GASEOUS SPECTRA IN VACUUM TUBES, 97 
with a magnifying power of 10 on the telescope. So low a prism power was 
chosen at starting, to be certain of including everything even of the faint 
terminations, as well as of the brighter, but often discontinuous, portions of the 
middle, of any spectrum ; and it is far too small to allow of accurate measures 
of place. But as spectroscopic place, with some slight reference to colour and 
brightness, is the great identifier of spectrum lines, I have attempted to 
measure the place of everything seen in the tubes (at least with care), 
and have reduced mere instrumental numbers to the absolute scale of wave- 
lengths of light. These wave-lengths are however for practical convenience 
given as the numbers of such undulations in an inch British; and go on 
increasing from 30,000 at the red, to 65,000 at the violet, end of the spectrum ; 
between which limits all our eye-observable phenomena will be found included, 
though photography takes account of many more. 
Examination of the observed quantities and elimination of “ impurities,” 
(Tables of 20 gas-vacuum tubes, spectroscopically examined. For these 
see Appendix I.) 
After the reader has examined each of the 20 gaseous spectra, as set forth in 
the Appendix above alluded to, the question may very properly be asked, how 
have they come out as to previous expectation, and to the promises made to 
inquire into the sufficiency of the makers’ labels, for positive information 
touching the physical and chemical contents of each of the tubes ? 
The answer is, unfortunately, that a very little contrasting of one tabular 
statement with another will show superficial, perhaps, but most unexpected 
and annoying contradictions. We may have admired in one spectrum a 
certain group of lines, and thought them characteristic of the particular gas said 
to be contained there, especially as that was an elemental gas ; but these tables 
show the very same set of lines in another tube, and another, and another still, 
no matter what their labels may declare for them, and whether they refer to 
elemental or compound gases. 
This however implies a difficulty already well known in spectroscopy :—viz. 
that the chemists have uot been able to prepare their products in sufficient 
purity, to resist the tests of the spectroscopist, But as they have done as well 
as they can, we, the observers, must now endeavour to eliminate the effects 
of practically unavoidable impurities by some virtual process ; and the first 
steps in this proceeding are fortunately very easy. 
Here for instance are four bright lines of Hydrogen, as measured by the 
late M. ANGSTROM, j 
Red with place at 38,707, 
Glaucous ,, 52.250, 
Violet > 99,025, and 
Lavender _,, 61,932, 
