Numerical simulations of close binary evolution were performed for five binary systems, using a newly developed evolutionary program. The systems have masses 3+2, 4+3.2, 6+4, 9+6, 12+8 M⊙ and periods 2d, 1d78, 3d, 4d, 5d respectively. The primary component was followed from the zero-age main sequence through the mass transfer phase to core-helium burning. Special care was given to the self-consistent determination of the mass transfer rate and the detailed treatment of composition changes. After the mass transfer phase the resulting systems consist of a main sequence star with a helium star companion of mass 0.36, 0.46, 0.82, 1.48, 2.30 M⊙ for the five systems respectively. Interesting “thermal pulses” were found in the 3+2 M⊙ system at the onset of helium burning.

It becomes more and more evident that for close binary evolution during Roche lobe overflow as well mass transfer as mass loss occurs. When a mass element ΔM is expelled from the primary during this phase, a fraction β is transferred to the secondary; the remaining part leaves the system. Moreover, angular momentum leaves the system, and also this fraction has to be specified; this fraction is related to a parameter α (Vanbeveren et al., 1979). For the computation of the evolution of massive close binaries also mass loss due to stellar wind of both components, prior to the Roche lobe overflow has to be taken into account. The mass loss rate Ṁ due to radiation driven stellar winds can be expressed as

The transfer of mass from one component of a binary system to its companion can involve ejection of the mass by radiation pressure, by Roche lobe overflow, by centrifugal action through the coupling of the upper atmosphere to the stellar rotation through magnetic fields, and by the dynamics of the atmospheric fluid under the action of an energy source at the photosphere or lower chromosphere. We will primarily address the latter mechanism, which commonly goes under the title of stellar winds. The major determinant of a stellar wind is the energy source. In the case of a binary system, the dynamic effects of the system on the flow pattern, particularly for close binaries, is a second major determinant, once an energy source is available to drive the wind. Roche lobe overflow might be thought of as a limiting case in which the system dynamics predominate.

A previously published model of a detached binary evolving into deep contact, under the assumption that both components fill the same equipotential surface, is tested for self-consistency. Hydrodynamic equations describing the flow between components at their interface indicate an enormous discrepancy in the magnitude and direction of the flow in the evolutionary calculation. This behavior strongly supports the view that binaries reaching contact in this way undergo a cyclic thermal instability as in many models of W Ursae Majoris systems, rather than overflowing an outer Lagrangian point.

SV Centauri is an early-type contact binary for which detailed physical data have been obtained from observations (Wilson and Starr 1976). The orbital period is p = 1.659 days and the separation is a = 16.1 R⊙. The radii of component Stars A and B are, respectively, R = 7.2 and 6.9 R⊙ so that they are in overcontact. The masses and the luminosities are M = 11.1 and 9.3 M⊙, and L = 2830 and 10900 L⊙ for Stars A and B, respectively. The more massive Star A is called the primary star but its luminosity is fainter than Star B. This implies that it is in the phase of rapid mass transfer and mass is outflowing from Star A. Indeed one of the most interesting and important points of this system is that the rate of period change is measured to be dP/dt = −9.4 × 10−8 (Irwin and Landolt 1972) which corresponds to the mass transfer rate of dMA/dt = −4 × 10−4 M⊙ yr−1.

During the evolution of a close binary system, there is a phase of mass exchange between its component stars. When the mass-losing star is an evolved star off the main sequence, the mass exchange is very rapid. Its rate may well exceed the critical accretion rate corresponding to the Eddington limit of the mass accreting star. Effects of such a rapid accretion were studied for the mass-accreting main-sequence stars (Flannery and Ulrich 1977; Kippenhahn and Mayer-Hofmeister 1977; Neo et al. 1977). The mass-accreting stars become quickly overluminous and their radii increase even to a red-giant size as mass is accumulated. Then the both component stars will fill their Roche lobes to form a contact binary system.

All proposals to explain the formation of blue stragglers advanced up to now seem to have inherent difficulties. One appealing possibility, the formation by coalescence of close binary stars, appears to be ruled out by the existence of blue stragglers with more than twice the turn-off mass of their cluster. We suggest that this contradiction does not exist, since the binary system on its way to coalescence passes through a phase in which a very extended common envelope with a size of several hundred solar radii forms. In this stage a third companion in the range of this envelope will be captured and finally merge to form a blue straggler that can have more than twice the turn-off mass.

Conventional computer models of close binary star systems usually start with at least one component on the main sequence. Models of premain sequence binaries have been computed to study mass transfer (Yamasaki 1971). However no pre-main sequence computations have been published that follow the evolutionary tracks of a binary system onto the main sequence, even though some observed systems appear to be premain sequence (Field 1969). The main purpose of this investigation is the evaluation of individual close binaries with a pre-main sequence model. The evaluation will be accomplished by comparing the positions of the observed binary on the Hertzsprung-Russell diagram with the evolutionary tracks generated by the pre-main sequence model. If both components appear to have the same age and fall near the tracks of the model, then the system is possibly pre-main sequence. Eleven semidetached binaries were considered, each with a total mass between 2.5 and 6 solar masses and with a period between 0.9 and 3.4 days.

The Hubble-Sandage variables are the most luminous stars in external galaxies. They were first investigated by Hubble and Sandage (1953) for use as distance indicators. Their main characteristics are high luminosity, blue colour indices, and irregular variability. Spectroscopically they show hydrogen and helium in emission with occasionally weaker Fell and [Fell], and no Balmer jump (Humphreys 1975, 1978). In this respect they closely resemble cataclysmic variables, particularly dwarf novae. In the quiescent state dwarf novae show broad H and HeI, together with a strong UV continuum. Weak FeII in emission has been observed in U Gem and SS Aur (Warner 1976). The Balmer jump is either not present, or weakly in emission. The principal spectroscopic difference is the increased breadth of the occasionally doubled emission lines.

