194 THE CHEMISTEY OF THE SUN. [CHAP. 



as Mitscherlieh, PI ticker, and Angstrom, and later of Salet and 

 Dr. Schuster, not to mention others, I am aware that though 

 there is a general consensus among spectroscopic workers that 

 double spectra cannot be ascribed to impurities, that consensus 

 is not pe haps even yet quite absolute. 



4. The Variations of Spectral Lines. 



We now approach details of another order details which have 

 a wide range. I refer to the different intensities of lines ob- 

 served in terrestrial light-sources under the different conditions 

 in which the phenomena occur, and in different celestial light- 

 sources. It is not my fault that for purposes of definition I 



FIG. 75. The varying intensities of the lines of Calcium as seen under different 



conditions. 



have to use language inconsistent with the received views of 

 spectrum analysis twenty years ago, when spectra were sup- 

 posed to be as changeless as the laws of the Medes and Persians. 



Let us first consider the facts in the case of calcium. 



The accompanying diagram will give an idea of the facts 

 observed and of the enormous variations exhibited. We can 

 readily differentiate the lines, and further see how great the 

 changes are when we pass from the laboratory either to the 

 general spectrum of the sun or to the spectrum of any part of 

 it. Here we have the reason why I have chosen to start with 



