Manchester Memoirs, Vol. I. (1906), No. 1. 17 



sphere (see Note III.) But the early promise of this 

 supposition is not followed up. Firstly, it would only give 

 us a corona one radius deep and the observed corona is 

 much bigger than this ; accordingly we must increase 

 our possible velocity so that particles may rise several 

 radii from the sun, and on coming to this more general 

 case we find quite a different law of distribution. Let 

 us take, for example, a velocity which would carry a 

 particle to the height of three radii vertically. If particles 

 were spouted from a point on the sun's surface in all 

 directions with this velocity, the densiiiy at the surface of 

 the sun comes out infinite as before ; at a little distance 

 it is finite but decreasing very slowly. The decrease 

 ultimately stops, and then becomes an increase ; and we 

 get anotJier shell zvith infinite density, due to the fact that 

 in the neighbourhood of a point on the other side of the 

 sun from the eruptive centre there is a great accumulation 

 of particles in a part of their orbits where they remain 

 for some time at nearly the same solar level (see Note IV.) 

 Outside this shell the density falls off at the rate we 

 desire, but fails altogether at a height of three radii. 

 This supposition accordingly is totally at variance with 

 the observed facts and we are led to the conclusion that 

 variations in direction of projection apart from variations 

 in amount do not give us an adequate explanation of the 

 corona. 



We must fall back on variations in magnitude of 

 velocity. But the work already done has not been thrown 

 away ; it has simplified the problem. We have seen that 

 with a given magnitude of velocity there is a certain shell 

 where the density becomes infinite compared with that 

 inside and outside, and hence in adding together the 

 results of velocities differing in magnitude, we need only 

 take account of this particular shell in each case, and we 



