in the Quarterly Journal of Science, 323 



which is instantaneously dissipated and produces no effect on 

 the fall of temperature. We may therefore put h = i + r ; 

 and, on account of the foregoing equation, we shall have, 



f = *-l. (a) 



The new symbol i stands for the difference between the absolute 

 heat, and the heat of temperature, answering to any change of 

 the bulk ; or it is the heat which enters into the air when its 

 volume increases, and is again extricated when the volume 

 decreases, without affecting the thermometer in either case. 

 I shall call latent heat, that portion of the absolute heat which 

 in ordinary circumstances is not indicated by the thermome- 

 ter, when a mass of air under a constant pressure changes its 

 bulk. This term was introduced by Dr. Black ; but it has 

 been objected to and proscribed by some chemists, who ex- 

 press the same thing by a phraseology certainly not more per- 

 spicuous. There can be no impropriety nor inconvenience in 

 using the term, since it is no more than the unambiguous ex- 

 pression of a fact, the existence of which is certain. It re- 

 mains now to discover the value of the number k; for when 

 this is done, the equation (a), taken always within the assigned 

 limits, contains the whole of the doctrine under consideration. 



In order to find the number k 9 or k — l, we must have re- 

 course to an experiment contrived by MM. Clement and Des- 

 brmes. These celebrated chemists have invented a very in- 

 genious process by which we can ascertain the proportion of 

 the latent heat to the heat of temperature, that is, the value of 



— , when a given mass of air under a constant pressure suf- 

 fers a small variation of bulk. Taking an average of many 

 experiments, it has been found that — and — , that is, /rand 



k— 1, are nearly equal to ~y~ and -| . Now, making &— 1 = §, 

 the equation (a) will coincide with the conclusion which I have 

 stated at p. 94< of this Journal for February 1827, and which 

 enables us to compute the latent heat for a given variation of 

 bulk. MM. Gay-Lussac and Welter repeated the experi- 

 ment alluded to on air under a great variety of pressures, and 

 in a range of temperature reaching from —20° to +40° on 

 the centigrade scale ; and the resulting proportion of the latent 

 heat to the heat of temperature came out in every instance very 

 nearly the same. By this means, not only are the numbers k 

 and & — 1 ascertained to a considerable degree of precision, 

 but it is likewise demonstrated that, within certain limits, they 

 are independent of the state of the air. Now this is a practical 

 proof in favour of the theory we have been explaining ; since, 

 2 T2 as 



