We decided to obtain a photometric catalogue of the galaxies in the
direction of the Abell 85 cluster of galaxies by first processing the field
681 in the SRC-J Schmidt atlas. This blue glass copy plate (IIIaJ+GG385)
was investigated with the MAMA (Machine à Mesurer pour l'Astronomie)
facility located at the Centre d'Analyse des Images at the Observatoire de
Paris and operated by CNRS/INSU (Institut National des Sciences de l'Univers).
In order to also get information on the neighbouring galaxy distribution, the
central 
area has been searched for objects
using the on-line mode with the 10 
step size available at that time.
The involved algorithmic steps are well-known. They can be summarized as
follows:
first a local background estimate and its variance are computed from
pixel values inside a 256 
256 window, then pixels with a
number of counts higher than the background value plus three times the
variance are flagged, which leads to define an object as a set of
connected flagged pixels; an overlapping zone of 512 pixels is used in
both directions for each individual scan. Although this method may appear
rather crude, its efficiency is nevertheless quite high for properly
detecting and measuring simple and isolated objects smaller than the
background scale. The region where ABCG 85 is located is not crowded by
stellar images (
), so that most of the objects
larger than a few pixels can indeed be detected this way.
The result was a list of more than 105 objects distributed over
the 
25 square degrees of the field bounded by 
and
(equinox 2000.0, as hereafter), with their
coordinates, their shape parameters (area, elliptical modelling) and
two flux descriptors (peak density, sum of background-subtracted
pixel values).
The astrometric reduction of the whole catalogue was performed with
respect to 91 stars of the PPM star catalogue
(Roeser & Bastian 1991)
spread over the field, using a 3
-order polynomial fitting. The
residuals of the fit yielding the instrumental constants were smaller
than 0.25 arcsecond and the astrometry of our catalogue indeed appears
to be very good, as confirmed by our multi-object fibre spectroscopy where
the galaxies were always found to be very close (< 2.0 arcsec, i.e.
3 pixels) to the expected positions.
Since the required CCD observations were not available at that time, a preliminary photometric calibration of these photographic data has been done using galaxies with known total blue magnitude. The magnitude of stars is certainly much easier to define, but such high-surface brightness objects suffer from severe saturation effects on Schmidt plates when they are bright enough to be included in available photometric catalogues. So, 83 galaxies were selected from the Lyon Extragalactic Database (LEDA) in order to compare their magnitude to their measured blue flux. A small region around each of these objects was scanned and this image has been used: i) to identify the object among its neighbours within the coordinate list and ii) to assess the quality of the flux value stored in the on-line catalogue with respect to close, overlapping or merged objects. The 74 remaining undisturbed objects identified with no ambiguity came from eight different catalogues in the literature. Whatever the intrinsic uncertainties about the integrated MAMA fluxes are, systematic effects were found with respect to the parent catalogue in a flux versus magnitude plot, as well as discrepancies for some objects between the LEDA and the Centre de Données Astronomiques de Strasbourg (CDS) databases. Consequently, three catalogues including 12 objects were removed and the LEDA magnitude of 5 objects was replaced by a CDS value which seems in better agreement with their aspect and with the overall trend when compared to similar objects. Later, 7 objects far from the overall trend were discarded. These successive rejections resulted in a set of 55 objects distributed over a six magnitude range. The magnitude zero-point for our photographic catalogue was obtained by plotting the flux of these objects against their expected magnitude. A rms scatter of 0.34 mag was computed around the linear fit.
Most of the diffuse objects included in our main catalogue were
automatically selected according to their lower surface brightness
when compared to stars. As usual for glass copies of survey plates,
the discrimination power of this brightness criterion drops sharply
for objects fainter than approximately 19
magnitude, and so does
the completeness of the resulting catalogue if no
contamination is allowed for. The number of galaxy candidates brighter
than this limit within the investigated area appeared, however, to be
already large enough to get a much better view of the bright galaxy
distribution than using the deeper but very incomplete catalogue
published by Murphy (1984). Moreover, including faintest objects was
not necessary for the redshift survey of the Abell 85 cluster of
galaxies we were planning (see Durret et al. 1997).
Hence, no attempt was
done to reach a fainter completeness limit. Nonetheless, in order to
select galaxies, the decision curve which has been computed in the Flux
vs. Area parameter space was fitted to the data so that
some objects identified by Murphy from CCD frames as faint galaxies
were also classified as galaxies by us.
Next, a further test based on
the elongation was performed in order to reject linear plate flaws or
artefacts, as well as to pick bright elongated galaxies first
classified as stars due to strong saturation effects. Finally,
spurious detections occuring around very bright stars (area greater
than 103 pixels) due to a wrong estimate of the local background were
tentatively removed by checking their location with respect to these
bright objects.
In this way, a list of more than 25000 galaxy candidates over the 25
square degrees of our SRC-J 681 blue field was obtained.
Figure 1: Spatial distribution of the 11862 galaxies brighter than
in the SRC-J 681 field. The large overdensities are
indicated by
superimposed isopleths from a density map computed by the method introduced
by Dressler (1980) with N=50; eleven isopleths are drawn from 850 to
2850 galaxies/square degree
The distribution of these galaxies is displayed in Fig. 1 (click here)
for objects brighter than . The Abell 85 cluster is
clearly visible, as well as several other density enhancements which
are mostly located along the direction defined by the cluster
ellipticity.
