9 Sample selection criteria and CGs properties

The different kinematical, morphological and environmental behaviour displayed by Ts and Ms allows us to relate commonly adopted CG selection criteria to the sample's properties.

Specifically our analysis has shown that multiplicity, velocity dispersion, large scale galaxy density and spectral/morphological mix are strongly linked together. Even if the explicit CG selection criterion constrains only one of these parameters, the link between them causes the remaining selection parameters to be constrained too. A constraint on compactness, such as the one we have adopted, biases a sample towards low multiplicity structures and thus indirectly towards low velocity dispersion, spiral rich, isolated structures. A small limit on the maximum velocity dispersion of CG members acts in the same direction. Therefore a parallel requirement for isolation though non affecting the low multiplicity CGs will severely reduce the number of high multiplicity CGs. Conversely, requiring a minimum of four members will bias a CG sample towards intrinsically embedded, spiral-poor groups.

Our analysis shows that Ms, whose parameters are statistically more reliable, happen to be more likely to constitute embedded subclumps whilst Ts might be more likely to be contaminated by systems which are just unbound projections of field galaxies. Interestingly, disregarding CGs in which faint members are counted (sample I), extremely compact systems with a minimum of 4 members which also fulfill the isolation criterion appear to be extremely rare indeed, thereby fitting predictions by numerical simulations claiming that compact configurations are rapidly destroyed (Mamon 1987; Barnes 1989).

Provided the fraction of non-physical CGs does not dominate the statistics, a large scatter in CG properties results when the analysis does not distinguish between Ts and Ms, as multiplicity appears to be a preliminary, robust discriminant between less evolved, field-systems and more evolved, embedded systems. Concerning HCGs, the suggestion that high and low $\sigma _{v}$ groups are intrinsically distinct can already be found in Mamon (2000), who further states that low $\sigma _{v}$ groups are either chance alignments or systems in their final stages of coalescence. The new point we add here is that low $\sigma _{v}$ systems are typically low multiplicity field structures.

That HCGs constitute a heterogeneous sample has previously also been stressed by de Carvalho et al. (1997) and Ribeiro et al. (1998), who, based on the analysis of 17 HCGs, identify 3 distinct CG families. They suggest that these correspond to 3 different dynamical stages, specifically they interpret embedded CGs as precursors of isolated and very dense systems. In comparison with Ribeiro et al. (1998) and based on our much larger CG sample we interpret low velocity dispersion, high overdense CGs (mostly Ts occurring along low density filaments) as the bottom level of the clustering process and embedded structures (either chance projections or collapsing cores within loose groups/poor clusters) such as systems in a more advanced evolutionary stage. Our interpretation, which explains the weak X-ray emission of field CGs in terms of their shallow potential wells (Heldson & Ponman 2000), requires that when X-ray emission is observed in small, gas rich CGs, it should be totally ascribable either to individual galaxies or to collisional shock-heating of the gas in low luminosity systems.

Our analysis indicates that interactions should be efficient mainly in the most overdense, low velocity dispersion structures, which are mainly Ts that include high fractions of gas-rich galaxies. Accordingly, it is not surprising that statistical analysis looking for interaction in HCGs (which include many n>3 CGs embedded within a common halo) globally reveals low fractions of merging remnants and blue Ellipticals (Zepf 1993). Actually, the suggestion that disturbances should be enhanced only among Ts better fits observations reporting that the most easily detected disturbed galaxies are spirals in small groups (Fried 1988) and that the most spectacular mergers, such as bright IRAS galaxies (ULIRGs), appear to involve strong interactions of gas-rich galaxies where the pairs are either isolated or part of small groups (Sanders & Mirabel 1996). It is also worth pointing out that the request for a minimum of 4 members which has biased the HCGs towards intrinsically embedded, gas-poor member groups, possibly explains why, despite the high expected interaction rate, HCGs as a whole present rather low evidence for strong AGN-starbursting episodes (Coziol et al. 1998; Kelm et al. 1998; Coziol et al. 2000).

Whether kinematical differences between Ts and Ms are generally compatible with hierarchical model predictions depends upon the specific assumptions one makes on the $(M/L)_{\rm gal}$ of Ts and Ms member galaxies and on the fractional group mass ( $f_{\rm gal}$) associated with its galaxies. Provided the luminosity of CG members is independent of multiplicity, and assuming $(M/L)_{\rm gal}$ and $f_{\rm gal}$ is the same for Ts and Ms, one predicts $r_{\rm ave}\propto N^{1/3}$ and $\sigma_{v}\propto N^{1/3}$. While the $r_{\rm ave}$ slope is roughly consistent with these expectations, the $\sigma _{v}$ slope increases much faster. Indeed, Fig. 8 suggests that the assumption concerning the same $(M/L)_{\rm gal}$ for Ts and Ms is probably not satisfied, as absorption and emission galaxies are expected to represent ellipticals and spirals, and the former are typically associated with higher (M/L) galaxies than the latter. While $(M/L)_{\rm gal}$ is expected to increase with multiplicity, $f_{\rm gal}$ might actually decrease, due to the fact that higher multiplicity CGs are more likely to be associated with gas-rich, X-emitting groups. Consequently, before assessing whether globally data on CGs are (or are not) compatible with hierarchical model predictions, more accurate models, taking into account the different $(M/L)_{\rm gal}$ and $f_{\rm gal}$ of Ts and Ms, should be investigated.