In order to obtain accurate proper motions (better than 3 mas/a
for a single star) it was clear that only first epoch plates
from earlier than 1940 from telescopes with focal lengths
larger than 2 m could be used. Another constraint for the first epoch
material was
that the available field should cover not less than 1.5 
1.5 
. The use of such old material implies
that only fields with bright (V<16) extragalactic objects
were suited. In the beginning we also restricted our search
mainly to fields with (optical) pointlike objects, for which also
an accurate radio position from VLBI was available. Our search
for bright candidates resulted in a first list of candidates,
which was published by Brosche (1980). In the following years we have
made an extensive search for first epoch plates in
different plate vaults mainly in
Europe starting with the plate vault of our double
refractor of Bonn (now located at Hoher List observatory).
Especially the quoted brightness of the objects
turned out to be quite uncertain. One of the reasons
is the strong optical variability of some of the objects.
First epoch plates of about
fifteen candidates were detected and partly measured, but only
for six objects the material seemed promising
for a complete measurement and reduction. A list of all
link fields used in
this project is given in Table 1 (click here). Moreover, plates of the
optical extended objects M 51 and M 94 were found in our own
archive, the M 51 data were later analysed with additional
observational data and a refined method of measurement and
reduction (Odenkirchen & Brosche 1995).
In addition to the fields with one extragalactic object we used data from the Bonn globular cluster program for the determination of absolute proper and space motions of globular clusters (e.g. Geffert 1987; Brosche et al. 1991b; Geffert et al. 1993).
Table 1: The fields for the calibration of the Hipparcos proper motion
system. Column 5 denotes the method of the
extragalactic calibration of the proper motions
(see Sect. 4)