W. Crookes—Spectrum of Helium. 303 
siderable purity as far as known contamination is concerned. 
These samples of gas were sealed in tubes of various kinds 
and exhausted to the most luminous point for spectrum obser- 
vations. In most cases no internal electrodes were used, but 
the rarefied gas was illuminated solely by induction, metallic 
terminals being attached to the outside of the tube.* For 
photographic purposes, a quartz window was attached to the 
end of the tube, so that the spectrum of the gas could be taken 
“end on.” 
My examinations have chiefly been made on five samples of 
as. | 
. 1. A sample from Professor Ramsay in March last. Pre- 
pared from cleveite. 
2. A sample from Professor Ramsay in May last. Prepared 
from a specimen of uraninite sent to him by Dr. W. F. Hille- 
brand. Gas obtained by means of sulphuric acid; purified by 
sparking. 
3. A sample from Professor Ramsay in June last. Prepared 
from broggerite. , 
4, A sample from Professor Lockyer in July last. Prepared 
by a process of fractional distillation from a sample of brég- 
gerite sent to him by Professor Brogger. 
5. A sample of gas from Professor Ramsay “helium puris- 
simum.” This was obtained from mixed sources, and had 
been purified to the highest possible point. 
In the following table the first four samples of gas will 
be called: 1. “cléveite, R.”; 2. “uraninite, R.”; 3. “ brég- 
gerite, R.”; and 4. “bréggerite, L.” Only the strongest of 
the lines, and those about which I have no doubt, are given. 
The wave-lengths are on Rowland’s scale. 
The photographs were taken on plates bent to the proper 
curvature for bringing the whole spectrum in accurate focus 
at the same time. The spectrum given by a spark between an 
alloy of equal atoms of mercury, cadmium, zine, and tin, was 
photographed at the same time on the plate, partially overlap- 
ping the helium spectrum; suitable lines of these metals were 
used as standards. The measurements were taken by means of 
a special micrometer reading approximately to the z7,!jpath 
inch, and with accuracy to the ;71,y th of an inch. The cal- 
culations were performed according to Sir George Stokes’s 
formula, supplemented by an additional formula kindly sup- 
plied by Sir George Stokes, giving a correction to be applied 
to the approximate wave-lengths given by the first formula, 
and greatly increasing the accuracy of the results. 
* Journal of the Institution of Electrical Engineers, Part 91, vol. xx, Inaugural 
Address by the President, William Crookes, F.R.S., Jan. 15th, 1891. 
