THE GREAT FLARE OF 1985 APRIL 12 ON AD LEONIS

Photometric and spectroscopic observations covering the wavelength range 1200-8000 angstrom are presented for a giant flare on the M dwarf star AD Leo. The flare radiated more than 10(34) ergs in this wavelength region and lasted for more than 4 hr. The magnitude and duration of the flare allowed us to obtain an unprecedented view of the response of the lower atmosphere. For the first time, a flare energy budget over the entire optical and ultraviolet wavelength region is constructed as a function of time during the flare. The continuum radiation is shown to be the dominant source of energy loss during both the initial "impulsive" phase and the later "gradual" phase. The emission lines contribute less than 10% of the total flare energy in this wavelength region but are about 4 times more important during the gradual phase than in the impulsive phase. The energy budget is compared in detail with another, less energetic, flare on AD Leo and found to be quite similar. A sample of nine well-observed flares on stars with spectral type ranging from dM3.5e to dM5.5e was then used to investigate the relationships between the integrated properties of several flare emission features. Good correlation between H-gamma and Ca II K emission, and H-gamma and U filter emission was found, extending the previously known correlation of H-gamma with soft X-ray emission. These correlations suggest that the observed emission features are produced under similar atmospheric conditions regardless of the total flare energy emitted. The range in total emitted energy may thus be primarily a function of flare area and duration, and not of large differences in flare heating rate and resulting atmospheric structure. In addition, the time evolution of the emission features is qualitatively similar for a large number of flares, and quantitatively similar for the two flares studied in detail. This suggests that the energy source and deposition mechanisms which operate during the impulsive and gradual phases are causally connected, in agreement with current models for solar flares.