The red is least turned by the prism from its original course and the violet most. If between the prism and the screen on which the spectrum falls there were interposed a gas of any kind that gas would absorb from the coloured rays passing through it the exact waves of light with which it would itself shine if it were made luminous by heat. It would not take out an entire section, or colour, from the spectrum, but only a small part of one or more of the colours, and the absence of these parts would be indicated on the screen by narrow black lines situated in various places; and these lines, in number and in situation, would differ with every different kind of gas that was interposed. If several kinds were interposed simultaneously they would all pick out their own peculiar rays from the light, and thus the spectrum would be crossed by a large number of dark lines, by the aid of which the nature of the various gases that produced them could be told. The effect would be the same if the gases were interposed in the path of the white light before it enters the prism;—and this, in fact, is what happens when the spectrum of the sun, or a star, is examined—the absorption has already occurred at the surface of the luminous body before the light comes to the earth. ] Fundamentally, spectroscopic analysis depends upon the principle of refraction, of which we have spoken in connection with the atmosphere. Although the most powerful spectroscopes are now made on a different plan, the working of the instrument can best be comprehended by considering it in the form in which it was first invented, and in which it is still most often used. In its simplest form the spectroscope consists of a three-angle prism of glass, through which a ray of light is sent from the sun, star, or other luminous object to be examined. Glass, like air or water, has the property of refracting, or bending, all rays of light that enter it in an inclined direction. In passing through two of the opposite-sloping sides of a prism, the ray is twice bent, once on entering and again on leaving, in accordance with the principle that we have already mentioned (see Part II, Sect. 4). Still, merely bending the ray out of its original course would have no important result but for another associated phenomenon, known as dispersion. To explain dispersion we must recall the familiar fact that white light consists of a number of coloured components which, when united, make white. It is usual to speak of these primary, or prismatic, colours, as seven in number. These are red, orange, yellow, green, blue, indigo, and violet. Physicists now assign a different list of primary colours, but these, being generally familiar, will best serve our purpose. Without going into an explanation of the reasons, it will suffice to say that the waves of light producing these fundamental colours are not all equally affected by refraction.