The discovery of ultra compact dwarf (UCD) galaxies has been very well reviewed by Michael Hilker in these proceedings. Here we wish to focus on the large-scale distribution of UCDs. Our study was originally motivated by a search for "M32-type" compact elliptical galaxies in the Fornax Cluster. We didn't find any of these, but we did demonstrate that morphological membership classification can overlook many compact dwarf cluster members [3]. This motivated the 2dF Fornax Cluster Spectroscopic Survey (FCSS, [4]), which measured redshifts of all objects, both "stars" and "galaxies" in the cluster centre, resulting in the detection of 5 UCDs ([5] and also [9]). High resolution imaging and spectroscopy have since established that the most luminous UCDs are a new type of object [6]. We have extended our UCD searches to fainter limits [71, finding a total of 62 compact objects filling a large region of intra-cluster space (see Fig. 1). We discuss the distribution of these objects below, noting that our working UCD definition (Fornax Cluster members barely resolved in photographic survey data with m(Bj) < 21.5 mag) may also include genuine globular clusters (GCS).

This paper investigates the possibility that ultra-compact dwarf (UCD) galaxies in the Fornax cluster are formed by the threshing of nucleated, early-type dwarf galaxies (hereafter dwarf galaxies).Similar to the results of Cote et al. for the Virgo cluster, we show that the Fornax cluster observations are consistent with a single population in which all dwarfs are nucleated, with a ratio of nuclear to total magnitude that varies slowly with magnitude. Importantly, the magnitude distribution of the UCD population is similar to that of the dwarf nuclei in the Fornax cluster.The joint population of UCDs and the dwarfs from which they may originate is modelled and shown to be consistent with a Navarro, Frenk & White (NFW) profile with a characteristic radius of 5 kpc. Furthermore, a steady-state dynamical model reproduces the known mass profile of Fornax. However, there are a number of peculiarities in the velocity dispersion data that remain unexplained.The simplest possible threshing model is tested, in which dwarf galaxies move on orbits in a static cluster potential and are threshed if they pass within a radius at which the tidal force from the cluster exceeds the internal gravity at the core of their dark matter halo. This fails to reproduce the observed fraction of UCDs at radii greater than 30 kpc from the core of Fornax.

Aims. We determine masses and mass-to-light ratios of five ultra-compact dwarf galaxies (UCDs) and one dwarf elliptical nucleus in the Fornax cluster from high resolution spectroscopy. We examine whether they are consistent with pure stellar populations or whether dark matter is needed to explain their masses. Methods. Velocity dispersions were derived from selected wavelength regions using a direct-fitting method. To estimate the masses of the UCDs a new modelling program has been developed that allows a choice of different representations of the surface brightness profile (i.e. Nuker, Sersic or King laws) and corrects the observed velocity dispersions for observational parameters (i.e. seeing, slit size). The derived dynamical masses are compared to those expected from stellar population models. Results. The observed velocity dispersions range between 22 and 30 km s(-1). The resulting masses are between 1.8 and 9.5 x 10(7) M-circle dot. These, as well as the central and global projected velocity dispersions, were derived from the generalized King model which turned out to give the most stable results. The masses of two UCDs, that are best fitted by a two-component profile, were derived from a combined King+Sersic model. The mass-to-light ratios of the Fornax UCDs range between 3 and 5 (M/L-V)(circle dot). The M/L-V ratio of the dwarf elliptical nucleus is 2.5. These values are compatible with predictions from stellar population models. Within 1-2 half-mass radii dark matter is not dominating UCDs and the nucleus. An increasing dark matter contribution towards larger radii can not be ruled out with the present data. The M/L-V ratios of some UCDs suggest they have intermediate age stellar populations. Conclusions. We show that the mass-to-light ratios of UCDs in Fornax are consistent with those expected for pure stellar populations. Thus UCDs seem to be the result of cluster formation processes within galaxies rather than being compact dark matter dominated substructures themselves. Whether UCDs gained their mass in super-star cluster complexes of mergers or in nuclear star cluster formation processes remains an open question. It appears, however, clear that star clusters more massive than about 5 x 10(6) M-circle dot exhibit a more complex formation history than the less massive "ordinary" globular clusters.

Using imaging from the Hubble Space Telescope, we derive surface brightness profiles for ultracompact dwarfs in the Fornax Cluster and for the nuclei of dwarf elliptical galaxies in the Virgo Cluster. Ultracompact dwarfs are more extended and have higher surface brightnesses than typical dwarf nuclei, while the luminosities, colors, and sizes of the nuclei are closer to those of Galactic globular clusters. This calls into question the production of ultracompact dwarfs via "threshing," whereby the lower surface brightness envelope of a dwarf elliptical galaxy is removed by tidal processes, leaving behind a bare nucleus. Threshing may still be a viable model if the relatively bright Fornax ultracompact dwarfs considered here are descended from dwarf elliptical galaxies whose nuclei are at the upper end of their luminosity and size distributions.