llieory of the Flicker Photometer. 723 



before brightness fusion/'' But the exact condition of 

 affairs has not been clearly indicated. Reasoning along 

 the lines adopted in the preceding section of this paper, a 

 clear conception of the action of the flicker photometer is 

 obtained. 



Let us assume, as before, that the brightness impression 

 derived from a fluctuating light source is, owing to its 

 manner of transmission, of much less amplitude than is the 

 stimulus. Two dovetailed stimuli are then reported to the 

 final receiver as shown in fig. 3. 



The amplitude of the fluctuations may be calculated in the 

 manner shown in the earlier portion of the paper. We have 

 thus far concerned ourselves merely with the brightness 

 factor of the two impressions and their resultant. So far as 

 our reasoning up to the present is concerned, two colours, 

 when the condition of brightness match is obtained (on the 

 flicker photometer criterion), could be alternated at any 

 speed whatever. But actually there is a lower limit to the 

 speed which may be used. This lower limit is determined 

 by the occurrence of " colour flicker," even if the photometer 

 remains in the position of brightness match. 



The explanation lies in this fact : — When two equally 

 bright colours are alternated at a given speed, resulting in 

 the condition shown in fig. 3, there exists at all times a 

 mixture of the two colours. But this mixture varies from 

 the maximum to the minimum of each colour. The two 

 extremes are represented in the drawing by a. and /3. At a. 

 the mixture has a larger proportion of colour A than of 

 colour B ; at ft the reverse is true. Hence, in the equal- 

 brightness condition there exists a rapid alternation of hue. 

 Now the higher the speed the closer the mixtures a and /3 

 approach each other in hue. Ultimately a point is reached 

 where the difference in hue becomes less than the just 

 noticeable hue difference *. This point is the lower limit of 

 speed. If the speed is raised above this point there is no 

 sensation of flicker at the point of photometric balance. If 

 it is raised much beyond this point loss of sensibility results, 

 while if the speed is made high enough a setting may be 

 made with any ratio of brightness. It follows at once also 

 that the greater the difference in colour between the two 



* The just noticeable hue difference in question is that between succes- 

 sively rapidly presented impressions. It is not necessarily or even pro- 

 bably as small a difference as can be perceived by simultaneous comparison 

 of adjacent coloured fields. The present general line of reasoning applies, 

 if the latter just noticeable hue difference is introduced, to a " mixture 

 photometer," toward which the Bunsen grease spot is one step. 



