Decembee 2, 1896.] 



KNOWLEDGE 



279 



shuts off the always-glowing interior, the star's continuous 

 spectrum vanishes, but the bright line spectrum will not 

 disappear, since it has its origin in the external atmosphere 

 overlaying the crust, where, under the influence of the now 

 rapid cooling, chemical association still continues and so 

 produces a discontinuous spectrum giving to the extin- 

 guished new star the appearance of a planetary nebula. 



This appearance has been discovered in Nova Cygni by 

 Lord Crawford, in Nova Aurigfe by Gothard, in Nova 

 Normffi by the astronomers of the Harvard College Obser- 

 vatory. According to Dr. and Mrs. Huggins, the spectra 

 of nebulio and novas have no real resemblance, for in the 

 position where the nebulie show their narrow and defined 

 lines they found in Nova Auriga?, February, 1893, extended 

 groups of lines {Froc. Bm/itl Soc. 54, p. 30). But since 

 February, 1893, the spectrum of Nova Aurigaj has shown 

 very interesting and significant changes, and now, according 

 to the numerous and complete observations of Campbell at 

 the Lick Observatory, is approaching the average type of 

 nebular spectrum {Astroph. Journal, Jan. 189o, page 50). 

 This remarkable assertion of Campbell has suggested to 

 me the following idea : If extinguishing stars show the 

 spectrum of gaseous nebuLie, then star clusters, where the 

 great majority of stars are extinguishing, will also show 

 such a spectrum. Thus if we assume that different clusters 

 can be of different ages or more or less approaching their 

 final extinction, then the younger irresolvable clusters will 

 show a substantially continuous spectrum, as does the 

 great nebula in Andromeda, and the elder irresolvable 

 clusters will show a gaseous spectrum, as does the great 

 nebula in Orion. If the spectra of gaseous nebulre, with 

 their bright lines and their more or less visible continuous 

 light, can be explained in this way, we have no longer to 

 deal with a very enigmatical nebulous matter showing in 

 its incommensurable aggregations neither any clear conden- 

 sation into centres of attraction nar an;/ relative motion of its 

 dijf'crent pnrta even in the line of sii/ht (Keeler: Nature, Dec. 

 27th, 1894). We can then comprehend also why nebulfe 

 described by Herschel as easily resolvable, as, for instance, 

 G.C. 4499, G.C. 4827, could show to Huggins a gaseous 

 spectrum. Such clusters, when spectroscopically observed, 

 lose almost entirely by spectral dispersion the white light 

 of theh small still visible stars, and show only clearly as a 

 gaseous spectrum the light of their countless already 

 invisible and extinguishing " new stars." It is clear, 

 moreover, that if the explanation suggested above is made 

 to depend a little upon my views as to the origin of new 

 stars, that dependence is not at all necessary. New stars 

 may be due to collisions of stars (Vogel), of stars and 

 cosmical clouds (Seeliger), of a star and smaller bodies 

 crossing its atmosphere (Beiopolsky),of meteoritic swarms 

 (Loekyer) ; they may be due to periastric tidal disturbances 

 (Huggins), to gaseous eruptions on a single star(Sidgreaves), 

 or to some other cause. My explanation demands nothin,' 

 else than that in some clusters the production of new sta;:i 

 might be very abundant and much more abundant than in 

 others. Since new stars, when small, may be easily over- 

 looked, it is probable that even in our own cluster they are 

 much more numerous than has been hitherto observed. 

 Again, all other bright line bodies, such as, for instance, 

 Beta Lyrfe, Gamma Cassiopeife, the Wolf-Piayet stars, our 

 sun, the red variables, and comets will eventually co-operate 

 to cause the spectrum of a cluster to have a gaseous 

 appearance. 



I know very well that, since novre show two super- 

 posed spectra, whose dark and bright lines are relatively 

 displaced, this appearance has led to the hypothesis of a 

 direct collision of two or more celestial bodies, or of one 

 body and a cosmical cloud, the bodies strongly heated by 



the collision showing afterwards by the displacement of 

 their lines their . relative motion in the line of sight. 

 But apart from the enormous amounts of the velocities 

 so suggested — velocities of -nhich no explanation is afforded, 

 and which are almost inconceivable in their character — 

 and apart also from many other difSculties that MM. ^'ogel 

 and Seeliger and others have expounded in their dis- 

 cussions on this subject, (a) we are not obliged to see in 

 the indicated displacement of lines a convincing proof of 

 motion in the line of sight. 



