We use a hybrid N-body program to study the evolution of massive black hole binaries in the centers of galaxies, mainly to understand the factors affecting the binary eccentricity, the response of the galaxy to the binary merger, and the effect of loss-cone depletion on the merger time. The scattering experiments from paper I showed that the merger time is not sensitive to the eccentricity growth unless a binary forms with at least a moderate eccentricity. We find here that the eccentricity can become large under some conditions if a binary forms in a galaxy with a flat core or with a radial bias in its velocity distribution, especially if the dynamical friction is enhanced by resonances as suggested by Rauch and Tremaine (1996). But the necessary conditions all seem unlikely, and our prediction from paper I remains unchanged: in most cases the eccentricity will start and remain small. As a binary hardens it ejects stars from the center of a galaxy, which may explain why large elliptical galaxies have weaker density cusps than smaller galaxies. If so, the central velocity distributions in those galaxies should have strong tangential anisotropies. The wandering of a binary from the center of a galaxy mitigates the problems associated with loss-cone depletion and helps the binary merge.

No abstract

We add black holes to nonrotating, spherical galaxy models, with the assumption that the black-hole growth is slow compared with the dynamical time but fast compared with the relaxation time. The outcome di ers depending on whether the core of the initial galaxy does or does not resemble that of an isothermal sphere. For the isothermal case the previously-known results are con rmed and sharpened: the black hole induces cusps in the density ( r 3=2 ) and velocity dispersion (v 2 r 1 ), and a tangential anisotropy in the velocity distribution away from the center. For the non-isothermal case the induced density cusp is steeper, and the induced anisotropy is larger and penetrates right to the center. The cusp around the black hole is insensitive to anisotropy in the initial velocity distribution, and also to the origin of the black hole, unless its mass comes exclusively from the stars of lowest angular momentum, in which case the cusp is suppressed. We discuss the implications for the interpretation of evidence for massive black holes in galactic nuclei.

No abstract

We study the effect of a massive central singularity on the structure of a triaxial galaxy using N-body simulations. Starting from a single initial model, we grow black holes with various final masses Mh and at various rates, ranging from impulsive to adiabatic. In all cases, the galaxy achieves a final shape that is nearly spherical at the center and close to axisymmetric throughout. However, the rate of change of the galaxy's shape depends strongly on the ratio Mh/Mg of black hole mass to galaxy mass. When Mh/Mg < 0.3%, the galaxy evolves in shape on a timescale that exceeds 100 orbital periods, or roughly a galaxy lifetime. When Mh/Mg > 2%, the galaxy becomes axisymmetric in little more than a crossing time. We propose that the rapid evolution toward axisymmetric shapes that occurs when Mh/Mg > 2% provides a negative feedback mechanism which limits the mass of central black holes by cutting off their supply of fuel.Comment: 27 Latex pages, 9 Postscript figures, uses aastex.sty. Accepted for Publication in The Astrophysical Journal, Nov. 26, 199