To calibrate our CCD magnitudes into the Johnson B, V and Cousins R standard system
(hereafter B, V, R),
we observed the faint standard star sequence NGC 300 from
Graham (1981) and
the bright standard star sequences MARKA, NTPHE, PG0231, SA92 from
Landolt (1992).
Several sequences were observed each night in order to obtain a zero-point for
each night.
We use the following linear transformation equations to convert instrumental
magnitudes 
into standard magnitudes 
:
where 
define the standard magnitudes and 
define
the instrumental magnitudes (in adu s
) through
respectively the three filters B, V, R.
The colour term COLOUR is the standard colour provided by
Graham or Landolt B-V or V-R.
The instrumental CCD magnitude of standard stars is calculated as the
"corrected isophotal'' magnitude (see Sect. 4.2.1), thus yielding a good
estimation of the total magnitude.
The extinction coefficients 
for the different filters B, V, R
are derived from the La Silla extinction curve provided in the ESO manual
(Schwarz & Melnick 1993).
The zero-points 
are specific to each combination of
Telescope/Instrument/ CCD/Filter/atmospheric conditions. These are easily measured
using a large number of standard stars observed each night (see Fig. 6 (click here)).
The colour coefficients 
allow the magnitudes from the
"observing'' filter (resulting from the Telescope/Instrument/CCD/Filter
combination) to be corrected into the B, V, R standard filters.
We assume that they remain constant during
each observing run, and use the calibrations of an entire run to calculate these
coefficients.
Two color coefficients for the V
band can be calculated depending on whether B-V or V-R colours are
used (denoted 
or 
respectively).
In practice, we estimate the colour coefficients for each different
configuration of Telescope /Instrument / CCD /Filter as shown in Fig. 5 (click here).
The resulting colour coefficients are shown in Table 3. The
quoted errors are the rms. uncertainties in the linear fit.
The measured colour coefficients are then used to determine the
accurate zero-point 
for each night.
Figure 5: Colour term as a function of colour for standard star
sequences of an entire observing run (NTT/EMMI-B, CCD TEK#31).
The error bars for each
star is the quadratic sum of the instrumental error resulting from several
measurements and the intrinsic error given by the authors. The solid line
represents the weighted least-squares regression whose slope provides the
value of 
. The 
are chosen
for each set of stars of each night to
minimize the scatter in the colour term 
Figure 6: Resulting zero-point value 
for one observing night
using the color coefficient 
calculated in Fig. 5 (click here).
The error bars for stars are the same as in Fig. 5 (click here), and the zero-point
is calculated by weighted least-squares regression by a constant term
Table 3: Measured colour coefficients