ON THE ATOMIC GROUPING IN ORGANIC BODIES. 
889 
capable of swinging at a considerable angle to the axis of the lens—a point of 
importance when it is remembered what a rapid alteration there is in the focus of 
the rays, as they descend in the infra-red of the spectrum. 
Reason for employing glass prisms. —From the nature of the research it was 
necessary to employ tubes to hold the various liquids, and it would have been impos¬ 
sible to have closed their ends with any material but glass owing to the solvent 
nature of some of them. From previous experience we knew that the glass used 
transmitted radiations to a wave-length of at least X 20,000 ; and it would have 
been utterly impossible by prismatic analysis to have distinguished any except very 
general absorptions even as low down in the spectrum as that wave-length. For these 
reasons we determined to employ glass prisms. Our results show that in the 
majority of cases no advantage would have been gained by using rock salt; though 
for thin layers of vapour we can well imagine that rays of still lower refrangibility 
would have to be studied. 
Number of prisms employed. —Three prisms were invariably employed when 
photographs were taken from which measurements had to be made. With one 
or two prisms the dispersion was insufficient to enable the details of the absorption 
spectra to be accurately observed ; though the general character of the absorption was 
always clearly marked. With five prisms, on the other hand, the absorptions in some 
cases became too undefined ; we therefore concluded that three prisms would be the 
best number with which to work. In nearly every case, however, photographs of the 
absorptions were taken with one or two prisms to give a preliminary idea of what we 
might expect to find with the greater dispersion. It is our intention to pursue the 
investigation with a diffraction grating, more especially to map the line spectra 
which have shown themselves. 
Measurement of the photographs. —The measurement of the photographs was care¬ 
fully made by means of a transparent scale applied to the film-side of the negatives, 
reading to ^ a millimetre by division, and to half that quantity or less by estimation. 
We attempted to take the measures with a micrometer, but the nature of some of the 
absorptions precluded its use, since much magnification more or less shrouded the 
phenomena; with the solar spectrum where the absorptions are in more definite 
lines as opposed to bands, measurements with a high magnifying power are com¬ 
paratively easy. The accuracy of our plan of measurement might be doubted, but it 
is really trustworthy after a little experience. We should not have trusted to it, how¬ 
ever, without referring some of the principal spectra to a comparison with the solar 
spectrum. The wave-lengths of the Fraunhofer lines in the infra-red region were 
taken from the map furnished by one of us with the paper already referred to. A 
test of the accuracy of the measurements was the fact that the same wave-lengths 
of the absorption of any particular substance could always be obtained on several 
photographs. 
We invariably compared all absortion spectra with that of ethyl iodide, which 
MDCCCLXXXI. 5 Y 
