president's address — SECTION A. 55 



imagine the back pressure of the ions in solution to be greater than it 

 really is. If, on the other hand, the solution pressure be relatively 

 small, we should be led to over-rate it for a similar reason. Now the 

 solution pressure ascribed to silver is one of the smallest we know, 

 being of the order 10'^'^ atmospheres — enormously smaller, that is to 

 say, than the back pressure of the ions in an aqueous or alcoholic 

 solution of silver nitrate even if it be as Aveak as one four-hundredth 

 of normality Again, Ag.NO,,, in aqueous solution, gives rise almost 

 exclusively to simple ions. ('') So we are restricted to the possibility 

 of the existence of such ions in the alcoholic solution. But their presence 

 there is very unlikely, as alcohol is a more associative substance than 

 water, and it is now generallv recognised that the formation of complex 

 ions becomes increasingly difficult as the degree of association of the 

 solvent increases. Still, they might be there. The important fact is 

 that, if they were there, they would only strengthen the case against 

 Nernst's theory. We should need to increase the ratio of the solution 

 pressures — not to diminish it — in order to include them. (6) Can the 

 result be due to ionisation of the alcohol '. If so, we must postulate 

 an osmotic pressure of positive ions derived from this source amounting 

 to at least 420 atmospheres. This would involve the ionisation of 

 considerably more than the whole of the alcohol present. It is rather 

 too large a demand on our credulity to ask us to believe that the solution 

 of less than half a gramme of Ag.NO,, in a litre of alcohol could ionise 

 the whole of the solvent, together with some hundred or so more ccm. 

 of alcohol which are not there, and that without perceptibly altering 

 the other physical properties of the liquid — such as density, viscosity^ 

 specific heat, and refractive index. After thus considering every known 

 possibility of this particular case, I felt myself compelled to admit that 

 the solution pressure of silver depends on the nature of the solvent in 

 contact with it. I could see no other alternative. It seemed, there- 

 fore, reasonable to conclude that the more obvious of the various 

 possible inferences from Kahlenberg's experiments — the inference 

 which he drew himself— is the right one, viz., that what Jones's experi- 

 ments prove for silver is true for other metals as well. 



The case is greatly strengthened by tlie recent work of Carrara 

 and d'Agostini. {q) These investigators performed experiments of the 

 same general character as Jones's, but on a much wider scale. The 

 metals they used were zinc, copper, cadmium, and silver ; the solvents, 

 water and methyl alcohol. Several salts of each metal were generally 

 used. All the electromotive forces and all the ionisation factors were 

 carefully determined. The results showed, for silver, a ratio of the 

 solution pressures of about 2,700 to 1 ; while the three divalent metals 

 agreed in giving a ratio just twice as great. The case here is at least 

 ten times as strong as that derivable from Jones's experiments — and 

 they alone are enough to settle the question. 



(/) For a proof that a very small proportion of the ions is complex see Guthe, 

 Bull. Bur. Stand., U.S.A., I. ; the proportion is far too small to affect the 

 argument in the text. 



(gf) Gaz. Chem., Ital. [35], I. See also Sucker's very careful review of this memoir, 

 Zeitsch. Elektroch., XI. 



