SPECTROSCOPIC OBSERVATIONS OF THE ROTATION OF THE SUN. 95 
seem perhaps somewhat pedantic, has certainly the advantage of referring the mean of 
the observations on each limb to one and the same moment of time. Besides, approach- 
ing a line alternately in opposite directions must decrease the amount of the personal 
error of the observer in his judgment of the bisection. After the completion of these 
measurements, the halves of the heliometer object-glass are reversed and the time noted. 
The observations are then repeated in the same order as before. A complete set there- 
fore consists of 32 pointings, 16 on each limb. When the set is finished, the heliometer 
is turned 10° in position angle, and the solar images are again brought into the position 
of fig. 2. The observation of a full set, including the necessary adjustments of the 
instrument, requires on the average from 10 to 15 minutes. 
The observations of Professor DunsEr refer to six equidistant points on the solar 
limb between the equator and 75° latitude. The selection of these fixed heliographic 
latitudes was perhaps necessitated by the arrangement of his instrument, which 
required a previous computation of the difference of declination between the north 
limb of the sun and the point to be observed. Since, however, a previous knowledge 
of the heliographic latitude is not required in our case, I have preferred to proceed 
from 10 to 10 degrees on the position circle of the heliometer, without considering 
at all the heliographic position of the points thrown upon the slit. The readings of 
the position circle, together with the time of the observation, are sufficient to evaluate 
afterwards the true heliographic latitude. In this way, a uniform and continuous 
distribution of observations over the whole quadrant, from the equator to the pole, may 
be secured, and we are, I think, in a better position to ascertain the character of the 
velocity-curve by this method than by the observations of six single points of the 
quadrant. 
After this general description of the instrument and the method of observation, I 
shall now turn to the results. The measurements were commenced on 13th August 
1901 and extended to 6th November 1903. - During this time 564 determinations of 
the rotational velocity were made. The values obtained, expressed in kilometres per 
second, were divided into two groups; the first group comprising the measurements 
made during 1901-2, the second those of the year 1903. The reason for this division 
is that the period 1901-2 was characterised by a low and protracted minimum of solar 
activity, while in the early part of 1903 the commencement of a new solar cycle was 
vigorously manifested by the appearance of large spots. It was therefore to be expected 
that, if indeed solar activity has an influence on the rotation of the sun, such a division 
into groups as I have made would show this influence more clearly than any other. In 
each group the individual values were arranged according to their heliographic latitudes, 
and from the materials thus collected normal values were formed by adequate combina- 
tions of single values into arithmetic means. The figures thus obtained are exhibited 
in the following table. 
