January 30, 1902] 



NA TURE 



30; 



The molecular weight of MgCl„ and of BaClo decreases with 

 increasing concentration until it becomes les-^ >han one-third the 

 theoretical value, but the conductivity in both cases increases 

 with the dilution. 



HgCU shows no ionisation by the boiling-point method. The 

 molecular weight increases with concentration. It might be 

 assumed that polymerisation takes place and, further, that some 

 of the molecules which are not yet polymerised are ionised and 

 thus account for the conductivity, which, though low, increases 

 with the dilution. 



In the case of KClOi, KNOjand AgNOj the molecular weight 

 increases with concentration, and there is good agreement with 

 the conductivity measurements. Thus AgNOj in the most 

 dilute solution tested appeared to be ionised to the e.xtent of 65 

 per cent., while the conductivity method showed 67 per cent. 

 For a normal solution the boiling-point method indicated 54 per 

 cent, and the conductivity 52 per cent. The agreement is 

 closer at the boiling point than at the freezing point. 



In the case of MgSOj the molecular weight begins, in the 

 dilute solution (2733 gm. in 100 gm. of water), with a value 

 above the theoretical, indicating no ionisation ; then it increases 

 with the concentration, and finally decreases after passing through 

 a maximum, the values in the concentrated solutions becoming 

 less than the theoretical. But there is no irregularity in the con- 

 ductivity values. What has been said of MgSOj applies also to 

 ZnSOj, Ni.SOj and CuSO^. The same general behaviour is also 

 exhibited by MnSOj, CdSOj, CoSOj and F'eSOj, except that 

 the molecular weights of these salts, while first increasing and 

 then decreasing with increase of concentration, always remain 

 above the theoretical values. The molecular weight of the sul- 

 phates is less by the freezing-point results than by the boiling- 

 point method. So that if it be assumed that the molecules are 

 polymerised, this polymerisation is greater at the higher 

 temperature. 



A series of boiling-point determinations was made on a 

 solution of cane sugar, as an example of a non-electrolyte. It 

 was found that the molecular weight diminished appreciably as 

 the concentration increased, becoming less than the normal 

 (212 in a solution of 2894 gm. in 100 gm. of water, as com- 

 pared with the normal 342). But, as is well known, the solution 

 does not conduct. A test with Fehling's solution showed that 

 no invert sugar had been formed by the boiling. Solutions of 

 HjBO;,, on the other hand, show practically constant molecular 

 weight with varying concentration. 



Discussion ol Results. — From the above results it appears that 

 there are solutions which are excellent conductors and which, 

 nevertheless, show a normal molecular weight of the solute. 

 While in some cases the molecular weight increases with the 

 concentration, thus agreeing qualitatively at least with the 

 ionisation theory, in other cases the molecular weight decreases 

 with increase of concentration, finally becoming less than what 

 it ought to be even for complete ionisation. In other cases the 

 molecular weight at first increases with concentration and then 

 diminishes. But the conductivity of these solutions continually 

 increases with dilution. There are cases, however, in which the 

 conductivity at first increases with the dilution and then 

 decreases, e.g. aqueous solutions of the alkaline hydroxides. 



It follows, therefore, that there is no such connection between 

 freezing points and boiling points of solutions on the one hand 

 and their conductivity on the other as is claimed by the ionisa- 

 tion theory. Often there is not even a qualitative agreement. 

 Want of agreement is to be found in the original table of 

 Arrhenius, but this was ascribed to experimental errors. 



Various properties of electrolytes have been explained by the 

 ionisation theory. Thus the various additive properties of salt 

 solutions are presented as .supporting the theory. But the 

 theory cannot be based on additive properties of this kind, for 

 such are known to exist in the case of true chemical compounds, 

 where, since there are no solutions under consideration and 

 since there is no electric conductivity observable, the possibility 

 of ionisation is out of the question. In the realm of physiology, 

 also, the theory cannot cope with the facts. 



The heats of neutralisation of acids and bases have been used 

 as an argument in favour of ionisation ; Crompton, however, 

 has shown that the theory is not only unnecessary, but that it is 

 inadequate. Again, the theory cannot be brought into harmony 

 with the law of mass action, which is one of the strongest 

 arguments against it. 



