7 Large scale environment of Compact Groups

Many embedded CGs are expected to be chance alignments of galaxies, not directly bound to one another along the line of sight, that form and destroy continuously within loose groups, whilst isolated CGs are generally assumed to be close dynamical systems, whose future evolution is a function of internal parameters only. Unfortunately, defining a CG as isolated is a non trivial problem, as one has to define boundaries (in space and luminosity) below which external galaxies perturb CG evolution and above which perturbations are negligible. Previous studies of CG environments yield contradictory results. Rubin et al. (1991) studying 21 HCGs find Ts to be more isolated than Ms. On the other hand Barton et al. (1996) do not confirm this result. However hardly any isolated CG should be retrieved, as bright galaxies are known to be strongly clustered, and faint galaxies are known to cluster around bright ones (Benoist et al. 1996; Cappi et al. 1998).

In order to properly investigate possible relations between small and large scale environments, the algorithm counts neighbours ( $N_{\rm env}$) for each CG within a distance $\Delta R=1~h^{-1}$ Mpc and $\Delta v^{\rm II}=\pm1000$ km s$^{\rm -1}$ from the CG center. To minimize distance uncertainties due to the relative incidence of peculiar motions neighbourhood richness is computed only for CGs at cz>1500 km s$^{\rm -1}$, thereby reducing the number of Ts and Ms in subsample I from 34 to 31 and from 21 to 17 respectively. In Fig. 10 the overall distribution of CGs with respect to $N_{\rm env}$ is shown. The solid line refers to Ts, the hatched area to Ms. It clearly emerges that Ts are more likely than Ms to be found in isolated environments, but that, compared to the much more numerous single galaxies (hatched line), their environment is denser. According to the KS test differences between Ts and Ms are significant at 98%, 97% and 94% c.l. in subsamples II, III and IV, whilst they are non significant in class I. In simulated samples Ms show no excess of neighbours with respect to Ts, so that no corrections for selection effects have to be applied to the environmental data. Thus we find three independent parameters (velocity dispersion, spectral properties and environmental density) suggesting that Ts and Ms constitute different populations. The sparser environment, the higher emission line fraction and the lower velocity dispersion of Ts all might result from high contamination by field interlopers. However they are also compatible with Ts being recently formed systems of field galaxies, not yet embedded within a common virialized halo, in which dynamical friction efficiently transfers orbital energy of the group into the internal energy of a single merger remnant.

Figure 10: CGs distribution as a function of large scale neighbours. Neighbours have been counted out to R=1 h-1 Mpc and $\Delta cz=\pm 1000$ from the CG center. The solid histogram refers to Ts, the hatched one to Ms. The dashed line shows the large scale environment distribution of single galaxies (on a scale that fits into the limits set by CGs).

It must also be stressed that although the probability of chance alignments decreases rapidly when going from Ts to Ms, the richer environment associated with Ms enhances the probability of Ms being chance alignments. Indeed, simulations indicate (Mamon, private communication) that a Multiplet with 5 neighbours is twice as likely to be a chance alignment than a Triplet with less than 2 neighbours. Finally, we underline that the relation between CGs multiplicity and large scale galaxy density indicates that when isolation is used as a CG selection parameter the sample is biased towards either luminous or low multiplicity CGs, the former including many early type galaxies, the latter a high fraction of late type galaxies. The requirement of isolation consequently induces large scatter in the spectral (and morphological) properties of CG samples.

Figure 11: Surface density contrast of Ts (empty triangles) and Ms (filled squares) versus velocity dispersion $\sigma _{v}$. Stars represent ACO $_{\rm CGs}$. The diagram clearly shows that CGs with different environmental and dynamical properties can be separated according to multiplicity.