We derive the rates of capture, Ndot, of main sequence turn off stars by the central massive black hole in a sample of galaxies from Magorrian et al. 1998. The disruption rates are smaller than previously believed with Ndot ~ 10^-4 - 10^-7 per galaxy. A correlation between Ndot and black hole mass, M, is exploited to estimate the rate of tidal disruptions in the local universe. Assuming that all or most galaxies have massive black holes in their nuclei, this rate should be dominated by sub-Lstar galaxies. The rate of tidal disruptions could be high enough to be detected in supernova (or similar) monitoring campaigns---we estimate the rate of tidal disruptions to be 0.01 - 0.1 times the supernova rate. We have also estimated the rates of disruption of red giants, which may be significant (Ndot ~> 10^-4 y^-1 per galaxy) for M ~> 10^8 Msun, but are likely to be harder to observe---only of order 10^-4 times the supernova rate in the local universe. In calculating capture rates, we advise caution when applying scaling formulae by other authors, which are not applicable in the physical regime spanned by the galaxies considered here.

We use data from the literature to constrain theoretical models of galaxy formation. We show how to calculate the dimensionless spin parameter $\lambda$ of the halos of disk galaxies and we compare the distribution of $\lambda$ with that observed in cosmological $N$-body simulations. The agreement is excellent, which provides strong support for the hierarchical picture of galaxy formation. Assuming only that the radial surface density distribution of disks is exponential, we estimate crudely the maximum-disk mass-to-light ratio in the $I$-band and obtain $<M/L> \la 3.56h$, for a Hubble constant of $100h~\kms \mpc^{-1}$. We discuss this result and its limitations in relation to other independent determinations of $M/L$. We also define a dimensionless form of the Tully-Fisher relation, and use it to derive a value of the baryon fraction in disk galaxies. For galaxies with circular velocity $\vmax>100\kms$, the median value is $md = 0.086 (M/L/3.56h)$. Assuming that the gas fraction in galactic halos is at most as large as that in clusters, we also conclude that $<M/L> \la 2.48h^{-1/2}$.

Using high resolution cosmological N-body simulations, we investigate the survival of dark matter satellites falling into larger haloes. Satellites preserve their identity for some time after merging. We compute their loss of mass, energy and angular momentum as dynamical friction, tidal forces and collisions with other satellites dissolve them. We also analyse the evolution of their internal structure. Satellites with less than a few per cent the mass of the main halo may survive for several billion years, whereas larger satellites rapidly sink into the center of the main halo potential well and lose their identity. Penetrating encounters between satellites are frequent and may lead to significant mass loss and disruption. Only a minor fraction of cluster mass (10 per cent on average) is bound to substructure at most redshifts of interest. We discuss the application of these results to the survival and extent of dark matter haloes associated with cluster galaxies, and to interactions between galaxies in clusters. We find that 35-40 per cent of galaxy dark matter haloes are disrupted by the present time. The fraction of satellites undergoing close encounters is similar to the fraction of interacting or merging galaxies in clusters at moderate redshift.

We examine the global stability of an exponential stellar disk embedded in a dark matter halo and constrain the mass-to-light ratio of observed disks in the I-band, Y_I. Assuming only that the radial surface density distribution of disks is exponential, we derive an analytic upper limit: Y_I <~ 2.4h, for a Hubble constant of 100h~\kms\mpc^{-1}. Using N-body simulations we derive a stability criterion significantly different from that of previous authors. Using this criterion we argue that, almost independent of the concentration of the halo, Y_I <~ 1.9h. We discuss this result in relation to other independent determinations of Y_I, its limitations and its implications for theories of disk formation and barred galaxies.

