M. Poisson on the Caloric of Gases and Vapojirs. 333 



from which we may determine k by giving to 6 any value for 

 which the corresponding one of H has been settled by obser- 

 vation. For example, by the table of M. Biot's Traite de 

 Physique, tome 1, p. 531, deduced from the experiments of 

 M. Dalton, H="''0887'i2 when 6 = 50°; and therefore the 

 preceding equation gives 



k-\ 



— = •0683 and ^'=1-073. 



k 



By employing values of H correspondmg to other values of fl 

 comprised between 0° and 100°, the value of h will scarcely 

 differ from the preceding by a hundredth at most, or a two 

 hundredth at least. We shall therefore retain this value of 

 k, to which joining the preceding values of C and y our for- 

 mula (8) becomes 



Q = 650+(-847) [ (266-67+9)(-Y-) -see-e?]' (^^^ 

 The application of this formula to temperatures far distant 

 from 100° shows us that the quantity Q varies but very little 

 in the case of saturation or when H = /2. For 9 = 0°, we have 



H = 5-059; whence Q = 658. For 9=— 19°-59 M. Gay- 



Lussac has found H=l-'3718; whence Q = 662. When 

 fl=140° many philosophers agree in giving to H nearly four 

 times its value at 100°, or four times '»-76; whence we get 

 Q=653. Again, M. Christian makes H nearly twice the last 

 value or eight times '»-76 when 9= 170° ; from which Q comes 

 out 661. These values of Q, as we perceive, differ but very 

 little among themselves, though they have ranged over a tem- 

 perature of nearly 200°, and a tension of vapours from almost 

 nothing to eight atmospheres *. This result shows that k in 

 the case of aqueous vapour is but very little greater than unity; 

 but we cannot, as we have shown above, suppose it precisely 

 equal to unity. We should not forget that Q is not sensibly con- 

 stant unless when the tension or vapour is a maximum. When 



» The evidence in favour of his formula which M. Poisson here adduces 

 in the supposed constancy of Q is illusive. It all results from the high 

 value which Q happens to have. Where would have been the evidence 

 had Q happened to have a much less value ? for instance, a value of about 

 .3 or 4 or even 10 ! I Did probability belong to the views producing tins 

 theorem, the coefficient of y being once nearly = should deviate but very 

 little from it. Its different values even under the range of temperature 

 M. Poisson mentions, have ratios from nothing to infinity. A greater proof 

 of the propriety and justice of my objections cannot be adduced than m 

 the very erroneous values of the tension H immediately following. No- 

 thing, it appears to me, can be a stronger argument of the insufnciency of a 

 theory, than the same formula in one instance coming up nearly to obser- 

 vations, and in another instance closely connected running almost m di- 

 rect opposition to thcin, — J. H. 



