October 27, 1904] 



NATURE 



635 



THE CLASSIFICATION OF THE STARS 

 ACCORDING TO THEIR TEMPERATURE 

 AND CHEMISTRYA 



II. 

 ■T^HE temperature relationships of the various groups in 

 ■'■ the classification of stars described in the previous 

 article are further illustrated in a diagram which is repro- 

 <tuced in Fig. 6, from which it will be seen that the stars 



Highest temperature. 

 10 Argonian 



Achernian «; 

 .■\lgolian £ 

 Markabian 5 



Sirian 

 Procvonian 



Antarian Stars with fluted spectra Piscian 



Fig. 6. — Temperature relationship of stellar groups. 



are arranged in sixteen groups along a temperature curve 

 having its apex in the middle. Apart from the inferences 

 as to low temperature which may be drawn from the occur- 

 rence of flutings, it will be remembered that while the 

 evidence for temperature was primarily based on the 

 strength of the continuous radiation in the violet, this de- 

 tailed arrangement of the stellar groups depended upon the 

 chemical sequence revealed by the successive predominance 

 of metallic, proto-metallic, and " gaseous " lines in the 

 different stars. 



This chemical sequence, however, so far as could be 

 judged from the photographs then available, was identical 

 with that derived from the investigation of the violet radi- 

 ation, thus showing that the apparent 

 chemical differences resulted from 

 changes of temperature. The 

 simplest explanation of the chemical 

 changes is that afforded by the dis- 

 sociation hypothesis, according to 

 which the step from metallic to proto- 

 metallic vapours is to be regarded 

 as a breaking up of complex mole- 

 cular groupings into simpler ones 

 still retaining characteristics which 

 permit the parent substance to be 

 distinguished, while the continuation 

 of the process results in the reduc- 

 lion of all substances to the finer 

 forms of hydrogen and helium. 

 The classification of the stars on a 

 temperature basis is therefore of the 

 utmost importance, not only for the 

 indications which it gives as to the 

 processes of stellar evolution, but 

 also on account of the light which 



it throws on the dissociation hypothesis and the evolution 

 of the chemical elements. Adopting the foregoing tempera- 

 ture sequence of the various stellar groups, this side of the 

 t|urstion has already been fully dealt with by Sir Norman 

 I.orkyer in his work on " Inorganic Evolution " (Mac- 

 niillan and Co., Ltd., iqoo). 



^ Continued from p. 614 



NO. 1826, VOL. 70] 



In view of the important issues depending upon the correct 

 determination of relative stellar temperatures. Sir Norman 

 Lockyer has endeavoured to confirm his previous results by 

 another piece of work, of which an account was recently 

 communicated to the Royal Society.' The previous con- 

 clusions involving the intensity of the violet radiations de- 

 pended upon photographs taken with optical appliances com- 

 posed of glass, which has a marked absorption for these 

 rays, and the relative intensities were judged by noting the 

 limits of the spectra in photographs which were of the 

 same intensity in the region about H7. Although there was 

 no reason to suppose that the general laws of continuous 

 radiation would be modified in still more refrangible parts 

 of the spectrum, it was possible to test the results further 

 by including this region in the discussion. An instrument 

 which was transparent to the ultra-violet radiations was 

 necessary for this research, and the one devised for the 

 purpose was a prismatic camera having a 2-inch 30° calcite 

 prism mounted in front of a si-inch quartz lens of 18 inches 

 focal length (Fig. 7). The prism is so cut that its first face 

 is perpendicular to the optic axis of the crystal, and it is 

 so arranged that the incident rays are normal to this face. 

 All the ravs, therefore, pass through the prism parallel to 

 the optic axis, and there is consequently no douWe 

 refraction. 



By this means it became possible to utilise not only the 

 length of the spectrum in the violet, but the relative bright- 

 ness of the different parts to a greater extent than before. 



To make the matter clear, it may be pointed out that the 

 temperatures of most of the stars are too high to permit 

 of their determination from the actual limits of the con- 

 tinuous radiations towards the violet, as might conceivably 

 be done in the case of a red-hot poker, since these limits 

 He bevond the wave-length for which our atmosphere is 

 transparent. The principle involved in the method employed, 

 however, is clearly indicated by Sir Norman Lockyer in a 

 quotation from Sir George Stokes," the substance of which 

 has been fuUv borne out by more recent work, namely, that 



" When a solid bodv such as a platinum wire, traversed 

 by a voltaic current, is heated to incandescence, we know 

 that as the temperature increases, not only does the radi- 

 ation of each particular refrangibility absolutely increase, 

 but the proportion of the radiations of the different re- 

 frangibilities is changed, the proportion of the higher to 

 the lower increasing with the temperature." 



In the case of stars, the radiation is of course modified 

 bv the continuous absorption of the stellar gases and 

 vapours ; but, so far as we know, the greater absorption is 

 always associated with reduced temperature " and increased 

 densitv, and regularly diminishes in intensity from the ultra- 

 violet towards the red. 



Fig. 7. — Quartz-Calcite Prismatic Camera. 



Conclusions as to the temperatures of the stars, as was 

 suggested by Crova in 1S78,* may therefore be based on 



1 Roy. Soc. Prac, vol. Ixxiii. p. 227 (igo^). 



2 Roy. Soc. /"r^c, vol. xxiv p. 35i(>876)- . . 



s It has been observed that various metals, including iron, produce a 

 conlinuous absorption at the blue end of the spectrum when reduced to the 

 state of vapour in the relatively cool oxy-hydrogen flame (Lockyer and 

 Roberls-Austen, Roy. Soc. Pnc , vol. xxiii. p. 344, 1875). 



■* Cotiiptes rcndusy vol. xvii. p. gSr. 



