340 Dr. Kohlrauscli. The Resistance of the Ions [Feb. 17, 



make the reduction to completely dissociated solutions very uncertain, 

 but now show only differences of irregular sign within the limits of 

 the errors of observation. 



From this we may assume that the true values of the coefficients 

 for the concentration 1/10000 would be practically the same as those 

 for complete dissociation. But, as a matter of fact, in the case of 

 such dilute solutions considerable uncertainty underlies the observations 

 themselves, on account of the variations of the solutions with time, 

 as well as the corrections for the conductivity of the water, on account 

 of its magnitude and somewhat uncertain theoretical basis. These facts 

 are especially true in the case of the acids. Therefore, the values found 

 for the concentration 1/1000, in which these two sources of error need 

 scarcely be considered, must be looked upon as experimentally much 

 more accurate. Then too, the condition of ionization of very dilute 

 solutions of weak salts is always somewhat uncertain on account of 

 possible hydrolytic action. In order to treat all electrolytes alike, 

 I shall therefore choose the coefficients a and /5 of the concentration 

 1/1000. 



The errors which may arise from the fact that these solutions lack 

 from 2 to 5 per cent, of complete dissociation cannot at present 

 be avoided. Since the values of the constants themselves differ but 

 little from those for a concentration of 1/10000, we may assume that 

 these errors are not great. The material is not yet complete enough 

 to allow the application of the corrections proposed by Mr. Whetham. 

 The values may indeed be capable of improvement, but I am of the 

 opinion that the following conclusions drawn from them are in the 

 main correct. 



In order to fill certain gaps in the series of the coefficients, I have 

 added to the electrolytes investigated by Deguisne my own observa- 

 tions on 1/1000 normal solutions of potassium and sodium fluoride, 

 sodium acetate and lithium chloride. The coefficients here are derived 

 only for the temperatures 10', 18°, and 26 3 , and will therefore be less 

 accurate than in Deguisne's results. 



(2.) Values of the Conductivity at Different Temjjeratares. — My con- 

 clusion, that by extrapolating the quadratic formula below 0°, the 

 conductivities of all electrolytes approach a zero value within a narrow 

 range of temperature, rested upon an empirically established connec- 

 tion between the coefficients a and f3. I will introduce here, however, 

 the more easily followed graphic representation (fig. 1) plotted from 

 the single coefficients. The extrapolated portions of all sixteen curves 

 converge toward the same region and pass through zero between 

 - 35° and - 41°.* 



* NH 4 C1 would nearly coincide with KCI ; 3IgS0 4 on account of the large 

 changes in its coefficient with increasing dilution is uncertain and therefore 

 omitted ; it would probably lie near Na 2 S0 4 . 



