2 Observations and data reduction

The observations presented here were obtained as part of an international collaboration at the observatories of the Canary Islands in several campaigns during the years 1990 and 1991. We used the Cassegrain foci of the 2.5 m Isaac Newton (INT) Telescope and Nordic Optical (NOT) Telescope at the Observatorio del Roque de los Muchachos (La Palma). Spectra in the region of the LiI line were obtained using the Intermediate Dispersion Spectrograph (IDS) and the IACUB echelle spectrograph at the INT and NOT, respectively. The H1800V grating and the 500 mm camera with a $385 \times 578$ GEC-CCD were used at the IDS, providing a dispersion of 0.22 Å per pixel. The adopted slit-width was usually 1 arcsec giving an effective resolution of 0.44 Å. The IACUB echelle spectrograph (McKeith et al. 1993) was used with a Thomson $1024 \times 1024$ CCD binned in the spectral direction to provide a dispersion of 0.1 Å per pixel. The 0.7 arcsec slit width employed gave a final effective resolution of $\sim$ 0.22 Å (in the case of G181-47 we used a 1 arcsec slit which gave 0.30 Å). Typically, we recorded spectra with exposure times $\sim 1800$ s. For the fainter stars, two or more exposures were combined in order to have a homogeneous sample with signal to noise in the range 100-200. Using the same instrumental configuration, several flat-field lamps were recorded each night. For wavelength calibration ${\rm Cu}-{\rm Ar}$ lamp spectra were used.

Most of the objects were selected from the sample of metal-poor stars in the works of Carney et al. (1987), Laird et al. (1988), and Schuster & Nissen (1988, 1989). Table 1 lists the stars observed, telescopes used, epoch of observation, exposure times and visual magnitudes. Several typical spectra in the region of the LiI line are plotted in Fig. 1. Table 2 lists the relevant photometric data and metallicities which will be used in a forthcoming paper to determine the stellar parameters $T_{\rm eff}$ and $\log g$ needed for the abundance analysis. Photometric magnitudes were obtained from the literature and the SIMBAD data base. Metallicities were also obtained from the literature using mainly the compilations by Schuster & Nissen (1989), and Carney et al. (1987). In general, the differences between these metallicities for the objects in common are of order 0.1-0.2 dex, but in a few of them (the most metal-poor stars) these differences can be as large as 1 dex.

  

Table 1: Programme stars

 

Table 1: continued

  

Table 2: Photometry and Li I equivalent widths of the programme stars

 

Table 2: continued

 

Table 2: continued

All the images were processed with the IRAF package following standard techniques of bias subtraction, flat-field corrections, optimal spectrum extraction and sky subtraction, wavelength calibration and linearization, and continuum normalization. The identification of the LiI $\lambda 670.8$ nm line was performed on the basis of the radial velocities measured by Laird et al. (1988), and Carney & Latham (1987), and/or the relative positions of the FeI $\lambda 667.8$ and CaI $\lambda 671.7$ nm lines. An unambiguous identification of the LiI spectral feature was always possible for the stars in Table 1. Those few very metal-poor stars in our original sample with poor signal to noise ratio and nearly featureless spectra for which the radial velocity was unknown have not been included in this paper.

  

Figure 1: Spectra of several stars of our sample in the LiI $\lambda$670.8 nm region. A three pixels box-car smoothing was applied to each spectrum