We present the results of a magnitude limited (K≤8.5 mag) multiplicity survey of T Tauri stars in the two nearest star forming regions, Taurus-Auriga and Ophiuchus-Scorpius (D = 150 pc), observable from the northern hemisphere. Each of the 70 stars in the sample was imaged at 2.2 μm using two-dimensional speckle interferometry resulting in a survey sensitive to binary stars with separations ranging from 0.″09 to about 2″.5.

The frequency of double stars with separation in this range is 46±8%. A division between the classical T Tauri stars and the weak-lined T Tauri stars shows no distinction. Furthermore, no difference is observed between the binary frequencies in the two star forming regions although the clouds have very different properties.

Given the limited angular separation range that this survey is sensitive to, both the spectroscopic and wide binaries will be missed. The rate at which binaries are detected suggests that most, if not all, T Tauri stars have companions.

Using infrared speckle imaging techniques, we have completed a comprehensive survey of all northern (δ ≥ −25°) M dwarfs within 8 parsecs for low mass companions. Of the 74 targets searched, six new companions were found. Included in the final census are four objects orbiting their primaries at sub-arcsecond separations which have masses near 80 Jupiters, making them viable brown dwarf candidates. Three of these — LHS 1047B, GL 623B and G 208-44B — are the faintest red objects for which masses have been determined and represent the limit of our current knowledge about the faint end of the mass-luminosity relation.

The complete sample includes 99 members, and under further analysis reveals fundamental facts about the red dwarf population that were unknown until the present study: 1) 30-40 % of M dwarf primaries have companions, 2) more companions are found orbiting 1-10 AU from the primary than in any other decade interval, and 3) there are 50% fewer red dwarfs known in the more distant half of the survey volume, presumably because the parallax and proper motion surveys are incomplete.

In addition, we find that the infrared luminosity function (LF) is fiat or rising toward the end of the main sequence, while the visible LF may be flat, and we illustrate that the determination of an accurate LF is critically sensitive to the resolution of binaries. A better description of the stellar population, the mass function, is found to be undoubtedly rising to the stellar/substellar break. Finally, we have developed a much-needed mass-luminosity relation for stars of mass 1.2 to 0.08 M⊙, and using these relations find that the M dwarfs contribute ~0.2 M⊙/pc3 to the galactic mass.

We surveyed all stars in Taurus (3h 45m < α < 4h 15m, 15° < δ < 35°) for multiplicity which are contained in the Herbig-Bell catalogue of young stars and have a 2 micron brightness of K ≤ 9.5 mag. This sample consists of 106 stellar systems (single or multiple), of which 43 are double or multiple according to the criteria of our survey, i.e. with separations of ≈0″.2 ≤ d ≤ 10″. Of these, 23 binaries are new detections found in this survey. The resulting degree of multiplicity, 43/106 = 41±6%, is higher than found for main-sequence stars. Provided that the period distribution is the same for young stars as on the main sequence, our result implies that the vast majority of stars are born in binary or multiple systems.

The pre-main sequence (PMS) binary frequency is a fundamental datum in the study of binary formation. It reflects on numerous basic issues, such as:

1. • The formation process. Binary stars are the primary branch of the star-formation process, and thus their frequency is an essential challenge to star-formation theories. (Indeed, the infrequency of single-star formation is likely as significant as the binary frequency.)

2. • The epoch of binary formation. Assessing whether the binary population exists in total by the pre-main sequence phase sets an upper limit on the binary formation timescale.

3. • Early period evolution. The frequency distribution as a function of period of PMS binaries, when compared to the distribution at the zero-age main sequence, can shed light on early orbital evolution.

4. • The interaction of binaries with disks. The formation and consequent dynamical evolution of a binary with semi-major axis less than typical disk radii must substantially modify disk structures and accretion flows. Thus the binary frequency might differ between PMS stars with and without associated disks.

