IONIC VELOCITIES. 385 



to compare the values of this dielectric constant for various 

 liquids with the conductivities of salts dissolved in them. 

 Since for our purpose a comparison with a definite constant 

 of the pure solvent is needed, the ratio of the conductivities 

 of very dilute solutions should be used. The conductivities 

 of very dilute solutions of calcium chloride in methyl and 

 ethyl alcohols have lately been determined by Fitzpatrick 

 and the writer of this article. There are not for these 

 cases such definite limiting values as for water solutions, 

 but the following relative numbers for the molecular conduc- 

 tivities (that of the infinitely dilute aqueous solution being 

 taken as 100) correspond to the greatest dilutions reached : — 



Water 100. Methyl Alcohol 70. Ethyl Alcohol 23. 

 Although the specific inductive capacities of water and 

 ethyl alcohol have been measured by several observers, the 

 only value for methyl alcohol is that given by Tereschin, so 

 for the sake of comparison his results for all three are used. 

 He gives water = 837 ; methyl alcohol = 32-65 ; ethyl 

 alcohol = 25*8. These numbers must be corrected for the 

 viscosities of the liquids (the relative values of which are 

 100 : 63 : 120). If we assume that the ionic velocities, and 

 therefore the conductivities, are inversely proportional to the 

 viscosities we get the following values : — 



Water 100. Methyl Alcohol 63. Ethyl Alcohol 26/ 

 The approximation of these numbers to those given above 

 for the conductivities, suggest that these properties are at all 

 events the chief factors in determining the "relative ionisa- 

 tion power" of a solvent. 



If the table of velocities on p. $jj be examined, it will 

 be seen that the values for the elements in a chemical group, 

 such as that of the alkali metals, increase with the atomic 

 weight. Thus we get Lz' = 36, Na = 45, K = 66. Studies 

 of these relations have been made by Ostwald, Walden and 

 others, and G. Bredig has collected all known results and 

 added others from his own observations in an elaborate 

 paper in the Zeitschrift fur Physikalische Chemie, 1 894, 

 xiii., p. 243. The ionic velocity, like so many other physical 

 and chemical properties, appears to be a periodic function 

 of the atomic weight of the elements, and if plotted on a 



