406 Cacly and Arnold — Electric Arc. 



the arc to spring back on to the first stage, so that for an instant 

 hissing ceases, current falls abruptly, and e.tn.f. rises several 

 volts. 



If one begins to increase the external resistance im mediately 

 after one of these changes, the first stage can sometimes be 

 maintained steadily for a short time, even though the current 

 is far greater than that at which the critical point normally 

 occurs. The reason evidently is that the anode becomes so 

 far cooled that evaporation ceases. A maximum of the charac- 

 teristic curve is often found in the neighborhood of this stage 

 (cf. fig. 5). 



The effect mentioned reminds one of Trotter's observation 

 of rotations in the carbon arc at the beginning of hissing, but 

 the cause can hardly be the same. It is more likely that it is 

 a case of electromagnetic rotations. Opportunity has not yet 

 been found to examine this more closely, but it is proposed to do 

 so with the aid of an oscillograph. That marked oscillations in 

 current are here taking place was shown by the fact that when 

 the arc circuit included the primary of an induction coil whose 

 secondary was connected to a quadrant electrometer, a large 

 deflection ensued. , 



§34. Tests for Continuity of Discharge. — : The following 

 experiments were made with a view to testing the assertion of 

 Lecher,* which has found general acceptance, that the arc 

 between terminals of silver, copper, or of carbon on the quiet 

 stage, is perfectly continuous, while that between terminals of 

 iron, platinum, or of carbon on the hissing stage is discontin- 

 uous or intermittent. 



The nature of the fluctuations between carbon terminals has 

 been investigated by Duddell, who found that on the hissing 

 stage the arc is not intermittent, but subject to irregular fluc- 

 tuations. 



Lecher's method was, to connect in parallel with the arc a 

 condenser and stretched brass wire in series. Changes in 

 potential difference between the arc terminals set up oscilla- 

 tions in the condenser circuit, thereby heating the wire and 

 causing it to expand. Now, recent work on the "singing" 

 metallic arc indicates the possibility of oscillations being set up 

 in such a circuit as this, in spite of its small self-inductance, 

 even though the arc of itself is not discontinuous. We decided 

 to investigate this, and at the same time to study the oscillations 

 in the neighborhood of the critical point. 



§35. To measure the intensity of the oscillations, we used a 

 bolometer similar to Fessenden's barretter, as this was easier 

 to construct and more sensitive than Lecher's hot wire. A 

 platinum wire about 4 cm long and - 024 mm in diameter, having 



*Wied. Ann., xxxiii, 609, 1888. 



