EEV. T. E. EOBINSON ON SPECTEA OF ELECTEIC LIOHT. 
949 
Air 0-39 
Nitrogen O' 32 
Oxygen 0*39 
Hydrogen 0*66 
Carbonic oxide 0*37 
Of this, two explanations may be given. It may be said that these lines are not seen 
at C.P. because they are overpowered by the brightness of the ground on which they are 
seen. In some cases (of which examples are given hereafter), especially in hydrogen, this 
cause does act in some degree. In that gas the centre of the C.P. spectrum was at first 
observed, where,, as I have stated, the finer lines are not easily seen with some metals ; 
and then the percentage was very high ; but at the negative boundary this bright haze 
does not interfere, and when observed there^ the proportion is almost exactly that of the 
other gases. Besides, it occurs frequently that in the C.P. spectrum, on each side of the 
place where the missing line should be, faint and narrow lines are seen with perfect 
distinctness. The other solution is this, that these lines do not depend entirely on the 
chemical character of the media and electrodes, but also on their molecular condition. 
On any other supposition it seems hard to conceive the passage from the C.P. to the 
transition spectrum, and the increase of brightness from that to the E. one, the same 
chemical elements being present in the three. 
In presenting the measures of these spectra, there is a difficulty arising from the 
impossibility of giving, within any practicable limits, the distinctive characters of the 
phenomena, though they are of considerable importance. For instance, the orange 
band y, which in the sodium spectrum is “ intensely bright,” is in Pt, CO “ faint, 
scarcely visible.” To tabulate them I must restrict myself to a few distinctive symbols. 
First, the lines which are far transcendent in brilliancy, and are not less broad than the 
image of the slit, I denote by a *. This implies merely that they surpass the others 
greatly ; one in red or violet may fully deserve this symbol, though it would seem dull 
beside I or 0. Not very frequently the same line has this mark in C.P. and E, but, 
except with sodium and potassium, less bright ; and sometimes one which is faint in C.P. 
is a * in E. Then follow 
Very bright, bright vb, b 
Conspicuous (from surrounding faintness rather than intense brightness) . c 
Faint, very faint f, vf 
Narrower than slit n 
Very narrow, like a hair vn 
Wider than slit (the exponent expressing how many times as wide) . . w“ 
When they are on an obscure ground o 
The symbol E implies that the pressure = 0'2 inch unless a different one is stated. 
I shall commence with aluminium, which I selected as the type because, from its small 
dispersion at the negative electrode, it may be assumed that its influence on the spectrum 
bears a small proportion to that of the gas. 
6 0 2 
