2 The data

2.1 Broad-band images

HCG objects were imaged using the 1.20-m telescope of the Observatoire de Haute-Provence, with a Tektronix $1024\times1024$ CCD. Each field covers a square of 11 $.\mkern-4mu^\prime$7 side, with a pixel size of 0 $.\!\!^{\prime\prime}$67. Exposure times of 600 s were secured through the V filter. The seeing conditions were mediocre, and the usage of our images has been limited to morphological considerations only.

We have also benefited from images of much better resolution, kindly communicated to us by D. Bettoni (for HCG 67), and by P. Hickson (for all three groups); hereafter, these R-band frames will be referred to as DB-Asiago and PH-CFHT images, respectively.

2.2 Spectroscopic observations

Long-slit spectra have been obtained at the 1.93-m telescope of the Observatoire de Haute-Provence, equipped with the CARELEC spectrograph. Spectra have a dispersion of 66 Å mm^-1 corresponding to 1.78 Å px^-1 on a $512\times512$ Tektronix CCD, with pixels of 27 $\mu$m; the actual spectral resolution is 3.2 Å FWHM. The wavelength range covered is $\approx4800<\lambda<5750$ Å. The scale perpendicular to the dispersion is 1 $.\!\!^{\prime\prime}$1, the slit width was 2 $.\!\!^{\prime\prime}$1. The exposure time and slit position angles of the different spectra are presented in Table 1. K0 III stars were also observed, as templates.

   

Table 1: Journal of spectroscopic observations

member	$\alpha$	$\delta$	exp. time	slit PA	seeing FWHM
ident.	(2000)	(2000)	(seconds)	(degrees NE)	(arcsec)
HCG 67a	13 49 11.3	-7 13 27.6	$3\times3000$	37	2.7
HCG 67d	13 49 09.7	-7 13 54.3	$3\times3000$	37	2.7
HCG 74a	15 19 24.7	21 06 20.3	$4\times3000$	21	2.1
HCG 74b	15 19 24.4	20 53 23.4	$4\times3000$	21	2.1
HCG 79a	15 59 11.4	20 45 14.9	$4\times3000$	65	2.9
HCG 79b	15 59 12.6	20 45 47.1	$4\times3000$	79	2.9
HCG 79c	15 59 10.9	20 45 41.4	$4\times3000$	79	2.9

2.3 Analysis of spectra

We have followed closely the technique used in Simien & Prugniel ([1997a], [1997b], [1997c], [1998]), whose main characteristics are summarized again, for convenience. Pre-reduction, cosmic-ray removal in the outer regions, wavelength calibration and sky subtraction were performed with ESO-MIDAS. Multiple spectra of the same object were co-added after applying a proper relative shift, i.e.: a) in $\lambda$, using the calibration spectra to monitor the mechanical flexures within the spectrograph, and b) perpendicularly to the dispersion, to match the nucleus position along the slit, down to a fraction of pixel.

The kinematical parameters are determined by a Fourier-Fitting technique similar to that described in Franx et al. ([1989]). This allows the simultaneous calculation of the rotation velocity V(r) and the velocity dispersion $\sigma(r)$, as a function of the distance r from the center.

The removal of the cosmic-ray hits in the central regions is performed on the galaxy spectrum within the same procedure; it is based on a three-sigma clipping of residuals between the galaxy spectrum and the properly shifted, convolved star spectrum. The V and $\sigma$ parameters are actually determined in two successive steps, using first the raw galaxy spectrum, and then the spectrum with cosmic rays removed as explained. Adjacent lines are averaged to enhance the S/N ratio of the spectrum in the outer regions. For a spectrum at radius r, a weight is assigned to the neighboring lines ( $r\pm\delta r$) entering the average; for this weight, a Gaussian fall-off as a function of $\delta r$ was chosen, with a FWHM width between 0 and 3.6 pixels (0 and $4\hbox{$.\!\!^{\prime\prime}$ }1$).

Our determinations of the heliocentric recession velocities $V_{0,{\rm hel}}$ and central velocity dispersions $\sigma_0$ are collected in Table 3 (also available in electronic form). The radial-velocity and velocity-dispersion profiles are displayed in Fig. 3 and Fig. 6, and proposed in electronic form in Table 4. Individual results are discussed in Sect. 3. MdOH presented the redshift survey of the entire Hickson's sample; our measurements of the systemic velocities have been compared with their measurements and, for the seven galaxies in common, the average difference is $\langle V_{\rm ours} - V_{\rm MdOH} \rangle = -24 \pm 99$ km s^-1, which is unsignificant.