It is, I believe, of fundamental importance to find out how far the structure and motions in the galaxies can be explained by gravitational theory. This does not deny that hydrodynamic and electromagnetic forces may play an important part. The special point which I have in mind is the possibility, from the standpoint of gravitational theory, for the spiral structure to be a kind of quasi-steady state of comparatively long lifetime.

I. The maps of the spiral pattern of neutral hydrogen in the outer and intermediate regions of the Galaxy resulting from the 21-cm studies at Leiden and Sydney may be compared with the optical structure of other galaxies. If, as seems probable, most hydrogen arms coincide with or follow closely the optical arms marked by the blue supergiant stars, a comparison of the HI pattern with photographs in blue light should be meaningful. Such a comparison may lead to the identification of the galaxy type to which our Galaxy belongs. The regularity and symmetry of the HI spiral pattern indicates that our Galaxy is a normal system, and one to which, in principle, a definite type in the revised classification scheme (de Vaucouleurs 1958, 1959) can be assigned. If this assignment can be made successfully it should assist in the interpretation of the radio observations of the inner regions. There is a good possibility of at least restricting the range of galaxy types consistent with the HI spiral pattern of the outer regions because there is a high degree of correlation between galaxy type and multiplicity of the spiral pattern.

I. Expansion of the Gaseous and Stellar Components of the Galaxy

If the gaseous component of the Galaxy is expanding as observed by Rougoor and Oort in the centre of the Galaxy and as postulated by Kerr in his early interpretation of spiral structure, the expansion must represent a phenomenon of fundamental importance in the Galaxy which has, in all probability, been operative for a significant fraction of the age of the Galaxy. Presumably, very young stars formed from this gas and having ages less than 1 % of the age of the Galaxy might be expected to retain in their motions the general character of the large-scale expansion of the gas from which they originated.

The great nebula in Andromeda, M31, is a particularly interesting external galaxy, since it is a giant spiral system, presumably much like our own Galaxy, and is close enough to be resolved with existing radio telescopes. During parts of November 1962, and from December 5, 1962, to January 20, 1963, the 300-foot transit telescope of the N.R.A.O. at Green Bank, W. Va., was used in conjunction with the Carnegie multichannel H-line spectrograph to study M31. The following is a preliminary account of these first observations.

Study of the radio continuum emission from the Galaxy is of interest both from the viewpoint of the physics of the radiating processes and the light which these shed on physical conditions within the Galaxy, and also because the large-scale organization of the emitting regions gives information about the structure of the Galaxy as a whole. We are concerned here particularly with the latter aspect, that is, the question of galactic structure, and for this we need to know the distribution of the emitting regions throughout the Galactic System.

A survey of the distribution of radio emission at wavelengths of 75 and 20 cm along an extensive section of the Southern Milky Way has recently been carried out by M. Komesaroff and myself using the 210-foot radio telescope at Parkes. The area surveyed ranges from lII = 280 to 355° and extends on the average to 6 degrees either side of the plane. The beamwidths of the aerial at 75 and 20 cm are 50 and 14 min arc respectively. Observations at the shorter wavelength offer a picture of this section of the Galaxy at considerably higher angular resolution than hitherto available, while the 75-cm observations are expected to make a useful contribution to our spectral information.

La structure de restes de supernovae galactiques de petit diamètre apparent a été étudiée sur 1420 Mc/s avec l'interféromètre à 2 antennes de Nançay, dont la résolution maximale est 18″ EW et 1′ NS (Lequeux 1962). Taurus A a des dimensions de 3′,9 × 2′,7 à demi-intensité, l'angle de position du grand axe étant 126°. L'émissivité dans cette radiosource croît rapidement vers le centre; d'autre part ses dimensions augmentent avec la longueur d'onde, ce qui peut s'expliquer d'après Woltjer (1958) par une création continue d'électrons relativistes au centre de la radiosource. Cassiopeia A, au contraire de Taurus A, est parfaitement circulaire; en admettant la symétrie sphérique, nos observations montrent que presque toute l'émission radioélectrique provient d'une enveloppe sphérique de 2′,0 de rayon et 0′,6 d'épaisseur à demi émissivité. Les résultats obtenus sur les Supernovae de Képler et Tycho Brahé sont en accord avec ceux de Cal Tech sur 960 Mc/s: la Supernova de Tycho Brahé est peut-être une enveloppe sphérique, mais on ne peut rien dire de celle de Képler.

