6.3 The Spatial Distribution of Metal-Poor Galaxies
Normal giant galaxies are found both in clusters and low density environments in the field. In particular giant ellipticals are strongly clustered and are predominantly found near the centres of rich clusters, while spiral galaxies are less clustered. The difference in the galaxy population in clusters and the field is an important observable, but not easy to interpret in view of the various effects that may affect galaxies in different environments: merging, interactions, harassment, dynamical disruption, stripping, cooling flows, pressure confinement by hot gas etc.
Dwarf elliptical galaxies are found predominantly in clusters or as companions of giant field galaxies (Binggeli et al. 1990) and there appears to be a lack of isolated field dEs. The last point may in part be due to selection effects, since low luminosity dEs have low surface brightness and may therefore have been missed by most local surveys. New dE members of the Local Group are still being discovered (Armandroff et al. 1998, Karachentsev and Karachentseva 1999) illustrating the point that our view of even local dEs are largely incomplete. Irregular dwarfs follow the same structures as those outlined by massive galaxies (Thuan et al. 1987; Comte et al. 1994).
The LSBGs are less clustered than ``normal'' (giant) galaxies (e.g. Bothun et al. 1986a, 1993; Mo et al. 1994), in the sense that they tend to avoid clusters, and are not found close to field galaxies. However, this is based on comparison with galaxy catalogues that are badly incomplete for dwarf galaxies, and especially for LSBGs and dEs. Therefore not much can be said from these studies about how LSBGs cluster with other LSBGs and faint dwarfs. Large volume limited samples of LSBGs are required for this purpose. Taylor (1997) finds from a VLA study of their environments, that about one quarter of the LSBGs appear to have H I companions, with a detection limit of ~ 107 M. These H I companions tend to have faint optical counterparts (Taylor, private communication), and are thus likely to be LSBGs themselves. This suggests that LSBGs need not to be extremely isolated. It is however clear that LSBGs avoid massive giant galaxies, which is understandable since LSBGs close to giant galaxies have a high probability to interact which would lead to increase their star formation rate, thus transforming the LSBG into a high surface brightness galaxy.
Iovino et al. (1985) presented results suggesting that H II galaxies are less clustered than normal giants. Salzer (1989) investigated the spatial distribution of ELGs in the UM catalogue, finding that the ELGs follow in most parts the structures outlined by bright normal galaxies, but tend to avoid the regions with the highest galaxy density. On the other hand Comte et al. (1994) showed that KISO ultraviolet excess galaxies were distributed in a similar manner to ``normal'' galaxies. The spatial distribution of the SBS sample was investigated by Pustilnik et al. (1995) who found similar results, but in addition a significant fraction (~ 20 %) was found in voids. Most H II galaxies/BCGs seem to be rather isolated when compared to existing galaxy catalogues and redshift surveys (Campos-Aguilar et al. 1993, Telles and Terlevich 1995). But these results mainly show that BCGs avoid giant galaxies, since these constitute the catalogues used for comparison. No constraint can be imposed on the clustering with other dwarf galaxies, e.g. LSBGs. The latter have indeed been found to be common companions to H II galaxies from the studies by Taylor and collaborators (cf. Taylor 1997). A recent study (Telles and Maddox 1999) attempts to address also the dwarf-dwarf clustering by comparing BCGs with APM (automatic plate measuring machine) catalogues, which contains more low luminosity galaxies, and comes to similar conclusions as previous studies in the sense that BCGs mainly populate environments less dense than normal galaxies. However, even the APM is badly incomplete for LSBGs, which are the likely companions in view of Taylor' results.
Given the apparent tendency of BCGs to avoid dense environments, such as rich clusters, it has been speculated whether the voids may be filled with BCGs and other faint dwarfs such as LSBGs. If so the view of the baryonic matter distribution would be very biased, and a large mass fraction would have been missed. This would be in agreement with ``biased'' galaxy formation theories (e.g. Dekel and Silk 1986) where dwarfs arise from low density peaks in the primordial density fluctuation spectrum. There have been some studies addressing this question: Pustilnik et al. (1995) found that 20% of BCGs may reside in voids. Popescu et al. (1997) finds some void galaxies, but show that the voids are not filled by an undiscovered population of BCGs. Similar results are reached by Lindner et al. (1998). BCGs in or near voids are predominately found near the borders (Lindner et al. 1996). The cases for other types of dwarfs, e.g. LSBGs, are less clear, but there is presently nothing that points toward a large density of any galaxy type in void, although this should be further investigated.
Dwarf irregular galaxies seem to be abundant in most environments, both in rich clusters and as pure field galaxies (Binggeli et al. 1980). However there are few studies addressing the dIs directly, and we have already seen that this class generally encompasses many different kinds of low mass irregular galaxies, including BCGs and LSBGs.
In conclusion, metal-poor BCDs and LSBGs avoid rich cluster environments, but may have neighbours of the LSBG type. Dwarf ellipticals are found in clusters or nearby luminous galaxies, while dIs are found in most environments (Binggeli et al. 1990).