v ( 276 ) 
spot-Jines are not properly “widened absorption lines 4 -, but rather 
absorption lines, embedded in dispersion bands ‘). 
Not all Fraunhofer lines are- widened or strengthened in the spot- 
spectrum. The individual differences of behaviour I hope to refer to 
at a later date. For the present we have only to remember the fact, 
that the composition of the mixture of gases cannot be the same at 
all levels of the solar atmosphere. It therefore also depends on the 
level in which the local rarefaction occurs, which lines of the spot- 
spectrum will be enveloped in strong dispersion bands. 
We shall now treat of the case that a rarefaction is projected on 
the solar disk at a certain distance from the centre. 
The R-light must show a distribution of brightness reminding of 
that which we observed in' our experiments with salt drops, that is: 
it is strengthened at the side of the spot, turned towards the centre 
of the disk, and very weak at the opposite side. 
The V-light is curved the other way ; the distribution will resemble 
that shown in our experiments with glycerine drops, that is: it is 
strengthened at the side, turned away from the centre, and very 
weak at the side, facing the centre of the disk. 
These rules do not involve any special supposition as to the arran¬ 
gement of the density gradients with respect to their magnitude. 
We may therefore make our idea of the manner in which the 
R-hght and the V-light are distributed in an excentrically placed spot, 
more precise, by again introducing our previous assumption that in 
the spot-region the gradient decreases from the interior ontwards. 
As a simplification we suppose that region to be nearly spherical. 
Our figures 4 and 5 show the paths of the R-light and the V-light 
respectively. (Waves, not suffering anomalous refraction, will conform 
to the scheme fig. 4, as with them R m A m is positive, but the degree 
of refraction will be small; so they will produce a spot of which 
the umbra is smaller than the one represented in fig. 4). We see 
that with R-light the umbra is shifted towards the limb, whereas 
at the opposite side of the spot, in c, a bright place appears, because 
there emerges the light which left the photosphere perpendicularly. 
With the V-light, on the contrary, the umbra is shifted towards the 
centre, and the brightest light, emitted along the normal of the 
photosphere, emerges from p. 
Two spectroheliograms of the same typical spot, simultaneously 
*) Pl ‘°c. Roy. Amst. VII, p, 134—140; IX, p. 343—859. 
Astrophysical Journal 21, p. 271-291 (1905) and 25, p. 95—115 (1907). 
