August 31, 1911] 



NATURE 



-99 



recombination are infinitely greater than that of the urea- 

 formation, as is undoubtedly the case. Other circumstances 

 make it highly probable that the ions are the active par- 

 ticipants in the transformation, but we may leave the 

 question open, and discuss the results on both assumptions. 

 Suppose, first, that the un-ionised cyanate is transformed 

 \ into urea. Then we have the successive reactions 



NH 4 - + CNO'— NHjCNO— CO(N H 2 )„. 



The -light reverse transformation of urea into cyanate may 

 for the present purpose be neglected, as it in no way in- 

 fluences the reasoning to be employed. 



If the un-ionised substance behaves normally, then the 

 ■conversion of the ammonium cyanate into urea, when re- 

 ferred to the un-ionised substance, will appear unimolecular 

 and obey the law of mass-action : when referred to the 

 ionised substance it will not appear to be bimolecular and will 

 not obey the law of mass-action. 



Suppose, now, that the direct formation of the urea is 

 from the ions. Then we are dealing with the actions 



NH 4 CNO— Nil/ + CNO'— CO(NH„) 8 . 

 Again, let us assume the un-ionised substance to be normal. 

 Once more, if the transformation is referred to the non- 

 ionised substance it will appear as monomolecular ; when 

 referred to the ionised substance it will not appear as 

 bimolecular, as it should if the mass-action law were obeyed. 



It is a matter of indifference, then, so far as the point 

 with which we are dealing is concerned, whether the ionised 

 or the non-ionised cyanate is transformed directly into urea. 

 If the non-ionised cyanate behaves normally the action when 

 referred to it will in either case appear to be strictly mono- 

 molecular. 



11 the ionised cyanate, on the other hand, behaves 

 normally, the reaction when referred to it will be bimolecular 

 and normal ; when referred to the non-ionised cyanate it 

 will not be monomolecular, and therefore will be abnormal. 



The actual experiments show that whether water or a 

 mixture of water and alcohol be taken as solvent, the 

 reaction when referred to the ions is strictly bimolecular ; 

 when referred to the non-ionised substance it is not mono- 

 molecular, i.e., proportional to c„, but rather proportional 

 to a power of c„ other than the first, namely, c,, -1 ' 4 . 



This is, to my mind, a very strong piece of evidence that 

 in the case of the abnormal electrolyte, ammonium cyanate, 

 the abnormality of the ionization equilibrium is to be attri- 

 buted entirely to the non-ionised portion. But ammonium 

 cyanate differs in no respect, with regard to its electrolytic 

 conductivity, from the hundreds of other abnormal binary 

 electrolytes with univalent ions : and I am therefore disposed 

 to conclude that it is to the non-ionised portion in general of 

 these electrolytes that the abnormality is to be attributed. 



A- 1 have already indicated, this conclusion is not alto- 

 gether novel, but in my opinion it has not been sufficiently 

 emphasised. Even in discussions where it is formally ad- 

 mitted that the divergence from the dilution law may be 

 " the non-ionised portion, yet the argument is almost 

 invariably conducted so as to throw the whole responsibility 

 on the ions. The point which ought to be made clear is 

 whether the constant k of the equation 



<* = , 



clt 



or the constant fe' of the reverse equation 



dt 



is really constant. If the former, then the ions are truly 

 normal, and primary explanations of the abnormality of the 

 strong electrolytes can scarcely be sought in high total ionic 

 concentrations and she like, though a connection between the 

 two no doubt exists, both bi ing '1' i< rmined by the same 

 cause. 



In my illustration I have assumed that there holds good 

 a dilution law of the kind given by Storch, of which 

 van 't Hoff's dilution law is a particular case. Here the 

 active mass is represented as a power of the concentration 

 other than the first power. The argument I have used is 

 altogether independent of this special a-sumption ; the active 

 ma>s of the abnormal substance may be any function of its 

 concentration, and the same conclusion will be reached. 



Nernst's principle of the constant ionic solubility product 

 affords additional evidence that the ions act normally in 

 solution. In deducing this principle it is generally assumed 

 that it is the constant solubility of the non-ionised salt that 

 determines the final equilibrium. This assumption, though 

 convenient, is not necessary. The equilibrium is a closed 

 one, thus : — 



j , Ions . , 



/ + 

 Solid salt.: 



;Non-ionised Salt 



The solid is not only in equilibrium with the non-ionised 

 salt but also with the ions. Now, in the deduction of the 

 change of solubility caused by the addition of a substance 

 having one ion in common with the original electrolyte 

 the mass-action law for ionisation is assumed. This is of 

 course justified when we deal with feeble electrolytes, but 

 in the case of salts and strong acids which do not follow 

 the mass-action law the experiments are found still to be 

 in harmony with the theoretical deductions. This is not 

 only so when the two substances in solution are both ab- 

 normal, but also when one is abnormal and the other 

 normal, no matter which is used to produce the saturated 

 solution. In fact, the principle of the constant ionic 

 solubility product may be employed with equal success to 

 calculate the effect on the solubility of one electrolyte of 

 the addition of another electrolyte with a common ion, 

 whether both electrolytes are normal, both abnormal, or 

 whether one is normal and the other abnormal. At first 

 sight, this apparent obedience of abnormal electrolytes to 

 the mass-action law seems strange, but a little consideration 

 shows that if it is only the non-ionised portion of a salt 

 that is truly abnormal, the theoretical result is to be ex- 

 pected. Suppose that the ions do behave normally in the 

 ionisation, then they must also act with normal active mass 

 with reference to the solid, with which they may be re- 

 garded as in direct equilibrium according to the closed 

 scheme referred to above. A change, then, in the concen- 

 tration of any one of the ions, brought about by the 

 addition of a foreign salt with that ion, will necessarily 

 bring about the change in soluBility of the salt calculated 

 from the mass-action law, so far at least as experiment 

 can tell us, for any variation from theory is caused by the 

 change in the nature of the solvent due to the addition 

 of the foreign substance. We ought, then, on the assump- 

 tion that the ions behave normally, to expect that the 

 principle of the constant solubility product would yield 

 results of the same degree of accuracy in dilute solutions 

 whether the electrolytes considered were normal or abnormal. 

 This, as I have said, is actually the case. 



To put the whole matter briefly, in the equilibrium be- 

 tween electrolytes agreement will be obtained between 

 theory and experiment whether we use the mass-action law, 

 or an' empirical law such as van 't Hoff's dilution formula, 

 provided only that we attribute the abnormality to the non- 

 ionised portion of the electrolyte. Thus we can deduce the 

 ordinary formulae for hydrolysis or for isohydric solutions as 

 readily for abnormal as for normal electrolytes, and find 

 the niost satisfactory agreement with experiment in both 

 cases. 



By this one simple assumption, then, for which I have 

 offered some direct justification, it is possible to find a 

 ba :- for call ulation with abnormal electrolytes. The 

 problem of why certain electrolytes should be normal and 

 others abnormal is, of course, in no way touched by this 

 assumption. That is a matter for further investigation and 

 research. 



Another great desideratum of the theory of solutions is 

 to find a general basis for the calculation of hydrates. The 

 present position of the theory of hvdrates in solution may 

 perhaps most aptly be compared to the theory of electrolytic 

 dissociation for solvents other than water. That hydrates 

 exist in some aqueous solutions is undoubted, but no general 

 rule or method exists for determining what the hydrates are 

 and in what proportions they exist. Similarly the theory of 

 electrolvtic dissociation applied to other than aqueous solu- 

 tions affords no general means of determining what the 



NO. 2l83, VOL. 87] 



