1 Introduction

For centuries, visual double stars have been studied with the goal of determining stellar masses and the statistics for binary and multiple star systems. Traditionally, such systems have been studied visually using a micrometer on a long-focus instrument. In this way separations and position angles can be measured with high precision. In more recent work (e.g., Mason et al. 2000a) speckle interferometry has successfully replaced the eyepiece micrometer.

The Washington Double Star Catalog, WDS, (Mason et al. 2001b) is the official IAU double star data base. At the time when we initiated the present project (July 2000), it contained $83\,287$ pairs of stars. Due to the nature of the catalogue, it gives only approximate coordinates; the RA, e.g., is given with a resolution of 0.1 min. This can lead to difficulties in identifying the systems or individual components unambiguously in other catalogues and on the sky. It is therefore desirable to provide accurate absolute coordinates for each component of a system. Accurate positions are also important for determining the mass ratio in orbiting pairs and for understanding the nature of a system, because proper motions can often tell a physical component from an optical.

The smaller Catalogue des Composantes d'étoiles Doubles et Multiples, CCDM (Dommanget & Nys 1994) gives more accurate coordinates, but contains only $34\,031$ systems, and we have therefore not used it in the present work. A new version of the CCDM is being prepared (Dommanget 2000), but will not include Tycho results.

The Hipparcos Catalogue (ESA 1997) includes an annex of components of double and multiple systems with very accurate astrometry for $12\,195$ systems of which about 3000 were new systems.

The Tycho-2 Catalogue (Høg et al. 2000b), contains no double star annex, but in an analysis of the Tycho-2 double star solutions, Mason et al. (2000b) found 1234 new systems, which were included in the WDS (Mason et al. 2001b), but 14 of these are left out of the Tycho Double Star Catalogue, TDSC, because they were found very uncertain. Also many known systems were extracted from the Tycho-2 Catalogue for insertion into the WDS. As described in Mason et al. (2000b) these came from two sources:

• 7485 Tycho-2 double star measures result from a special processing called "double star treatment'' (Col. 201 of the Tycho-2 Catalogue);

• 5283 double stars come from more widely separated systems given a "normal treatment''.

The identification of doubles in Tycho-2 was based on a rather narrow, but safe, search window and several thousand Tycho-2 doubles were left behind.

When the Tycho-2 Catalogue was prepared, the rather slow duplicity analysis was only carried out for a small subset of the stars, suspected to be non-single. We have now looked at all Tycho-2 stars and resolved $17\,000$more systems, including about $13\,000$ discoveries. The large number of new systems has been reached not only by analysing more stars, but also through allowing solutions with smaller separations than the limit of 0.8 arcsec used in Tycho-2. It should be noted that for the smallest separations, the risk of spurious detections or large errors in  $\rho, \theta$ is increased.

Having noticed that the identification of double stars is often difficult, and that the WDS at the time did not include all Tycho results, we decided to make the Tycho-2 doubles more easily accessible. We have therefore used this opportunity to include in the present TDSC, not only the newly discovered systems, but also the doubles already in the Tycho-2 Catalogue. All double star information from the TDSC will be included in the WDS.

For most double stars it takes quite a few measures spanning several years to classify them as having motion which is either Keplerian or rectilinear; typically ten is the minimum. Of the systems observed by Tycho, $13\,800$ of them have between two and nine measures, other than the Tycho measure. For these, the Tycho measure will aid in eventual motion characterization. Tycho also made measures of 6441 systems having between 10 and 50 historical observations. It is usually these which would be considered as the "ripest'' for orbit analysis. Of the (now) $84\,486$ WDS systems, only 1430 (less than 2%) have orbit determinations. Work on this continues. A small number (1509) of the Tycho double star measures are of systems which are well characterized (i.e., number of measures greater than 50). Many of these are quite useful in giving an independent assessment of Tycho double star measures.