TOLMAN. — ELECTROMOTIVE FORCE PRODUCED IN SOLUTIONS. 127 



of water. In the case of hydrogen iodide, owing to the oxidation of 

 the acid, there was a gradual increase in the amount of I2 present in the 

 sohition. A rough colorimetric analysis of the solution as finally used 

 showed the presence of about 2/100 mol U per kilogram of water. 



The glass tubes which contained the rotating solutions have already 

 been described. In the experiments as finally performed only one tube 

 was used, an approximate counterbalance being placed in the other 

 arm of the rotator. By connecting two tubes in series, it would have 

 been possible to double the electromotive force to be measured. This 

 would not have greatly increased the certainty of the measurements, 

 however, since, on stopping the rotator, the residual potential differ- 

 ences between the two electrodes were always found to be in the same 

 direction, and by connecting two tubes in series the size of the error 

 as well as that of the potential difference to be measured would have 

 been increased. 



Before filling the tubes, they were carefully rinsed with some of the 

 solution to be used. The electrodes were heated to incandescence in a 

 blast lamp and placed in the solution without being touched by the fin- 

 gers. ^^ By using care, it is possible in this way to reduce the original 

 electromotive force between the two electrodes to the neighborhood of 

 0.2 millivolt or less. The small variable electromotive forces which do 

 persist are probably partly due to differences in temperature between 

 the two ends of the solution. As already described the tubes were 

 sealed with purified ceresin. This was almost the only material found 

 whose presence near one of the electrodes did not produce a large 

 electromotive force. 



After the apparatus had been set up ready for rotation it was tested 

 for insulation. This is very important, since any leakage between the 

 leads coming from the solution would apparently have decreased the 

 size of the electromotive force produced. The test was carried out by 

 disconnecting one of the leads from its binding post on the hub of the 

 rotator and applying a drop of 1^ volts at the other end of the leads 

 where they joined the measuring system. A galvanometer was in series 

 with the potential drop to measure the current leaking from one lead 

 to the other. ^3 



In general, for the final experiments the galvanometer was absolutely 

 stationary, and if there was more than a trace of a deflection, the 



" Laurie, Zeits. f. phys. Chem. 64, 617 (1908). 



^3 In order to make the test more thorough the lead which had been dis- 

 connected from the binding post was connected to the steel rotator. Until 

 considerable experience had been gained, it was very difficult to eliminate 

 leaks between the wires coming from the electrodes and the rotator. 



