But if the light looked at be white, the light of a candle for example, or of a jet of gas, instead of having a series of bright and dark bars, we have the bars _coloured_. And see how beautifully this harmonizes with what has been already said regarding the different lengths of the waves which produce different colours. Looking again at Fig. 17 we see that a certain obliquity is necessary to cause one ray to be a whole undulation in advance of the other at the point _l'_; but it is perfectly manifest that the obliquity must depend upon the length of the undulation; a long undulation would require a greater obliquity than a short one; red light, for example, requires a greater obliquity than blue light; so that if the point _l'_ represents the place where the first bar of red light would be at its maximum strength, the maximum for blue would lie a little to the left of _l'_; the different colours are in this way separated from each other, and exhibit themselves as distinct fringes when a distant source of white light is regarded through a narrow slit. By varying the shape of the aperture we alter the form of the chromatic image. A circular aperture, for example, placed in front of a telescope through which a point of white light is regarded, is seen surrounded by a concentric system of coloured rings. If we multiply our slits or apertures the phenomena augment in complexity and splendour. To give some notion of this I have copied from the excellent work of M. Schwerd the annexed figure (Fig. 18) which represents the gorgeous effect observed when a distant point of light is looked at through two gratings with slits of different widths.[B] A bird's feather represents a peculiar system of slits, and the effect observed on properly looking through it is extremely interesting. [Sidenote: COLOURS OF THIN FILMS.]