My contribution to this Joint Discussion can hardly be termed very novel or original. Rather, I just want to call your attention here to two headaches - one perhaps only a hangover, but the other a real migraine. Sooner or later these two headaches beset anyone who persists in asking how the random motions of stars in the only vicinity where we tolerably observe them can possibly jibe with the sensible presumption that not only our neighborhood but this entire Galaxy should by now be reasonably stable.

There is of course nothing strictly ‘local’ about any problem involving the far-reaching gravity. Yet if some conceivable instability of our collection of nearby stars comes even close to deserving such a label, it is surely the Jeans instability or tendency toward gravitational collapse. After all it is not difficult to estimate (e.g., Toomre, 1964) that the most troublesome of such incipient clumpings ought to have significant dimensions (such as half-wavelengths) of the order of 3 or 4 kpc. Even the latter scale is probably not too large a fraction of our distance from the galactic center for ‘local’ analyses in the WKBJ spirit to remain coarsely trustworthy.

This paper elaborates the postulate that the Earth and Moon became a binary system during their accretional development and that the Moon’s growth was essentially completed before the assumed solar nebula dissipated. The solar nebula was still hot enough at the formation of the two bodies that both consisted largely of the refractory and relatively low-density minerals now characteristic of the Moon. During the subsequent condensation, agglomeration and accretion of siderophile and more volatile higher density minerals, the Earth grew very much faster than the Moon because of (a) its much greater gravitational capture area coupled with retention by a sizeable atmosphere and (t>) the Moon’s velocity, with respect to the solar nebula, which produced a wind that aerodynamically blew away volatiles and smaller debris resulting from hypervelocity impacts of larger planetesimals. This ‘impact differentiation’ process favored the retention of the refractory minerals on the Moon (Figure 1). The Moon’s surprisingly high moment of inertia follows naturally from the basic postulate.

Objects of the solar system, in addition to the Sun, can be classified into four groups -the planets, objects of lunar mass, smaller objects of variable mass and the comets.

If the solar proportion of gases relative to non-volatile compounds of the variety in the terrestrial planets, namely about 300 times the mass of these elements, were added to the terrestrial planets, they would have masses comparable to those of the major planets. Mercury is low in mass but has a high density, indicating that it has lost several times its mass of silicate materials relative to high density metallic iron. If this were restored and then the component of gases were added, it would also fall into the group rather naturally. Mars appears to be rather small. Uranus and Neptune have rather high densities indicating some loss of gases, probably hydrogen and helium. When we attempt to estimate the mass of primitive solar material from which the planets were evolved, we conclude that they evolved from very similar masses. Later, I shall argue that the process was a very inefficient one.

Satellites are a common feature in the solar system, and all planets on which satellite orbits would be stable possess them. (For Mercury the solar perturbation is too large, and the retrograde spin of Venus would cause satellites to spiral in to the planet through tidal friction.) An explanation of the formation of satellites must hence be one which makes the phenomenon exceedingly probable at some stage in the solar system formation processes, and very improbable processes like a capture cannot be the answer in most cases.

Small particulate matter must have been very abundant in the early solar nebula. Such particulate matter must have existed both from the first condensation of the low vapor pressure components of the gas in the first round, and it must also have been composed of material scattered from impacts after some major bodies had begun to form, frequently finding themselves no doubt on collision orbits.

Most theories of the origin of the solar system begin by assuming that, at the time of its formation or shortly afterwards, the Sun somehow acquired a disc-like ‘solar nebula’ in which the planets later formed by accretion. The presently available gravitational collapse calculations, while not yet sufficiently sophisticated to delineate in detail the processes involved, at least support the plausibility of this idea and suggest that the solar nebula may have formed by the accretion or capture of leftover proto-stellar material into orbit around the early Sun. Infall of matter into the disc probably continued for roughly 105 to 106 yr, and may have produced transient strong heating effects through the thermalization of the infall energy in shock fronts. It seems likely that the satellite systems formed in a basically similar way through the capture of residual solid matter into orbit around the planets.

A more complete discussion of gravitational collapse and the formation of the solar system has been given in The Origin of the Solar System, ed. H. Reeves, p. 142. (Paris: CNRS).

The intuitive notion that a satellite system will change its configuration rapidly when the satellites come close together, and slowly when they are far apart, is generalized to ‘The Principle of Least Interaction Action’, viz. that such a system will most often be found in a configuration for which the time-mean of the action associated with the mutual interaction of the satellites is a minimum. The principle has been confirmed by numerical integration of simulated systems with large relative masses. The principle lead to the correct prediction of the preference, in the solar system, for nearly-commensurable periods. Approximate methods for calculating the evolution of an actual satellite system over periods ˜ 109 yr show that the satellite system of Uranus, the five major satellites of Jupiter, and the five planets of Barnard’s star recently discovered, are all found very close to their respective minimum interaction distributions. Applied to the planetary system of the Sun, the principle requires that there was once a planet of mass ˜ 90 Mθ in the asteroid belt, which ‘disappeared’ relatively recently in the history of the solar system.

