96 DRS. BALM ON 
TaBueE I. 
1901-2. 1903. 
Heliographic | Linear Velocity | No. of | Heliographic | Linear Velocity | No. of 
Latitude. per sec. Obs. Latitude. per sec. Obs. 
km. km. 
Gis 1:908 33 ule 1:898 16 
5-9 1°894 30 9:1 1°883 15 
8:3 1871 30 15°5 1831 16 
12°8 1:802 30 21:1 1:753 15 
18:0 1:720 30 Deer 1°631 16 
23°8 1:594 30 34:0 1512 16 
30°6 1:488 30 40°3 1°365 15 
38°7 1-265 30 47:1 1-201 16 
NGO 1:061 30 54:7 1-014 16 
55°8 0:840 30 62°3 0-797 15 
65:1 0-560 30 69°2 0°564 16 
75-5 0°307 30 76°0 0°408 15 
83:0 0187 14 
The values of the linear velocities were now plotted upon squared paper as ordinates, 
with the heliographic latitudes as abscissze, and in each group a curve was drawn repre- 
senting, as closely as possible, the observed ordinates. In this graphical form the results 
are represented in the accompanying fig. 3. It is seen at first glance that the two 
groups differ materially from each other. They agree fairly well only at the equator, 
but in higher latitudes the observations of 1903 show considerably larger values than 
those of the two preceding years. There is not a single exception to this rule. The 
evidence of these curves leaves therefore no doubt as to the reality of this systematic 
difference, which is also apparent from a consideration of the probable error. An in- 
vestigation of the observations of 1903 shows the probable error of a single observation 
to be +£0°070 km., a value which appears to be very nearly constant for all helio- 
graphic latitudes. From this we find the probable error of a mean value for the first 
eroup +0°013 km., and for the second group +0°018 km., while the observed 
systematic differences between the groups attain the considerable value 0°16 km. in 
middle latitudes. 
Professor Dun&R has also computed the probable error of his observations. I find 
from his figures that of a single observation to be +0°138 km., also nearly constant for 
all latitudes. Considering that his single observations comprise twelve to twenty-four 
pointings of each line, as compared with only eight in my observations, we may conclude 
that the latter show at least double the accuracy of those of Professor Dunir. As 
already pointed out, this favourable result is in my opinion only due to the ereater 
stability and the more comfortable management of the instrumental plant. 
The influence of a possible error in the position angles shown by the heliometer has 
