Separation of the Least Volatile Gases of Atmospheric Air, &c. 391 



way stop-cock d, by which c could be put in communication with the 

 closed vessel, D, or with the tube e, by which also D and e could be 

 connected. The tube e passed down nearly to the bottom of the 

 vacuum jacketed vessel E, and out again through the cork ; and on to 

 a gauge /, and through a sparking tube g to a mercury pump F. 

 The stopcock n being still closed, the whole of the apparatus between 

 n and the pump, including the vessel D, was exhausted, and liquid 

 hydrogen introduced into E. The three-way cock d was then turned so 

 as to connect c with D, and close e, and then n opened. B was thereby 

 put in communication with D, which was at a still lower temperature 

 than B, and the gas dissolved in the liquid in B, along with some of 

 the most volatile part of that liquid, distilled over, and the latter 

 condensed in a solid form in D. When a small fraction of the liquid 

 in B had thus distilled, the stop-cock d was turned so as to close the 

 communication between D and c, and open that between D and e. 

 Gas from D passed into the vacuous tubes, but in so doing it had to 

 pass through the portion of e which was immersed in liquid hydrogen, 

 so that condensible matter carried forward by the stream of gas was 

 frozen out. 



For separating the least volatile part of the gases, the vessel E, with 

 its contents, was dispensed with, and the tube c made to communicate 

 directly with that connected with the gauge, sparking tube, and pump ; 

 and generally several sparking tubes were interposed between the 

 gauge and pump, so that they could be sealed off successively. The 

 bulk of the liquid in B consisted of nitrogen and oxygen. These were 

 allowed gradually to evaporate, the temperature of B being still kept 

 low so as to check the evaporation of the gases less volatile than 

 oxygen. When a great part of the nitrogen and oxygen had thus 

 been removed, the stopcock n was closed, and the tubes partially ex- 

 hausted by the pump, electric sparks passed through g, and the gases 

 examined spectroscopically. More gas was then evaporated from JB t 

 and the spectroscopic examination repeated from time to time. 



The general sequence of spectra, omitting those of nitrogen, hydro- 

 gen, and compounds of carbon, which were never entirely removed 

 by the process of distillation alone, was as follows : The spectrum of 

 argon was first noticed, and then as the distillation proceeded the 

 brightest rays, green and yellow, of krypton appeared, and then the 

 intensity of the argon spectrum waned, and it gave way to that of 

 krypton until, as predicted by Eunge, when a Leyden jar was in the 

 circuit, the capillary part of the sparking tube had a magnificent blue 

 colour, while the wide ends were bright pale yellow. Without a jar 

 the tube was nearly white in the capillary part, and yellow about the 

 poles. As the distillation proceeded, the temperature of the vessel 

 containing the residue of liquid air being allowed to rise slowly, the 

 brightest of the xenon rays began to appear, namely, the green rays 



VOL LXVIII. 2 E 



