250 Notes 



a pencil, etc., one from the candle, and another from the day- 

 light received at a small opening of one of the window-shut- 

 ters ; the light of the candle will appear orange-coloured in the 

 daytime, and so will that shadow of the body which belongs 

 to or is made by the daylight; but the shadow of the body 

 made by the candle will surprise any person, by being of a fine 

 blue. 



19. More than once I have been agreeably struck with this 

 appearance, produced unintentionally when I have been writing 

 by candle-light on a winter's morning ; upon the daylight being 

 let in, the shadow of my pen and fingers in the orange-light 

 of the candle, were beautifully blue. 



20. I suppose there is such a thing as the harmony of 

 colours, of which painters speak so much ; according to the 

 explanation here given, our key to the solution of every case 

 of harmony and of contrast is to consider what is the other 

 colour, simple or compound, which, joined to a given one, 

 simple or compound, will constitute white. Thus red requires 

 green ; yellow, purple ; blue, orange ; and vice versa^ the mix- 

 tures in proper proportions will be white. 



21. Sir Isaac Newton (prop. 6, part 2, of book i. Optics) 

 has given a method forjudging of the colour of the compound 

 in any known mixture of primary colours, but it is not easy, 

 even for mathematicians, to put his rules in practice. The 

 gentleman who consulted me on this subject of shadows has 

 been accustomed, for a long time, to assist his memory, when 

 he is painting, by the use of the simple diagram [Fig. 28]. 

 Let R, Y, B represent the three uncompounded colours, red, 

 yellow, blue ; and let o, G, P represent the compounds orange, 

 green, and purple ; it is evident that, to make a deeper orange, 

 we must add more red ; and to make a bluer green, we must 

 add more blue ; and to make the purple redder, we must add 

 more red, and vice versa : but besides this, the diagram puts us 

 in mind that g is the contrast to r, and that, therefore, those 

 two colours cannot be mixed without approaching to a dull 

 whiteness or greyness ; and the same may be said of Y and P 

 and of B and o. These colours are also contrasts to each other ; 



