( 1270 ) 
VoscEL’s observations on average violet light. This means, that for 
instance such waves as correspond to the places d, or d, of fig. 1, 
will reach a point M with an intensity that varies in a stronger 
measure with the direction in which they left the photosphere, than 
does tbe intensity of the average light. 
Those very waves (near d, or d,) are at the same time especially 
liable to changing their direction of propagation when there is a 
density gradient at M. If we were not concerned with anomalous 
refraction, but only with anomalous scattering, we should find 
that a definite kind of light would have a constant intensity, at any 
definite distance from the limb. But the refraction causes inequality. 
And the chance of observing a conspicuous variety of brightness, 
when Jooking at a region M of the solar atmosphere, which projects 
on the disk at a given distance from the limb, increases with the abso- 
lute value of n—1. By this circumstance we can e.g. explain a 
prominent feature of the spectroheliograms obtained by Hare and 
ELLERMAN '), viz. the gradual increase of the contrasts in a series of 
photographs of the same region, taken with waves, selected within 
K, at decreasing distances from A, 
§ 5. The co-operation of the two before-mentioned influences. 
The above discussion of the consequences of anomalous dispersion 
suggests the following interpretation of the variable, bright X,-line, 
generally supposed to be an emission line. 
There are two causes by which the brightness may increase on 
approaching A, . 
One of them depends on the presence of the two small (unequal) 
maxima of the R,-curve in fig. 1, indicating a diminution of the 
loss of light by scattering. This influence acts almost equally in all 
points situated at equal distances from the centre; it is strongest on 
the middle parts of the disk; for in proportion as we approach the 
limb, the scattered light itself will contribute a greater share in the 
total emergent beam, thus levelling the curve and reducing the 
importance of the two small tops. The part of the effect, due to this 
first cause, is weak any how, even on the central parts of the disk, 
but of a constant nature: it always yields a narrow, double K,-line, 
displaced toward the violet, and produces a displacement of A, 
toward the red. 
The other cause depends on refraction in irregular density gradients. 
On the middle of the disk, refraction can only result in diminution 
1) Hale and Evterman. The Rumford Spectroheliograph of the Yerkes Observatory. 
Publications of the Yerkes Observatory, Vol. III, Part I, 1903. 
