the spectrum of the same substance under different conditions, which was initiated by Sir Norman Lockyer. In many stars which exhibit spectra of earlier type than that of the sun, it has been found that a large proportion of the lines correspond with metallic lines which are specially developed in the electric spark, whereas in the solar stars the lines occurring in the electric arc spectra are the predominant feature. The same element in different types of stars thus shows different spectral lines, and some of the most striking differences in stellar spectra have thus received a completely satisfactory explanation without the aid of hypothetical elements. It is interesting also to note that in some of the earlier types of stars there is a special development of lines produced by the absorption of gases, including nitrogen and oxygen, which are very inadequately exhibited in the spectrum of the sun.
The outcome of the study of the various groups of stellar spectra, from our present point of view, is to show that there are but few indications of stellar elements which do not visibly reveal themselves in the sun also, either by the same or different families of lines. Sulphur, which has been traced by Lockyer in some of the stars, is about the only one which can be pointed to with certainty. There are certainly some stellar lines which await explanation, but they are now relatively few, and one can scarcely escape the conviction that they will eventually yield to experimental reproduction from terrestrial matter, as has so often happened in the past.
Valuable evidence is furnished by binary stars. The probability that these systems have originated in the fission of a single mass of matter necessitates identity of composition of the components. Yet there are many binary stars in which the spectra of the two members are of different types, and this fact greatly strengthens the view that stars having different spectra do not necessarily differ in composition. The generally accepted view that the ancestors of the stars are represented by existing nebulae demands that these bodies should contain all the materials of which stars are known to be composed. And yet the nebulae have a very simple spectrum of bright lines, among which only those belonging to hydrogen and helium have been certainly identified. One possible explanation of this seeming anomaly is that the nebulae may be in so primitive a state that they contain
