Professors Liveing and Dewar , On the most volatile gases . 107 
On the most volatile gases of the atmosphere. By Professors 
G. D. Liveing and J. Dewar. 
\Read 18 February 1901.] 
Atmospheric air, at ordinary pressure, was liquefied directly 
by contact with the walls of a vessel cooled below — 200° C. ; 
when about 200 cc. of liquid had condensed, communication with 
the atmosphere was closed, and a fraction of the liquid, still kept 
at — 210° C., allowed to distil into a second, still colder, vessel 
immersed in liquid hydrogen. When about 10 cc. had collected 
in the second vessel, in the solid state, communication between 
the two vessels was closed, and the gas above the solid in the 
second vessel was found to have a pressure of 10 to 15 mm. of 
mercury. Some of this gas was pumped out, and 43 per cent, of 
it was found to be hydrogen. Also tubes previously exhausted 
and sparked to remove hydrogen from the electrodes, and then 
filled with gases from the liquid air, at atmospheric pressure, 
gave the spectrum of hydrogen very strongly. Calculated as a 
percentage by volume of the atmosphere, hydrogen is present in 
very small quantity, nevertheless it forms a sensible percentage, 
and accords with the supposition that there is an interchange of 
gases by diffusion between our atmosphere and interplanetary 
space. 
In other experiments the gas above the solid in the second 
vessel was allowed to pass through a U-tube cooled in liquid 
hydrogen into tubes previously exhausted ; and the spectra given 
by electric discharges through them examined. These spectra 
are brilliant with the red, orange and yellow rays of helium and 
neon, but shew besides a vast number of other rays, belonging to 
substances hitherto unknown, which are most brilliant at the violet 
end, so that they can be photographed, notwithstanding the 
opacity of the glass tubes, up to a wave-length 3142. Their 
brilliance depends on the character of the discharge, and is 
greatest about the negative pole, and with no leyden-jar in 
circuit. 
These spectra were searched for the characteristic nebular, 
coronal, and auroral rays. Tubes filled as above described shew 
no ray at about the wave-length 5007, the brighter green ray 
of nebulae, though they give a weak ray near, but not in the 
exact position of the other green nebular ray, X 4959, and 
a strong ray close to the strongest ultra-violet nebular ray X 3727. 
A tube which had been filled with gas from the second vessel, 
without first passing through the U-tube cooled in liquid 
