ON ARGON AND ITS COMPANIONS. 
49 
table. This conclusion, as will be shown when tlie question of their atomic weights 
is discussed, is wholly untenable. 
The separation of these gases from air, and from one another, is a very trouble¬ 
some one, and demands much time and jjatience. Moreover, the tests of purity are 
not easily applied. The pocket spectroscope is, of course, a useful instrument, when 
it is required to decide as to the nature of the main constituent of a gaseous mixture. 
It is convenient, in using the instrument, to remove the lenses and the slit, and to 
examine the capillary portion of the Pliicker tnbe at a distance of 6 or 8 feet. The 
brightly illuminated capillary then itself forms a sufficiently good slit, and the 
spectrum is a very luminous one. This plan, however, fails when one gas is con¬ 
taminated by traces of others. By altering the pressure in the Pliicker tul)e, and by 
interposing a jar and a spark-gap, it is often possible to reveal the presence of a 
small proportion of one of these gases in a large amount of anotlier. But at the 
best it is only a rough guide. The final criterion of purity was always the balance. 
If two portions of gas Avere found to possess nearly the same density, they were 
fractionated and their density again compared. If the density was not altered by 
fractionation, they were regarded as homogeneous and identical. 
Apjjaratus employed in these Researches, 
It will conduce to In’evity and clearness, if we begin with an account of the 
apparatus and methods which we found advantageous in preparing the licjtiid air, 
in fractionating the gases, in determining their density, their refractivity, tlieir 
molecular volumes, their compressibilities, and in measuring their vapour-pressnres. 
In some cases, tlie methods have been em})loyed before, and where this has l)een tlie 
case, we will confine our description to any modifications vdiich we have found useful 
for the particular purpose in view, giving references to the papers previously pub¬ 
lished on the subject. 
The Liquefaction of Air* was carried out l)y means of an apparatus Imilt from the 
design of the Inventor, Dr. Hampsox, by the “ Brin ” Oxygen Company. Its 
jorinciple, which is common to Dr. Linde’s machine, consists in the expansion of air 
compressed under a pressure of 150 to 175 atmospheres, through a valve L 
regulated by the spindle M; the “ Thomson-Joule ” effect is thus utilised, and 
the expanded air passes upwards over the surface of the coils KK, through which 
it has passed down’svards in a state of compression, thereby cooling the coils pro¬ 
gressively, and absorbing more and more heat in expanding, until the temperature 
is ultimately reduced to below the temperature at which the air liquefies. The 
liquid collects in the vacuum-vessel N and can be drawn off at the stop-cock O. The 
* Up till April, 1899, we obtained our supplies of liquid air from Brin’s Oxygen Company, and we 
take this opportunity of expressing our sincere thanks to Mr. Murray and Dr. PIampson for their 
kindness in furnishing it. 
VOL. CXCVMI.-—A. 
II 
