The distribution of magnetic field strengths in the interstellar gas is an important datum which relates to the formation of dense cloud complexes as well as to the distribution and propagation of cosmic rays. Brown and Chang (1983) found the brightness temperature of the radio continuum emission to be correlated with the HI column density NH and deduced a field strength which increases with particle density as <H> ~ <n>0.44. Troland and Heiles (1986) critized their approach and pointed out that Zeeman measurements suggest no increase in field strength for particle densities n ≤ 100 H-atoms cm-3.

A technique for separating the radio continuum emission of the Galaxy into its thermal and nonthermal components has been recently developed by Broadbent, Haslam and Osborne (1989). In this the thermal component is identified by its detailed correlation with the 60 μm infrared emission as observed by IRAS after the subtraction of zodiacal light and the HI-associated dust emission. This technique has been applied to the 408 MHz allsky survey of Haslam et al. (1982). A model of the distribution of synchrotron emissivity in the galactic disk has then been derived including information on the other tracers of spiral structure (HI, CO and giant HII regions) in order to account in detail for the observed nonthermal emission. The spiral arm pattern has two pairs of arms emanating from a central ellipse. The function describing the underlying variation of synchrotron emissivity with galactocentric radius is zero at the centre, rises to a sharp peak and then falls off slowly beyond 3 kpc. Using a scale length of the variation of cosmic ray electron density derived from γ-ray observations, we find that the scale length of the magnetic field variation must be as long as 22 kpc. This agrees with the scale length derived assuming equipartition between energy densities of magnetic field and cosmic rays. The variation of emissivity with height above the plane deduced by Phillipps et al. (1981) when included in our model gives good fits to the observed cuts across the plane. We have modelled the variation of the galactic magnetic field across a spiral arm as a gaussian. In order to fit the peaks in the galactic plane profile σ=0.2 kpc and a maximum compression of the field in the arm of 3.5:1 is required. These parameters are compatible with the results of N-body simulations of spiral arm formation which treat gas clouds as the unit particles. A detailed description of the model is about to be submitted for publication.

Polarization data at 390 and 826 MHz were obtained with the 300-foot telescope in February 1987. A survey of selected regions of sky planned for December 1988 had to be postponed. However, our limited data at 390 MHz show that the 30′ beam detected polarization temperatures between four to six times larger than found in surveys with a 1.3 arcmin resolution. This was true in both the highly polarized region around 1=140 degrees and in the North Polar Spur where polarization structures appear to be unresolved (<0.9 pc at the distance of the spur). High resolution observations will be critical to our understanding of the interstellar magnetic field and the scale-length of depolarizing structures.

Data from the Effelsberg Galactic plane survey at 11 cm wavelength have been used to produce a survey of polarized intensity in the first Galactic quadrant. Besides polarized sources (Supernova remnants and extragalactic objects) extended polarized emission features are visible, which are not connected to distinct radio sources. To decide whether these features reflect characteristics of the local field or the distant spiral structure of the Galaxy we performed an integration of polarized intensities as a function of Galactic longitude. An anticorrelation with the thermal background component suggests a distance of more than 6 kpc for some components of the polarized emission. This enables us to derive an upper limit for the uniform component of the Galactic magnetic field.

The results of a study of the magnetic field of the Galaxy up to ~3.5 kpc from the Sun by the analysis of interstellar polarization of optical radiation of catalogue stars (Mathewson et al., 1978) are presented. For the regions 90° < l <180° and 300° < l < 360°, where the polarization degree P is maximum, the empirical form is found.

Observations of polarization in and around Galactic supernova remnants, the deduced Faraday rotation and magnetic fields and their interpretation are reviewed. There are some indications of interaction with the Galactic magnetic field, but the conclusions are that this is an aspect which would benefit from considerably more observation and theoretical modelling.

We present a comparison of the Galactic magnetic field and the magnetic field of three supernova remnants (SNRs) of bilateral symmetry derived from linear polarization measurements. These data are best explained if a quasi-perpendicular propagation of the SNR-shock wave is assumed.

The remnant of Tycho's supernova of 1572 has been polarimetrically imaged using the 4 configurations of the VLA at wavelengths of 22 and 6 cm. A few data were also obtained at wavelengths of 21, 19, and 18 cm in the B configuration to check for any ambiguities in the Faraday rotation measurements and to look for deviations from the λ2 dependence of the Faraday rotation which would indicate significant internal Faraday effects. Although the total intensity shows structure on a scale of about 1 arcsecond the polarized emission appears to be resolved into cells of perhaps 10-arcsec size (1 arcsec corresponds to 0.011 pc at the 2.2 kpc distance to Tycho's SNR). We have therefore convolved the map shown in Figure 1 to a resolution of 4 arcsec to improve the signal-to-noise ratio. The contours represent the total intensity at 6-cm wavelength and the vectors show the position angle of the magnetic field with a length corresponding to the polarized intensity at 6 cm. In order to be seen, the vectors have a separation of 6 arcsec but no significant information is lost on this spacing.

