The subject allotted to me in the program is the ‘Summary of Symposia I–III: problems considered and present status of them’. Perhaps I should have been more cautious when I agreed to this title about two years ago. For it means that, if I do my job well, you will know the present status of all problems and the Symposium is finished today. However, I am sure that, no matter how I try, there will remain enough for other speakers and for other days of discussion. For this reason, in surveying the field, I shall not consciously try to avoid the topics of other speakers. Overlap will do no harm.

This discussion on the opening day of the Symposium started in the afternoon and was consequently rather short. The version presented here is even shorter since some extensive remarks by Zel'dovich and Sunyaev have been transferred to a more proper location, the discussion following the Reports by Weaver and by Field (p. 77).

The task assigned for this review is to discuss structural features of the Galaxy as they may involve or relate to gas dynamics. Since a topic of such great breadth would permit discussion of essentially any aspect of the Galaxy, we shall narrow the view and direct attention primarily towards the larger-scale features of the system. In particular, effort will be made to place emphasis on those areas of the subject in which problems of interpretation exist and in which new theoretical models are needed.

As excellent reviews have appeared recently (Spitzer, 1968a; Pikel'ner, 1968a) we shall consider here only some special topics where progress is being made – thermal properties of Hi and Hii clouds, and shock waves. The components of the medium which are dynamically important are neutral atoms, charged ions and electrons, grains, cosmic rays, magnetic field, and light. In Table I we indicate the degree of momentum coupling among these components. The main mass of the medium, in ions and atoms, is well coupled through collisions, with a mean free path λ ≈ 10−3 pc, and may usually be treated hydrodynamically. Parker (1965) has shown that for many purposes the cosmic-rays can be considered as a separate gas, strongly coupled to the magnetic field. The coupling of the atom-ion gas to the magnetic field is through the magnetic forces on the ions, which transmit the forces to the neutral atoms by collisions unless the differential stresses on the two components are unusually large. Grains are not coupled strongly to any other component, as their stopping distances are of the order of parsecs and their gyro-periods of the order of 104 yr. The coupling of grains to radiation pressure can lead to interesting differential effects. Light is important in carrying away thermal energy but does not critically affect the momentum balance of the gas. A possible exception is the Ly-α radiation field discovered by Kurt and Sunyaev (1967) which has a pressure of 10−13 dyne cm−2. If this phenomenon is common in interstellar space, it could have effects.

This long, very intensive, and partially very confused discussion, has been rearranged in six sections: (1) Direct and Indirect Evidence of X-rays and Low-Energy Cosmic Rays; (2) The Boundary Layer between the Stable Gas Phases; (3) Theoretical Aspects of Interstellar Gas Dynamics and the Formation of Clouds; (4) Observational Aspects of Interstellar Gas Dynamics and the Formation of Clouds; (5) Observations of the Rarefied, Neutral Intercloud Medium and of the Interstellar Electron Density; (6) The Dynamical Theory of Hii Regions. Section 2 has been transferred from the Discussion on Monday, September 8 (Chapter 2). To Section 6 have been added remarks made during various discussions. A couple of remarks have been transferred to other Discussions. Part of the Discussion (in Section 5) was very confused; an attempt has been made to condense and to make as much sense as possible out of what was said. For the convenience of the reader I recapitulate a few concepts, the (mis)-use of which lead partially to the confusion:

1. (i) The hydrogen surface density or column density NH = ʃ nH dl (NH is sometimes called the hydrogen measure HM).

2. (ii) The dispersion measure DM = ʃ ne dl (DM is often called the electron surface density Ne).

3. (iii) The rotation measure RM = c1 ʃ neB‖ dl.

4. (iv) The emission measure EM = ʃ ne2 dl.

   In these definitions, nH represents the hydrogen density, ne the electron density, B‖ the component of the magnetic field strength along the line of sight and l the distance along the line of sight. Conventionally nH and ne are expressed in cm−3, B in μG and l in pc. In addition there are two combinations of these quantities (and of the electron temperature T) involved in the Discussion:

5. (v) The free-free absorption coefficient k(v) = c2(v)ne2T−3/2 (v is the frequency).

6. (vi) The free-free emissivity ɛ(v)=c3(v)ne2T−1/2 (only at radio wavelengths). c2 and c3 depend also on T, but only rather weakly.

On Friday afternoon, September 12, an informal meeting was held to discuss the distribution of neutral interstellar matter. Chairman of the meeting was Thomas. Summaries were presented at the formal discussion meeting, later the same afternoon. No change of importance has been made in the original discussion. The text has been inserted where it seemed most functional, namely closely following the Discussion on the Reports by Weaver and Field.

The First Symposium on cosmical gas dynamics clearly demonstrated that magnetic fields play an essential role in the dynamics of the interstellar matter. The Symposium emphasized that the behavior of interstellar gas should be described not by traditional gas dynamics but with the help of magneto-gas dynamics. As we now realize this was only the first step, since in plasmas a very broad range of phenomena can occur which are coupled not with large-scale magneto-gas dynamic motions but rather with more subtle collective plasma motions. To describe properly the dynamics of an ensemble of charged particles we should use not only average characteristics such as density, pressure and mean velocity, but also more detailed features of the particle distribution function.