Our survey for companions of the young stars now includes 45 systems in the Taurus star-forming region (SFR) and 21 in the Ophiuchus SFR. It is carried out by lunar occultation and imaging in the IR and can identify binaries in the separation range 0″.005 to 10″ in systems brighter than K=10 mag. The observed multiplicity in Taurus is ~1.6 stars/system which is comparable to that of the nearby solarlike stars but corrections for incompleteness increase the multiplicity to at least 1.8.

Inner active disks are equally represented among the single and multiple systems. The multiple systems have less massive outer disks than the single systems, but there are significant exceptions. The binary UZ Tau W contains a circumstellar disk or disks of mass ~0.024 M⊙ and size ~13 AU. The quadruple GG Tau system has a remarkably extensive circumbinary disk of mass ~ 0.07M⊙. These mass estimates are comparable to the minimum values required for the proto-planetary disk of the Solar System.

The specific angular momenta of the most widely spaced binaries in our sample adjoin the lowest values that can be measured for molecular cloud cores. The actual distributions probably overlap which suggests that the origin of the angular momentum of binaries is in their molecular cloud birthplaces.

We have observed a complete sample of southern Hα emission line pre-main sequence (PMS) stars associated with the nearby dark clouds Chamaeleon and Ophiuchus in an effort to detect binary companions. We have used the high resolution CCD camera SUSI (0.13″ /pixel) at the ESO 3.5m New Technology Telescope (NTT). The observations were carried out in the Gunn z filter (0.95μ). As a result, we have discovered 28 companions out of 160 objects surveyed in the range of angular separation 0″.5 – 10″. We present histograms of semi-major axes and of brightness ratios of these PMS binaries. Taking small incompleteness corrections into account, the frequency of PMS binaries with semi-major axes in the range 100-1500 AU is 20%, while it is only 15% in same range for solar-type Main Sequence (MS) stars. Extrapolation yields a total PMS binary frequency of 70%. Evolutionary and environmental implications are briefly discussed.

We present infrared imaging results of young stars in Cha I and II clouds that show either extended structure or a nearby neighboring star. Both regions appear to show high incidence rates of pairs. After excluding possible background stars, as judged from their brightness, color, or the local stellar surface number density, a frequency of 10% is deduced for binaries in Cha I with separations 2-5″, which is comparable to that in Taurus and to that of main sequence field stars.

High resolution (R ~ 20000) spectroscopic observations of visual pairs of T Tauri stars (TTS) in the spectral range 655-675 nm, and 385-405 nm for some systems, are reported. Good seeing allows us to resolve pairs with minimum angular separation of ~ 2″. The radial velocities, overall spectral properties and detection of the Li I line are used to decide whether the components of the binaries are likely to be gravitationally bounded. In this paper we present first results on a subset of our sample: 3 visual companions are not TTS, namely DL Tau/c, GG Tau/c(SW) and NTTS 45251+3016/c. The star GG Tau/c(S) is confirmed as a T Tauri star. The system UZ Tau is probably composed of at least 4 components. The star 1E0255.3+2018 (Fleming et al. 1989), previously thought to be single, is found to be a close visual binary. Finally, lithium abundances for the PMS components of 6 binaries are presented and their consistency with theoretical expectations is briefly discussed.

Interferometric observations of the 2.6 mm dust thermal emission around the T Tauri triple system UZ Tau show that most of it is equally divided between UZ Tau W (the close binary) and UZ Tau E. The emission is at least partially optically thick at 2.6 mm which implies an origin in disks of size ~ 13 AU and mass ~ 0.024 M⊙. The 2.6 mm emission of the GG Tau system, a hierarchical quadruple, is partially resolved. Strong emission extended over ~ 3″ × 5″ is associated with the close binary GG Tau. Weak emission is detected at GG Tau/c, also a close binary. Evidently extensive dusty disks can survive in the environment of close binaries for at least ~ 105 y, and their structure can vary markedly from system to system.

Current techniques for the detection of long-term, low-amplitude (<50 m s−1), radial velocity variations are briefly reviewed together with some of their most successful programs. In the era of 8- to 10-m telescopes we must strive for a precision of < 1ms−1.

