1898.] on the Theory of Colour Vision, &c. 805 



In the small spectrum we can place three slits and make a patch of 

 white light. By altering the width of one or two of the slits we can 

 form colours of any hue. [White was here matched, and three other 

 colours made, and again whiteJ\ Instead of light being diminished 

 by alteration of the width of the slit, we can cut off varying quantities 

 from each ray and allow them to impress the retina for different 

 times, tbe persistence of vision blending the impressions together. 

 In fact, by an artifice of the kind I have here, which consists of a 

 long band of paper punctured along the three lines of the slits with 

 holes of different sizes, and passing the strip in front of the slits, we 

 can play a regular tune in colour. [^Shoian!] 



But we can get these same tunes of colour, though not quite so 

 pure (i.e. unmixed with white) if we use considerable parts of the 

 spectrum. The slits are withdrawn, and all those parts of the 

 spectrum which can come through the holes are mixed together and 

 the colours are reproduced, but not in quite such purity as before. 



There is another method of altering the intensity of the rays, and 

 that is by placing in front of the slits photographic deposits of 

 different opacities, and you see that we have, as before, different 

 colours produced. The diagram (Fig. 3) shows a print of the de- 

 posits employed. The three rows represent the transparencies of 

 white, violet, blue, peacock blue, dark green, orange and red, taken 

 through an orange, a green, and a blue screen respectively. The 

 three left-hand squares in the transj^arency covered the three slits, 

 and white was formed on the screen. The next three squares gave a 

 violet, the next three a blue, and so on. This is the foundation of 

 colour photography. Having learnt that the colours mixed together 

 need not be single rays of the spectrum, but may occupy adjacent 

 parts on each side of the single ray and still produce approximately 

 the same results, we can go a step further, as it shows that we may 

 use the light coming through media such as coloured glasses instead 

 of pure spectrum colours. 



An interesting experiment is to imitate the sj)ectrum by means of 

 a red, a blue and a green glass. A slit is placed in the lantern, 

 and an image of it thrown on the screen. We have a disc formed of 

 these three glasses, each being shaded by an appropriate mask, to 

 imitate the extent of each sensation in the spectrum. This disc 

 rotates in front of the slit. The varying combinations give a large 

 range of colour, and we have a tolerable representation of the 

 spectrum produced. 



I think now we are in a position to realise what is required in 

 order to reproduce by photography the spectrum with all its colours. 

 TVe must get three photographic negatives, each one of which will 

 take in only so much of the spectrum as is represented by the colour 

 sensations as shown in the diagram, and secure that the brilliancy of 

 the light coming through the transparencies or positives taken from 

 the negatives at each part shall be represented by the heights of the 

 curves, the maximum height in the positive being represented in each 



Vol. XV. (No. 92.) 3 g 



