those who, with Mr. Ivor}^, 

 insist on an air-thermometer being the standard ! 



What a striking contrast between such absurdities and the 

 following formula, in which not the slightest inconsistency can 

 be detected, viz. 



i = (448° + t) (p^ - 1). 



* It is obvious that we may proceed in strict conformity to this law, and 

 yet vary the result considerably ; nay, in some cases, enormously, at 

 pleasure, according to the number or magnitude of the parts into which 

 we subdivide the whole change of density as taking place at successive steps, 



