For more than two decades, the class of cD-type galaxies that reside at the centers of rich galaxy clusters - the ``brightest cluster members'' or BCGs - have attracted considerable attention because of their exceptionally rich globular cluster populations. M87 in Virgo is the prototype of this class, and the GCSs in more than 30 such galaxies have now been studied (see Blakeslee et al. 1997; Harris et al. 1998 for summary lists). These galaxies are of special interest because they provide uniquely strong - and challenging - tests for our various galaxy formation models.
In M87, the basic shape of the MDF and the spatial distribution of the GCS are remarkably similar to what we see in NGC 4472. The special issue we are confronted with in M87 and the other BCGs is the classic ``high specific frequency'' problem: stated quite simply, why do they have so many globular clusters? In the 20 or more years we have been aware of the problem, no obvious answer has emerged. In any of the competing scenarios (in situ, mergers, accretions), appeals have been made to rather arbitrary conditions (such as an outstandingly high cluster formation efficiency in high density protogalactic gas; or especially violent compression shocks during mergers; or accretion of dwarf E galaxies which themselves had high SN). All of these alternatives have tended to be uncompelling.