Galactic nuclei are now generally thought to have density cusps in their centres, and the evidence is mounting that as a consequence they are unlikely to be triaxial. Self-consistent stellar dynamical models of non-axisymmetric cusps would be an interesting counter-argument to this conclusion. We consider 2-d analogues of triaxial cusps: a sequence of non-axisymmetric, cuspy discs first described by Sridhar & Touma (1997). Scale-free models with potential {}$\Phi\propto r^\alpha$ are examined in detail. It is shown analytically for $0<\alpha \la 0.43$ that self-consistent models with positive phase-space density do not exist. Numerical solutions of the combined Vlasov and Poisson equations suggest that the whole sequence of models with $0<\alpha<1$ are also unphysical. Along with existing work on cusps, we conclude on purely theoretical grounds that galactic nuclei are not expected to be triaxial.

We argue that a universal density profile for dark matter halos arises as a natural consequence of hierarchical structure formation. It is a fixed point in the process of repeated mergers. We present analytic and numerical arguments for the emergence of a particular form of the profile. At small radii, the density should vary as $r^{-\alpha}$, with $\alpha$ determined by the way the characteristic density of halos scales with their mass. If small halos are dense, then $\alpha$ is large. The mass-density relation can be related to the power spectrum of initial fluctuations, $P(k)$, through `formation time' arguments. Early structure formation leads to steep cusps. For $P(k)\sim k^n$ we find $\alpha\simeq 3(3+n)/(5+n)$. The universal profile is generated by tidal stripping of small halos as they merge with larger objects.

We describe an algorithm for constructing N-body realisations of equilibrium stellar systems. The algorithm complements existing orbit-based modelling techniques using linear programming or other optimization algorithms. The equilibria are constructed by integrating an N-body system while slowly adjusting the masses of the particles until the time-averaged density field and other observables converge to a prescribed value. The procedure can be arranged to maximise a linear combination of the entropy of the system and the $\chi^2$ statistic for the observables. The equilibria so produced may be useful as initial conditions for N-body simulations or for modelling observations of individual galaxies.

Galaxies exhibit a variety of non-axisymmetric structure (bars, spiral structure, lopsided structure, etc.). These suggest the following general problem: what are the possible stationary configurations of a two-dimensional self-gravitating fluid other than an axisymmetric razor-thin disc? We address a modest component of this problem: we seek non-axisymmetric razor-thin discs of two-dimensional barotropic fluid that are stationary in an inertial frame. We distinguish between `razor-thin' and `two-dimensional,' applying the latter term to the equation of state. Furthermore we assume that our systems are scale-free, which reduces the partial differential equations describing the system to ordinary differential equations. We also allow for the presence of an axisymmetric background potential. This simple and highly idealized problem already exhibits a rich variety of solutions, the richest being for m=1 symmetry.

We show that outside-in disc instabilities, which can produce rapid-\null rise FU Orionis outbursts in T~Tauri systems, are a natural consequence of the existence of protoplanetary protostellar companions. The \pcomp{} would be formed through gravitational instability on a dynamical timescale ($\sim 10^5\y$), and provided it was formed at less than a few \AU{}, it would be swept in to $\sim15\Rsun$ in less than $10^5\y$. A companion of mass $10^{-2}$\Msun at this radius would act as a flood gate which stores material upstream and periodically releases it following the ignition of the thermal ionisation instability.

We show how to develop an expansion of nearly oblate systems in terms of a set of potential-density pairs. A harmonic (multipole) structure is imposed on the potential set at infinity, and the density can be made everywhere regular. We concentrate on a set whose zeroth order functions describe the perfect oblate spheroid of de Zeeuw (1985). This set is not bi-orthogonal, but it can be shown to be complete in a weak sense. Poisson's equation can be solved approximately by truncating the expansion of the potential in such a set. A simple example of a potential which is not one of the basis functions is expanded using the symmetric members of the basis set up to fourth order. The basis functions up to first order are reconstructed approximately using 10,000 particles to show that this set could be used as part of an $N$-body code.

