The Virgo cluster is a typical cluster in every sense, and we would not expect it to be otherwise. Not even the fact that it is made of (at least) three smaller clusters (one of which is grossly dominating in mass, however), which are on the verge to merge, turns out to be very special. Many other clusters, at a closer look, show almost exactly the same features. A good example is the somewhat more distant Centaurus cluster, where we distinguish between a massive, X-ray emitting and dE-rich subclump (centered on the cD NGC 4696) and at least one less massive, spiral-rich unit, which is falling into the dominant structure with a high relative velocity (Jerjen 1995, Stein et al. 1997). Even the Coma cluster, formerly the prototype of a hypthetical class of ``regular'', ``relaxed'' clusters, has given way to this picture: it also is an aggregate of three subunits, two of which must have merged long ago, but still show traces (as the respective central dominant galaxies, NGC 4874 and NGC 4489, are not yet merged), and the third of which (centered on NGC 4839, of Virgo cluster size and mass!) is in the process of merging (White et al. 1993, Colless & Dunn 1996). We seem to live in the epoch of rich cluster formation.
What is special about the Virgo cluster, for us, is of course its proximity. There is no other cluster of comparative richness lying that close (Fornax, Ursa Major, and Coma I are all much less rich than Virgo). That the cluster is harbouring an active galaxy is not unusual (in fact, as Blandford emphasizes in this volume, M87 is a fairly lousy, i.e. inactive AGN). But again: it is the proximity which may render M87 a kind of Rosetta stone for AGN astrophysics. Indeed, enormous efforts are spent in the attempt to unveil the secrets of the center of M87, of which this volume bears ample witness. The efforts spent to investigate the extragalactic environment of M87 are very small in comparison. However, as I tried to show here, the central pc of M87 may be intimately connected with the structure and dynamics of the Virgo cluster as a whole.
Of the many features of the Virgo cluster yet to be studied, I mention two which I regard as especially relevant for the present discussion. (1) As mentioned before, we still lack radial velocity data for ca. 800 Virgo members. Among them are the hundreds of dEs which form the cocoon around M87 / M86. It would be highly desirable to know the velocities of as many of these dwarfs as possible, to get a more complete picture of the cluster kinematics. Present-day technology should allow the measurement of at least the brighter of these objects. (2) Based on such data, with the projected positions of all, and the radial velocities of nearly all Virgo galaxies, plus (possibly) the X-ray gas distribution taken as model input (initial conditions), one could ``run'' the whole Virgo cluster. In particular, one should be able to simulate the M87 / M86 subclump interaction, or at least put useful constraints on its dynamics. For a full 3D simulation, the computing time might be prohibitively large. However, even with very simplifying assumptions to save computing time, such a simulation might be very rewarding. A cluster simulation with realistic, i.e. observed quantities as input parameters (not only in the statistical sense, but galaxy-by-galaxy) has not yet been carried out. The Virgo cluster would be the obvious first choice for such a project (for a first, crude attempt, see Schindler & Binggeli 1994).
Acknowledgements: My involvement with the Virgo cluster was initiated by an illuminating and pleasant collaboration with Dr. Allan Sandage and Prof. G. A. Tammann, to whom I'm very grateful. I thank Dr. Hans Böhringer for the nice X-ray image of the Virgo cluster (Fig. 5). This work was supported by the Swiss National Science Foundation.