DILUTE SOLUTIONS AT Till! KKKKXINc POINT 



causing !i rise in the apparent equivalent conductivity till the quantity of salt added 

 was so great that the error became unimportant. 



The apparatus was eventually moved to a room in which no gas flames were 

 allowed, and the cell more carefully shielded from the atmosphere. Much greater 

 constancy in the resistance of the solvent was thus obtained, though even after the 

 removal a slight deterioration in quality sometimes appeared. As an example of a 

 rather bad case the following numbers may be given. They show the resistance of 

 the water used as solvent for a series of solutions of potassium chloride. 



December 16, 1898. KC1 at 0. Resistance of water. 



Time. Temperature. Resistance. 



10 h 0"' -0-10 34400 



10 12 -0-02 34260 



10 26 + 0-01 34140 



11 26 +0-11 33890 

 11 29 +0-12 33850 

 11 42 +0-13 33810 

 11 46 +0-13 33790 



The resistance of the first solution was measured at 1 1 '50, at which time the slight 

 falling off still noticeable in the water would have brought its resistance to about 

 33770 at + 0'13. The freezing point of water on the thermometer as used was 

 + 0'08, and this resistance corresponds to 33818 at that temperature. The 

 reciprocal of this number, viz., 2'957 X 10~ 5 , may be taken as a measure (in arbitrary 

 cell units) of the conductivity of the solvent ; it has to be subtracted from the 

 corresponding value for the solutions. 



The constancy was usually greater than this, and it was found that water could be 

 added from the filling machine without changing the measured resistance, except by 

 the small amount due to the increase in volume of the liquid and the slight rise in 

 temperature caused by the addition of warmer water. This is shown by the following 

 figures : 



July 30. Resistance of Water. 



Time. Temperature. Resistance. 



12.32 -09 34420 



12.35 -06 34340 



12.39 '05 34310 



12.44 '05 34260 



12.47 -06 34190 



