THE SPECTRA OF STARS AND NEBULA. 



151 



convenience of description, for it was found that the different 

 types were not abruptly divided, but were connected by sped ra 

 representing well-marked transition stages. 



Here we get the first definite indications of an evolution of the 

 stars somewhat analogous to that which Darwin enunciated for 

 organic life. The differences in the spectra of the stars are not 

 to be attributed primarily to differences in composition, but to 

 their having reached different stages in an evolutionary process. 

 Continuit y of the spectral series practically compels us to believe, 

 for example, that our Sun was once a star like Sirius, and that 

 in due course Sirius will become a star like the Sun, the Sun 

 meanwhile having become a red star with a spectrum of bands. 

 We now speak quite freely of young or early-type stars, and of old 

 or late-type stars. On the Harvard system of classification, the 

 successive stages are designated 0, B, A, F, G, K, M, N, in the 

 order of early to later types. 



The inclusion of solar, or G-type, stars in the evolutionary 

 scheme necessarily implies that all the stars are similar in 

 chemical constitution to the Sun, but independent evidence of 

 the universal distribution of terrestrial kinds of matter is to be 

 found in abundance in the analysis of individual stars. It is 

 especially instructive to begin at the lower end of the stellar 

 sequence, where there is every reason to believe that the 

 temperatures of the stars involved are relatively low, so that the 

 reproduction in our laboratories of the lines and bands of which 

 they are characterised should present the minimum of difficulty. 

 This expectation is completely realised. In the relatively cool 

 stars of classes N and M, we find bands of carbon and titanium 

 oxide respectively, together with such metallic lines as can be 

 produced at comparatively low temperatures. As we go upwards 

 in the series, from the red stars, through the yellow and white 

 ones, to the Wolf-Eayet stars of class 0, it has been found that 

 while we continue to deal in the main with familiar elements, 

 the reproduction of the stellar lines demands gradually increased 

 energy of the exciting sources. Beginning with flames and 

 electric discharges of low intensity, we end near the upper 

 limit of the series with the most powerful discharges to which 

 our spectrum tubes of glass or quartz will submit. The sur- 

 prising thing is that the resources of our laboratories are already 

 adequate to reproduce so many of the lines which occur in the 

 spectra of the stars, even of those which are believed to be at 

 the highest temperatures. There are still some celestial lines of 

 unknown origin, but previous experimental success encourages 

 the hope that they may yet be reproduced from terrestrial matter. 



