Annu. Rev. Astron. Astrophys. 1991. 29: 239-274
Copyright © 1991 by . All rights reserved

Next Contents Previous

3.5.1 CENTRAL BLACK HOLES Most models of the energy production in QSO's invoke the presence of supermassive nuclear black holes. Predicted masses are in the range of 108 to 109 Msun, depending on efficiencies and the average lifetime of the QSO phenomenon (284). If the lifetimes are short, many galaxies may have supermassive black holes.

High spatial resolution observations of nearby galaxies have indicated M / L values that increase sharply in the nuclear regions. These are usually interpreted with supermassive black holes, but a cluster of dark remnants cannot be excluded for most cases (291). Nearby well-studied systems are M31 (98, 191, 291) and M32 (98, 291, 330). The inferred black hole masses are 106 - 107 Msun and 107 - 108 Msun, respectively. A similar increase in M / L was found for NGC 4594 (176, 192), and corresponds to a black hole mass of around 109 Msun. Surprisingly, the case is weakest for M87, a known active galaxy. The early claims of a detection of a central black hole (301, 377) were based mostly on surface photometry and velocity dispersions. It was soon realized that anisotropies in the velocity dispersion, and a non-stellar contribution to the light, can explain the observations with a constant M / L model (45, 101, 261). More recent data can be modeled with only a modest anisotropy in the velocity distribution (73, 99). The requirement that these models are stable provides an extra constraint (238), but the stability of the most recent models has not been studied.

The evidence for nuclear black holes in M31, M32 and NGC 4594 rests on curves of velocity dispersions and rotation. Modeling this type of data is not straightforward, as it involves the coupled deconvolution of the surface brightness, radial velocity and velocity dispersion profiles. This makes it difficult to search the complete solution space. The models employed are usually assumed to be spherical or axisymmetric (46), but to date no axisymmetric models based on explicit distribution functions f geq 0 have been constructed. Additional information is contained in the line profiles (71, 233, 290), which can now be derived from digital spectra. Application of this powerful technique may well put much tighter constraints on the behavior of M / L.

Gerhard has argued that the rapid central rotation in M31, M32 and NGC 4594 can be produced by a small bar which is viewed down its major axis (138). This hypothesis can be tested by observing a larger sample of galaxies, for which one would expect random viewing angles.

Next Contents Previous