214 CONDUCTIVITY AND VISCOSITY IN MIXED SOLVENTS. 



phate, where the S0 4 ions tend to produce a positive viscosity coefficient to such 

 a great extent as to overcome entirely the negative effect of the potassium ion ; 

 and the resultant action is the production of a positive viscosity coefficient by 

 potassium sulphate when dissolved in water. The above facts show very clearly 

 that there is some close connection between the physical properties of caesium, 

 rubidium, and potassium and the negative viscosity phenomena. In dis- 

 cussing viscosity in general, Thorpe and Rodger l state that " Viscosity is, no 

 doubt, the net result of at least two distinct causes. When a liquid flows, 

 during the actual collision or contact of the molecules, a true, friction-like 

 force will be called into play, opposing the movement. But in addition to 

 this cause, even after the actual collision, molecular attractions will exercise 

 a resistance to forces which tend to move one molecule past another." 



Our study of viscosity has led us to the same conclusion as that arrived at 

 by Thorpe and Rodger, viz, that viscosity phenomena are largely a function 

 of the frictional surfaces of the ions, molecules, and molecular aggregates in 

 any given solution. If now by any possible means the total frictional surface 

 should be decreased, then the viscosity of the solution would also be decreased. 

 Suppose that into a solvent containing molecules with comparatively small 

 molecular volume we bring a salt which yields particles having relatively 

 large atomic or ionic volumes. The larger particles of the dissolved substance 

 would be distributed among the smaller molecules of the solvent, and these 

 smaller molecules, instead of coming in contact with each other so frequently, 

 would come in contact with the larger salt particles, and thus the total fric- 

 tional surface involved would be decreased and, consequently, the viscosity 

 of the solution would also be decreased. In other words, the effect of bringing 

 those salts whose ions have large atomic volumes into pure water, would be 

 the same as if some of the molecules of the water had combined into larger 

 spheres, and thus caused a diminution in the total frictional surface. Since 

 most salts do not produce such an effect on water, we must assume that their 

 atomic volumes are too small, and that only the salts of those metals having 

 very large atomic volumes could produce a diminution in the viscosity of 

 water. The explanation proposed above for the negative viscosity phenomena 

 can then be easily tested by a glance at the atomic volume curves of the ele- 

 ments. When this test is applied, we find that csesium, rubidium, and potassium 

 stand at the very maxima of the atomic volume curve, and that none of the other 

 elements have anything like as large atomic volumes as the three elements just 

 named. Furthermore, if our hypothesis is correct, then that ion which has 

 the largest atomic volume should produce the greatest decrease in the viscosity 

 of water, and the ion having the next smaller atomic volume should produce 

 smaller decrease in the viscosity of water. Here again we find that our theory 



1 Loc. cit. 



