We have carried out four different visible campaigns to observe the H, H, H, Hand H Balmer lines, Ca II H and K, Na I D and Mg II 4481 Å lines. The observed spectral ranges are listed in Table 3. We summarize in Table 4 the information available for each star as well as the instrumental set-up (telescope+spectrograph) used to obtain the data and the Julian date of the observations.
The first campaign was carried out in September 1988 with the Isaac Newton Telescope of the Roque de los Muchachos Observatory in La Palma. All the spectra were taken with the Intermediate Dispersion Spectrograph (IDS) equipped with a GEC CCD which has 22 m pixels arranged in an 578 385 array. We observed H,Ca II H and K, Na I D and the Mg II4481 lines with a resolving power around .
The second campaign (October 1988) was carried out with the Coudé Spectrograph (resolving power around ) of the 2.2 m telescope at the Calar Alto Observatory. The detector used was an RCA CCD (, pixels). We obtained the H,H, Ca II(H, K) and Mg II4481 profiles.
The third campaign was carried out in the Pic Du Midi Observatory (France) in August 1994. We used the MUSICOS (for MUlti- SIte COntinuous Spectroscopy) Spectrograph (Baudrand & Bhm 1992) coupled to the 2 m Bernard Lyot Telescope (TBL) by means of a double optical fiber of m core diameter (due to the mechanical separation of the spectrograph from the telescope, excellent stability is guaranteed during the night). The MUSICOS spectrograph works in cross-dispersion mode allowing the observation of whole visible range (from 380 nm to 880 nm) in two exposures with a resolving power around . Six stars of our sample were observed with this instrument.
The last campaign was carried out in September 1994 with the Jacobus Kapteyn Telescope of the Roque de los Muchachos Observatory in La Palma. We used the Richardson Brealey Spectrograph (RBS) with a TEK CCD which has m pixels arranged in a array. In this campaign we paid special attention to the H and H lines which were observed simultaneously with a resolving power around 3700.
The spectra of the two southern stars (HD 59612 and HD 102878) were kindly provided to us by Andreas Kaufer. These spectra were taken in La Silla with the Heidelberg Extended Range Optical Spectrograph (HEROS) at the ESO 50 cm telescope. HEROS is a portable fibre-linked two-channel echelle spectrograph which covers from 345 nm to 560 nm and from 580 nm to 865 nm in one exposure. The resolving power is around over the complete wavelength range and, as with MUSICOS, stability is guaranteed by the separation from the telescope.
The spectra obtained at La Palma and Calar Alto have been processed using standard astronomical data reduction packages (IHAP (Middelburg 1981), MIDAS and IRAF). In all cases the spectra have been extracted from the CCD frame after dark noise correction and flatfielding. At this stage, each spectrum was rebinned into a wavelength scale by using a polynomial fit to the positions provided by the comparison lamps (ThAr, ThNe or CuNe). A calibration spectrum was obtained immediately after the stellar spectrum, without any change in the spectrograph configuration. The long-term accuracy achieved for the wavelength calibration is of the order of 1 km s-1.
In the final step the resulting wavelength-calibrated spectrum was rebinned to heliocentric velocities and the continuum was normalized to unity by fitting a low-order polynomial and dividing the spectrum by this function.
The MUSICOS spectra were semi-automatically reduced with the dedicated software MUSBIC. A complete description of this procedure can be found in Baudrand & Bhm (1992). Briefly, as the acquisition, the reduction is divided into two spectral domains: the red (550-890 nm) and the blue (390-550 nm). Both domains are precisely defined with regard to the order positions. The first step carried out by the program is to determine the exact position of the orders. The procedure is performed on the stellar, flatfield and calibration spectra. Then the program extracts each order of these spectra. The half-widths of the nearly Gaussian order profiles (perpendicular to the dispersion) vary between 2.5 and 3.5 pxl for the blue and for the red domain respectively. Therefore the extraction width is usually 7 pxl for the blue domain and 9 pxl for the red, to contain most of the signal. The program integrates the extracted signal above the baseline defined by the mean value of the signal at the adjacent interorder positions. Two data tables contain wavelengths and pixel positions of all significant emission lines for each order of a thorium reference spectrum. Starting from guessed positions read in the tables, the program finds the exact positions and automatically deduces for each order the coefficients of the 3rd order polynomial giving the wavelength calibration.
The stars observed with the IUE satellite are all the A-supergiants in the 1986 IUE Archive, studied before by Talavera & Gómez de Castro (1987), as well as the new set observed by us in campaigns during 1986, 1991, 1994 and 1995. The new spectra were taken using the prime camera of the long-wavelength IUE spectrograph (LWP) in high-resolution mode. The spectral range observed with this configuration is from 1900 Å to 3200 Å with a resolution of 0.2 Å; the most prominent indicators of wind in A-supergiants, Mg II and Fe II lines, are observed in this range.
The IUE observations carried out by us are summarized in Table 5. For each star the date of observation, the image number and the exposure time are provided.
All IUE data have been processed at VILSPA using the IUE Spectral Image Processing System (IUESIPS). IUESIPS produces data as free as possible from instrumental effects. These include geometric distortion correction, photometric nonlinearity correction, spectral order extraction and wavelength and flux calibration.
The standard IUE data reduction has an uncertainty in the zero of the wavelength scale which can amount up to 0.5 Å. To correct for this we have measured the observed wavelength of some selected interstellar lines listed in Table 6 and compared with their laboratory wavelength. This method cannot be used in some stars for which the lines selected are not clearly interstellar (lines wider than the IUE instrumental profile). In such a case we compared the observed spectrum with that of HD 46300 which shows several narrow lines, and henceforth can be used as a reliable template.
Copyright The European Southern Observatory (ESO)