282 MORRIS LOEB 



a(q-bv) b-aq 



gain or loss will be represented by 6 = . 



l-av I-av 



As b-aq is the value given in the column of differences, the 



factor represents the necessary correction. Conse- 



1 av 



quently, after completing the deficit of portion 4 by the 

 addition of the small deficits in 2 and 3, we have for the 

 rate of transference of the negative ion, 



16.78 , 16.82 



1.0017 and - -1.0017, or a mean value of 0.524. 



32.10 32.10 



As already stated, the determination was usually simplified 

 by dividing the liquid into two portions instead of four. 



In this experiment there was a second apparatus placed in 

 series in the circuit; hence the mean intensity of the current 

 was only 0.0012 ampere, and the electrolysis lasted seven 

 hours. We frequently saved time and used our battery to 

 better advantage by performing two simultaneous elec- 

 trolyses in parallel circuits, each of which must, of course, 

 contain its own resistance-box; by means of a properly 

 constructed switch, we could apply the measuring circuit to 

 either box at will, thus securing the measurement of both 

 electrolyses with the same galvanometer and element. We 

 may also state here that the last was frequently compared 

 with another cell of the same construction and was always 

 found to be exactly equivalent to it in electro-motive force, 

 a proof that it did not vary during the course of our 

 experiments. 



5. RATES OF TRANSFERENCE OF THE NEGATIVE ION 



Hereafter t implies the temperature during electrolysis; 

 m the molecular concentration (gram-molecules per liter); 

 /i the milligrams of silver precipitated on the cathode accord- 