Preliminary results are given for a systematic survey of K-dwarf stars in the solar vicinity. Nine companions to the G and K dwarfs have very low M2 sin i, less than 0.08 M⊙. These detections from a sample of 540 G and K primary stars support the reality of the existence of companions with mass below 0.08 M⊙: brown dwarfs exist.

A comparison of the relative mass function distribution f(m)/M1 between low-mass (M1 < 1.3 M⊙) and intermediate-mass (2 < M1 < 5 M⊙) binaries suggests a dependence of the mass-ratio distribution on the mass of the primary: f(q, M1).

By combining the photometric information with orbital elements constraints, we have derived the mass-ratio distribution for intermediate-mass stars. As is the case for low-mass stars, this distribution does not have any maximum close to M2/M1 = 1.

From a new study of 113 F8-G1 IV or V primaries in an apparent-magnitude limited sample, measured with a CCD for about 20 radial velocities of accuracy 0.2-0.5 km s−1 each, we collected data for 28 binary orbits and 39 visual systems. We again found a flat distribution (or slightly decreasing toward lower masses) of secondary masses, unlike the Salpeter distribution for B2-B5 stars. But we find that the different distributions for these field stars and for open clusters of various ages can all be explained by a capture mechanism of binary formation.

Over the past 16 years at the McDonald Observatory and the Kitt Peak National Observatory, I have obtained high-dispersion spectroscopic observations of roughly 3 dozen multiple systems and 7 speckle or astrometric binaries. The current status of this obviously long-term project is discussed. Although some systems have periods of about 50 to 85 years, most have periods in the range of 1 to 20 years. Thus, many now have been observed for at least one cycle and preliminary spectroscopic orbital elements have been determined. For a number of the systems fundamental data such as the distance to the system, as well as the individual spectral types, absolute magnitudes, and masses can be obtained if accurate visual/speckle orbital elements also are known. These elements and the derived fundamental parameters may be used as constraints on multiple-star formation theory.

There can be few subjects that benefit more than binary stars do from complementary approaches. Consider their orbits, for an initial example. By ‘visual’ methods — by which I mean all methods of obtaining angular resolution on the sky, including the modern techniques of optical interferometry — one can determine an orbit, but with the scale known in terms of angular measure only. By spectroscopic (radial-velocity) methods one obtains the linear scale of the orbit, but with those techniques the inclination to the line of sight is indeterminate. In order to find the complete characteristics of the orbit in three dimensions — as is essential in order to find those most important data, the masses of the component stars — one needs to utilize both methods. It is, accordingly, clearly of great value to have a meeting like this that is object-oriented rather than technique-oriented; it is a long time since there was a meeting specifically devoted to binary stars, and especially in view of the great advances that have taken place in observational techniques of all sorts in recent years it is more than timely that we should have this meeting now. We are certainly very grateful to Dr. McAlister and his colleagues for arranging it and holding it at this beautiful venue.

For more than a decade we have been measuring stellar radial velocities with three almost identical digital speedometers on telescopes in Arizona and Massachusetts. By now we have accumulated nearly 100,000 measurements with a typical precision of better than 1 km s-1. One of the main scientific applications has been surveys of binaries in several different stellar environments, to study the frequency and orbital characteristics of binaries in a variety of stellar populations. A main goal is to confront theories of binary formation and evolution with observational results. With various collaborators we have investigated the binary populations among pre-main-sequence stars, in the Hyades and M67 open clusters, and in the Carney-Latham proper-motion sample. Thus, we have data for coeval samples of binaries covering a wide range of ages. One result is clear evidence for evolution of binary orbits. The orbital period at which there is a transition from circular to eccentric orbits gets longer for older samples of binaries, presumably due to tidal circularization. Another result is that the frequency of binaries does not seem to depend on the stellar population. Binaries are just as common among the oldest stars in the halo of our Galaxy as among the younger stars in the disk.