THE NEW THEORY OF SOLUTIONS. 389 



According to Briihl the reason why oxy-solvents do not 

 as a rule yield associated solutions is due to their unsatur- 

 ated nature ; oxygen is here regarded as playing the part 

 of a tetravalent atom, the "free valencies " of which being 

 the real cause of the dissociative action of the solvents on 

 molecular aggregates. 



Comparison of abnormal molecular weights given by 

 different methods. — It follows from the preceding discussion 

 that, assuming Avogadro's hypothesis applies to dilute 

 solutions, in certain solvents the molecular weights of 

 certain dissolved substances are not identical with their 

 gaseous molecular weights. The alcohols which have 

 normal gaseous molecular weights give abnormal solution 

 molecular weights. Both the gaseous and solution molec- 

 ular weights of the acids are abnormal, but the two are 

 not identical under conditions which, according to the 

 gaseous analogy, are similar. Thus Nernst calculates that 

 in solution in benzene, at 8o°, acetic acid is dissociated to a 

 greater extent than when it exists as gas at the same 

 temperature and concentration ; the reason he gives being 

 that the dielectric constant of benzene (2*2) is greater than 

 that of free space (1). This reasoning would not apply, 

 however, to the molecular weights of the alcohols in solu- 

 tion and in the state of gas. Indeed, if we deal with 

 substances which frive abnormal molecular weights when 

 gasified, in no case are these identical with the solution 

 molecular weights, and the latter vary with the solvent. 

 For example, chloral hydrate and alcoholate are partly 

 dissociated on vaporisation, but when dissolved in benzene 

 they give practically normal molecular weights in dilute 

 solutions. I n acetic acid they are partly dissociated into their 

 proximate constituents in dilute solutions, and in water 

 chloral hydrate gives normal values while the alcoholate is 

 almost completely dissociated at all concentrations up to 

 18 per cent. The conclusion to be drawn from the fore- 

 going facts is, therefore, that if the dissolved substance has 

 an abnormal gaseous molecular weight, or if in the liquid 

 state it consists of molecular aggregates, the solution 

 molecular weight differs from the gaseous molecular 



