New data concerning observational aspects of Wolf-Rayet binaries are summarised. WR+O and WR+c binary systems are considered. All the data concerning WR binaries agree well with the modern understanding of WR stars as the helium remnants formed as a result of mass loss by massive O-type stars.

Preliminary radial velocity orbits are presented for three binary systems containing Wolf-Rayet stars, namely Sk-71° 34 in the LMC, and WR8 and WR98 in our galaxy. Sk-71° 34 is found to be a WN3 + O6 double-lined binary with an elliptic orbit of period about 34 days. WR8 and WR98 both have WN type spectra with carbon lines. In WR8 the N and C lines appear to move in antiphase, while in WR98 all emission lines have the same orbital motion.

In recent years very massive single stars have been found to be upward of 90 M⊙. Massive contact binary systems have been found among the early-type systems, but their masses are far less than those reported for single stars. The most massive component found is about 60 M⊙.

It is generally believed that no late-type very massive stars have been detected (Humphreys and Davidson). This may be due to the large amount of mass loss from stellar wind. Recently, several extremely long-period late-type binary systems have been found to be contact systems. Two systems, UU Cnc and 5 Cet, have their primary components with masses exceeding 40 M⊙, and K spectra. This result tends to suggest that close or interacting binary stars may be able to preserve the mass loss from stellar wind within the binary systems.

The different channels for the formation of WR-stars, as suggested in the literature, are investigated. The presently available tools, in terms of evolutionary recipies, are reviewed and the results investigated of the use of these tools, with respect to the WR-binaries.

Two of the three basic formation channels, mentioned in the literature, may serve as ways to obtain the presently observed group of WR-binaries: stellar wind mass loss in O-stars during the pre-WR phase and mass transfer in a massive close binary system. Discrepancies with observations necessitated the incorporation of stellar wind in the binary scenario.

Theoretical developments in the last years are reviewed and confronted with the observations. The necessity of mass transfer in the formation process is argued for a major fraction of the double-lined WR-binaries. Using specific evolutionary ingredients, the original parameters of the systems are calculated. We also discuss the appearance and disappearance of double-lined WR-systems as well as the existence of eccentric WR-systems after mass transfer. The evolution of a number of specific systems is discussed individually.

A method of light curve analysis is described which allows the study of an eccentric partially-eclipsing system containing one component possessing an extended atmosphere. The effects of transparency as well as limb-darkening are taken into account. Preliminary results obtained for the Wolf-Rayet eclipsing binary HD 5980 in the SMC are presented.

The role of IUE in the determination of orbital elements of massive stars is reviewed, emphasising the complementary value of ultraviolet observations when combined with archival optical data. Results on eight systems are described and the potential of the IUE archive for radial velocity work in general is advocated.

Now it is clear that there exist some long period (periods exceed a year) Wolf-Rayet binaries. This paper presents the results of our radial velocity study for three WR binaries: HD193077 (WN6 + O9), HD193793 (WC7 + O5-6) and HD192641 (WC7 + O9). Our spectroscopic observations confirm that HD193077 and HD193793 are long period systems with the periods of 1538 and 2886 days. Most likely, HD192641 is also a long period WR binary (P ≥4400 days).

The formation of the eccentric, long period binary WR 140 (HD 193793, WC7 + 04-5V, e= 0.8, P = 7.9 yr), in terms of an eccentricity modulated late case B of mass transfer, is investigated. Using only stellar wind models, it is impossible to obtain a combination of a spectral type O4-5 together with a WC star.

We report on progress in the theoretical modelling of colliding winds in WR binary systems, concentrating on two WR+O systems, namely V444 Cygni, and HD 193793.

In order to synthesize the eclipse light curve of V444 Cygni, we adopt the following model. The O star revolves round the WR star still within the outer regions of its extended atmosphere. The O star shadows a distinct volume of the WR atmosphere which thus cannot contribute to the total flux seen by the observer. On the other hand, additional radiation emerges from the surface of the O star. Its contribution to the total flux is more or less diminished by absorption when the rays pass through those parts of the WR atmosphere which lie between the O star and the observer. The WR atmosphere is given by our usual models (cf. Hamann and Schmutz, 1987; Wessolowski et al, 1988).

Radial-velocity and photometric observations of eclipsing binaries allow one to estimate the masses of the stars in such systems, their relative brightnesses and sizes. If polarization variations are detected throughout the binary period, one may confirm the inclination of the system. All three types of information are available for V444 Cygni and CQ Cephei, see Underhill, Yang, and Hill (1988a,b), Underhill, Gilroy, and Hill (1990), and the references quoted in those papers.