Professor G. R. A. Ellis reviewed the wide range of radio emission from Jupiter. At centimetric wavelengths the thermal radiation corresponds to a blackbody at 130K. Between 2 m and 10 cm wavelength there is a powerful component of synchrotron radiation from the electrons trapped in the radiation belts. At longer wavelengths there is a great variety of impulsive radio emission from coherent plasma oscillations.

The magnetic field of Jupiter is known from the polarisation of the synchrotron radiation to be situated centrally (within one tenth of the radius) and inclined at 10° to the rotation axis. The radiating electrons have energies of the order of 10 MeV, and a density of 10”-3 cm-3, much greater than in the case of the Earth’s radiation belts.

The decametric radiation varies with the rotation of Jupiter, possibly analogously to pulsar radiation. Bursts at around 4 MHz reach very high brightness temperatures, exceeding 1017 K. The occurrence of these strong bursts is closely related to the position of the Jovian satellite Io, which must have an interaction with the main magnetic field.

There are two methods to study the chemical composition of the envelopes ejected by novae: (a) the analysis of the absorption spectra of novae; (b) the analysis of emission lines during the nebular stage.

Nova V603 Aql 1918 was uniquely suited for studies of the structure of its shell because of the very favorable orientation of the shell in space. Additionally, almost by chance, spectroscopic observations were made in such a way that they permitted derivation of a three dimensional model of the shell.

V603 Aql, as we see from its light curve (Figure 1), was a typical fast nova which showed strong periodic fluctuations in light during the transition stage. The period of fluctuation was approximately 10 days, during which time the visible light changed by roughly 50%.

Absolute spectral energy distributions for supernovae of both types I and II have been obtained. These observations demonstrate three facets of supernova spectra. First, both SN I’s and SN II’s have a continuum that varies slowly and uniformly with time, and which carries the bulk of the radiated flux at early epochs. Second, some lines in both SN I’s and SN II’s have P Cygni profiles: broad emissions flanked on their violet edges by broad absorptions. Third, some lines are common to SN I’s and SN II’s and persist throughout the evolution of the spectrum. The continuum temperatures for both SN I’s and SN II’s are about 10000 K at the earliest times of observation and drop in one month’s time to about 6000 K for SN II’s and about 7000 K for SN I’s. After several months, the continuum may cease to carry the bulk of the flux, which might be in emission lines, but continues to exist, as shown by the presence of absorption lines. The P Cygni line profiles indicate expansion velocities of 15000 km s-1 in SN II’s and 20000 km s-11 in the SN I 1972e in NGC 5253. Line identifications for SN II’s include Hα, Hβ, H and K of Ca II, the Ca II infrared triplet at λ8600, the Na I D-lines, the Mg I b-lines at λ5174, and perhaps Fe II. The [O I] lines λλ6300, 6363 and [Ca II] lines λλ7291, 7323 appear after eight months. For SN I’s, the lines identified are H and K of Ca II, the infrared Ca II lines, the Na I D-lines, and the Mg I b-lines. There is some evidence that Balmer lines are present two weeks after maximum. The strong and puzzling λ4600 features drifts with time from λ4600 near maximum light to λ4750 after 400 days.

One of the principal sources of information about supernovae are the spectra of these stars. Thus we are going to discuss mainly the spectra of type I and type II supernovae around light maximum (tmax).

The present structure and recent changes in Cas A are discussed. On the deepest available exposures the optical remnant of this supernova is seen to consist of an almost complete shell. The southern part of this shell is outlined by knots that have developed during the last decade. It is pointed out that moving nebulosity in which [O III] radiation is particularly strong is distributed differently from nebulosity in which the [O III]/[S II] ratio is more nearly normal. Observation of the motions of individual knots suggests a model in which dense blobs of matter, which have a highly anomalies composition, become luminous as they plough through a stationary interstellar (or circumstellar) medium.

Observations of a number of supernova remnants have been carried out with the low energy X-ray telescope on the Copernicus satellite. Data are presented on the X-ray structure of the remnants Cassiopeia A and Puppis-A. Marginal detections or new upper limits are reported for the remnants IC443, DR4, MSH15-52A, Downes 83, Downes 84 and 3C392.