clude that they represent non-terrestrial matter in the sun, as they may simply be uncatalogued faint lines of the metals; or, with equal probability, they may represent the constituent lines of a banded spectrum, the origin of which has not yet been traced. It has sometimes been supposed that these unidentified lines are the spectra of familiar elements which have been dissociated under the influence of solar temperatures, but the circumstance that the stronger lines can be produced under laboratory conditions renders this very improbable.
In the present state of our knowledge of spectroscopy, it would be very unwise to suppose that any non-terrestrial matter whatever is indicated by the solar dark lines, and the discovery of the terrestrial equivalent of the bright helium lines of the sun's chromosphere encourages the hope that another gas may yet be unearthed which will reproduce for us the bright line spectrum of the corona which is observed during eclipses.
One of the most striking facts relating to this analysis of the sun is the poor display which is made by the non- metallic elements. Silicon and carbon are the sole representatives of the metalloids which have been identified with certainty, while among the gases, helium, hydrogen, and oxygen are all that are directly inscribed, so far as our present knowledge extends. In addition, nitrogen is known to be there in combination with carbon in the form of cyanogen. This seeming absence of most of the non-metals from the sun has been explained, in accordance with experimental work, as a result of their spectroscopic ineffectiveness through admixture with metallic vapours; and, more recently, the Count de Gramont has shown that in the case of tellurium, phosphorus, arsenic, antimony, and boron, the most sensitive lines are situated in the ultra-violet, in that part of the solar spectrum which, is cut off by the absorption of our own atmosphere. The latter explanation also applies in the case of gold, and possibly also to some other metals which do not disclose their presence by Fraunhofer lines. Recent studies of the spectra of sun spots are also very suggestive in this connection. Lithium, for example, fails to show its characteristic lines in the ordinary solar spectrum, but the red line is strongly marked in the spectrum of spots, where the physical conditions are apparently more favourable
