2 The photographic plate data

The photometric catalogue of the galaxies in the direction of the Abell 496 cluster of galaxies was obtained by processing the field number 621 in the SRC-J Schmidt atlas. Part of the glass copy of this blue plate (IIIaJ+GG385) was investigated in June 1993 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). Due to the cluster location with respect to the plate boundaries, only a 2.5 $^{\circ}\times$2.5$^\circ$ square region roughly centered on the cluster coordinates could indeed be searched for objects. At the cluster redshift, this limited scan is however large enough to secure a correct investigation of the overall cluster area (radius of $\sim$ 4.5  h₅₀^-1 Mpc at least).

The algorithms involved in the MAMA on-line mode available at that time for detecting and measuring astronomical sources are summarized in Slezak et al. (1998). Basically, objects appear to be defined as a set of connected pixels with intensity higher than typically a 2 sigma threshold above the local sky background and they are described by their flux (the sum of background-subtracted pixel values, which lead to what is called hereafter a "plate'' magnitude), their area and their elliptical shape parameters. This approach is efficient in most cases. However, one must expect that this simple vision model fails for crowded fields where the probability to get blended objects increases drastically when low detection thresholds are applied. The galactic latitude of Abell 496 is $b_{\rm II} \simeq -34.4^{\circ}$and some blends with stars are indeed included in the 35541 individual sources listed in the on-line catalogue we were provided with (cf. the final visual check of the galaxy candidates which is described below).

The astrometric reduction of the whole catalogue was performed with respect to 98 stars of the PPM star catalogue (Roeser & Bastian 1991) spread over the plate using a 3$^{\rm rd}$-order polynomial fitting. The residuals of the fit yielding the instrumental constants were smaller than 0.20 arcsecond and the astrometry of our catalogue indeed appears to be very good, as confirmed by cross-checking galaxy coordinates with the literature or the APM database (mean average separation: 0.7 arcsec) and by our multi-object fibre spectroscopy follow-up where the galaxies were always found to be very close to the expected positions.

The CCD observations needed to calibrate accurately the Schmidt plate were not available at that time. Hence, a preliminary photometric calibration of these photographic data has been done using galaxies with known total blue magnitude. The use of catalogued stars was rejected since such high-surface brightness objects suffer from too severe saturation effects (coming both from the emulsion itself and from the electronic settings of the MAMA facility). So, 40 galaxies available in the Lyon Extragalactic Database (LEDA, Paturel et al. 1997) were selected and their magnitudes compared to their measured blue fluxes providing that no close or overlapping objects were present as checked from a small scan around each LEDA galaxy. These 31 undisturbed objects span a 3 magnitude range, but in a very non uniform way. Hence, the 3$\sigma$ clipping routine used to compute the best fit further discarded 15 objects. The final rms on the zero-point is unfortunately not better than 0.4 mag owing to: i) the quite large uncertainty quoted for the total magnitude estimate in the database along with their irregular distribution, ii) the limited magnitude range, and iii) differences in the involved flux estimates ("plate'' vs. total magnitude).

A basic star-galaxy separation has been performed mainly with respect to a classical surface brightness criterion. As usual for glass copies of survey plates, the ability of this criterion to discriminate drops sharply for objects fainter than approximately 19$^{\rm th}$magnitude. However completeness and purity of any catalogue are most of the time competing goals. So, another test based on the elongation was then performed in order to reject linear plate flaws, as well as to pick bright elongated galaxies first classified as stars due to strong saturation effects. Finally, spurious detections that occur around very bright stars (area greater than 10³ pixels) due to an incorrect estimate of the local background were tentatively removed by checking their location with respect to these bright objects (the detection processing included no smoothing). Down to the detection limit, we obtained a list of more than 4000 galaxy candidates over our SRC-J 621 blue field to the detection limit.

The differential luminosity distribution of our catalogue of diffuse objects indicates that this sample is complete down to the $b_{\rm J}=$ 19.5 magnitude (see Fig. 2). However, its purity is less than the usual 95% level for high galactic latitude fields (cf. Slezak et al. 1998). Our spectroscopic run indeed indicates a contamination level by stars at least three times higher ($\simeq$ 20%, see Durret et al. 1999). Such a low success rate may partly be ascribed to the overall image quality at the plate corner where Abell 496 is located. As usual for Schmidt plates the PSF may indeed be quite poor at the borders, which randomly increases the fuzziness of otherwise point-like objects and thereby leads to parameter estimates closer to those of diffuse objects than to genuine star-like ones (cf. Fig. 1 where an overdensity is clearly visible at the edges of the plate).

Figure 1: Spatial distribution of 3879 galaxy candidates in the field

More generally, one can also question the efficiency of the classification procedure itself when the number density of stars is very high. In fact, a selection mostly based on a surface brightness criterion implies keeping $\simeq$ 5% of the total number of stars in order to select most of the galaxies. Hence, in most cases, the contamination level of a quite complete galaxy sample unavoidably increases with the star number density. The galactic latitude of Abell 496 is $b_{\rm II}=-$34.4$^\circ$, which is quite close to the Galactic plane and may explain the high absolute number of misclassified single stars. But, for the present data, blended stars identified as a single diffuse object is the main explanation, as first noticed during the spectroscopic run and partly expected from the involved detection software. This is confirmed by a check of our list of bright ( $b_{\rm J}<$ 17.5) galaxy candidates against the APM list for the #621 SRC-J field. Among our 449 candidates within the same celestial zone, 389 objects are described as nebular in the APM catalogue, 52 are classified as star-like, 1 is related to a plate flaw while 10 have no close counterpart. As evidenced by a visual check using the DSS, among these 62 discrepant objects, there are 8 genuine galaxies, 2 asymmetrical objects, 40 blended stars, 8 single bright stars (important saturation, diffraction spikes), and 4 star-like objects. So, it appears that only 52 objects are misclassified (12% of the total number of candidates), out of which 75% are merged images, plus two asymmetrical objects.

Figure 2: Magnitude distribution in the $b_{\rm J}$ band of the 3879 galaxy candidates in Abell 496 photographic plate field

On one hand, this $\simeq$ 15% contamination level is disturbing for studies involving individual objects picked among galaxy candidates. On the other hand, its net effect for statistical studies is only a decrease of the contrast for the signal of interest providing that the misclassified stars are randomly distributed. So, for such applications, the present photometric catalogue of Abell 496 certainly remains valuable.

Table 1 lists the catalogue of galaxy candidates obtained from the SRC-J 621 plate in the 2.5 $^{\circ}\times$2.5$^\circ$ field of Abell 496. Note that the 101 misclassified stars we were able to identify during our spectroscopic follow-up have been rejected, as well as the 54 objects selected by a visual check (the 52 misclassified objects and 2 asymmetrical objects described above), yielding 3879 entries. The meaning of the columns is the following:
(1) running number;
(2) to (4) right ascension (equinox 2000.0);
(5) to (7) declination (equinox 2000.0);
(8) half-major axis (arcseconds);
(9) excentricity e defined as $\sqrt{1-({b\over a})^2}$, where a and b are the major and minor axes respectively;
(10) position angle of the major axis (from North to East);
(11) $b_{\rm J}$ magnitude;
(12) and (13) X and Y positions in arcsecond relative to the centre defined as that of the diffuse X-ray emission of the cluster (see Sect. 3.1);
(14) distance to cluster center in arcseconds;
(15) MAMA catalogue reference number.