There are many types of solar flare, but the classic type is a two-ribbon flare with three phases — a preflare phase, a rise phase and a main phase. The properties of these phases are described, together with some recent observational advances in understanding the conditions for solar flares. Such flares are thought to be caused by an eruptive MHD instability which drives reconnection and therefore energy conversion. A review is given of our current understanding of the nature of this instability and the resulting reconnection process, including a recent attempt to describe its three-dimensional nature.

The systematic study of stellar flares in now more than 30 years old. Throughout that time the solar flare paradigm has been used, highly successfully, to understand various aspects of the stellar phenomenon. In this introductory overview I will attempt to justify the use of the solar model in understanding the stellar equivalent. But I will also point out differences between the two which may require extension of the basic solar model and speculate briefly on what these extensions might be.

There are many excellent previous reviews of stellar flares in the literature (Haisch 1989, Byrne 1983, 1992), to which the reader is referred, along with the other reviews in this volume, for further reading.

Radio observations provide the most direct information on non-thermal electrons in stellar flares and in the coronae of late-type stars. Radio emissions of single main-sequence F, G, and of many K stars have recently been discovered, in addition to the well-known dwarf M stars. Their long-duration radio flares with low circular polarization, slow variations and broad bandwidth can be attributed to gyrosynchrotron emission of mildly relativistic electrons. The same holds for the low-level (‘quiescent’) radio emission. On the other hand, highly polarized radio flares of M stars have been interpreted by coherent emissions from loss-cone instabilities of magnetically trapped electrons. These conjectures are consistent with recent VLBI observations. The identification of the radio emission process allows to estimate the high-energy component of the flare and compare it to the total flare energy. The weakly polarized radio emission may serve as a proxy for hard X-ray signatures of relativistic electrons. The fraction of primary energy released into energetic electrons then appears to be large and similar to solar flares.

AD Leonis is a very active, single dMe flare star. The similarities between this type of star and the Sun has led to study their radio radiation, which originates from their corona. The high brightness temperatures and other characteristics of most dMe radio bursts can be attributed to a non-thermal, coherent mechanism: plasma radiation or a cyclotron maser instability (CMI) are both plausible explanations. Even for the strongest burst of AD Leo which reached 940 mJy at 21 cm, it was not possible to discriminate between these two mechanisms (Bastian et al. 1990).

Here we present an intense burst from AD Leo, exhibiting strong spikes for which the CMI seems to be the only reasonable explanation. In Sect. 2 we describe the observations, and in Sect. 3 we give an interpretation for this event.

RS Cvn systems have been studied very extensively at radio wavelengths and a large fraction of them have been found to be strong radio sources, with 6 cm radio luminosities in the range between 1015 and 1018 erg s−1. The radio flux density is highly variable and it usually shows two different regimes: active periods, characterized by a continuous strong flaring which can last for several days, and quiescent periods, during which the flux density goes down to a few mJy. These characteristics indicate the non-thermal nature of the radio emission which is driven by magnetic activity of these systems, seen in other spectral regions. In this scenario, the radio flux arises from the interaction between the magnetic field of one or both components with mildly relativistic particles, i.e. gyrosynchrotron emission.

Due to the erratic nature of radio flares in RS CVn binaries, most VLBI observations have been performed during quiescent periods. Only the 1983 UX Ari observation (Mutel et al. 1985), carried out during an intense outburst, allowed the first, and still the unique, hybrid map of a RS CVn binary, which revealed a core-halo morphology of the radio source.

It is still not clear if the flare events are localized in very compact regions of the corona of one component or if the entire binary system is involved. VLBI measurements in both active and quiescent periods are extremely important to determine the characteristic dimensions and the structure of the radio source.

