In a general way it is found convenient to remember, or rather to assume, that three parts of red, five parts of yellow, and eight parts of blue form together white, and, therefore, that the pencil of white light contains three rays of red, five of yellow, and eight of blue. To produce the other prismatic colours, we must mix red with a little yellow to form orange; yellow with some blue to form green; much blue with a little red to form indigo, and a little blue with some red to form violet. In performing experiments on colour it is convenient, instead of a drop of water, to substitute a prism of glass in decomposing the rays of light. We may thus produce at will a convenient image, called a _prismatic spectrum_, which, when thrown on a wall, is a broad band of coloured lights, having all the tints of the rainbow in the same order. Looking at this image, the red is at the top and the violet at the bottom, and it may be asked, How does the red get amongst the blue to form violet, if the red rays are bent up to the top of the spectrum? The answer is, that a quantity of white light not decomposed, and a part of all the colour rays, reach all parts of the spectrum, however carefully it is sheltered, but that so many more red rays get to the top, so many more of the yellow to the middle, and so many more blue to where that colour appears most brilliant, that these are seen nearly pure, whilst where the red and yellow or yellow and blue mix they produce distinct kinds of colour, and where the blue at the bottom is faint, and some of those red rays fall that do not reach the red part of the spectrum, the violet is produced. In point of fact, therefore, all the colours of the spectrum, as seen, are mixtures of pure colour with white light, while all but red are mixtures of other pure colours with some red and some yellow as well as white. Primitive and pure colours, therefore, are not obtained in the spectrum, and a question has arisen as to which really deserve to be called pure, Dr. Young upholding green against yellow, and even regarding violet as primitive, and blue a mixed colour. A consideration of the results of this theory would lead us farther than is necessary for the purpose we have now in view. We also find philosophers now-a-days calmly discussing a question which most people considered settled very long ago, namely, whether blue and yellow together really make green.