president's address-- section a. 53 



4it a liquid-metal boundary affected by the back-pressure of positive 

 ions other than those of the metal itself ? Nernst says " N© " ; but C 

 could never discover his reason. The only surmise I have been able to 

 make on the point is that he must have tacitly assumed Dalton's Law 

 of Partial Pressures to hold in this case. But in view of the mutual 

 solubility of metals, as exemplified by alloys, I doubt the applicability 

 of Dalton's Law. It has always seemed to me to have no more to do 

 with the present case than it has with the mixed vapors of alcohol and 

 water in the presence of one of those liquids. This view is, to my mind, 

 strongly supported by Roberts- Austen's experiments, to which I have 

 already referred. We may fairly suppose that the passage of particles 

 of one metal into the mass of the other takes place ionically — at any 

 rate Nernst would have to suppose so in order to be consistent with 

 himself. But if such ionic transfers between solid metals are possible, 

 similar transfers between a metal and a liquid must be a fortiori possible. 

 It follows, then, that Kahlenberg's result is only what might be reason- 

 ably expected on the basis of the ionic hypothesis, and that Nernst's 

 theory would be improved — even from his own standpoint — if the 

 restriction which he laid down were withdrawn. An important de- 

 duction may be drawn from this argument, viz. : — Calculations of solution 

 pressure are quite unmeaning unless based on measurements of the 

 electromotive force between the metal concerned and a fure solution 

 of one of its own salts. 



The third objection to Nernst's theory is based on the behaviour 

 of non-aqueous solutions, the allegation being that the solution pressure 

 of a metal depends on the nature of the solvent in which it is immersed, 

 and that quite independently of the ionising of that solvent or the 

 formation of complex ions in the electrolytic solution. The examina- 

 tion of the evidence tendered on both sides of this question has been 

 a thankless task, as so much of it proved to be valueless. To take one 

 reason alone : many workers selected pyridine as the solvent ; but 

 their observations must, for the present, be ruled out of court, because 

 the nietals and salts used act chemically on the pyridine. One elaborate 

 research of Kahlenberg's (a) must, on this ground, be discarded. 

 Measuring the potential differences between some 20 metals and 

 solutions of Li CI in pyridine, in water, and in mixtures of the two, he 

 found that the voltaic order of the metals changed from solvent to sol- 

 vent. His argument would be conclusive were it not for the fata 1 

 fact, demonstrated seven years earlier by Lasczinski and Gorski, (6 ) 

 that Li CI combines with pyridine to form the non-electrolyte 

 Li CI + 2CaH5N; the pyridine solution is, in consequence, nearly free 

 from Li ions, and a very bad conductor : indeed, it is difficult to say pre- 

 cisely what it is composed of. We might fairly expect some abnormalities 

 of a scarcely calculable kind in the electromotive force relations of such 

 a liquid, and when we remark that gold in contact with it is positive to 

 iron, we see that we get them ; but these abnormalities neither confirm 

 nor confute Nernst's view — they have simply nothing to do with it. 



(a) Lot. cit. I mention this set of experiments because Kahlenberg attaches 

 special importance to it. See Trans. Int. El. Congr., St. Louis, 1904. 

 (b) Zeitsch. Electro-Chem., IV. 



