190 
containing oxygen. Alcohol and ether are both hydro- 
carbons—they contain hydrogen, carbon, but also oxygen ; 
and we noticed that in these oxygen compounds these bands 
of absorption were shaded bands, and net sharp and 
defined, as it were. We then went on to another series of com- 
pounds, or rather, part of the same group which centained no 
oxygen atall, Thus we worked with methyl iodide, ethyl iodide, 
and propyl iodide, and we found a very marked difference 
between the two spectra. We found that in the cases where 
there was no oxygen there were no shaded bands—that is to 
say, that if there were bands, they were sharp bands without 
shading at the sides (Fig. 11), What was the signification of this? 
i100 1200 
UT 
i LiL tl Avian 
Meee eek ae) 
<=" ebb 
TTT 
Mmm Th 
Fic. 11.—Absorpuon spectra of hydrocarbons, &c. 
There must have been some meaning init. Were the lines or 
the shaded bands due to carbon or to hydrogen? The lines 
could not be due to oxygen, because they were not present when 
oxygen was present. To what, then, could they be due? They 
_ must be due to carbon, hydrogen, or iodine ; and which of these 
it became important toascertain. In Fig. 11 we haveamap of a 
selection of the different alcohols and the iodides, which were 
photographed. You will see that what I have said about alcohol 
is correct ; that we have the shading off of the bands in the case 
of the alcohols. But when we come to the iodides we have a 
marked difference; we have lines springing up, as in ethyl 
iodide—distinct lines—which are also found in the other iodides. 
The question then was to trace these lines to their origin. If 
they were due to carbon we ought, of course, only to find them 
in carbon compounds ; if they were due to hydrogen we ought 
only to find them where there was hydrogen. So we tried a 
series of substances, the absorptions of which I throw on the 
screen. When we tried chloroform, which contains only one 
atom of carbon and one of hydrogen, but three of chlorine, the 
whole spectrum became one of lines, nothing else but lines. 
These lines might be due tothe carbon and hydrogen, or they 
might be due to the chlorine; so the next substance we tried 
was hydrochloric acid, which contains only hydrogen and 
chlorine. Here we have lines again which were coincident 
with some of the lines in the chloroform: those lines might 
still be due to hydrogen or chlorine, The next substance we 
tried was ammonia, which contains no chlorine, but three 
atoms of hydrogen and one of nitrogen. That gave lines again. 
We next tried carbon disulphide and carbon tetrachloride 
which contain neither hydrogen nor oxygen, with the result that 
we had neither bands nor lines. Evidently we had tracked the 
lines to their source ; they were due to the oscillation of hydrogen 
in these particular compounds we had examined. When we took 
sulphuric acid we found the same result again; the bands were 
rather shaded, and to a certain extent it was the same in water 
also. The oxygen, as we shall see, formed these bands, but at 
the same time at the limit of the bands a distinct line was formed. 
Thus then we found in all the absorption spectra which contained 
lines, that those lines were due to hydrogen and nothing else. 
I should next like to show you the further information we 
gained from these photographs, In the diagram we have the 
NATURE : 
[Dec. 22, 1881 
alcohol absorption spectrum, with a chloroform spectrum he- 
neath it. The question to be answered is—Why should we 
have bands with sharp edges and fine lines in one case, and 
bands with sharp edges and bands shaded off in the other? In 
both cases the bands with sharp edges seem to be due to the base 
of the compound. Thus, in the case of the chloroform, the 
thick line or sharp-edged band seems to be due to the combina- 
tion between carbon and hydrogen, and those other lines seem to 
be due to the vibration of hydrogen and nothing elre. 
What was the meaning however of the shaded bands as (say) 
in alcohol? When we came to the photographs it was found 
without exception that hydrocarbons containing oxygen, when 
not contained in the radical or base of the compound, 
always gave some shaded bands, and on measurement it was 
found that the shading always stopped at points where, in 
other spectra, we had marked the hydrogen lines. This coin- 
cidence was very remarkable, and could not be fortuitous; in 
fact, it seemed that there must be some connection between the 
position of these lines and the termination of the bands, The 
bands must be due to the oxygen in the compounds. What we 
eventually arrived at at last was this, that the oxygen blotted out 
the spectrum between two hydrogen lines; that is to say, if you look 
at it in one way, the oxygen oscillated between two hydrogen lines 
and cut out that particular portion of the spectrum. When we 
came to the benzine series, or in fact any other series, we found 
the same hold good: where we had hydrogen and no oxygen 
we had lines; where we had oxygen with the hydrogen we 
had bands. Where we had carbon, hydrogen, and oxygen, 
you see we had a shaded band and few lines; where we 
had carbon, hydrogen, and chlorine, or carbon, hydrogen, and 
bromine, or carbon, hydrogen, and iodine, or carbon and hydrogen 
alone, we had sharp-edged bands and many lines. Where we 
had carbon and chlorine, or carbon and nitrogen, or carbon 
and sulphur, we had no absorption whatever. That is to say, 
if you place bisulphide of carbon or cyanogen before the slit in 
one of these tubes we shall see no absorption take place except 
a general absorption. In other words, the absorption gradually 
increased and mounted up from the least refrangible end of the 
spectrum towards the blue end of the spectrum, and that was 
the only absorption which could be traced. The following table 
concisely shows the above :— 
C 
Carbon, Hydrogen, and Oxygen gir ae Speci DEeaTe 
Carbon, Hydrogen, and Chlorine 
Carbon, Hydrogen, and Bromine 
Carbon, Hydrogen, and Iodine... 
Carbon and Hydrogen is 
Carbon and Chlorine ... 
Carbon and Nitrogen ... 
Carbon and Sulphur ... 
Hydrogen and Oxygen 
Hydrogen and Chlorine 
Hydrogen and Nitrogen 
Sharp-edged bands and many 
lines. 
1. $ No special absorption. 
Bands and lines. 
Lines only. 
Lines only. 
The character of absorption then was general and special. 
Where we had special absorption bands they were due 
primarily to hydrogen atoms vibrating ; whilst the general ab- 
sorption was due to the molecules ; the heavier the molecule 
the more it let the ultra-red through the spectrum, If bases 
of two series were present we found the absorption due to each 
base of these particular compounds present in the spectrum. 
Thus, by taking the absorption spectrum of, say, a compound 
of ethyl and benzine, we are able to say that the ethyl base was 
there and the benzine base was there also ; the bases of these 
particular series being denoted by these thicker bands with 
sharp edges to which I have already referred. 
Thus then spectrum analysis opened a way for the chemical 
analysis of these organic compounds, not of course in their 
entirety, but so as to get a qualitative idea of what they may 
contain. 
The length of liquid generally employed was six inches, and the 
natural question to ask is, What is the difference caused by in- 
creasing or diminishing the thickness of the liquid between the 
slit and the source of light? If you increase the thickness of the 
liquid between the slit and the source of light, it is this: Where you 
have oxygen bands you have them shaded off; increasing in in- 
tensity, but spreading further—they do not spread further in a 
nondescript way, but in a very marked manner they spread out 
to the next hydrogen line, and so on. Therefore, supposing 
with a very small thickness we have oxygen bands extending 
between two hydrogen lines, in the next six inches of liquid the 
