282 



NA TURE 



[Janlakv 19, 1899 



and each other. One of these two identical tubes could then 

 be subjected to the hydrogen cooling, following the directions 

 already given, and the two vacuum tubes now compared. If 

 there was a marked difference in resistance to the passage of 

 the discharge in the frozen tube, then something must have 

 condensed, and by a few tentative trials a limit might be 

 reached when the initial exhaustion was unaffected by the 

 hydrogen cooling. Such experiments have not yet been made. 

 The presence of any vapour of mercury would require to be 

 carefully eliniin.ited, otherwise the method would not be satis- 

 factory. Tubes that are prepared without taking special pre- 

 cautions to exclude organic matter and water from the glass, 

 deteriorate, especially with electrodeless tubes after the dis- 

 charge has taken place for some time. 



The rapidity with which the vacua are attained is such as 

 theory would suggest, assuming a hole of a square millimetre in 

 section through which the air rushes into the condenser and 

 that a velocity of current between 600 and 700 feet a second is 

 attained, then a vessel of 20 c. c. capacity could be reduced in 

 pressure in l second to l/io of the initial pressure, and if the 

 same rate is continued at the end of 60 seconds to d's)*. Sir 

 George Stokes has been good enough to consider the problem 

 and writes as follows : — 



"Let \ be the volume of the vessel, A the area of an 

 aperture by which the air is conceived as rushing out with the 

 velocity :•, p the density of the air in the vessel at the time /, 

 D the initial density, that is, the atmospheric density. 



"Then, according to our hypothesis, Kv.dl is the volume of 

 air, and, therefore, \vp.dl, the mass of air, which rushes out in 

 the time dt. But this equals the loss of mass of air in the 

 vessel during the time dl, and, therefore, 



Avp .dt= - \de, 



a differential equation of which the integral is 



p = Dt— ^"'*. 



"Suppose now V to be 20 c.c, or 20,000 c.mm., .\ to be the 

 area of a circle of I or 2 mm. diameter, say 2 sq. mm. , v to be 

 333 m., or 333,000 mm., / (in seconds) to be 60 ; then 



, p ^2x333,000x60 ^ g^ 

 p 20,000 



? = 5254 X lo-"*. 

 P 

 " This would give a density of almost inconceivable sniall- 

 ness. Doubtless the supposition made above as to the rate of 

 discharge is very wide of the mark, being much too great. If 

 the velocity of rushing is about half the velocity of sound, the 

 ratio _of densities would become 72 x 10'-'". If so it is satis- 

 factory to find that the mathematical following out of the 

 hypothesis leads to a density of the residual air in the vessel 

 which is enormously below what suffices to account for the 

 observed result." .\ practical mode of rapidly attaining a high 

 vacuum in any vessel is to displace the air with carbonic or 

 .sulphurous acid, either at the atmospheric or under diminished 

 pressure, and then to freeze out the remaining gas by the use 

 of liquid air, just as in the experiments with liquid hydrogen. 



The first vacuum lube was an electrodeless one, the air had 

 not been dried, nor the elass specially cleaned. On spectro- 

 scopic examination it showed hydrogen lines bright along with 

 the seconil or compound line spectrum of the same gas, and a 

 series of bright bands defined on the less refrangible side, 

 diffuse on the more refrangible, which occur in the yellow, 

 green, blue, and indigo. These bands were found to be 

 identical with the carbonic oxide spectrum. With a Leyden 

 jar in the secondary circuit the line spectrum of hydrogen dis- 

 appeared, leaving the second .spectrum fainter : but the carbonic 

 oxide bands remained bright, and there was no appearance of 

 the hydrocarbon spectrum. The second tube had aluminium 

 electrodes, and, like the last, had no special treatment in filling 

 in the air. This tube showed also the line spectrum and the 

 second spectrum of hydrogen : the latter being bright along 

 with the carbonic oxide spectrum ; but on sparking the latter 

 <lisappeared. No appearance of the hydrocarbon spectrum could 

 be detected, but there was a suspicion of bands in the indigo 

 like the negative pole spectrum of nitrogen. The addition of a 

 J^eyden jar brought out nothing new, only intensifying the line 



NO. 1525, VOL. 59] 



spectrum of hydrogen, while leaving the second spectrum bright. 

 In neither of the above tubes could any lines of nitrogen or 

 oxygen be recognised. The third tube was filled with air drawn 

 over cotton wool, red-hot copper oxide, and phosphoric pent- 

 oxide, no rubber joints being employed. The spectrum showed 

 the carbonic oxide bands and the hydrogen line spectrum as 

 before. Only the second hydrogen spectrum was feeble. There ' 

 was a yellow line W.L. 5849, identical with one occurring in the 

 natural gas from the King's Well at Bath. In a paper on " The 

 Liquefaction of .Air and the Detection of Impurities" (Chtui. 

