67 



moon is absent, ma}^ be inferred from the foct that when 

 either of the phmets Venus or Mercury is about to cross 

 the sun, it can be seen before touching the disk — a dark 

 body on the coronal background. In eclipses also, a few 

 minutes before or after the total phase, the whole outline 

 of the moon can be seen. It is impossible to believe 

 that the moon could have an atmosphere of sufficient size 

 and density to cause the phenomena witnessed in the 

 corona, and 3^et this atmosphere not be apparent in many 

 other ways. 



The examination of the corona by instrumental means 

 strengthens the belief in its solar origin, at least in the 

 solar origin of the portions nearest the sun. The obser- 

 vations with the polariscope in different eclipses have been 

 quite conflicting. In some, no traces of polarization 

 could be found, in others they were ver}^ marked, while 

 the direction of the plane of polarization varied. A dis- 

 cussion made previous to the last eclipse by an English 

 astronomer showed that the greater part of the observa- 

 tions might be harmonized on the supposition that the 

 corona contained reflected sunlight, which might be caused 

 by meteoric or other matter in the vicinity of the sun. 

 The observations made last summer for the most part con- 

 firm this opinion. But there are certain observations in 

 several of the eclipses which indicate that part of its light 

 comes through certain substances and not by reflection, 

 since the plane of polarization did not pass through the 

 sun but tangent to it. This discordance in the polaris- 

 copic observations has not yet been explained. 



Let us examine the testimony of the spectroscope. 



[Upon the screen were shown the spectrum of the 

 sun, the solar spectrum together Avith that of iron show- 

 ing the coincidences of sixty-five of the bright lines in 

 the spectrum of iron with dark lines in the solar spec- 



