2 Observations and data reduction

The basic data on the observed stars have been gathered in Table 1. Star identifications are given in Cols. 1 and 2, the spectral types from the BSC and from Abt and Morrell (1995) (hereafter AM) in Cols. 3 and 4.

  

Table 1: The programme stars

The list of the established double stars in this sample can be extracted from the informations given in Annex 1 of the Hipparcos Input Catalogue (Turon et al. 1992), as well as from the Hipparcos and Tycho Catalogues (ESA, 1997) for the new ones discovered during this mission. For visual doubles the magnitude difference and the angular separation, given in Cols. 5 and 6, are for the systems for which these values can affect the measures. These informations are given in the last column of Table 1. Results from search for binarity by speckle observations (Col. 7) are taken from the update of the Bright Star Catalogue (Hoffleit & Warren 1994) [hereafter named BSC1994], kindly communicated by W.H. Warren. Projected rotational velocities, according to the BSC and to AM are added in Cols. 8 and 9.

The spectra were obtained with the ECHELEC spectrograph mounted at the ESO 1.5 m telescope, in the period Jan. 89 - Jan. 95. The linear dispersion is about 3.1 Å mm^-1 in the chosen wavelength range 4210-4500 Å. The slit width of 320 ${\mu}$m corresponds to 1.52 arcsec on the sky. From unblended thorium lines, the FWHM of the instrumental profile was measured; its value is 0.17 Å, in agreement with the size of the entrance slit. The original frame covers 11 orders of the echelle spectrum; only 9 of them have been retained and reduced, while the 2 truncated orders on both sides of the central wavelength have been discarded. Flat field correction with a tungsten lamp and wavelength calibration with a Th lamp have been made with classical procedures. Great care was devoted to remove instrumental defects which arose from time to time: the most important ones were the two ghosts present as transversal bars on the original frames in the 1991-1994 period; the method is based on the accurate analysis of the intensity in the inter-orders. The resolution is about 28000 and the S/N, highly variable from the center to the edges of each order, covers the range 50 to 200.

The analysis of the H$_\gamma$ profile being one of the goals of the present study, we had to produce a single spectrum by interconnecting the echelle orders. A particular attention was given to the method adopted for this connection. The complete reduction procedure is described in Burnage & Gerbaldi (1990, 1992). The broad spectral domain covered by the H$_\gamma$ line in A0 dwarf stars imposes to give a particular attention to the procedure adopted to normalize the spectra. The normalization is made by trial and error method; first the spectra have been normalized in the classical way by drawing the continuum through the highest points of the spectrum and then overplotted on a template synthetic spectrum. This allows to choose a number of continuum points common to both; these points are used as references in the final normalization procedure.

Since the observations span over 5 years, we have systematically observed four stars. The number of spectra obtained for each of them are: HD 15371: 19 spectra, HD 48915 (Sirius): 31 spectra, HD 149348: 13 spectra, HD 193924: 15 spectra. We checked the instrumental stability of the spectrograph and the consistency of the reduction procedures through the comparison of the spectra of these stars.

The reliability of the photometric calibration given by the flat field correction has been investigated through the measurement of the equivalent width (hereafter written EW) of lines selected over the observed spectral range; the number of lines measured in each spectrum was: 9 for HD 15371, 29 for HD 48915, 15 for HD 149348 and 6 for HD 193924. For each of these stars, we controlled that the distribution of the EWs of the selected lines is Gaussian; the mean value of the EW for each of these lines has been computed and the mean value of all the standard deviation is smaller than 0.005 Å.

The mechanical stability of the instrument has been tested through the measurement of the position of the same set of lines in the same stars. The final result is that the rms of line wavelengths corresponds to 1 $\mathrm {km\, s^{-1}}$.

We verified also, using all the spectra of HD 48915, that the profile shape, after the normalization of the spectra, is conserved whatever the setup of the spectrograph has been.

These necessary tests allow us to use confidently the whole set of observations as an homogeneous sample of data in the following analysis.