Abstract

We present a comprehensive study of the luminosity function and spatial structure of the globular cluster system of NGC 4472, the brightest galaxy in Virgo, based on deep wide-field Washington CT, CCD images. The globular cluster luminosity function shows a peak at T-1 = 23.3 +/- 0.1 mag, about 1.5 mag brighter than our 50% completeness limit. Comparing this value with that of the Galactic globular clusters, we estimate the true distance modulus to NGC 4472 to be (m - M)(0) = 31.2 +/- 0.2 (corresponding to a distance of 17.4 +/- 1.6 Mpc). With our large sample (approximate to 2000) of bright globular clusters over a wide field, we make a definitive investigation of the spatial structures of the metal-poor and metal-rich cluster populations and find that they are systematically different: (1) the metal-rich clusters are more centrally concentrated than the metal-poor clusters; and (2) the metal-rich clusters are elongated roughly along the major axis of the parent galaxy, while the metal-poor clusters are essentially spherically distributed. In general, the metal-rich clusters closely follow the underlying halo starlight of NGC 4472 in terms of spatial structure and metallicity, while the metal-poor clusters do not. The global value of the specific frequency of the globular clusters in NGC 4472 is estimated to be S-N = 4.7 +/- 0.6. The local specific frequency increases linearly outward from the center of NGC 4472 until similar to 5'.5, beyond which it levels off at S-N similar to 8.5 until the limit of our data at 7'. The specific frequency of both the metal-rich and metal-poor populations shows similar behavior. However, S-N of the metal-poor clusters is about a factor of 2 greater than that of the metal-rich clusters in the outer regions. Implications of these results for the origin of the globular clusters in NGC 4472 are discussed. These results are consistent with many of the predictions of both the model of episodic in situ formation plus tidal stripping of globular clusters given by Forbes et al. and the Ashman Zepf merger formation model, but each of the models also has some problems.