So, also, the light from any metal sufficiently heated will give a spectrum, just as sunlight gives a spectrum, under the needful conditions. That is to say, there must be slit and prism, or slit and diffraction-grating, for the light to pass through. But the kind of spectrum is by no means always the same. Putting aside for a few minutes the thought of sunlight and starlight, let us look at the kind of rays or beams which are given forth by heated earthly substances. Any very much heated substance sends forth its light in rays or beams, and any such rays or beams may be passed through a prism, and broken up or “analyzed,” just as easily as sunlight may be “analyzed.” Suppose that we have a solid substance first--a piece of iron or of steel wire. If it is heated so far as to give out, not only heat, but also light, and if that light is made to travel through the slit and prism of a spectroscope, the ray will then be broken up into its sub-rays. They, like the sub-rays of a sunbeam, will form a continuous row of soft color-bands, one melting into another. This is the characteristic spectrum of the light which is given forth by a burning or glowing _solid_. Next, suppose we take a liquid--some molten iron or some molten glass, for instance. If you have ever been to a great plate-glass manufactory, like that at St. Helens, not far from Manchester, you will have seen streams of liquid glass pouring about, carried to and fro in huge caldrons, bright with a living light of fire from its intensity of heat. If a ray of _that_ light had been passed through slit and prism, what do you think would have been the result? A continuous spectrum once more, the same as with the glowing iron or steel. The light-ray from a heated and radiant _liquid_, when broken up by a prism, lies in soft bands of color, side by side. Both of them a good deal like the solar spectrum, you will say. Only here are no mysterious dark lines crossing the bright bands of color. But how about gases? Suppose we have a substance in the state of gas or vapor, as almost every known substance might be under the requisite conditions, and suppose that substance to be heated to a glowing brilliance. Then let its light be passed through the slit and prism of a spectroscope. What result shall we find this time? Entirely different from anything seen before. Instead of soft, continuous bands of color, there are _bright lines_, well separated and sharply defined.