4 Observations and data reduction

The stellar library was observed in a total of 13 observing runs from 1991 to 1996 using the JKT, INT and WHT at the Roque de los Muchachos Observatory (La Palma, Spain), the 3.5 m telescope at Calar Alto Observatory (Almería, Spain), and the 2.12 m telescope at San Pedro Mártir Observatory (Mexico). The bulk of the spectra (320 out of a total of 650 spectra) were taken in run 6, where the Richardson-Brealey Spectrograph was used to obtain 2.8 Å resolution spectra in the blue spectral range. A description of each observing run, including relevant observational configuration parameters, is given in Table 2. The number of stars observed in each run is quite variable since some of the runs were not devoted to the calibration of the break and, in these cases, only bright stars of the library were observed during twilight periods. The last column in Table 1 list the run numbers in which each star was observed.

  

Table 2: Observational configurations

The reduction of the data was performed with our own reduction package REDucmE [11, (Cardiel & Gorgas 1999)], which allows a parallel treatment of data and error spectra (see below). We followed a standard reduction procedure for spectroscopic data: bias and dark subtraction, cosmic ray cleaning, flat-fielding, wavelength calibration, C-distortion correction, sky subtraction, atmospheric extinction correction and flux calibration (we did not attempt to obtain absolute fluxes since, as most line-strength indices, the break only requires relative fluxes). Cluster stars spectra were also corrected from interstellar reddening, using the color excesses quoted in Tables 4 and A3 from G93 and W94, respectively, and the averaged extinction curve of [50, Savage & Mathis (1979)].

In order to optimize the observing time during most of the runs, comparison arc frames were not taken next to each star observation. Instead of this, we only acquired arc exposures for a selected subsample of stars, which comprised objects with a complete coverage of all the spectral types and luminosity classes observed in each run. The wavelength calibration of the rest of the stars was performed by comparing them with the reference spectra. The repetition of this procedure with different reference spectra allow us to guarantee that wavelength calibration errors were always  pixels.