SENSATIONS IN TERMS OF LUMINOSITY. 281 



(XXII.) Confirmation of the Observations by Colour-blind Persons. 



In confirmation of the positions of A and D, fig. 1, as stated before, complete red- 

 and green-blind eyes were called in to make observations. A totally green-blind 

 person gave the following readings for that position in the spectrum where white was 

 matched. The Scale Numbers used are those of the diagram. 



The white was placed alongside the spectrum colours, and the slit through which 

 the colour issued was gradually moved from the red towards the neutral point. The 

 same procedure was adopted in moving from the blue towards the same point. 



The readings were 



From the red From the blue. 



38-5, 37-5, 38, 37'1 36'8, 37'1, 37'5. 



The mean of these readings is 377 as the neutral point of the green-blind. The 

 difference is only '2 of the scale. A look at the curve will show that there is likely 

 to be greater variation when moving the slit from the red, as the curve is there less 

 steep. A mean of the readings " from the blue" gives 37'1, the position fixed by the 

 preliminary trials, and which answers to SSN 288'5. 



Two red-blind persons marked the point in the spectrum where the colour matched 

 white. One read 34'2, 35, 35'4, or a mean of 35'1 ; the other read 35'2, 35, 35'6, 

 35*8, or a mean of 35'4. The mean of the two means is 35'2. 



From the coincidence of the areas with the colour equations, and from the position 

 of neutral points of the colour-blind with the points where the curves met, we may 

 conclude that the observations are correct within the limits of the errors of observation. 



It may be stated that the nearest approaches to the colour sensations in pigments 

 are : Vermilion, to which a little blue has been added ; emerald green and true ultra- 

 marine, to which a slight trace of red has been added. GREVILLE'S cyanine blue is 

 not far from the colour. All are slightly paler, however, and in using them as colour 

 discs this paleness must be allowed for. 



(XXIII.) The Observations applied to the Normal Spectrum of the Electric Arc 



Light. 



Before proceeding further, I have thought that it would be of interest to show the 

 colour sensations of a normal spectrum. The compression of the red in the prismatic 

 spectrum and the extension of the blue does not enable a comparison to be easily made 

 between the sensation curves of this spectrum and the results obtained by K<ENIG, 

 which are based on the normal spectrum. 



The following table is calculated from observations made with a grating spectrum 

 in 1891. The grating was ruled on speculum metal, and had about 14,000 lines to 

 the inch. 



\ol. . \ in. A. 2 O 



