Figure 1: Spectrum of PKS 2126-158 in the wavelength
range 
covered by the new observations showing
fits to identified lines overlaid upon the normalized spectrum.
The spectral interval 
and the corresponding
absorption lines are shown in GCFT.
The dashed line represents the noise per resolution element.
Upper vertical ticks correspond to Ly
lines, lower
ticks correspond to metal lines
The determination of the continuum in the QSO spectrum is a critical step
because it affects the measurement of the absorption line parameters. While
the continuum redward of the Ly emission can be drawn without difficulty, the
high line density in the Lyman forest complicates the operation
(Young et al. 1979; Carswell et al. 1982).
In the present work, the task has been fulfilled with the help of a
procedure, allowing to select automatically and in a
reproducible way the regions of the spectrum free of strong absorption
lines or artificial peaks (e.g. due to cosmic rays), i.e. where the RMS
fluctuation about the mean becomes consistent with noise statistics.
The continuum level has been estimated by spline-fitting these regions
with quadratic polynomials.
The normalized spectrum is shown in Fig. 1 (click here).
The detection and measurement of absorption lines in the spectrum have
been carried out as in GCFT and we refer to this paper for details of the
procedure. In particular, the lines have been fitted with Voigt profiles
convolved with the instrumental spread function, making use of a
minimization method of 
. This step has been performed within the
MIDAS package with the programme FITLYMAN (Fontana & Ballester 1995). The values of
redshift z, Doppler parameter 
(where 
is
the velocity dispersion) and column density N
have been determined for isolated lines and individual components of blends.
The number of components of each absorption feature is assumed to be the
minimum required to give a reduced 
(corresponding to a
confidence level 
).
The identification of the metal systems is described in Sect. 5 (click here)
All the lines shortward of the Ly emission not identified as due to metals
have been fitted as Ly and Ly
.
For a further control, we used the
blue, lower-resolution spectra to search for
Ly, Ly
and Ly
lines
in correspondence to the stronger Ly lines (
).
In few cases, we could ascertain that the absorption was not Ly.
Such lines are listed as
unidentified in Table 2 and they probably belong to metal
systems still to be recognized.