The occultation and spectroscopic approaches to the discovery of multiple stars are reviewed, and the technique of spectroscopic measurement recapitulated. A notation for multiple stars is proposed. An examination of the gravitational interaction levels suggests an average value of ∆ log P for one hierarchical step close to 3.0 and ∆ log a near 2.0 consistent with a binding ratio near 1 : 10000. The distribution for B stars seems to differ from other types and it is conjectured that they cannot be old enough to be residues of clusters. In general few multiple systems seems to show evidence of age sufficient for advanced evolution of the components.

The difficulties of discovering and observing multiple stellar systems are reviewed and discussed with especial reference to those systems which are best observed by a combination of visual and spectroscopic or visual and photometric means. Attention is drawn to the difficulties that often exist in reconciling the results of two different methods of observation. A brief discussion of the possibility of detecting spectroscopically invisible companions of low mass is included in the paper.

A method is described which, through the use of an optical image rotator in conjunction with an area scanner, allows the accurate determination of the magnitudes and relative position of the components of a double star even if the angular separation is significantly smaller than the radius of the seeing disk.

A new 1.87 m fixed-baseline interferometer designed for binary star observation is now in operation at the Italian outstation of the Royal Observatory Edinburgh. It has a theoretical resolving power of 0.015 arcsee and a limiting B magnitude of 6.5. Atmospheric seeing is dynamically corrected, allowing a quantitative measurement of fringe visibility.

A historical summary of stellar-perturbation analyses is presented, followed by a discussion of the attainable accuracy and the sources of observational errors. A provisional value for the mass-density of astrometric companions is found to be 0.063±0.018 (p.e.) solar masses per cubic parsec.

A few examples of the resolution of very close multiple stars by the method of lunar occultation are described. In this way both the power and some of the limitations of this technique may be demonstrated.

A systematic search for binaries in a sample of 123 bright field stars of types F3-G2 IV or V is described. Combination of the results for 25 newly discovered spectroscopic binaries with those of 21 spectroscopic, 23 visual, and 25 common-proper-motion pairs previously known brings to 88 the total number of companions identified in the sample. The distribution of the 88 periods shows a single maximum; the median period is 14 years.

The frequencies of singles: doubles: triples: quadruples are found to be 42: 46: 9: 2. Less than half of the stars are thus observed to be single. Estimates of the completeness of this search lead to a determination of the number of systems missed and to a determination of their secondary masses.

Analysis of the secondary-mass distributions, observed and predicted from completeness estimates, indicates the existence of two types of binaries.

For systems with periods less than about 100 years, the distribution of secondaries varies with the cube-root of the secondary mass. These systems must be the result of fission. Two-thirds of their primaries have companions of stellar mass. Extrapolation of the cube-root relation implies that the remaining one-third of the primaries have non-stellar secondaries, i.e., close companions whose masses are less than 0.07 solar masses.

For systems with periods larger than roughly 100 years the distribution of secondaries follows that given by the van Rhijn function. They must be the result of condensations that contracted separately but are bound gravitationally. Roughly three-fourths of all primaries in the sample have such distant companions, all of which are likely to be themselves close fission-systems. Evidently single stars are rare among solar-type dwarfs.

There is much current interest in flare stars of the type dMe with particular concern as to their ages and mechanism responsible for flaring. Some fifty Me stars are on the Sproul Astrometric program. They have been discussed as flare stars or potential flare stars (Lippincott 1971). At least half are known to be members of binary or multiple systems. All are on the Sproul program to obtain accurate parallaxes and to analyze the long interval observations for non-linear proper motions indicating unseen companions. The astrometric results are discussed and statistics presented for the multiplicity among these nearby systems.

A brief review of the history of multiple-star discovery is followed by a discussion of selection effects, discovery probability, and detection techniques. The incidence of multiplicity and the properties of multiples in various samples of binaries are investigated.

En étudiant les systèmes multiples issus d’un couple de découverte ancienne grâce à l’emploi ultérieur d’un instrument plus puissant, on constate qu’un grand nombre d’étoiles doubles n’ont jamais été examinées avec un tel instrument; il en résulte que parmi ces couples plusieurs centaines sont des systèmes au moins triples que nous ignorons.

