218 CONDUCTIVITY AND VISCOSITY IN MIXED SOLVENTS. 



of water. (2) A much greater velocity of the ions in acetone than in water. 

 This has been demonstrated to be correct, since the fluidity of acetone is 

 much greater than that of water, and especially because the solvates formed 

 in water are much more complex than those in acetone, as is shown by the 

 great difference in the temperature coefficients of conductivity. These are 

 about ten times as great in aqueous solutions as they are in acetone solutions. 



(10) A negative coefficient of viscosity has been found for potassium sulpho- 

 cyanate in aqueous solution, and we have offered the following explanation 

 to account for it, after having called attention to the fact that all previously 

 offered explanations are inadequate. 



From the experimental data recorded by Wagner we find that caesium, 

 rubidium, and potassium are practically the only ions that produce the nega- 

 tive viscosity in aqueous solution. We have further pointed out that all 

 potassium salts do not give a negative viscosity, and that this is due to the 

 fact that viscosity is an additive function of both the ions involved ; and that 

 in some cases the action of the anion, tending to produce a positive viscosity, 

 is sufficient to overcome the action of the potassium ion which tends to produce 

 a negative viscosity in aqueous solution. We have pointed out the fact that 

 the total frictional surface in a liquid would be diminished by the introduction 

 into that liquid of a substance which gives ions with atomic volumes much 

 larger than the molecular volumes of the molecules of the liquid itself; and 

 we find that such is probably the case when caesium, rubidium, and potassium 

 salts are dissolved in water. 



We have tested our suggestion by reference to the atomic volume curve of 

 the elements as drawn by Lothar Meyer, which, as is well known, shows that 

 potassium, rubidium, and caesium have by far the largest atomic volumes 

 of any of the elements. We have still further tested our hypothesis by 

 showing that the amount of negative viscosity produced by the chlorides of 

 these three elements is in the same order as their relative atomic volumes. 



