Aug. 2^, 1872J 



NATURE 



359 



by Angstrom,* who found that when the spark passing through 

 a hydrogen vacuum giving the baml spectrum, and at the same 

 time, the three well. known lines (as is generally the case), is 

 examined in a revolving mirror, two different images are obtained 

 corresponding to the two spectra. Two different discharges 

 must therefore take place in the tube, one of very short dura- 

 tion, which corresponds to the line-spectrum, and one which lasts 

 much longer, and corresponds to the band spectrum. As this 

 spectrum is, moreover, according to him, identical with the 

 Pljectrum of acetylene obtained by Berthelot, Angstrom thinks 

 himself justified in ascribing this band spectrum to acetylene. 



My own experiments lead to the confirmation of Angstrom's 

 supposition that the second spectrum of hydrogen is due to hydro- 

 carbon. Generally two distinct causes, which we shall consider 

 separately may introduce a hydrocarbon into the vacuum tube : — 



1. The gas passing through india-rubber tubes will carry with 

 ii small pieces of india-rubber. 



2. All the tubes are more or less greasy, and it is almost im- 

 possible to remove tliis greasiness entirely. 



These two causes are sufficient to produce such effec's as are 

 observed by Phicker, but the spectrum obtained in consequence 

 of these impurities will necessarily vary sd much with different 

 tubes, that the presence of impurities will soon be recognised as 

 it was by Pliicker. Such constant effects as those observed by 

 Wullner must, however, be due to a more constant cause, and 

 we liave not to look out long for such a cause. ^Yhen WiiUner 

 examined the spectrum of oxygen, he discovered two new spec- 

 tra, and even before Angstrom had expressed his doubts as to 

 the correctness of these spectra, WiiUncr had found out that 

 they were due to carbon or carbon compound, which were in- 

 troduced into his vacuum tubes by the greased stop-cocks by 

 which they were shut. Thus by the results obtained with oxy- 

 gen it is proved that carbon compounds were introduced into 

 his vacuum tubes, and we have only to consider whether these 

 carbon compounds, which were sufficient to change entirely the 

 spectrum of oxygen, are sufficient to make themselves percep- 

 tible in hydrogen. But Berthelot has shown that benzole mixed 

 with a large quantity of hydrogen gives a spectrum of bands, and 

 that acetylene when mixed with a sufficient quantity of hydrogen 

 is not decomposed by the electric current. 



It seems at first sight improbable that a gas passing through 

 an india-rubber tube should carry with it a sufficient quantity of 

 solid particles to change the appearance of the spectrum ; but 

 Tyndall.t in his experiments on actinic clouds, mentions the 

 effect produced by an india-rubber joint through which the gas 

 subjected to examination had passed. The quantity of matter 

 carried away by a current of air passing through an india-rubber 

 tubing is not so small as might at first sight appear. The follow- 

 ing experiment shows this. Take a piece of such tubing, and fix 

 a small piece of glass tube to one of its ends. If the air is now 

 sucked in through the glass tube the taste of india-rubber will 

 make itself at once perceptible, although the mouth is not in 

 direct contact with it. When I examined the spectrum of hy- 

 drogen, which had thus passed through au india-rubber tubing, 

 the spectrum did not vary sensibly under difi'erent pressures. 

 The colour of the spark was whitish, even under a pressure of 

 2 mm. The spectrum, evidently the band spectrum, 

 described by Wullner, did never show itself so well developed as 

 is described by him, and resembled much more the spectrum as 

 oliserved by Pliicker. In different vacuum tubes the spectrum 

 had the same appearance ; and I concluded therefore that the 

 cause of impurity must lie in the tubing, which was therefore 

 removed. 



The vacuum tube was now fitted into the receiver containing 

 the hydrogen, and sealed off at one end as soon as filled with the 

 gas. This way of proceeding, however, presented many prac- 

 tical dilficulties, and generally, therefore, another way was adop- 

 tcil. A drop of water was introduced in the vacuum tube, and 

 after the vacuum had been made, the water was boiled, and the 

 tube exhausted again, this being repeated until all the water had 

 been evaporated and the air driven out. The spark decomposed 

 the vapour, and the spectrum of hydrogen was thus obtained. 

 Its appearance now varied much with the different vacuum tubes. 

