54: president's address — section a. 



Others of Kahlenberg's experiments are more suggestive, even if they 

 do not prove his point. He finds that solvents of similar chemical 

 constitution give nearly the same electromotive force for a given metal 

 and electrolyte, while solvents of different chemical nature give very 

 different electromotive forces. This would be quite conclusive if the 

 ionisation factors in the different solvents could be assumed as even 

 roughly equal ; but we know that, where such factors have been 

 determined, the assumption would, in general, be wildly erroneous. 

 As Kahlenberg did not measure his ionisation factors, and they seem 

 to be undiscoverable elsewhere, he has only made out half a case. 



Two sets of experiments are, however, conclusive on the point — 

 conclusive against the generality of Nernst's theory. The first of these 

 is due to H. C. Jones,(r) who examined cells of the type 



IN I N . I 



— Ag. NO 3 in alcohol I — Ag. NO, in water I Ag, 

 m \ m I 



(where m had values ranging from 10 to 1,000), in order to measure 

 the difference of potential between silver in alcohol and silver in water, 

 and thus to arrive at the ratio of the solution pressures in the most 

 direct way possible. He used the electromotive forces corresponding 

 to m = 40 and m = 400, because for these concentrations the ionisation 

 factors of Ag.NOg in water are known from Kohlrausch's work, and 

 those in alcohol from that of Vollmer, {d) while the potential difference 

 at the junction of these aqueous and alcoholic solutions is known to 

 be extremely small. To each of the two specified combinations, there- 

 fore, Jones was able to apply Nernst's simplified equation — 



E = -0002 T flog. ~ — log.^^l 



P P 



where everything was known or measured except log. ~. Thus — i 



"2 "2 



was determined without leaving much room for cavil, save for an 



arithmetical slip which I have found in the reductions. Correcting 



p 

 this slip, we obtain for — the values ^\^ and 2 i -0 (e) for the two experi- 

 ments, the difference between the two fractions being within the limits 

 of experimental error, but their average value a long way outside those 

 limits. We see, then, that the apparent value of the solution pressure 

 of silver in water is at least 200 times that in alcohol. Jones, however, 

 failed to consider two alleged possible explanations of his results. I 

 venture, therefore, to supply the omission, (a) Can the result be due 

 to the presence of unknown complex ions ? To answer this we must 

 see what effect such ions could produce. The effects will differ in diffe- 

 rent cases ; but very little reflection is needed to indicate their general 

 nature, which is, luckily, all that we need in this case. If the solution 

 pressure of a metal be large compared with the osmotic pressure of the 

 positive ions supposed to exist in the electrolyte, the presence of un- 

 known complex ions will cause it to be underestimated, since we should 



(c) Zeitsrh. Phys. Chem., XIV. {d) Inaug. Diss, 



(e) Jones gives xa and rs ; but this is due to the slip in arithmetic. 



