April, 1903.] 



KNOWLEDGE. 



79 



From these few examples alone we might conclude that 

 " nature abhors self-fertilisation " in civil as in natural 

 history, and that no sociological species, as no biological 

 species, is perpetually self-fertilised. Where a nation, 

 like the Chinese, maintains its vigour intact for centuries 

 of refusal to cross, it owes its tenacity to being one of the 

 most widely migratory peoples on record. Dispersal over 

 a large area may protect a species against its own 

 exclusiveness. 



THE CHEMISTRY OF THE STARS. 



II.— VARIETIES OF SPECTRA. 



By A. Fowler, f.r.a.s. 



The first recorded observations of the spectra of the stars 

 and planets were made in 1814 by Fraunhofer, who then 

 discovered the important facts that Venus and Mars had 

 spectra resembling that of the sun, while the spectrum 

 of Sirius was very different. After increasing his pre- 

 viously very small optical equipment — a prism of 60° 

 in front of the telescope of a theodolite — it was found that 

 the spectrum of Castor was like that of Sirius, the spectra 

 of Capella and Pollux similar to the spectrum of sunlight, 

 and the spectrum of Betelgeuse somewhat different. 

 Fraunhofer, in short, recognised that there were at least 

 three varieties of stellar spectra. 



It was not until 18511, however, that these observations 

 received the notice to which they were entitled. In that 

 year Kirchhoff published his researches on the origin of 

 dark lines in spectra, and it soon became evident that the 

 methods employed in the analysis of sunlight could be 

 equally applied to the determination of the chemical 

 constituents of the stars. Rutherfurd, Secchi, and Sir 

 William Huggins were most prominent among the earlier 

 workers in this direction, and while the latter observer 

 chiefly concerned himself with the minute analysis of a 

 few stars, the others observed a greater number of stars 

 with a view to their classifioatiim into groups. Following 

 up this pioneer work, detailed investigations of the spectra 

 of the stars have been carried on with conspicuous success 

 by Prof. Pickering in America, Sir Norman Lockyer, Sir 

 Wm. Huggins, and Dr. McClean, in England, and 

 Professors Vogel and Scheiner in Germany. 



As a general result of the work which has been done, it 

 is found that the great majority of the thousands of stars 

 which have been studied either^fall into one of the types 

 recognised by the earlier observers, or may be considered 

 as intermediate between two of the types ; that is, the 

 different varieties merge into each other by almost 

 insensible gradations. For this reason it is now generally 

 believed that stars are not fundamentally different in 

 chemical composition, but that the differences in their 

 spectra an; mainly due to differences in the physical states 

 which the stars have reached in an evolutionary process. 



In classifications of stellar spectra, an attempt is 

 accordingly made to indicate them-der in which the various 

 spectra succeed each other in the history of a star, and not 

 merely to put similar stars into the same group. There 

 are some points, however, on which authorities differ, 

 though there is a general agreement in regard to some of 

 the ])rincipal types ; and it is doubtless because of the 

 divergencies of opini^m that the classification suggested bv 

 Secchi, with slight additions, is even now tlie only one 

 which can be considered an universal language of the 

 subject. It is, in fact, still a very convenient classification 

 so far as concerns the observations which can be made 

 visually, though it is hopelessly inadequate to deal with 

 the finer differences of spectra revealed by modern photo- 



graphic methods. There is one feature of Sir Norman 

 Lockyer's classification which has much to recommend it 

 as a tentative nomenclature for the greater number of 

 groups which must now be separately distinguished. In 

 this, the various groups are no longer denoted by a dis- 

 tinctive number or letter, but by a name derived from that 

 of a typical star, a plan which had been previously 

 adopted in a small way by other writers ; thus if a star 

 gives a spectrum like that of Sirius, it is called a Siriaii 

 star; if like Antares, an Aiitarian star; if like 19 Piscium, 

 a Piscian star, and so on. This nomenclature has the 

 advantage of admitting of easy extension as future 

 researches may require, without disturbing any arrange- 

 ment of numbers or letters. 



It is proposed to discuss later the more important 

 varieties of spectra in some detail, but meanwhile it will 

 facilitate matters to take a broad general review of the 

 various types. In the light of our present knowledge, 

 the classification of Secchi may be conveniently exteaded 

 as follows, retaining his enumeration so far as it goes, and 

 indicating the varieties of each which have been separately 

 named by Lockyer : — 



Type O. — Spectra in which the dark lines of hydrogen 

 and helium are of moderate intensity and nearly all 

 other lines relatively feeble. 



From their frequent occurrence in the constellation 

 Orion, the stars of this type are often called " Orion 

 stars," but they are by no means confined to that part 

 of the heavens. 



" Helium stars " is a name also sometimes applied 

 to stars of this class. 



A few stars of this type, such as y Cassiopeise, 

 show bright lines in addition to the dark ones. These 

 are frequently referred to as stars of the " Orion type 

 with bright lines." 



Includes the stars classified by Lockyer, according 

 to differences in detail, as Alnitamiaii (so named from 

 £ Orionis), Crucian (S Crucis), Tauriau (^ Tauri), 

 Achendnn {a. Eridani), Algolian and Rujeliaii. 

 Type I. — Spectra in which the dark lines of hydrogen 

 are very strong and all other lines relatively weak. 

 The weaker lines include both arc and spark lines of 

 various metals. 



Includes the groups named by Lockyer Markabiau 

 (at Pegasi) and Siriaii. The Cygniaii (a Cygni) 

 stars may be regarded provisionally as a special 

 variety in which the hydrogen lines are somewliat 

 subdued, while the spark lines of the metals are 

 especially developed. 

 Type I. -II. — Spectra intermediate between Types I. and 

 II. ; the hydrogen lines are of moderate strength, and 

 a great number of metallic lines, both arc and spark, 

 are also conspicuous. 



Includes Lockyer's group of rrocyunian stars. The 

 Polariaii gi'oup, also having numerous dark lines, 

 may be provisionally regarded as a special variety ; 

 the hydrogen lines are less intense than in Procyon. 

 Type II, — Spectra resembling that of the sun, in having 

 a multitude of strong lines corresj)ouding with lines 

 seen in the arc spectra of the metals, the hydrogen 

 lines having lost their predominance. From their 

 resemblance to the sun, the sturs of this class are 

 often called stars of the " solar tyjK.>." 



Includes Lockyer's groups of Ahlebariau and Arc- 

 tiirian stars. 

 Type III. — Spectra with dark flutiugs, fading towards 

 the red end of the spectrum, and having also many 

 metallic lines. The hydrogen lines are very feeble. 

 Includes Lockyer's Antarian group. 



