Visibility of the Different Colours of the Spectrum. 719 



second, third, fourth, arid fifth columns, that is, the distri- 

 bution of the energy in the spectrum as seen through the 

 eyepiece of the spectroscope. 



470 pn 



2-08 



2-57 



1-58 



•44 



3-74 



490 



1-87 



4-1.7 



1-63 



•57 



7-27 



530 



1-67 



9-33 



1-80 



■80 



224 



570 



1-59 



18-2 



1-98 



l'OO 



57-3 



610 



1-50 



30-9 



2-03 



114 



107 



650 



1-46 



51-3 



2-15 



1-33 



215 



In order to measure the influence of the Purkinje effect 

 luminosity curves were taken at two intensities for every 

 observer. The procedure on making a test was as follows. 

 First of all the gentleman supervising the test saw that the 

 voltage on the two lamps was right, that the instrument was 

 in adjustment, that the lamp B was in the first standard 

 position 92 cm. from the white comparison surface, and that 

 the lamp A was in its box in a standard position near the 

 ground glass. The intensity of the illumination falling on 

 the eye was then about one half metre-candle. The slit was 

 then set in succession at four standard widths, and for 

 each of these widths, on each side of the brightest part 

 of the spectrum, wave-lengths were sought for which the 

 intensity of the spectrum matched the intensity of the 

 white surfaces. Thus eight points were obtained on the 

 luminosity curve. Then for a wave-length midway between 

 the middle points already obtained the width of the slit 

 was diminished until the intensities matched, and an addi- 

 tional point thus obtained. This whole procedure was then 

 repeated twice for two other positions of the lamps A and B, 

 for which the intensity of illumination was ^J-q metre- 

 candle, and finally was repeated a fourth time for the 

 original intensity, so that in all four luminosity curves were 

 obtained each with nine points on it, two for a strong- 

 intensity and two for a weak intensity. At the end of every 

 test the position of the sodium line was read. The obser- 

 vations were entered on cyclo-styled sheets and the curves 

 plotted to a standard scale, strong intensity in black and 

 weak intensity in red. The complete set of readings took 

 three hours or more. 



In order to get the average luminosity curves the mean 

 of each observer's two u strong " and the mean of each 