We present and describe recent radio observations of the Crab Nebula, which allow us to determine the magnetic field orientation and depolarization at unprecedented resolution. The observations were made in 1987-1988 using all four configurations of the VLA, at 1410,1515,4625, and 4885 MHz. The resulting maps were all convolved with a clean beam of 1.8″ × 2.0″, elongated in P.A. 80°, and the residuals added back in.

There are two very important properties of the magnetic fields in supernova remnants which we wish to emphasize: 1) The net radial orientation of the magnetic vectors throughout the shell and 2) an apparent gradual increase in the fractional polarization with evolutionary age.

The fast dynamo acting in a turbulent flow explains the origin of magnetic fields in astrophysical objects. Stellar cycles and large-scale magnetic fields in spiral galaxies reflect the behaviour of a mean magnetic field. Intermittent magnetic structures in clusters of galaxies are associated with random magnetic field.

Large-scale axisymmetric and bisymmetric spiral (ASS and BSS) structures are found of magnetic fields in spiral galaxies by measuring the Faraday rotation of polarized radio emission. Dynamo theory is introduced to explain the field structures, and strong magnetogravitational interaction is suggested to occur between the BSS magnetic fields and spiral density waves. Up-to-date data about the rotation measures RM and redshifts z of QSOs and distant radio galaxies are given for discussing large-scale intergalactic magnetic fields.

Accretion disks around compact objects as well as the gaseous components in galaxies often have the form of a torus. To study the structure and behaviour of magnetic fields generated in such rings, a dynamo is investigated, which is working inside a torus embedded into vacuum. The equations for the kinematic αω-dynamo are written down in toroidal coordinates (see Figure 1). Besides loss of magnetic flux by Ohmic diffusion (characterized by the magnetic diffusivity D) they describe its production by the inductive effects of differential rotation and of turbulent matter, which we have chosen as ω(r) = ω′0r and α = α0 sin θ, respectively. These equations are solved by series expansion into the exponential decay modes of slender tori, which are available in analytical form. A linear homogeneous system of equations follows for the expansion coefficients; its eigenvalues determine the time-dependence of the solutions, the dynamo modes.

The present status of galactic dynamo theory is discussed. A new concept which allows the determination of marginal dynamo numbers for axisymmetric as well as non-axisymmetric large-scale magnetic field modes in axisymmetric disks is applied to a simple galaxy model. The results obtained so far show that a preference of non-axisymmetric fields can only be expected if the α-effect is highly anisotropic and the differential rotation is not too strong. Mostly axisymmetric-spiral fields have to be expected.

A concept of mean-field models of galactic dynamos is explained in which α-effect and differential rotation act in an infinite slab of electrically conducting fluid surrounded by free space or by conducting matter at rest. Some results for α2 -dynamos are presented. In the limiting case in which the α-coefficient does not depend on the radius, axisymmetric and non-axisymmetric magnetic fields can be maintained with equal ease. For reasonable α-distributions a preference of axisymmetric fields is suggested as long as an isotropic α-effect is considered but it may disappear with anisotropic α-effect. The effect of differential rotation, which allows αω-dynamos, is briefly discussed.

Besides the mean flow the alpha is the other input quantity for any mean-field dynamo model. It describes the generation of turbulent electromotive force <u × B> from a large-scale field <B> for a given turbulence. The necessary helicity of the turbulence results from the joint action of Coriolis force and density stratification. The standard estimate of 1 km/s for alpha in galaxies is a surely well-established approximation. One of the essentials, however, remains open. Due to the extremely anisotropic structure of disks the tensorial character of alpha can no longer be ignored. In stellar applications anisotropy in the α-tensor leads to a preferred excitation of non-axisymmetric magnetic fields. That is true for α2 -dynamos if the alpha parallel to the rotation axis, α||, is much smaller than that in the equatorial plane, α⊥. The idea is that also for disk-like configurations a similar behaviour makes the existence of the observed large-scale non-axisymmetric magnetic BSS modes understandable within the frame of the mean-field dynamo theory.

Radio polarization observations of spiral galaxies suggest the existence of large-scale galactic magnetic fields which are of either axisymmetric -spiral (ASS) or bisymmetric-spiral (BSS), i.e. non-axisymmetric, structure (cf. Beck, 1939). Clear evidence for a BSS field was indicated for M31 by M. Krause et al. (1989).

I give some results of calculations of the magnetic fields generated by dynamo processes in galaxy models consisting of a turbulent disc embedded in a low-conductivity halo. The field configuration is found to be strongly dependent on the turbulent structure of the disc.

Kinematic models of galactic dynamo in axisymmetric disk show remarkable agreement with observations. We argue that nonlinear effects are relatively weak in galactic dynamos and consider the properties of linear dynamo models inherited by global magnetic configurations in spiral galaxies as well as nonlinear distortions of the linear solutions.

Linearized equations of motion are solved for the self-gravitating gaseous disk in interaction with the Lorentz force due to the bisymmetric spiral (BSS) magnetic fields. Since the pattern velocity of the BSS fields is very close to that of the spiral density wave, a nearly-resonant interaction occurs between these two waves to enhance spiral condensation of gas. The BSS fields seem to compensate the spiral density waves for their secular dispersion.