98 ; DR J. ‘'HALM ON 
our first point would be nearer the pole and the second nearer the equator than is 
shown by computation, and the arithmetic mean of both must therefore agree, with the 
actual latitude. If, then, we arrange, as has indeed been done in these observations, the 
measurements so that equal numbers of points are observed on opposite ‘sides of the 
poles, the mean values of the heliographic latitudes are practically uninfluenced by a 
systematic error of the position angle. The question, however, whether such an error 
actually exists, is readily answered by the observations. We have only to separate the 
Tig.4- showiny the valuss v- see B iw the liso periods 1go1-2 
and 1905 
observations into two groups in such a manner that the heliographic latitudes increase 
with the position angle in one group, and decrease in the other. The error would thus 
show itself by the fact that the same values of the velocity would not correspond to 
exactly the same latitudes in the two groups; or, to put it in other words, that corre- 
sponding latitudes should show slightly different values of the velocities. I have made 
such investigation for the whole time from 1901 to 1903, with the following result :— 
TaB.eE II. 
Heliographic | Group I. | Group II. 
Latitude. km. km. 
82°5° 0:21 0°22 
750 0°34 0-41 
65:0 0°65 0°64 
55'0 0°92 0-91 
45°0 1:18 Gtr 
35°0 1°43 1:42 
25:0 1°64 1:63 
150 UG LiTE Sits) 
50 1:89 1:89 
