320 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 63 



To make what use they can of the minute amounts of light that we 

 get from most stars, astronomers have developed their primitive- 

 sounding methods of photometry through colored glass filters and 

 spectral classification from spectra taken with low resolving power; 

 nowadays spectral classification is often done in a more quantitative 

 way by isolating a narrow band of the spectrum containing the absorp- 

 tion line one is interested in with the aid of an interference filter or 

 a spectrometer, but the principle is still not so very different from the 

 older and cruder method of looking at the spectrum under a micro- 

 scope and saying that line A is stronger than line B but weaker than 

 line C. All these methods can be broadly described as methods of 

 stellar classification, and they have two great advantages. The first 

 advantage is that relatively little light is required and so one can 

 examine stars, clusters of stars, and even galaxies at very great dis- 

 tances; and methods have been developed of estimating the age of a 

 galaxy from the color distribution or spectrum of the galaxy as a 

 whole, without even looking at the individual stars belonging to it. 

 The other advantage of such methods is their rapidity. Naturally 

 if the measurements are comparatively simple, one can observe a large 

 number of stars in a reasonably short time and so by now extensive 

 lists are available giving spectral types and color indices for thou- 

 sands of stars. On the other hand, the experience of the last few 

 years has shown that one can quite easily be fooled by the results of 

 these classification methods, because a given result can come about 

 from different causes. An example of this is that the color and the 

 spectral type of a star depend on its chemical composition as well as 

 on its surface temperature; and another example wliich complicates 

 the issue further is the fact that, if the distance of the star is not 

 known at all, we still have the problem of deciding just how bright 

 it is. 



The moral of this is that the extensive lists of comparatively simple 

 observations on many stars have to be supplemented by an intensive 

 attack on a relatively small number of cases by using the more power- 

 ful but laborious approach of taking spectra with as high a resolving 

 power as possible and examining the weak lines as well as the strong 

 ones. This has been done so far for no more than about 100 stars, 

 which is far too few, and of course there are difficulties, in particular 

 the fact that it requires huge telescopes to collect enough light for 

 the purpose, even for the study of most of the comparatively nearby 

 stars. A very hopeful technical development in this direction is that 

 of image converter or image intensifier tubes, which have already 

 been brought to a considerable degree of perfection by Professor 

 Andre Lallemand of Paris Observatory, and which promise to be 

 about 100 times as sensitive as ordinary photography. Even this 

 development, however, does not mean that we shall be able to do 