The differential luminosity distribution of the galaxy candidates
indicates that the sample appears quite complete down to the
19.75 magnitude (see Fig. 2 (click here)). To go further, we first
tested the completeness of this overall list by cross-identifying it
with three catalogues from the literature (Murphy 1984;
Beers et
al. 1991; Malumuth et al. 1992) with the help of images obtained from
the mapping mode of the MAMA machine.
Figure 2: Differential magnitude distribution of the 25 103 galaxy
candidates in the SRC-J 681 field
It appeared that: i) all but one
galaxy of the Malumuth et al. (1992) catalogue of 165 objects are actually
classified as galaxies, with a mean offset between individual
positions equal to 1.10 
0.06 arcsecond; ii) 94% of the 35
galaxies listed by Beers et al. (1991) inside the area are included in our
catalogue, only 2 bright objects which suffer from severe saturation
being misclassified. Note that such an effect also caused 5 of the 83
galaxies chosen as photometric standards to be misclassified, which
gives the same percentage as for the sample by Beers et al. The
comparison with the faint CCD catalogue built by
Murphy (1984) in the so-called
band (quite similar to that obtained using a photographic IIIaF
emulsion with a R filter) was performed only for objects which were visible
on the photographic plate with secure identification (only uncertain X
and Y coordinates are provided in the paper) and classified without any
doubt as galaxies from our visual examination. There remained 107
objects out of 170, among which 88 are brighter than 
19.75
(
18.5).
Down to this flux limit, 82 objects ( 93%)
are in agreement, thereby validating the choice of our decision curve in the
Flux vs. Area parameter space. These cross-identifications therefore indicate
that the completeness limit of our catalogue is about 95% for such objects,
as expected from similar studies at high galactic latitude.
In order to confirm this statement and to study the homogeneity of our
galaxy catalogue, we then decided to verify carefully its reliability
inside the region of the Abell 85 cluster of galaxies itself. The
centre of ABCG 85 was assumed to be located at the equatorial
coordinates given in the literature, 
and 
, and a square
region of 
around this position was defined; such an
angular distance corresponds to 2.7 Mpc h100-1 at the
redshift of the cluster (z=0.0555).
However, let us remark that the
position of the central D galaxy is slightly different,
and
, and so is the centre we found from
our X-ray analysis of the diffuse component of this cluster, i.e.:
and
(Pislar et al. 1997). For all our
future studies, we then chose to define the cluster centre as that of
this X-ray component.
The distribution of the 
candidates within the area has
been first of all visually inspected to remove remaining conspicuous
false detections around some stars as well as some defects mainly due to a
satellite track crossing the field. This cleaned catalogue contains a
little more than 4000 galaxy-like objects, half of which brighter
than . The intrinsic quality of this list has then been
checked against a visual classification of all the recorded objects
within a 
area covering the region already
observed by Murphy (1984) around the location 
and 
. The
inspection of the corresponding MAMA frame of 
pixels enabled us to give a morphological code to each object, as well
as to flag superimposed objects and to deblend manually 10 galaxies
(new positions and flux estimates for each galaxy member). Of course,
the discrimination power of this visual examination decreases for
star-like objects fainter than 
due to the sampling involved (pixel size of 0.67''), and an exact
classification of such objects appeared to be hopeless above the
a priori completeness limit of our automated galaxy list guessed to
be . Down to this limit, our results can be summarized as
follows: i) 
of the selected galaxies are true galaxies
(including 7 multiple galaxies and 2 mergers with stars), while 4%
may be galaxies; ii) 7 genuine galaxies are missed (4%).
Since these
contamination and incompleteness levels of 
were
satisfactory, we decided to set the completeness limit for our
automated galaxy catalogue at this magnitude .
For objects fainter than our completeness limit, the visual check of
the inner 
part of our object list has enabled us to
confirm the galaxy identification of 135 galaxy candidates as well as
to select 214 misclassified faint galaxies. The total number of
galaxies included in the visual sample down to the detection limit is
541, whereas the initial list only contains 338 candidates within the
same area. Keeping in mind that both catalogues are almost identical
for objects brighter than , we decided to replace the
automated list by the visual one inside this central
area. Note that about 150 objects remained unclassified, including 26
galaxies from the CCD list by Murphy. We added these 26 galaxies to
the final catalogue whose galaxies are plotted in Fig. 3 (click here).
Figure 3: Positions of the 4232 galaxies detected on the photographic plate
relative to the centre of the cluster defined as the centre of the diffuse
X-ray component. North is to the top and East to the left
Table 1 lists the merged catalogue of 4,232 galaxies obtained from the
SRC-J 681 plate in the field of ABCG 85, with V and R
magnitudes computed using the transformation laws obtained from our CCD
data (see Sect. 3.3). This table includes the following information: running
number; equatorial coordinates (equinox 2000.0); ellipticity; position
angle of the major axis; 
, V, and R magnitudes; X and Y
positions in arcsecond relative to the centre defined as that of the diffuse
X-ray emission of the cluster (see above); cross-identifications with the
lists by Malumuth et al. (1992);
Beers et al. (1991) and Murphy (1984).