Electromotive Forces of the Electric Arc. 



513 



The essential stipulation is made that whatever means be used to 

 measure the resistance and E.M.F.'s of the arc, the conditions of the 

 arc must not be in any way changed by the test. It is considered that 

 the main phenomena of the arc depend on the exact thermal conditions 

 of its different parts, and on the distribution of the heated gaseous 

 and other particles, so that it is necessary to maintain these constant 

 during the test. This leads to the condition that not only must the 

 testing current used be very small, but also that the test must be 

 completed in an exceedingly short time after applying the same. 



As illustrating how very short a time may be allowed to elapse, it 

 was found that an appreciable change in the thermal conditions of an 

 arc had taken place in 1/10,000 second after changing the arc current 

 by as little as 3 per cent. 



Historical. 



A brief historical resume' is given showing that previous experi- 

 menters have not succeeded in measuring the true resistance and back 

 E.M.F. of the arc, due to their not having realised the importance of 

 completing the test before the conditions of the arc have had time to 

 be altered by the testing current. 



Those methods, similar to the Kohlrausch method of measuring the 

 resistance of an electrolyte, in which'an alternating testing current is 

 superposed on the direct current, such as that employed by Messrs. 

 Frith and Eodgers, who found that what they measured as the resist- 

 ance of the arc had in some cases a negative value, are shown to have 

 failed owing to the frequency of the alternating testing current not 

 being high enough. This frequency should be, instead of a few 

 hundred periods per second, as used by previous observers, many 

 thousand periods per second, in order that the conditions of the arc 

 may not vary, and the true resistance may be obtained. 



Preliminary Experiments. 



In the preliminary experiments the oscillatory discharge of a con- 

 denser was superposed on the main direct current through the arc, and 

 was used as the testing current, the wave-forms of the superposed 

 oscillatory P.D. and current being recorded by means of an oscillo- 

 graph. If the arc behaved as a non-inductive resistance, the waves of 

 P.D. and current should be similar curves, and in phase. This is found 

 not to be the case with frequencies up to 5000 periods per second. 

 The author concludes from these experiments that, each increase made in 

 the frequency of the superposed alternating testing current has led to 

 the arc conditions being less affected by it, and, in consequence, to the 

 arc behaving more and more like an ordinary non-inductive resistance, 

 and therefore that much higher frequencies are required to obtain the 



2 o 2 