We present VLA observations of RS CVn binaries which indicate that the low-frequency polarization inversion usually ascribed to the quiescent radio emission of these systems is in fact due to the presence, much of the time, of a coherent burst emission. Fig. 1 shows two examples of this coherent emission in the behaviour of the light curves of HR1099 in left and right circular polarization at 1.4 GHz. The emission is highly circularly polarized, with one polarized component steady or slowly varying in a way similar to the higher frequency emission (Fig. 1a), while the other shows rapid fluctuations on timescales of ∼10s and a slower modulation on longer timescales. The similarity of the time profiles in two 50 MHz-wide sidebands 30 MHz apart indicates that the fluctuating component is broadband; moreover, the sense of polarization at 1.4 GHz is opposite to that of the quiescent emission at higher frequencies (attributed to gyrosynchrotron emission), which is exactly the behaviour attributed to the 1.4GHz quiescent emission of these systems by Mutel etal. (1987).

The rapid fluctuations and high polarization are characteristic of a coherent emission process. We argue that a combination of two sources of emission is generally present at 1.4 GHz: a very highly (possibly 100%) polarized coherent component, and a weakly-polarized, steady gyrosynchrotron component. Since the polarization at higher frequencies should be the x-mode of (optically thin) gyrosynchrotron emission, we infer that the low frequency polarization probably represents the o-mode. The observed characteristics of the emission then suggest that the most likely mechanism for the highly-polarized component is plasma emission at the fundamental of the plasma frequency. Such plasma emission is relatively rare on the Sun at 1.4 GHz, but could occur in the hotter coronae of RS CVn systems because the higher temperature reduces the free-free opacity which is thought to absorb 1.4 GHz plasma emission on the Sun.

UX Arietis is an active binary system with an orbital period of 6.44 days. The presence of long lasting large spots on the surface of the more active star has been deduced from optical observations (Vogt & Hatzes 1991).

The observations presented here have been performed from December 1992 until October 1994 using the Effelsberg 100 m telescope. The instrument was available for this program in the gaps between previously scheduled observations. For this reason our observations span a frequency interval between 1.4 GHz (21 cm) and 43 GHz (7 mm), depending on the scheduled receiver at each observing time. These data, which are an extension of those published by Neidhöfer et al. (1993), are shown in Fig. 1, plotted in terms of the known orbital period (6.44 days). The search for different periodicities (by using the Phase Dispersion Minimization method) has shown the existence of two main other periods: at 156 days and at 24 days (Massi et al., in preparation).

We present the results of the first three years of a single dish radio monitoring of HR 1099 we are carrying out using the the 32m radiotelescope at the Noto VLBI station of the Istituto di Radioastronomia of the Italian Consiglio Nazionale delle Ricerche (C.N.R.).

These data, together with other observations available from literature, constitute a database which shed light on the radio emission from this system.

HR 1099 tends to undergo extended periods of strong activity, interrupted by periods of low flux density level, as can be seen from the radio behaviour on a timescale of several months. The switching between flaring and quiescent periods is not random and, at the moment, it is not possible to associate a characteristic duration to each phase.

Flare activity of solar active regions is generally believed to depend on a sheared configuration of magnetic fields (Hagyard et al. 1984). There are cases when the shear necessary for a flare can be attributed to the emergence of a new flux in the spot group (Wang 1992). But, perhaps, a newly born active region can also influence the magnetic field configuration in a nearby active region (Poleto et al. 1993, Gesztelyi et al. 1993). In this paper we are interested primarily in the influence of a newly emerging spot group on a nearby one.

The three neighbouring active regions NOAA AR 6412(B-C), 6413(A) and 6415(D) have been studied between 13-22 December 1990. White-light pictures for studying sunspot proper motion and area evolution were taken at Gyula Observing Station (Hungary), Debrecen Heliophysical Observatory (Hungary) and Helwan Observatory (Egypt). Times and positions of the flares were taken from the Solar Geophysical Data (No. 558, part 1, February 1991).

Observational results on flare stars in the Galaxy and their interpretation, based on the observational approach to early stages of stellar evolution at Byurakan, are considered. The evolutionary path of red flare stars is determined. Stellar flare activity is explained as an internal phenomenon of stars. The significance of photographic observations of stellar flares in star clusters and associations is emphasized.

Red dwarf stars in the Galaxy occur in three forms: as flare stars, T Tau stars and red dwarf stars of constant brightness. Haro (1957) suggested that all these stars present regular evolutionary stages of red dwarf stars.

