WORK OF E. G. MAHIN. 117 



great dilution, to break down these associated molecules into single molecules, and 

 subsequently into ions; in which case the molecular conductivity might reach a 

 value in accord with the rule that has been established in practically all other cases 

 that have been tested, namely, that molecular conductivity is inversely proportional 

 to viscosity. The curves representing molecular conductivity and fluidity, respec- 

 tively, would then assume similar forms, minima in the one corresponding to minima 

 in the other, etc. 



An effort has been made to determine the conductivities of the solutions in ques- 

 tion at high dilutions, approaching the region of complete dissociation. We can see 

 no reason why there should be any constant relation between molecular conductivities 

 and viscosity in different solvents until this point is reached, bearing in mind the 

 widely differing degrees of dissociation existing in solutions of the same concentra- 

 tion in different solvents. In very dilute solutions the difference between the vis- 

 cosity of the solution and that of the solvent is generally much less than the experi- 

 mental error, hence we have directly compared our values for the conductivity of 

 the solution with those for the viscosity of the solvent. 



The questions for which we have sought the answer are, then, the following: 



1. Will those salts that have, at ordinary concentrations, abnormally low values 

 for molecular conductivity possess, when completely dissociated, values which are 

 inversely proportional to the coefficients of viscosity? 



2. If so, is the product of molecular conductivity and viscosity constant for mixed 

 solvents and at different temperatures? 



3. Is the value of the constant the same for different electrolytes? 



4. Are the abnormal conductivities in acetone and mixtures of acetone with other 

 solvents due to association of the salt? 



EXPERIMENTAL. 



The determination of conductivity in very dilute solutions is attended with con- 

 siderable difficulty. Dutroit and Levier, 1 and Dutoit, 2 commenting upon the lack 

 of agreement between conductivity values in acetone as obtained by different inves- 

 tigators, attributed the differences chiefly to the use of platinized electrodes, to the 

 action of light, and to the use of impure solvents. Cohen 3 had already noticed that 

 electrodes which are coated with platinum black cause changes in the solutions, 

 resulting in fictitious conductivity values; and he supposed this to be due to cata- 

 lytic action of the platinum black. Dutoit and Levier showed that such electrodes 

 gave rise to errors, through adsorption of the dissolved substance; they also observed 

 that the action of light upon acetone solutions causes a decrease in the conductivity, 

 this change being reversed if the solution is subsequently placed in the dark. 



In working with very dilute solutions, it becomes necessary either to make compar- 

 atively large amounts of each solution by direct weighing of the electrolyte, or to 

 obtain the highly dilute solution by successive additions of solvent to the more con- 

 centrated solutions. The first method becomes impracticable when using solvents 

 which are relatively difficult to prepare in the pure state. The second method intro- 

 duces the error due to the large number of volume measurements which are neces- 



'Journ. Chim. Phys.,3,435 (1905). =Zeit. Elektrochem., 12, 642 (1906). % Loc. cit. 



