eta Carinae is in the midst of a radio outburst which has made it one of the brightest stellar radio sources in the sky. This paper reviews our current understanding of this outburst, and discusses how the radio emission may be used to study the mass loss history of eta Car in very recent history (tens df years).

On five occasions between 1999 June 29 and 1994 May 3, we have used the Australia Telescope Compact Array to image eta Carinae at a wavelength of 3 cm and a resolution of 1''. These observations have revealed remarkable activity. Since 1992 June, the total flux density has increase from 0.8 to 2.2 Jy, and the original single compact source has grown to a complex of sources spread over an area of about 16 arcsec(2) (Fig. 1). Strong hydrogen recombination-line spectral emission has appeared at the site of the strongest of these new sources. This recombination emission has the largest spectral width ever observed from a star, +/-250 km s(-1) and reveals gas with turbulent velocities as great as 250 km s(-1) approaching us at an average velocity of about 200 km s-(1). We believe that this radio outburst has been caused by a more than threefold increase of ultraviolet luminosity, and consequent ionization of previously neutral gas clouds.

We describe a set of millimeter interferometric observations of solar flares carried out in conjunction with GRO experiments during the 1991 June Campaign of the Max'91 Program. We show evidence that millimeter emission probes the most energetic (MeV) electrons in solar flares; we also find that in the same flare there can be both impulsive nonthermal and gradual thermal millimeter emission. Millimeter emission usually occurs at the steep rise phase of the hard X-ray emitting electrons (25-100 KeV). There appears to exist some delay between BIMA mm-emission onset and GRO-BATSE 25-100 KeV X-ray emission. Both results have implications for the particle acceleration process.

We present high spatial resolution radio observations of the peculiar southern star eta Carinae, made with the Australia Telescope. The images, at 8 and 9 GHz with a resolution of 1.0'', show a source of dimension 10'' and total flux of 0.7 Jy dominated by a strong central peak. The radio emission is unpolarized and offers no support to models which invoke degenerate stars or more exotic objects within the core of eta Car. In these data we find no evidence for more than one energy source in the core with arcsecond separations as some infrared observations have suggested. Several levels of structure are evident in the radio image, which shows symmetry on the larger scales. Conventional formulae for stellar wind radio sources give a mass loss rate of order 3 x 10(-4) M. yr-1 based on the radio flux in the central peak, which yields a wind momentum flux of order 20% of the momentum flux available from the star's radiation field. The radio emission at these frequencies is consistent with thermal emission from gas flowing away from a ''luminous blue variable'' star (LBV). Eta car is probably the brightest thermal stellar wind radio source in the sky.

We report on the observations of a number of flares at a wavelength of 3.5 mm during the 1991 June solar campaign. Many flares, including small ones, show an impulsive phase at milllimeter wavelengths which indicates the presence of MeV electrons, and the millimeter observations are far more sensitive to such electrons than are current γ-ray detectors. However, these energetic electrons do not always show a good correlation with the lower-energy electrons which produce hard X-rays below 100 keV. The production efficiency of MeV electrons seems to vary considerably from flare to flare. An extended phase similar to the soft X-ray behaviour is also seen at millimeter wavelengths, which we attribute to dense hot material radiating thermal bremsstrahlung. In the impulsive onset the millimeter emission seems to be consistently delayed with respect to the hard X-rays.

We have carried out high-spatial-resolution millimeter observations of solar flares using the Berkeley-Illinois-Maryland Array (BIMA). At the present time, BIMA consists of only three elements, which is not adequate for mapping highly variable solar phenomena, but is excellent for studies of the temporal structure of flares at millimeter wavelengths at several different spatial scales. We present BIMA observations made during the Gamma Ray Observatories (GRO)/Solar Max 1991 campaign in Jun. 1991 when solar activity was unusually high. Our observations covered the period 8-9 Jun. 1991; this period overlapped the period 4-15 Jun. when the Compton Telescope made the Sun a target of opportunity because of the high level of solar activity.

We compare the millimeter, microwave, and soft X-ray emission from a number of solar flares in order to determine the properties of the high-frequency radio emission of flares. The millimeter observations use a sensitive interferometer at 86 GHz which offers much better sensitivity and spatial resolution than most previous high-frequency observations. We find a number of important results for these flares: (i) the 86 GHz emission onset appears often to be delayed with respect to the microwave onset; (ii) even in large flares the millimeter-wavelength emission can arise in sources of only a few arc sec dimension; (iii) the millimeter emission in the impulsive phase does not correlate with the soft X-ray emission, and thus is unlikely to contain any significant thermal bremsstrahlung component; and (iv) the electron energy distributions implied by the millimeter observations are much flatter (spectral indices of 2.5 to 3.6) than is usual for microwave or hard X-ray observations.