ANALYSIS OF STARLIGHT — PAGEL 315 



between the selective absorption at the line wavelength of our atoms 

 of hydrogen, calcium, or iron on the one hand, and the continuous 

 absorption of the atmosphere, at all wavelengths, on the other hand. 

 If the atmosphere is transparent, then we can see a large number of 

 atoms of calcium or hydrogen and the absorption lines will be corre- 

 spondingly strong; whereas if the surface layers are more opaque, 

 we shall see weaker lines even if the number of hydrogen and calcium 

 atoms per gram is the same. 



Now the continuous absorption of light in cool stellar atmospheres 

 is due to a negative ion of hydrogen, H", which consists of a proton 

 surrounded by two electrons and absorbs light by losing the extra 

 electron to form ordinary atomic hydrogen. The number of H" ions 

 in the atmospliere itself depends on the rate at which neutral hydrogen 

 atoms can capture free electrons, and these electrons are supplied in 

 turn by ionization of the metals; the degree of ionization of the 

 metals is here about a half, so that there is one electron for every two 

 metallic atoms. If we now compare a dwarf and a subdwarf, the 

 subdwarf having fewer metal atoms by a factor of a hundred or so, we 

 see that the subdwarf has not only fewer metal atoms capable of pro- 

 ducing a dark absorption line, but also fewer electrons to provide 

 general opacity in the atmosphere, so that the subdwarf atmosphere is 

 considerably more transparent. Consequently the metallic absorp- 

 tion lines do not become fainter to the same degree as one might nor- 

 mally liave expected, certainly nothing like the factor of some hundreds 

 shown by the abmidances, and even the weakening that does occur in 

 the metallic lines can be more or less got rid of by choosing your sub- 

 dwarf at a lower surface temperature ; you will recall from the picture 

 of the spectral sequence how rapidly the intensities of the dark lines 

 increase with diminishing surface temperature toward the end of the 

 sequence. 



These points are strikingly displayed by a series of spectra of K- 

 type dwarfs taken by Olin C. Wilson at Mount Wilson and Palomar 

 Observatories (pi. 7). 



The interpretation that I have placed upon these spectra, which is 

 not necessarily accepted by other people, is along the lines of what 

 I have just said. In each pair, the lower spectrum represents a sub- 

 dAvarf with fewer metal atoms in its atmosphere than are present in 

 the normal dwarf shown above, though not by the large factor of 100 

 that I have been quoting up to now. The difference is probably by 

 a factor of 5 or 10. The subdwarf, however, which is marked "b," 

 has a more transparent atmosphere than the normal dwarf marked 

 "r," so that the absorption lines due to metals look about equally strong 

 in the two spectra. However, the lines due to hydrogen — H gamma 

 and H delta— are quite different as can be seen ; they are stronger in the 

 spectrum of the subdwarf because its atmosphere is more transparent 



