Theory of the Flicker Photometer. 



709 



is best done by reference to figs. 1 and 2, of which the first 

 is copied on a large scale from the diagrammatic represen- 

 tation in the Philosophical Magazine, p. 358, Sept. 1912 ; 



Fw. 1. 



Logarithms of Brightness 



Diagram in atic representation of experimentally found relationship 

 between critical frequency and brightness for various colours. 



a and b, relative intensities of green and blue, which have the same 

 critical frequency^ of disappearance of flicker. 



a and c, relative critical speeds for the same intensity of green and 

 blue. 



while the second is the figure appearing on page 301 of the 

 same issue. Fig. 1 shows the definite rectilinear relation 

 which exists between the logarithm of the brightness (log I) 

 and the critical frequency of disappearance of flicker. It 

 shows the different slope of the straight lines for different 

 colours, and the abrupt change of slope at low intensities. 

 In accordance with the first fact we have the reversed 

 Purkinje phenomenon at high and medium illuminations. 

 In consequence of the second we have a true Purkinje effect 

 at very low illuminations. 



A reversed Purkinje effect had previously been found with 

 the flicker photometer. The question was therefore raised 

 as to whether the flicker photometer could be considered as 

 a means of dovetailing two pure flickers. One assumption 



