( 1267 ) 
the limb of the sun, we must fix our attention on another peculi- 
arity of the propagation of light through matter. Indeed, anomalous 
dispersion implies not only anomalous scattering, but also, wherever 
the density of the medium is unequal, anomalous refraction. 
Let us, for the present, leave out of consideration those “large-scale 
irregularities’, characterized by more or less ‘systematically arranged”’ 
density gradients (such as probably exist in sun-spots) ‘), and imagine 
the average condition of the solar atmosphere to be like that of hot 
gases above a fire, i. e. a complex of irregular density gradients 
strongly varying from point to point in direction, magnitude, and 
sign. A very extensive layer of gas, thus constituted, must in some 
way act as a turbid or scattering medium. The optical effect pro- 
duced by such an atmosphere will be comparable to what we observe 
when viewing a luminous surface through a plane-parallel glass tank 
in which, for instance, water and glycerine have just been stirred, 
but are not perfectly mixed yet. This “scattering by refraction” is, 
like molecular scattering, specially strong for kinds of light from the 
vicinity of absorption lines. So the effect of anomalous refraction is 
in many respects very similar to that of anomalous scattering, and 
will in so far only strengthen the latter. 
But we should not overlook the difference between the two 
processes. The intensity of the effects of anomalous refraction depends 
on the degree in which there happens to exist variety of density in 
the medium. So it may be quite different at different places on the 
solar disk *), whereas the intensity of the effect of molecular scattering 
is more equally distributed, only increasing gradually from the 
centre to the limb. And, secondly, we must notice that the direction 
of the density gradients may strongly influence the intensity of the 
light emerging from the solar atmosphere along a given line. This 
circumstance too eauses a kind of inequality in the distribution of 
the light, such as molecular scattering could not produce. 
We are now prepared to inquire into the changes, which the 
distribution of the light in general, and the aspect of the calcium 
lines in particular, must show when we pass from the centre to the 
limb of the solar disk. 
Those changes are of course closely connected with the fact, that 
in the central parts of the disk the main source of light has an 
almost symmetrical position behind the atmosphere, but not in the 
non-central parts. 
1) Proc. Roy. Acad. Amst. XII, p. 273, (1909). 
°) The irregular distribution of the light in spectro-heliograms can be explained 
n this basis. 
