THE CIRCUMSTELLAR GLEAM FROM THE T-TAURI STAR RY-LUPI

The T Tauri star RY Lupi was followed simultaneously spectroscopically and photometrically over 7 nights covering one of its deepest brightness minima ever recorded. As reported earlier the star first becomes redder with decreasing brightness. At V = 12.5 this trend is changed towards the blue and this time the object was as blue as in maximum at the deepest minimum. The spectral type (G8) remains constant over the entire range of 3 mag in V. During the deep minimum, when the star brightened from V = 13.12 to V = 12.45, the line-to-continuum ratios and equivalent widths of absorption features in the U- and B-bands did not change with time. We show that the additional light component, present when the star is heavily occulted by circumstellar dust, can not be due to an emission component that either is produced at low brightness levels or is always present. Instead it is due to stellar radiation scattered into the line-of sight by dust in the immediate surrounding of the star. We test this result by assuming that the scattering occurs in a circumstellar disk which is inclined so that the line-of-sight passes its huffy upper layers with variable degree of foreground circumstellar dust. Standard Values of such a disk model together with standard values of the properties of the dust particles give a good theoretical agreement with the observed colour pattern, both for the location of the turning point in the colour diagrams and the energy distribution and flux at minimum brightness. At this point the optical flux from the object is totally dominated by the gleam from the disk. Polarization data are in agreement with this scenario. It is also pointed out that even though the component of scattered light may stay constant over one night, there are considerable long-term changes in the details of the scattering. We also discuss the behaviour of the emission lines, which most likely form in a halo/wind, not occulted much by circumstellar dust. Complementary Far-UV data and one CO-observation are discussed and future tests of the model are proposed.