214: Speyers — Molecular Weights of some 



tion. The only explanation I can suggest, and one which 

 Zawidski also suggests, is that the two components react chem- 

 ically. 



On reviewing the deduction of van't Hoff's law, it 'seemed 

 that neither the compression of the solution nor the heat of 

 solution of the solute had to be considered. Both of these 

 quantities seemed to cancel out.* It is to be understood of 

 course that Q belongs to a reversible change. It has nothing 

 to do with the heat of a chemical change such as the chemical 

 reaction of a solute with solvent. For example : If Q refers 

 to a heat of dilution, then when we remove or add solvent, 

 which by the nature of the cycle we can do at will, then Q is 

 either absorbed or rejected according to the nature of the 

 solvent and solute, and when we reverse the operation, the sign 

 of Q is reversed. But if Q refers to a heat of chemical combi- 

 nation, then the addition of solvent in the cycle to the solu- 

 tion will produce an evolution of heat and this evolution 

 cannot be compensated in some other operation because of the 

 nature of the cycle. See resorcinol and ethyl alcohol. With 

 this understanding, the application of equation (2) to concen- 

 trated solution seems to be allowed by theory. 



All the systems considered in the preceding papers were 

 composed of liquids, and as there might be some disturbing 

 action overlooked in the van't HofE development when the 

 solute was a volatile liquid, it seemed quite desirable to meas- 

 ure the relative depression of the vapor pressure for concen- 

 trated solutions of non-volatile solutes. 



A few preliminary trials showed that the boiling point 

 method was unsatisfactory, even under reduced pressure,! with 

 such concentrated solutions. The change in temperature was 

 too great. Moreover, theory asserts very positively that the 

 formula for calculating the molecular weight from the change 

 in boiling point is only valid when the heat of vaporization Q 

 does not vary between the two boiling points. Unless these 

 lie close together, or unless the variation of Q is taken into 

 account, the values for the molecular weights so obtained are 

 more or less wrong. Similarly for cryoscopic measurements. 



The requirements were best met by the method of Walker % 

 changing the arrangement a little, for in its original form 

 satisfactory results could not be obtained. § The solution was 

 contained in a bulb tube shown by the full lines in the dia- 

 gram. The solution filled the tubes to about d, about 40 cc 

 of solution. The larger part of the solvent was evaporated in 



* Journ. Am. Chem. Soc., xxi, 725, 1899. 

 f Journ. Phys. Chem., i, 766, 1897. 

 jZeitsch. phys. Chem., ii, 602, 1888. 

 § Journ. Phys. Chem., i, 766, 1897. 



