FORMING THE LOW TEMPERATURE SPECTRUM OF OXYGEN. 425 
ultimate constitution of the a (alpha), to be perfectly different from that of A 
and B; for in place of many regular and symmetrically arranged powerful lines, 
it was made up of little doublets and triplets of both inconceivable minuteness 
and of very irregular occurrence (see the said Lisbon spectrum as printed in 
Vol. XXIX. of the Society’s Transactions). These Lilliputian foundation 
stones of the a (alpha) band appeared to me at the time almost ridiculous in 
their smallness ;—but now I recognise them as having a close family resem- 
blance to the triplicity of the low-temperature oxygen lines which I have been 
trying hereinbefore to describe. They are not indeed the very same, for they 
are in different spectrum places ; but they do give the idea (suggested also by 
some points in both the aurora spectrum and cometary spectra not yet repro- 
duced in any laboratory electrical experiments) that there is a temperature- 
level in Nature for the incandescence of gases, much lower than the 
low-temperature oxygen spectrum of these pages. Wherefore, if we could 
artificially produce that kind of ultra low-temperature illumination, electric 
probably, we might find that the a (alpha) band in the solar sunset spectrum 
represented the oxygen, while the A and B bands showed us the nitrogen 
gas thereof; they two being the mighty gas constituents known so well to 
everyone, except telluric-line solar spectroscopists, to exist in the earth’s 
atmosphere; and in such overpowering quantity as to practically exclude 
everything else except watery vapour. 
