4 Conclusions

Although we have mainly used already published data, we did revise the parallaxes and the individual masses of some of those systems. That was sometimes possible thanks to the integration of unpublished observations kindly provided by the authors (e.g., HIP 14328). However, just fitting published data more efficiently provided considerable improvement (e.g., HIP 677) as shown in Fig. 2. For some systems, we simply confirmed an available solution (e.g., HIP 104858). But, even in those cases, we usually supplied the first simultaneous orbital solution and, therefore, the first homogeneous standard deviations based on all observations. Usually, that yielded narrower confidence intervals, more consistent with the estimated errors on the observations.

Being hypothesis-free, orbital parallaxes are also useful to check the reliability of other determinations (e.g., spectroscopic parallaxes). For instance, the agreement between the Hipparcos results and those we derive is quite good.

Acknowledgements

I thank Patricia Lampens for her careful reading of a very preliminary version of this manuscript and for her valuable suggestions. I also thank my former colleagues from the Royal Observatory of Belgium and from the University of Liege for the fruitful discussions we had about stellar masses. Michel Mayor, Stephane Udry and Roger Griffin are thanked for supplying some radial velocities prior to their publication. This research has made use of the Washington Double Star Catalog maintained at the U.S. Naval Observatory and of the Simbad database operated at CDS, Strasbourg, France.