Mixtures of Liquids and of Solutions. 



Wijkander's Experiments (continued). 



Benzine and Alcohol, at 25° C. 



is; 



P. cent. Alcohol. 



Viscosity. 



Weight. 



Vol. 



Obs. 



•00641 



832 

 1275 



Calculated. 



Vis. 

 forml. 



p. cent, 

 error. 



Log. 

 forml. 



p. cent, 

 error. 



Mob. 

 forml. 



p. cent, 

 error. 







50 



100 



53-0 



•00977 



4-17 



•00923 



! 



+ 11 



•00870 



-5 



An examination of these tables shows that no one of the 

 three formulae represents correctly the variation of the vis- 

 cosity with constitution in every case. The most unsatis- 

 factory of the three is that which makes the viscosity a 

 linear function of the constitution, and of the other two that 

 which makes the logarithm of the viscosity follow the linear 

 law seems somewhat less unsatisfactory than the one which 

 makes the mobility follow that law. In general the observed 

 viscosity is less than that given by the logarithmic and 

 greater than that given by the mobility law, i. e. the con- 

 stitution shown in fig. 3 or 4 gives too great, and that in 

 fig. 1 too small a value for the viscosity. If the observed values 

 for different mixtures divided the interval between these two 

 calculated values in some fairly constant ratio, it would have 

 justified the conclusion that the viscosity of a mixture was a 

 definite function of the amounts and viscosities of the con- 

 stituents present, but it will be seen from the tables that the 

 observed values are to be found in a different part of the 

 interval for each mixture, and for some even outside it. 

 Hence, even in mixtures of liquids which are generally 

 understood to have no chemical action on each other, i. e. to 

 be physical mixtures purely, there is some interaction of the 

 liquids which prevents the viscosity of the mixture being a 

 function of the amounts and viscosities of its constituents 

 only. 



In those cases where the observed viscosity is less than 

 that calculated for fig. 3 or 4, a reduction of the calculated 

 value might be made by assuming that there is a finite slip 

 at each surface of contact of the two constituents. The 



