GASEOUS SPECTRA UNDER HIGH DISPERSION. 445 



Society's printed paper alluded to. The letterpress thereof certainly speaks of 

 four lines only, and gives the places only of 4, in figures; but in the map 

 accompanying them they are made into 8; viz., each line of the four is made a 

 double line; two, much more distinctly so than the other two. It is not for me 

 to pronounce on the accomplished fact of the Society's thus doubling the 

 number of lines in a very scantily furnished spectrum; and making the original 

 single lines of observation conform more nearly to the theory they publish, by 

 representing them " compound " to the extent of doubling each one, — but it is 

 absolutely necessary for truth's sake to warn all into whose hands the Philo- 

 sophical Transactions may come, of the absolute falsity of the (London) Royal 

 Society's Oxygen spectrum plate, in that respect. 



The talented author of the paper, moreover, has never claimed to have 

 seen more than four single lines, placed as described; has made an immense 

 number of most admirable experiments to assure himself that they belong to 

 pure Oxygen, and not to any accompanying impurity, — and that there is 

 nothing else in the O spectrum of equal visibility. That degree of visibility 

 however, being something very small ; for Oxygen gas is what is generally 

 known, as a bad illuminator, in all vacuum tubes. 



There then, with only four truly observed lines, the tube spectrum of 

 Oxygen might have remained, had I not in 1879, independently of the late 

 energetic Dr Van Monckhoven, both struck, and worked out, the idea of using 

 vacuum tubes end-on, in place of transversely to their capillary part, as others 

 seem to have done universally before that time; some of those earlier observers 

 even using the tube's upright line of light, in place of the slit of a spectroscope 

 proper. But with the new end-on vacuum tubes, and equally when they were 

 made for me in Paris, or in London, I immediately, through the greater bright- 

 ness of their light, saw the presence of many fainter features constant in, or 

 evidently belonging to, the O Spectrum of the four lines. 



First, for instance, I found that three, out of the four, primitive lines were, 

 each of them, a triple. Each triple a long way from its nearest neighbour, but 

 of precisely similar build ; and I have since then discovered three other such 

 triples, one of them further away towards the Red, than the longer known 

 Orange one; and two others further towards the Blue, than the older Green 

 one. 



They make moreover a remarkably connected, though wide apart, and only 

 faintly luminous system altogether, extending through so great a range of the 

 Spectrum as from Red to Glaucous; for the six triples are arranged in three 

 pairs, whereof the mean place of the third pair is from the mean place of 

 the second pair, close on half the distance that the mean place of the second 

 is from the mean place of the first; while at the same time the much smaller 

 distance apart of the sixth triple from the fifth is just about half that of the 