 Soc. Proi., November 1S97), the separation of helium from 

 this gas is described by liquefaction and fractionation, and it was 

 observed that during the sparking the helium lines were well 

 marked along with " otlurs, the origin of wliiclt must be settled 

 later." It was further observed, " fVith a modified form of 

 apparatus it will he possible to collect any residuary gas from 

 the use not of 3 cubic feet of air or Bath gas, but from hundreds 

 of cubic feel of su-rli products." The helium and other associated 

 material was shown to be more volatile than nitrogen. Pursuing 

 this course of investigation in the summer of this year, the volatile 

 portion of air was examined, when the presence of material 

 giving the same lines as Bath helium was recognised. While 

 this investigation was in progress, Prof Kamsay and Dr. Travers 

 observed the same spectrum in the more volatile portion of argon 

 which they have associated with a new element called neon. 

 The use ol liquid hydrogen, as described, proves that the most 

 characteristic line of neon in the yellow, about W.L. 5849, can 

 be detected in 25 c.c. of ordinary air, and the presence of helium 

 in the atmosphere is confirmed.' 



A fourth tube, filled like the preceding one, had a phosphoric 

 pentoxide tube left on. This showed again the carbonic o.xide 

 bands, but no hydrogen lines could be detected ; while the 

 oxide of copper ought to have removed all free hydrogen and 

 transformed all the organic matter into carbonic acid and water. 

 Yet it appears that the spectrum of the carbon compounds is diffi- 

 cult to remove from electrodeless tubes, probably owing to car- 

 bonic acid coming from the glass. There were some broad 

 diffuse bands that may arise from the drying agent. The absence 

 of hydrogen in this tube suggests that its presence in the third 

 tube was due to vapour of water coming slowly from the glass. ^ 

 I am greatly indebted to Prof. Liveing for making a careful ' 

 examination of the spectra of these tubes. 



Sir William Crookes was good enough to prepare two lubes 

 with platinum electrodes, which he sparked in vacua till all 

 hydrogen disappeared, and then filled with dry air, but without 

 the use of red-hot copper oxide or any agent for the absorption 

 of carbonic acid or the destruction of organic matter. .-Vfter the 

 cooling with liquid hydrogen, he found on spectroscopic exam- 

 ination, in one no hydrogen, but two faint lines, one about 

 5852 W. L. and the other 5676 W. L. The second tube showed 

 the same yellow about 5852, the helium line along with 5939 

 and 6145, the hydrogen lines C and F, and some red lines. 

 The observations of Crookes confirm the presence of neon, 

 helium, and hydrogen. The absence in his tubes of the carbonic 

 oxide spectrum is important, seeing all the electrodeless tubes 

 gave this spectrum. In these tubes the vacuum was very high, 

 and it was difficult to observe the gaseous spectrum. Still, the 

 fact of finding hydrogen in one and not in the other, leaves the 

 presence of free hydrogen in the atmosphere as a question for 

 further inquiry. The tube that did not contain hydrogen was 

 heated very hot in ortler to get a discharge, and then the 

 spectrum showed some bands like the negative glow of nitrogen. 

 Occasionally, a jar discharge was got to pass, and when this 

 took place the nitrogen lines could be seen. .-Xn electrodeless 

 tube filled carefully with oxygen made from fused chlorate of 

 potash, which was contained in an extension of the vacuum tube, 

 gave nothing but the carbonic oxide bands. In future experi- 

 ments it will be easy to concentrate all the least volatile material 

 in air or other g.ases, and ,thereby to make a more thorough 

 examination of the spectrum. In the meantime my object is to 

 show one of the scientific uses of liquid hydrogen. 



I have to thank Mr. Robert Lennox for efficient aid in the 

 conduct of the difficult experiments. Mr. Heath has also helped 

 in the work. 



' In a p.apcr .Uong with Prof. Liveing. " On ihi- Spcclrum of the Electric 

 Disch.-vrgc in l.iijnid Oxygen, Air, alid Nilrogcn " {J'/iit. .\fag., 1894). we 

 noted that durin>; the distill.-ition and conccnlratiun in X'oene of liquid 

 oxygen .-ind air under diminished prcvurc, thai iwn liright lines appcircd in 

 tbc spectrum at wave-length 557 and 555,, and th.il one of these lines was 

 very near ihc position of the auron.il line. These lines .ire nDW attributed by 

 ihesximc chemists to a new clement, cryplon. 



