6i8 



NATURE 



[February 4, 19 15 



NOTES ON STELLAR CLASSIFICATION. 

 II. 



IN the first series of notes ^ I stated how I was 

 trying- to find whether a combination of the 

 meteoritic hypothesis and my system of chemical 

 classification of stars into two series of ascending 

 and descending temperatures might land us in a 

 method of detecting physical differences. The 

 importance of this inquiry depends upon the fact 

 that any system of chemical classification along 

 one line must land us in confusion, seeing that 

 equal or nearly equal temperatures, and therefore 

 chemistry — for chemistry is the child of tempera- 

 ture- — mark two very different physical states in 

 the life of a star in its progress from nebula to 

 extinction as a cold solid globe. 



In order to show the method of attack I give 

 here a diagram showing my classification and that 

 of Harvard (H.3) based upon chemistry. The 

 Harvard symbols given are those of the type stars 

 used in my classification, but the letters B, F, K, 

 etc., are generally used alone. 



The difference in physical state as- 

 sumed on the meteoritic hypothesis is 

 indicated by the wide dotted curve to 

 represent nebulous conditions at the 

 bottom of the ascending arm, and the 

 firm dark line to represent condensed 

 stars at the bottom of the descending 

 one. 



I first deal with the evidence of 

 bright lines in certain chemically 

 classified stars as constituting a 

 difference from the normal stars in 

 which bright lines do not occur. The 

 inquiry will show us the stellar 

 species in which these bright lines 

 occur. 



In 1876 I suggested that bright 

 lines in stellar spectra might be 

 added to a dark line spectrum by our 

 spectroscopes revealing to us the 

 presence of a gaseous envelope above the most 

 valid absorbing region. ^ I believe Pickering was 

 the first to see such an envelope, and it has been 

 seen since by others. 



How such an envelope can strengthen its story 

 when its temperature is increased is shown in a 

 valuable table given in Miss Cannon's cata- 

 logue. ^ There we find eighteen stars in which 

 bright H)8 alone is superposed on the absorption 

 line in eight stars. In the others Hy is added in 

 ten cases. In six stars we get HS added to Hj8 

 and Hy. In one star we get a fourth line, He, 

 added to the preceding three. As lines are added 

 the intensity of the earlier lines is increased. The 

 extreme case is fx Centauri, where the intensity of 

 the hydrogen lines is €2, S3, y 6, and jS 10. In 

 some cases the bright lines are variable, but the 

 same law holds ; the work always begins at 

 the red and decreases towards the violet (Fig. 2). 



But this is not all the story. In /x Centauri, 



1 Nature, November 12, 1914. 



- See Proc. Roy. Soc, 1878, p. 49. 



'•> Harvard Annals, vol. xxviii., part ii., p. 228. 



where we get the greatest development of the 

 bright hydrogen, there are other lines. All these 

 lines but one are enhanced lines — and of iron ! 



Among the facts to be borne in mind in con- 

 sidering these questions are those revealed by the 

 recent study of the Madrid photographs of the 

 recent Nova (Geminorum 2), recorded in the last 

 volume of Kensington researches. 



The photographs showed a well-marked stage 

 as the Nova cooled down, principally indicated by 

 a bright line near A 4640 (4639-2) ; and it was 

 further noted that in previous Novae the appear- 

 ance of this line preceded that of the ordinary 

 nebula line near A 500, and further that it was the 

 brightest line in the whole spectrum where H^, 

 Hy, and A 4472 were visible. 



The 4640 line has been noted in several planet- 

 ary nebulae, and among the bright lines of the 

 Wolf-Rayet stars not yet chemically classified. 



This line is associated in Novae and elsewhere 

 with another at A 4688, not produced by the same 

 substance for the hnes vary inversely ; it 



Sequence of Stellar Temperatures. 



Arconian 

 Alnitamian. 



O. 

 B 



Crucian 

 Taurian 



RiCELIAN 



Cycnian 



POLARIAN FSPk 

 ALDEBARIAN.K5. 



Antarian 



AcHERNIAN . 



Alcolian . 

 Markabian . 



SiRIAN . 



Procyonian. 

 Arcturian . 



PiSCIAN . 



Bs. 

 Ba. 

 A 



A . 

 Fs. 

 K . 

 N 



Fig. I. — The temperature curve. 



also is frequently seen in the Wolf-Rayet 

 stars. 



We seem justified, then, in assuming, to begin 

 with at least, four stages of bright lines : — 

 H + enhanced lines 

 H + 4688 

 H + 4640 

 lowest temperature H + 500. 

 and these are common to the stellar envelopes a 

 novae. 



If we include P Cygni, we find another change, 



highest temperature H + He + lines of O + N. 

 Of the substance which produces 4640 we know 

 as yet nothing, but it is possible that 4688 repre- 

 sents carbon, of the existence of which in nebulae 

 and the Wolf-Rayet stars there is evidence.* 



On the other hand, in 1905 ^ Mr. Baxandall 

 and myself recorded a line -at A 4685*97 in the 

 spectrum of a helium tube, and its occurrence 

 in stars was noted. This would be another origin; 



4 See especially my paper in Proc. R.S., vol. xlvii., p. 40, 18 



5 Pioc. Roy. Soc, vol. l*xiv., p. 546. 



\ 



NO. 2362, VOL. 94] 



