We report submilliarcsecond-precise astrometric measurements for the late-type star AB Doradus via a combination of VLBI (very long baseline interferometry) and HIPPARCOS data. Our astrometric analysis results in the precise determination of the kinematics of this star, which reveals an orbital motion readily explained as caused by gravitational interaction with a low-mass companion. From the portion of the reÑex orbit covered by our data and using a revised mass of the primary star (0.76 M _) derived from our new value of the parallax (66.3 mas \ n \ 67.2 mas), we Ðnd the dynamical mass of the newly discovered companion to be between 0.08 and 0.11 If accurate photometric information can M _. be obtained for the low-mass companion, our precise mass estimate could serve as an accurate calibration point for di †erent theoretical evolutionary models of low-mass objects. This represents the Ðrst detection of a low-mass stellar companion using VLBI, a technique that will become an important tool in future searches for planets and brown dwarfs orbiting other stars.

The expected distributions of eclipse-depth versus period for eclipsing binaries of different luminosities are derived from large-scale population synthesis experiments. Using the rapid Hurley et al. BSE binary evolution code, we have evolved several hundred million binaries, starting from various simple input distributions of masses and orbit-sizes. Eclipse probabilities and predicted distributions over period and eclipsedepth (P/∆m) are given in a number of main-sequence intervals, from O-stars to brown dwarfs. The comparison between theory and Hipparcos observations shows that a standard (Duquennoy & Mayor) input distribution of orbit-sizes (a) gives reasonable numbers and P/∆m-distributions, as long as the mass-ratio distribution is also close to the observed flat ones. A random pairing model, where the primary and secondary are drawn independently from the same IMF, gives more than an order of magnitude too few eclipsing binaries on the upper main sequence. For a set of eclipsing OB-systems in the LMC, the observed period-distribution is different from the theoretical one, and the input orbit distributions and/or the evolutionary environment in LMC has to be different compared with the Galaxy. A natural application of these methods are estimates of the numbers and properties of eclipsing binaries observed by large-scale surveys like Gaia.

Aims. The magnitude differences for Hipparcos visual binaries give important information about the upper part of their mass-ratio distribution, and in previous studies, the author found a narrow peak at q = 1 This excess of equal-mass, wide pairs has not been confirmed in other studies, and the present work aims to strengthen the Hipparcos results. Methods. We construct a Galaxy model with binaries of known properties, filter it through a "Hipparcos-like" selection model, and then compare the model with the actual Hipparcos observations. By changing the input distributions in the model, the fit to the observations can be improved, enabling some conclusions about the mass-ratio distribution f q (q) and the distribution of semi-major axes f a (a). The most important weak point is the modelling of the selection effects in the Hipparcos Input Catalog, but by testing different biases, typical effects can be quantified. Results. From fittings in different color(mass)-intervals, I find that the shapes of the distributions of orbit-sizes and mass-ratios vary with mass, even in the small (1.9−5.5 Msun) mass-sum interval well covered by the Hipparcos data. In particular, the excess of "twins" in f q (q) is definitely present, but the peak amplitude diminishes at higher masses. There is normally also a decrease in the number of binaries with increasing mass-ratio in the interval 0.5 < q < 0.9, but below 2 Msun, this slope suddenly reverses, giving a broad peak towards 1.0 below the narrow one. For typical wide orbits (100−1000 au), the f a (a)-distribution shows a steeper decline for lower masses, while almost reaching the scale-free 1/a power-law for the highest ones.

No abstract