722 Messrs. H. E. Ives and E. F. Kingsbury on the 



the Purkinje phenomenon. The two colours in question would 

 start to diverge from each other in just this way as the illumi- 

 nation dropped, due to that cause. At the higher illumina- 

 tions, as above noted, brightness is the predominant factor, 

 hence the Purkinje effect should show itself. But at lower 

 illuminations the diffusivity factor is the predominating one, 

 and the effects peculiar to flicker would overwhelm every- 

 thing else. It is to be remembered in this connexion that the 

 logarithmic scale of intensities of fig. 1 refers to physical 

 intensities, so that the Purkinje effect is not included. The 

 true explanation, however, may lie in the breakdown of the 

 critical frequency-log I relation at high illuminations, 

 evidences for which were found in the previous paper. 



The choice of the Fourier equation for study in this con- 

 nexion needs no defence. It does not involve the assumption 

 that the transmission is either thermal, electric, or chemical 

 diffusion, since the equation applies to any form of trans- 

 mission. It is the logical equation to test. The choice of a 

 simple cosine type of stimulus is dictated, of course, by the 

 necessity for avoiding too great complexity in the mathema- 

 tical work. By increasing the number of terms a series can 

 be chosen to represent any desired stimulus. It remains for 

 future work to show whether the introduction of such extra 

 terms will affect the results in any appreciable degree. 

 Practically the question is this : Does an abrupt alternation 

 of the two fields in a flicker photometer give different values 

 from a "dissolving" alternation ? kSucIi a difference has not 

 been found experimentally. 



An interesting development of this theory is the expression 

 connecting the diffusivity Kwifch the brightness. Equation (7) 

 states that the diffusivity varies as the square of a distance 

 (X) and as the logarithm of the intensity. The latter is a 

 measure of the depth of penetration of light into an absorbing 

 medium ; the former a statement of the characteristics of 

 the medium. For instance, it represents the area of a cone 

 as a function of distance from its apex. The use of the 

 Weber-Fechner law in deriving the relation between K and I 

 is of course valid only over the region where that law r holds. 

 Deviations from that law have been found both at very high 

 and at very low illuminations. 



6. Colour Flicker and Brightness Flicker. 



The mode of action of the flicker photometer has generally 

 been considered something of a mystery. The most descrip- 

 tive statement has probably been that " colour fusion occurs 



