7 Conclusions

As already stated before, the usage of a 1m telescope equipped with a professional CCD camera allows to obtain high quality relative astrometric data on double stars. Mean errors are 0.01 $^{\prime \prime }$ in angular separation and 0.07$^\circ$ in position angle. This has been obtained by applying a strict observational protocol throughout the various campaigns, notwithstanding many instrumental changes. We have obtained for a large sample of double stars observed in an all-sky mode standard V and I magnitudes as well as (V-I) colour indices with precisions in the instrumental system of 0.006 and 0.009 mag respectively for the brightest companion and 0.009 and 0.024 mag for the faintest one. Accuracies (external errors obtained through comparison) are generally below 0.03 mag but degrade quickly for the smallest separations observed ( $\rho < 3\hbox {$^{\prime \prime }$ }$). Comparison of our data with the photoelectric photometry in the UBV, Geneva and Strömgren systems shows a systematic difference of one hundreth of a magnitude on the total V magnitudes. The same comparison illustrates very well the degradation (higher scatter of the differences) coupled with larger angular separations and the larger differential magnitudes. For systems with angular separation between 5 $^{\prime \prime }$ and 12 $^{\prime \prime }$,  CCD photometry appears to be more adequate than photoelectric photometry: systematic effects appear due to the chosen diaphragms which are generally too small with respect to the observed separations and magnitude differences. For systems of even smaller separation and with almost equally luminous components, we found another (small) systematic effect for which however we still have no explanation. This effect implies that total V magnitudes are slightly fainter when measured with CCD photometry than with photoelectric photometry. The measurement of the difference $\Delta(V-{\it i})$ is an efficient tool to detect subgiant or giant companions among double stars. The analysis of the colour differences shows that 40 systems have a negative $\Delta(V-{\it i})$. From these, 17 out of 23 systems with known spectral types and with $\Delta(V-{\it i}) < -0.1$ may be in an advanced stage of evolution, i.e. 74% comprise a subgiant or giant primary component. Such systems are of particular astrophysical interest and deserve further attention.

Acknowledgements

We gratefully acknowledge the allocation of telescope time by the European Southern Observatory during the full length of the key programme. We appreciate the support of and the many discussions with our colleagues Drs. J. Cuypers (Brussel), D. Sinachopoulos (Athens) and W. Seggewiss (Bonn). We thank the referee for helping to improve this paper with several detailed comments and judicious remarks. The network Réseau Européen des Laboratoires: Étoiles Doubles Visuelles was supported in 1992 by the French Ministère de la Recherche et de la Technologie. P. Lampens also acknowledges funding by project G.0265.97 of the Nationaal Fonds voor Wetenschappelijk Onderzoek (België).