4.2. MDFs and Specific Frequencies: How Many Components?
Integrated photometry of large samples of clusters in giant E galaxies like the Virgo members has made huge strides in the past few years. The especially detailed photometric study of NGC 4472 by Geisler et al. (1996) and Lee et al. (1998) is an excellent example of such work, and represents a major step in our understanding of the MDFs and their implications. Among the most important of their results, from a comprehensive and accurate metallicity-sensitive (C - T1) photometric survey of the GCS, is that the two components (MPC, MRC) have very different central concentrations (with the MRC closely following the spatial structure of the integrated halo light and the MPC following a much shallower distribution). Each of these components by itself displays little radial gradient in color. Perhaps most significant of all is the observation that the mean color of the MRC clusters as a function of radius precisely matches the color of the halo light - to well within the observational uncertainties of both. By inference, just as we found earlier for NGC 5128, the halo stellar population appears to be dominated by a metal-rich population whose MDF is very similar to the metal-rich globular cluster component. A reasonable conclusion is that the MRC clusters and the majority of the halo stars formed together.
Let us take this argument one step further (see Forbes et al. 1997 for similar reasoning). Suppose we assume that the halo has just two distinct components, whose mean metallicities are identical with those of the MRC and MPC clusters. Close inspection of the actual color indices and their internal uncertainties (see H99) then shows that 6% of the halo stars can belong to the MPC, otherwise the integrated color of the entire halo would deviate from that of the MRC level beyond their error bars. Just as we did above for NGC 5128, we can translate this ratio into separate specific frequency estimates for the MPC and MRC, knowing that SN = 5.6 for the whole galaxy (Harris et al. 1998). We obtain SN(MRC) 2.4 and a lower limit for the MPC of SN 50! The specific frequency for the metal-rich component is similar to the average of many other normal ellipticals (see above), but the lower limit for the metal-poor component is already higher than the measured SN for any known galaxy. If we interpret this result within a conventional in situ formation picture, it means that either the cluster formation in the first, metal-poor burst was outstandingly efficient, or that a very large fraction of the initial gas went unused, just as we speculated earlier for NGC 5128. The data for NGC 4472 suggest, apparently, a still more extreme difference between the relative amounts of metal-rich and metal-poor stellar components.
Very similar results emerge from a new study of the GCS in NGC 1399, the central giant galaxy in the Fornax cluster (Ostrov et al. 1998). Once again, a clear bimodal MDF emerges, in which the metal-richer clusters match the mean color of the galaxy halo light extremely well. The specific frequency of the MRC alone they estimate to be SN 3, within just the same range that we found for NGC 4472.