328 ON THE CALCULATION OF THE CONDUCTIVITY OF 
copper wire. In order to draw the correction curve more 
accurately, the wire was calibrated by this means in a number 
of different fractions. The possible error of a conductivity 
measurement, when the reading was made at the middle of the 
bridge wire, was shown by a number of determinations to be 
about one-tenth of one percent. For solutions of BaCl, and 
NaCl measured in the cell first described, the reading would be 
made at this part of the wire if the concentration was about 0.3 
gramme-equivalent. For 0.5 gramme-equivalent solutions, with 
the reading at or near .64 of the length of the bridge, two-tenths 
per cent, for normal solutions four-tenths. With the cell for 
dilute solutions, the possible error, wherever the reading might 
be made, was about six-tenths. Here other sources of error, 
such as change of capacity of the cell, were greater than the 
bridge error. In order to obtain a good minimum also a high 
clear note from the induction coil was necessary when a very 
dilute solution was in the cell, and this could not always be 
obtained. 
The capacity of each cell was determined by a comparison of 
the conductivities of the solutions measured in it, with the values 
given for corresponding solutions by Kohlrausch. The numbers 
given below, under the headings NaCl and BaCl, are the ratios 
of the conductivities of solutions measured in the first cell, to the 
specific conductivities of corresponding solutions, as measured 
by Kohlrausch. 
RATIO. 
Concentration. ! 
NaCl. + BaClg 
5 pABAY | UiisisierWalaerenvesen 
3) SOTO DS terets Neteciive csv 
1 2135 Dds sg 
0.5 2126 2131 
