When the massive component in a close binary system evolves to fill its Roche lobe, mass transfer occurs and gas is accreted onto the companion star. Recently, the response of the unevolved secondary to accretion has been studied by a number of authors, but the emphasis has been on relatively massive stars which have a radiative envelope (Ulrich and Burger 1976; Flannery and Ulrich 1977; Kippenhalm and Meyer-Hofmeister 1977; Neo et al. 1977). The results show that the mass accepting star becomes overluminous and grows in radius until rapid mass transfer ultimately brings the two stars into contact. Such changes in the structure are caused by the steep increase in the specific entropy in the outermost layers and only a small amount of mass (about a tenth of the initial mass of the star) is accreted before contact is made. Thereafter, the expansion of the common envelope will lead to mass loss from the system. It is also found that, for a given accretion rate, the radial increase is much more conspicuous for a smaller mass star.

Mass loss can affect the evolution of binaries in various ways, during different stages of the evolution.

1. For massive stars stellar wind mass loss will change the masses of the components during their main sequence evolution.

2. During the Roche lobe overflow phase (or tidal interaction phase) matter can leave the system.

3. For low mass stars matter can leave the system during the mass exchange phase and can be stored in envelopes, disks or rings.

4. Sufficiently massive stars(>8-15M0) undergo at the end of their life a supernova explosion, where most of the matter is blown away and a compact object, a neutron star or a black hole can be left.

5. For intermediate stars one of the components can evolve into a degenerate He or CO dwarf; a reverse mass transfer can dump matter on this degenerate dwarf. If the conditions are favorable the white dwarf can explode with loss of matter, and a neutron star can be the result.

6. The chemical abundances in the outer layers change.

Mass outflow from interacting close binary systems, accompanied by loss of orbital angular momentum, appears to be a very important process affecting the evolution of binary stars. Together with accretion on the mass-gaining component, it is the least understood aspect of the general complex process we call “evolution with mass transfer and/or mass loss”, or, more briefly, “interaction”. It is therefore very imperative to assemble and examine all available facts or hints about mass loss.

High resolution ultraviolet spectroscopy of the early-type X-ray binary Vela X-l/HD 77581 provides a unique example to study the ionization structure of the stellar wind which varies under influence of the X-ray source. New results for other early-type X-ray binaries will be presented and compared with the Vela X-1 model. These results show that the observed variations strongly depend on the intensity of the X-ray source and on the density structure of the stellar wind.

The observations give us on one hand the mass ratio (q) distribution of OB+OB type binaries in a stage prior to Roche lobe overflow (RLOF), and on the other hand the q distribution of WR binaries which are considered to be in a stage after RLOF. By comparing both distributions one may hope to be able to say something about the mass loss during RLOF.

The remaining core hydrogen burning lifetime after a case B of mass exchange is computed for the mass gaining component in massive close binaries. Effects of stellar wind mass loss and mass loss during Roche Lobe OverFlow (RLOF) are included. Consequences for the evolutionary scenario are discussed.

R CrB at minimum light shows three type of spectra: 1) an absorption line spectrum similar to that observed at maximum light; 2) a sharp emission line spectrum mainly due to singly ionised metals which is displaced to the blue by 3 to 10 kms-1 with reference to the absorption spectrum observed at maximum. This is supposed to be a kind of permanent feature which exists even at the time of light maximum; 3) a broad emission line spectrum consisting of mainly λ3889 of Hel, H and K lines of Call and the D lines of Nal. The widths of the broad emission features indicate expansion velocity of 200 kms-1, much greater than the escape velocity (Gaposchkin 1963, Rao 1975).

We have examined the observational data of 100 Algols in order to check the validity of several simple models of non-conservative mass transfer. Strong evidence of mass and angular momentum loss has been found at least in about 20% of our Algols. Case B mass exchange is favoured for low-mass Algols, while case A predominates, though not so widely as expected, in Algols of higher total mass.

V 861 Sco (HD 152667) is a well known single line binary(1,2).The proposed association with a variable X-ray source(3,4) has triggered renewed interest in the system. Recent optical, infrared and ultraviolet observations(5 - 12) have provided a better knowledge of its characteristics. A first estimate of the mass loss was given by Hutchings(13).

Here we report on a series of photometric observations in the infrared (from 1.25 to 4.8 µ) taken at various orbital phases which improve and extend previous measurements by Tanzi et al.(12)

The results give evidence of a phase modulated infrared excess which can be interpreted in terms of a non isotropic mass flow in the system or, alternately, of a contribution from a colder secondary component.

The physical parameters of the components of V861 Sco are derived from light-curve analysis and published spectroscopy. Good agreement with evolutionary models is obtained. The stellar wind is investigated using IUE data; the results include no large phase dependence of the mass loss rate and insensitivity of the velocity of the wind (measured with respect to interstellar lines) to changes in the photospheric velocity, even near the base of the wind. However, small random changes in velocity near the base of the wind are amplified to larger changes further out.

