94 Journal of the Mitchell Society. [June 



ula then becomes too low it must be because the vapor pres- 

 sure curve as calculated from Biot's formula is below the true 

 vapor pressure curve. We accordingly subtracted the observed 

 vapor pressures from the calculated in the immediate neigh- 

 borhood of the critical temperature. (These results are 

 shown in detail in Table 2, Jour. Phys. Chem. 9, p. 402 [1905]). 

 In every case except normal octane, stannic chloride, and the 

 associated substances (methyl alcohol, ethyl alcohol, and 

 acetic acid), the calculated minus the observed vapor pres- 

 sures does give negative differences at the highest tempera- 

 ture compared (usually the critical temperature itself). With 

 regard to these five exceptions, the associated substances can- 

 not give exactly correct results by the use of equation 4 and 

 their evidence does not therefore bear upon that equation. 

 Normal octane and stannic chloride are the only substances 

 whose divergence can be considered as evidence against the 

 conclusion above derived and it is sufficient to point out in 

 explanation a remark made by Dr. Young 1 when the vapor 

 pressures of normal octane were published. He there states 

 that the observed values of the vapor pressure above 280° C 

 for normal octane, stannic chloride, and acetic acid, are prob- 

 ably too Io-ju owing to an error in the temperature scale, i. e., 

 in the boiling points of the liquids used as a heating jacket. 

 A close examination of the observed and calculated vapor 

 pressure curves reveals the fact that no matter how the con- 

 stants for Biot's formula be altered they cannot exactly repre- 

 sent the true vapor pressures in the neighborhood of the crit- 

 ical temperature. The deviation is very slight, usually neg- 

 ligible when the vapor pressure is considered, because the 



8 P 

 proportional error is very small. But when the -=-=- is con- 

 sidered the proportional error is very large. 



Since Biot's formula was empirical and in the immediate 

 neighborhood of the critical temperature was forced to fit a 

 curve it could not exactly follow it usually happens that the 



i Jour. Ohem. Soc., 77, 1147 (1900). 



