about the mass ratio must be delayed ical third component is not only proved early stars can be enhanced by inclu- until a spectroscopic determination of q by the light-time effect or third light, but sion of interacting close binaries with will be available. It is planned to collect also by a time change of the orbital hitherto unexplored or uncertain system the necessary rv data in July 1991 at the inclination (Drechsel et al., 1989). parameters. The current programme not ESO 1.52-m telescope (with ECHELEC). only aims at the measurement and solu- It should be noted that in a review of lU Aurigae tion of eclipse light curves, but will also early-type binaries by Hilditch and Bell complement radial-velocity data (1987) two objects with very similar pa- IU Aurigae (BOVp + B0.5, P = Id. 81 1) necessary for independent spectro- rameters appear: V Puppis (BIV, P = is a semi-detached system with the sec- scopic determinations of mass ratios, Id. 495) and TX Aurigae (BIV, P = ondary filling its critical Roche volume. which are necessary for reliable photo- 1910); these are semi-detached sys- As in the case of AH Cephei, IU Aurigae metric results. An essential subject of tems, with secondaries filling their is a rare case for which the presence of future investigation will be an adequate Roche lobes. The detached system a third body can be confirmed beyond treatment of radiation pressure effects, V1331 Aquilae therefore seems to be an any doubt by light-time effect, third light which have already been shown (Drech- important addition to the sample of the of about 20 per cent, and in addition by sel et al., 1991) to be of major impor- earliest binaries. In the following, a few the time variation of the orbital inclina- tance for the shape and configuration of more complex interacting binaries are tion due to the precessional motion of early-type close binary stars. briefly described, for which recent solu- the binary orbit triggered by the third Even at the end of the pre-VLT era tions were derived in the scope of this body (Mayer and Drechsel, 1987). and certainly also in the future, bright programme. stars are far from being exhaustively explored. For good reasons, large in- L Y Aurigae struments are not available for such ob- LY Aurigae (09.5111 + 09.5111, P = jects, while medium- to small-size tele- AH Cephei 4d. 003) is a massive contact system. A scopes are still able to provide a wealth AH Cephei (BOV + B0.5V, P = 1 d. 775) close visual field star with an angular of important new data - especially if is an early-type close detached system. separation of 0.5 arcseconds con- they are located at such marvellous Light curves from two widely separated tributes appreciable third light to the sites like La Silla. epochs (1930 and 1984) with clearly dif- total flux. This might be one reason for ferent depths of eclipse minima were partly contradictory previous results analysed. The derived time variation of concerning the mass ratio and system the orbital inclination suggests the pres- configuration. The recently obtained ence of a third body in the system, light-curve solution yields a third-light References which manifests itself not only by light- contribution of 10 per cent and a mass Drechsel, H., Lorenz, R., Mayer, P.: 1989, time effect, but also by the precession ratio of about 0.6, which is compatible Astron. Astrophys. 221, 49. of the orbital plane of the eclipsing bi- with the spectroscopic value (Drechsel Drechsel, H., Gayler, S., Lorenz, R., Mayer, P.: 1991, AG Abstract Ser. 6, 74. nary. This finding is further confirmed by et al., 1989). Hilditch, R.W., Bell, S.A.: 1987, Monthly Not. a fraction of third light of about 5 per Roy. Astr. Soc. 229, 529. cent, as is evident from the solution of 4. Future Prospects Kallrath, J., Linnell, A.P.: 1987, Astrophys. J. the light curves. Besides IU Aurigae, AH 313, 346. Cephei is a unique example for a triple We have shown that the sample of Mayer, P., Drechsel, H.: 1987, Astron. Astro- system, where the presence of a phys- absolute dimensions known for very phys. 183, 61 . Hunting the Brown Dwarf J, -M. MAR1077-I, DESPA, Observatoire de Paris, France, and C. PERRIER, Observatoire de Lyon, France and ends as a "failed star", faintly shin- achieved in astronomical instrumenta- 1. What are "Brown Dwarfs"? ing in the infrared due to the release of tion, specifically at infrared wave- During the seven last years there has gravitational energy associated with its lengths, which opened the door to the been a considerable interest developing progressive contraction. Observations possible direct detection of at least the in view of the discovery and observa- of some members of this new class of