Radio studies of galactic HII regions are best carried out at the two ends of the radio spectrum. At high frequencies, of hundreds or thousands of megacycles per second, HII regions are seen by virtue of their thermal emission against a weak nonthermal background. Since radio waves are unaffected by the obscuration along the plane, it is possible in principle to see right through the Galaxy, and the high resolution which can be achieved in the thousands of megacycles range enables us to study at least the nearer regions in considerable detail. At low frequencies, below about 20 Mc/s, ionized hydrogen is seen in absorption against a bright nonthermal background. Since quite tenuous regions may be almost opaque at the lower frequencies, the technique provides quite a sensitive method of detecting them. The absorption increases with decreasing frequency so that studies at different frequencies enable us to see to varying depths along the line of sight and could permit the derivation of rough distance estimates.

It is well known that the Large Magellanic Cloud contains in addition to a considerable number of ordinary O-associations a certain number of large objects which, however, are similar in nature to the associations. These objects were named “constellations” by Shapley. But the large complex 30 Doradus surpasses notably all of these objects both in diameter and in absolute brightness. The latter is of the order of −15m0 while its diameter is of the order of 600 pc. If we take the average absolute brightness of associations in our Galaxy as equal to −10m0 then it turns out that 30 Doradus is 100 times more luminous than the ordinary associations. The photographic images of more distant galaxies reveal that sometimes complexes occur in them of the same order of luminosity and dimensions as 30 Doradus. Therefore it seems to us useful to regard these complexes as a special class of objects and call them superassociations.

van de Hulst (1958), in giving representative sizes of the structural details of the interstellar medium, lists the dimensions of a typical cloud, seen in 21-cm emission, as 20–70 pc but warns that we should not be surprised to find different “sizes of interstellar clouds” determined from different sets of data.

In a study made by Miss Gail Smith at Leiden a cloud was found with a rather exceptional velocity and a large velocity spread. The cloud is elongated, with a ratio of length to breadth of about 3 : 1. Its length is about 10°. The average velocity relative to the local standard of rest is +90 km/sec; the width of the profiles is considerable, corresponding to a halfwidth of about 30 km/sec. The mass can be estimated only if we know the distance. Taking 200 pc as a reasonable estimate of the minimum distance of the cloud the minimum mass is about 200 solar masses. The computed mass varies as the square of the distance.

The 300-foot transit telescope at the N.R.A.O. at Green Bank, West Virginia, was used in conjunction with the Carnegie 90-channel H-line spectrograph to study the far parts of the Galaxy between l = 11° and l = 50° (lII and bII are to be understood throughout). A travelling feed was constructed at the Carnegie Institution which enabled us to track one point in the sky for approximately 5m40s. This made it possible to make two completely independent frequency scans of a given point each day. On successive days, duplicate curves were taken with the frequency scale shifted, in order to guard against single channel anomalies and pointing error. Line profiles were taken at 160 points in the region under study, in meridional sections every 5° of l, from b = −0°8 to +2°0 in steps of 0°. In addition, a single longitudinal section was taken at b = +0°4 in steps of 1°.

The Zeeman effect can be used to measure directly the longitudinal component of the magnetic field in interstellar neutral hydrogen clouds. The frequency separation between the two circularly polarized components is 28 c/s for 10–5 G and can be inferred from measurements of the intensity difference between left- and right-hand circular polarization as a function of frequency. Earlier experiments at Jodrell Bank showed that the mean field in the interstellar medium was less than 10–5 G (Davies et al. 1960). Recent work using more sensitive techniques has provided a positive measurement of a weak general magnetic field and of fields of varying intensity in different interstellar clouds.

