from those of Taylor treated in such a way, that opposite three-hour-groups get equal weights (see (28) on page 37)
D = + 5i°.4
The correction A/uz necessary to bring the two determinations to agreement is found to be
(34 ) = + O".oo87
the common value of D being
(35 ) D = + 25°*3-
On the other hand the correction needed by both Newcomb and Taylor in order to reduce them to Auwers' system is
(36 ) = -f o".oo38 (88 st.)
which, though considerably smaller is still in the same sense. In deriving this correction from the values of fig of the stars common to Auwers and Newcomb a small supplementary correction has been applied to eliminate the influence of the difference of precession. The correction (36) leads to
(37 ) D = + 33°-2
for Newcomb's stars and to
(38 ) D = + 42°*i
for those of Taylor. The divergence from Bradley is thus greatly diminished. Lastly, in order to reduce the /ug of Bradley in this zone to Auwers' system, we have to add the value (36) to the difference (27). The value
A /ug — + o".ooi6
thus obtained leads to
(39 ) D = + ï9°-8.
The reduction to Auwers is thus seen to reduce considerably the divergences from the result (30) found from the whole of the Bradley-stars. The mean of the three values (37), (38) and (39), combined with the respective weights 1, £and 1 per star is D = + 30°-5, a value accidentally in nearly perfect agreement with (30) Zone — 520 to — 90°.
For the sake of completeness the stars of Auwers in the region between — 5 20 and — 90° have also been subdivided in two groups. From the larger proper motions we obtained
D = + 37°.o (101 st.)
from the smaller ones
D = -f 44°.6 (124 st.).