378 



Transactions of the Canadian Institute. 



[Vol. VII. 



colour of any part of the spectrum, and, from what has just been said, 

 can therefore act as standards indicating the correct opacity patches 

 on a set of negatives of the spectrum. The projection, upon the screen, 

 of three negatives of the spectrum on one plate (Fig. 3) shows the 

 approximate agreement of the deposits with the curves ; while a trans- 

 parency from these negatives shows the agreement between their 

 relative transparencies and the curves. 



The method of obtaining such negatives is that of selective 

 absorption. The effect of interposing red, green, and blue glasses or 

 stained films in a beam of white light is to absorb, roughly speaking, all 

 but the red, green, and blue light, respectively ; this can be easily shown 

 by interposing them in the beam producing the spectrum \shown\ If 



FlU. 4 — Action of the Spectrum on Photographic Plates. 



they be placed in the path of the light entering the camera, near the lens 

 say, they will perform the same office of picking out or selecting, hence 

 the name, the red, green, and blue parts of the object, and allowing only 

 these parts to be registered on the respective plates. This seems simple 

 enough, but in practice the question of colour screens or filters, as the)- 

 are termed, is much complicated by the fact that no photographic plate 

 yet produced represents the spectrum in degrees of monotone corre- 

 sponding to the visual intensity. It will probably be remembered, b}- 

 those who heard the ])aper presented by the writer to the Institute last 

 year on " Colour Values in Monochrome," that, only after tedious 

 experimenting, could orthochromatic plates be made to render the 

 spectrum approximate!)' correctly. The curves representing the action 

 of the spectrum on sucVi plates are shown in Fig. 4, by the irregular 

 curves in each part of the diagram, the regular curves indicating the 



