1898.] on Liquid Air as an Analytic Agent 817 



To test the working with an unoonden sable gas added to air, a 

 volume of 10 cubic feet was taken in the gasholder, and to that 

 500 c.c. of hydrogen were added. This is in the proportion of less 

 than 1 in 500. Even after two hours' cooling, the tube B could only 

 be filled four-fifths. In order to prove that the gas accumulated in 

 the upper part of B was hydrogen, the three-way stopcock at E was 

 turned, and the temperature allowed to rise, so that tlie gas was 

 expelled from the evaporation of the liquid air and collected over 

 mercury as shown at F. The gas thus collected was easily com- 

 bustible and consisted chiefly of hydrogen. The amount of hydrogen 

 was then reduced to 1 part in 1000 of air, and it was found that 

 after one-and-a-quarter hours' cooling, the bulb B had filled to 

 within a half c.c. of the capillary tube. A new sample of air con- 

 taining 1 part of hydrogen in 10,000 of air, filled the bulb B com- 

 pletely as if it were ordinary air. 



It appears from these experiments that 1 part of hydrogen in 

 1000 of air is just detectable in the form of an uncondensable residue. 

 As the 80 litres of air coiidens -^d contained some 80 c.c. of hydrogen, 

 it appears that 100 c.c. of liquid air at from —200° to —210° C. had 

 dissolved nearly all this gas ; in fact, that 20 c.c. of hydrogen at 

 the low temperature is dissolved in 100 c.c. of liquid air, and can 

 only be detected by examining the first sample of gas boiled off or 

 extracted by lowering the pressure on the liquid. In the paper on 

 ' The Liquefaction of Air and Research at Low Temperatures,' * it 

 was shown that if hydrogen containing a small percentage of oxygen 

 were employed for the purpose of getting a hydrogen jet, the liquid 

 collected from it was oxygen, containing, however, so much hydrogen 

 dissolved in it that the gas coming off for a time was explosive. 



Coal gas, which is a mixture of hydrogen, marsh gas, carbonic oxide, 

 and various illuminating gases and impurities, after passing through 

 a coil of pipe surrounded with solid carbonic acid for the purpose of 

 condensing the vapours of benzol, naphthalene, &C.5 when supplied to a 

 tube similar to B, surrounded by boiling liquid air, gave a liquid and 

 gaseous portion at the lowest tem[ erature. It was possible to con- 

 dense in this way all the constituents of coal gas, and to separate 

 them after liquefaction by fractional distillation, except carbonic 

 oxide and hydrogen. 



Ultimately, however, the carbonic oxide would be condensed, and 

 hydrogen be left alone in the gaseous state. Similarly, any gas less 

 easily condensed than air could be separated from a mixture of the 

 same with air. Hydrogen present in air to the extent of one in a 

 thousand is just detectable, but smaller quantities escape direct obser- 

 vation owing to solution in the liquid. In order to press this inquiry 

 a little further, some natural gas known to contain a different con- 

 stituent, like helium, suggested itself as being worthy of trial. Lord 

 Rayleigh's analysis of the gas from the King's Well, at Bath, gave 



* Proc, 1895, vol. xi. p. 221. 



