678 REPORT— 1897 



Chemical effects were practically excluded by working the arc in an air-tight 

 metallic hood, filled with a gas which exercised no chemical action upon the 

 electrodes. An air-tight glass window in this hood enabled the observer to 

 examine the arc either with the naked eye or with the spectroscope. 



To exclude the electric current for an instant, and to examine the arc imme- 

 diately afterwards, the following device was used : — 



A high-speed, 100-volt, alternator was employed to feed the arc. But, in 

 series with the armature and the arc, were placed two interrupters, which cut out 

 either all the positive or all the negative parts of the alternating current. In 



Fig. 1. 



either case the current was broken just as the current curve crossed the axis of X. 

 In case the positive currents were cut out, the break occurred at A, A.', A", &c., 

 the make at B, B', &c., as indicated in the figure. 



By cutting the current off just as the curve crosses the axis, self-induction 

 effects are practically avoided. 



The intervals of time indicated by the shaded portion of the current curve were 



employed to photograph the arc or to examine it with the eye. This examination 



was made through openings in a large steel disc of the form indicated in the figure. 



This occulting screen and the two large interrupter rings were 



Fig. 2. placed on a common shaft with the armature of the dynamo. 



The interrupter rings were insulated from the shaft ; each 

 had two slate sectors keyed into it, and each carried two brushes 

 set 90° apart. 



It was found that, in the case of the iron arc, in an atmos- 

 phere of air, oxygen, coal gas, or hydrogen, there are two 

 distinct luminosities having very different properties. 



Of these luminosities one is a cloud of light, strongly 

 coloured with yellow, and floating at a distance of some milli- 

 meters from the electrodes, one of which was an iron rod, the- 

 other a rotating iron disc. 



This is apparently the ' flame ' of the ordinary carbon arc. 

 This yellow cloud persists from one one-hundredth to one two-hundredth of 

 a second after the current has been broken. 



The other luminosity is the blue sheet of light which most impresses the eye on 

 looking at the ordinary iron arc. 



This light disappears in less than one five-thousandth part of a second after the 

 current is cut off. One is certain that the interval during which the blue light 

 persists is, however, still less than this. For the actual instant at which the 

 current is shut off is not the instant at which the brush passes on to the slate 

 sector, but an instant later than this, on account of the spark which remains at 

 break of current. 



So that, after the arc is broken, practically the only light that remains is this 

 yellow cloud. 



The light from the red-hot iron poles, giving a continuous spectrum, is, oi 

 course, here not considered. 



[Photographs of the two parts of the arc shown to the section.] 



On making the current the first light to appear is an intense blue right at the 

 point of contact of the two electrodes. The yellow cloud, the ' flame,' comes later. 

 We have succeeded in photographing the blue arc of one current before the yellow 

 cloud of a previous current had died out, thus obtaining the two kinds of arc on 

 one plate mth a single exposure. 



