viction that the matter distributed throughout the visible universe is fundamentally of the same composition as that of the earth was forced upon the majority of those who considered the subject in its various aspects. Although little can be advanced in the shape of new arguments, it will perhaps be of interest to reconsider the chemical unity of the cosmos in the light of some of the new facts with which the science of astrophysics has been enriched.
It is scarcely necessary to explain that the presence of a substance in a luminous source, whether this source be on a laboratory table or in a remote part of the universe, is indicated by the coincidence of spectral lines. In some of the cosmical sources, the spectral lines are produced by the direct emission of luminous gases and appear bright ; in others, the lines appear dark on a background of continuous spectrum, the reversal depending upon the property possessed by a vapour of absorbing precisely those radiations of which its emission spectrum consists at the same temperature. In either case the mode of identification is the same.
In this connection, it is a fortunate circumstance that the spectra of the various elements consist of more than one line, ranging from about a dozen in the case of lithium to some thousands in iron, uranium, and certain other elements. Hence, the identification of an element in a celestial source is in general independent of the establishment of absolute coincidences in the positions of the lines with those produced terrestrially, and rests rather upon the apparent agreement in position of a succession of lines having the same relative intensities in the celestial and terrestrial sources.
There are indeed several disturbing causes which combine to make it in most cases impossible to prove absolute identity of wave-length in the two sources. If a star be under observation, its spectral lines will be displaced by an amount depending upon its velocity in the line-of-sight (« radial velocity ») relatively to the earth; to the red if the earth and star be receding from each other, and to the violet if they be approaching. The displacement due to the earth's orbital motion can easily be calculated, and appropriate allowance made for it, but that caused by the motion of the star itself will always remain. It is only by assuming that the true wave-lengths of the stellar lines are identical with those of
