260 J. SAKÜRAI; 



Since the index, d, at once gives the ratio, —j, the resistance of the vessel 



containing the solution is obtained by simply multiplying the resist- 

 ance of the rheostat by this ratio. In all the determinations I so 

 arranged the resistances that the index should always lie between 0.9 

 and 1.0 divisions on the scale when the telephone was silent ; in this 

 manner, any inaccuracy arising from possible errors of calibration was 

 reduced to the minimum. This part of the scale is subdivided into 

 ten parts, and it is easy to read to one-half of these divisions. 

 Instead of attempting to catch the sound-minimum of the telephone, 

 I determined the limits of the region of silence and took the arithme- 

 tical mean of these limits as the true sound-minimum, as for example, 

 0.955-0.975 = 0.965. For each dilution I made two sets of deter- 

 minations by slightly altering the resistance of the rheostat, each set 

 consisting of two separate readings, and took the mean of these four 

 readings. 



The resistance vessel for holding the solution, which I employed, 

 is of the Arrhenius .form and was made by following the directions 

 given by Ostwald. (Handbuch für pliysilco-chemischc Messungen). 

 Burettes and pipettes were also accurately calibrated according to the 

 methods described in the same valuable work. 



The temperature, at which the determinations were made, was 

 25.°00 — 25.°05 throughout, this constant temperature being maintained 

 in a water thermostat worked by a small water-turbine. 



Resistance capacity of the vessel. For determining the resistance 

 capacity of the vessel, I employed -^r- A r solution of potassium chloride, 

 whose specific conductivity is accurately known to be /=2.594x 10" 3 , 

 at 25°. If /= measured conductivity of the vessel containing -=^r N 

 solution of potassium chloride and c=its resistance capacity, then — 