SV Cen is a well known early-type contact binary. The less massive component is a slightly evolved B1 star. The secondary is of spectral type B4 III. The period of the system is P = 1.6585 days. SV Cen shows with Ṗ/P = -2 10-5 yr-1 one of the largest known period decreases. We carried out ground-based photoelectric and spectroscopic (coudé) observations at ESO, and high dispersion UV observations at VILSPA, with the aim

a) to derive new orbital parameters of SV Cen

b) to determine further epochs of minima (i.e. to study the period variations)

c) to look for possible mass loss indicators in the ultraviolet range.

In recent years the short period eclipsing binary RT And has been studied extensively by a number of authors. From photo electric observations the light curve and the photometric elements were derived by Dean (1974) and Mancuso et al. (1979). The geometric dimensions of the binary confirm that the system is detached consisting of two stars of spectral type F8 V and KO V. The mass ratio is 0.66. In spite of this classification, effects appearing among semidetached systems have been observed, namely, variations of the amplitude and the shape of the light curve and changes of the length of the orbital period.

Conditions are derived for supercritical winds, accelerated by the radiation pressure of an object above the Eddington limit. A relation between the different energy fluxes is found. The conditions seem to be satisfied by novae, and perhaps by η Car. They are not satisfied however by the symbiotic star Z And.

It is now known that strong stellar winds develop in stars mostly at the red and blue sides of the HR diagram. However, although the mass loss rates observed in O and M stars are comparable, the corresponding wind velocities are vastly different. It would thus be of great interest to find a binary system, containing both a cool and a hot star each with its own wind, and observe the resultant interaction. For a long time, α Sco (M1.5 Iab + B2.5 V) was the only known example (Kudritzki and Reimers 1978, van der Hucht et al. 1980). The situation in this case is best illustrated by a VLA map made by Gibson (1979) who finds that a shock develops at the surface of interaction of the two winds. In this paper I shall describe another binary system in which two stellar winds are interacting with dramatic effects.

Numerical simulations of the evolution of close binaries with primary mass between 3 and 15 Mʘ and mass ratios of ~ 1.5 were carried out. These results show that after mass transfer the system consists of a bright main-sequence.star together with a faint He-star in a rather wide orbit (p> 60d) Evolution of this kind predicts isotope anomalies for the CNO-elements with respect to their solar values.

In 1963, a diagramme has been established by the author, for visual and spectroscopic binaries, showing the existence of a real correlation between total mass and eccentricity.

The consideration of an up-dated material leads to-day to an identical diagramme confirming the reality and the stability of the mentioned correlation.

This diagramme is discussed with the assumption of a substantial secular mass-loss of the components of the considered binaries.

IUE observations at high and low resolution of Nova Aql 1918 show neither evidence of outflow nor the presence of nebular lines. This indicates that the shell ejected at the time of the outburst and surrounding the system for many years (Mustel and Boyarchuck, 1970) has by now disappeared.

The high excitation spectrum presents rapid variations in the far UV and eclipse effects in the near UV that seem well correlated with the orbital phase.

The observations can be interpreted in terms of phenomena occurring in or near the accretion disk surrounding the white dwarf. However, the small inclination of the orbital axis raises serious problems in trying to model the system.

The problem of the interaction of a strong stellar wind with a massive collapsing neutral hydrogen cloud has been considered and a new class of steady flow solutions found.

The detailed comparison of these models with one filamentary shell, N70 in the LMC, shows that they successfully explain all the observed properties of the HII region and enable mean mass-loss rates to be derived for the central stars.

We summarize here some of the results that we have obtained from chemical evolution models. In particular, we discuss those aspects related to stellar mass loss, helium production, a varying heavy element yield and the very light stars.

The observed value of the helium to heavy element abundance ratio ΔY/ΔZ is about 3 (e.g. Lequeux et al. 1979). First theoretical estimates gave ΔY/ΔZ= 0.4 (Hacyan et al. 1976).

Three types of spectra have been extensively observed among mainsequence B stars. B-normal is the classical absorption spectrum which defines a normal main-sequence B star. Be is a B-normal spectrum except : (i) some absorption lines, notably the first hydrogen Balmer series members, are replaced by emission lines; (ii) some lines from some singlyionized metals, not normally present in B stars, sometimes appear, either in emission or absorption. Be-shell is a Be spectrum with narrow and deep absorption cores in the Balmer and singly-ionized metal lines. A fourth type, B-shell, has been identified as a B-normal, absorption, spectrum except for the presence of FeII lines, and narrow, deep absorption cores in these and the hydrogen Balmer lines. Once thought to each represent a different kind of star, these spectra are now realized to simply represent different temporal phases, which one and the same star can traverse, apparently in no (as yet) fixed order. Some of the brightest stars --- eg γ Cas, 59 Cyg, Pleione --- have been observed in all of the 3 prominant phases; some stars, in only some of them; 70 % of the B stars have been observed only in the B-normal phase.