222 Speyers — Molecular Weights of 



some 



Aeenaphthene in propyl alcohol. 7 per cent aeenaphthene. 

 Association is indicated. Corrections very large. Neither 

 formula to be preferred. 



Naphthalene in propyl alcohol. 10 per cent to 22 per cent 

 naphthalene. The molecular weight is below the normal at 

 29'8° but the corrections are very large. At 45° association is 

 strongly indicated, though the corrections are too large for this 

 conclusion to be altogether satisfactory. Neither formula to 

 be preferred. 



Naphthalene in toluene. 35 per cent to 54 per cent naph- 

 thalene. The molecular weights are normal by formula (1) 

 but abnormal by formula (2). Considerable association is indi- 

 cated by the second formula. The corrections are very small. 

 Formula (1) much to be preferred. 



Aeenaphthene in toluene. 17 per cent to 45 per cent aee- 

 naphthene. The average molecular weight is 151 ± 3 by 

 formula (1) but quite abnormal, 172 ± 5, by formula (2). Asso- 

 ciation is indicated by the latter formula. The corrections are 

 small. Formula (1) is much to be preferred. 



In not a single case then does formula (2) give very satis- 

 factory results, whereas in several cases formula (1) does so and 

 in no case is it less satisfactory than formula (2). 



Now formula (1) is altogether too arbitrary. What right 

 have we to use the molecular weight of the solvent in the 

 gaseous state in preference to the molecular weight in the 

 liquid state. In formula (2) the molecular weight in the 

 gaseous state is to be used according to van't IToft's theory. 

 We have good reason to think that the molecular weight of 

 water in the liquid state is twice or perhaps four times 18. 

 Were we to put some such value for M in (1), the results would 

 be very objectionable. We can, however, justify the normal 

 value for M, that is its value in the vapor state, by adopting a 

 view published some years ago*; namely, that a liquid giving 

 oft* vapor contains some vapor in the dissolved form, or other- 

 wise expressed, that the simple molecules corresponding to the 

 vapor molecules are formed inside the liquid instead of at the 

 surface. So long as the solvent gives oft simple molecules in 

 the vapor state, so long there are dissolved simple molecules in 

 the liquid, and these are the molecules whose concentration is 

 changed by the solute. The complex molecules produce the 

 simple dissolved molecules but have no direct vapor pressure 

 of their own. 



Let us consider a liquid consisting of simple molecules only 



whose vapor pressure is p. To this liquid, equation 3 applies 



directly. Now suppose that the liquid is mixed with so much 



of some other inert, non-volatile, liquid that the vapor pressure 



* Journ. Am. Chem. Soc, xviii, 724, 1896. 