The following discussion was long and chaotic, in spite of the energetic efforts of the Chairman. Apparently the subject was unfamiliar to many participants. I did quite some reshuffling. A long contribution by Ozernoi on photon whirls in the primeval Metagalaxy has been transferred to Chapter 13.

I will review the impressive advances in the observations of the galactic magnetic field made since the time of van de Hulst's review at the 1966 Noordwijk Symposium (van de Hulst, 1967). Most of these observations are so recent that the consequences have not yet been worked out very well and are in need of discussion. Luckily the emphasis in the present Symposium, unlike that in many others, is on discussion. For this reason I will not hesitate to include in my review provocative speculations. I will base my discussion on Mathewson's (1968) elegant magnetic field model, consisting of a local field in the form of a sheared helix, superimposed on a large scale longitudinal field. I will include the criticism of this model by Gardner et al. (1969b). Mathewson has succeeded in accounting for such data as the distribution of background polarized radiation from the Galaxy, the distribution of rotation measures of extra-galactic radio sources and even the spurs and ridges in galactic continuum emission. There are critics who inherently distrust models that account for too many things at one time, but I feel that we should try to account for as many things as possible with the least number of models. Mathewson has succeeded in uniting much data and I will only add a few pieces to his model.

The topic of this presentation is the origin and dynamical behavior of the magnetic field and cosmic-ray gas in the disk of the Galaxy. In the space available I can do no more than mention the ideas that have been developed, with but little explanation and discussion. To make up for this inadequacy I have tried to give a complete list of references in the written text, so that the interested reader can pursue the points in depth (in particular see the review articles Parker, 1968a, 1969a, 1970). My purpose here is twofold, to outline for you the calculations and ideas that have developed thus far, and to indicate the uncertainties that remain. The basic ideas are sound, I think, but, when we come to the details, there are so many theoretical alternatives that need yet to be explored and so much that is not yet made clear by observations.

The Discussion has been rearranged into four sections: 1. Various Questions; 2. The Galactic Halo; 3. The Origin of the Magnetic Field; 4. Summary. Section 4 is the outcome of another meeting (on September 17). One very extensive remark has been condensed. The latter part of the discussion on September 12 dealt with the informal meeting on Hi gas and is presented separately as Chapter 6.

The term accretion originally referred, in astronomical contexts, to the capture of mass by stars, and, later, to mass capture by other centers of gravitational force. As such, the process has not proved to be of general importance, in spite of early hopes. However, there are other aspects of the problem which may yet prove worthy of attention in interstellar gas dynamics. In particular, the effects of stars, galaxies, or even clusters of galaxies, on ambient matter streaming by them may be detectable, directly or indirectly, and it is on such possible effects that I shall concentrate here. These effects are related to the original accretion problem but may be thought of separately; nevertheless, I retain the use of the term accretion to refer to all aspects of the motions induced in an ambient medium by a gravitating object.

[The text of this Chapter contains the first part of the Discussion on Thursday, September 18, 1969; the second part is given in Chapter 26. I added a long contribution on cosmogonical problems, presented by Dr. Ozernoi on September 10.]

In this paper I will deal almost exclusively with the interchange of matter between the interstellar gas and the stars. Other possible forms of interchange, such as mass-loss to and accretion from intergalactic matter e.g., Oort's (1969) views on mass flowing into our Galaxy, will be discussed later I hope.

The nature of the supernova event is still poorly understood. A variety of models has been proposed and the interpretation of the observations remains ambiguous. About all that is certain is that in the supernova event much of the matter is ejected at speeds of typically 10 000 km sec−1 and that about 1049 erg of visible light is emitted. The amount of matter ejected and the bolometric correction to be applied to the optical radiation are very uncertain. The discovery of a pulsar in two supernova remnants (Crab Nebula and Vela X) suggests that frequently a neutron star or other condensed object results following the outburst.

During the second half of this Discussion van de Hulst commented on the overall flow of energy in the interstellar medium. That part of the Discussion has been transferred to the Final Discussion (p. 362). The rest has been rearranged into the form presented here.

When Parker developed the theory of the solar wind, he suggested that other stars might also be surrounded by an expanding atmosphere. This suggestion has been worked out by Deutsch (1958, 1960, 1961, 1969) and by Weymann (1960, 1963), who discussed coronal evaporation as a possible mechanism for mass loss in red giants and used hydrodynamical equations for describing the phenomena. During the last few years detailed observations of the solar wind have been made, and the theoretical approach has also been further developed. Both might help us in our understanding of problems of a quasi-stationary mass loss of stars. But besides its relevance for the study of other stellar winds, the solar wind is important, as it is the best observed cosmical plasma available for guidance in the study of other cosmical, collisionless plasmas. What makes the interplanetary plasma so important is that beyond several solar radii from the Sun there are no collisions among the gas particles (one solar radius [R⊙] is 7.0 × 1010 cm; 1 AU is 215 R⊙). Only near the Sun the collision rate is sufficiently high to maintain isotropy of the thermal particle motions and to produce ordinary fluid behavior. At larger distances from the Sun the fluid behavior is maintained by irregularities in weak magnetic fields and by micro-instabilities in the plasma.