The symmetric multistep methods developed by Quinlan and Tremaine (1990) are shown to suffer from resonances and instabilities at special stepsizes when used to integrate nonlinear equations. This property of symmetric multistep methods was missed in previous studies that considered only the linear stability of the methods. The resonances and instabilities are worse for high-order methods than for low-order methods, and the number of bad stepsizes increases with the number frequencies present in the solution. Symmetric methods are still recommended for some problems, including long-term integrations of planetary orbits, but the high-order methods must be used with caution.

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.

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.

Massive black holes are thought to reside at the centres of many galaxies, where they power quasars and active galactic nuclei. But most galaxies are quiescent, indicating that any central massive black hole present will be starved of fuel and therefore detectable only through its gravitational influence on the motions of the surrounding stars. M32 is a nearby, quiescent elliptical galaxy in which the presence of a black hole has been suspected; however, the limited resolution of the observational data and the restricted classes of models used to interpret this data have made it difficult to rule out alternative explanations, such as models with an anisotropic stellar velocity distribution and no dark mass or models with a central concentration of dark objects (for example, stellar remnants or brown dwarfs). Here we present high-resolution optical HST spectra of M32, which show that the stellar velocities near the centre of this galaxy exceed those inferred from previous ground-based observations. We use a range of general dynamical models to determine a central dark mass concentration of (3.4 +/- 1.6) x 10^6 solar masses, contained within a region only 0.3 pc across. This leaves a massive black hole as the most plausible explanation of the data, thereby strengthening the view that such black holes exist even in quiescent galaxies.

The collapse time for a cluster of equal-mass stars is usually stated to be either 330 central relaxation times ($\trc$) or 12--19 half-mass relaxation times ($\trh$). But the first of these times applies only to the late stages of core collapse, and the second only to low-concentration clusters. To clarify how the time depends on the mass distribution, the Fokker-Planck equation is solved for the evolution of a variety of isotropic cluster models, including King models, models with power-law density cusps of $\rho\sim r^{-\gamma}$, and models with nuclei. High-concentration King models collapse faster than low-concentration models if the time is measured in units of $\trh$, but slower if it is measured in units of $\trc$. Models with cusps evolve faster than King models, but not all of them collapse: those with $0<\gamma<2$ expand because they start with a temperature inversion. Models with nuclei collapse or expand as the nuclei would in isolation if their central relaxation times are short; otherwise their evolution is more complicated. Suggestions are made for how the results can be applied to globular clusters, galaxies, and hypothetical clusters of dark stars in the centers of galaxies.

The stellar ejection rate and the rates of change of the binary semimajor axis and eccentricity are derived from scattering experiments for the restricted three-body problem. They are used to study the evolution of binaries in simple models for galactic nuclei, starting soon after the black holes become bound and continuing until the evolution is dominated by the emission of gravitational radiation, or until the ejected mass is too large for the galaxy to be considered fixed. The eccentricity growth is found to be unimportant unless the binary forms with a large eccentricity. The scattering results are compared with predictions from Chandrasekhar's dynamical-friction formula and with previous work on the capture and scattering of comets by planetary systems. They suggest that a binary with masses $m_1\geq m_2$ should not be considered hard until its orbital velocity exceeds the background velocity dispersion by a factor that scales as $(1+m_1/m_2)^{1/2}$.

We present a method for simulating numerically the effect of the adiabatic growth of black holes on the structure of elliptical galaxies. Using a parallel self--consistent field code, we add black holes to N--body realizations of model distribution functions for spherical galaxies, using a continuous mass--spectrum. The variable particle mass, combined with a simple multiple timestep integration scheme, makes it possible to evolve the models for many dynamical times with $N \sim 10^6-10^8$, allowing high spatial and mass resolution. This paper discusses verification of the code using analytic models for spherical galaxies, comparing our numerical results of the effect of central black holes on the structure of the galaxies with previously published models. The intrinsic and projected properties of the final particle distribution, including higher order moments of the velocity distribution, permit comparison with observed characteristics of real galaxies, and constrain the masses of any central black holes present in those galaxies. Our technique is promising and is easily extended to axisymmetric and triaxial galaxies.

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 differs 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 confirmed and sharpened: the black hole induces cusps in the density ($\rho\sim r^{-3/2}$) and velocity dispersion ($v^2\sim r^{-1}$), 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.