This survey was performed in the context of an ESO-key programme, thus guaranteeing
the necessary observing time to perform the full programme. The observations began
in 1989 at the 3.60 m telescope and were transferred to the NTT-3.5 m in 1991.
At a pace of two observing runs per year, the full time allocation was consumed
by the fall 1995.
The observing strategy is designed to make optimal use of telescope time.
The great advantage of the multi-mode instruments used (EMMI at the NTT,
D'Odorico
1990; Dekker et al. 1991, and EFOSC at the 3.6 m,
Buzzoni et al. 1984) is the
possibility to switch between spectroscopic and photometric observations during
the course of the same night, using the same instrument.
It allows one to adapt to variable weather conditions: when the conditions are
photometric with good seeing quality, priority is given to imaging; if
the weather degrades, one can switch to the spectroscopic mode by simple rotation of
two wheels placing an aperture mask and a grism into the optical path.
Because the number of observing nights is spread over many runs, the imaging data
of a given run can be reduced in preparation for the subsequent run, thus yielding
finding charts of the galaxies with 
to be observed spectroscopically.
During each run, we observe our main survey region in first priority, and two auxiliary
regions at the beginning and end of the night when the airmass for the main survey
is large.
Here we only report on the data for the main survey region. The data in the auxiliary
fields will be used as control samples.
The position, galactic latitude and solid angle of
the main field are given in Table 2 (click here).
In this region of the sky, we acquired 25-30
frames in each of the B, V, R
bands with the 3.6 m telescope.
These data cover the right ascension range
, and correspond to an area of 
.
With the NTT telescope, we acquired 20-25 frames with
(
).
The 
width in declination is covered by 3 EFOSC fields and 2 EMMI
fields.
Table 1: Survey characteristics
Because we performed the photometric observations over a period of six years, different CCD cameras were used over the years to follow the improvement in CCD technology, so that the size and the characteristics of the individual CCD frames vary across the mosaic (made of typically 50 frames in each band). We summarize the principal CCD characteristics in Table 2.1 (click here). The first column gives the telescope and instrument used with the different channel for EMMI on the NTT (blue imaging channel: BIMG, and red imaging channel: RIMG). The useful area is the used part of the CCD which differs for large CCDs (LORAL, TEKTRONIX) from the total area because of vignetting.
Table 2: CCD characteristics and photometric observations
At the 3.60 m telescope with EFOSC, the photometric observations were performed in a range of
seeing between 1.10 and 1.65 arcsec. When the FWHM is < 2 pixels, the
profile of unresolved objects is poorly sampled, which degrades the star/galaxy
separation. With EFOSC, we observed the photometric fields with airmasses between
1.0 and 1.4, and the exposure times for the
B, V, R bands are respectively 30 , 25, 20 minutes.
The majority of the fields were obtained in only 1 exposure in each band.
Bright stars (
) are thus saturated.
Since 1991, we have observed at the NTT with an average seeing of 
arcsec.
As the pixel size of the CCDs is smaller, we have a better sampling of the
p.s.f. and therefore our star/galaxy separation is limited by seeing.
The airmasses are also in the range 1.0 - 1.4. The exposure times for the
B, V, R bands were respectively 25, 20, 15 minutes.
To avoid saturation of the CCDs by bright stars, the images are summed from
2 or 3 exposures.
The photometric strategy adopted to probe the different strips is to obtain a mosaic of
CCD frames regularly offset by 
of the CCD size providing many overlaps
for subsequent checks of the photometry.
In Table 2.1 (click here), we summarize the characteristics of the photometric observations during
the different runs. For each run,
the photometric zero-points (apparent magnitude of a star with an
absolute magnitude M = 0. and null color term exposed during 1 second, as
defined in Eq. (1)) are given.