Small galaxy systems such as pairs and Compact Groups (CGs) constitute the very lowest end of the clustering hierarchical scale. Given their high galaxy density and small velocity dispersion most CGs are expected to separate from their underlying background, become bound systems and ultimately collapse within a few crossing times. Actually the high frequency (or extreme longevity) of CGs can match the rather short lifetimes predicted by merger simulations (Barnes 1989) simply by varying the fraction of dark matter distributed through the group (Mamon 1987; Athanassoula et al. 1997; Zabludoff & Mulchaey 1998), assuming continuous accretion of infalling galaxies (Governato et al. 1996), or assuming that CGs are dense configurations that form temporarily within loose groups (Diaferio et al. 1994). An alternative scenario requires that merging CGs are continuously replaced by new forming ones (Mamon 2000).
To date it is difficult to further constrain the relative importance of parameters and correlations entering the modelling of CGs, essentially because no definite conclusions concerning fundamental properties of CGs have been achieved. A large unbiased sample is needed to provide statistically reliable answers to questions such as: Do isolated CGs really exist? And how does the request for minimum multiplicity depend upon magnitude and morphological classification of member galaxies? Hence, questions related to a proper choice of CG selection parameters become fundamental, whilst actually, these parameters are generally chosen according to criteria aiming at reducing contamination by non-physical structures. Indeed, the bound status of CGs is difficult to establish. CGs, unlike galaxy clusters, though presenting adequate mass density profiles, are generally too close (z<0.1) to induce efficient gravitational lensing phenomena (Mendes de Oliveira & Giraud 1994; Montoya et al. 1996), while concerning X-ray properties, diffuse emission tends to be associated only with embedded CGs in loose configurations that contain at least one early-type galaxy (Ponman et al. 1996; Mulchaey 2000; Heldson & Ponman 2000). Other tracers of a common potential well, such as HI or CO, are at present available only for a limited number of CGs (Williams & Rood 1987; Oosterloo & Iovino 1997; Verdes-Montenegro et al. 2001).
Therefore proximity in projected and redshift space, although affected by small number statistics, peculiar motions and interlopers (Moore et al. 1993; Diaferio et al. 1994), still remains the main tracer of physical association between galaxies in CGs. Interaction patterns and kinematical peculiarities between member galaxies constitute an a posteriori probe of physical association.
Because of the difficulty in identifying high redshift CGs, only low redshift CG samples are so far available. The best studied CG sample (HCGs, Hickson 1982, 1997) contains 92 CGs showing extremely heterogeneous characteristics. HCGs have been visually selected (according to multiplicity, isolation and luminosity concordance of member galaxies) and thus reflect some of the systematic biases intrinsic to identification of systems on the basis of their bidimensional distribution only. In order to overcome these biases automatic identification of CGs has been performed on a deep 2-D southern catalogue (SCGs, Prandoni et al. 1994; Iovino et al. 1999) and on 3-D catalogs (RSCGs, Barton et al. 1996, 1998). These studies intended to produce large CG samples by (partial) parametrical reproduction of Hickson's selection criteria. Hickson's isolation criterion has been slightly relaxed by Iovino et al. (1999) and not included at all by Barton et al. (1996), who additionally, included triplets among CGs. Triplets are structures generally excluded in bidimensional selected CG samples because, apart from the expected high contamination by superposed fore/background galaxies, they might represent a collection of unrelated field galaxies, rather than a physical structure (Diaferio et al. 1994). The Catalogue of Triple Galaxies (Karachentseva et al. 1979; Karachentseva & Karachentsev 2000) constitutes the exception, but because of poor number statistics affecting dynamical parameters, Triplets have so far been investigated mainly in relation to their high peculiar galaxy content. Recent availability of a 3-D large galaxy sample, including nearly 20 000 redshifts for northern galaxies brighter than mB=15.5 (UZC, Falco et al. 1999), allowed us to construct a large CG sample selected on the basis of their compactness only. In selecting the sample we did not try to reproduce any of Hickson's criteria except compactness, in order to check if and at which level the properties of CGs are linked to multiplicity, to the large scale environment and to the luminosity and spectral properties of member galaxies.
The CG selection algorithm is described in Sect. 2. In Sect. 3 we describe the UZC catalogue and the prescriptions for the algorithm input parameters. The analysis of the characteristics of Triplets (Ts) and Multiplets (Ms) are presented in Sect. 4. In Sect. 5 statistical reliability of the CG sample is discussed. In Sect. 6 spectral properties of CG galaxy members are presented. CGs large scale environment and surface density contrast are analysed in Sects. 7 and 8 respectively. In Sect. 9 the relation between adopted selection parameters and the properties of the resulting CG sample are discussed. Conclusions are drawn in Sect. 10.
A Hubble constant of
km s
Mpc
is used throughout.
Copyright ESO 2002