2. The sample
2.1. The photometric catalogue
The starting photometric catalogue is the COSMOS/UKST galaxy catalogue of the southern sky (Yentis et al. 1992), obtained from automated scans of UKSTJ plates by the COSMOS machine. We extracted a circular region of diameter, centered on and , containing 11525 objects to the limiting magnitude .
Fig. 1a shows the isodensity contours obtained binning the data in 2 2 arcmin cells and smoothing with a Gaussian of 6 arcmin of FWHM. For the two clusters circles of one Abell radius have been drawn around their nominal center. Note that it is already evident that A 548 is not a smooth cluster with a single central nucleus, but presents multiple condensations. Inside the Abell circle of A 3367 we note a single condensation shifted northward with respect to the nominal center. In Fig. 1b the same isodensity contours are shown with superimposed the OPTOPUS fields positions.
The coordinates of the centers of these fields are listed in columns (2) and (3) of Table 1, together with the observation date in column (4).
Spectroscopic measurements were obtained using the ESO 3.6m telescope at La Silla, equipped with the OPTOPUS multifiber spectrograph (Lund 1986), on the nights of 1993 February 25-26 and October 16-17.
The OPTOPUS multifiber spectrograph is formed by a bundle of 50 optical fibres at the Cassegrain focal plane of the telescope; this field has a diameter of 32 arcmin, and each fibre has a projected size on the sky of 2.5 arcsec. We used the ESO grating 15 with 300 lines/mm and a blaze angle of . This grating allows a dispersion of 174 Å/mm in our wavelength range (3700-6100) Å. We used the detector Tektronic 512 512 CCD with a pixel size of 27 µm, corresponding to 4.5 Å, i.e. a velocity bin of 270 km/s at 5000 Å; the resolution is Å. Four of the 50 fibres were dedicated to sky measurements, leaving 46 fibres available for the objects.
Fields f51a, f51b, f52 and f53 were observed for 1 hour, split into two half-hour exposures in order minimize the presence of cosmic hits. Fields f61 and f62 were observed only for 1/2 hour. The observing sequence was a 30-s exposure of a quartz halogen white lamp, a 60-s exposure of helium vapour lamp for fields f61 and f62, or 60-s exposure of helium + neon vapour lamp for fields f51, f52 and f53; then the scientific field, and again the arc and the white lamp.
2.3. Data reduction
The extraction of the one-dimensional spectra was performed using the APEXTRACT package as implemented in IRAF 1.
Positions and tracing solutions of lamps and objects were determined on the flat field exposures. The procedure we adopted to estimate the relative transmission of each fibre is based on the fitting of a Gaussian profile to the [OI]5577 sky line in each spectrum and on computing the continuum-subtracted flux of this line (Bardelli et al. 1994). If we assume that the flux and the shape of the spectrum of the night sky remain constant in the telescope field, this value is the same in each spectrum apart from the transmission of the fiber, which is a multiplicative factor. After having normalized the spectra, we can subtract the `mean sky' obtained as the average of the 4 sky spectra.
2.4. Redshift data
We have obtained a total of 276 spectra: 45 were not useful for redshift determination ( of the total), because of poor signal-to-noise ratio or badly connected fibers, and 51 turned out to be stars ( of the reliable spectra), leaving us with 180 galaxy redshifts. The galaxies whose spectrum presents detectable emission lines are 79, corresponding to a percentage of of the total.
The radial velocities of galaxies with spectra with absorption lines have been determined using the program XCSAO in the IRAF task RVSAO (Kurtz et al. 1992), which is based on the cross-correlation method of Tonry & Davis (1979). The determination of redshift is done by fitting a parabola to the main peak of the cross-correlation function. Sixteen different templates (eight stars and eight galaxies) were used for the determination of the radial velocities, choosing as better estimate the one which gave the minimum cross-correlation error, defined as:
where w is the FWHM of the cross-correlation peak and r is the ratio between the height of the correlation peak and the rms of the antisymmetric part of the correlation function (Kurtz et al. 1992).
To estimate the redshift of spectra with strong emission lines we used the EMSAO program in the IRAF task RVSAO.
The top panel of Fig. 2 shows an example of a spectrum with strong emission features, with [OII], [H], [OIII], lines, while in the bottom panel of Fig. 2 a spectrum with only absorption lines is presented.
In Tables 2-7 we list the galaxies with redshift determination. Columns (1), (2) and (3) list the right ascension, the declination and the magnitude respectively; column (4) and (5) give the heliocentric velocity () and its internal error (in km/s), from absorption and emission lines respectively. The code in column (6) indicates the presence of emission lines: the symbols a, b, c, d, refer to [OII]3727Å, [H]4861Å, [OIII]4959Å and [OIII]5007Å, respectively.
We remember that the cross-correlation errors are only internal formal errors. In order to have true statistical errors, these values have to be multiplied for the factor 1.53 found by Vettolani et al. (1998) comparing multiple observations of the same galaxies: after this correction, the average statistical error on our velocities is 95 km/s. If one wants to take into account also the uncertainties introduced by the different reduction procedures, the factor is slightly larger and has the value of (see Bardelli et al. 1994).
In order to check the zero point precision of our velocity scale, we considered the histogram of the measured velocities of the stars misclassified as galaxies (Fig. 3), which are expected to have a zero mean velocity. Considering only the 41 spectra with the higher signal-to-noise ratio, we found km/s ( km/s): this small systematic effect will be neglected in the following analysis, since the errors associated to the galaxy velocities are larger. However, we cannot exclude that the value of is completely due to bulk motions of stars in this region of the sky.
Very recently, Cappi et al. (1998), analysing the ESP survey (Vettolani et al. 1997, 1998), noted a systematic difference between the velocities estimated from the emission lines and the cross-correlation for the same galaxy, with an average difference of km/s (obtained from about 750 galaxies). Our observations are taken in the same instrumental configuration of the ESP survey and can give an independent estimate of this effect, although with a smaller sample. On the basis of 10 galaxies, we find km/s, consistent within the errors with the result of Cappi et al. (1998).
© European Southern Observatory (ESO) 1998
Online publication: August 6, 1998