WRIGHT— RECENT SPECTROSCOPIC OBSERVATIONS. 519 



means we have of studying their physical conditions is by spectro- 

 scopic analysis of their light, and we must therefore look to that to 

 furnish the greater part of the material for whatever study can 

 be made of the processes of stellar change. 



Among the spectra of the heavenly bodies there is the greatest 

 diversity ; still, as Secchi first showed, they can nearly all be segre- 

 gated into a comparatively few fairly homogeneous groups. While 



Fig. I. 



the general validity of Secchi's classification is recognized , it is not 

 sufficiently comprehensive nor exact to describe the great variety of 

 spectra that have, in recent years, been made available for study 

 through the aid of photography. The system referred to here will 

 be that of the Draper classification, developed at the Harvard Col- 



1 The writer wishes to acknowledge his indebtedness for this illustration, 

 excepting the first spectrum, to the Draper Catalogue (Ann. H. C. O., 91), 

 from which it is copied and where it serves as frontispiece to illustrate typical 

 stellar spectra. 



