96 Potts — Rowland's New Method. 



In the measurement of electric absorption and hysteresis 

 loss, it is necessary to know the period of the current used. 

 In this investigation a speed counter and chronograph were 

 used ; the speed counter was placed at the end of the dynamo 

 shaft, and directly connected to it, a contact was so arranged 

 that for every one hundred revolutions of the dynamo arma- 

 ture a circuit was closed and a record made on a chronograph 

 sheet. On a table beside the electro-dynamometer was a key, 

 which also could be used to make a record on the chronograph 

 sheet. As soon as the bridge had been balanced, this key was 

 pressed. And by the measurement of this sheet the period of 

 the current at the time of the observations was quite accu- 

 rately determined. The error from this source was usually not 

 more than one part in 1000, never more than 1 in 100. 



After each adjustment, the resistance of each arm was deter- 

 mined by the use of a " Post-office Box " (when the current 

 through any part of the apparatus was not large, it was not 

 necessary to measure its resistance after each adjustment). The 

 resistance of that part which was affected the most by heating 

 was measured first, and in this way the actual value at the 

 time of adjustment was quite closely determined. In cases 

 where the heating was large, the error from this cause might 

 reach several parts in 1000 in R\, and consequently a con- 

 siderably greater amount in the value of the electric absorption 

 resistance. 



Errors due to induction and electrostatic action of the dif- 

 ferent portions of the apparatus were carefully guarded against 

 by the arrangement. And induction was tested for by 

 reversal of the relative directions of the currents in different 

 portions of the apparatus. Usually no effect was noticed or at 

 most it was very small. The errors introduced by the self- 

 induction of the electro-dynamometer coils and also that 

 caused by the electrostatic action of the turns of a large coil 

 on one another, were determined and corrected for when suf- 

 ficiently large in amount. 



Apparatus. 



Electro-dynamometer* — The self-induction of the fixed 

 coils was - 0165 henry, and of the hanging coil '0007 henry. 



Dynamos. — The current used in this investigation was fur- 

 nished by one of three dynamos. The Westinghouse alternator 

 in power house of the University furnished a current of period 

 •0075, i, e. 133 complete periods per sec. This was used for 

 only a few observations. In most of the work two small 

 dynamos constructed in the University workshop were 



*See this Journal, July, 1897, p. 35. 



