154 WORK OF J. SAM GUY.' 



that redistillation did not essentially improve this glycerol. Its specific gravity 

 showed that it contained about 0.02 per cent of water. The two lots obtained from 

 Kahlbaum showed slightly different viscosities, as was indicated by the experimental 

 results. 



Water. The water was purified by the method of Jones and Mackay, 1 with the 

 modification as mentioned by Schmidt, and had a mean specific conductivity of 

 1.5 XlO -6 at 25. 



Ethyl and Methyl Alcohols. The ethyl alcohol was purified by several distillations 

 from the very best quality of lime, and block-tin condensers were always used. It 

 had a mean conductivity of 1.8 X 10~ 7 at 25. The methyl alcohol was first distilled 

 from a small amount of dilute sulphuric acid and then several times from lime. It 

 had a mean specific conductivity of 2.0 X 10~ 6 at 25. 



SALTS. 



In all cases Kahlbaum's purest specimens were used, and these were recrystallized 

 at least three times from conductivity water, carefully dried at 125, and the solu- 

 tions made by direct weighing. 



VISCOSITY. 



The viscosity measurements were made by means of the Ostwald viscosimeter as 

 modified by Jones and Veazey, 2 and the size of the capillary so regulated as to be 

 best adapted to glycerol measurements. The method of calibration has been dis- 

 cussed in detail by Schmidt. Viscosity wascal dilated from the formula; 



V St 



Vo So to 

 in which rj is the viscosity coefficient for the liquid in question, r\o that of water, S the 

 specific gravity of the liquid, t the time of flow of the liquid, S the specific gravity of 

 water at the given temperature, and / the time of flow of the water. Fluidity was 

 calculated from the formula 



V 



where d represents the fluidity. The values of 170 are taken from the work of Thorpe 

 and Rodger, 3 being 0.00891 at 25, 0.00720 at 35, 0.00598 at 45, 0.005057 at 55, 

 0.004355 at 65, and 0.003786 at 75. 



TEMPERATURE COEFFICIENTS. 



The temperature coefficients, both in per cent and in conductivity units, have been 

 calculated, the latter being simply the actual increase in molecular conductivity per 

 degree rise in temperature, while the former were calculated from the formula; 



Temp, coeff. of p v = ^^ 



M^o 10 



The temperature coefficients of fluidity were calculated in the same way. 



Viscosity measurements were made only with the tenth-normal solutions, since at 

 higher dilutions the difference between the viscosity of the solution and that of the 

 solvent was very slight. 



lAmer. Chem. Journ., 17, 83 (1895). *Zeit. phys. Chem., 61, 641 (190S). >Loc. cit. 



