L28 PROCEEDINGS OF THE AMERICAN ACADEMY. 



:rouble was eliminated before making a run. A deflection of one milli- 

 aieter on the galvanometer scale would have corresponded to a resis- 

 tance of about 1.5 X 10^ ohms between the leads. 



In the experiments a number of readings of the residual electromo- 

 tive force were taken with the rotator stationary. It was then brought 

 lip to speed and a new series of readings commenced as soon as possi- 

 ble. After running for several minutes, the steam was shut off and 

 cQore readings of the residual electromotive force commenced as soon 

 IS the rotator had stopped. The voltmeter attached to the magneto 

 was placed near the potentiometer so that a determination of the speed 

 3f the apparatus was made immediately after each measurement. The 

 approximate time when each measurement was taken was also 

 recorded. 



At the end of a day's measurements, the magneto was standardized 

 with a stop watch and counter. Usually, about a half a dozen stand- 

 ardizing runs were made at speeds in the neighborhood of those used 

 in the actual measurements. Each run lasted about a minute. The stop 

 watch was started when some even figure appeared on the counter, and 

 stopped similarly about a minute later. In general, five readings of 

 the voltmeter were made, one before the stop watch was started, the 

 next three at intervals of fifteen seconds, and the last one after the 

 watch had been stopped. It was not easy to hold the speeds constant 

 enough to make standardizing runs of over a minute desirable. 



6. The Experimental Results. 



In the following tables are given the data on which the calculations 

 for each of the salts investigated are based. The first column gives 

 the time when the observation was taken, the second column, the read- ^ 



ing of the voltmeter, F (in decivolts), which indicated the speed of the * 



apparatus, and the third column, the potential difference in millivolts, 

 E X 10^ this being called negative when in the opposite direction 

 from that produced by the centrifugal force. The radii of the two 

 electrodes r^ and ri are also given, and the data for the standardiza- 

 tion of the magneto. In the case of all salts investigated, the inside ^ 

 electrode was positive with respect to the outside one during rotation. | 



It is a striking fact that, upon stopping the rotator the residual * 



electromotive force is always found to be in the opposite direction from r 



that produced by the centrifugal force. It will also be seen from an 

 examination of the data that there is a general tendency for this resid- •■ 



ual electromotive force to increase somewhat in magnitude and then | 



in the course of a few minutes gradually to disappear. The average | 



^ 1 - 



