486 J. H. van't Hoff on the Origin 



III. Deviations, 



The question now arises, Is the foregoing sketch to be 

 considered as a theory of solution ? 



Not necessarily. It is rather a necessary correlation of 

 experimental results and laws, some of which have been known 

 for a long time, whilst others are of quite recent origin, and 

 I scarcely think that serious objections can be raised against 

 this correlation. 



It is not even necessary to select osmotic pressure as the 

 starting-point; the whole may be just as easily deduced from 

 Henry's or Raoult's law, but the osmotic law is a simple 

 tangible expression of the whole behaviour of dilute solutions, 

 and its physical meaning can be easily stated in words, and 

 graphically represented. 



When rarefied matter exists under such conditions that it 

 tends to spread itself out by diffusion, then the pressure, at any 

 given temperature, which prevents it from so doing depends 

 only on the number of molecules and is independent of their 

 nature and of the medium in which they exist. 



Where does the theory of solutions begin ? In my opinion 

 it begins where the attempt is made to explain the deviations 

 as being due to secondary causes. 



Such deviations exist, and immediately show themselves by 

 non-agreement with all the above correlated laws, for example, 

 in abnormal values of the osmotic pressure, deviations from 

 Henry's law, unexpected lowerings of freezing-point, &c. 



In now passing on to the real theory of solutions, my object 

 is to mention shortly in what way these deviations have been 

 explained, and with what success. 



One thing in particular is to be noted : the above-mentioned 

 relations, if they are to remain at all valid, necessitate only 

 that we are dealing with great dilutions, and are thus limiting 

 laws like those of Boyle, Gay-Lussac, and Avogadro. But it 

 is well-known that exceptions are also to be met with in the 

 case of very dilute solutions. Some lowerings of the freezing- 

 point are unexpectedly small, whilst others are unexpectedly 

 great. That this does not go so far as to conceal the main laws 

 is best proved by the fact that Baoult, long before the theory 

 of solutions, always referred to the present calculated values 

 as " normal." 



The Small Lowerings of the Freezing-point, corresponding 

 to a smaller osmotic pressure, are of two kinds. 



First, and especially in the case of compounds containing 

 the hydroxyl group, e. g. the alcohols and acids in benzene 

 solution, values are obtained which are obviously only about 



