(in ) 
displacements of different lines range from 0 to more than 2 kilometers 
per second, so that the elements would move independently. And 
Mr. Evershed himself remarks, that the hypothesis of a radial move¬ 
ment in spots seems entirely out of harmony with the conclusions 
arrived at by Prof. Hale in his researches on vortices and magnetic 
fields in sun-spots. 
To my mind, the difficulties involved in Mr. Evershed’s explanation 
of the important facts which he has discovered are insuperable. 
Motion in the line of sight cannot be the cause of those systematic 
displacements. Neither can they be accounted for by pressure or by 
the influence of a magnetic field, without introducing very improbable 
hypotheses ad hoc. On the other hand I could not find any serious 
difficulty in describing the newly-discovered phenomena as mere 
consequences of the anomalous refraction of photospheric light in 
local rarefactions of the solar atmosphere. 
On the origin of the Fraunhofer lines. 
We considered a sun-spot as a region in which the density gradient 
regularly decreases in ail directions from the centre outwards, thus 
preserving the same sign unto the outer limit of the penumbra. 
Density gradients of the same order of magnitude are supposed 
to occur just as well outside the spots, at many places throughout 
the solar atmosphere, but there they are not arranged by the syste¬ 
matizing influence of an important vortex, and often change their sign. 
Now if the widening of the Fraunhofer lines in the spot-spectrum 
must be ascribed to anomalous dispersion in extensive regions con¬ 
taining a deep “depression”, why should not the width of the lines 
in the ordinary solar spectrum also be caused for a considerable part 
by dispersion in minor rarefactions and condensations ? 
At first sight the evidence seems unfavourable to this view. For 
one might contend that, if the R-light and the V-light were not 
really absorbed in the solar atmosphere, but only dispersed, such waves 
must leave the sun after all, be it at points very distant from those 
where they emerged from the photosphere, and could not, therefore, 
be wanting in the average sun-light. 
But this argument is incorrect. 
To understand this, imagine the whole solar atmosphere divided 
by surfaces into alveoles or cells in such way, that each cell encloses 
either a maximum or a minimum of density. The shape oi the cells 
*ill vary to the extreme. 
Let us first consider a nearly spherical cell enclosing a minimum 
(fig 7). All the light which, coming from the photosphere SS', enters 
