txiith Elastic Fluids. 167 



ture afterwards communicated. Now by both processes the 

 air is ultimately brought to the same condition, and conse- 

 quently it must have acquired the same quantity of heat. It is 

 therefore proved that the whole heat acquired by air which 

 expands under a constant pressure, is composed of two inde- 

 pendent parts, namely, the latent heat and the heat of tenqje- 

 rature. 



Let us next compare the dilatation of a mass of air under a 

 constant pressure, or, which is the same thing, an air-thermo- 

 meter, with the indications of a mercurial thermometer : tlie 

 whole heat acquired by the air will be proportional to the 

 ascent of the mercury or to the increase of the bulk of the 

 air : and again, according to what is shown above, the heat of 

 temperature will likewise be proportional to the ascent of the 

 mercury, or to the increase of the bulk of the air ; wherefore 

 the difference of these two heats, that is, the latent heat, must 

 be proportional to the ascent of the mercury, or to the in- 

 crease of the bulk of the air. It thus appears that the three 

 heats, namely, the whole heat acquired by the air and its two 

 parts, the heat of temperature, and the latent heat, receive, 

 each, equal additions for equal increments of the bulk of the 

 air; and consequently for any given dilatation, they will al- 

 ways bear the same constant proportions to one another. And 

 this must be admitted as true for a very extensive range of 

 temperature, or so long as thermometers of air and mercury 

 continue to measure heat exactly. 



When air expands under a constant pressure, a rise of one 

 degree of Fahrenheit's thermometer has been found to cor- 

 respond to an increase of volume equal to ^^^th of the bulk 

 possessed by the air at the freezing of water. It follows, there- 

 fore, that, in order to know the latent heat absorbed in any 

 dilatation, or disengaged in any condensation, we have only 

 to investigate the invariable proportion it bears to the heat of 

 temperature capable of producing the same change of volume 

 under a constant pressure. Now this invariable proportion 

 has been deduced in the last Number of this Journal, from an 

 experiment of MM. Clement and Desormes, and it comes out 

 equal to \ nearly. Such is the nature of the experiment men- 

 tioneil, that it leads to a proportion rather below the truth ; 

 but we may correct the result by the velocity of sound in the 

 atmosphere, which agrees better with | than §. Hence it ap- 

 pears that, when air expands under a constant pressure, the 

 whole heat it acquires for any increase of volume, ihe heat of 

 temperature and the latent heat, are as the numijers 11, 8, 3, 

 or, more nearly, as 7, 5, 2. If we apply to u thermometer of 

 Fahrenheit's construction a scale having the distance between 



the 



