374 Mr. H. E. Ives on a 



o> M being the critical speed for the equality point *. 

 Calculation of Sh therefore presupposes a knowledge of S B > 

 and any error or variation in Sb will be reflected in the 

 value found for Sh t- It must not be overlooked that while 

 the brightness difference discrimination fraction has only 

 one value, the hue fraction is a function of the colour 

 difference constituting the ends of the scale ; the larger 

 the colour difference the smaller will this fraction be. 



Experimental values of Sh fc>r the red-green comparison 

 used in this test are given in the next section. 



8. Relative Values of Brightness and Hue Discrimination 

 Fractions under Steady and Fluctuating Conditions. 



The question of crucial importance in the flicker photo- 

 meter is not so much the actual values of the discriminating 

 fractions, but whether their ratio is different for fluctuating 

 and steady conditions. If this ratio is the same, then the 

 mixture by rapid alternation possesses no advantage over 

 any other form of mixture photometer, in which each field 

 is diluted with light from the other until the colour difference 

 is no longer troublesome. 



In order to examine these ratios for both steady and 

 fluctuating conditions, it was necessary to arrange for the 

 measurement of both discrimination fractions with juxta- 

 posed fields under steady conditions. This was simply 

 arranged by viewing the mixture photometer glasses 

 through a good bi-prism, size of field and brightness of 

 retinal image being made the same as for the flicker photo- 

 meter work. The procedure in obtaining the brightness 

 discrimination fraction was to use one light only, for which 

 the green was taken, and turn the two opal glasses relative 

 to each other until a difference of brightness was just notice- 

 able between the two halves of the bi-prism, when the angle 



* Equation (11) can be used to calculate the critical speeds for all 

 hue differences lying between any two for which a» M has been deter- 

 mined, by noting that halving the colour difference between the 

 compared fields is equivalent to doubling the value of 8 B , &c. Critical 

 speed hue difference curves when calculated in this way have the 

 general characteristics of the experimental one just published by 

 Troland (" Apparent Brightness," Illuminating Engineering Society 

 Convention, 1916). The critical speed for elimination of colour flicker 

 between equally bright fields is a measure of their hue difference which 

 may prove easier to determine than the present method of stepping off 

 just noticeable differences. It should be possible to connect these 

 critical speeds with the colour triangle. 



f £ H could also be obtained directly by a similar method to that used 

 for ti B by finding the critical speeds for various intermediate hue 

 differences. 



