March 29, 1906] 



NA TURE 



507 



association with titanium llutings or bands, then to 

 a Herculis, in which " the bands have acquired 

 strength through the efflux of time, it is supposed, 

 and the progress, of cooling." Further than this 

 she does not go, Secchi's fourth type of stars, showing 

 absorption flutings of carbon in their spectra, being 

 regarded as a class apart, and not united to the sun 

 and its congeners in an evolutionary series. 



In this classification of stellar spectra, temperature 

 is not considered as an essential factor or a con- 

 comitant of the changes described ; the question of 

 the temperatures of the stars forms the subject of a 

 separate chapter, new to the present edition. Heat 

 is not entirely discarded, but "luminescence," 

 "radiology," and "electrical excitement" now 

 appeal to Miss Gierke's affections, and she flirts with 

 them whenever she has the opportunity, though little 

 is known of their resources. By this course she is 

 able to believe, with Sir William and Lady Huggins, 

 that solar stars are hotter than Sirian stars, while 

 she acknowledges, as has been shown, that arc-lines 

 are characteristic of the former type of spectra and 

 spark-lines of the latter. Whether the temperature 

 of the electric spark is actually higher than that of 

 the arc has yet to be decided, but solar stars can only 

 be placed at a higher temperature-level than white 

 stars by neglecting much circumstantial evidence in 

 support of the superiority of the spark. As a matter 

 of fact, laboratory observations, so far as they are 

 available, show that the same spectra can be pro- 

 duced by thermal or electrical action. Under con- 

 ditions from which electrical influences were probably 

 excluded, and the spectra obtained were due solely 

 to high temperature, nitrogen at temperatures above 

 3000" C. has been found to give an emission spectrum 

 in which the principal lines characteristic of the 

 element were visible (Nasini and Anderlini, Atti del 

 Lined, July, 1904). The agency by which gases and 

 vapours are rendered incandescent seems, indeed, to 

 be inconsequential, and high temperature is probably 

 competent to produce the same spectroscopic results 

 as those observed when incandescence is caused by the 

 oscillatory discharge. 



Electric and thermal effects cannot, however, be 

 distinguished from one another in stellar spectra; 

 therefore any attempt at a temperature classification 

 must provide for possible electric influences. It was 

 recognised by Sir Norman Lockyer more than thirty 

 years ago (Roy. Soc. Proc, vol. xxii., p. 372, section 

 ii., June, 1874) that the action of electricity must be 

 included in the term " temperature "; and while the 

 chemical changes effected by thermal and electric 

 forms of energy cannot be discriminated in celestial 

 spectra, the aim should be to construct a chemical 

 classification without waiting for a complete under- 

 standing of the active causes of atomic vibrations or 

 molecular combinations revealed by the spectroscope. 



From whatever point of view stellar spectra are 

 studied, little support can be found for the conclusion 

 that the solar stars are at a higher temperature-level 

 than the white stars. From a comparison of the 

 spectra of Capella and Vega — typical solar and white 

 stars respectively — Sir William and Lady Huggins 

 no. 1900, VOL. 73] 



concluded that " The solar orb seemed intrinsically 

 the bluer, and was inferred to be the hotter of the 

 two"; but neither the observation nor the inference 

 can be regarded as established. Adopting the relative 

 length of the ultra-violet spectrum as a criterion of 

 stellar temperature, observations show that solar stars 

 are really weaker in ultra-violet rays than white stars. 

 In a paper on " Radiation through a Foggy Atmo- 

 sphere " (Astropliysical Journal, vol. xxi., No. 1, 

 January, 1905) Prof. Schuster accepts " the compara- 

 tive weakness of the ultra-violet radiation in solar 

 stars " as a fundamental fact which he attributes to 

 molecular scattering in the photospheric regions of 

 these bodies; but whatever the explanation, it is clear 

 that spectroscopists have not adopted Sir William 

 and Lady Huggins's view as to the ultra-violet spec- 

 trum of solar stars, but hold an opinion directly 

 opposed to it. The extension of spectra into the 

 ultra-violet may be regarded as the result of increased 

 temperature, but by this standard white stars are 

 placed above solar stars, and not below them. 



All standards of comparison should, however, lead 

 to the same spectroscopic succession if they are true 

 tests of evolutionary development of celestial species. 

 The sequence derived from comparisons of the lengths 

 of ultra-violet spectra is the same as that revealed 

 by the presence of gaseous and metallic lines of 

 helium and hydrogen, and the enhanced and arc lines 

 of the metals. In stars which have relatively the 

 longest ultra-violet spectra there are few absorption 

 lines ; iron is represented practically by the enhanced 

 lines alone, and the lines characteristic of the arc 

 spectrum are almost or entirely absent. Only by 

 considering the length of the ultra-violet spectrum 

 together with the presence or absence of iron lines and 

 lines that are intensified in passing from the condition 

 of the arc to that of the spark can a useful classifi- 

 cation of stellar spectra be established. When this 

 principle is adopted a chemical classification of spectra 

 becomes possible, and a reasonable scale of stellar 

 thermotics is arrived at. From a hot star like 

 Bellatrix a descending series can be arranged through 

 $ Persei, y Lyrse, Sirius, Castor, and Procyon to 

 Arcturus — a relatively cool star — by considering the 

 changes in the spectrum of any constituent in passing 

 from one grade to another. A continuous ascending 

 series of spectra reaching up to Bellatrix can also be 

 arranged from the spectrum of nebulae of the Orion 

 type through planetary nebula? and the Wolf Rayet 

 stars showing the line a. 4688 in their spectra to Orion 

 stars in which this line is dark. This arrangement 

 of spectra corresponds also with the sequence which 

 would be expected as the result of various degrees of 

 chemical dissociation at different temperatures. Stars 

 having the smallest number of chemical elements re- 

 presented in their spectra are probably the hottest, 

 while an increased number of lines in other spectra 

 is probably due to the existence of an increased 

 number of chemical elements as the result of lower 

 temperatures, the inferior position on the temperature 

 scale being indicated also by the reduction of the 

 relative length of the spectrum, increase of the relative 

 intensity of red radiation and general absence of 



