As the plate covering the maximum overlap field of all plates around the cluster and at the same time being one of the deepest, plate No. 5821 was selected as reference plate in the plate matching and astrometric reduction. The proper motion determination had already been described in Scholz et al. (1996). The absolute proper motions were obtained by a plate overlap solution using about 1200 background galaxies as reference objects. Before that an extensive error removal technique was applied in order to exclude systematic positional plate-to-plate distortions of the different Schmidt plates and the known errors of the measuring machine (cf. Evans 1988; and Evans & Irwin 1995).
Figure 2: Calibration of the APM magnitudes a) on the V plate
5821 and b) on the B plate 8348 with photoelectric standards
from Arp (1962). A linear polynomial fit was used. Objects
with stellar classification are plotted as +, those with
non-stellar classification (not used in the fitting) are
plotted as 
A vector-point diagram of the absolute proper motions of all stars in the cluster region had already been included in the previous paper (Fig. 4 (click here)a in Scholz et al. 1996). The histogram of the total proper motion errors in dependence on the object classification is shown in Fig. 1 (click here). Except for the faintest stars and for the globular cluster stars in the central part of the cluster strongly affected by image crowding typical proper motion errors (peak in Fig. 1 (click here)b) of 2 mas/yr were obtained. The median of the distribution of proper motion errors for stellar images shown in Fig. 1 (click here)b is at about 4 mas/yr.
The deepest Tautenburg plates in B (8348)
and V (5821) have been used here for a two-colour photographic photometry
in order to compare the CMD with the astrometric
membership determination. The APM magnitudes are already internally calibrated
(see Bunclark & Irwin 1983) so that they are almost linearly related to
the standard photometric system. For the external calibration we used the
photoelectric standards from Arp (1962) well covering the magnitude
interval 11 < V < 19.5 of our study. Among the 21 identified stars from Arp
(1962) there were several objects classified as non-stellar on our plates due
to crowding effects. The accuracy of our photographic two-colour photometry
can be roughly estimated as 
0.1 getting worse at the faint end.