4 The spectral atlas

The lines studied in this work have been selected because:

1. Either they are good tracers of winds and mass outflow. For instance not all the UV lines obtained in the 1900-3200 Å range with IUE are represented. We have selected a sub-sample based on the previous work of Talavera & Gómez de Castro (1987).

2. Or they are useful spectroscopic indicators for the determination of stellar parameters: effective temperature, gravity and rotational velocity.

The selected lines are listed in Table 7 (visible) and Table 8 (ultraviolet) together with their basic atomic parameters. For the ultraviolet lines, the atomic data come from the semi-empirical calculations of Kurucz & Peytremann (1975).

  

Table 7: Visible lines selected for this atlas

  

Table 8: Ultraviolet lines selected for this atlas

The atlas, which is accesible from the CDS via ftp, is presented in a comprehensible way by means of two different kind of plots. The spectral atlas of stars with visible spectra only is presented with three stars per page. The line profiles of each star are represented in 7 different panels. These plots are designated collectively as Fig. 1. The spectral atlas of stars with both visible and IUE spectra or with IUE spectra only is presented with one star per page. All the line profiles corresponding to the same star are displayed in 12 different panels within the same page. These plots are designated collectively as Fig. 2. We provide the HD/BD number, the spectral type and the absolute visual magnitude for each star.

The profiles (normalized to the nearby continuum) are plotted versus wavelength (in Å). The visible lines are displayed in the first seven panels corresponding to: the lines of H I Balmer series, He I and Mg II and the doublet lines of Ca II and Na I. The ultraviolet lines are shown in bottom panels: Mg II (uv1) and Fe II (uv1, uv2, uv3, uv62, uv63, uv161). The laboratory wavelengths of the lines are marked by a dashed vertical line. Empty panels indicate that there are no observations at this wavelength range for a given star.

The flux scale varies from source to source to better display the observed profiles. We have selected for each star the best quality spectra obtained (best S/N ratio and spectral resolution) of all these available from the different campaigns. The IUE spectra are often very noisy, especially at the shortest wavelength. Where that was the case we smoothed the spectrum with a boxcar of two, three or five points in order to show a decent profile.

We have observed most of the stars several times in various spectral ranges to study their variability. Variations have been detected in H$\alpha$, Mg II (uv1) and Fe II (uv1, uv62, uv63). No evidence of variability has been detected in the other lines. We have represented in Fig. 3 a summary of this study. Figure 3 is presented in the form of a single plot for each star containing two or more spectra. The spectra are labeled with the date of observation and with the HD/BD number of the star. All the H$\alpha$ profiles are displayed for each variable A-supergiant. We have also selected some few representative examples of variations in the UV range. As in Fig. 1 the laboratory wavelength of the lines is marked.