brightest, i.e. the youngest, brown tions of presumed sub-stellar objects, celestial bodies would of course be of dwarfs: high efficiency infrared arrays also named "brown dwarfs". This illus- the highest importance for the theory of and diffraction-limited imaging tech- trative term denotes a class of objects very low mass star formation, and mod- niques in the IR are prime weapons in that appears naturally in the theory of els have been proposed which aim at this hunting. star formation: recent models predict predicting the photometric and spec- trophotometric characteristics of brown that the collapse and fragmentation of a 2. First Attempts molecular cloud should produce clumps dwarfs and their evolution with respect down to 0.02 solar masses. Between to their birth mass and age. The first observers to spot something this lower limit and 0.07 solar masses, Another reason for the revival of this were McCarthy et al. (1985). They used the fragment is not massive enough to observational activity is of course to infrared speckle interferometry to detect ignite nuclear reactions inside its core be found in several breakthroughs brown dwarf companions possibly or- on-going northern hemisphere survey. Among others, our target list included some famous sources, such as Proxima Centauri, Barnard and Kaptein's stars, as well as some extreme low-luminosity stars such as Ross 154, Wolf 359 and some Van Biesbroeck's stars (VB4 and VB5). Alas! Our first observations turned out to be rather disappointing: not only not a single new sub-stellar companion ap- peared in our visibility curves (see Fig. I), but VB8B itself, the unique can- didate specimen of the class, proved to be only a mere artifact, caused by a tenuous calibration problem (Perrier and WOLF 359 0bj.K SM 86-03-21 ESO 3.57rn PA 90d 1232 Scans Mariotti, 1987): the number of detected brown dwarfs was reduced from 1 to O! Further observations in 1987 and 1988 did not allow us to invert this un- fortunate result. In the meantime, sever- al discoveries of new candidates were reported, but up to now none of them is clearly confirmed as sub-stellar, be- cause an unambiguous determination of the mass is always lacking. A by-product of our first observation campaign has been, however, to lead to new separation measurements of sever- al low-mass red-dwarf binaries, includ- .oo ing the determination of the masses and .OO .50 1.0 1.5 2.0 2.5 3.0 orbital parameters of Gliese 570B Frequency (l/arcsec) (Mariotti et al., 1990), a binary system independently discovered as a spec- Figure 1 : Long-exposure image and visibility of Wolf 359 at 2.2 microns. troscopic SB2 system by Duquennoy l.a: Superimposed East-West profiles of the source and of a reference star (dashed line) and Mayor (1988). Indeed, infrared enlarged 20 times with respect to peak value. The total scan amplitude is 18.4 second of arc. The r.m.s. noise value measured on the sky (right half of the scan) corresponds to a magnitude speckle interferometry turned out to be K = 13.2. an extremely powerful technique when 1.b: Corresponding visibility showing that the image core is not resolved: any companion with applied to the observations of binaries Kt11 and at a separation larger than 0.5 would produce a detectable sinusoidal modulation previously detected or suspected be- of the visibility. cause of their radial-velocity variations: biting the nearest red dwarfs in the northern hemisphere. The reason to search for brown dwarfs in binary sys- tems is that, unlike Lewis Carroll's Snark - that can be recognized thanks to "five unmistakable marks", one, and only one, parameter allows yet to confirm the sub-stellar nature of an candidate - brown dwarf, namely its mass. The reason to use speckle-interferometry, a technique achieving imagery at diffrac- tion-limited angular resolution, is that it - gives access within a distance of 10 parsecs to separations of the order of 1 A.U., and therefore to binary systems with typical periods of a few years. 3. Observations at La SiHa Figure 2: A sub-stellar companion of Gliese 440? Following these tracks, we have In January 1990, we detected a very faint and red object a few seconds of arc away from GI 440, one of the closest white dwarfs with a distance of 5 pc. Further observations in March started in 986 a Of Obser- 1991 revealed that the separation of the two objects has increased by about 2 seconds, i.e. vation of the 'Outhern hemi- exactly the amount and the direction due to the proper motion of GI 400: the "companion" is sphere stars with the infrared speckle hence not a brown dwarf, but rathera background "norma1"red star lying for a while on the line system available at the La Sills 3.6-m of sight.