The WC star HD 164270 (=WR 103 in the catalogue by van der Hucht et al., 1981) is one of the stars to which Smith (1968) ascribed variability, although she was not able to investigate the light variations completely. The star has been regarded as a single star (i.e. no evidence of absorption lines) but it has been proposed by e.g. van den Heuvel (1976) that “single” WR stars might have collapsed companions. Eclipses by such short-period companions should be observable. Thus we designed an observational programme to look for hitherto unobserved companions. The programme contained 16 “single” WR stars. Some of them, including WR 103, had previously shown light variations as mentioned.

ED 50896 (=EZ CMa) has long been known as a variable star. Recently, Firmani et al. (1980) derived a period of 3.763 days from spectral and photometric variations. This period was later confirmed by McLean (1980) and by Cherepashchuk (1981).

At present we know that the Wolf-Rayet stars do not form a homogeneous group of objects with similar masses, but that their masses span a considerable range. Whether or not there is a correlation between the masses and the different subtypes of WR stars, can only be answered by more observations of WR stars in binarysystems yielding estimates of the masses of their components.

Radial velocities are determined by Fabry-Pérot interferometry at 131 points of the ring nebula S308. The velocities have yielded a kinematic distance of 1.5 kpc for the object, and an expansion velocity of 45–60 km s−1. The ring nebula has a diameter of 13 pc and the age is estimated to be about 1.5 to 2×105 years.

High and low resolution spectra of the spectroscopic triple star θ Muscae were studied for time-variability. Within 7 min no variations were found in the low resolution data. On a time scale of 8 days and longer, variations in the P Cygni profiles of the resonance lines occur. These may be explained as variable mass loss from the system, or as aspect effects in an asymmetric wind flow.

The Wolf-Rayet star HD 5980, which is probably associated with the bright HII region NGC 346 of the Small Magellanic Cloud, was found to be an eclipsing binary by Hoffmann, Stift and Moffat (1978). Breysacher and Perrier (1980) determined the orbital period, P=19.266±0.003d, of the system whose light curve reveals a strongly eccentric orbit (e=0.47 for i=80°). The behaviour of the light curve outside the eclipses shows that we are dealing with a rather complex binary system. An analysis of the spectroscopic data is presented here.

In this paper we summarize the basic observational properties of WR stars, and the current status of the evolution of massive stars, giving particular emphasis to those physical mechanisms (mass loss, overshooting, turbulent diffusion…) that may lead O type stars to the WR stage. We consider only those scenarios for WR formation and evolution that have been developed during the last few years and that appear to be sufficiently substanciated by observational and theoretical arguments. Finally, the theoretical results are compared with the observations, pointing out several still unsettled aspects of the problem.

A large part of the stars of the galactic disk are binaries. Abt and Levy (1976,1978) estimate that more than 50% of B type stars are members of binary or multiple systems, and about 50% are close binaries. According to Conti et al. (1980) the total fraction of certain or probable binaries among O stars is about 36 percent. This means that during the evolution their components can fill their Roche volumes, and consequently have to lose mass in order to keep the star within the allowed surface. Several review papers dealing with the evolution of binaries were published recently: Paczynski (1971,1980), van den Heuvel (1976), de Loore (1980,1981), Webbink (1979).

The conclusion by Maeder et al.(1980) that the strong gradient in NR/NWR results from the effect of metal content on the rate of mass loss is reconsidered. We suggest that the above gradient simply reflects a simi lar gradient in the progenitor stars, although effects of metallicity cannot be excluded.

According to the classical evolutionary scenario for massive close binariesit is qenerally assumed that two WR phases are encountered, i.e. a WR+OB stage and a WR + compact star stage. The survival probability of a massive binary after an instantaneous supernova (SN) explosion critically depends on the assumed asymmetry of the SN explosion, i.e. on the “kick” velocity Vk (Sutantyo, 1978). The purpose of this paper is to make a statistical study on the expected WR + compact star binaries compared to the WR+OB type binaries based on non-conservative massive binary evolution, as a function of Vk which, as we will see, is very poorly known for these types of stars.

Grids of evolutionary models for single stars in the range 9–120 M⊙ have been computed up to the end of the central C-burning phase with up-to-date stellar physics. In addition to other results these models allow us to study the main properties of the formation of WR stars as post-red supergiants (Maeder, 1981c). Three grids of models have been computed: one for constant mass evolution (grid A) and the other two (B and C) with mass loss. As there exists no unique parametrization representing, over the various spectral types, the mass loss rates as a function of basic stellar parameters, we have used mass loss rates fitted on the observations, applying different laws for a) OB stars, b) B-G supergiants, c) red supergiants (RSG) and d) WR stars. At a given luminosity, the rates in grid C are twice as large as in grid B and both grids encompass the central bulk of the observed mass loss rates.

