28 Mr. W. Sutherland on Weak Electrolytes and 



So far we have neglected possible changes in the alcohol 

 as cause of part of the contraction. The various empirical 

 methods of determining the factor of association for alcohol 

 at 15° C. give values of the order 2 + 0*2. For the most 

 part then alcohol consists of (C 2 H 5 OH) 2 . Possibly some of 

 this changes into C 2 H 5 OH. But my reasons for treating 

 the amount of this change as negligible are first that the 

 change of trihydrol into dihydrol accounts for the maximum 

 contraction, that at the value of p x for maximum contraction 

 a hydrate separates out which must have the composition 

 (C 2 H 5 OH) 2 + 3(H 2 Q) 2 because only dihydrol is present. 

 At this concentration of alcohol then there is probably no 

 C 2 H 5 OH present ; so it seems safe to assume that C 2 H 5 OH 

 is not present at ordinary values of p lt Perhaps in very 

 dilute aqueous solutions of alcohol it may be found. But by 

 the method of the lowering of the freezing-point it is 

 currently held to have been ascertained that in all aqueous 

 solutions of alcohol from p x =- 0*0000185 up to 0'46 the 

 formula of alcohol is C 2 H 5 OH, If my previous arguments 

 are sound, the contradiction here shown proves that the 

 method of the lowering of the freezing-point breaks down 

 for measuring the mass of the associated molecules of 

 alcohol. It appears that in water the two C 2 H 5 OH groups 

 are so loosely joined in (C 2 H 5 OH) 2 that they lower the 

 freezing-point as if they were free. Useful as the method 

 may be in other ways, it is not adequate to decide whether 

 the molecules of solutes in water are loosely associated or 

 not. In connexion with acetic acid we shall see that the 

 matter is important. 



From p 1 = 0-46 to 1*0 the law of the second process, the 

 conversion of (H 2 0) 2 into H 2 is given by the equation 



(A + /O A 2 /2)/ B p 2/9 =0-2979-0-59jt> 2 (p 1 2 + 0-29 i?2 ). (12) 



The coefficients in this equation are so related as not to 

 give a real value of p 2 for the minimum value of the left 

 side, but they nearly allow a minimum close to p 2 = 0'57 to 

 join on to the maximum of (9) at p y z=l— p 2 = 0*46. The 

 equation gives the values of (A + pA 2 /2)/2?ip 2 p with a maximum 

 error of the order 0*8 per cent., and the densities with a 

 maximum error of 3 in 10,000. Its interpretation in a 

 general way is that when p± is greater than 0*46 the 

 equilibrium between dihydrol and hydrol is determined by 

 the occurrence of favourable conjunctions amongst the 

 molecules. A broad approximate way of summarizing the 

 information in Table I. is to take the mixture containing 

 50 per cent, of alcohol as a standard in which the water all 



