A spectroscopic and photometric study of the pulsating R Coronae Borealis type variable, RY Sagittarii.
Results are given of an intensive study of RY Sgr during the period . The rapid decline in brightness at the beginning of this period and the gradual rise to maximum light' are covered. 382 emission lines (mostly identified) were found during the initial decline as was a continuum attributed to the electron attachlnent spectrum of CN. The development with time of this emission line chromospheric spectrum indicates both a lowering of the excitation and a decrease in the effects of self-absorption. The chromosphere decays on a time scale of 22 days. The initial decrease in photospheric radiation has a time scale of 5 days or less. The' absence of [Fe ii] emission suggests that the density in the emitting region is greater than io14 atoms in the early stages of the decline if collisional excitation is dominant. During the decline displaced shell lines of Ca ii (-200 `WI s-1) were present. On the rise to light maximum and at maximum, the light and colour curves show smooth variations (AV 0m.5 A(B-V) 0m.3 A(U-B) 0m.5) with a period of days. A similar periodicity is present in the radial velocities (amplitude 30 km s-') and it is concluded that RY Sgr is a pulsating variable (as originally suggested by Jacchia) as well as an RCB star. At some epochs on the rising branch the spectrum is normal (i.e. as at maximum) but the star is faint and red. This together with the infra-red observations of Lee and Feast strongly supports the hypo?hesis of extinction by a circumstellar dust shell. Values of AvIEB-v = and / = I 3 are found for the shell. Abnormal spectra and colours follow a drop in light on the rising branch. They are apparently due to the filling in of the photospheric absorption lines by chromospheric emission. The primary cause of this phase seems to be a decrease in photospheric light, possibly due to the formation of a fresh dust shell near the photosphere. Displaced shell absorption lines (Ca ii and Na i) are present at this stage. The pulsations of RY Sgr are broadly consistent with recent work by Trimble on the pulsation of helium stars.