September 5, 1895] 



NA TURE 



449 



(Stage 2.) The hydrogen lines will continue to thin out, and 

 when the absorption of the hotter lower layers makes itself felt 

 the spectra will show the high temperature spectra of the various 

 chemical elements, showing many more lines. The difference 

 hetwecn these and the lines seen in stars of increasing tempera- 

 ture should be one due to the different percentage composition 

 of the absorbing layers, so far as the known lines are 

 concerned. 



With this increasing line absor])tion there will be a recurrence 

 '■f the continuous absorption in the ultra-violet. 



(Stage 3.) With the further thinning of the hydrogen lines 

 and reduction of temperature of the atmosphere, the absorption 

 lUuings of the compounds of carbon should come in. 



So much, then, for what we should exix:ct, assuming the 

 liypothesis to be true. 



I now proceed to show how far these requirements are satisfied 

 liy the mass of new facts now at our disposal. 



TiiK Ai:tuai. I'liEN'OMEXA Rf.corded on the 

 Photographs. 



Nehuhe. 



The photograjjhs of the spectrum o( the Orion Nebula show- 

 lines at wave-lengths which approximate very closely to the lines 

 I'f hydrogen, to (lutings which appear in the spectra of coui- 

 povmds of carbon, to a fluting of magnesium at 5006, and to the 

 longest flame lines of iron, calcium, and magnesium. 



The chromospheric line designated Dj has been recorded in 

 the visual spectrum of the Orion Nebula by Dr. Copeland,' and 

 the observation has since been confirmed by Mr. Taylor. - 



The line which is always associated with D3 in the spectrum 

 of the chromosphere, viz. that at A. 4471 (Lorenzoni's f), is 

 also shown in the photograph of the spectrum of the Orion 

 Nebula. 



The reiiuirements of the hypothesis with regard to nebuhi; 

 are therefore met in every point so far considered by the new 

 facts. 



Dividing up tlie lines into the three groups of origins sug- 

 gested, we have in the case of the Orion Nebula : — • 



(<j) Spectrum of large interspace (= that of non-condensable 

 gases driven out of the meteorites) = lines of hydrogen ; 

 tlutings of carbon. 



(b) Spectrum of vapours produced by the large number of 

 partial collisions = fluting of magnesium at A 5006 ; low tem- 

 perature lines of iron, calcium, and magnesium.'' 



(( ) Si)ectrum of the vapours produced at a very high tem- 

 perature by the relatively small number of end-on collisions. 

 The solar chromosphere may be taken as indicating the spectrum 

 associated .with this very high temperature = chromospheric 

 lines, D3 -I- X4471."' 



Bright-I.tne Stars, 



Prof. Pickering has shown that the Draper Memorial Photo- 

 graphs (copies of which he has very kindly forwarded uk') 

 l>rove that bright-line stars are intimately connected with 

 ihe planetary nebula;, the lines in the spectra being almost 

 iilentical. 



The main point of difference is that the chief nebular line near 

 A 5006 is not seen in the spectrum of bright-line stars, and this 

 no doubt is due to the relative absence of feeble collisions as 

 condensation goes on. The brightening of this line in the spectra 

 uf Nova Cygni and Nova .\uriga;, as the stars faded away, is 

 sufficient cvicience that it is associated «itb low temperature, and 

 lience it is not surprising to find that it is absent from the spectra 

 of the bright-line stars, which on this hypothesis are hotter than 

 ilie nebuhe, since they are more condensed. 



I have stated that we should expect the hydrogen lines to be 



^ yfonthty ^/aliccs, vol. xlviii., p. 360. 



- /A/(/., vol. -xlix. p. 124. 



3 1 have previously given evidence deduced from eye observ.itions, 

 indicating tlie presence of other low temper.-\ture flutings of manganese and 

 magnesium. 



■* Since the lecture-i were delivered (and in this I summarised a paper 

 1 had previously sent in to the Royal Society), this part of the 

 hypothesis has been enormously strengthened by the discovery of a new 

 series of g.-LSes which the spectrum indicates are jussociated with the one 

 giving the line 1)3 which I discovered in 1868 and named helium. These 

 new gases contain many lines in addition to D3 and 4471, which appear both 

 in the soL'ir chromosphere and nebula of Orion and stars of increasing tem- 

 perature. 



NO. 1349, VOL. 52] 



fainter, and the carbon flutings, and the continuous spectrum to 

 be brighter than in nebula;. 