We demonstrate, using a simple analytic model, that the presence of a massive satellite can globally modify the structure and emission properties of an accretion disc to which it is tidally coupled. We show, using two levels of numerical approximation, that the analytic model gives reasonable results. The results are applicable to two astrophysical situations. In the case of an active galactic nucleus, we consider the case of a $\sim 10^3\Msun$ compact companion to the central black-hole and show that it could modulate the emitted spectrum on a timescale of $\sim10^5$ years. In the case of a T Tauri accretion disc, a satellite such as a sub-dwarf or giant planet could modify the disc spectral energy distribution over a substantial fraction of the T Tauri star lifetime.

We describe a technique for solving the combined collisionless Boltzmann and Poisson equations in a discretised, or lattice, phase space. The time and the positions and velocities of `particles' take on integer values, and the forces are rounded to the nearest integer. The equations of motion are symplectic. In the limit of high resolution, the lattice equations become the usual integro-differential equations of stellar dynamics. The technique complements other tools for solving those equations approximately, such as $N$-body simulation, or techniques based on phase-space grids. Equilibria are found in a variety of shapes and sizes. They are true equilibria in the sense that they do not evolve with time, even slowly, unlike existing $N$-body approximations to stellar systems, which are subject to two-body relaxation. They can also be `tailor-made' in the sense that the mass distribution is constrained to be close to some pre-specified function. Their principal limitation is the amount of memory required to store the lattice, which in practice restricts the technique to modeling systems with a high degree of symmetry. We also develop a method for analysing the linear stability of collisionless systems, based on lattice equilibria as an unperturbed model.

(To appear in the Astrophysical Journal) Holdom and Malaney (1994) have suggested a mechanism for gamma-ray bursts which requires that stars be captured by a neutrino ball. Neutrino balls would be, for the most part, denser than main sequence stars, but their density would decrease as their mass increased. We show that small neutrino balls would subject stars to tidal forces sufficient to disrupt them. We thus argue that if neutrino balls existed at the centres of galaxies, only the largest would be able to act as a source of gamma-ray bursts. Such neutrino balls would have a mass of order $10^7\Msun$. Tidal capture of stars by a neutrino ball would not be important, but dynamical friction against the neutrinos or star-disc interactions could both be important capture mechanisms. We find that a gamma-ray burst would occur in a galaxy containing such a neutrino ball roughly every $10^2\y$, and the fraction of all galaxies contributing to the gamma-ray burst flux would be $\sim 10^{-4}$, assuming that this was the mechanism of all gamma-ray bursts. These numbers have implications for neutrino ball models of active galaxies, assuming that all gamma-ray bursts and all AGN come from neutrino balls. Either a small fraction $\sim 10^{-2}$ of the lifetime of such an object could be spent as an AGN, or that the probability of a neutrino ball becoming an AGN would be $10^{-2}$. It is not possible to rule out the possibility that neutrino balls might exist at the centres of galaxies through direct ground-based observation of stellar kinematics.

A test mass, $M$, moving through an ambient medium of light particles with lower average kinetic energy than itself suffers a deceleration caused by its scattering of the light particles. The phenomenon is usually referred to as dynamical friction. The velocity, $\v$, of the test mass decays on a timescale independent of $\v$ in the non-relativistic case. We derive expressions for dynamical friction in the case that the test mass and the light particles are relativistic, and that the scattering is weak (with impact parameter, $b\gg M$). In the case that the light particles are ultra-relativistic, and isotropic in the frame in which $M$ moves with velocity $v$, we find an explicit expression for the dynamical friction. The well known factor of 2 correcting the Newtonian scattering of photons to give the Einstein angle, $4M/b$, has the largest effect on the resulting friction, which is modified by a factor of roughly $16 / 3\gamma_v$ over the simple non-relativistic case. In the non-relativistic case, the largest contribution to the friction comes from light particles moving slower than $v$. We find that this is not the case for ultra-relativistic scattering, essentially because the scattering angle is independent of $\v$. Some astrophysical implications are discussed. (Accepted for publication in Monthly Notices.)