One part of refraction, namely, deviation, enables us to obtain, but the other half, dispersion, prevents our obtaining, except under certain conditions, an image we can make use of. By dispersion is meant the property of splitting up ordinary light into its component colours, of which we shall say more in dealing with spectrum analysis. If we wish to get more light by increasing the aperture of the telescope, the deviation of the light passing through the edge of the object-glass is increased, and with it the dispersion, the result of this increase of deviation. If the light of the sun be allowed to fall through a hole into a darkened chamber, and then through a prism, Fig. 41, it is refracted, and instead of having an exact reproduction of the bright circle we have a coloured band or spectrum. The white light when refracted is not only driven out of its original course—deviated—but it is also broken up—dispersed—into many colours. We have a considerable amount of colour; and this the early observers found when they increased the size of their telescopes, for it must be remembered that a lens is only a very complex prism. [Illustration: FIG. 41.—Dispersion of Light by Prism. ] First, they increased the size by enlarging the object-glasses, and not the focal length; but when they had done that they had that extremely objectionable colour which prevented them seeing anything well. The colour and indistinctness came from an overlapping of a number of images, as each colour had its own focus, owing to varying refrangibilities. They found, therefore, that the only _effective_ way of increasing the power of the telescope was by increasing its focal length so as to reduce the _dispersing_ action as much as possible, and so enlarging the size of the actual image to be viewed, without at the same time increasing the angular deviation of the rays transmitted through the edges of the lens. The size of the image corresponding to a given angular diameter of the object is in the direct proportion of the focal length, while the flexure of the rays which converge to form any point of it is in the same proportion inversely. [Illustration: FIG. 42.—Diagram Showing the Amount of Colour Produced by a Lens. ]