376 Transactions of the Canadian Institute. [Vol. VII. 



these spectrum primaries, he distinguished visually from spectrum or 

 natural colours. By closing the green and blue-violet slits, the red of 

 the spectrum, through the third slit, colours the patch red ; the gradual 

 opening of the aperture in the green produces, by the mixture of red 

 and green, orange-red, orange, and yellow ; the green aperture 

 remaining open and the red being closed gradually, gives yellow-green 

 and green. The same procedure followed with the green and blue- 

 violet slits produces blue-green, blue, and blue-violet ; and with the 

 blue-violet and red slits forms violet, purple, and red, the whole range 

 of spectrum colours including also the purples. (Matching of colours 

 shozvn). Any colour in nature may be matched in like manner, the 

 tints being produced by first matching the hue, and then opening all 

 the slits sufficiently to add the required amount of white ; the shades 

 being produced by making the slits narrow enough to sufficiently 

 diminish the luminosity. 



The correct proportions of the primary colours necessary to 

 reproduce the colours of the spectrum are indicated graphically by 

 Maxwell's sensation or, more properl}', colour-mixture curves (Fig. i). 

 These curves, determined b}^ the aid of his colour box, a modified form 

 of spectroscope, and in principle similar to the experiment just 

 described and shown, do not so nearly represent the stimulation of the 

 fundamental nerve sensations or processes of the eye to produce any 

 colour sensation, as the amount of the primary colours required to 

 match the same colour. Although the positions of Maxwell's primaries 

 are not quite correctly chosen, the amount of error introduced in his 

 colour-mixture curves is practically negligible ; and hence the amount 

 or intensities of the primaries necessary to reproduce any spectrum 

 colour are immediately given by measuring the lengths of the corre- 

 sponding ordinates to the curves. The relative intensities of the 

 primaries necessary to produce any colour whatever can be determined 

 by matching the colour by the three slits, and then measuring their 

 relative widths ; unit width being determined by the relative apertures 

 required to give white light. 



The possibility of matching any colour by the mixture of the three 

 primary or reproduction colours forms the basis of the three-colour 

 process of photography. If, by any photographic process, three 

 coloured images of any coloured object can be obtained, one red, one 

 green, and one blue-violet ; and if each image contains in its various 

 parts, the correct proportions of its own colour required to match the 

 colours of the object, then the optical superposition or mixture of these 



