1891.] Visibility of the different JRays of the Spectrum. 517 



about X 4510, as shown in Table II, is due to the fact that equal 

 luminosities of each colour have been considered as being reduced. 



Some interesting experiments were carried out by placing slits in 

 different parts of the spectrum, and forming a mixture of light on the 

 ground glass of the apparatus. An intense D light mixed with a 

 faint light near F formed a colour patch, and this mixed light was 

 extinguished and found to require 9 of aperture of the sector. The 

 D light was then shielded and the single ray of blue-green light was 

 extinguished, when it was found that the same aperture was required 

 to extinguish this beam alone. Red and green and various other 

 mixtures were tried, all showing that in the extinction of light the 

 green-blue light was the last visible, and was equivalent to extin- 

 guishing that light alone, although it might be mixed with very much 

 brighter light in the red or yellow. In the blue the conditions 

 somewhat change, as will be seen in the diagram, but if slits of equal 

 aperture were used the same results were obtained. 



The diagram shows that in the spectroscopy of feeble light the rays 

 in the blue and green are the first to be perceived, and that rays of 

 far greater intensity in the yellow and red may exist without ex- 

 citing the sense of light. This may account for some of the varied 

 results recorded in eye spectroscopic observations of sources of feeble 

 luminosity, in which the yellow and red lines are absent. 



In extinguishing white light, the fact of the total extinction of the 

 blue-green light is of importance. 



It is not the light at that particular wave-length which disappears 

 last, but some one sensation which is principally existent at that point, 

 but which extends over a great portion of the spectrum which has 

 to be extinguished. For instance, in extinguishing the light from 

 the reflected beam of the electric light already alluded to, it was 

 found that the light illuminating the ground glass was 720 times 

 brighter than that reaching the screen. To extinguish 0'014 

 of the light from an amyl lamp on the ground glass the sector had 

 to be closed to 21, that is the light of one amyl lamp luminosity, 

 falling on the screen at 1 foot distance, had to be reduced to 



14 1 21 1 



x ^571 x T57^ or 77T-t ^ o the original light. Had the luminosity 



1UOO 720 180 441,000 

 of the unit of luminosity been due entirely to the colour at X 4776, it 

 would have had to be reduced to about oooVoo of its luminosity before 

 it became invisible. Thus the electric light gives about half the 

 sensation of this light that the monochromatic light of that colour 

 and luminosity would give, and hence we may conclude that about 

 half the luminosity of the white light is due to this sensation, of 

 course distributed unequally through its spectrum. This is a very- 

 close approach to the area of the green sensation curve of the 

 spectrum when the luminosity is taken into account. 



