GOODWIN AND HASKELL. — CONDUCTIVITY OF SOLUTIONS. 405 



such an agreement is actually found to exist even with relatively small 

 values of the initial concentration is evidence of the correctness of the 

 principle on which the computation is based. 



The second method of computation is based on the assumption that 

 the true values of the equivalent conductivity increase regularly to a 

 maximum at tlie highest dilution, and that the decrease in the observed 

 values from the maximum obtained at a moderate dilution is due to a con- 

 stant error in the specific conductance arising from the effect of the 

 impurities in the water. Therefore it is justifiable to add to the o1> 

 served values of the specific conductance such a quantity as will cause 



.6 .S /.O U /.4 



Concentration in milliTnols per liter. 



Figure 2. 



/<? 



the maximum in the so-corrected equivalent conductivity values to occur 

 at the lowest concentration at which the error just referred to has at- 

 tained its constant value. The method of procedure is best illustrated by 

 a reference to the accompanying diagram (Figure 2). In this the curve 

 AB represents the " observed" values of the equivalent conductivity A, 

 of the acid computed from the results of an actual series of experi- 

 ments (first series, in nitric acid, page 410) by subtracting from the 

 observed specific conductivity of the solution k, the observed conduc- 

 tivity of the water /c„, and dividing the difference by the concentration 

 c; that is, Aq^ = {k, — /<,„) / c. Now the error in the value A at A 

 must be at least as great as the ordinate AC, C being equal to the 