A study of the binary frequency among the O-stars finds that only 40% are binaries. They show a strong tendency to have mass ratios near unity and none with values less than 0.3. We discuss the evidence that mass loss is responsible for the observed mass ratios.

Some of these results are part of a more extended investigation on WR stars that will appear in four papers, i.e. Vanbeveren and Packet (1979), Vanbeveren and Conti (1979), Vanbeveren and Doom (1979), Vanbeveren and de Loore (1979).

In many recent papers dealing with computations of evolutionary sequences for massive stars with mass loss, the evolutionary status of Wolf- Rayet stars has been discussed. When one whishes to compare these theoretical computations with observations, it is of great importance to know the masses of Wolf-Rayet stars, which unfortunately are well determined only in a few cases. Estimates of the masses are provided by binaries with WR type components. I will report here preliminary results of spectrographic observations for 4 WN+OB systems, with new information about the masses of their components.

An updated list of known spectroscopic double-line and single-line Galactic Wolf-Rayet binaries is presented. From this we discuss e.g. mass ratios, the binary frequency and the evidence for low-mass unseen (compact?) companions. Spectroscopic binary orbits of several WR stars in the Magellanic Clouds are noted for the first time.

As recently as six years ago it was generally agreed that all Wolf-Rayet (WR) stars were the product of mass loss induced by Roche lobe over-flow (RLOF), and that thus they were all members of binary systems. In particular, the presence of absorption lines in the spectrum of a WR star was taken as a definitive indicator that the star was a binary, as it was well known that emission alone (other than the absorption present in P Cygni profiles) was a WR characteristic. However, Niemela (1973) showed that in the WN 7 binary HD 92740 that the absorption and emission lines move in phase, proving that in at least one case the absorption spectrum originates in the WR star itself. Conti (1976) has meanwhile suggested that WR stars could form by mass loss due to a stellar wind rather than RLOF. Obviously it is well worth examining the belief that all (or even most) WR stars are binaries. The currently popular feeling is that the late WN stars (WN 7, 8, and 9) are the only possible exceptions to the traditional picture, while the members of the other subclasses are all hydrogen-poor and probably members of binaries. I hope to convince you today that this is not the complete story.

The structure of the extended atmosphere and the physical characteristics of the WN5 star V444 Cyg have been determined from the analysis of light curves of this system in the range λλ2460 − 35000. The radius of the core of the WN5 star at τ = 1 is ro = 2.9R⊙; the temperature at the surface of the core is equal to 90000oK, the electron density at its surface is 0.9 · 1013 cm−3. The bolometric luminosity of the WN5 star Lb = 2.1039 erg/s is close to the critical Eddington luminosity 2.6 · 1039 erg/s. The electron temperature decreases monotonically with height through the extended atmosphere from the value Te = 80000° − 90000°K at r = 2.9R⊙ to the value Te = 20000° − 30000°K at r = 8R⊙. The outflow velocity at the surface of the core (ro = 2.9R⊙, τ = 1) is vo = 400 km/s; the outflow in the region r ≥ 12R⊙ becomes stationary with the constant velocity v = 1000 − 1200 km/s. The total interstellar absorption for the system V444 Cyg is Av = 2m.66. The results obtained confirm strictly the model of Wolf-Rayet phenomenon proposed by Beals (1944). Chromospheric-coronal effects in the extended atmosphere are absent until r = 20R⊙. Characteristics of the WR star are consistent with the predictions of the evolution theory for massive close binary systems with mass exchange. The WR star is in a late evolutionary stage after the main sequence stage, being a helium remnant that has been formed as a result of the mass exchange.

The ultraviolet spectrum of the close binary UW CMa (= HD 57060; 07f Ia + 0-B), obtained with the COPERNICUS (OAO-3) and IUE satellites, is analyzed in terms of a comparison between observed P Cygni resonance line profiles and theoretical calculations in an expanding model atmosphere.

He II (λ 4686) emission has been detected in the spectrum of the binary Be star ɸ Persei. The He II emission has velocity variations which are 180° out of phase with the velocity variations of the primary. The primary velocity curve is determined from the broad He I and H lines, and the Fe II emission lines. The amplitude of the primary velocity curve (K1) is 16 ± 2 km s−1. The amplitude of the He II velocity curve is 121 ± 9 km s−1.

If one assumes that the He II emission arises from the secondary, minimum masses for the primary and secondary are 29 ± 14 M⊙ and 3.8 ± 1.7 M⊙, respectively. We suggest that the secondary is a hot compact star, either the remaining helium core of a once more massive star, or a neutron star. The primary is too cool, Bl IIIe, to have any appreciable He II emission.

Preliminary values of masses and radii are presented for twelve Algol systems with cooler components of class F or later. The results are new for nine of the systems and are based on measures of the D lines for the faint cooler components. For most of the systems with periods longer than six days the orbits of the hotter components are uncertain because of absorption by non-photospheric matter in the systems. The masses of the hotter components are approximately normal for their spectral types, while the masses of the Roche-lobe filling components range from 0.2 to 1.4 solar masses.

We have examined the observational data of 60 Algols in order to check some theoretical views on mass loss and mass transfer in Algols. Not negligible contradictions between some theoretical non-conservative attempts and the observational scenario have been emphasized.

Evolutionary calculations show that during the last, slow stage of mass exchange in close binary systems, matter with gradually increasing molecular weight is transferred. If no mixing occurs the accreting star moves towards a position in the HR diagram hotter than the main-sequence band, in clear contradiction with observational evidence in Algols. So some efficient mixing mechanism must be present. Several mechanisms are examined, some of which can be definitely ruled out. A closer examination of some others is necessary.