OXYGEN IN TtiE SUM } 



each element in the solar mass. In some cases, the quantity 

 of a material necessary to give unmistakable spectroscopic 

 evidence is singularly small ; insomuch that new elements, 

 as thallium, caesium, rubidium, and gallium, have been actu- 

 ally first recognized by their spectral lines when existing in 

 such minute quantities in the substances examined as to give 

 no other trace whatever of their existence. But it would be 

 altogether a mistake to suppose that some element existing 

 in exceedingly small quantities, or, more correctly, existing in 

 the form of an exceedingly rare vapour in the sun's atmosphere, 

 would be detected by means of its dark lines, or by any other 

 method depending on the study of the solar spectrum. When we 

 place a small portion of some substance in the space between 

 the carbon points of an electric lamp, and volatilize that 

 substance in the voltaic arc, we obtain a spectrum including 

 all the bright lines of the various elements contained in the 

 substance; and if some element is contained in it in ex- 

 ceedingly small quantity, we may yet perceive its distinctive 

 bright lines among the others (many of them far brighter) 

 belonging to the elements present in greater quantities. But 

 if we have (for example) a great mass of molten iron, the 

 rainbow-tinted spectrum of whose light we examine from a 

 great distance, and if a small quantity of sodium, or other sub- 

 stance which vaporizes at moderate temperatures, be cast into 

 the molten iron so that the vapour of the added element 

 presently rises above the glowing surface of the iron, no trace 

 of the presence of this vapour would be shown in the spec- 

 trum observed from a distance. The part of the spectrum 

 where the dark lines of sodium usually appear would, un- 

 doubtedly, be less brilliant than before, in the same sense 

 that the sun may be said to be less brilliant when the air is 

 in the least degree moist than when it is perfectly dry ; but 

 the loss of brilliancy is as utterly imperceptible in the one 

 case as it is in the other. In like manner, a vapour might 

 exist in the atmosphere of the sun (above the photosphere, 

 that is), of whos'e presence not a trace would be afforded in 

 the spectroscope, for the simple reason that the absorptive 



