326 ON THE CALCULATION OF THE CONDUCTIVITY OF 
not quite to that accuracy. In using them for making analyses 
volumes of 30 ¢.c¢. or over were delivered. 
Analyses. 
The strength of NaCl solutions was found by volumetric 
determination of the chlorine according to Mobr’s method. 
Many of them also, particularly the stronger ones, were made 
up by weighing dry NaCl, which had been kept in a desiccator, 
in calibrated flasks, and filling up with water. 
The BaCl, solutions were analyzed by precipitating the 
barium with Na,SO,. The barium sulphate was collected on 
filters and its amount determined in the ordinary way. In 
many cases also the amount of chlorine in the filtrate was 
determined volumetrically. 
The volumetric analyses were not so reliable as the gravi- 
metric, the error of the former sometimes reaching five-tenths 
of one per cent, though generally in the direct analyses of the 
chlorine, one or two tenths. 
Conductivity Measurements. 
The conductivity of the solutions whose strength had been 
determined in these various ways was measured by Kohlrausch’s 
telephone method. The bridge wire, made of German silver, 
was divided into thousandths, which again admitted of easy 
subdivision by the eye into tenths. Of a set of four platinum 
resistances in the instrument, viz., 1000 ohms, 100, 10, and 1, 
the first two only were used. These were certified by Queen 
& Co., of Philadelphia, to be correct to one-tiftieth of one per 
cent. A number of solutions were compared with both of these 
resistances, and the difference between the conductivities thus 
measured lay within the limits of error. 
To contain the solutions during the measurement of condue- 
tivity, two cells of different type were used. One was in the 
shape of a U-tube, the middle part being about 2 inch in diameter 
and 5 inches long, while the two arms had each a diameter of 17 
