EEV. T. E. EOBIE'SON ON SPECTEA OE ELECTEIC LIGHT. 
983 
looks like a row of luminous pillars, from No. 53 to No. 77.- In general the E. spectra 
are much fainter than the G.P. ; the brilliant No. 3 a, 15 y, 22 44 and the k group 
are inconspicuous, though often present. On the other hand, some are occasionally 
brilliant; thus 38 n occurs as ^ twenty-five times; 23 and 24 have more n^s in E. than 
C.P. ; and the E. spectrum, Te, N, is positively splendid with thirteen -^s. But I think 
there is no instance of a linef occurring in any E. which is not found in some C.P., or 
lice versa. These differences may be attributed to three causes. First, the discharge 
which at C.P. passes in a flash of, probably, evanescent section, must be very much brighter 
than when diffused in a tube of 0-2 inch diameter, which it fills completely. This would 
account for the lines being fainter, but not for the difference of character. In a com- 
pound tube of two pieces, 0-5 inch diameter connected by one 0-05 inch, with lead elec- 
trodes, A., E., the spectrum in the narrow part had more lines than in either of the wider, 
as being brighter; but it contained all theirs, and was quite distinct from the C.P. 
Secondly, it may be thought that the air is less heated, because of the less resistance and 
the greater heat-capacity of rarefied gas. Were this the principal cause, the E. ought 
never to contain more lines than the corresponding C.P., much less have any as a 
which is faint in the other ; the transition-spectrum, too, should be more luminous than 
the E. Thirdly, the mere increased distance of the gas molecules tna^ modify the light- 
vibrations. This is mere guess ; but so is much more of our speculation on this myste- 
rious subject, and it may at least point out the road for future research. Experiments 
at pressures greater than the atmospheric, produced by mechanical force or heat, and 
with electrodes in various states of aggregation, would probably throw light on the sub- 
ject. One thing must be remembered : in these E. discharges the light, when viewed 
in a revolving miiTor, has a certain permanence ; but in the C.P. it is, so to say, instan- 
taneous. 
3. It is generally believed that each of the bodies present in the track of a discharge 
has its own independent spectrum, coexisting with those of the others and in nowise 
interfering with them. This is not universally true, as will be seen by comparing the 
spectra of air, both C.P. and E., with those of its components N and O. Without going 
through the entire series, but taking the first twenty-five numbers of the Table, and the 
last, I find that out of 
353 lines found in 
A., C.P., there are neither in N nor 0 . 
71 
156 
5? 
A., E., ,, ,, 
62 
404 
?? 
N, C.P., there are wanting in A . . 
186 
226 
55 
N E 
XI..5 55 5, 
124 
314 
?5 
0, C.P., there are wanting in N and A. 
133 
133 
^5 
0, E., 
52 
Many of those wanting in A are found both in N and O. From these numbers it is 
t Of course I mean the lines of this Table, without including in the statement those of special character 
and peculiar origin. 