(a) The hydrogen lines are decidedly less prominent. Indeed 

 they were not recorded at all in the eye observations of 

 7 .\rgus(.\rg.-Oeltz., 17681), of Wolf and Rayet's second and 

 third stars in Cygnus, ' but they are shown in Prof. Pickering's 

 photographs. 



(*) In my previous discussion of these bodies - I showed that 

 there was evidence of a very considerable amount of carbon 

 radiation in the visible region of the spectrum. Subsequent 

 work and an examinition of Prof. Pickering's photographs have 

 strengthened this view. 



(c) There can be no question as to the continuous spectruin 

 being brighter in bright-line stars than in nebula;. 



Slars of Increasing Temperature. 



(Stage I.) We should expect the spectra to show — 



\a) Absence of bright lines. 



{b) The presence of dark metallic flutings. 



((•) The presence of bright carbon flutings. 



{d) Continuous absorption in the violet. 



Many of the stellar ])hotographs answer these requirements. 



(a) They show no l>right lines under normal conditions, but 

 if the .stars are variable, the disturbances which bring about the 

 change of lutninosity at maximum, produce bright lines in the 

 spectrum as in the case of the spectrum of Mira Ceti photo- 

 graphed by Prof. I'ickering. 



{b) Dark flutings have been photographed in several spectra. 



((■) The photographs appear to show the actual pre.sence of 

 carbon radiation ; further photographs are being obtained to carry 

 on the inquiry.' 



The stars of this class which have already been photographed 

 at Kensington are well advanced in condensation, as indicated 

 by the numerous dark lines, and all the flutings, both bright and 

 dark, are confined to the region less refrangible than G. We 

 should therefore not expect to get the more refrangible carbon 

 flutings. It is among the least condensed stars that we should 

 expect the bright carbon to be more manifest, and, indeed, in 

 the spectrum of Mira Ceti photographed by Prof. Pickering, 

 there is strong evidence of the presence of one of the more 

 refrangible carbon b.tnds comtnencing at \ 4215. 



(rf) The photographs fully detiionstrate that there is a very 

 considerable amount of continuous absorption in the ultra-violet 

 or violet. 



It must be added that the sequence of the spectra photographed 

 resembles that deduced from eye observations, and the won- 

 derful thing is that the observations of Dunerwill bear the severe 

 test which has thus been applied to them. 



(.Stage 2.) .-Vt this stage we should expect — 



(a) Diminution in the amount of continuous absorption. 



(/') Spectrum consisting of dark metallic lines, but possibly 

 differing from the solar spectrum. 



These conditions are fulfilled by the .stars of which a T.auri 

 and 7 Cygni may be taken as types. The continuous absorption 

 is least in the latter. These spectra show numerous metallic 

 lines, but they do not exactly resemble the solar spectrum. 

 The hydrogen lines are comparatively thin, while other lines 

 have very difl'erem intensities as compared with lines in the solar 

 spectrum. 



In these stars we have to deal with the varying volatilities of 

 the meteoretic constituents of tlie SMarm, while in the case of 

 stars which are cooling we have to deal with successive com- 

 binations rendered possible by the fall of temperature in a 

 gaseous mass. Hence diflerences in the spectra are to be 

 expected. 



(.Stage 3.) The phenomena which would be expected on the 

 hypothesis, at this stage, are fully satisfied by such stars as 

 a Cygni, J3 Orionis, f Orionis, « Persei. In these stars there is 



I Roy. Soc. Proc. vol. xliv., pp. 33-43. 



•-' Ihid. 



3 Subsequent eye observations by myself and Mr. Fowler seemed to leave 

 no doubt .IS to the presence of these bright carbon (lutings {Roy. .Soc. Proc, 

 vol. .\lvii. p. 40). I)r. Copeland had previously made important observatiotfs 

 of '"Nova" Orionis with reference to this point {Montltly Sotut's. vol. 

 .\lvi. p. it2), and he identified one of the bright bands .is " the gre.it hydro- 

 carbon band seen in the spectrum of every comet that has been examined 

 under favourable circumstances." Referring to his observations of o Orionis, 

 Mr. Maunder ("Greenwich .Spect. Observations," 1889, p. 22) slate.s that 

 "the carbon band at 5164 w.as coincident (within the limits of observation 

 with this dispersion) with the bright space tow.irds the blue of Duntir's 

 band 7." 



