WORK OF E. G. MAHIN. 119 



Burettes and graduated flasks were calibrated by direct weighing of distilled water, 



on the basis of the true liter, and corrected to use at 20. Pipettes were not used 



in any part of the work. 



SOLVENTS. 



Water was purified by the method of Jones and Mackay, 1 ordinary distilled water 

 being twice redistilled from chromic acid and once from barium hydroxide. The 

 stills and bottles were protected from the entrance of carbon dioxide by a tube of soda- 

 lime. The average specific conductivity of the purified water was 1.2 X10 -6 at 0. 



Methyl alcohol was prepared by boiling over lime and subsequent distillation 

 into a bottle containing fresh lime, where it was allowed to stand for several weeks. 

 It was redistilled immediately before using, and its average specific conductivity 

 was 1 X 10" 6 at 0. 



Ethyl alcohol was prepared in the same manner as methyl alcohol. The average 

 specific conductivity at was 5X10 -7 . 



Acetone was allowed to stand for a month or more over fused calcium chloride, 

 and was distilled immediately before using. Our experience with acetone was quite 

 different from that of Dutoit, 2 in that we found that either acetone or solutions in 

 acetone increased in conductivity very decidedly when exposed to sunlight or even 

 fairly bright diffused light. After distillation, therefore, the acetone was kept in the 

 dark except when being used. It is believed that any error due to impurities in the 

 acetone has been small in most cases, since acetone was obtained having a specific 

 conductivity as low as 0.57 X10 -7 reciprocal Siemens units at 0, and 0.75 X10 -7 

 at 25. Dutoit and Levier stated that Benz 3 obtained acetone having a specific 

 conductivity of 0.23 X10~ 7 reciprocal Siemens units at 18, this being the lowest 

 of which we have found any record. 



SOLUTIONS. 



In all cases conductivity measurements were made with solutions from solvents 

 which were distilled on the same or the preceding day. Solvents and solutions 

 were kept in bottles of Jena glass. In mixing solvents and in making solutions, they 

 were placed in a 20 bath for some time before diluting to the mark on the flask. 

 In designating mixed solvents percentage by volume is understood. 



An examination of the published viscosity data for mixed solvents will show that 

 there are slight differences between the values given by different investigators, 

 although those given by any one person are generally consistent among themselves. 

 It is probable that such discrepancies are caused chiefly by differences in the amount 

 and kind of unavoidable impurities in the solvents. For this reason we have deter- 

 mined the viscosity of all of our solvents, and the data given in table 85 are those 

 which have been obtained in this investigation. Figures 44 and 45 are drawn from 

 these data. 



In table 85 the symbols have the following significance : r? is the coefficient of 

 viscosity at 0, t? 25 that at 25, $ the fluidity at 0, and <i> 25 that at 25. The 

 temperature coefficient of fluidity is the change in fluidity per degree, divided by 

 the fluidity at 0. 



'Zeit. phys. Chem.,22, 237 (1897). "Loc. cit. 'Dissertation, Lausanne, 1905. 



