We thus arrive at a picture of the solar atmosphere as consisting of layers of very hot gases, which are continually changing their relative positions and forms; sometimes ejections of intensely hot, glowing gases occur,--we call these prominences; sometimes down-rushes of gaseous matter occur,--we call these spots. Among the substances which compose the gaseous layers we recognize hydrogen, iron, magnesium, sodium, nickel, chromium, etc., but we also find substances which can at present be distinguished only by means of the wave-lengths of the light which they emit; thus we have 1474 stuff, 5017 stuff, 5369 stuff, etc.
Let us now turn to another part of this subject. By a special arrangement of apparatus it is possible to observe the spectrum of the light emitted by a glowing vapour, parts of which are hotter than other parts, and to compare the lines in the spectrum of the light coming from the hottest parts with the lines in the spectrum of the light coming from the cooler parts of the vapour. If this is done for sodium vapour, certain lines are apparent in all the spectra, others only in the spectrum of the light coming from the hottest parts of the sodium vapour: the former lines are called "long lines," the latter "short lines." A rough representation of the long and short lines of sodium is given in Fig. 7.
[Illustration: Fig. 7.--Long and short lines of sodium.]
Now, suppose that the lines in the spectrum of the light emitted by glowing manganese vapour have been carefully mapped, and classed as long and short lines: suppose that the same thing has been done for the iron lines: now let a little manganese be mixed with much iron, let the mixture be vaporized, and let the light which is emitted be decomposed by the prism of a spectroscope, it will be found that the long lines of manganese alone make their appearance; let a little more manganese be added to the mixture, and now some of the shorter lines due to manganese begin to appear in the spectrum. Hence it has been concluded by Lockyer that if the spectrum of the light emitted by the glowing vapour of any element--call it A--is free from the long lines of any other element--say element B--this second element is not present as an impurity in the specimen of element A which is being examined. Lockyer has applied this conclusion to "purify" various elementary spectra.