200 TRANSACTIONS OF THE [MAB. 1, 



five per cent) it was found almost impossible to distinguish be- 

 tween flashing and flickering when two gray disks were com- 

 bined. If two disks differing considerably in color were com- 

 bined pair-wise and rotated the transition from the flashing back- 

 ground to a background of uniform hue with flicker was very 

 well marked even when the difference in luminosity was as small 

 as one or two per cent.* 



The experiments seemed to indicate that, while the phenomena 

 described under I and II could be obtained for little or no dif- 

 ference in luminosity provided the disks differed sufficiently in 

 color, for the production of flickering it was necessary that there 

 should be a diflTerence in luminosity. 



The rates of rotation given in the above description of the 

 various phenomena hold true only when the difference in lumi- 

 nosity is more than about forty per cent. If the diff'erence in 

 luminosit}' between the two disks used is less than this amount 

 the rates of rotation at which the various phenomena occur will, 

 in general, be lower than the rates above given. 



In the application of the phenomenon of flickering to photo- 

 metric work it may be of importance to know the rate of speed 

 at which the disks must be revolved to give the most accurate 

 results. The speed required for the greatest accuracy will evi- 

 dently be the speed at which the true flickering due to very 

 slight diff'erences in luminosity is the most violent, and this, in 

 general, was found to be the lowest rate at which true flickering 

 could be observed. For the determination of this speed a num- 

 ber of pairs of disks were selected so that the disks of each pair 

 diflfered in color but did not differ in luminosity by more than 

 two per cent. It was found that if the illumination was very 

 low, about sixteen candle power at a distance of one meter, the 

 lowest speed at which the flashing vanished was about fourteen 

 revolutions per second. The lowest speed at which the flicker 

 disappeared was about twent^'-five revolutions per second. The 

 best speed, therefore, for observation at this illumination 

 would be about fourteen revolutions per second. When the il- 

 lumination was about fifteen hundred candle power at a distance 

 of one meter the lowest speed at which all flashing disappeared 

 was about twenty-four revolutions per second and the lowest 

 rate required to extinguish the flickering was about thirty-four 

 revolutions per second. Therefore, under this illumination a 

 speed of about twentj'-four revolutions per second would give 

 the best result. 



* For the methods employed In measuring the luminosities of colored disks see 

 next section. 



