( 278 ) 
small, that there is no sufficient asymmetry in the supply of light, 
the inclination of the lines becomes imperceptible. 
With any other direction of the slit the effect must be less; t 
vanishes when the slit bisects the spot in a direction at right angles 
to a line joining the spot and the centre. 
Is it necessary to admit radial movement in sun-spots? 
Bulletin N°. XV of the Kodaikanal Observatory (Febr. 1909) 
contains a remarkable communication by J. Evershed, entitled: 
“Radial movement in sunspots”. The new facts there described 
chime perfectly and in every detail with the above necessary conse¬ 
quences of refraction. In the spectra of all the spots examined Mr. 
Evershed found the lines displaced according to the following law. 
Wherever a spot be situated on the solar disk — provided that the 
distance from the centre surpasses iO 3 — the majority of the Fraun¬ 
hofer lines of its spectrum are slightly inclined when the slit bisects 
the spot in the direction toward the centre.of the disk. The inclination 
is less with other directions of the slit, and disappears when the 
slit is at right angles to the sun’s radius. The displacement of the 
lines is always to the red on the side of the spot, turned towards 
the limb, and to the violet on tlie side facing the centre; it differs 
in magnitude with the individual lines 1 ). 
To explain these phenomena Mr. Evershed invokes Doppler s 
principle. He attributes the displacements to a radial movement 
outwards from the spot-centre. The motion must be essentially 
horizontal, or parallel to the sun’s surface; the velocity seems to 
increase, from the centre of the spot outwards, yet at the limits of 
the penumbra the motion apparently ceases abruptly. The lines of 
the true umbral spectrum seem usually to be almost entirely unaffected 
by motion in the line of sight. So there is not a kind of spring m 
the umbra, and we are puzzled with the question, whence the matter 
comes, which is permanently spreading out in the spot, and wheie 
it accumulates. Moreover, the maximum velocities indicated by the 
l ) Hp was not inclined. We presume that the anomalous refraction produced 
by the hydrogen of the solar atmosphere is too considerable to become apparent 
in the same way as the refraction caused by other elements. Even in the average 
solar spectrum the dispersion band of is more than one Angstrom unit wi e, 
in the spotspectrum it may perhaps extend so much farther, that one is 
to look upon it as not belonging to the line, and, therefore, to consider the i 
as weakened. Now, the narrow central line is the real absorption line Hp, w 1C 
of course cannot be displaced or curved by refraction. 
