OBSERVATIONAL TESTS ON THE DYNAMICS OF GLOBULAR CLUSTERS

Galactic globular clusters are nearly gas-free, self-gravitating stellar systems characterized by an apparently simple geometry, with finite size probably determined by tidal truncation. These unique properties make them excellent laboratories for studies of stellar dynamics, and ideal targets for N-body simulations. For a long time, they have been treated as spherically symmetric, nonrotating, isotropic systems. The spherical King (1966) models, constructed to match this physical picture, are usually considered as the correct zeroth-order dynamical reference model, and are sometimes successful in representing the observed characteristics of these systems. In reality, this simple physical picture suffers from a number of limitations, which become more evident now that much improved observations have become available. In particular, deviations from sphericity have been observed and should be explained. Three physical ingredients are expected to affect the observed morphologies of stellar systems: internal rotation, pressure anisotropy, and external tides. A proper identification of the physical ingredients that shape the internal dynamics of globular clusters will lead to draw conclusions on their origin, and on the origin of their host systems. In particular, a detailed characterization of the role played by internal rotation and pressure anisotropy in present-day globular clusters would be a crucial element to discriminate among different formation scenarios for this class of stellar systems; indeed, the main goal of this Thesis is to clarify the role of these two important dynamical factors. For the purpose of giving a detailed and more realistic description of globular clusters, dynamical studies are an important counterpart to the stellar populations analyses often carried out for these systems. Dynamical studies are meaningful only when both photometric and kinematic data are taken into account, but unfortunately for globular clusters the application of dynamical models is frequently carried out only in relation to the available photometric profiles. This Thesis addresses this issue, and strongly supports the view that accurate kinematic data are crucial to provide a satisfactory description of these systems.