There are, indeed, countless experiments that have 

 proved in our laboratories that, without any motion of the 

 source of light, a displacement of lines can occur ; firstly, 

 by changes of pressure (b) ; secondly, by changes of 

 temperature {<•) ; thirdly, by changes of admixed matter 

 (d) ; fourthly, by changes of chemical combination (e i. 

 Now, if one thing is absolutely certain in our know- 

 ledge of sun and stars, it is this, that in their superposed 

 layers there must be great differences of pressure, 

 temperature, admixed matter and chemical combina- 

 tions. And it is therefore very strange that in general, 

 in explaining the celestial spectra, all line displacements 

 are simply interpreted as if all the experiments indicated 

 above had never been made, and as if the only possible 

 cause of such displacements was a hypothetical motion in 

 the line of sight. 



Our sun often presents dissymmetrical reversals having 

 some analogy with those of the novas. M. Deslandres 

 has discovered that the dark reversal of the bright H and 

 K lines does not always occupy the centre of these lines. 

 Whatever may be the cause of such a dissymmetry, 

 its occurrence in a single body proves that it does not in 

 every case require the hypothesis of a collision of different 

 bodies in order to explain it ( /). M. Lockyer's photographs 

 of the arc spectrum have shown, moreover, that even in the 



(a) H. C. Vogel, Astr. and A., December, 1893, January, 1894, 



February, 1894; '' On the New Star in Auriga." H. Seeliger, iiirf.,p. 

 142 ; " On the New Star in Auriga." 



Fr. Sidgi-eaves, Mem. of the S.A.S., 61, p. 34. W. H. Pickering, 

 Aslr. and A., 94. p. 201. 



" It may be also that the spectral lines are liable to be displaced by 

 other causes than motion in the line of sight. The spectrum of the 

 nora in Am-iga is rather startling, if such Telocity is the only admissible 

 explanation of it." (W. H. Monck. Fulliraiiom of the Ast. Soc. of 

 the Pacific, VII. p. 38). If the sudden iip-blazing of the novfe wen 

 due to heat evolved by some collision of stars, their afterwards rapid 

 reduction of light would be inconceivable (Scheiner, Spectr. d. 

 Gesfinie. p 300 ; Locky r ; The Meteoritic Blip. p. 326). How 

 much easier that rapid reduction can be underst od when, admitting 

 that the temperature remains almost uneliauued, the phenomenon is 

 entn-ely due to the temporary dispersion of a dark veil in the outer 

 layers of the star, wliich there necessarily recoudeuses again by loss 

 of heat as soon as there no longer a sufficient heat is evolved b\ 

 chemical combination. The rapid reduction of light being so 

 explained, we have there (as well as in the red variables) a phenomenon 

 of the same order as tlie darkening of our sun by even the smallest 

 clouds in our sky. 



(6) ZoUner, Ser. d. Sac/is. Ges. d. Wiss, 31st October, 1870, p. 

 233. Miiller, Fogg. Ann. 150 p 311. 



Loekver, Phil. Tram., 74, p. 8' '5. P'oc. R. S 79, p. 428. Liveing 

 and Dewar. Proc. R. S., 28, p. 367.' Schuster, Nature, 75, p. 148. 



Ciamicean, H'iener Per., 78, p. 886. Schuster, Die sptctral- 

 analgse v. Eoscoe, 90, p. 148. Scheiner, Spectridatvilyse d. Gestirne, 

 p. 143. 



(e) Eoscoe, Die Spectralanali/se 1890, p. 148. Proc. R. Soc. 21. 

 p. 282. 



(d) Loekyer, Chemistry of the Sun, pp. 369-370. Vogel, Piacf. 

 Spectr. Anal. p. 248. 



Mon. d. All. d. Wiss. (Mai, 1878). Spectr. Anal. v. Roscoe, 1890, 

 fig. 63 (4, 5, 6 ) 



Claes , JFied. Ann. 3, p. 389. Kundt., ibid. 4, pp. 34. 



(c) Eussell, Phil. Trans. 81, p. 887. Proc. R. Soc. 32, p. 258. 

 Die Speclralanali/se ». Roscoe, 1890, pp. 163-173. Mitscheriicli, Ann. 

 Pki/.i. Chem. 121," p. 3. 



(/) Deslandres, Comptes Rend., 29 Aout., 1894, p. 457. 



