Most stars are members of binary or multiple systems. Nevertheless, double stars are in general more poorly known from an astrophysical point-of-view than their "single" counterparts - partly, because the employed techniques introduce biases and selection effects and partly, because the interpretation of the data is more complex. A large part of the available information on binary stars comes from ground-based astrometric surveys made by visual observers during the 19^th and 20^th centuries, most of which were based in the northern hemisphere. This concerns close visual binaries with separations less than 1" as well as intermediate pairs with separations between 1" and 10". In the second half of the 20^th century, the photographic technique was used to complementarily study the motions of wide and even very wide binaries (separations up to 1'). Unseen companions of astrometric binaries as well as wide common proper motion pairs were thus detected. The first high-precision search on a large scale and of the complete sky was obtained by the space astrometric mission Hipparcos. From a systematic monitoring of a sample of 118.000 stars over 3 years, some 3000 newly resolved doubles and several thousands of suspected doubles have been detected. In the separation-vs.- difference in magnitude plane, the distribution of these new discoveries shows a high concentration in the practically unexplored regime below one arcsec (separation < 1", Hp < 4 mag)  (Fig. 1 in Lindegren et al., 1997, A&A 323, L53).
It was also the first time that such a vast material of differential magnitudes with a precision of ~0.10 mag has been obtained for double stars of close and intermediate separation. The results of the Hipparcos mission illustrate the impact of a systematic survey, even though quite limited in time. It also illustrates that the frequency of binary and multiple stars is probably still largely underestimated. A good knowledge of close visual binaries is needed for the determination of stellar masses and related calibrations while an improved determination of the distribution functions of true separations , relative motions, mass and luminosity ratios, temperature differences is necessary for the intermediate and wide binaries. The investigation of the fundamental properties of various classes of binary and multiple systems is indeed crucial to our understanding of the galactic content and its structure.

STRONG POINTS of the LMT
 

WHY THE LMT ?
 

SOME IMPORTANT FEATURES
 

DISTANCE LIMITS FOR OBSERVING ANY CLASS OF BINARIES
 

The formula:

Par = R(arcsec) / A(A.U.)

Par is the limit parallax for a given resolution R. A is the true separation between the components. At the detection level, A must be replaced by  (see hereunder).

There are two distinct levels:
 

=> RESOLUTION OF VISUAL BINARIES (with sub-arcsec resolution): a > 0.2"

Accuracy ~ probably better than one pixel: (between) 0.2" (and 0.4") = R

=> DETECTION OF ASTROMETRIC BINARIES:  >= 0.012"

Accuracy ~ 3 * Accuracy of astrometry ~ 0.012" = R

CONCLUSIONS FOR DOUBLE AND MULTIPLE STARS
 

 Astrometric binaries (with apparent semi-axes of the photocentric orbit  >= 0.012") can be detected and monitored if:

(arcsec) = (k - ). a(arcsec) >= 0.012",

where k = fractional luminosity and  = fractional mass.

With a mass ratio of 2-3, m ~ 3 mag, the factor (k - ) represents a maximum contribution of ~0.25. The apparent semi-axes, a, should thus not be smaller than 0.048". Orbital periods less than 25-30 years are relevant for detecting orbital motion. This concerns astrometric binaries in short-period systems up to a distance of 100 pc. At this distance, the sum of both masses should then be larger than 0.18 solar mass. For example, a system at 100 pc could consist of two equally massive components of 0.1 Mo with an apparent V magnitude of 21 mag or of two 0.2 Mo components (k=0.5;  m ~ 0). For a system at a distance of 50 pc, the same constraints imply that a >0.1 Mo stellar companion to a 0.2 Mo star (k=0.33; m ~ 2.9) or a > 0.18 Mo stellar companion to a 0.5 Mo star (k=0.26; m ~ 4.2) can be detected.