Cady and Arnold — Electric Arc. 405 



meter reading recorded before the electrodes touched. Little 

 reliance is to be placed on the precision of these values, on 

 account of the variable resistance of the globules themselves, 

 as stated above. Taken all together, our observations indicate 

 a value of from 21 to 23 volts as the critical potential differ- 

 ence for zero length of arc. Probably the best method would 

 be to record the drop in voltage by means of an oscillograph, 

 which would indicate the true residual drop after the globules 

 had touched and before they had time to cool perceptibly. 



Observations of the critical voltage for copper by the same 

 method gave about 22 volts at 4 amp., in fair agreement with 

 v. Lang's 23*9. Child (1. c.) found as the sum of the positive and 

 negative drops for copper, using a carbon exploring electrode, 

 at 6 amp., 25 volts. 



From the observations represented in fig. 5, together with 

 other data -covering a wider range, we have constructed a 

 system of curves connecting voltage with length of arc, for 

 the arc between iron terminals. These make it evident that 

 the linear law represented by equation (1) does not hold, 

 especially for small lengths. The curves are convex upward, 

 suggesting that an equation of the form 



E= a + M + c V I. 



in which a, b, and c may be functions of the current, would be 

 nearer the truth. We prefer not to assign numerical values 

 to these constants until more complete data have been obtained. 

 For the iron arc in nitrogen, the E-l curves have the same 

 general form as in air. For the same current and length of 

 arc, the potential difference is about twelve volts less than in 

 air, as was found by Arons.* The critical voltage (zero length) 

 was perhaps two or three volts less than in air. 



IV. Electrical Oscillations in the Metallic Arc. 



§33. Rotations in the Iron Arc. — In the first of the prelim- 

 inary papers already referred to, mention was made of an effect 

 noted with the iron arc in free air. After the current has been 

 increased on the second stage to a little below two amperes, 

 the positive end of the arc begins to rotate, so that on the 

 anode a ring instead of a spot of light appears. This is accom- 

 panied by a high-pitched whistling sound, which as the current 

 is still further increased degenerates into a sputter, and this in 

 turn into a steady, strong hiss, the ring meanwhile having 

 disappeared. The arc is in great agitation, swaying about 

 from side to side, and sometimes separating into several arcs. 

 At the beginning of whistling, there is a curious tendency for 

 *Ann. Phys., i, 706, 1900. 



