l8 Turner, Total Solar Eclipses. 



arrive at a conception of the corona as made up of a series 

 of shells one inside the other, corresponding to different 

 velocities. But this does not accord at all well with its 

 appearance ; it would give a stratified corona, whereas the 

 observed corona is distinctly radial in character. 



For reasons which would be tedious to give at length, 

 the evidence seems to indicate that variations in direction 

 of velocity must be very small, and that we may assume 

 the direction to be approximately vertical. The corona 

 near the sun is formed from loiv vertical jets ; that at a 

 distance, from jets which reach a great height ; and since 

 the former is so much denser, there must accordingly be 

 many more velocities of small size than of large. The 

 law of degradation of brightness is accordingly now 

 become a law telling us how many more small velocities 

 there are than large ones, and we accordingly return to 

 what was originally said about the magnitude of these 

 velocities. 



It may not really be the velocities themselves which 

 vary, but the force of the sun's attraction, as diminished 

 by light pressure. When the light pressure and attraction 

 nearly balance one another, a very slight change in the 

 adjustment wiil double the difference, and it is obvious 

 that we have here great possibilities for variation in mag- 

 nitude. When we introduce this variation into our 

 equations (see Note V.), we find a notable effect, some- 

 what of the kind required, on the distance to which the 

 particles are ejected. If we assume a series of attractive 

 forces, in the ratios lo, 9, 8, 7, 6, 5, then a velocity which 

 would, under force 10, eject a particle to the height of i 

 radius from the surface (or 2 from the centre — for com- 

 parison with our law it is better to measure from the 

 centre), would, with the forces 9, 8, 7, 6, eject it to 2'2, 27, 

 3"5, 6'0 radii from the centre ; while under force 5, the 