The chemical reactiveness of electrolytes has been explained 



by attempting to ascribe to the ions a peculiarly strong chemical 

 activity on account of the electrical charges that are supposed to 

 reside upon them. In this connection attention is drawn to the 

 action of water in frequently facilitating chemical action. While 

 this fact may be in agreement with the ionisation theory, it 

 cannot be used to support it ; for there are many pure substances 

 and mixtures that are very active, although there is no ground 

 for assuming the presence of ions ; t'.^'. many explosives. It is 

 well recognised that many bodies unite with the solvent, and 

 interaction then takes place between the new products, reactions 

 taking place which might easily not occur between the original 

 anhydrous bodies. 



It has been supposed by Nernst and by J. J. Thomson that the 

 higher the dielectric constant of a solvent the greater its ionising 

 power. But many exceptions are now known, e.g. liquid NHj, 

 butyronitrile and pyridine (H. Schlundt), liquid SO., vWalden), 

 liquid HCN, and amylamine. 



That the ionising power of solvents is dependent upon the 

 polymerisation of their molecules, as claimed by Dutoit and 

 Aston, has been shown to be incorrect in many cases by 

 ICahlenberg and Lincoln. 



The ionisation theory is at its best in explaining electrolysis, 

 but there are many phenomena which the theory does not 

 explain. For example, why are the deposits of silver from 

 some solutions poorly adherent and from others dense and 

 well adhering, the potential difference and current density being 

 the same ? 



.\s to the ionisation theory being required by thermodynamics, 

 Clausius, who showed the discrepancy in the Grotthus theory, 

 did not find it necessary to put forth such a radical hypothesis as 

 that of Arrhenius. Nor did Hittorf find it necessary to frame 

 such a theory. Finally, no marked improvements or discoveries 

 in electrolysis are due to the theory. It has led to Nernst's 

 theory of the E.M.F. of galvanic cells and a formula which really 

 involves the assumption that the law of mass action is 

 applicable to electrolytes in the sense required by the ionisation 

 theory. That this law does not hold has already been men- 

 tioned. By maintaining the correctness of this formula and 

 thus assuming that the law of mass action holds for electrolytes, 

 Jahn has arrived at the conclusion (as clearly he must) that the 

 ratio of the equivalent conductivityat a given concentration to that 

 at infinite dilution does not correctly represent the degree of 

 ionisation, and that the ionic velocities vary in dilute solutions. 

 This has given rise to a discussion in the Zeils. PItys. Chem. 



If the ionisation theory is not true, then the original difficulty 

 with the van't Hoff theory of solutions recurs, viz. the 

 theoretical interpretation of the factor i in the gas equation. 

 Of course this equation is supposed to hold strictly only for 

 ideal gases. A normal solution, however, is rather dilute for 

 many of the practical purposes of life. Not that one expects the 

 gas equation to hold strictly for a normal solution, but what one 

 has a right to expect from the modern theory of solutions is 

 that, with increasing concentration, a solution should behave at 

 least qualitatively as a gas does with increase of pressure. The 

 ionisation theory does not satisfactorily explain the significance 

 of the factor i. In any case this factor should never be placed 

 equal to unity without experimental evidence, whether in con- 

 nection with electrolytes or non-electrolytes. 



Substances of similar chemical composition, when dissolved in 

 the same solvents, behave similarly so far as boiling points or 

 freezing points are concerned ; this shows that the influence of 

 the chemical nature of the solute aft'ects these variations. 



The analogy between gases and solutions has been pressed too 

 far, so that it has been forgotten that we are dealing only with 

 an analogy. The solution of a substance and the expansion of a 

 gas are really very different. A gas will expand in vacuo or mix 

 with any other gas, but a substance will not dissolve in every 

 liquid. And here lies the difficulty of the theory. It neglects 

 the all-important role of the solvent. It fails to emphasise the 

 fact that the process of solution takes place because of a mutual 

 attraciion between solute and solvent, and this attraction is the 

 essence of the so-called osmotic pressure, which is closely related 

 to, if not essentially identical with, chemical affinity. The 

 attraction between solvent and solute should be recognised. 

 Each solution should be e.^amined separately, beginning with 

 the most concentrated, the behaviour of the most dilute solutions 

 appearing as a limiting case ; then we shall see the present 

 theory of solutions in its true relation to the facts. 



W. R. C. 



NO. 1683, VOL. 65] 



