The Sun is a typical star among billions of others in our galaxy, but it is the only one that can be well resolved and for which the physical processes can be explored in detail. Much of our solar research focuses on the magnetic field, its nature and its role in causing solar and stellar activity (star spots, flares, cyclic activity) and in determining the structure and dynamics of stellar atmospheres. An example of the magnetic structuring is shown in Fig. 1, an X-ray image of the Sun obtained from the Japanese Yohkoh satellite. The physical processes at work here are common to many other objects in the universe. The ever changing magnetic field is produced by dynamo processes inside the Sun, which are not yet sufficiently understood. At the surface of the Sun (in the photosphere) they can be measured via the Zeeman effect through polarimetry in selected spectral lines. Such observations and interpretations are challenging, since the magnetic flux is highly fragmented with a nearly fractal fine structure that cannot be fully resolved. We have developed various indirect methods to overcome the spatial-resolution limit and deduce the intrinsic properties of the unresolved magnetic structures. 

For magnetic-field diagnostics accurate polarimetry is needed. We have developed a new technology that allows imaging polarimetry of the Stokes vector with CCD-type detectors with a precision of 0.001 % in the degree of polarization. One such polarimeter system that we have built is now in semi-permanent use at the world's largest solar telescope, at Kitt Peak (Arizona). At this level of polarimetric accuracy, which is better by a factor of 10-100 than previous systems, the entire solar spectrum is polarized also in the absence of magnetic fields, because scattering processes in the Sun's atmosphere are a source of polarization. The big surprise when first looking in this new parameter domain was that the linearly polarized spectrum is extremely structured with little similarity to the intensity spectrum. An example of a portion of this newly discovered ``second solar spectrum'' is seen in Fig. 2. Here most polarized features are due to scattering by the C_2 molecule, which hardly leaves any ``finger prints'' in the ordinary intensity spectrum. Other portions of the polarized spectrum show strong signatures of quantum interferences, Raman scattering, and hyperfine structure, effects that are otherwise hidden from view in the ordinary spectrum. 

Our stellar observations have been done with a variety of international facilities on ground and in space, e.g. with ESO, IUE, ROSAT, and ISO, but lately we have had special emphasis on exploiting the unique capabilities of the Hubble Space Telescope ( HST). Some of our main topics are close binary stars, in particular symbiotic stars and the physics of hot stars, in particular Wolf-Rayet stars. In many of the close binaries a hot, compact star accretes mass from a red giant star, which leads to nova-like outbursts. Figure 3 gives an example of one of our HST spectra, showing P Cygni like line profiles which are characteristic signatures of stellar mass loss and the presence of a strong stellar wind from the hot white dwarf in the old symbiotic nova AG Peg. The observations from X-ray to infrared and radio are interpreted with detailed physical models that account for the radiation and ionization processes as well as the dynamics. 

The physics of particle acceleration and plasma instabilities can be studied in great detail in the solar corona with radio astronomy methods. Our institute operates a radio spectrometer about 50 km from Zurich, which has a time resolution of about 1 millisec and covers the frequency band 0.1 - 3 GHz. With dynamical radio spectra various physical processes and emission mechanisms can be diagnosed. An example is our dynamical radio spectrum in Fig. 5, which shows a ribbon-like structure that is a signature of moving particle beams. The steepness of the ribbons gives us the propagation speed, which in this case is about the speed of light. This means that the particle beams are relativistic. The small patches below the ribbons contain direct information on the acceleration mechanism. While the ETH radio spectrometer can identify the type of physical process, coordinated observations with the VLA (Very Large Array) in New Mexico allow us to locate the source on the Sun. Combination with X-ray data (e.g. from the Yohkoh satellite) gives the amount and distribution of released energy. 

The coronae of solar-like stars are studied with observing programs on the VLA as well as with VLBI (Very Long Baseline Interferometry). The observations are correlated with X-ray observations from ROSAT and ASCA to untangle the contributions from thermal and non-thermal sources. The VLBI observations have resulted in the first case of a spatially resolved main sequence star. Its VLBI radio image is shown in Fig. 6. For comparison the predicted size of the optical image (the stellar photosphere) is shown by the dashed circle to the lower left of the figure. The radio corona is much more extended and has an irregular shape. With future VLBI observations it may be possible to follow the rotation and evolution of individual active regions on other stars.