437 



that the sum of its sensible and latent heat is at every tem- 

 perature a constant quantity, — equation (II) becomes 



/" =/' - -01135 (I - O X J -^f- ; (IV) 



and equation (III) becomes 



/" =/' - -01017 (t - f) X ?^f-. (V) 



" The theory which has led to these conclusions is now 

 universally admitted to be correct ; but as doubts may 

 be entertained respecting the exactness of the coefficient, 

 whose value, as has been seen, depends on the numbers by 

 which a and e are represented, (numbers which are, in all 

 probability, not as yet known with great precision,) it would 

 appear desirable to deduce its value directly from experi- 

 ment. This is the immediate object of the present commu- 

 nication. 



" In my second paper on the dew-point, I have given three 

 distinct series of experiments, applicable to such a purpose; — 

 the first relating to air whose dew-point was determined by 

 Daniell's instrument ; the second to air perfectly dry ; and 

 the third to air whose dew-point is known with certainty, 

 and without the aid of any form of condensation hygrome- 

 ter. From these, in all of which t' is greater than 32°, I 

 have calculated 5i values of the coefficient, by methods to 

 the explanation of which I now proceed. 



" 1. Representing the coefficient in question by in, the 

 hygrometric formula becomes 



f"=:f-m(t-t<)X^f. 



Now if air, in reference to which t and t' have been accu- 

 rately noted, be raised to any higher temperature, and the ob- 

 servation repeated, we obtain data for determining the value 



P — f 

 of m. For/" being constant, /' - m (t - t') X ^~^~ > for 



p — f' 

 one observation, will be equal to f' — m (t — t') X — ^— 



