AND ITS APPLICATION TO THE DETERMINATION OF COLOUR SENSATIONS. 351 



the spectrum, L the lens throwing the image of the face of the prism on the screen G. 

 In the path of the ray X a plain glass mirror is inserted reflecting a proportion of the 

 beam to a second silvered mirror M a , which in its turn reflects the beam Y to C. 

 Sectors can be inserted in one or both of the beams X and Y. 



If the colour to be measured is that of a piece of (say) coloured glass, it is inserted 

 at D in the path of the beam X ; or if it be a pigment whose colour has to be measured, 

 it is placed at C, so that it is illuminated by X, and a white square placed adjacent to 

 it is illuminated by Y, a rod R being placed in a proper position to throw two shadows 

 touching each other at C. I have found that instead of using one plain mirror at M, 

 it is better to have a bundle of glasses, so that the intensities of the beams X and Y 

 are more equal than when a single glass is employed. 



The readings are made by equalizing the brightness of the illuminated shadows 

 first with the colour in position and then without. The two measures give the 

 percentage of light reflected or transmitted from the coloured medium or surface. 



(21.) Measurement of Emerald-green and Chrome-yellow. 



As examples of the way in which Table III. is to be used, the light reflected from 

 emerald-green, Table VI., and from chrome-yellow, Table VII., has been tabulated. 



In both tables Column I. shows the standard scale numbers. 



In both tables Column II. the relative intensity of the light reflected from the 

 colour compared with that reflected from a white surface. 



In both tables Columns III., IV., and V. are copied from X., XL, XII. , Table III. 



In both tables Columns VI., VII., and VIII. are III., IV., and V. multiplied by the 

 intensities in Column II. 



The areas of the curves of RS, GS, and BS in VI. , VII., and VIII. for emerald- 

 green are taken and found to be on an empiric scale (which is the same as that of 

 the luminosity of the naked spectrum of the crater of the arc light), RS 202, GS 133, 

 BS 1-418. 



GS and BS are multiplied by 2 '3 and 178 (the factors for making the sensation 

 curves of equal area) respectively, and found to be 306 and 252 respectively. The 

 lowest of the ordinates is RS 202. This must be deducted from GS x 2'3 and BS x 178, 

 and we have as the remainders 104 and 50. These must be divided by 2 '3 and 178, 

 and from these (which are 45'2 and '28) the percentages of GS and BS are calculated, 

 and are found to be 99'38 and '62. This, from the diagram and from calculation, 

 gives the dominant colour as SSN 35'64 or X 5070. 



The area of the spectrum curve is 830 on the same scale, and the sum of the three 

 curves is 336. The luminosity of the emerald-green, when white is taken as 100, is 

 Jrtt x 100, or 40'5. (This is the same as was made by direct measurement.) 



The amount of RS and GS and BS used to form the white is 290. The sum of the 

 areas of the three curves is 336. The percentage of white is therefore f~f x 1 00, 



