Although dwarf galaxies are the most common type of galaxy thus far discovered and may be substantial contributers of mass to the universe, very little is known about them. With masses 10 to 1000 times smaller than 'typical' galaxies and often of low surface brightness, these galaxies long remained undiscovered due to selection effects. These effects selected against both small pointlike and large diffuse galaxies favoring instead extended galaxies with a surface brightness well above the sky (which has a typical surface brightness of 21.5 V mag / square degree). To survey these galaxies outside the Local Group requires a large detector area and long integration time and it took the advent of the ccd and wide field imaging telescopes for searches to become feasible. A large external dwarf population was first detected in the Virgo cluster (Reaves 1983, Bingelli et al. 1984) and found to contain 900 dwarf ellipticals. Fornax, another nearby cluster was subsequently studied and also found to exhibit a substantial dwarf population (Caldwell, 1987). Recently work has also progressed on a third rich cluster, Coma (Thompson & Gregory 1993), where again, a large population of over 700 dwarfs has been found. Almost everything we know about these significant small galaxies has proceeded from this sample of only 3 rich clusters. Some work has also been done on a handful of other clusters and small groups with the goal of determing the galaxy luminosity function.
Dwarf galaxies may be a key to understanding galaxy structure, formation, and evolution. One theory as to their origin is that they formed from small peaks in the initial power spectrum. If this is the case, one would expect to find them in greater numbers and to be less clustered than bright galaxies. On the other hand, they may have formed from tidal debri from galaxy mergers or from galaxies that underwent RAM pressure stripping as they fell toward the center of clusters and, therefore, more would be expected at cluster centers. It is also important to determine why these galaxies have a low surface brightness. Supernovas may have stripped these low mass galaxies of gas and halted star formation early on. Stripping may also have occurred from environmental influences such as nearby galaxies or by an intracluster medium. Tidal interactions with neighbors, however, may have the effect of triggering star formation. Hence, determining the spatial distribution of the dwarfs within the clusters will have important ramifications on all theories of galaxy evolution.
Considering the importance of these low mass galaxies, it is unfortunate that so little is known about them. Since most of the dwarfs have such a low surface brightness and lack a bright nucleus, a spectroscopic determination of velocity dispersion is often impossible. This also inhibits the confirmation of cluster membership which instead must be done using purely morphological means and is therefore subject to great uncertainty. Metallicities and ages of stellar populations are also poorly known. Much can be learned, however, from studying the numbers and locations of the dwarfs within clusters of differing environments.
We are currently completing a sample of all nearby (z < .02) rich clusters. We have so far obtained deep (reaching limiting surface brightnesses of 26.0 mag/square degree) CCD V-band data for 9 clusters (A1367, 2634, 426, 569, 262, 1656, 3537, 3656, and 3526) covering up to 1 square degree.
The goal of this project is 3-fold. First, we will determine the faint-end slope of the galaxy luminosity function in each cluster. A thorough understanding of the faint-end galaxian luminosity function (LF) is important to firmly establish whether dwarf galaxies can plausibly account for much of the inferred dark matter content in clusters. A steep LF (a = -1.8 for the faint-end slope of a Schechter function) would imply that dwarfs - in particular, their associated dark halos - contribute significantly to the overall dark matter fraction. Attempts to obtain luminosty functions in the past were made with photographic plates and only work on three clusters - Virgo, Fornax, and Coma - reached faint limiting magnitudes down to Mr ~ -13 (Sandage et al., 1985, AJ, 90, 1759; Ferguson et al. 1988, AJ, 96, 1520; Thompson et al. 1993, AJ, 106, 2197). Deeper CCD imaging has recently been used to establish faint-end slopes of the LF in several clusters (Bernstein et al. 1995, AJ, 110, 1507; De Propris et al. 1995, ApJ, 450, 534). However, both studies were restricted to cluster cores, whereas we plan to attain faint surface-brightness limits over the central square degree of the clusters. Results from these earlier studies found very different values for the faint-end LF slope, perhaps due to different treatments of the selection effects inherent in detecting these low surface brightness galaxies. We have carefully and systematically addressed incompleteness and selection effects in our cluster observations through false galaxy analysis. Our final completeness corrected luminosity functions for 5 clusters are fit with Schechter functions. We find similar slopes of -1.40 for A1367, A262, and A3526 while steeper slopes are found for A3537 and A3656.
Second, because we are covering large fractions of the total cluster area, these data will also be extremely useful for studying the effect of environment on dwarfs. This will be accomplished by examining dwarf morphology as a function of location, dwarf clustering properties as a function of luminosity and type, and the global dwarf-to-galaxy ratios. Previous work has suggested that the dwarf/giant ratio increases with increasing cluster richness and that dwarfs may be more clustered than bright galaxies (Ferguson & Sandage, 1991, AJ, 101, 765; Feguson & Bingelli 1994, A&ARv, 6, 67). Three of our clusters (A3526, A2634, and A426) do not appear to follow these simple trends. Instead, the dwarfs appear less clustered, and the dwarf/giant ratio does not vary significantly with richness class. We also find in all three clusters that the dwarf/giant ratio increases with distance from the cluster center before falling at distances > 30arcminutes. This suggests that environmental influences may suppress dwarf formation, affect dwarf survival, or induce fading via tidal stripping (or harassment) in the cluster core. In contrast, our work on a fourth cluster, A1367, finds only a monotonic decrease outward from the center. It is apparent that dwarf galaxy populations do not appear to follow simple trends and are heavily affected by differing cluster environments. Because of this, dwarfs are perfect candidates in probing environmental effects, but we require a larger sample of clusters and dwarfs to isolate specific physical effects that might govern dwarf evolution.
Finally, our comprehensive survey of dwarfs in rich nearby clusters of differing environments is also useful to determine the cluster-to-cluster variations (in dwarf morphology, numbers, and eventually color) expected at a single epoch. Such data are required as a control when probing to greater redshifts where color and abundance changes provide important clues to dwarf formation and evolution. Our data are also a prerequisite for any future fiber redshift surveys to study dwarf kinematics.