Fig. 8 is given in order to make this "purifying" process more clearly understood. But when this process has been completed there remain, in many cases, a few short lines common to two or more elementary spectra: such lines are called by Lockyer _basic lines_. He supposes that these lines are due to light emitted by forms of matter simpler than our elements; he thinks that at very high temperatures some of the elements are decomposed, and that the _bases_ of these elements are produced and give out light, which light is analyzed by the spectroscope. Such short basic lines are marked in the spectra represented in Fig. 8 with a positive sign. Now, if the assumption made by Lockyer be admitted, viz. that the short lines, or some of the short lines, which are coincident in the "purified" spectra of various elements, are really due to light emitted by forms of matter into which our so-called elements are decomposed at very high temperatures, it follows that such lines should become more prominent in the spectra of the light emitted by elements the higher the temperature to which these elements are raised. But we know (see p. 308) that the prominences around the sun's disc are hotter than the average temperature of the solar atmosphere; hence the spectrum of the light coming from these prominences ought to be specially rich in "basic" lines: this supposition is confirmed by experiment. Lockyer has also shown that it is the "basic," and not the long lines, which are especially affected in the spectra of light coming from those parts of the solar atmosphere which are subjected to the action of cyclones, _i.e._ which are at abnormally high temperatures. And finally, a very marked analogy has been established between the changes in the spectrum of the light emitted by a compound substance as the temperature is raised, and the substance is gradually decomposed into its elements, and the spectrum of the light emitted by a so-called elementary substance as the temperature of that substance is increased. But it may be urged that Lockyer's method of "purifying" a spectrum is not satisfactory; that, although all the longer lines common to two spectra are eliminated, the coincident short lines which remain are due simply to very minute quantities of one element present as an impurity in the larger quantity of the other. Further, it has been shown that several of the so-called "basic" lines are resolved, by spectroscopes of great dispersive power, into groups of two or more lines, which lines are not coincident in different spectra.