The larger part of close binary components with initial mass exceeding ∼20 Mo becomes WR stars in the core helium burning stage. Some of the most massive WR stars may be products of evolution of single massive stars with initial masses exceeding ∼50 M0 if the mass loss in the infrared supergiant stage is effective enough. The Ledoux criterion of convective stability seems more promising to explain the observed properties of WR stars.

We have analyzed Ciatti et al's (1980) UBV lightcurves of the massive eclipsing binary RY Set. The BO primary, being considerably over-luminous and oversized for its mass, is certainly appreciably evolved. That this star is nearly filling its Robe lobe strongly supports the idea that it is undergoing a phase of mass transfer (case B), through Robe lobe overflow, towards its more massive secondary companion. In view of its properties, RY Set, which represents an uncommon evolutionary stage, is lively to evolve into a Wolf-Rayet binary systems.

Sanduleak (1971) has listed five stars, not apparently associated with planetary nebulae, which show very strong 0 VI 3811,34 å emission. He pointed out that two of them are in the Magellanic Clouds and have absolute magnitudes comparable to those of classical (Population I) Wolf-Rayet stars. 0 VI emission is know to occur in some classical Wolf-Rayet stars, but not with the extreme strength shown by the Sanduleak stars. We have obtained absolute optical spectrophotometry (3100–7400å) of all five of these stars, using the UCL Image Photon Counting System and RGO Spectrograph on the Anglo-Australian Telescope. Figure 1 shows their relative flux distributions. Inspection shows that Sand 1 is very lightly reddened, Sand 2 and 3 have intermediate reddening, and Sand 4 and 5 are heavily reddened. IUE ultraviolet spectrophotometry has been obtained of the first three stars; Sand 4 and 5 are too heavily reddened for IUE spectra to be feasible.

During the last years, many authors have recognized the existence of some OB stars that show the nitrogen absorption lines enhanced (Walborn 1970, 1971; Conti 1973; Jashek and Jashek 1974).

The treatment of close binary evolution changed from conservative (total mass and angular momentum constant) to non-conservative (a certain fraction of the matter expelled from the mass losing star, the primary, leaves the system, carrying away a fraction of the total angular momentum). Only the evolution of the mass losing star was computed in detail; the evolution of the accreting star was estimated in a very simple way by computing the evolution in the normal way just like for single stars, but taking into account the mass increase.

In a recent paper (Doom and De Grève, 1981) the remaining main sequence lifetime of the mass gaining component in massive close binary systems was computed. Using results of that paper and the definition of the four important events in the evolution of a massive close binary system (RLOF(M1), RLOF(M2), SN(M1), SN(M2)), four evolutionary stages in the life of the system can be defined: OB+OB, WR+OB, c+OB (or WR+WR) and c+WR. The two possibilities for the third stage depend on the initial mass ratio of the system. The final stage c+c, is not considered here.

Over the last 40 years WR stars have been associated with almost all possible evolutionary stages. Hereafter are listed those scenarios which may be considered as most relevant in view of recent surface abundance determinations (cf. Willis, this meeting).

It is well known that the Wolf-Rayet phenomenon is not restricted to some bright and massive stars, presumably in their core hydrogen-burning or helium-burning phase, but that it is also encountered among the central stars of some planetary nebulae (PNe). The PN nuclei are generally regarded as the evolutionary product of low and intermediate mass stars (with initial masses M.1 below ∼5 M⊙), which have lost most of their hydrogen-rich envelope during the so-called Asymptotic Giant Branch (AGB) phase. Correspondingly, their present mass cannot exceed the Chandrasekhar limit (∼1.4 M⊙), and their internal structure consists of a highly degenerate carbon-oxygen core containing most of the stellar mass, surrounded by an intershell region of mass ΔMCSH, and by a very low-mass envelope (Me < ∼10−3 M⊙).

Fully ten percent of the Wolf-Rayet stars known in our galaxy are central stars of planetary nebulae. These stars are highly-evolved stars of approximately one solar mass. The existence of WR characteristics in the spectra of central stars as well as in massive, younger stars proves conclusively that WR spectra are a phenomenon that occurs in hot stars of widely different masses and evolutionary histories. Just what this phenomenon is — the nature of the instability that drives the wind — is not known, but the phenomenon is associated with a prior event in the star's evolutionary history. The common event that links Wolf-Rayet stars is loss of most, if not all of the outer, hydrogen-rich envelope just prior to the onset of WR characteristics. As Tutakov (this volume) and others have emphasized, there are many ways for a star to lose its H-rich envelope. These ways include mixing, ejection via a wind, or, in the case of a close binary, ejection via Roche-lobe overflow. The existence of WR-type central stars implies simply that we should include one more path leading to the Wolf-Rayet phenomenon, and that is, ejection to form a planetary nebulae.