ON COLOUR PHOTOMETRY 



561 



vermilion appeared to him as dark green, the only effect of introducing the red sector 

 into the coloured disc was to lower the total luminosity, and to diminish the quantity 

 of green necessary to produce with the blue a balance against the black and whit* 

 discs. Thus, with 126 of the vermilion card, 103 blue and 131 green matched 

 99'5 white; and, with 188 vermilion, it required 110 blue and 62 green to match 

 91-5 white. 



In our fonner paper we gave R.'s spectrum curve, which indicated the proportion of 

 light which he receives from each part of the spectrum, as compared with normal 

 sight. Fig. 24 shows the curves of luminosity of the light reflected from the coloured 

 sectors, and the same reduced so as to correspond to R.'s sight are indicated by dotted 

 lines. The next table gives the numerical results of this reduction. 



The areas of the reduced curves are green 172, vermilion GO, and ultramarine 36. 

 The angular values of the sectors, it will be remembered, were 133'7, 96'6, and 1297. 

 Taking the areas to represent values of luminosity, R.'s values per 1 of sector are 

 emerald green T286, vermilion '621, French ultramarine blue '278. The area of the 

 normal curve for white, as stated above, is 534, representing 133'7. The area of 

 R.'s curve is 343, giving a value of 2'566 for 1. 



Applying these values to the observations, we find a very close correspondence when 

 the two colours were used, but not quite so close when the red was introduced. The 

 angle of white in R.'s observations was 1357, which, multiplied by 2'566, gives 346 

 as the value of the luminosity. The value of the blue luminosity is 115 X '278 or 

 32, of the green 245 X T286 or 315 ; and these added together make 347, which is 

 very close to the value obtained for the white. 



In the second observation similar calculations will make the value of the white 255, 



MDCCVLXX X VIII. A. 



4 c 



