346 PROCEEDINGS OF THE AMERICAN ACADEMY. 



current which had passed through the interrupters. This latter we shall 

 call, for convenience, the " intermittent " current. A voltage of approxi- 

 mately 70 was ordinarily used. 



To the naked eye the appearance of these two arcs, the alternate and 

 the intermittent, is the same except in size. But they present the follow- 

 ing striking difference to an observer on the side of the occulting disk 

 next the arc. Through the openings in the occulting disk, a lens projects 

 an image of the arc upon a white screen on the side of the disk away from 

 the arc. This screen is visible through the disk all the while. The 

 alternating arc is visible upon the occulting disk and upon the screen at 

 all times ; but the intermittent arc is seen only on the occulting disk. 

 The image appears to hang in mid-air, and although the white screen is 

 seen with perfect distinctness the intermittent arc which is projected upon 

 it bears little or no resemblance to that projected upon the occulting disk. 

 The light which reaches the screen (that is, the light which penetrates 

 the occulting disk) is that which has persisted after the current has been 

 cut off. Such an occulting disk in connection with the interrupters, then, 

 forms a most convenient means of separating radiations according to their 

 length of life after the exciting current has been cut off. 



II. The Persistent Luminosity of the Arc. 



In the case of the iron arc, we have measured the duration of the 

 luminosity after the current has been cut off and find that with a voltage 

 ranging from 70 to 80 and a current varying from 10 to 20 amperes 

 the life of the white luminous cloud that floats above the junction of the 

 electrodes covers a period of from two to five thousandths of a second 

 after the disappearance of the current. 



By the use of the sliding table top we have photographed the rise and 

 decay of the arc. 



In Figure 8 are reproduced three series of seven photographs each. 

 All of the series arfe taken under the same conditions. Each series, begin- 

 ning at the left and going to the rights, depicts the decay of the current. 

 The first of the series represents the arc as it appeared toooo ^^ ^ second 

 before the current was interrupted. The last photograph, the one at the 

 extreme right, represents the arc as it appeared lootro <^^ ^ second 

 after the current was interrupted. All the photographs were taken at 

 equal intervals of tqooo ^^ ^ second. It is evident, therefore, that the 

 third photograph was taken just as the brush of the interrupter passed 

 on to the slate sector. It will be observed that the arc proper (which is 



