CADY AND MCFARLAND.| Composition of Natural Gas. 263 
is very powerful and quite satisfactory. Readings can be taken 
rapidly and the.duplicates agree closely. 
The new form of end-on tubes made by Goetze (see H, fig. 6), 
when filled with gases of the argon group, give beautifully 
sharp spectra. 
The helium obtained as above practically always shows hy- 
drogen lines and carbon bands, the latter, most likely, from a 
trace of carbon monoxid, but these two impurities are present 
in such small quantities that the contraction produced on ex- 
ploding a mixture of the helium with pure oxygen and electro- 
lytic gas is too small to measure. During the analysis of the 
gas no neon was detected, but later, during the preparation of 
larger quantities of helium from some of the gases, neon was 
found. The quantity was not determined, but it is certainly 
small compared to that of helium in the same gas. Argon was 
also found, as will be described. 
ISOLATION OF NEON. 
The method for the isolation of this gas is briefly as fol- 
lows: The natural gas is largely condensed in a bulb sur- 
rounded by liquid air. The uncondensed portion is then passed 
into bulbs filled with cocoanut charcoal and cooled with liquid 
air. As has already been said in the above description of the 
separation of helium, all the gases in the mixture except hy- 
drogen, neon and helium are completely absorbed. Of these 
three gases hydrogen is rather freely absorbed, neon somewhat, 
and helium scarcely at all. In working with small quantities of 
gas, such.as is done in an ordinary analysis for helium, the neon 
is very largely taken up by the charcoal and consequently does 
not show in the spectral tube. With larger amounts of gas the 
charcoal finally becomes saturated with neon, and then its 
lines are plainly visible in the spectrum of the residue. The 
difference in the absorption of helium and of neon is great 
enough so that neon spectroscopically free from helium may be 
prepared as follows: A charcoal bulb is heated to about 400° 
C. and exhausted as completely as possible, using the mercury- 
pump. It is then cooled with liquid air and the unabsorbed 
gases from another charcoal bulb are passed in, allowed to 
stand for a time, and removed with a mercury-pump. After a 
sufficient amount has been passed through the bulb, the helium 
is removed by pumping until a cathode-ray vacuum is obtained. 
The bulb is then warmed slightly and some gas removed. After 
this, by warming the bulb more strongly, neon spectroscopic- 
