Observational determinations of density, dimensions, temperatures, and velocities of circumstellar features are surveyed and discussed, with a view to establishing limiting values that could be useful in any theoretical treatment of circumstellar structure. Densities of the order of 1013 particles cm-3 are found for streams and disks, although there is evidence for regions of much lower density in many systems, and some systems may have much denser circumstellar matter.

Dimensions of disks seem to be remarkably constant (fractional radius ∼ 0.3) in a wide variety of systems. The total mass of circumstellar matter is always a small fraction of the mass of the system. Temperatures are usually similar to the temperatures of the stars in the system, although hot spots certainly exist in some systems, and ‘flare’ activity is evidence of localized very high temperatures. On the other hand, some systems may contain solid circumstellar matter. Observed velocities are usually several hundred km s-1. The empirical relation between the velocity of rotation of disks and the orbital period is described. Finally, the stability of these features is briefly discussed.

A distinction is drawn between radial and tangential modes of ejection from stars, and the possible flow patterns are described. They are: expanding streams, falling streams, jet streams, circulatory streams, and gaseous envelopes. Motion around Be stars is discussed at some length, as a preliminary to studying more complicated flow in binary systems. The rotational velocity of the Be star is insufficient to form the ring. It appears likely that radial instability is temperature sensitive. Rings and disks in binary systems are discussed from the point of view of periodic orbits for particles within the gravitational field of such a system. The formation of these rings is discussed. The expected relation between rotational velocity of the ring and the orbital period is discussed. The relation of circumstellar streams to period changes is mentioned. Finally, the influence of magnetic fields on the circumstellar material and the system is discussed.

Observations of Be stars are summarized and discussed, with a view to establishing dimensions of, and physical conditions within, their envelopes. Shell stars and Be stars are considered together, the observed differences between them being considered as due to the size of the envelope and the angle that the rotational axis makes with the line of sight. Rotational ejection of matter is insufficient to account for the formation of the envelopes. Theories of the structure of envelopes are also discussed. The envelopes of Wolf-Rayet stars are considered in a similar way. The interpretation of these is complicated by the known binary nature of many Wolf-Rayet stars. Other groups of stars possessing envelope-like structures – U Gem stars and old novae, symbiotic stars, and T Tauri stars – are briefly mentioned. The problems of stellar envelopes, first listed by Struve thirty years ago, are still unsolved.

Although the Struve Symposium was planned to discuss conditions in binary stellar systems, it is necessary to consider the observations of extended atmospheres in single stars in order to compare them with those of binary systems. The Sun, being the closest star, has the first chromosphere that was observed, and we can assume that the many features observed on the Sun will be much enhanced in giant stars with more extended atmospheres. Among the early-type stars, the hydrogen emission lines in the spectra of Be stars are known to vary and this suggests the atmospheres are unstable. Similar, even more pronounced effects, are seen in the spectra of shell stars such as 48 Librae. The supergiant stars have also been shown to have spectral lines that change, often quite rapidly, with time. Smolinski has observed in the spectra of some F and G-type supergiants much-broadened absorption lines that also vary with time. Wilson's systematic studies of the H and K emission lines in late-type stars have greatly extended our knowledge of stellar chromospheres, and the ionized magnesium lines at 2800 Å should give additional information as more extra-terrestrial observations are obtained. Numerous observations of the ζ Aur stars have been obtained near eclipse; ζ Aur and 31 Cyg seem to be similar systems and show similar phenomena – a rise in temperature outward from the photosphere and an extensive, variable calcium chromosphere, but a viable theory for these chromospheres has not yet been given. VV Cep and 32 Cyg probably are larger and more complex systems. The emission lines in the ultraviolet spectra of the M-type stars may be indicators of coronae in these cooler atmospheres. Most variable stars seem to have extended atmospheres and, when detailed studies are made, show changes in both light and spectrum. Shock-wave phenomena are undoubtedly important in the analysis of many of these stars, as also are fluorescent, selective-excitation processes. Flare stars, though generally dwarfs, are of great interest in studies of stellar atmospheres.

Extended atmospheres can be detected from observed gross deviations from local thermodynamic equilibrium, as in the study of emission lines, large velocity fields, and dilution effects. It is probable that the presence of a companion affects the atmosphere of the primary star in close double stars, but in stars like ζ Aur and 31 Cyg, the atmospheric effects produced near eclipse may be among the best clues for the interpretation of extended stellar atmospheres.

(A) The possibilities and the difficulties of a theoretical study of extended atmospheres in binaries are briefly discussed.

(B) We try to summarize and discuss critically the present status of the theory of three types of extended atmospheres (i.e. atmospheres in which the average photon mean-free-path is the same order of magnitude or larger than the stellar radius):

(1) Extended atmospheres in hydrostatic and in grey or non-grey radiative equilibrium.

(2) Dynamic (expanding) atmospheres which occur if the radiative acceleration is slightly smaller than the acceleration of gravity.

(3) Stellar coronae which are formed in the presence of a mechanical energy flux.

In (1) we study the importance of the ‘forward peaking’ of the radiation field in the outer layers of the atmosphere. The possibilities for the solution of the non-grey transfer problem in an extended atmosphere are discussed.

In (2) we pay special attention to Marlborough's and Roy's (1970) result that the atmospheric gas cannot be accelerated directly to supersonic velocities by the action of the radiation force.

In (3) the large differences in the coronal properties of stars of different chemical composition are emphasized. We draw attention to the partially unexplored but probably very interesting properties of coronae of helium-rich stars.

Several groups of close binary stars are considered in an attempt to explain their present state as a consequence of a large-scale mass transfer or mass loss in the past: Algol-like semidetached binaries, some shell stars (AX Mon), some binary X-ray sources (Cen X-3, Her X-1), the recurrent nova T CrB, helium-rich binaries v Sgr and KS Per, and the symbiotic variables.

Algol-like binaries like U Sge cannot be products of a conservative case A of mass transfer; rather, mass loss from system and/or a temporary contact stage must be invoked. Nova T CrB as well as the symbiotic variables probably contain a mass-losing giant and a helium star, which again may be a product of a previous mass transfer of type B. Similarly, some of the X-ray sources may actually be binaries undergoing a second process of mass transfer. The systems v Sgr and KS Per may contain helium stars expanding to the right of the helium main sequence, while the other component may be a rather inactive main-sequence star. Some shell stars may be products of mass transfer. Mass loss from convective envelopes is also discussed.

Loss of mass and of angular momentum from many binary systems must be anticipated. Behavior of the mass-accreting stars may often be decisive for the appearance and evolution of the system.

I have been assigned the task of presenting a summary of the Conference that has just finished. Actually during the Symposium the review papers have touched upon many items, some of them apparently not strictly connected with the subject of the meeting but helpful to clarify the picture related to close binaries and aimed at establishing differences with single stars, whenever possible. Therefore, my task would be somewhat difficult it if were not that all the review papers were extremely clear, very well given and fully discussed and that there were so many contributions from the floor presented without any pressure from time or otherwise. Furthermore most of what I would say has been already said. All this makes my task a lot easier as it appears clear that I should confine myself to offer only a few comments. Let me start by saying that Dr. Struve would have felt most pleased to have been present at this meeting and realize how many astronomers are now interested, strongly interested, in the field of close binaries and how much the field has and is expanding taking also up the efforts of researchers that work in the radio and X-ray regions and attracting the attention of those who are interested in problems of general relativity.