Measurements of polarization of radio emission from the galactic background were started in 1957 and seemed to be partially successful, but so uncertain that not much weight was attached to them. Only in the last two years have such polarization measurements been fully successful. They were carried out by radio astronomers at Leiden, Cambridge, and Sydney. I shall report here on the measurements made by the Leiden group, and draw some general conclusions. The first year of observations was described by Westerhout et al. (1962) and Brouw, Muller, and Tinbergen (1962); the work has been continued since by Muller, Brouw, Tinbergen, and Berkhuysen (1963). The observations made by the Cambridge group are substantially in agreement with the Leiden work (Wielebinski, Shakeshaft, and Pauliny-Toth 1962), and so are the Sydney observations, on which Dr. Gardner reports in another paper (this volume, paper 35).

Polarization effects involving our Galaxy have been studied in two ways: firstly, by observations of the polarization of the galactic radiation itself, and secondly, by the effects of galactic magnetic fields and ionization on the polarization of extragalactic radio sources. For an aerial of fixed dimensions, the intensity of the background varies at about λ2·5 (λ is wavelength) and that of the sources at about λ0·7. Thus background measurements are restricted to the longer wavelengths where the radiation is sufficiently intense, while sources are most easily studied at the shorter wavelengths where background confusion is small. The background radiation includes a component from small unresolved extragalactic radio sources in addition to the galactic radiation which is of interest in the first type of measurement. Both types of observation have been made with the 210-foot aerial at Parkes; the observations of sources were done in collaboration with J. B. Whiteoak and that of the galactic radiation with J. A. Roberts, assisted by J. B. Whiteoak and A. G. Little for part of the time.

It has become clear in recent years that the spiral features of our Galaxy — like those of all galaxies — are of recent origin and are presumably short-lived phenomena. To trace them optically, we need to confine ourselves to concentrations in the interstellar gas and to stars and star groupings recently formed from these. We are hence limited primarily to OB associations and star clusters in which the earliest spectral types for the stars are not later than B2, preferably O5 to B1. It is most important that radial velocities be measured for a fair sampling of these stars, especially so for directions in which the radial velocity effects due to galactic rotation vary appreciably with distance. Radial velocities of the stars in question and those found from interstellar absorption lines are not only useful as indicators of distance, but they are very much needed for the identification of star groups and their associated HI clouds, found by 21-cm techniques. They assist also in the study of HII regions, which can now be located by either radio or optical methods.

Four main methods of delineating spiral structure in the Galaxy have been developed in the past 10 years.

In an earlier article* it was pointed out that the galactic radial motions ΔE (R, l) of the very young stars did not show the uniformity of motion to be expected from a smooth regular expansion of the Galaxy. Instead, the very young stars were found to show large-scale regional peculiar motions; these regional peculiar motions are displayed in Figure 1. In addition to regional peculiar motions and the space distribution of stars, Figure 1 also shows the spiral structure delineated by neutral hydrogen gas. As is customary in such diagrams, the space distribution of gas and the space distribution of the stars are not in good agreement. As various investigators have mentioned, stars and gas appear to define different spiral arms. However, such a conclusion is not warranted by data such as those employed in construction of Figure 1. In Figure 1 (as is invariably the case in earlier published diagrams of the same sort) two distance scales have been employed in the construction of the diagram. The distances of the stars have been derived from photometric data; the distances of concentrations of neutral hydrogen gas have been derived from measured hydrogen gas radial velocities and a galactic rotation curve. It should therefore come as no surprise if there are disagreements between hydrogen spiral arms and star spiral arms. Any regional peculiar motion of a gas concentration directly becomes an error in the inferred distance of the gas concentration.

The determination by optical means of the galactic structure at great distances from the Sun is becoming more and more important. Most optical investigations, using ordinary techniques — objective-prism survey in the blue spectral region + UBV photometry — do not reach very far out owing to the heavy obscuration in most directions in the galactic plane. Only a few very luminous OB stars and some rare supergiants have been identified in some selected regions at distances exceeding 4 kpc.

The region surrounding the galactic centre has been surveyed with the 210-foot telescope of the Australian National Radio Astronomy Observatory at a wavelength of 10·0 cm. At this wavelength the telescope beamwidth is 6.7 between half-intensity points.