Fundamental data on OB supergiants and binary systems are presented. Mass loss is reviewed, and its correlation with close binary membership. The evolution of massive binaries is discussed and several classes of observed binaries compared with the theoretical picture. Individual systems of significance are pointed out and the chief problems in massive binary research are indicated.

Most of the strong galactic X-ray sources must be low mass, close binary systems, such as Her X-1 and Sco X-1. Two evolutionary scenarios are discussed, both involving type I supernovae that occur when mass-accreting white dwarfs are driven over their mass limit. In one, accepting the correctness of the idea that a neutron star or black hole is the seat of the X-ray emission, the SN occurs before the system is an X-ray source. Another possibility is that the white dwarf is the X-ray source, just prior to its collapse and the ensuing SN.

It is well known that the outcome of case B evolution of the primaries of massive close binary systems (M1 ≥ 9 M⊙) depends on the initial primary mass. The most massive primaries finally ignite carbon, form iron cores and presumably end in a supernova explosion, whereas the lighter ones presumably end as white dwarfs, without carbon ignition. This paper derives an estimate of the mass boundary separating these two kinds of evolution.

As an example of the first case, the evolution of a 20 M⊙ + 14 M⊙ system was computed; after the mass exchange, the primary star (with M = 5.43 M⊙) evolves through the helium-burning (Wolf-Rayet) stage towards a supernova explosion; finally the system evolves into an X-ray binary (BWRX-evolution).

As a representative for the second case the evolution of a 10 M⊙ + 8 M⊙ system was examined. After the first stage of mass exchange, the primary (with a mass of 1.66 M⊙) approaches the helium main sequence; during later phases of helium burning the radius increases again, and a second stage of mass transfer starts; after this the star (with a mass of 1.14 M⊙) again evolves towards the left in the Hertzsprung-Russell diagram and ends as a white dwarf (BSWD-evolution). A system of 15 M⊙ + 8 M⊙ is found to evolve very similar to the 20 M⊙ + 14 M⊙ system. The mass Mu, separating the two types of evolution, must therefore be situated between 10 and 15 solar masses. An initial chemical composition X = 0.70, Z = 0.03 was used for all systems.

The expected final evolution of massive close binaries (CB) in case B is reviewed. Primary stars with masses ≳ 12–15 M⊙ are, after loosing most of their envelope by mass exchange, expected to explode as supernovae, leaving behind a neutron star or a black hole.

Conservative close binary evolution (i.e. without a major loss of mass and angular momentum from the system during the first stage of mass transfer) is expected to occur if the initial mass ratio q0 = M20/M10 is ≳ 0.3. In this case the primary star will be the less massive component when it explodes, and the system is almost never disrupted by the explosion. The explosion is followed by a long-lasting quiet stage (106–107 yr) when the system consists of a massive main-sequence star and an inactive compact companion. After the secondary has left the main-sequence and becomes a blue supergiant with a strong stellar wind, the system becomes a massive X-ray binary for a short while (2–5 × 104 yr).

The numbers of Wolf-Rayet binaries and massive X-ray binaries observed within 3 kpc of the Sun are in reasonable agreement with the numbers expected on the basis of conservative CB evolution, which implies that several thousands of massive main-sequence stars with a quiet compact companion should exist in the Galaxy. About a dozen of these systems must be present among the stars visible to the naked eye. During the second stage of mass exchange, large loss of mass and angular momentum from the system is expected, leading to a rapid shrinking of the orbit. The supernova explosion of the secondary will in most cases disrupt the system. If it remains bound, the final system will consist of two compact stars and may resemble the binary pulsar PSR 1913 + 16.

In systems with q0 ≲ 0.2–0.3 large mass loss from the system is expected during the first stage of mass exchange. The exploding primary will then be more massive than its unevolved companion and the first supernova explosion disrupts the system in most cases. In the rare cases that it remains bound, the system will have a large runaway velocity and, after a very long (108–109 yr) inactive stage evolves into a low-mass X-ray binary, possibly resembling Her X-1.

It has been proposed that tidal interaction is the mechanism which restored the circular orbits of X-ray binaries after the supernova explosions which produced the compact components in those systems. (Lea and Margon, 1973; Sutantyo, 1974). This implies that the primaries must possess large viscosity in order to provide large energy dissipation rates required in this process. At the end of this process the system reaches a stable synchronous and circular orbit. If this hypothesis is true, there will be further consequence of the strong tidal interaction during the later stages of evolution of the system. The gradually changing structure of the primary during the post main-sequence stages will continuously disturb the stability of the orbit. There will be transfer of angular momentum from the orbital motion into the rotation or vice versa through the tidal interaction. At a certain stage, the radius of the primary exceeds a critical radius, the orbit becomes unstable and synchronism becomes impossible. The compact component will be spiralling down onto and into the primary as proposed by Sparks and Stecher (1974).

Accumulated evidence on the existence of collapsed stars (black holes) in X-ray and non-X-ray binaries is considered. At present Cyg X-1 is the only strong case. The determinacy of solutions for mass ratios and other parameters of X-ray binaries is discussed, and some effects of tidal distortion and X-ray heating on measured radial velocities and ellipsoidal light curves are described.

