272 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1931 



atmosphere is at low pressure. When we combine this method with 

 what can be learned from the colors of the stars we find temperatures 

 ranging all the way from 1,600° in long-period variable stars at mini- 

 mum brightness up to over 30,000° C. for the hottest blue stars. 



The pressures in the atmospheres turn out to be surprisingly low, 

 and, although there are wide variations from star to star, the pres- 

 sures do not in general exceed one-thousandth of that of air at the 

 earth's surface. On some stars the atmospheres are at pressures 

 much lower than this. The entire atmosphere of the sun as far 

 down as we can see is about 50 to 100 miles deep, but there is so little 

 stuff in all this depth that it corresponds to only about 5 or 6 feet 

 of ordinary air. An amount of material which is absolutely trans- 

 parent here becomes so foggy on the sun as to be nearly opaque, be- 

 cause of the great number of free electrons and ionized atoms which 

 can stop passing light darts and send them flying off in other direc- 

 tions and even turn them into heat. 



I have said that there are great differences in the pressures at the 

 surfaces of different stars. Differences in pressure must mean differ- 

 ences in the force with which gravity packs down the material. And 

 differences in the force of gravity must mean one of two things: 

 Either differences in the masses of the stars or differences in the 

 distance from the surface to the center. We have already found 

 that the masses of all the stars are more or less alike, but that their 

 sizes do differ enormously. And so the pressure differences which 

 we can interpret from the line-intensities in the spectra of these 

 stars really mean differences in the sizes and therefore in the actual 

 brightness of the stars. Of course, we want to know the true candle- 

 power of as many stars as possible, in order that we may determine 

 their distances, which, as we have seen, may be done with the aid 

 of their apparent brightness. 



If we could really count accurately the number of calcium atoms 

 of the three kinds shown in Figure 4 by measuring the strengths 

 of the spectral lines which they produce, and if the simple theory con- 

 necting the numbers of the three kinds of atoms with the tempera- 

 ture and pressure were entirely satisfactory, we could, by studying 

 the spectrum of a star, calculate exactly what pressure must exist 

 in its atmosphere to give this spectrum and could go from that to its 

 true brightness. Some day we hope that it may be possible to do this, 

 but at present the results can be only approximate because the condi- 

 tions in stellar atmospheres are more complex than Figure 4 indicates. 



Fortunately a powerful empirical method can be used. In 1914 

 Doctor Adams and Doctor Kohlschiitter studied in detail the differ- 

 ences in the spectra of nearby stars whose true brightness was already 

 known. We have seen that the spectra of two stars may be very 



