370 Mr. H. E. Ives on a 



No setting can be made at a lower speed than that fixed by 

 the intersection of these curves. The actual observation of 

 a flicker-limit curve such as that indicated by the small 

 diagram, with its very definite truncation of the brightness 

 flicker curves, should constitute a strong argument for the 

 correctness of the line of reasoning followed. 



For the experimental test of these relations, red and green 

 lights were taken as furnished by two point source 100 watt 

 tungsten lamps shining through red and green glasses of 

 fairly narrow spectral transmissions. The two lamps were 

 operated in parallel, one terminal of each going to the 

 battery, the other terminals being connected to the opposite 

 ends of a resistance, the sliding middle point of which was 

 joined to the other battery terminal. By moving the sliding 

 contact the intensity of one lamp increased while the other 

 decreased, and vice versa, this constituting the best way of 

 operating a flicker photometer, as previously pointed out. 

 The first operation in this experiment consisted in finding, 

 by the flicker photometer criterion, the equal brightness 

 voltages for the lamps through their coloured glasses, each 

 illuminating one of the opal glasses, placed normal to it, and 

 these voltages were used throughout the work with them. 



In fig. 7 are shown the critical frequencies obtained on 

 carrying through the entire series of comparisons of the two 

 coloured fields, the various relative intensities being obtained 

 by moving the two lamps to various calculated positions on 

 the photometer bar. It will be seen that the points follow 

 the curves calculated from equation (2), (S = '0002) right 

 down to the sharp intersection of the colour flicker limiting 

 curve and then follow this across from one brightness flicker 

 limiting curve to the other. As they should be, the critical 

 frequencies at the end points are not exactly the same for 

 two coloured lights equal by the flicker photometer criterion. 

 The experimentally found relationships are in every respect 

 in agreement with those indicated by the theory. 



7. Measurement of the Hue Difference Discrimination 

 Fraction for Fluctuating Impressions. 



Exact determination of the speed at which colour flicker 

 appears, which is made possible by the new flicker photo- 

 meter, opens the way to the measurement of the just 

 perceptible hue difference corresponding to the brightness 

 difference fraction discussed in section 5. 



In the first theoretical paper it is shown in equation (9) 

 and its discussion, that if two coloured fields R and G of 



