28 THE ROYAL SOCIETY OF CANADA 
The lines in the yellow green region were selected to include as many 
elements as possible among the limited number of measurable lines 
in the region. Some, such as the lines of Mn, Ti, Si, are not of very 
good quality for measurement, but were included in order to give 
evidence in regard to question c., Section 3 above. In the violet 
region No. 4to No. 13 inclusive, are the ten lines selected to be measured 
by all observatories co-operating in this work and the other five are 
lines which Adams and Lasby* found gave systematically higher or lower 
values of the rotation than the general reversing layer. The column 
“Velocity Constant” gives the half value of the multiplier required to 
reduce the millimetre displacement to kilometres per second, and will 
evidently give the observed velocity of the sun’s hmb. These multipliers 
are readily determined, in the well known way, when the linear dispersion 
at the region is known. As the grating gives practically a normal spec- 
trum over the narrow limits used, it is sufficient to determine this disper- 
sion, which is about 0.70° A. per millimetre at \5600 and 0.75 A at 
44250, for five or six lines over the region used. When these values 
and the multipliers are plotted on cross section paper they are found 
to lie within the errors of observation on a straight line, and the con- 
stants for all the lines measured can be at once read off. 
REDUCTION OF MEASURES. 
10. The observed or measured velocities are the radial com- 
ponents of the actual velocities at certain points on the sun’s disc 
whose latitudes can be readily computed, and it is hence necessary to 
know the angle of inclination between the radius vector and the direc- 
tion of motion at the point in order to apply the necessary corrections, 
the further correction for the motion of the earth in its orbit being 
made to obtain the sidereal rate. In the early observations, by Dunér 
and Halm, of the rotation of the sun by the spectroscopic method, 
the measurements were made at the limb and the computations and 
corrections were straightforward. When, however, as in Adams’ 
work and our own the observed points are some distance within the 
limb, the matter is not quite so simple. Adams’ method of reduction** 
depends upon projecting the observed points radially to the limb and 
obtaining the corrections by Dunér’s methods and tables, but this 
assumes the rotation of the sun to be that of a solid body, which is of 
course not the case. A further correction is therefore necessary for the 
difference in angular velocity at the observed and computed points. 

* Adams and Lasby—An investigation of the Rotation Period of the Sun by 
Spectroscopic Methods, p. 119. 
** Adams and Lasby, p. 13. 
