66 Mr. J. S. Dow on the Physiological 



brightness method and then, by setting the motor in rotation , 

 to produce the proper-sized band of flicker. 



Although, as stated above, very approximate, these readings 

 certainly seem to warrant the suggestion that the Purkinje 

 effect cannot be observed to the same extent in the case of 

 the flicker sensation. A typical result is that shown in 

 fig. 4. 



Comparison of Purkinje Effect for Flicker and Equality of 

 Brightness Photometers. 



1-0 



0*9 



l 3 





0-8 



6 0-7 



0-6 





, 











•« 



• 





Fuck Eft 





• 



* 





■ — * T" 



s^** 





x /X 



/ x 



X 



"" ft? 6 

 X 

































1-0 2-0 3-0 4-0 5-0 



Illumination of Photometric surfaces in Lux. 



60 



It may be added that other observers have recorded similar 

 experiences. For instance, in the discussion following a 

 paper by Dr. Steinmetz before the Illuminating Engineering 

 Society of New York in December 190G, Mr. P. S. Millar 

 referred to some tests on a mercury arc-lamp, which were 

 carried out on a long photometrical bench with different 

 distances between the sources compared and with a corre- 

 sponding range of variation in the illumination of the photo- 

 meter-screen. A Lummer-Brodhun photometer was first 

 used, and it was then found that the candle-power of the 

 mercury -vapour lamp, compared with a carbon-filament 

 standard, apparently changed gradually, as the illumination 

 of the photometer-screen was reduced, to a value higher by 

 30 per cent, than that at normal illuminations. But when a 

 flicker photometer was used this difference did not occur. 



