448 



NA TURE 



[September 5, 189- 



eye obseirations. and it is only by photographs that a classifica- 

 tion from this point of new can be made. 



Some spectra show a remarkable continuous absorption either 

 in the ultra-\-iolet or Wolet, in others this absorption extends to 

 about K, whilst in a third class it reaches as far as G. 



These considerations gave four marked groups. Each of these 

 main groups are next sub-di\nded into sub-groups by the most 

 marked differences in the sjiectral lines. 1 do not projwse to 

 give the detailetl inquiry in this place. 



The important fact which stood out when the photc^raphic 

 attack had got so far was that, whether we take the varying 

 thicknesses of the hydrogen lines or of the lines of other sub- 

 stances as the basis for the arrangement of the spectra, it was 

 not possible to place all the stars in one line of temperature, 

 but it was necessary to arrange the stars in two series. 



When this sorting was completed, I was in a position to con- 

 sider the various di\isions of the photographic sjieclra thus 

 arrived at, in relation to the groups which were previously sug- 

 gested from a discussion of eye obserN'ations. It is clear that if I 

 got the same results the first conclusions would be strengthened. 



We have, therefore, to inquire how far this condition is satis- 

 fied by the mass of new facts at our disposal. This involves the 

 consideration of some points in connection with the meteoritic 

 h)'pothesi5, and it must sjiecially Ix; borne in mind that the 

 fundamental difference between mine and other classifications is 

 that it demands the existence of bodies of increasing as well as 

 bodies of decrea.sing temperatures. 



Since in my classification the connection between nebulce and 

 stars is insisted on, it was necessary to obtain a spectrum of one 

 of the brightest of the neljula; as a term of comparison. The 

 nebula of Orion was selected, and a photograph taken with a 

 30-inch silver on glass reflector in February 1890. This photo- 

 graph contained 54 lines, which were carefully tabulated for the 

 purposes of the comparison to which reference has been made. 



Tht Complex Origin of the Spectra of Nebula. 



On the hypothesis, the bright lines seen in the nebula? should 

 have three origins 



(1) The lines of those substances which occupy the greatest 

 volume (or largest area in a section) ; in other words, the lines 

 of those sulKtances which are driven furthest out from the 

 meteorites and occupy the interspaces, when possibly they may 

 be rendered luminous by electricity. Chief among these, from 

 laboratory experiments, we should expect hydrogen, and next, 

 from the same experiments, we should expect gxscous com- 

 [munds of carlxjn. 



(2) We are justified in a.ssuming that the most numerous 

 collisions will be partial ones — grazes — sufficient only to pro- 

 duce com)>aralively slight rises m temperature. The nebula 

 spectrum, so far as it is producc<l by this cause, will therefore 

 depend upon the phenomena produced in greatest numlier, and 

 wc may hence ex|)eot to find the low temperature lines of 

 various metallic substances. 



(3) In addition to the large numlicr of partial collisions there 

 will lie a relati\'ely small number of end-on collisions, prwlucing 

 very high tcm|K*ralure,' and, so far as this cause is concerned, 

 there will \k some lines pro<luce<l which are a.ssociatcd with very 

 high temperatures. 



Combining these conclusions, in the spectra of nebula: we 

 should expect tf> find evidence of 



Hydrogen and compounds of carbon. 



Low temperature metallic lines and flutings. 



Lines which are only produced at very high tem|x:ratures. . 



The Passage to Bright-liiu Stars. 



On the hy|xiihcsis, the lines seen in the spectra of bright-line 

 star>shiiul>l. in the main, resemble those which appear in ncbuke. 

 They will differ, however, for two reasons : — 



(I) fJwing t<i ixirtial condensation of the swarm the hydrogen 

 area will be restricted, and the bright lines nf hydrogen will 

 lose Ihcir prominence ; the volume occupied by the carbon com- 

 poimd» will !«• relatively increased, and the brightness of the 

 carl- ill lie enhanced. 



U nt cif the increaiscd number of collisions, more 



mcloiiit, ,. iii Im; rendered incandescent, and the continuous 

 .ipcctrum will lie brighter than in nebul.-e. 



' A'-y. .Sflt. Prttc.j voL xltii. p. 150. 



