We have used the High Resolution Camera at the Canada-France-Hawaii Telescope to obtain VRI photometry of the globular clusters in the innermost 140 ^ { \prime \prime } of the M87 halo .
The results are used to discuss several issues concerning the formation and evolution of globular cluster systems in supergiant ellipticals like M87 .
Our principal results are as follows : ( 1 ) From our deep R - band photometry of the cluster population , we find no significant change in the globular cluster luminosity function ( GCLF ) with galactocentric radius , for cluster masses M { } ^ { > } _ { \sim } 10 ^ { 5 } M _ { \odot } .
This result places constraints on current theoretical predictions of the rate of cluster evolution by tidal shocking and evaporation , indicating that the main effects of dynamical evolution may be only on lower-mass clusters ( { } ^ { < } _ { \sim } 10 ^ { 5 } M _ { \odot } ) that are below the faint limit of most current observations .
( 2 ) Combining our ( V - I ) color indices with other data in the literature , we derive the metallicity gradient and mean metallicity of the clusters from r = 9 ^ { \prime \prime } out to r \simeq 500 ^ { \prime \prime } ( \sim 0.7 to 35 kpc ) .
Within the core radius r _ { c } = 1 ^ { \prime } of the globular cluster system , the metallicity distribution is uniform , but at larger radii the mean metallicity declines steadily as Z / Z _ { \odot } \sim r ^ { -0.9 } .
( 3 ) The various options for explaining the existence of high specific frequency galaxies like M87 are evaluated .
We argue that alternatives involving secondary evolution ( such as the formation of many globular clusters during mergers , or the existence of a population of intergalactic globular clusters centered on the same location as the central elliptical ) are capable of modifying the characteristics of the GCS in distinctive ways , but are not likely to be the primary explanation for high - S _ { N } galaxies .
( 4 ) We offer a new explanation for the large observed S _ { N } range among brightest cluster ellipticals ( BCGs ) .
It is suggested that these central supergiant E galaxies formed in an exceptionally turbulent or high-density environment which favored a very rapid initial star formation burst .
As a result , a much higher than average fraction of the protogalactic gas was driven out in a galactic wind not long after the first globular clusters were formed , thus artificially boosting the specific frequency of the remaining galaxy .
From a total sample of 30 BCGs , we derive empirical scaling relations which relate to this hypothesis .
Our analysis favors the view that BCGs began forming at redshifts z { } ^ { > } _ { \sim } 5 , distinctly earlier than most other galaxies .