1835.] Composition of WJiite Light. 115 



siderable depth of the crimson glass to effect this, the red 

 will be more or less inclined to yellow. 



In making these and similar experiments, we must 

 constantly recollect that no colour can be a compound of 

 more than two of the primitive colours : we must not, 

 therefore, when, for instance, we add red to blue and 

 produce violet, suppose that violet is a compound of blue 

 and red ; the fact is, that in this, and all similar cases, a 

 colour is discovered which happens to be in excess after the 

 other colours are exhausted in the formation of white light. 

 Blues dyed with indigo, always transmit and reflect a larger 

 portion of violet than of green light ; hence, the peculiar 

 colour of this dye, which is distinguished by forming a 

 part of Newton's scale ; and hence, the violet colour which 

 is discovered when the green and part of the violet are 

 neutralized by the addition of red. The usual method of 

 dyeing violet, is to give a light blue ground with indigo, 

 and then convert the blue to violet, with cochineal or some 

 other crimson dye. It is remarkable, that the different 

 simple colours are all produced in the greatest perfection 

 by a mixture of two compound colours ; thus, crimson and 

 yellow form scarlet, blue and yellow form green, and blue 

 and crimson form violet. We have, therefore, just as much 

 reason to infer, that red is a compound of crimson and yel- 

 low, and violet a compound of crimson and blue, as that 

 green is a compound of blue and yellow. 



Before I conclude this paper, I wish to say a few words 

 relating to the theory of accidental colours. 



The usual method of obtaining ocular spectra, by impres- 

 sing the eye with coloured wafers placed upon a white 

 ground, is well known. 



Not being able to produce colours of any great intensity 

 by this method, I have tried various others, among which 

 the following has been found both convenient and in- 

 structive : 



In a card, white on one side and black on the other, I 

 cut an aperture, about the sixteenth of an inch in breadth 

 and half an inch in length, and place before it coloured 

 glass or glasses, which transmit light of the colour upon 

 which I want to make the experiment. I then look at the 

 aperture, the card being held before a window, until the 



I 2 