It is generally suspected that multiple systems and galactic clusters are objects of different kinds. Their frequency distribution as a function of the number of components, n, as well as their sizes and their masses, show important differences between both categories and also the probable existence of a limit lying between n = 10 and n = 20 components. The basic difference between these objects seems to be that the multiple systems finally appear to be components of the stellar medium in th same way as single and double stars, whereas the galactic clusters are groups of components of a particularly dense part of this medium.

Using both measured values and model atmospheres, the equivalent widths of the Na D lines of hypothetical late type main sequence binaries are calculated. Enhancements in the D lines of 90% over the value for a single star whose B — V color is the same as the composite B — V are possible. It is suggested that this effect may be used to identify binaries in galactic clusters, in particular those binaries whose components differ by more than 3 mag in brightness.

We describe a method which interposes a relation Mass-Luminosity in order to calculate the dynamic parallax and the masses of the components in a system of known orbit which involves a spectroscopic binary. We apply the method to a few cases.

Analyses of early Lick Obs. spectra and recent high dispersion Kitt Peak National Obs. spectra indicate that k Peg is a quadruple system having all four spectra visible under favorable conditions. Preliminary spectroscopic orbital elements are presented for the AB system, and for Bb. The Bab system has been found to be the previously known 5.97-day spectroscopic system. The total mass is found to be between 2 and 3 solar masses only. Thus the brightest of the four stars Ba(F5IV) —or possibly all four of the stars—must have an abnormally small mass for its luminosity. This system appears to be similar to several others which have apparent overluminous components. Such systems are of interest in regard to stellar evolution.

The visual binary Psi Sagittarii (ADS 12214, HR 7292) is a spectroscopic triple system with orbital periods of 10.78 days and 20.0 years. Periastron passage in the long period crbit occurred on JD 2442410 ± 30 days. Masses derived from the parameters of the spectroscopic and visual orbits are 2.7 m⊙ for the visual primary, 2.1 m⊙ for the short period pair primary, and 1.7 m⊙ for the short period pair secondary. The distance to the system is 100 parsecs. The orbital inclination of the close pair is nearly the same numerically as that of the visual orbit. Suggested spectral classifications indicate that all three components have evolved off the main sequence. Mass exchange in the close pair has not yet occurred.

An investigation is currently being made of the parallax, proper motion, orbital motion, and mass ratio of the nearby binary star Stein 2051 [04h 11m.4, + 58°49’; (1900); mv = 11.08 and 12.44] based upon photographs taken with the Sproul Observatory 61-cm refractor, the 155-cm astrometric reflector of the U.S. Naval Observatory, and the 33-cm Vatican Carte du Ciel astrograph.

A parallax of 0”185 reduced to absolute has been obtained from the material of the Sproul and Naval Observatories which covers the period 1965 to 1975.

Since an arc of only 64° in the orbital motion of the visible components has been described from the 1908 epoch of the earliest Vatican plate, and since the separation has increased from 6.” 3 at that time to 7.“ 3 at present and still on the increase, the period of the orbit can only be roughly estimated to be in excess of 300 years.

With all plates measured in the same reference frame it has been possible to determine the individual motions of the two components and the mass ratio between them. Preliminary results indicate a perturbation in the orbital motion of the red component with a period of 20 years and a semi-major axis of 0’.’065. The mass ratio between the red dwarf system and the white dwarf is 0.5.

A mass of 0.22 M⊙ of the red component has been determined based upon comparison of its luminosity (Mv = 12.42) with the red dwarfs Krüger 60 A (Mv = 11.82), o2 Eri C (Mv = 12.73) and Krüger 60 B (Mv = 13.48) which have the same spectral type (M4) and known masses of 0.26, 0.19, and 0.16 respectively in units of solar masses. Its dark compainion has a mass of 0.02 M⊙ The white dwarf of spectral type DC and Mv = 13.78 has a mass of 0.48 M⊙ as determined from the mass ratio of 0.5 mentioned above.