 One of them was carefully washed, first with sulphuric acid and 

 then with distilled water. The spectrum obtained in this tube 

 showed the continuous spectrum only so faintly that I think it 

 would have escaped the notice of any observer who was not 

 looking out for it. The influence on the colour of the spark was, 



* Phil Trans. 1865. 



t Tyndall, " Radiant Heat," p. 341. 



however, much greater. The spark did not show that saturated 

 red colour characteristic of hydrogen, but it was always some- 

 what whilish. This, however, is the nearest approach to the 

 pure spectrum of lines which I could obtain. It is ihus ren- 

 dered at least highly probable that organic impurities are the 

 cause of the observed spectrum of bands, and Augstiom's sup- 

 position that it is acetylene is very plausible. W^iillner examin- 

 ing the spectrum of acetylene, says that it is identical in the 

 red and yellow with the second spectrum of hydrogen, but that 

 it differs from it in the green and blue. He concludes that the 

 acetylene is decomposed into hydrogen and carbon, that the 

 hydrogen shows its yellow bands, and the carbon its green and 

 blue bands. Now these green and blue bands are not at all due 

 to carbon, but to carbonic acid, as stated by Wiillner himself, 

 and only prove that carbonic acid was present. That the acety- 

 lene was decomposed is a supposition, which is in contradiction 

 to Berthelot's results, mentioned above. 



In order to ascertain whether the last traces of air remrining 

 in the tube might not have an influence upon the spectrum of 

 hydrogen, I tried to obtain the spectrum of ammonia, for this is 

 the only compound which might possibly be formed, as aqueous 

 vapour is always decomposed. Plucker says that ammonia also 

 is decomposed into its elements, but I succeeded in obtaining its 

 spectrum by the following arrangement : — A few drops of a strong 

 solution of ammonia in water are introduced into the vacuum 

 tubes, and the induction current is allowed to pass while the 

 pump is being worked. Thus a vacuum is obtained sufficient to 

 allow the passsage of the current, and at the same time the gis 

 is constantly renewed. The spectrum of ammonia is very re- 

 markable ; while nitrogen shows a spectrum of more than 70 

 lines, and the four lines of hydrogen are distributed over the 

 whole spectrum ; a combination of these two gases when lumi- 

 nous gives otit a perfectly homogeneous light. The colour of the 

 spark is greenish yellow, and when examined by the spectroscope, 

 it shows a broad band of exactly the same colour. The red and 

 blue part of the spectrum is completely dark, while in the yellow 

 and some parts of the green a faint continuous spectrum is seen. 

 The wave-length of this band was determined by interpolation 

 to be 5,686 to 5,627 .V* metres, and the band is therefore placed 

 at the more refrangible end of the strong terrestrial absorption 

 band called by Brewster 5. The spectrum of ammonia bears a 

 strong resemblance to the spectrum of sodium, when at a high 

 temperature its lines begin to widen. 



WiUIner has discovered a third spectrum of hydrogen, which 

 is a spectrum of lines, and appears under the lowest pressure 

 which can be obtained. Not having a Sprengel's pump at my 

 disposal, I could, unfortunately, not obtain this spectrum, and I 

 merely mention here that it appears, from various remarks of 

 Pliicker, that he examined hydrogen under the precise circum- 

 stances under which this third spectrum ought to have come out. 

 Thus, he says," that hydrogen allows the induction spark to pass 

 at a lower pressure than any other gas. At another place 

 he says that when a tube filled with dry air is exhausted 

 as far as possible, the lines of Iiydrogm and silicium appear, the 

 lines of hydrogen being due to the hygroscopic condition 

 of the glass. Angstrom believes this spectrum of lines 

 to be due to sulphur. This is in harmony wiih the fact that 

 small pieces of caoutchouc (containing sulphur) form part of the 

 sources of error ; but Wullner has shown that although all the 

 lines of his spectrum of hydrogen seem to coincide with strong 

 lines of sulphur, some strong lines of sulphur do not appear, and 

 according to Wullner, the spectrum has in its general character 

 no resemblance to that of sulphur. At any rate, it seems impro- 

 bable that if this spectrum is really due to hydrogen, it should 

 have escaped the notice of Pliicker. 



If we inquire now what bearing these results have on the 

 general question of double spectra, we must remark that tJvo 

 different siilijects have been mixed together. W'e have first 

 bodies which are gases at the ordinary temperature, such as ni- 

 trogen, hydrogen, oxygen. The question, whether these bodies 

 can give different spectra under different circumstances, must, I 

 think, be answered in the negative. This was the opinion ex- 

 pressed by Angstrom from the beginning ; and although this 

 physicist clearly obtained all the results mentioned by Phicker 

 and Wullner, they seem always to have left the conviction in his 

 mind, that they are due to impurities. But there are other bodies, 

 such as iodine, sulphur, and bromine. The existence of two 

 spectra, in the case of iodine and sulphur, seems to be satisfac- 

 torily established by the researches of Mr. Salet. One of the 



* Phil. Trans. 1865. 