After CQ Cep, the system CX Cep has the shortest orbital period known among Galactic WR+O binaries. However, no definitive spectroscopic study is yet available for CX Cep, probably because of its relatively faint magnitude (B ≈ 13). We have therefore obtained and analyzed some 60 CCD spectra (3700-4900 Å, S/N ⋍ 100, 5 Å/2 pixels) in August and October 1987.

We have recently used the PISCO and VATPOL Polarimeters at the MPI (Chile) and CASLEO (Argentina) 2.2 m telescopes to monitor over several weeks the brightest WR+O systems fainter than 9th magnitude in the Galaxy (V filter) and LMC/SMC (GaAs tube without filter). Each data point is accurate to ~ 0.02% in polarization P.

We have obtained some 15 CCD spectra at different orbital phases for each of the SBl systems HD193928 (WN6), HD197406 (WN7), and CQ Cep (WN7), in an attempt to detect the companion, determine its orbit and estimate the masses. The wavelength range of the spectra is λλ3600 - 4200Å, with 1.4Å/pixel and < S/N > ≈ 100/pixel. We compare the best emission-line (N IV λ4058Å) with the best absorption-line (H9 λ3835Å).

RY Scuti (HD 169515) is a massive 9thmag eclipsing binary (Porb = 11.12d) surrounded by a peculiar nebula. High resolution spectroscopic observations in Hα, He I (λ5876), [N II] (λλ6548, 6584), [A III] (λ7136) and [S III] (λλ9069, 9532) are presented. All emission lines show a complex profile with two main structures: a sharp and strong red component and a very broad fainter complex blue one in which at least three sub-components can be detected. We have analysed these profiles in order to derive information on the velocity field in the nebula. Permitted and forbidden lines show the same velocity field indicating a common line forming region. The presence of a multi-structured blue component in all profiles gives evidence that velocity gradients are present within the nebula. Systemic corrected expansion velocities of + 30 and −45,−30,−9 Km/s are found for the red and the three blue components respectively. Weaker emission structures, reflecting the same asymmetries of the strong emissions, are observed close to the permitted lines at larger velocities (Vb = −189 Km/s and Vr = 158 K m/s). The observed velocity field indicates an asymmetric mass outflow from the system very likely through the second Lagrange point rather than via stellar wind. The system should loose mass during the mass exchange phase. Using the masses (M1 = 39M⊙, M2 = 49M⊙) and radii (R1 = 37R⊙, R2 = 41R⊙) of the two components (Milano et al., 1981), and locating RY Scuti in the [log R, log M] diagram for early type contact binaries from Leung and Schneider (1979), we find that this system has just evolved-off the Terminal Age Main Sequence. Moreover, placing RY Scuti in the [log P, log M1/M2] diagram for O+O and WR+OB binaries of Massey (1981) we find that it is in an evolutionary phase just preceeding a WR+OB stage.

Current knowledge of the stellar winds and mass loss rates for WR stars is reviewed. Recent IR spectroscopy and reassessments of UV resonance line P Cygni profiles have led to revisions of terminal velocities, with v∞ ≃ 0.75 × previous estimates. Radio and IR (10μm) free-free emission for well-established thermal sources, coupled with recent considerations of the wind ionisation balance and chemistry, leads to WR mass loss rates lying in the range 10-5 − 10-4 M⊙ yr–1. This scale is confirmed by independent analyses of optical polarisation modulation in WR+O binaries. No significant differences are apparent between the mean mass loss rates of: (a) single and binary WR stars; (b) WN and WC stars, and (c) the WN and WC subclasses. The overall mean WR mass loss rate is ~ 5 × 10-5 M⊙ yr–1. Although WR radiative luminosities are uncertain, there may be a rough scaling of MWR with L∗, with a spread of up to an order of magnitude at a given L∗. WR winds have the highest momenta of the hot luminous stars, with values of M v∞ c/L∗ in the range 1-30 (WN7,8 and WC9 stars may lie near the lower bound). An additional mechanism to radiation pressure may be required to initiate the high WR mass loss, although thereafter the winds may be radiatively accelerated. Intrinsic variability in optical light, polarisation and emission lines, and in UV P Cygni profiles, indicate significant instability in the WR winds. For extragalactic WR stars in the Local Group, optical line strengths and widths do not suggest substantial differences in wind velocities and mass loss rates of subtypes compared to galactic counterparts.

The future use of space-borne IR spectroscopy to determine the ionization structure and abundances in the outer winds of WR stars is described. A mass loss rate of 1.7×10-5 M⊙ yr–1 has been derived from 10 μm photometry of the WO2 star Sanduleak 5 (WR 142). The He/H number ratios in the winds of P Cyg and AG Car have been derived from a recombination line analysis of their 1-4 μm spectra and mass loss rates of 2.2×10-5 M⊙ yr–1 and 3.7×10-5 M⊙ yr–1 have been respectively derived for them.