forming the Low-Temperature Spectrum of Oxygen. 335 



what shall be said for its continued silence as to the presence 

 of any free oxvgen in the earth's atmosphere, when some 

 gases that are therein do make themselves most signally con- 

 spicuous on the solar spectrum, in the shape of dark lines or 

 bands, growing as the sun descends in altitude, until they 

 become at last more grandly thick and black than all the truly 

 solar markings put together ? 



Amongst these gigantic eases of Fraunhofer lines of some 

 kind of telluric gaseous origin, we do indeed know, by a sort 

 of inductive process rather than by any positive proof, that 

 the band too hastily called " little a " and other bands and Hues 

 near C and near D are the spectroscopic proofs of watery 

 vapour, as an invisible but potent gas in our atmosphere. 

 But what makes the far greater A or great B and the a (alpha) 

 band, no one pretends to know. 



At first sight it might well be suggested that they must be 

 formed by one or other, or both, of the two grand constituents 

 of the earth's atmosphere as established by the chemists, viz. 

 oxygen and nitrogen. But there we are instantly met by 

 Angstrom's inflexible law, promulgated by him in 1853, and 

 repeated at p. 39 of the description of his " formal Solar 

 Spectrum " in 1868, viz. that " a gas in the state of incandes- 

 cence emits luminous rays of the same refrangibiliry as those 

 which it can absorb" — or, conversely, when it acts by absorp- 

 tion, it produces dark in all the spectrum places where, when 

 in a state of incandescence, it produced bright lines. 



Now we know by laboratory experiments what bright lines 

 are given out by both oxygen and nitrogen in various states 

 of incandescence : and not one of those lines is in the place, 

 and at the same time endowed with the physiognomy, of great 

 A or great B, or of the « (alpha) band of the sunset solar 

 spectrum. 



To get over this astonishing difficulty, Angstrom — who held 

 that all these three bands are of a similar visible constitution, 

 viz. a thick line and then a band of thin lines stretching out 

 towards the red end of the spectrum — suggested that they 

 might be produced by " carbonic acid." But, over and above 

 the difficulty that carbonic acid is an almost insensible impurity 

 in the open air, and could hardly be expected to extinguish 

 every sign of the existence of the atmosphere's two great con- 

 stituents, we do know the spectrum given by incandes- 

 cent carbonic acid in the laboratory, and it cannot claim to 

 having a band in the place of either A or B or a. (alpha), 

 besides its series being turned in the opposite direction, or 

 every band vanishing towards the violet, in place of the red, 

 end of the spectrum. 



