65G 
PROFESSOR W. RAMSAY AND DR. J. SHIELDS ON THE 
In seeking for other evidence pointing to similar conclusions, attention is naturally 
drawn to deductions of molecular weight from the depression in freezing-point of a 
solvent. But, in such cases, the concentration of the dissolved substance is compara¬ 
tively small. Data by E. Beckmann are available' for ethyl alcohol dissolved in 
acetic acid and in benzene, and for acetic acid dissolved in benzene (‘ Zeitschr. Phys. 
Chem.,’ vol. 2, p. 728). The molecular weight of CgHgO is 46, but BECKArANN found 
that a solution of alcohol in benzene, containing 32’45 per cent, of alcohol, froze at a 
temperature which implied a molecular weight of 318; ie., 6’9 simple molecules of 
CgHgO had coalesced to form complex molecules. Here, however, it was not proved 
that the crystals separating from the solution consisted of pure benzene, and not of a 
compound. The same solvent gave a molecular depression, when it contained 22'8 
per cent, of acetic acid, corresponding to a molecular weight of 153 for acetic acid; 
by dividing this number by 60, the molecular wmight of it appears that 
approximately three simple molecules have coalesced to form a complex molecule. 
The molecular weight of alcohol, determined by the depression in freezing-point of a 
solution of 14’2 per cent, of alcohol in acetic acid, gave 58 as the molecular weight of 
alcohol; and this implies a commencement of association. 
Substances which have no, or only a slight, tendency to associate simple to complex 
molecules, however, such as naphthalene (molecular weight, 128), show no such large 
deviations, even in a concentrated solution. Thus Beckmann found the molecular 
weight of naphthalene to be 132, even when 20‘5 per cent, of naphthalene was 
dissolved in benzene. 
That such abnormal results are not unknown, even in the gaseous state, is proved 
by the density of sulphur vapour at temperatures not far removed from its boiling- 
point under normal pressure ; Biltz’s results (‘Zeitschr. Phys. Chem.,’ vol. 2 , p. 920) show 
that whether the formula be accepted as Sg, or whether the association be regarded as 
proceeding to an indefinite limit, the fact remains that association takes place with 
fall of temperature. (See also Thorpe and Hambly’s results with hydrogen duoride 
‘Chem. Soc. Trans.,’ vol. 55, p. 163.) 
As to the reason of such condensation, it is, perhaps, premature to speak. A recent 
paper on Valency, by Flavitzky (‘J. Prakt. Chem.,’ N.F., vol. 46, p. 57), however, 
appears to suggest a clue. It may be that the oxygen in such molecules as acetic 
acid and the alcohols has not exerted its maximum vmlency, on account of the com¬ 
paratively high temperature, and the distance between the molecules when alcohol is 
in the gaseous state ; and that valency increases either with fall of temperature, or 
with closer approach between the molecLdes. , 
A further proof that association is furthered by the liquid condition has been 
furnished by Cundall (‘Chem. Soc. Trans.,’ 1891, p. 1076). The association of nitric 
peroxide molecules (NOg) is greatly promoted by the liquid condition, for while a 1‘44 
per cent, solution in chloroform contains only 0'274 per cent, of NOo, the gas at similar 
temperature and pressure corresponding to such dilution contains 0‘96 per cent. iSfO,,. 