When a contact binary expands so much that the stellar surface moves beyond the outer Lagrangian point, a common envelope binary is formed. The suggestion is made that while the two dense stellar nuclei spiral towards each other, the envelope expands and is eventually lost. Most of the angular momentum is lost with the envelope, and therefore the final orbital period may be orders of magnitude shorter than the initial period. V471 Tau could have formed from a binary with a ten year orbital period. Most probably, cataclysmic variables are products of the evolution of systems like V471 Tau. Observational discovery of a short period binary being a nucleus of a planetary nebula would provide very important support for the evolutionary scenario presented in this paper.

All attempts to fit the binary pulsar PSR 1913 + 16 into an evolutionary scenario consider the effect of an instantaneous mass ejection on the orbital elements of a circular binary.

An analytic model has been constructed for the change of orbital elements because of tidal drag in a binary system where one of the components is compact (Lecar et al., 1976). The basic assumptions are (i) tides on the compact object are neglected (ii) only P2 deformations of the non-compact star are considered (iii) the lag angle is linearly proportional to the difference between the instantaneous orbital angular velocity and the spin angular velocity of the non-compact star (iv) the semi major axis a and eccentricity e do not change significantly in a single orbit. The results are exact solutions for a and ė as a function of e and the ratio of primary star to orbital angular velocity, Ω/ω.

Tidal instability in this model is manifest by the dynamical inability to attain or sustain synchronism. An analysis to second order in e and first order in (Ω/ω−1) shows that the criterion for instability is the same as in the work of Counselman (1973), namely ma2 > I where m is the reduced mass and I the moment of inertia of the primary star. This criterion is shown to apply to eccentric orbits and is thus more general than the criterion derived from the energetics of circular orbits. Stable systems may overshoot the condition of synchronism before settling into the final circular synchronous orbit.

Spectroscopic and photometric observations relevant to the structure of classical novae, recurrent novae and dwarf novae are collected together. Details are given of optical, infrared, satellite ultraviolet, X-ray and radio observations. Most of these can be understood in terms of the hot spot model of cataclysmic variable stars. Estimates are made of luminosities, temperatures and rates of mass transfer. Observations made during eruptions of dwarf novae imply large changes taking place in the accretion disc.

From spectroscopic and photoelectric data BD +16° 516 (V 471 Tau) is inferred to consist of a white dwarf of mass 0.72M⊙ and radius 0.012R⊙, with a detached K2V companion of mass 0.70R⊙ and radius 0.78R⊙; the separation of the two components is 3.02R⊙. The evolutionary history of the system is discussed.

The first image-tube spectra of five dwarf novae and one old nova are described and compared with the colours of these stars.

Radial velocities of U Gem have been measured from individual components of the emission lines of the Balmer series, Ca II K, and He 14471, originating in the gaseous disk and in the hot spot of this binary system. These new data give K1 ≅ 143 km s−1 and Vdisk sin i ≅ 530 km s−1. The resulting mass of the primary component is m1 = 1.1 ± 0.4 m⊙ and the mass-ratio m2/m1 = 0.6 ± 0.2. Radial velocities and intensities of the emission components originating in the hot spot are also discussed.

We have evolved a large number of models of thermonuclear runaways in the hydrogen-rich envelopes of 1.00 M⊙ carbon-oxygen white dwarfs. Models characterized by enhanced CNO abundances satisfy the observations of the common nova outburst. They eject from 1027 to 1029 g moving with velocities of 200–2400 km s−1: kinetic energies of 1044 to 1045 erg. The theoretical light curves are similar to the observed light curves of common novae during the early stages of the outburst. As a consequence of our studies, we can present an explanation for the continuous ejection of mass which is observed for long times after the initial outburst and for the relationship between maximum magnitude and the decline to minimum. We are also now able to present an explanation for the constant luminosity phase of the outburst and the oval shapes of the ejected nebulae.

We have extended our earlier work to models with extreme enhancements of 12C and to models which include the infalling material in the evolution. The extreme 12C studies result in outbursts which reach near supernova proportions while the accretion models result in light curves which resemble observed light curves quite closely. Finally, the accretion studies also demonstrate the need for enhanced abundances.

Recent work on the physical processes resulting from mass transfer between the red and blue components of dwarf nova binaries is reviewed. The optical behaviour of the blue component's accretion disc suggests that it may be the infall, accretion energy which is being liberated during outbursts. Theoretical results which suggest that the red component may suffer quasi-periodic mass transfer instabilities are discussed. The resulting accretion disc properties are considered and compared with the observed optical outburst behaviour for the simplest steady state disc models. The complexity of the interaction between the two stellar components in these systems is emphasized.

Gravitational radiation of energy and angular momentum can modify and in some cases, control the evolution of a close binary system. The region of interest is briefly delineated. Recent work of the author and colleagues of relevance to this area is discussed, including theoretical studies of accretion, mass loss and mass transfer, and an observational study of a system, HZ 29 where gravitational radiation may dictate its behaviour.