NO. 1349, VOL. 52] 



Stats 0/ Jiicrcasiiig Temperature. 



Initially, each pair of meteorites in collision may be regarded 

 as a condensation. 



Ultimately, when all the meteorites are volatilised, there will 

 only be one condensation, in the shape of a spherical mass of 

 vapour. Between these points there must be other conditions. 



(Stage I.) At the stage of condensation immediately follow- 

 ing that of the bright-line stars, the bright lines from the inter- 

 spaces will be masked by corresponding dark ones produced by 

 the absorption of the same vapours surrounding the incandescent 

 meteorites. One part of the swarm will give bright lines, 

 another dark lines at the s;ime wave-lengths, and these lines will 

 therefore vanish from the spectrum. The interspaces « ill be 

 restricted so that absorption phenomena will be in excess, and 

 the first absorption will be that due to low-temperature vapours, 

 that is, fluting absorptions of various metals. The radiation 

 spectrvim of the interspace will now be chiefly that of the com- 

 pounds of carl/on. Under these conditions we know from 

 laboratory experiments ' that the amount of continuous absorp- 

 tion at the blue end will be at a maximum. 



(Stage 2.) With further condensation the radiation spectrun* 

 of the interspaces will gradually <lisappear, and the fluting ab- 

 sorptions will be replaced by dark lines, for the reason that the 

 incandescent meteorites will be surrounded by vapours pro- 

 duced at a higher temperature, the number of violent collisions 

 per unit time and volume being now greatly increased. This 

 dark line spectrum need not necessarily resemble that of the Sun. 



(Stage 3.) The line absorption and the continuous absorption 

 at the blue end of the spectrum will diminish as the condens,a- 

 tions are reduced in number, for the reason that only those 

 vapours high up in the atmospheres surrounding the condensa- 

 tions will be comjietent to show absorption phenomena, in 

 consequence of the bright continuous spectrum of the still dis- 

 turbed lower levels of those atmospheres. 



.Vmong the more important lines which will disappear at this 

 stage will be those of iron, for the reason that there will be 

 bright lines from the interspaces occupying the same positions 

 as the tiark lines produced by the absorption of the vapour 

 surrounding the stones. 



The number of violent collisions per unit time and volume 

 being further increased, we should expect the absorption of very 

 high temperature vapours. 



The Hottest Stars. 



Ultimately, then, we should expect llvit the order of the 

 absorbing layers will follow the original order of the extension 

 of the vapours round the meteorites in the first condition of the 

 swarm, and the lines seen bright in nebuke, whatever their 

 origins may be, should therefore appear almost alone as dark 

 lines in the hotter stars, and the hydrogen especially should 

 have its lines broadened with each increase of depth in the 

 atmosphere. The continuous absorption at the violet end ol 

 the spectrum will be at a minimum. If, when the hydrogen 

 lines are thick the swarm is not yel completely condensed, 

 that is, if there l>e nebulous matter surrounding the central 

 mass of vapour, a fine bright line will be seen down the centre ot 

 each dark one. 



Stars of Decreasing Temperature. 



When we consider the cooling condition, that is, what hap- 

 pens when the temiK-ralure of the mass of vapour is no longer 

 mcreased by the fall towards the centre of meteorites composmg 

 the initial swarm, wc should exjiect to find the phenomena 

 indicated belt>w. 



(Stage I.) The hydrogen lines will begin to thin out, on 

 account of the diminishing depth of the absorbing atmosphere, 

 and new lines will appear. 



The new lines will nut necessarily be the same as those 

 observed in connection with the stars of increasing temper.ature.''' 

 In the latter there will lie the |>erpetual explosions of the 

 meteorites affecting the atmosphere, whereas in a cooling moss- 

 of va|M)ur we have to deal with the absorption nf the highest 

 layers nf vapours. Those lines whirlj will first make their ap- 

 pearance, however, will be the longest low temperature lines ol] 

 the various chemical elements. 



t lyockycrand Kobcns-Atutcn, Koy. Soc. J*rvc., 1S75, p. 344. 

 1 A'ty. Soc. I'roc.f vol. xlv, p. 3BJ. .^„ 



