8 
HAYF0RD. 
is being observed group No. 1 is also being observed, and 
during the latter half of the period group No. 3 is being ob¬ 
served at the same time as group No. 2. Similarly, group 
No. 3 is observed first in connection with No. 2, and then in 
connection with group No. 4, and so on. The difference be¬ 
tween the mean latitude from group No. 1 and the mean 
from group No. 2 during the period when both groups are 
being observed is obviously due to the difference of the mean 
error in declination of the stars of the two groups and to ac¬ 
cidental errors of observation. This method of observation, 
after being extended throughout the whole cycle until group 
No. 12 overlaps group No. 1, furnishes a means of determin¬ 
ing the declination correction to each group to reduce it to 
the mean of all of the groups, and thus to eliminate the 
declination errors from the computed change in latitude. 
As an additional precaution, the same list of stars is to be 
observed at all the stations. 
The effect of an error in the assigned proper motions of 
the stars observed will be to make the latitude of any one 
station appear to increase or decrease with lapse of time, but 
will have no appreciable effect on the value finally derived 
for the motion of the pole ; for, the same list being used at 
all stations, all the latitudes will appear to increase or de¬ 
crease together. Both the Japan and the Maryland station 
would appear, say, to have increasing latitudes, although 
they are nearly on opposite sides of the pole, and therefore 
this result could not be mistaken for an actual motion of the 
pole. 
Aside from the precautions already indicated against ab¬ 
normal refractions due to local conditions, the observations 
themselves have been planned so as to guard against errors 
arising from this cause. Each group contains six pairs 
which are used directly for computing the latitude variation, 
and of which very few have zenith distances exceeding 20 
degrees. Each group also contains two pairs introduced for 
the special purpose of studying the actual refraction, and of 
which the zenith distance is about 60 degrees. The normal 
