28: 



NATURE 



[November 12, 1914 



NOTES ON STELLAR CLASSIFICATION. 

 1 N all my early work at Kensing-ton (1887) on 

 -»- the classification of stars according- to their 

 spectra, the best basis for a classification was 

 chiefly considered. The data available were the 

 chemical facts obtained by a detailed inquiry into 

 the chemical origins of the various lines and 

 flutings, using higher dispersion and more labora- 

 tory work than were employed in other observa- 

 tories. 



The first and most important result obtained 

 was that it was found necessary to divide the 

 stars into two groups, one set increasing their 

 temperatures while in the other the temperature 

 was running- down. 



This, indeed, added a physical difference to the 

 chemical ones, and there was another, for on the 

 meteoritic hypothesis the stars getting hotter 

 were sparse swarms of meteorites being vaporised, 

 while those getting- colder were rapidly condensing- 

 into solid bodies. 



These were fundamental changes of front. Up 

 to that time only a line of decreasing- temperature 

 had been considered. Passing over the earlier 

 classification of Rutherfurd (1863), which was only 

 modified by Secchi (1867), though by those ignor- 

 ant of the subject he is credited with the origina- 

 tion of it, I may mention that in 1873 I referred to 

 the question of temperature in a Bakerian Lec- 

 ture, ^ and in 1874 Vog-el brought out a consider- 

 able classification. This was also based upon a 

 line of descending- temperatures. On this clas- 

 sification I wrote in another Bakerian Lecture 



(1882) 2; — 



"The idea which underlies the classification is 

 that a star of Class L on cooling- becomes a star 

 of Class II., and that a star of Class II. has as it 

 were a choice before it of passing- to Class Ilia, 

 or Class 1 1 lb. Thus under certain conditions its 

 spectrum will take on the appearance of Secchi 's 

 third type, Class Ilia. (Vog-el); in certain other 

 conditions it will take on the appearance of 

 Secchi 's fourth type. Class I lib. (Vogel). There 

 is now, however, no doubt whatever that Secchi 's 

 Class III., represents stars in which the tem- 

 perature is increasing-, and with conditions not 

 unlike those of the nebulae — that is to say, the 

 meteorites are discrete, and are on their way to 

 form bodies of Class II. and Class I. by the 

 ultimate vaporisation of all their meteoric con- 

 stituents. There is also no doubt that the stars 

 included in Class Illb. have had their day; and 

 that their temperature has been running down, 

 until owing- to reduction of temperature they are 

 on the verge of invisibility brought about by the 

 enormous absorption of carbon in their atmo- 

 spheres." 



" Pechiile was the first to object to \'og-ers 

 classification, mainly on the ground that Secchi's 

 types III. and IV. had been improperly brought 

 together; and my work has shown how very just 

 his objection was, and how clear-sighted was his 



1 Phil. Trans , vol. clxiv., p. 292. 

 ■-! Proc. R.S., vol. xliv., p. i. 



NO. 2350, VOL. 94] 



view as to the true position of stars of Class 

 Illb." 



Prof. Keeler wrote in 18942 " Lockyer's system 

 of stellar classification provides for both an as- 

 cending and a descending branch of the tempera- 

 ture curve, and in this respect it certainly has 

 advantages over other systems which claim to 

 have a rational* basis." 



Prof. Frost * thus alludes to it : " Many spectro- 

 scopists are unwilling to admit that the only 

 course of stellar development is along a line 

 of descending temperature." He does not name 

 the "many spectroscopists," but he adds: " The 

 further question may be raised— Is it not possible 

 that a celestial body, after it has reached the 

 condition in which we may define it as a star, 

 may twice be red and exhibit a spectra of the 

 third type, once in its youth and again in its 

 old age? " 



In my papers are many, statements giving the 

 answer to this question, years before Prof. Frost 

 put it. 



In the Bakerian Lecture (1888) I published 

 a temperature curve with ascending and descend- 

 ing arms, along which stars could be arranged 

 according to their swarm- and condensing-condi- 

 tions and their temperature as revealed by the 

 spectra. On it were rearranged stars as classi- 

 fied by Vogel and stars with bright lines and 

 nebulae. Another curve was giv^en in the cata- 

 logue of 470 of the brighter stars which I after- 

 wards published in 1902, in which I classified 

 stars only, giving to each group a distinctive 

 name. 



Of the stellar spectra used, some had been 

 photographed at the Solar Physics Observator} 

 and others by Mr. McClean at the Cape and 

 Tunbridge Wells. 



A few years after the publication of X'ogel's 

 memoir the work of classification was taken 

 up at Harvard. The first one adopted was an 

 alphabetical, from A and B onwards in one line. 

 This we may mark as H.i. 



At about the same time (1888) that I published 

 my temperature curve at Kensington, the classifi- 

 cation of northern stellar spectra in special new 

 groups was being carried on at Harvard by Miss 

 Maury as part of the Henry Draper Memorial. 

 The classification was on one line of temperature 

 in groups numbered from i to xxii. (H.2.). 



But Miss Maury did not content herself with 

 the group classification ; over and above her 

 twenty-two groups of spectra she gave us three 

 main divisions of stars depending upon the char- 

 acteristics of the individual lines. Here we find 

 a physical touch superadded. In Division a were 

 included lines narrow and clear, in h lines relatively 

 wide and hazy, and in c, among other conditions, 

 unusually intense metallic but not solar lines. I 

 found- these divisions of the utmost value to me 

 when I was preparing my classification. 



The Southern stars were undertaken by Miss 



s "Ast. and Ast. Phys." 1894, p. 60. 



* "Astronomical Spectroscopy," p. 318. 1894. 