The space distribution of UV Cet flare stars in the solar vicinity indicates that these low luminosity stars belong to the disk population of the Galaxy. Therefore one can suppose that all red dwarf stars have a space distribution which is similar to that of the flare stars (Mirzoyan et al. 1988a). Only a small part of them is found in star clusters and associations. The T Tau stars are an exception: because of their very short lifetime, all of them are found in stellar associations.

The space distribution of red dwarf stars is determined by the fact that all of them are formed in star clusters and associations and are finally lost from their stellar systems and merge into the general galactic field (Mirzoyan 1995).

The kinematic properties of 93 UV Ceti stars of the solar neighborhood are studied, based on a list of flares within 25 pc of the Sun (π ≥ 0".04). With updated values for their distances, proper motions and radial velocities (Gliese & Jahreiss 1991) space velocity dispersions are calculated for these stars. It is found that the total velocity dispersion of the flare stars (σ=30±3 km s−1) is similar to that of the F5 V stars from the same catalogue, for which the conventionally estimated mean age is about 3 · 109 years. A number of flare stars are identified as members of the Hyades, Sirius or Pleiades groups. The velocity dispersions found for the nearby flare stars, as well as their scale height and the membership of some of them to young kinematic groups, indicate that they belong to the young disk population. A small number (7) of UV Ceti stars have kinematics corresponding to the thick disk or halo population. Their long-lived chromospheric activity is interpreted as due to coalescence of old contact binaries. The question of the age of Proxima Centauri is examined in the context of our results, and found to be compatible with the ages of a Centauri A and B.

I selected the common objects from our Data Base on 234 UV Cet Flare Stars and Related Objects and the Third Catalogue of Nearby Stars by Gliese & Jahreiss (CNS-3). It is found that manifestation of flare activity (FA) (flares in any spectral region or photometric variability caused by spots) is observed on 143 flare stars (FS) out of the 3803 CNS-3 stars within 25 pc of the Sun with absolute magnitudes Mv from 5.2 (Sp G2 V) to 18.8 (Sp dM5e − dM6e).

Magnitudes and spectra of flare stars: For each MV-interval, Table 1 gives the “mean” spectral class (Sp), the number of FS (Nf) and the portion of FS among other red dwarf stars from CNS-3 (Nf/N). Absolutely fainter stars include a higher portion of known FS. However, observational selection effects can cause this dependence, because the probability is higher to register flares over the background radiation of a faint star. The observed portion of FS among all red dwarfs from CNS-3, which are fainter than Mv=5, is about 5%.

Photographic observations with wide-field telescopes of flare stars in the Orion association and the Pleiades cluster have been carried out since 40 years. About 500 flare stars were found in each of these systems. The ages of these stellar systems differ by one or two orders of magnitude (Mirzoyan 1991). This fact allows us to compare them, in order to show that they fit into the concept, suggested first by Haro (1957), that the evolutionary status of flare stars is a stage of red dwarf star evolution which follows that of the T Tau stars (Haro 1976, Ambartsumian & Mirzoyan 1970).

The basic difference between the Orion and Pleiades subsystems of flare stars has been known for a long time: the coexistence of flare stars and T Tau stars in the Orion association, and the absence of the latter group in the Pleiades. Some of the T Tau stars show flare activity (Haro 1964). This fact and the existence of multiple systems of trapezium type, which are dynamically unstable, in the Orion association, are arguments in favour of an evolutionary connection between these two types of stars.

Mirzoyan (1976) showed that the concentration of flare stars around the center of the Pleiades cluster (Alcyone) was the same, irrespective of their proper motions. At that time, however, proper motions of only a few flare stars were known. The result was confirmed on the basis of more extensive observational material (Chavushian 1979, Mirzoyan 1983). Photographic observations of stellar flares in the general galactic field during 181 hours yielded the detection of only a single flare, i.e. the percentage of flare stars in the general galactic field is about 10% of the total number of flare stars detected in the regions of stellar clusters and associations (Chavushian 1979, Mirzoyan et al. 1988).

This result shows that flare activity can be considered as a definitive cluster membership criterion, which appears to be a better one than the cluster membership probability, which is based on proper motions. To prove this, cluster membership probabilities (Stauffer et al. 1991) for 408 Pleiades cluster flare stars from the catalogue by Haro et al. (1982), are used (Table 1).

