314 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 63 



and stronger absorption lines. Therefore the first problem that 

 arises in interpreting the position of subdwarfs in the color-luminosity 

 diagram is to sort out the effect of temperature on the color from 

 the effect of metal abundance, that is to say we have to decide whether 

 star A is more blue than star B because it is hotter, or because it has 

 a lower abundance of the metals. For stars that are about as hot as 

 the Sun, this problem has been solved by making photometric obser- 

 vations through a third glass filter that lets tlu'ough the ultraviolet, 

 making use of the fact that comparison of the three different values of 

 the brightness obtained by measuring through the three different glass 

 filters gives an indication of the strength of the absorption lines and 

 hence of the metal content. 



Wlien we come to somewhat cooler dwarfs, with surface tempera- 

 tures of the order of 4,000° and corresponding to spectral type K, 

 the whole effect becomes more subtle and complicated, and it is this 

 section of the main sequence that we have recently been considering 

 from the viewpoint of the theory of stellar atmospheres. As the stars 

 become cooler, a high proportion of the metal atoms settle down into 

 their lowest energy states and produce numerous very strong absorp- 

 tion lines which overlap, so that in the blue part of the spectrum it 

 is no longer meaningful to speak of a continuous background at all. 

 Another difficulty is that there is little obvious difference between the 

 spectra of dwarfs and subdwarfs when they are as cool as this. 



One factor having an important influence on the intensities of ab- 

 sorption lines is the degree of transparency, or of opacity, of the stellar 

 atmosphere. So far, we have dealt only with the dark absorption 

 lines in the spectra of stars and said nothing about the continuous 

 backgromid light on which they are superposed, except to compare the 

 latter to the black-body type of radiation emitted by a heated poker. 

 In the 19th century, stars like the Sun were believed to be essentially 

 solid or liquid bodies, owing to their high density, and the continuous 

 radiation from them was believed to be quite analogous to the light 

 radiated from an ordinary solid or liquid heated to mcandescence. 

 The dark absorption lines are, of course, characteristic of matter in 

 the gaseous state, and so it was supposed that they were produced by 

 a gaseous atmosphere very similar to the atmosphere of the Earth, 

 only hotter of course. Nowadays we know that the Smi is a gaseous 

 body throughout, and when we talk about the solar atmosphere or 

 stellar atmospheres in general, we merely mean the layers of the star 

 which we can see because they are close to the surface, making no dis- 

 tinction between the layers where the continuous radiation comes from 

 and the layers producing the dark lines, since both are essentially the 

 same surface regions. This introduces some extra complication into 

 the problem of predicting how intense we should expect a given absorp- 

 tion line in the spectrum to be, since there will now be a competition 



