206 SCIENCE PROGRESS 



and by Nernst (Zeit.phys. Chem. 1893, ii. 220), that the coefficient 



of ionisation depends mainly on the specific inductive capacity 



of the solvent, has much to commend it, both on the theoretical 



and on the experimental side. As a matter of theory, it was 



pointed out that the ordinary law governing the attraction 



between charged bodies 



F -?' ? 2 

 er 2 



shows that the force between them varies inversely as the 



specific inductive capacity e of the medium ; the attractive force 



between a pair of ions immersed in water as a dielectric would 



therefore be only one-eightieth of that in air or one-fortieth of 



that in a hydrocarbon medium such as benzene or petroleum. 



On the experimental side it is to be noted that the small group 



of substances which possess in a high degree the power of 



producing ionisation are for the most part substances of high 



specific inductive capacity, e.g. : 



Water specific inductive capacity, 80 



Hydrogen cyanide ... „ „ „ 96 



Ammonia .... „ „ ,,22 



and that in homologous series the specific inductive capacity 



and the ionising power of the solvent decrease together as the 



molecular weight increases, e.g. : 



Water . . e = 80 



Methyl alcohol . 35 



Ethyl alcohol . 26 



Ethyl ether . . 4 



Methylamine . 10 

 Ethylamine . . 6 

 Ammonia . . 22 



Hydrogen cyanide . 96 



Methyl cyanide . 36 



Ethyl cyanide . . 26 



Propyl cyanide . 20 



Attention has frequently been called to the existence of 

 exceptions to the Nernst-Thomson rule, but in many of the 

 cases that have been quoted the comparison has been invalidated 

 by the use of the molecular conductivity of a solution in place 

 of its coefficient of ionisation as a measure of the ionising power 

 of the solvent. Thus it has frequently been noticed that 

 solutions in liquid ammonia at — 37 have a higher electrical 

 conductivity than the corresponding solutions in water at + 18 , 

 although the specific inductive capacity is much higher in the 

 latter solvent (Goodwin and Thompson, Physical Review, 1899, 

 8, 43) ; but the more recent experiments of Dutoit {Zeit. 

 Elektrochemie, 1906, 12, 642) have shown that the coefficients of 

 ionisation are relatively low, and that the high conductivity 

 of the ammoniacal solutions is due to the exceptionally great 

 mobility of the ions in this solvent, and not to any excess of 



