|Annu. Rev. Astron. Astrophys. 1991. 29:
Copyright © 1991 by . All rights reserved
The wide-field starcount programs provided new data on the large-scale spatial structures of GCSs, and were complemented later by CCD imaging with its ability to reveal the inner GCS structures within the high background light of the nuclear regions. The original paradigm that globular clusters traced out much the same halo structure as any other old Population II component was broken by the discovery, in the Virgo giants M87 and M49, that the GCSs are globally less centrally concentrated than the halo light. On this basis, Harris & Racine (108) and Racine (171) first suggested that the GCS might be older than the bulk of the surrounding halo.
Not all galaxies, however, share this phenomenon. In the Milky Way itself, the GCS seems to match the halo star distribution fairly well (97), and to first order the same agreement is found within other galaxies of both disk and elliptical types (82, 87, 105, 106, 111). Interestingly, two of these are the cD systems NGC 3311 (110) and NGC 1399 (105), whose GCSs are structurally like those in M87 and M49 but whose halos are much more distended than average. No case has been found in which the GCS is more centrally concentrated than the halo.
If the projected GCS surface density is approximated by a power-law curve cl ~ Rgc-, in giant ellipticals is typically 1.5-2.0 for Rgc 5 kpc (see 94 for a comprehensive discussion). Figure 6 shows two examples. For M87, suggestions have often been made that its large cluster population and extended spatial distribution might have arisen from extra clusters that were somehow added to its outer halo (see, e.g., 56, 63, 221). However, any such views are incorrect. The M87 system is virtually identical in structure with the prototypically normal M49 system at all radii; that is, M87 contains three times more clusters than normal at all locations, both inside and out. The high-SN anomaly is thus a global characteristic, not easy to explain by arguments special to one radial region (such as the high mass in the extended X-ray halo of M87, or cluster formation in cooling flows).
Figure 6. Spatial structure of GCSs in two giant Virgo ellipticals, M87 and M49. The number of globular clusters per unit area projected on the sky is plotted against radial distance from the center of the galaxy. The filled circles are from wide-field starcount data extending out to r ~ 20' (see 94); the starred symbols are central projected cluster densities from CCD imaging programs (76, 103, 128). Because the two sets of data have been shifted vertically for clarity, the vertical scale is arbitrary. For Virgo, 1 arcmin 5 kpc, so the profiles cover the projected radial distance range from 2 to 100 kpc. Even though M87 has a far larger cluster population, the two GCSs are structurally quite similar at all radii.
Another quite general phenomenon in large galaxies seems to be that cl(r) flattens off to a near-constant level within Rgc 2 kpc. Observations right in to the nucleus of M87 (76, 114, 128) show the effect, and similar conclusions have been reached for several other systems including the Milky Way (82, 103, 106, 172). Dynamical destructive processes (Section 7.1), which are enormously more effective near the nucleus, may have virtually removed whatever clusters were originally in these central regions. In principal, these same processes might also generate radial differences in the GCLF which should be detectable. Observationally, little is yet known about any such trends, though again in M87 (76, 128) and the Milky Way (5), no obvious variation of cluster luminosities with radius is seen.
In large disk galaxies, the GCS appears to be nearly spherical in shape and, perhaps, rounder than the spheroid isophotes [(82, 106), but see H. Harris et al. (87) for a different comment on NGC 7814]. For giant ellipticals, the GCS contours agree well with the ellipticity and orientation of the galaxy isophotes (38, 109, 117). The smaller E6 system NGC 3377 has a GCS which is even more elongated than the galaxy, perhaps as a residual effect of dynamical evolution (106).
GCS kinematics and dynamics can be explored through direct radial velocity measurement of the clusters. Until recently such data were available only for the Milky Way (e.g. 205) and M31 (122, and references cited). However, velocities of sufficient precision (± 50 km s-1) have now been obtained for some dozens of clusters in NGC 5128 (88, 116, 191), M87 (119, 151, 152), and M49 (151). These studies show that the velocity dispersions are nearly constant with radius and confirm the existence of dark-halo mass extending outward with galactocentric distance as M(R) ~ Rgc to several tens of kiloparsecs. In M87, the GCS velocity dispersion profile thus forms a smooth dynamical connection between the central galaxy light and the still more extended X-ray halo. Mould et al. (151) also find that the dispersion (vr) is clearly higher for the GCS than for the underlying halo light at the same location, and they argue that this is probably a consequence of the larger spatial scale of the GCS. Suitably anisotropic velocity distributions are an alternate possibility.