The ROSAT Wide Field Camera all-sky EUV survey of 1990-91 identified over 200 active late-type stars (Pounds et al. 1993). Most of these were not known to be active prior to the survey and follow-up optical spectroscopy has been successful in selecting the most active dMe stars. These stars are characterized by their Balmer line emission which is direct evidence for an active chromosphere. We present photometric monitoring data for 4 new dMe stars, including the discovery of flaring activity on all 4 stars.

The observations were made at the South African Astronomical Observatory at Sutherland near Capetown in October 1994, using the 0.5m and 1.9m telescopes for photometry and spectrometry, respectively. The 0.5m telescope is equipped with an automated, pulse-counting photometer. Flare monitoring of all the stars was done in the U-band with either 5 or 10 s integration times. All the stars monitored were between magnitude V = 11 and 13. A total of 15 flares were detected, including at least one on each star (Table 1). Comparison stars were measured in standard UBVRI photometric bands for each star.

In a field of 100 square degrees around 26 Com, 238 X-ray sources have been discovered by ROSAT during the All-Sky-Survey observations. These sources were identified using objective prism spectrograms taken with the Hamburg Schmidt telescope on Calar Alto (courtesy N. Bade) and including the positional accuracy of typically 30" and the X-ray to optical luminosity ratio. In an ongoing investigation of these X-ray sources on more than 400 plates of the Sonneberg astrographs 400/1600 mm and 400/2000 mm, a total of 5 of the M type counterparts were found to be flare stars. In Table 1 we give the ROSAT name (column 1), the new designation as a Sonneberg variable (2), the position of the optical counterpart (3), the magnitude range of the newly discovered flare stars (4), the ROSAT PSPC countrate during the All-Sky-Survey (5), the distance D between X-ray and optical position (6). All 5 objects had not been known to be flaring. The real amplitudes must be larger than those given in column 3 because the (unknown) duration of the flares is usually much shorter than the exposure time of the plates. The figures give APM finding charts of these new flare stars.

In 1986-94 a series of coordinated monitorings for the flare star EV Lac has been performed. The observations were aimed to collect homogeneous data on the star both in its active and quiet states, to study its flares and starspots. These photometric patrols were accompanied by monitorings of EV Lac of different other kinds: polarimetric, spectral, radio, UV and high energy studies. In total, these monitorings covered a range from decametre radio wavelengths through IR, optical and UV windows up to very high energy γ-rays. The list of participants of these studies is very large. Up to date the results of 1986, 1987, 1990 and 1991 are published (Gershberg et al. 1991a,b, 1993, Alekseev et al. 1994), the data of 1989 are submitted (Berdyugin et al. 1995), the data of 1992 (Abdul-Aziz et al. 1995) will be submitted in the near future. In 1993, photometric observations were accompanied by monitoring observations of EUVE and IUE, in 1994, VHE γ-ray observations were included. The publication of data collected in 1993 and 1994 is in preparation. In this communication I shall concentrate on several new findings related to the EV Lac flares only.

During the 1994 coordinated observations of the red dwarf flare star EV Lac, the star was monitored in the very high energy (VHE) γ-ray range around 1012 eV with the Crimean ground-based γ-ray telescope GT-48. This telescope consists of two identical optical systems (Vladimirsky et al. 1994) which were directed in parallel on EV Lac.

The detection principle of the VHE γ-rays is based on the Čerenkov radiation emitted by relativistic electrons and positrons. The latter are generated in the interaction of the γ-rays with nuclei in the Earth’s atmosphere that leads to an appearance of a shower of charged particles and γ-quanta. The duration of the Cherenkov radiation flash is very short, just about a few nanoseconds. The angular size of the shower is ∼ 1°. To detect such flashes we use an optical system with large area mirrors and a set of 37 photomultipliers (PMs) in the focal plane. Using the information from these PMs which are spaced hexagonally and correspond to a field of view of 2°.6 on the sky, we can obtain the image of an optical flash. The electronic device permits us to detect nanosecond flashes (40 ns exposure time and 12 μs readout dead-time).