Table 2 gives positions and proper motions of 48 stars from our measurements. As mentioned above the system of the proper motions is defined in each field independently by the link procedure. Although these data were mainly provided for the calibration of the Hipparcos proper motions, they may be also used for the calibration of any astrometric catalogues, especially also for future satellite projects like DIVA (Bastian et al. 1996) and GAIA (Lindegren et al. 1994). All of our data were provided to the working group on the extragalactic link of the Hipparcos system. Besides the fact that a complete discussion is given in the papers by Tucholke et al. (1997) and Kovalevsky et al. (1997) a rough estimate of the accuracy of our data is given as follows.
Table 3: Mean estimated uncertainty of our proper motions
determined by the comparison with H37 stars
We have compared our
final proper motions with the corresponding ones from the
H37 catalogue. We obtained a rms of the differences
of 3 mas/a, which in the worst case may be taken as an
upper limit of the accuracy of our data. For this
comparison we were able to use data of 47 common stars.
Due to a possible rotation of the H37 system
with respect to our data, a complete comparison has
to include a rotation of the two systems. We have applied
a program given by Sinachopoulos (1986), which
has been used before for the determination of systematic
differences between radio and optical positions (Brosche
& Sinachopoulos 1987). A modification of that program
was also the base for the calculation of the rotation of
the Hipparcos data for all Bonn fields. Since the rotation
angles of the solutions are discussed elsewhere we will
restrict ourselves here to the accuracy of our data,
which may be estimated by the rms of the deviations.
The rms of the comparison including a possible
rotation was 2.8 mas/a for all 47 stars. This result
improves to 2.5 mas/a by omitting 6 outlying stars
on the 3
level. Taking only
the fields which had been reduced to the H37 catalogue
improves the result slightly to 2.3 mas/a.
Table 3 (click here) gives the results of our comparison.
Since we expect that the accuracy of our data will
be mainly responsible for these deviations, Table 3 (click here)
gives mainly the uncertainty of our proper motions.