attending sudden Changes of Illumination. 373 



current could be supplied to the two electric lamps in the following- 

 manner : During half a turn of the commutator, no current to 

 either lamp ; during the succeeding one-sixth of a turn, current to 

 the interior lamp only ; during the remaining one-third of a complete 

 turn, current to the exterior lamp only. 



Starting with darkness, and turning the commutator quickly 

 through 180, the observer saw, as soon as the interior lamp was 

 lighted, the shadow of the tinfoil, which was, as usual at the initial 

 stage, of a bright red hue ; but a small fraction of a second later, 

 before it had time to lose its rodiiess and become black, the image 

 was obliterated by a flood of light from the exterior lamp, while at 

 the same moment the other lamp was extinguished. 



When the commutator was caused to make four or five turns per 

 second, the image of the tinfoil was almost continuous, and was at 

 once recognised by inexperienced observers to be red.* 



This experiment was repeated in another form, the arrangement 

 being such that the light of two lamps was interrupted by screening, 

 instead of by breaking the current ; the changes in the illumination 

 could thus be made more rapidly. 



Two black cardboard disks, from each of which a sector of 60 had 

 been cut out, were mounted 3J in. (9 cm.) apart at the ends of a 

 horizontal axle, being so fixed that the posterior edge of the opening 

 in one of the disks was exactly opposite to the anterior edge of that 

 in the other. Between the disks, and in a parallel plane, was sus- 

 pended a sheet of white paper, across the middle of which a narrow 

 strip of tinfoil was gummed. Two clear glass electric lamps were 

 placed near the outer faces of the disks at the same height as the 

 axis, the incandescent filaments being directed horizontally. To an 

 observer looking at the plain side of the paper across the edge of one 

 of the disks, while they were rotating slowly in the proper direction, 

 the paper first appeared dark all over, then it was illuminated from 

 behind by one of the lamps, the dark strip becoming visible ; finally, 

 it was illuminated from the front by the other lamp, and the strip 

 could no longer be seen. When the angular velocity was sufficiently 

 increased, the strip was seen continuously, or nearly so, and its colour 

 was, as before, bright red. 



Experiment VI. 



From a disk of white cardboard 6 in. (15 cm.) in diameter a sector 

 of 60 was cut out ; the remainder of the disk was divided into two 



* The lamps used in this experiment were made to my order. They are of 

 8-candle power and have very thin filaments, the efficiency being 2'5 watts per c.p. 

 They were worked at a pressure of 6 per cent, above their marked voltage, and the 

 incandescence responded very quickly to the current. 



