GENERAL DISCUSSION OF RESULTS. 205 



The discovery of this fact enabled us to explain many things which had hitherto 

 been found in connection with the relative velocities of the ions, but which could 

 not be explained. If we compare the atomic-volume curve with the curve of the 

 migration velocities of the ions, we shall see that those ions which have the greatest 

 hydrating power have the smallest migration velocities. Sodium and lithium, whose 

 atomic volumes are less than half that of potassium, have velocities which are only 

 about two-thirds that of potassium. This for a time was not understood. Sodium 

 and lithium form salts which crystallize with two and three molecules of water; 

 salts of these elements are therefore hydrated in solution, and these hydrates around 

 the sodium and lithium ions decrease the velocities of those ions. 



A large number of lines of evidence for the above relations are discussed in the 

 preceding pages. The values of a are calculated from the molecular lowerings 

 of the freezing-point, for all concentrations less than that at which the freezing-point 

 curve passes through the minimum. In the more concentrated solutions a can not 

 be calculated from freezing-point lowerings, on account of hydration. 



It was predicted from the solvate theory of solution that dissociation as measured 

 by the freezing-point method would have higher values than when measured by elec- 

 trical conductivity. This was confirmed experimentally without a single exception. 



In the case of every salt studied the dissociation as calculated from the freezing- 

 point lowering is higher than the dissociation as calculated from conductivity. 



Since the above prediction was based upon the solvate theory of solution, its 

 verification is in keeping with that theory. 



RESULTS OBTAINED BY KREIDER, 



The conductivity method as left by Kohlrausch could not be satisfactorily used 

 to measure dissociation in any solvent other than water. The reason for this is 

 almost obvious. Take a solvent with small dissociation power. The dilution at 

 which complete dissociation would be reached in such a solvent would be so great 

 that the Kohlrausch conductivit}' method could not be applied to it. It would 

 thus be impossible to determine fx^ accurately for the substance in a slightly dis- 

 sociating solvent. The result was that the conductivity method, as a means of 

 measuring dissociation, could not be used with any reasonable degree of accuracy 

 even with a solvent with the dissociating power of ethyl alcohol. We had, up to 

 this time, no thoroughly reliable method for measuring the dissociation in such com- 

 mon and important solvents as the alcohols. The freezing-point method, obviously, 

 could not be used for this purpose, since the alcohols do not freeze at temperatures 

 that can be accurately measured. 



The boiling-point method was the only one available for the purpose under dis- 

 cussion, and this could be used only with fairly concentrated solutions. Dilute 

 solutions raised the boiling-point of alcohol so little that the change could not posr 

 sibly be measured with any degree of accuracy. This is especially true since the 

 boiling-point method is affected very seriously by barometric changes. Further, 

 the rise in the boiling-point of a solvent like the alcohols, by a dissolved substance, 

 is very slight; and, consequently, the error in measuring this small quantity is rela- 

 tively large. The hope of measuring with reasonable accuracy dissociation in non- 



