ON COLOUR PHOTOMETRY. 557 



disc, about 24 inches in diameter, was tuken, and commencing about 4 inches from its 

 centre the scale of the spectrum was laid off on a radius. Concentric circles were 

 drawn through the different points of the scale, and, from the centre as origin, angles 

 were laid off proportionate to the height of the ordinates of the emerald green curve 

 (fig. 16) ; the intersections of the lines forming these angles with the circles through 

 the corresponding abscissae gave a figure which was cut out of the disc (Plate 23, 

 fig. 19). The latter being rotated in a proper position in front of the spectrum 

 allowed the proper proportions of the different rays of the spectrum to pass through, 

 and then, being recombined on the screen, produced the colour of the pigment. The 

 identity of the colour with emerald green was proved by reflecting white light on to 

 a square of paper coloured with that pigment placed near the colour patch. A similar 

 mask (fig. 20) being cut out to correspond to the ordinates of the curve No. IV. 

 taken above the tangent to its lowest point produced a grey patch on the screen. 



(In the prismatic spectrum the rays are more or less curved ; but, as their curvature 

 will not correspond at all points to that of the disc, it is necessary, in order to obtain 

 a correct result, to reduce the breadth of the spectrum by shortening the collimator 

 slit.) 



A further proof of the accuracy of the measurements was also adopted. The three 

 discs were rotated with a larger white * disc on the same axis ; a card, having the 

 curvature of the outside circle of the coloured discs and a breadth the same as that 

 of the rod usually employed, replaced the rod, and, as before, the rays from the 

 two spectra cast shadows, one on the white rotating disc and the other on the 

 rotating coloured sectors, which, it should be recollected, to the eye gave a grey. 

 Measurements were made as before, and the readings being reduced proportionally to 

 the white which would have been present when the black sector was rotated with the 

 white sector, the curve No. V. in fig. 16 was obtained. At first it would appear that 

 the brightness of the mixture was too great ; but, as a matter of fact, it was not, for 

 the white card employed in the experiment was slightly greyer than that used in the 

 white disc which, when rotated with the black, gave the grey. It will be noticed 

 that this curve is practically parallel to that obtained by the summation of the three 

 luminosities ; this appears to confirm its correctness. If any curve give* a grey, then 

 any other curve parallel to it will do the same. 



XXXVI. Experiments with Yellow and Blue Discs. 



There are two favourite colours which are often used in class demonstrations to 

 show the formation of a grey on their rotation in proper proportions. One is pale 

 yellow chrome, and the other a French blue, also very pale. Discs were prepared 

 with these colours, and the grey produced matched by black and white. The colours 

 were subsequently measured by the plan already described ; the following tables give 

 the results of the measurements 



The beam lighting this liad passed through the rotating sectors black and white sectors, therefore, 

 could not be used, since scintillation was produced. 



