210 
2°, That the appearances of polarization are presented 
when light shines through a thin cloud, whether the said cloud 
be visible or invisible. 
3°. That this effect is not produced by clouds above a cer- 
tain thickness. (N.B. I estimate the thickness of a cloud by 
its darkness: the quantity of light it absorbs.) 
4°, I believe that the polarization of the corona which I 
detected was simply due to the intervening atmosphere. 
I am not perfectly certain of the absolute success of the 
black velvet arrangement. I think that there may have been 
some slight reflexion in the tube; had the light been stronger 
this might have had a perceptible effect; as the light was 
so much cut off by clouds, I believe that it was practically 
successful. 
I believe the light was quite sufficient and the instrument 
sufficiently sensitive to enable me to speak with great con- 
fidence when I determined that there was polarization, and to 
prove that when I failed to detect it it must have been very 
small. My determination of plane is certainly within 5° and 
probably closer. 
So much for my own observations. I now add a few words 
on the general results of the expeditions as far as they are yet 
gathered up. 
1. From the spectroscope we learn that the light of the 
corona contains the hydrogen lines and also the unknown line 
1474. 
2, The shape of the corona is not circular, but roughly 
quadrilateral; it is broken by indentations and by a conspicuous 
V-shaped gap towards the S.E. corner, seen not only at three 
stations in Spain but also in Sicily. 
3. The inner brighter corona, which is supposed to have 
a defined outer boundary separating it from the outer corona or 
glory, is probably an optical or subjective effect. 
4. The question of the polarization of the corona still re- 
