May 5, 1904] 



NA TURE 



What is the general process of change in a solution while 

 it is being vaporised? The answer is quite distinct: the 

 residue is al-jiiays less volatile than the original solution, 

 and the distillate more volatile. If there were an example 

 of a solution behaving in the contrary way, then the process 

 of vaporisation at constant temperature would be an 

 explosive one. For the vapour begins to form at a given 

 pressure ; if by this the vapour-pressure of the residue were 

 lowered, the vaporisation would continue of itself at a 

 continually accelerated rate until all the liquid would be 

 vaporised at once. It would be, in other words, a labile 



equilibrium. These equilibria are, however, only mathe- 

 matical fictions, and have no experimental existence. If, 

 on the contrary, the residue has a lower vapour-pressure, 

 then the process is self-limiting, and shows the character- 

 istics of a stable equilibrium. With hylotropic bodies we 

 have an indifferent equilibrium, because the state is in- 

 <lependent of the progress of the transmutation. 



This being granted, we can ask : if we continue the 

 separation of a solution into a less and a more volatile 

 part by repeated distillation, what will finally become of 

 it? Generally considered, two cases may happen. First 

 the residue may become less and less, and the 

 xlistillate more and more volatile, and there is 

 no end to the progress. This case we may 

 •exclude from experimental evidence of a most 

 general character, for we may take it as a 

 general law that it is impossible to enhance 

 any property beyond all limits, even by the 

 lunlimited application of our methods. We 

 must conclude, therefore, that we shall ulti- 

 mately meet with a limit of volatility on both 

 sides, that finally we shall have separated our 

 •solution into a least and a most volatile part, 

 and that both parts will not change further 

 ty repeated distillation. This is a most 

 interesting result, for it means that every 

 solution can be resolved into components, 

 which are hylotropic bodies. For simplicity's 

 sake we have considered only the case that 

 two hylotropic components are generated by 

 •the process of separation ; generally more than 

 two may be formed, but in every case only a 

 limited number of such components is possible. 

 We may formulate therefore as a general 

 Jaw : — 



// is possible in every case, to separate solu- 

 tions into a finite number of hylotropic bodies. 



From the components, we can compose the 

 solution again with its former properties. This 

 is also a general experimental law ; if ex- 

 ceptions seem to exist, it is only because 

 the case is not one of true equilibrium. Still we may limit 

 our consideration to those cases where the law holds good. 

 Then we have a relation between the properties of any 

 ■solution, and the nature and relative quantity of its hylo- 

 tropic components, which admits of only one interpretation. 



NO. I 80 1, VOL. 70] 



Every solution of distinct properties has also a distinct com- 

 position and vice versa. 



If we consider for simplicity's sake solutions of only two 

 components, we may represent any property as depending 

 upon the composition in a rectangular coordinate system, 

 the abscissae giving the composition and the ordinates the 

 value of the property considered. In this way, we get a 

 continuous line of a shape dependent on the particular case 

 chosen. 



If we consider the boiling points of all solutions formed 

 by two hylotropic components, the most simple forms of 

 curves (indeed the only experimental ones 

 known) are given by the types I, II, and III, 

 Fig. 3. For any solution, for example, the 

 solution with the abscissa a, we can foretell 

 its variation on distillation by the slope of the 

 curve. For, as the residue must be less 

 volatile, the residue will change to the ascend- 

 ing side of the curve. This is for I and III 

 to the right, for II to the left side of She 

 diagram. The change of the distillate is the 

 opposite. 



If we try to apply this criterion to the points 

 in of the curve II and III, where there is a 

 maximum and a minimum of the boiling point, 

 we arrive at no decisive answer, for if the 

 boiling point is already the highest possible 

 it cannot rise, and if it is the lowest possible 

 it cannot fall. We are forced therefore to 

 conclude that the boiling point cannot change 

 at all, that is, that this special solution must 

 behave as a hylotropic body. 



This is a well known theorem of Gibbs and 

 Konovaloff, to wit, that a maximum or a 

 minimum, generally spoken of as a dis- 

 point in the boiling-curve, is necessarily con- 

 nected with the property of distilling without change in 

 the composition of the solution. A similar law holds good 

 for the transitions from liquid to solid and from solid to gas. 

 Now this looks like a contradiction ; while a few minutes 

 ago we placed solutions in a class e.xclusive of hylotropic 

 bodies, we have here solutions, that is, mixtures, which 

 behave like hylotropic substances. But the contradiction 

 vanishes if we consider a series of boiling-point curves 

 corresponding to various pressures. We then find that the 

 composition at the distinguishing point does not remain 



^uishim 



constant under different pressures, but shifts to one side, with 

 alteration of pressure. This fundamental fact was discovered 

 and experimentally developed in an admirable way by Sir 

 Henry Roscoe, and has since proved itself a most important 

 criterion in recognising a chemical individual. 



