and Flicker Photometer Speed. 103 



a fraction proportional to B K (assuming that successive pre- 

 sentation hue discrimination is proportional to juxtaposed 

 presentation discrimination and proportional in the same 

 way for different colours). 



Suppose this calculation gone through for the whole 

 spectrum ; there can then be plotted a series of values of 8 K 

 against wave-length, the value of the S H units being un- 

 determined. If then the exact value of 8 H be found by the 

 method above outlined, for any one wave-length, its value 

 for other wave-lengths is at once obtainable from the plot 

 just assumed made. The value of the constant into which 

 the other terms of (1) can be combined (M) is obtained from 

 the same observations, so that all the factors necessary to 

 furnish the critical speeds have been found. 



When it comes to carrying through this indicated process 

 it is found that existing data are not adequate to an exact 

 solution, and that the work to put what we have into the 

 shape required is rather great. In order to translate the 

 distances on the ordinary equal sensation-sum triangle, 

 which are proportional to 



VC^-RO' + CG.-GO'+CBi-BO*, . . . (4) 



where R, Gr, and B are the three sensation coefficients *, to 

 distances on the equal luminosity triangle, it is necessary to 

 know the luminosity values of the three sensations. If we 

 call these L E , G G , and L B , then the distances on the equal 

 luminosity triangle are proportional to 



I have carried through the calculation of the distance of 

 white (centre of the ordinary triangle) from the various 

 parts of the spectrum, using a modified triangle obtained 

 from Koenig's colour sensation data, which has been recently 

 correlated with luminosity, only to find that for all practical 

 purposes the result is the same as though the distances had 

 been measured in the ordinary triangle, the mixture point 

 being given by the point distant from the two colours 

 inversely as their luminosities, as given by (5). This fact, 

 which is principally due to the enormous exaggeration of 



* A discussion of colour-mixing diagrams, upon which the following 

 is "based, is given by the writer in the Journal of the Franklin Institute, 

 Dec. 1915, p. 678 : " The Transformation of Colour Mixture Equations 

 from One System to Another." 



