Dec. 22, 1881 | 
NATURE 
189 
in this ultra-red region, more particularly as regards the resolu- 
tion of bands into lines. You saw there were very few lines 
apparently, but there were bands, and the question asked was, 
Could we resolve these bands into lines? You recollect that 
Draper had in his photograph three lines below the limit of the 
red end of the spectrum taken by the oxidising process. They 
did not go very far down as it turned out, but still, there they 
were, and I think I can show you that those lines and bands 
are resolvable into lines. To do that, of course, we have to use 
a diffraction grating. On that stand I have a diffraction grating 
similar to the one Mr. Lockyer showed you, which was used in 
all the researches on the spectrum. We have on the screen the 
spectra produced by the grating; you will see that even the first 
two which lie next to the bright central image of the slit are 
much feebler than the spectrum you ordinarily see on the screen, 
as Mr. Lockyer pointed out. If you turn the grating further 
round you will see that another spectrum comes on, and by 
turning it still further we get a third, and soon. They are all 
feeble, but the two last very feeble indeed, but still they are 
present ; of course by turning the grating in the other direction we 
should get similar spectra on the other side of the central image 
of the slit. By holding the screen up rather closer to the source 
of light, we shall be able to see the spectra better. I want you 
to notice that the violet of the third spectrum overlaps the red 
of the second spectrum, In order to photograph the ultra-red 
of the first spectrum it was necessary to use some artifice to cut 
out those invisible rays which lie between the violet and the 
red, and belong to the ultra-violet of the second spectrum, and 
also the violet and the blue, and the green of the same spectrum. 
In order to do that we used various absorbing media, but the 
most practicable for the purpose we had in view was a solution 
of bichromate of potash in water of about 1-25th of an inch in 
thickness. You will see that bichromate of potash cuts off the 
violet and the blue, and leaves the red and yellow intact. This 
solution was used with the diffraction spectrum to photograph 
the ultra-red regions. I will throw a diagram on the screen to 
show the overlapping of the different spectra, to make it more 
clear. You see in the second order the H lines comes a little 
beyond A (Fig. 8), and in the third line they come as far as the 
ve 
FED CB Ace 21 7: P 
Un | 
By 
Fic. 8.—Overlap of the diffraction spectra of the rst, 2nd, 3rd, and 4th orders, 
D line. You will also notice that the bichromate of potash 
cuts off certain rays in the first order-spectrum, the same rays 
of the second-order spectrum, and those also in the third-order 
spectrum, 
I will now throw on the screen some photographs taken with 
this diffraction spectrum [shown and explained]. 
That is as far as we have been able to distinguish with the 
diffraction grating up to the present time, although we have 
hopes that with more labour we shall be able to get further 
down, not to the theoretical limit of the spectrum as shown by the 
diffraction grating, since that is infinitely far down, but at all events 
towards that way. In order to show how we can plot the wave- 
lengths it is only necessary to use the same plate of green bromide 
and to expose half the slit to the second-order spectrum for the 
blue end, and the other half to the first order of the red end of the 
spectrum, using, of course, proper absorbing media. In this 
photograph we adopted this artifice. The top half of the slit 
was exposed to the red end, and the bottom half to the blue, 
and so you see two spectra superposed one above the other 
(Fig. 9). Now we know that in the second order the wave-length 
of a line will be exactly half that of the wave-length of the next 
order which is above it. Thatis to say, suppose the wave-length 
of the H line to be 3900, the ultra-red ray which lies over it would 
be exactly 7800, and so on. By these means, by the coincidences 
of these lines one with the other, one is able to ascertain the 
exact wave-length of lines which lie in the ultra-red rays of the 
spectrum, 
Then came the question, were we able to separate Draper’s 
lines into bands, and were we able to separate these bands which 
a4 
80 
Fic. 9 —Method of determining the wave-lengths in the supra red region. 
we photographed into lines ? Draper’s three lines were separated 
into 250 distinct lines, and the bands on the screen into some- 
where over 500. 
Having obtained means of photographing in the ultra-red 
region of the spectrum, what was the natural use to make of it ? 
To introduce it into the photographic art? Not so, because there 
were considerations which prevented our doing so ; but it seemed 
that there were other problems which might be settled very 
readily by recourse to another investigation. It seemed probable 
that colourless liquids ought to exercise absorption in the ultra- 
red regions. Nothing was known regarding them beyond the 
remarkable and well-known experiments made by Prof, Tyndall 
with a thermopile, with some source of radiation ata comparatively 
low temperature. He used a red-hot platinum spiral ora cube of 
hot water, and noted the radiation which was allowed to pass 
through different liquids and gases, But the knowledge obtained 
by this method was very much the same as if we were told that 
so much total visible light was cut off when examining the absorp- 
tion spectra of coloured bodies. No definite knowledge was 
obtained as to the parts of the spectrum where the absorption of 
the liquids took place ; in other words, Tyndall gave us a noti- 
fication of the absorptions, and not their locality—a most 
important point. 
Col. Festing joined with me in investigating this question, and 
we commenced, as might naturally be expected, by testing the 
absorption spectrum of water, and then we went on to a variety 
of hydrocarbons, such as the alcohol series, benzine, and so on. 
I need not recount to you all the various difficulties we found in 
our way ; they were varied, but ultimately we were able to over- 
come them. Early in our work we had glimmerings of the 
truth that subsequently burst upon us in its full and truest light. 
The method we adopted was as follows :—You may imagine a 
source of light—the positive pole of the electric light forms a very 
brilliant source when cast by a lens upon the slit of the photo- 
spectroscope(Fig.10). A tubeofliquid, 2, was placed between 
F 
Fic. 1o.—Apparatus used in photographing the absorption spectra of liquids. 
the lens and the slits : the rays were passed through prisms P, and 
eventually were received on the photographic plate F such as we 
have here, Passing sunlight through the top half of the slit, 
and then using the electric light to get the absorption spectrum 
of the liquid through the bottom half of the slit, we were able 
to compare either the solar spectrum with the absorption of the 
electric light after passing through any liquid ; or by placing two 
different liquids before the top and bottom half of the slit we 
| were able to compare their absorption spectra with each other. 
Some of the first results we obtained were with hydrocarbo % 
