OXYGEN IN THE SUN. 25 



the other hand, the experiment suggested by Dr. Draper 

 would not have the effect he supposes, if it were correctly 

 made. Doubtless, if the light from a considerable area of 

 dully glowing sodium vapour were received by the spectro- 

 scope at the same time as the light of a candle-flame seen 

 through the sodium vapour, the light of the sodium vapour 

 overcoming that of the candle-flame would indicate its 

 presence by bright lines ; but if light were received only 

 from that portion of the sodium vapour which lay between 

 the eye and the candle-flame, then I apprehend that the dark 

 lines of sodium would not only be seen, but would be con- 

 spicuous by their darkness. 



It is in no cavilling spirit that I indicate what appears to 

 me erroneous in a portion of Dr. Draper's reasoning on his 

 great discovery. The entire significance of the discovery 

 depends on the meaning attached to it, and therefore it is 

 most desirable to ascertain what this meaning really is. 

 There can be no doubt, I think, that we are to look for the 

 true interpretation of the brightness of the oxygen lines in 

 the higher temperature of the oxygen, not in the great depth 

 of oxygen above the photospheric level. The oxygen which 

 produces these bright lines need not necessarily be above 

 the photosphere at all. (In fact, I may remark here that 

 Dr. Draper, in a communication addressed to myself, men- 

 tions that he has found no traces at present of oxygen above 

 the photosphere, though I had not this circumstance in my 

 thoughts in reasoning down to the conclusion that the part 

 of the oxygen effective in showing these bright lines lies 

 probably below the visible photosphere.) Of course, if the 

 photosphere were really composed of glowing solid and 

 liquid matter, or of masses of gas so compressed and so 

 intensely heated as to give a continuous spectrum, no gas 

 existing below the photosphere could send its light through, 

 nor could its presence, therefore, be indicated in any spectro- 

 scopic manner. But the investigations which have been 

 made into the structure of the photosphere as revealed by 

 the telescope, and in particular the observations made by 



