What has been shown to be true for calcium vapour in this special case is doubtless equally true for all the elements which have been spectroscopically identified in the sun and other cosmical bodies by the usual methods. The mode of identification of the chemical elements in the celestial bodies thus rests on a sure foundation.
Investigations of the spectra of the heavenly bodies long ago revealed the wide diffusion in space of some of our chemical elements, but complete uniformity has even yet not been established. Several of the known chemical elements have not been traced in any celestial source, and some of the lines which are found in the spectra of the heavenly bodies have not yet been identified with known chemical elements, nevertheless, as observational and experimental work has progressed, there has been a gradual removal of this inequality. The number of elements recognised in celestial bodies has steadily increased, and the number of « unknown lines » in celestial spectra is constantly being diminished, as the outcome of continued research. We may well look forward with confidence to the time when all celestial spectra will come within the scope of laboratory reproduction, either by the further production of new lines in the spectra of known substances under altered conditions, or by the discovery of new elements. An interesting example of one of these processes is the writer's discovery that carbonic oxide at extremely low pressures changes its more familiar spectrum to one which is identical with that of the tails of comets; and an instance of the other process is the recent detection by Ross of the comparatively new element Dysprosium in the sun's chromosphere.
Let us consider how the matter now stands with regard to the sun, which can be most completely studied, partly on account of its relatively small distance, and partly because of the fortunate circumstance that the apparent size of the moon is such as to make possible a total eclipse of the sun's most luminous surface, so that the fainter outlying parts become visible. Of the twenty thousand dark lines in the solar spectrum catalogued by Rowland, the great majority of the stronger lines have already been matched by spectra produced in the laboratory, chiefly from common metals such as iron, manganese, calcium, and so on. The unidentified lines are mostly very attenuated, and it is too early to con-
