2 Mr. W. C. Dampier Whetham on the 



already been carried out, and an attempt to connect such 

 results quantitatively with the dielectric constants has been 

 made by the present writer*. Since then further experi- 

 ments have been made by Vollmerf on solutions in methyl 

 and ethyl alcohols, and he finds that the ratios of the conduc- 

 tivities of various salts in water, methyl alcohol, and ethyl 

 alcohol are 100 : 73 : 34. If we assume that the frictional 

 resistances which the three solvents oppose to the passage of 

 the ions through them are in the same ratios as their vis- 

 cosities (a more or less plausible assumption for which there 

 is no experimental evidence), we can find relative values for 

 the ionizing powers by taking them to vary as the dielectric 

 constants and as the ionic fluidities. The numbers for the 

 three liquids come out as 100 : 63 : 26, a rough agreement 

 with observation, which goes to show that we are probably 

 working somewhat in the right direction. 



The experiments now to be described were undertaken in 

 order to investigate the same idea by another method. It 

 seemed possible, if a liquid could be found of a suitable nature 

 and used as solvent, that water, used as solute, should itself 

 be ionized, and make a solution of high conductivity. 



A search for data on dielectric constants did not, however, 

 discover any liquid of higher constant than water. Never- 

 theless, it seemed likely that an examination of the conduc- 

 tivities of mixtures of water and other liquids having various 

 constants would throw some light on the question at issue. 



In the choice of liquids for this purpose three things had 

 to be borne in mind. Firstly, the substances must be liquid 

 at convenient temperatures and pressures ; secondly, they 

 must, when pure, have high specific resistances ; and thirdly, 

 their dielectric constants must vary as much as possible. 

 The three acids — acetic, trichloracetic, and formic — were 

 finally selected. Many dielectric constants were determined 

 by Thw,ing|, whose numbers for water, formic acid, and 

 acetic acid are 75*5, 62'0, and 10' 3 respectively. The con- 

 stant for trichloracetic acid is not given, but from the numbers 

 for aldehyde (18"55) and chloral (5*47) it appears that the 

 replacement of three hydrogen by three chlorine atoms lowers 

 the dielectric constant considerably. The value for trichlor- 

 acetic acid is, therefore, probably less than 10. Formic acid 

 appears to have a higher dielectric capacity than any liquid 

 other than water which has yet been examined. 



The work was done in the summer of 1896, but its publica- 

 tion was delayed because it seemed desirable that the 



* Phil. Mag. 1894, xxxviii. p. 392. 



t Wied. Ann. 1894, lii. p. 328. 



X Zeits, physikal. Chem. 1894, xiv. p. 286. 



