We have now described the light we have adopted, and the reasons for adopting it; and having obtained our light, we can now consider by what plan we shall form our spectrum. There are two ways open to us--one by glass prisms, and the other by a diffraction grating. Glass prisms separate white light, or indeed any light, into its components, from the fact that the refraction of each coloured ray differs from every other. Thus the red rays are least refracted, and the violet the most, and the yellow, green and blue are intermediate between them, being placed in the order of least refrangibility. Between these there is of course every shade of simple colour, one melting into the other. In order to form a pure and bright spectrum with prisms, in a room of limited dimensions, we have to use certain auxiliary apparatus which are not positively essential, though convenient. The real essentials to form a spectrum are a narrow slit, a glass prism, with perfectly plane faces, and a lens. If this be the only apparatus available, the slit must be placed at a long distance from the prism, the beam of light must pass through the slit on to the prism, and the lens must be placed at such a distance from the slit that it forms a sharp image on a screen. When the light passes through the prism, the screen will have to be rotated in the arc of a circle, so that its distance from the slit measured along the line of the ray to the prism, and from the prism to the screen, is the same as it would be without the intervening prism. An apparatus of this description is not convenient, however, as it requires much more space than is often available. If a lens be placed between the slit and the prism, at exactly its focal length from the former, the light entering the slit will, after passage through the lens, emerge as parallel rays, that is, they will emerge as they would do if the slit were placed at an infinite distance from the observer. The focal length of this collimating lens need not be greater than twelve to eighteen inches, so that the great space required by the cruder apparatus is very much curtailed. The lens and slit are mounted one at each end of a tube of the necessary length, and are thus handy to use. Instead of one prism two or three may be used, giving an angular dispersion of the spectrum two or three times respectively greater than that which would be given by only one prism; consequently to obtain a given length of spectrum with the increased dispersion, the focal length of the lens used to focus the image on the screen may be diminished.