Theory of the Flicker Photometer. 121 



the data being taken from fig. 9. Relative positions along 

 A — B represent various relative intensities of illumination 

 from A and B on the photometer; the lines a, /3 represent 

 the proportions of A and B in the mixture alternated in the 

 flicker photometer. Speeds are indicated by the ordinates. 

 Below a critical speed the alternated hues for any speed 

 become more different than the least noticeable hue dif- 

 ference *. The two heavy lines in fig. 9 have been drawn 

 to select arbitrarily a certain mixture difference. The points 

 in fig. 10 corresponding to this difference form the dashed 

 line limiting " colour flicker." As either colour prepon- 

 derates the hue difference for any speed becomes less, so 

 that a lower speed is possible before colour flicker appears. 

 A setting either side of the centre at the speed of colour 

 fusion merely causes brightness flicker. 



Colour flicker, caused by the alternation of noticeably 

 different hues, may perhaps be identical with " coarse flicker," 

 brightness flicker with "fine flicker/' — terms to be found in 

 flicker photometer literature. 



7. Summary and Conclusion. 



In this paper we have shown how the behaviour of the 

 flicker photometer is directly deducible from the rectilinear 

 critical-frequency-illumination relationship. It is assumed 

 that a fluctuating stimulus is transmitted as a considerably 

 damped fluctuating impression whose form and amplitude 

 are calculable by using the Fourier linear diffusion equation. 

 Values of the diftusivity factor are obtained from the critical 

 frequency data on the assumption that the Weber-Fechner 

 law determines the disappearance of flicker. These values 

 are then used to determine the amplitude of the fluctuations 

 of the transmitted impression. On the assumption that the 

 instantaneous values of brightness of two alternated colours 

 are additive, the condition for no resultant flicker has been 

 found. Calculated values show all the characteristics of the 

 experimentally determined flicker photometer data, such as 

 the reversed Purkinje effect, the approximation to the critical 

 frequency criterion at low intensities, to equal brightness at 

 high. Application of the same line of reasoning explains 

 the relationship of colour flicker to brightness flicker. 



* The just noticeable hue difference corresponds to different wave- 

 length intervals or mixture proportions in different parts of the spectrum, 

 -so that the constant mixture differences selected in the diagram would 

 not correspond to exactly the same hue differences for different dominant 

 hues in the mixed colours. 



